Polymer Latices

Polymer Latices
JOIN US ON THE INTERNET VIA WWW, GOPHER, FTP OR EMAIL:

WWW: http://www.thomson.com
GOPHER: gopher.thomson.com A service of IWP®
FTP: ftp.thomson.com
EMAIL: [email protected]
Polymer Latices
Science and technology
Second edition
Volume 2: Types of latices
D. C. Blackley
Formerly Reader in Polymer Science and Technology, Polytechnic
0/ North London
lal11 SPRINGER-SCIENCE+BUSINESS MEDIA, B.V

First edition 1966
Second edition 1997
© 1966, 1997 Springer Science+Business Media Dordrecht
Originally published by Chapman & Hall in 1997
Softcover reprint ofthe hardcover 2nd edition 1997
Typeset in 10/12 Times by AFS Image Setters Ltd, Glasgow
ISBN 978-94-010-6479-8 ISBN 978-94-011-5866-4 (eBook)
DOI 10.1007/978-94-011-5866-4
Apart rrom any fair dealing for the purpose of research or private study, or
criticism or review, as permitted under the UK Copyright Designs and
Patents Act, 1988, this publication may not be reproduced, stored, or
transmitted, in any form or by any means, without the prior permission in
writing of the publishers, or in the case of reprographic reproduction only
in accordance with the terms of the licences issued by the Copyright
Licensing Agency in the UK, or in accordance with the terms of Iicences
issued by the appropriate Reproduction Rights Organisations outside the
UK. Enquiries conceming reproduction outside the terms stated here should
be sent to the publishers at the London address printed on this page.
The publisher makes no representation, express or implied, with regard to
the accuracy of the information contained in this book and cannot accept
any legal responsibility or Iiability for any errors or omissions that may be
made.
A catalogue record for this book is available from the 8ritish Library
Library of Congress Catalog Card Number: 96-70800
00 Printed on permanent acid-free text paper, manufactured in accordance

with ANSI/NISO Z39.48-l992 and ANSI/NiSO Z39.48-l984 (permanence
of Paper).
Contents
Preface xi
9. Natural latices 1
9.1 Introduction 1
9.2 Cultivation of natural rubber 3
9.2.1 The principal rubber tree 3
9.2.2 The Hevea brasiliensis plantations 7
9.2.3 Propagation of Hevea brasiliensis 11
9.2.4 Tapping 18
9.2.5 Improvement of latex yield 22
9.3 Preservation of natural rubber latex 24
9.3.1 Preliminary considerations 24
9.3.2 The ideal preservative for natural rubber latex 26
9.3.3 Ammonia as a preservative 27
9.3.4 Low-ammonia preservation systems 35
9.3.5 Other preservatives for natural rubber latex 40
9.4 Concentration of natural rubber latex 42
9.4.1 Preliminary considerations 42
9.4.2 Concentration by evaporation 43
9.4.3 Concentration by creaming 44
9.4.4 Concentration by centrifugation 55
9.4.5 Concentration by electrodecantation 66
9.4.6 Properties of natural rubber latex concentrates 69
9.4.7 Redispersible natural rubber from natural rubber
latex 77
9.5 Constitution of fresh and ammonia-preserved natural
rubber latex 78
9.5.1 Introduction 78
9.5.2 The rubber phase 79
9.5.3 The aqueous phase 96
9.5.4 Lutoids and other particulate phases 101
9.5.5 Changes in the chemical constitution of natural rubber
latex as a consequence of ammoniation and storage 105
VI Contents
9.6 Other naturally-occurring latices 115
9.6.1 Summary of characteristics of other naturally-
occurring latices 115
9.6.2 Sources of other naturally-occurring latices 115
9.6.3 Comments concerning some other naturally-
occurring latices 117
9.7 Biogenesis of polyisoprenes 118
9.7.1 Biological function of the polyisoprenes 118
9.7.2 Outline of polyisoprene biosynthesis 119
9.7.3 More detailed consideration of mechanism of
polyisoprene biosynthesis 121
References 132
10. Synthetic latices: general principles of production 137

10.1 Introduction 137
10.1.1 Definition of a synthetic latex 137
10.1.2 Distinguishing features of synthetic latices relative
to natural latices 139
10.1.3 Literature pertaining to emulsion polymerization,
and to synthetic latices in general 143
10.2 Emulsion polymerization reactions 145
10.2.2 Mechanism and kinetics of emulsion polymerization
reactions 154
10.2.3 Principal components of practical aqueous emulsion
polymerization reaction systems, other than
monomers and water 176
10.2.4 Methods of carrying out emulsion polymerization
in practice 199
10.3 Preparation of functionalized latices by emulsion
copolymerization 238
10.3.2 Advantages and disadvantages which accrue from
the functionalization of synthetic latices 241
10.3.3 Production of carboxylated latices by emulsion
polymerization 243
10.3.4 Distribution of carboxylic-acid groups in
carboxylated latices 248
10.3.5 Effects of carboxylic-acid monomers upon rate of
polymerization and particle nucleation 255
10.4 Agglomeration and concentration of synthetic latices 260
10.4.1 Some problems associated with the production of
high-solids synthetic latices 260
10.4.2 Methods of partially agglomerating synthetic latices 264
Contents Vll
10.4.3 Concentration of synthetic latices 289

10.5 Production of non-aqueous synthetic latices 293
References 296
11. Synthetic latices: individual types 303

11.2 Latices of styrene-butadiene copolymers 305
11.2.1 General considerations 305
11.2.2 Effect of styrene-butadiene ratio in copolymer 309
11.2.3 Production of styrene-butadiene copolymer latices 312
11.2.4 Types of styrene-butadiene copolymer latex 317
11.2.5 Styrene-vinylpyridine-butadiene terpolymer latices 320
11.2.6 Production of masterbatches based upon styrene-
butadiene rubbers for use in dry-rubber technology 321
11.3 Latices of acrylonitrile-butadiene copolymers 323
11.4 Polychloroprene rubber latices 326
11.4.2 Production of polychloroprene rubber latices 328
11.4.3 Hydrolysis of polychloroprene rubbers in latex
form 330
11.4.4 Effects of polymer hydrolysis during storage upon
the properties of polychloroprene rubber latices, and
upon the properties of films derived from the latices 332
11.4.5 Some types of polychloroprene rubber latex
currently available 334
11.5 Latices of vinyl acetate polymers and copolymers 334
11.5.1 Types of vinyl acetate polymer and copolymer
latices currently available 334
11.5.2 Production of vinyl acetate polymer and copolymer
latices 338
11.5.3 Plasticization of vinyl acetate polymers by
copolymerization 342
11.5.4 Hydrolysis of vinyl acetate polymers and copolymers
in latex form 345
11.6 Latices of polyacrylate and polymethacrylate esters:
acrylic latices 353
11.7 Latices of vinyl chloride-vinylidene chloride copolymers 357
11.8 Functionalized synthetic latices 358
11.8.2 Additional chemical reactions to which
functionalized latex polymers are susceptible 359
11.8.3 Alkali-thickenable synthetic latices 368
11.9 Miscellaneous synthetic latices 371
11.9.1 Cationic and amphoteric synthetic latices 371
viii Contents
11.9.2 Synthetic latices which contain microvoid particles 375
11.9.3 Synthetic latices which contain polymers
functionalized such that they strongly absorb
ultraviolet light 377
11.9.4 Synthetic latices which contain unusual polymers 379
11.9.5 Redispersible synthetic latices 383
References 387
12. Artificial latices 391

12.2 Effect of latex particle size upon rate of creaming
or sedimentation 393
12.3 Methods of producing artificia11atices 394
12.3.2 A selection of processes which have been proposed
for the production of artificial latices 397
12.4 Particular types of artificial latex 419
12.4.1 Artificial latices of isobutene-isoprene rubbers 419
12.4.2 Artificial latices of synthetic cis-l,4-polyisoprene
rubbers 420
12.4.3 Artificial latices of cis-l,4-polybutadiene rubbers 430
12.4.4 Artificial latices of reclaimed rubbers 431
12.4.5 Artificial latices of polyurethane polymers 434
12.4.6 Artificial latices of other polymers 435
References 439
13. ChemicaUy-modified latices: 1. Prevulcanized latices 441

13.2 Sulphur-prevulcanization of natural rubber latex 445
13.2.1 History of sulphur-prevulcanization of natural
rubber latex 445
13.2.2 Preparation of sulphur-prevulcanized natural rubber
latex 446
13.2.3 Methods for assessing the degree of vulcanization
of prevulcanized natural rubber latex 450
13.2.4 Effect of sulphur-prevulcanization upon properties
of natural rubber latex, and of films derived from
natural rubber latex 454
13.2.5 Effect of some important procedural variables
upon sulphur-prevulcanization of natural rubber
latex 460
13.2.6 Comparison between sulphur-prevulcanization of
natural rubber latex and sulphur-vulcanization of
dry natural rubber 465
Contents ix
13.2.7 Kinetics and mechanism of sulphur-prevulcanization
of natural rubber latex 469
13.2.8 Properties of films derived from sulphur-
prevulcanized natural rubber latex 483
13.2.9 Structure of films derived from sulphur-
prevulcanized natural rubber latex 485
13.2.10 Industrial grades of sulphur-prevulcanized natural
rubber latex 491
13.3 Prevulcanization of natural rubber latex by reaction with
organic peroxides (including hydroperoxides), and with
hydrogen peroxide 492
13.3.2 Older investigations 494
13.3.3 More recent investigations 495
13.4 Prevulcanization of natural rubber latex by exposure to
high-energy radiation 498
13.4.2 Prevulcanization by irradiation with high-energy
electromagnetic waves 499
13.4.3 Prevulcanization by irradiation with high-energy
electrons 507
13.5 Prevulcanization of other rubber latices 508
13.5.1 Styrene-butadiene rubber and acrylonitrile-
butadiene rubber synthetic latices 508
13.5.2 Isobutene-isoprene (butyl) rubber artificial latices 509
References 510
14. Chemically-modified latices: 2. Types of chemically-modified

latex other than prevulcanized latices 512
14.2 Epoxidation of natural rubber in latex form 513
14.3 Graft-copolymerization reactions in latices 518
14.3.1 Outline of graft-copolymerization reactions 518
14.3.2 Graft-copolymerization reactions in natural rubber
latex 523
14.3.3 Graft-copolymerization reactions in other
latices 535
14.4 Cis-trans isomerization of natural rubber in latex form 537
14.4.2 Experimental details 539
14.5 Cyclization of natural rubber in latex form 543
14.5.2 Preparation of cyclized natural rubber from
x Contents
14.5.3 Mechanism and kinetics of cyclization of natural
rubber in latex, and structure of the product 544
14.6 Halogenation of olefinicaIly-unsaturated polymers in latex
form 548
14.7 Hydrohalogenation of olefinicaIly-unsaturated polymers in
latex form 553
14.8 Halogenoalkylation of natural rubber in latex form 555
14.9 Depolymerization and oxidation of natural rubber in
latex form 558
14.10 Latices containing interpenetrating polymer networks 561
14.11 Modifications involving only the surface of the latex
particles 563
14.11.1 Reversal of charge of latex particles 563
14.11.2 Surface dyeing of latex particles 566
14.11.3 Surface modifications of latex particles for
immunological applications 567
References 573
Index 577
Preface
This book is effectively the second edition of my book High Polymer

Latices which was first published as two volumes in 1966 by Maclaren
& Sons. At the time of its publication, no other book had appeared
which attempted to cover the same ground. Rather surprisingly perhaps,
it appears that no other similar book has appeared in the intervening
years. Since its publication, High Polymer Latices has become accepted
as a comprehensive survey of the science and technology of polymer
latices. Although still providing a reasonable account of the funda-
mentals of the subject, both theoretical and practical, it is now
somewhat out of date, and certainly out of print. It is hoped that this
second edition will be similarly accepted as an updated survey covering
the same ground. Of course, the depth of treatment varies throughout
the book: inevitably, some areas are treated in more detail than others,
and the treatment of some areas is more up-to-date than is that of
others. Of course, some readers will wish that there was greater
emphasis on some aspects and less on others. However, it is hoped that
the overall balance will be found to be broadly satisfactory and still
comprehensive within the bounds of a single book written by a single
author. In any event, those who do not find this revision satisfactory
are free to write their own books on the subject!
In preparing this second edition, the text has been completely re-
written and expanded to take account of many of the developments in
the subject which have occurred since the mid-l 960s. Some errors have
also been corrected. As in much of polymer science and technology,
there has been a proliferation of publications in the area of latex science
and technology over the years since the mid-l 960s. Many of these
publications are to be found in appropriate scientific and technical
journals; many others are confined to the patent literature still others
have appeared in the form of trade publications. Deepening understand-
ing of the factors which affect the colloid stability of lyophobic sols has
assisted greatly in interpreting the behaviour of polymer latices under
various conditions. It is to be hoped that the present work has benefited
Xll Preface
from study and reflection upon much of the literature relating to latex
science and technology, and to colloid science, which has appeared since
High Polymer Latices was published. In addition, it is hoped that this
work has also benefited from experience gained by the author in
supervising numerous research students who have worked in this area,
and also from experience gained from continuing involvement in
industrial consultancy in this field.
For convenience of handling, the book has now been divided into three
volumes instead of two. The first deals with certain fundamental matters
such as the fundamental characteristics of latices, the colloid stability of
latices, colloidal destabilization, flow behaviour and handling. The matter
of the determination of various latex properties is also dealt with in this
volume. The second volume deals with the latices themselves. The subject
matter for this volume falls naturally under the headings of natural latices
(of which natural rubber latex is pre-eminent), synthetic latices, artificial
latices and chemically-modified latices. The treatment of synthetic latices
has been divided between two chapters: the first deals with the general
principles of production, the second with the individual types of synthetic
latex. Of the various chemically-modified latices, prevulcanized rubber
latices are by far the most important industrially; they are therefore
considered in a separate chapter, the remaining types being the subject of
a further chapter. The third volume deals with the various applications of
latices, and also with the related matters of compounding ingredients and
the preparation of solutions, dispersions and emulsions for addition to
latices. Each of the principal applications of latices is given a separate
chapter. The final chapter deals with diverse miscellaneous applications
which are not covered elsewhere.
It is, of course, impossible and inappropriate to give comprehensive
references to the literature pertaining to a subject such as latex science and
technology. However, a selection of references is given for most chapters.
It is hoped that these will provide an entree into the wider scientific and
technical literature for those readers who require more information than
can be given here. To assist with the location of journal references, full
journal names are given. Wherever possible, reference to patents in the
text is by way of the names of the inventors, because they are deemed
to have created the intellectual content of the patent. The name of any
assignee company is then given as part of the reference at the end of
the chapter. The dates given for patents are usually those for publication
or acceptance, not those for application. An exception is a few British
patents which were granted early in this century, at a time when the
practice was to number patents in the year of application, starting afresh
each year; for ease of locating these patents, it is necessary to know the
date of application as well as that of acceptance.
The readership of this book is assumed to have a level of knowledge
Preface xiii
approximating to that of a first degree in physical sciences. The
emphasis throughout this book is upon rational interpretation of the
properties and processing of polymer latices in terms of fundamental
physical and chemical concepts, insofar as this is possible in our present
state of knowledge. To this end, particular significance is attached to the
fundamental matters dealt with in Volume l. It is hoped that the
approach which has been adopted throughout the book as a whole will
encourage scientists and technologists working in this field to attempt a
similar interpretative approach to their labours.
It will be evident that the subject matter of this book is presented in
relatively few chapters, some of which are very long. This has been a
deliberate policy; it is believed that the overall structure of the subject
can be the better made clear in this way, rather than by increasing the
number of chapters. Detailed structures have been provided for each of
the chapters, with decimalized numbering of the sub-sections; it is hoped
that these structures will be found to be especially useful for the lengthy
chapters. Occasional repetition of statements and principles may be
noticed. This again has been deliberate policy. As well as emphasizing
important concepts, the intention has been to assist as far as possible
those readers whose interest is confined to selected parts of the book
rather than to the whole. It is suspected that most readers will come
into this category; this is certainly not the type of book to read from
cover to cover! For the same reason, extensive cross-referencing has
been provided within and between the various chapters.
Most of the original diagrams have been re-drawn for this edition;
many more diagrams, all specially drawn, have been added. In contrast
to the first edition, experimental points have generally been inserted in
the graphs. Although data points inevitably cause graphs to become
more cluttered than if they are omitted, this disadvantage is offset by
the opportunity offered for readers to judge for themselves the strength
of the experimental evidence upon which conclusions derived therefrom
are based. In those cases where graphs are shown without supporting
points, the reason is either that points were unavailable or that they are
inappropriate.
SI units have been used throughout. Although this system has several
obvious deficiencies, it is best we have, it is at least coherent, and it is
unlikely to be superseded in the foreseeable future. Where results are
quoted from sources in which results were expressed in other than SI
units, those results have been converted to SI units. As far as possible,
numerical information in tables and graphs is presented throughout in
the form of the ratio 'quantity/unit'. By this means, it is possible to
clarify unequivocally the significance of numbers which have been
tabulated or appended to the axes of graphs. It has been the experience
of the author, and of many others, that this is by far the clearest way
xiv Preface
of indicating the precise significance of numerical entries, especially where

numbers are expressed in decimal-exponent form. One particular SI unit
which has caused some confusion in the past amongst technologists is
that for density. The basic SI unit is kgm- 3 • However, this gives numbers
which are WOO-times larger than the numbers for density expressed in the
unit gcm- 3 to which many had become used. The preferred unit in this
book in Mgm- 3 , with mgmm- 3 as an alternative; the numerical values
are then identical to those expressed in g cm -3.
There are various matters of terminology which call for comment.
Terms such as latex compound and latex compounding are used where
strictly latex mixture and latex mixing should be used. The excuse is
that the former terms have become sanctified by common usage, and are
unlikely to be superseded by the latter. Mass has been used in
preference to weight where this is the correct term to use, e.g. in
formulations for latex compounds.
Molecular mass has been preferred to the older term molecular weight
or to the modem alternative relative molecular mass. There then arises
the matter of the unit in which molecular mass is to be expressed.
Whereas relative molecular mass is a pure number, molecular mass has
units of mass per molecular entity. Of the various alternatives for the
latter unit, the dalton is preferred over the atomic mass unit or the unit
gmol- I . Although in common usage now, and numerically equal to
molecular mass in daltons, the latter has been rejected because it
focusses attention upon the mole of entities rather than upon the
individual molecular entity. However, I have refrained from using the
abbreviation '0' for dalton, and from attaching SI prefixes when
denoting large values of molecular mass. In this connection, attention is
also drawn to the fact that the basic unit of molar mass is kgmol- I ,
and not gmol- I . Thus in converting molar mass in kgmol- I to
molecular mass in daltons (or, indeed in gmol- I ), it is necessary to
multiply by a factor of 1000.
As is conventional in rubber technology, the term modulus is used to
denote the stress at a given strain for rubber vulcanizates which are
undergoing deformation; unlike engineering practice, it does not denote
the ratio of stress to strain. Furthermore, the moduli of rubber
vulcanizates are conventionally calculated on the basis of the initial
(unstrained) cross-sectional area, and not upon the cross-sectional area
in the deformed state. Extension at break has been preferred to the more
usual equivalent term elongation at break. Moduli are usually specified
for some particular extension, rather than for the equivalent elongation;
there seems to be no good reason for using the one term in the context
of moduli and the other when referring to deformation at rupture.
Endeavours have been made to achieve consistency throughout in
matters such as these, but no doubt some inconsistencies will be found.
Preface xv
Some difficulty has been experienced with the use of the two words
coacervant and coagulant, and their cognates. After careful deliberation,
the decision has been taken to follow the usage adopted in the· first
edition: in general, the word coacervant is used to denote any additive
which causes gross and obvious colloidal destabilization in a polymer
latex, whereas the word coagulant is restricted to additives which bring
about the particular form of gross colloidal destabilization known as
coagulation. The distinction is certainly one which needs to be made, and
this is one way of making it. However, rather illogically, the expression
coagulant dipping has been retained because it is so widely used in certain
sections of the rubber latex industry; to use a more precise term such as
coacervant dipping or gelation dipping would be to invite confusion.
The nomenclature of organic compounds in a book such as this
inevitably presents problems. Trade names have generally been avoided;
they usually give little information as to chemical structure, and in any
case may be somewhat ephemeral. The policy has been generally to give
familiar names by which the compounds are commonly known in
industry, but also to give at an appropriate point the systematic name,
at least of the major component. The chemical structure can then be
derived unequivocally, even if no structure is given in the text.
Furthermore, it is frequently the case that industrial grades of chemicals
are mixtures comprising a principal component together with minor
amounts of other substances. As an example, although it is helpful to
know that oleic acid and the oleates are principally cis-n-9-octadecenoic
acid and cis-n-9-octadecenoates respectively, the latter names are too
clumsy for general use, and in any case minor amounts of other
substances are present in industrial grades of these substances.
Another matter which inevitably presents problems in a book such as
this is that of mathematical notation. An attempt has been made to
simplify this as far as possible. Unfortunately, some symbols have had
to be used with different meanings in different parts of the book,
because of the multiplicity of quantities which it has been necessary to
represent in the book as a whole. However, every effort has been made
to ensure that no confusion can arise within any given discussion.
Occasionally it has been thought helpful to introduce as mathematical
symbols distinctive letters from the Cyrillic alphabet, as well as the more
usual letters from the Greek alphabet; it is hoped that this occasional
practice will not be found to be distracting.
Common abbreviations for latex additives are noted but not generally
used. The following abbreviations are used where appropriate:
pphr parts by mass per 100 parts by mass of rubber
pphls parts by mass per 100 parts by mass of latex solids
pphp parts by mass per 100 parts by mass of polymer
pphm parts by mass per 100 parts by mass of monomer.
xvi Preface
The expressions parts per million and parts per hundred million and their
abbreviations ppm and pphm, are not used. They are replaced by the
units mg kg-I and ~g kg-I respectively, the latter differing form parts per
hundred million by a factor of 10. Abbreviations such as m/m, m/v and
mol/mol are used frequently to indicate clearly the nature of quantities
such as concentrations, percentages and composition ratios.
The values of the common physical constants used in making the
calculations reported in this book have generally been taken from the
15th edition of Tables of Physical and Chemical Constants, originally
compiled by G. W. C. Kaye and T. H. Laby, published by Longman
Scientific & Technical, Harlow, in 1986. Data for polymers have
generally been taken from the 3rd edition of Polymer Handbook, edited
by J. Brandrup and E. H. Immergut, published by John Wiley, New
York, in 1989.
There is one further matter to which any author writing on a
technological subject in these days must draw attention. This concerns
the health hazards which may accompany the use of some of the
substances which he mentions. We are becoming increasingly aware of
these hazards. Seldom, if ever, does the reputation of a substance in this
respect improve as a consequence of increased knowledge and wider
experience. Usually greater knowledge informs us of the existence of a
hazard where none was thought to exist, or of a greater hazard than
was believed formerly. For various reasons, not the least of which is the
rate at which knowledge and opinion are changing in this field, it has
been considered inappropriate to deal with this aspect of the subject in
this book, except to make brief reference to the protein-allergy problem
associated with the use of natural rubber latex, and to the problem of
dermititic sensitization which can occur when some of the compounding
ingredients used with rubber latices come into contact with human skin.
Rather, it has been considered advisable to give a general warning that
many, if not all, the substances mentioned here are potentially liable to
become suspect in this respect in the light of increasing knowledge.
Those who propose to make use of substances mentioned in this book
are therefore strongly advised first to seek up-to-date authoritative
opinion on possible health hazards. For those who propose to direct
others to use these substances, this precaution is obligatory.
It remains to acknowledge with thanks the benefit I have received
from those with whom I have worked professionally in the area of
latex science and technology over the years. In particular, I thank my
former colleagues of the Polytechnic (now University) of North London,
Dr. W. F. H. Burgar and Mr. T. T. Healy, and also Mr. T. D. Pendle and
Mr. A. D. T. Gorton, formerly of the Malaysian Rubber Producers'
Research Association (now the Tun Abdul Razak Research Centre), for
many helpful discussions. I have also received much help and stimulation
Preface xvii
from colleagues from industry and other academic institutions with whom
I have interacted over the years; in particular, I mention Prof. M. Porter
of London International Group. At the same time as acknowledging help
from many colleagues and professional acquaintances, I alone must accept
full responsibility for any errors which this book may contain. It is also a
great pleasure to acknowledge the unfailing assistance of the library staff
of the University of North London in obtaining much of the literature
which has been consulted during the course of preparing this second
edition. Finally, I acknowledge with deep gratitude the continuing
patience and support of my wife during the period I have been engaged
in completing this work.
D. C. Blackley
February 1997
9
Natural latices
9.1 INTRODUCTION
The subject of this chapter is those lyophobic colloidal dispersions of polymers
which occur naturally as the products of the metabolism of certain types of
plant. Although all such latices are in principle of great intrinsic interest to
the botanist and the biochemist, industrial interest has been overwhelmingly
in one type of naturally-occurring latex, namely, natural rubber latex.
Consequently, most of this chapter is concerned with this latex, although
brief mention is made of some of the other types. Natural rubber latex is
obtained from particular types of tree which grow in tropical regions. The
principal rubber-producing trees in cultivation at the present time (and for
the foreseeable future) are members of the species Hevea brasiliensis, of the
family Euphorbiaceae. It is from this species that the world's supply of natural
rubber latex (and hence of natural rubber) is obtained almost exclusively.
One ofthe most important developments in the synthetic-rubber-producing
industry since the end of World War II has been the development in the
mid-1950s of a general-purpose synthetic rubber having almost identical
chemical composition to the polymer contained in natural rubber latex. This
material has been called 'synthetic natural rubber'. Unlike most of the
general-purpose synthetic rubbers produced up until that time, this rubber
is not obtained from the corresponding monomer by free-radical emulsion
polymerization, but by solution polymerization in a hydrophobic medium
using ionic initiators. There is thus no question of this polymerization reaction
being used to produce a 'synthetic natural rubber latex'. However, artificial
latices of 'synthetic natural rubbers' have been produced by dispersing the
solid polymer in an aqueous medium, as described in Chapter 12 (see
especially section 12.4.2). Notwithstanding hopes in the late 1960s, all attempts
to produce synthetic analogues of natural rubber latex by aqueous emulsion
polymerization have failed so far; industrial interest in developing processes
for the production of a latex of this type has now waned. The monomer
which is used for the production of 'synthetic natural rubber' is obtained
from petroleum. It was widely forecast that the development of this polymer
D. C. Blackley, Polymer Latices

© Springer Science+Business Media Dordrecht 1997
2 Natural latices
would herald the demise of the natural-rubber-producing industry; 20 years

was typical of the predicted time-scale over which demise would occur. The
non-fulfilment of this prophecy is attributable to various factors, some of
which are economic and some environmental. These factors include:
1. the sharp rises in world petroleum prices, and in the prices of all products
derived from petroleum, which have occurred since the early 1970s;
2. the ability of the Hevea brasiliensis tree to produce rubber from renewable
starting materials;
3. the ability of the Hevea brasiliensis tree to produce rubber by a process
which does not pollute the environment, and which produces oxygen as
a by-product.
However, although the production of rubber in the tree is pollution-free,
the subsequent processing of the product obtained from the tree may, of
course, cause pollution.
Some interest has been shown in the production of natural rubber latex
outside the environment of the Hevea brasiliensis tree, by means of a
biotechnological process which uses the same natural precursors and the
same biochemical pathway as is used by the tree. The idea has been to extract
the rubber-producing enzyme system from the tree, and supply that system
with the necessary nutrients. In principle, this type of process offers several
advantages relative to the production of natural rubber latex within the tree.
Major advantages include greater control of the production process, with
consequent reduction in the variability of the product, together with great
improvement in the efficiency of collection of the latex. Other advantages
include the possibility of varying and controlling the molecular characteristics
of the rubber which is formed. Whether or not a biotechnological process
for the production of natural rubber latex is workable is not clear at present.
What does seem to be clear is that a process of this type would not be feasible
economically in the present market situation. The Hevea brasiliensis tree is
a remarkably efficient producer of natural rubber, although the process of
obtaining the rubber from the tree is tedious and labour-intensive (see section
9.2.4 below). It therefore seems unlikely that any biotechnological process
will compete successfully with the tree in the foreseeable future.
The structure of this chapter is as follows: a brief description is first given
of the cultivation of the rubber tree, together with the methods used to obtain
the latex from the tree. Of the latter, the most usual is known as tapping.
The next main topic to be considered is that of preservation of the latex.
Preservation is essential if the latex is to be kept for any length of time. Being
derived from a living organism, fresh natural rubber latex is subject to
degradation by microorganisms. This has to be prevented if the product is
to remain useful. Consideration is then given to the concentration of natural
rubber latex. This matter is important because the proportion of rubber in
the latex as obtained from the tree is insufficient for economic transportation
Cultivation of natural rubber 3
and for application in the large-scale manufacturing processes in which it is

commonly used. A summary is then given of the constitution of fresh and
preserved natural rubber latex. The chapter concludes with brief reference
to other types of naturally-occurring latices, and to the biochemical pathway
by which the hydrocarbon polymer is formed in natural latices containing
polyisoprenes.
There is an extensive literature concerning natural rubber latex, both as
it is obtained fresh from the Hevea brasiliensis tree and in the preserved and
concentrated state in which it is usually encountered industrially. A useful
and authoritative summary of the structure, composition and biochemistry
of the latex is available in a chapter by Archer et al. [1]. Colloidal properties
are summarized in a chapter by Cockbain and Philpott [2]. The composition
and properties of natural rubber latex concentrates are reviewed in chapters
by Gazely, Gorton and Pendle [3] and by Calvert [4]. Books by Polhamus
[5], by Bobilioft' [6] and by Dijkman [7] deal in detail with various botanical
aspects of the Hevea brasiliensis tree. A recent publication edited by D'Auzac,
Jacob and Chrestin [8] provides detailed information concerning the
biochemistry and physiology of the latex obtained from this tree. A manual
edited by Edgar [9] deals with various aspects of the rubber-plantation
industry. A booklet published by the Rubber Research Institute of Malaya
[10] deals specifically with the vegetative propagation of Hevea brasiliensis
by bud-grafting.
9.2 CULTIVATION OF NATURAL RUBBER
9.2.1 The principal rubber tree
9.2.1.1 General description

The common rubber tree is said to be a hybrid strain which is rather hardier
and easier to propagate than are the pure strains. Many well-defined varieties
of the tree, known as clones, have been, and continue to be, developed with
specific improvements as the objective. Hevea brasiliensis is a large and
ungainly tree which can grow to a height of over 30m. Typically, mature
specimens which have not been tapped have heights ca. 20-25 m and trunk
circumference ca. 3-4 m. Tapping tends to restrict the growth somewhat,
heights ca. IS-20m being typical. The foliage of the tree is rather sparse,
and dull green in colour. The polymer obtained from this hybrid tree has
been known as para rubber.
The salient features of the leaves, flowers and fruit of the Hevea brasiliensis
tree are shown in Fig. 9.1. The leaves are trifoliate, i.e., three-lobed. Each
lobe is long and narrow at both ends. The leaves are shed once a year; this
takes place while the new leaves are forming, so that the tree is never bare.
4 Natural latices
leaves Flowers
Fruit Transverse section of fruit Seed.
Figure 9.1 Salient features of the leaves, flowers and fruit of the Hevea brasiliensis tree.
The time of the year at which leaf-fall occurs varies somewhat with the
locality. The flowers are borne in inflorescences towards the ends of the leafy
twigs where the new season's growth is starting. Each inflorescence is much
branched, and bears both male and female flowers. The female flowers are
smaller than the male, and occur especially on the lesser shoots. The flowers
are small, yellowish-green, and have a pleasant scent. Hevea brasiliensis is
not self-fertile. Cross-pollination is therefore required if fruit is to form. Even
then, only a small fraction of the female flowers set seed; it has been suggested
that this is in part a consequence of the scarcity of suitable pollinating insects.
The fruit comprises a large capsule which normally contains three seeds in
separate cavities. The longitudinal axis of the fruit is normally one of three-fold
symmetry. The seeds are quite large (ca. 2.5 em in length), oval in shape, and
have smooth mottled brown coats. When ripe, the seed capsule dehisces, i.e.,
explodes violently under the influence of sunlight, thereby scattering the seeds
some considerable distance. The fruit requires ca. 5 months from fertilization
for full development. Of this period, ca. 3 months is required for growth to
full size, and a further 2 months for ripening.
9.2.1.2 More detailed structure of the mature trunk

The principal regions of the mature trunk of the H evea brasiliensis tree are
illustrated schematically in Fig. 9.2. The innermost region is known as the
pith or heartwood. It is dead, being the vestigial remains of what was once
Pith or Heartwood
Wood orSapwood _ _
.---- r-- I<"-
r-
Cambium
Inner cortex or bark
Outer cortex or bark----1
Transverse section Longitudinal section
Figure 9.2 Schematic illustration of principal regions of mature trunk of the Hevea brasiliensis
tree.
living tissue. The bulk of the trunk comprises wood, or sapwood, from which
the tree derives its mechanical strength and rigidity. It also serves as a matrix
in which are embedded various conductive elements. The outermost region,
or cortex, is subdivided into an outer cortex, which is a protective layer
consisting mainly of cork cells, and an inner cortex, which consists mainly
of stone cells. Embedded in the matrix of stone cells are specialized cells
known as latex vessels. These vessels are living cells, the sap of which is the
latex. The thin layer of cells which separate the cortex from the sapwood is
known as the cambium. The cells of the cambial layer are in an active state
of subdivision. Only by their activity can the tree expand in girth. If the
health and vigour of the tree are to be maintained, this layer must not be
damaged during tapping or otherwise.
The principal characteristics of the latex vessels are illustrated in Figs
9.3(a)-(f). The internal diameter is ca. 20 J.1m, and the wall thickness ca.
5-7 J.1m. The manner in which the vessels are formed is illustrated in Fig.
9.3(a). Parenchymatous cells differentiate by gradual dissolution of the walls
of adjacent cells, so as to form continuous vessels. Eventually these walls
disappear altogether to form the latex vessels. The latex vessels wend their
way up the trunk of the tree in a direction which is approximately, but not
exactly, vertical. They make an angle of ca. 4° to the vertical, the direction
to the vertical being such that the system of vessels as a whole forms a
right-handed spiral of very long pitch. Figure 9.3(b) shows a transverse section
of the cortical region of the trunk. The latex vessels form concentric mantles
which surround the cambium. The density of vessels diminishes with
increasing distance from the cambium. It is greatest at ground level and
decreases in both directions, i.e., both up the tree into the branches and down
6 Natural latices
(a) (0)
(c) (tl)
"'-,IT '
R0ti~i;~a(oo R;~&jM$ oj ;utdd% R~$r
pa!1:1C!~ cm~$ w<sll paih:les
{e} (1)
Figure 9.3 Illustrating the principal characteristics of the latex vessels: (a) manner in which
latex vessels are formed; (b) transverse section of cortical region of trunk; (c) longitudinal
tangential section of cortical region of trunk; (d) longitudinal radial section of cortical region
of trunk; (e) three-dimensional view of interior of cambial and cortical regions of trunk; (f)
structure of a typical latex vessel in greater detail.
into the roots. The average density of vessels at any level appears to correlate
well with the thickness of the outer cortex at that level. There are numerous
anastomoses, i.e., connections, between the vessels in anyone mantle. This
is illustrated in Fig. 9.3(c), which shows a longitudinal tangential section of
the cortical region through a ring of latex vessels. Figure 9.3(d) shows a
longitudinal radial section of the cortical region, from which it is evident
that anastomoses between successive mantles are rare. These various features
are summarized collectively in Fig. 9.3(e), which gives a three-dimensional
view of the interior of the cambial and cortical regions of the tree. Figure
9.3(f) shows in greater detail the structure of a typical latex vessel. The latex
vessels exhibit all the usual characteristics of plant cells, in addition to having
certain specialized features. They have a cellulosic envelope, against which
is pressed a cytoplasmic membrane. Nuclei derived from the parent
parenchymatous cells are present, although they are now retained against
the walls of the cell and are separated from the cytoplasm by thin membranes.
There are mitochondria present, recognizable as such by their morphology.
There is an endoplasmic reticulum associated with the microsomes.
9.2.2 The HeJ'ea brasiliensis plantations
9.2.2.1 Conditions required for the growth of Hevea brasiliensis

Hevea brasiliensis is a tree which flourishes under tropical conditions. Its
principal climatic requirements can be summarized as an average temperature
of ca. 80°F (27"C) and an average annual rainfall of ca. 80 in. (200 cm).
Drainage of the soil is important. With this proviso, a low-level site is preferred
to an elevated one, although Hevea brasiliensis has been grown successfully
at altitudes in excess of 600 m. Swampy, undrained, exposed localities are
especially unsuitable for the culture of this tree. Hevea brasiliensis will grow
satisfactorily in most types of soil, but the clay-loam type is preferred. It is
relatively insensitive to soil pH, although an acidic soil (pH 4.0-4.5) is preferred.
9.2.2.2 Regions of the world where Hevea brasiliensis is found

The Hevea brasiliensis tree is found in three principal regions of the world:
1. tropical America, notably the Amazon area of South America, which is
the original home of the tree, and the place where it was first discovered
by European explorers;
2. India, Sri Lanka and the tropical Far East, where the principal cultivated
plantations of the tree are to be found;
3. tropical Africa, where there are a few large plantations of fairly recent
establishment in places such as Liberia and the former Congo.
For many years now, the principal areas of the world in which natural
rubber is produced have been Malaysia, Thailand, Indonesia, and that part
of South East Asia formerly known as Indochina. Much of the latex obtained
from Hevea brasiliensis trees is used for the production of dry natural rubber,
and is exported from the natural-rubber-producing countries in this form.
Of the natural rubber which is exported in the form of preserved natural
rubber latex concentrate, most originates from Malaysia. Much of the natural
rubber in South East Asia is produced at latitudes less than 15° above the
Equator: these regions provide particularly advantageous climatic conditions.
Thus the plantations in Thailand are found mainly in the south and south-east
of the country. The extent to which the production of natural rubber is
8 Natural latices
encouraged in such countries depends upon several factors, notably the world
price for natural rubber, the extent to which local manufacturing industry
can take advantage of the proximity of a ready supply of natural rubber and
natural rubber latex, the extent to which other more profitable enterprises
- in particular, agricultural enterprises - are available, and the extent to
which competing agricultural enterprises are beset with difficulties arising
from climate, pests and diseases. Thus in Malaysia, important industries in
recent years have included tin mining, the production of oil-palm, petroleum
exploitation, and the manufacture of electronic goods, in addition to the
production of natural rubber. The relative importances of these industries
for the economy of the country has varied from time to time. In the rural
areas of Malaysia, the production of oil-palm has competed with that of
natural rubber for dominance. Both industries suffer from the disadvantage
relative to many others that investment and development are long-term; it
is not possible to develop a rubber or oil-palm plantation on a time-scale
comparable to that required for the building of a factory or a chemical plant.
In this respect, the natural-rubber-producing industry is disadvantaged
relative to the synthetic-rubber-producing industry. The way in which the
development of a local rubber-manufacturing industry can influence the
demand for natural rubber is illustrated by the late-1980s surge in demand
for medical examination gloves produced from natural rubber latex, following
increased perception ofthe human immunodeficiency virus as life-threatening
(see further sections 17.1 and 17.6 of Chapter 17 (Volume 3». Many new
manufacturing facilities for these gloves were set up in South East Asia,
thereby increasing considerably the demand for natural rubber latex concentrate.
At the time of writing, the demand for these gloves appears to have been
largely satisfied. Many of the intended manufacturing facilities either were
never fully commissioned or have now ceased production.
Much of the soil in the tropical Far East is rather poor, and would probably
be barren if transferred to a temperate region. Fertility depends upon the
prevailing conditions oftemperature and rainfall. Perpetual intensive weathering
processes make available to plants a continual supply of nutrients. This
intensive weathering, together with the disturbances to natural equilibria
which inevitably occur when attempts are made to cultivate tropical soils,
give rise to the problem of soil erosion. Some measure of soil conservation
has been attempted by interplanting Hevea trees with cover-crops. The
requirements for a satisfactory cover-crop are somewhat exacting, but there
are several obvious advantages which accrue from the planting of one which
is satisfactory. A cover plant which has been used successfully appears to be
Centrosema pubescens. This is a creeping legume which forms a loose mat
about 45 cm deep upon open ground, but which climbs any available support.
It is rather a slow grower at first, but forms an excellent cover when
established. Being a leguminous plant, it contributes to the available nitrogen
content of the soil by fixation of atmospheric nitrogen.
9.2.2.3 Outline of the history of the Hevea brasiliensis plantations

As early as 1834, Thomas Hancock drew attention to the high price of natural
rubber from Brazil, and suggested the possibility of growing rubber-
producing trees under more closely-controlled conditions in the similar
tropical climate of certain regions of the Far East. The first attempt to
implement this suggestion was made by Sir Joseph Hooker in 1873. Some
2000 Hevea brasiliensis seeds were obtained from the Amazon region of Brazil,
and planted at Kew Gardens in London. Only 12 seeds germinated. Of these
seeding plants, six were sent to Calcutta, but did not thrive there. In 1876,
a second attempt was made by Sir Henry Wickham. Some 70000 seeds were
brought from Brazil by himself, and planted at Kew. About 2000 young
plants were obtained. Most of these were sent to Ceylon (now Sri Lanka),
where they were successfully established. It was principally from Ceylon that
the Hevea brasiliensis was established in other parts of the Far East. An
important contribution to the development of the plantation industry was
made by Ridley. Whilst working at the Botanical Gardens in Singapore, he
studied the problem of how to preserve the viability of Hevea seeds during
periods of storage and transportation. The germination of figures cited above
indicate the magnitude of the problem. Ridley discovered that viability was
retained over longer periods if the seeds were packed in moist, powdered
charcoal. This procedure was widely used at the beginning of this century
for the distribution of Hevea seeds throughout the tropical Far East.
The development of Hevea plantations in the Far East proceeded only
very slowly for about 30 years after Wickham had initially introduced the
tree in the late 1870s. Four factors seem to have militated against the
expansion of the plantation industry:
1. The methods being used to obtain the latex from the trees were crude,
resulting in low yields and debilitated trees.
2. The early planters were under the misapprehension that the Hevea tree
flourished best in swampy ground. This was understandable, because the
tree had first been discovered growing in the swamps of tropical South
America. This misunderstanding of the optimum cultural requirements
for the tree led to lower yields of rubber than would otherwise have been
obtained.
3. There was no economic requirement for any new crop in the tropical Far East.
4. Consumer conservatism in Europe during the first decade of this century
resulted in a preference for wild natural rubber from Brazil over the
product of the plantations. The wild product was perceived to be more
uniform than the plantation product, and to have superior elasticity.
The first native rubber estates were established in 1900, and the first estate
owned by a European company was established in 1902. Plantation rubber
was first offered in substantial quantity in 1910. Since that time, rapid
10 Natural latices
development of the plantation industry occurred. By 1938, some 97% of the

world's supply of natural rubber came from Asia. Since 1912, the production
of wild natural rubber in tropical America has declined steadily. It is of no
importance now, except possibly to a few South American rubber-processing
companies. The rapid expansion of the rubber-plantation industry subsequent
to 1910 is attributed to reversal of the four factors outlined in the preceding
paragraph. The development of superior methods of obtaining the latex from
the tree, and a better understanding of the cultural conditions required for
the well-being of the tree, led to higher yields of rubber and greater
attractiveness of rubber plantations as a commercial enterprise. A series of
disastrous failures of established crops, such as coffee, cocoa, tobacco, etc.,
made planters more favourably disposed than hitherto towards the establishment
of new crops such as natural rubber. The development of the motor industry
in Europe and North America increased the world requirement for natural
rubber to such an extent that by 1912 Brazil was unable to satisfy all the
demands. Consumers were therefore forced to use the plantation product,
thereby discovering its two-fold superiority over the wild product in respect
of cleanliness and uniformity. The consequence of all these factors has been
the establishment in several parts of the world of extensive areas of plantations
of the Hevea brasiliensis tree, together with the ancillary facilities necessary
for working the latex from the trees up into either dry rubber or preserved
natural rubber latex concentrate. Such plantations are of broadly two types:
numeral small-holdings managed by local workers, and large estates owned
and managed by large companies.
In an attempt to protect themselves against a fall in world prices for natural
rubber, the Far East growers entered into a series of restrictive agreements
amongst themselves. This trend started in 1917 with an agreement between
the English growers. It was responsible for various attempts by the large
American rubber-manufacturing companies to establish plantations of their
own elsewhere in the world. Thus it was that the Firestone Tyre and Rubber
Company established in Liberia in 1923 a plantation which was to become
one of the largest in the world. Similarly, the Goodyear Company established
plantations in the Philippines in 1928. American companies have always
shown an understandable interest in the possibility of establishing plantations
of cultivated Hevea trees in tropical America. More than one attempt to do
this has been made. In 1928, the Ford Motor Company tried to establish a
plantation (,Fordlandia') in Brazil; in 1934 Goodyear tried to do the same
in central America. A second Ford plantation was established at Belterra in
Brazil in 1934. These attempts were, however, thwarted by a fungal infection
to which the Hevea tree is prone, known as South American leaf blight. So
far, this disease has not spread to other areas of the world, although, with
widespread air travel, it seems likely that it will eventually. South American
leaf blight has been described by McIndoe [11] as 'a sword of Damocles
which hangs over the head of the [plantation] industry, and a challenge to
the plant breeder'. The high-yielding strains of tree, which are, of course, of
great interest to the plantation industry, are especially susceptible to this
infection. Attempts have been made to develop high-yielding strains which
are resistant to this infection, but so far without any apparent success.
The natural rubber plantation industry has been subject to several
vicissitudes over the years. Much ofthe Far East was occupied by the Japanese
during World War II. As a consequence, supplies of natural rubber from
much of the tropical Far East were denied to the Allies. This underlined the
wisdom of having Hevea plantations established elsewhere in the world. Even
so, the Allies suffered serious shortages of natural rubber during World War
II. These shortages prompted a massive American cooperative (and very
successful) programme of research into the manufacture of satisfactory
general-purpose synthetic rubbers for use in products such as tyres and tyre
inner tubes for motor vehicles. Since World War II, the balance between
natural rubber and its competitors, with its consequences for the natural
rubber plantation industry, has been determined (at least in non-communist
countries) principally by normal commercial factors such as demand, supply,
costs and performance. The balance between supply and demand was
seriously distorted in the mid-1950s by stockpiling on the part of the United
States of America Government during the Korean war, in an attempt to
prevent repetition of the problems which had been encountered in World
War II in respect of supplies of natural rubber. The price of natural rubber
in the western democracies rose sharply as a consequence of this stockpiling.
This in turn seems to have provided added stimulus for the development of
synthetic alternatives; it was shortly after this that the successful development
of 'synthetic natural rubber' was first announced.
9.2.3 Propagation of Hevea brasiliensis
9.2.3.1 Introduction
Plants are propagated in two distinct ways: by seed, and by various vegetative
methods. Propagation by seed requires the fertilization of female flowers by
pollen from the male flowers of either the same plant or a different plant.
Either way, the characteristics of the resultant plant are a hybrid of the
characteristics actual or latent in the plant(s) from which the seed was derived.
Genetic mixing occurs, and the progeny is subject to the statistical laws of
Mendelian inheritance. This is the appropriate method of propagation when
it is desired to develop new varieties of a given type of plant. Vegetative
methods of propagation utilize those parts of the plant - leaves, stems and
roots - which are concerned primarily with supporting the life of the present
plant, and not with the establishment of future generations. The product of
vegetative propagation is usually genetically homogeneous with the parent
plant, although some deterioration in vigour may occur. Vegetative methods
12 Natural latices
are therefore preferred when it is desired to replicate a variety true-to-type,

without hybridization from other varieties. A group of plants resulting from
the vegetative propagation from one individual is known as a clone. Many
well-known clones of Hevea brasiliensis have been established over the years.
Not only the trees themselves, but also the latices obtained from them, often
have distinctive clonal characteristics.
9.2.3.2 Propagation by seed

The structure of the Hevea brasiliensis seed, and its mode of germination,
are illustrated in Figs 9.4(a)-(d). Figure 9.4(a) shows that the seed is protected
by a tough outer coat, and that immediately inside is a membrane known
Radicle
...--Stem
(a)
Hypocotyl
Radicle emerging
from germ pore
Secondary roots /Taproot

(b)
(c)
Cd)
Figure 9.4 Illustrating the structure of the Hevea brasiliensis seed and its mode of germination:
(a) shows a longitudinal section of a mature seed before germination; (b), (e) and (d) show
successive stages of germination.
as the inner seed coat. The germ pore, through which the plant will eventually
emerge, is a region where the outer seed coat is not so strongly developed.
The bulk of the seed consists of endosperm, which functions as a reserve food
supply to support the seedling plant until it has developed roots and leaves,
and so can obtain nutrients for its continued existence externally. The
structural elements of the embryo plant, namely, the radicle, hypocotyl and
plumule, do not at this stage appear to be characteristically different. They
do, however, appear as differentiated structures during the early stages of
germination. The successive stages of germination are illustrated in Figs
9.4(b)-(d).
Interest today is mainly in the germination of seed which has been obtained
by the controlled crossing of one clone with another. In this way, it is possible
to explore the characteristics of specific crosses· which are thought may offer
advantageous features. Controlled pollination is a delicate operation which
is carried out by hand. It is essential to prevent prior unwanted pollination
of the female flowers. The staminal columns of the selected male flowers are
inserted into the calices of selected female flowers in such a way that the
anthers and stigma are in close proximity. The female flowers should be
mature but hitherto unopened. After insertion of the staminal column, the
mouth of the corolla of the flower is closed with a plug, e.g., of cotton wool.
The purposes of this plug are to keep the staminal column in place, and to
prevent unwanted pollen from entering the flower. A drop of latex may be
placed between the petals and the plug to prevent the latter from being blown
away by the wind. Two or three days after pollination, the plug and the dried
petals fall off. If pollination has been successful, the ovary will begin to grow
into a fruit; if not, the flower will dry and fall off. The fruit matures 4-5
months after pollination. It is then enclosed in a bag to catch the seeds when
dehiscence occurs as the capsule ripens. The fertility of any given cross varies
with the varieties which are crossed, and with the direction of the cross. A
5% success rate for controlled pollination is considered to be reaonable.
The viability of Hevea brasiliensis seeds rapidly decreases with storage. It
has been recommended that the seeds be sown within 24 hours of collection,
the germination rate then being ca. 50-60%. Although seeds could be planted
in the positions to be occupied by the subsequent trees, this procedure clearly
tends to uneconomical use of the available space, because the young seedling
requires several years to come to maturity. The more usual practice is first
to germinate the seeds in special beds, and then, as soon as the radicle is
seen to be emerging, to transfer the seeds to a nursery bed. Here they remain
for a considerable time. The majority of them are used for vegetative
propagation by bud-grafting (see section 9.2.3.3 below). Those which are not
rna y be used as seedling stumps after a period of 1-2 years. Seedlings intended
to be used in this latter way are cut down to a height of ca. 60 cm and
defoliated. They are then left for some 10 days to allow the dormant buds
at the top of the cut stem to swell and sprout. The stump is then dug up
14 Natural latices
and its root lightly trimmed prior to being transplanted. Several shoots
develop on the cut stem. All but the strongest are removed, and this remaining
shoot regenerates the stem. There are two main objectives for this drastic
stumping: to stimulate the growth of the plant, and to harden the stem so
as to reduce its vulnerability to attack by small animals. An alternative
procedure is to allow the seedling plant to grow for ca. 4 years, and then
cut it back to a height of ca. 3 m. Such a plant is described as a high-stumped
seedling.
9.2.3.3 Vegetative propagation

The most common form of vegetative propagation practised for Hevea
brasiliensis is bud-grafting. This involves the bringing together of a wood-bud
scion and a rootstock in such a way that they fuse together. Ultimately sprouts
develop from the bud to generate the top of the resulting tree, which then
has the characteristics of the tree from which the bud was taken. The source
of the bud is selected with regard to such factors as the yield and quality of
the latex which is produced, and the resistance of the tree to leaf and bark
diseases. The root of the resultant tree retains the characteristics of the
rootstock; the latter is selected primarily for its vigour and its resistance to
root diseases.
The principal steps of a typical bud-grafting process are indicated in Fig.
9.5. The wood buds are obtained from the young succulent branches of the
mother tree. These branches should be about one year old, upon which cork
has just begun to form. They should be cut whilst in a state of vigorous
growth, i.e., when in the stage of putting forth a fresh increment of growth.
The wood buds form in the axils of the previous year's leaf stems. They may
be recognized as an irregular patch above, and distinct from, the scar left
where last year's leaf fell. The bud is removed by cutting out a small piece
of the surrounding bark which is backed by a thin sliver of wood; the
function of the latter is to protect the cambium and facilitate handling.
Traditionally, the rootstock has been a seedling plant 6-12 months old,
having stem diameter 2.5- 5 cm at a height of 10 cm above ground level.
There is now a tendency to use younger rootstocks (ca. 3 months), in which
case the process is known as green budding. Typically, the budding process
commences by opening a small panel of bark (ca. 2 cm x 9 cm) at an
appropriate height from the ground, the opening being such as to form a
receptacle into which the piece of bark bearing the bud can be placed. The
incisions for this panel are made well before the bud is inserted, so as to
allow time for any exuded latex to dry to such an extent that no more liquid
latex will exude and contaminate the cambium of the rootstock when the
panel of bark is pulled away. This panel may be cut so that it is opened
either from the top or from the bottom. The incisions should be deep enough
Branch
bud
Leaf
scar
Cutting of budpatch Budwood after

removal of budpatch
Cutting of bud-panels
Insertion of budpatch
into pana!
Development of viable bud-graft 'Elephant's foot'
Figure 9.5 Illustrating principal steps of a typical bud-grafting process for the vegetative
propagation of Hevea brasiliensis.
16 Natural latices
just to touch the sapwood. The piece of bark containing the bud from which
the sliver of wood has been taken is now inserted into the opened panel so
that it faces outwards and upwards, and is pushed in as far as possible. The
piece of bark is trimmed beforehand so that a gap of ca. 2.5 mm is left all
the way round. If no gap is left, then the piece of bark tends to be pushed
out of the panel by the wound callus which forms subsequently. The panel
is then closed and bound up with a material, such as muslin, which has been
treated with paraffin wax, or with coconut leaves and string. The binding is
tight, so as to prevent penetration by any rainwater which runs down the
trunk. The panel is inspected 2-4 weeks after bud-grafting, depending upon
the amount of rain which has fallen during that time. The viability of the
bud may be tested by making a small incision: the cut surface thereby exposed
should be green and healthy. The bud is inspected again after a further 1-2
weeks, when definite signs of growth should be apparent. Shortly after this
second inspection, the trunk above the bud is cut off obliquely to within
2.5-tO cm of the top of the budding panel, if the bud appears to be viable.
No sealing of this cut is necessary. The effect is to activate all the dormant
buds around the periphery of the cut, as well as the grafted bud. Shoots
appear some 3 weeks after the trunk has been cut. Those from the ungrafted
buds are darker in colour than that of the grafted bud. All the roots from
the ungrafted buds must, of course, be removed. The general progress of the
shoot from the grafted bud is indicated in Fig. 9.5. After about 5 months,
the emergent stem becomes densely foliated, and the residual trunk above
the graft begins to decay. In fully-grown trees, the grafted union gives rise
to a characteristic appearance at the base of the tree to which the term
elephant's foot is appropriately applied.
Young budded trees may be transferred to a field site ca. 2-3 months after
budding, and then stumped to a height of ca. 2.5 m ca. 2 years after budding.
They may then be transplanted into their final positions ca. 3 weeks later.
They must be planted so that the union is above soil-level, otherwise the
scion may throw out its own roots. Some interest has also been shown in
high-stumped buddings. To produce these, the budding is left to grow to a
greater height before being stumped. The advantages which have been claimed
for this procedure include a tendency for the tree to grow straighter, greater
uniformity of growth with more even development of the tree, a tendency
for the tree to come into tapping about 1 year earlier after planting out, and
the absence of any tendency for the scion to commence rooting should the
union be buried below soil-level. Interest has also been shown in recent years
in three-part trees, i.e., trees which have been produced by a second grafting
at the level where the canopy of the tree will form. By this means, it is possible
to produce trees which have selected root, trunk and canopy characteristics.
The latter are selected with such attributes as resistance to wind damage and
resistance to leaf diseases in view. It is not clear to what extent this concept
has been exploited at present. Against the advantages accruing from a tree
having selected canopy characteristics must be set the obvious economic

disadvantage of an additional grafting stage.
Other methods of propagating Hevea brasiliensis vegetatively have been
explored. There are two principal reasons for interest in vegetative propagation
by methods other than grafting:
1. vegetatively-propagated rootstocks of predictable characteristics could
thereby be made available;
2. the performance of clonal trees growing on their own roots could be
in vestigated.
Propagation from branch cuttings is possible, but the main difficulty with
this method is rapid loss of moisture from the cutting by transpiration. To
overcome this difficulty, it is necessary to apply a water mist for ca. 12 hours
during the day. The following procedure for striking leafy softwood branch
cuttings has been described by Tinley [12]: The rooting medium comprises
a mixture of equal volumes of sand and partially-rotted sawdust. It is laid
15 cm deep over a porous substrate, such as a mixture of granite chippings
and broken bricks. This combination of rooting medium and substrate
materials provides a well-drained and aerated medium which possesses
suitable water-retaining characteristics. The shoots are obtained from bushes
which have been cut back hard so as to encourage vigorous growth. They
should be cut with fully-expanded leaves and dormant terminal buds, which
are cut off. The shoots are then trimmed to ca. 25 cm in length by a slanting
cut just below a node. The lower leaves are removed, and the shoots stood
in water for a few minutes. Any coagulated latex is removed from the cut
end, which is then dipped in a powder containing 20% m/m iron(III)
dimethyldithiocarbamate. This substance is primarily a fungicide, but it may
also have a slight growth-promoting effect. Conventional rooting aids do
not appear to be of much assistance in this case. The shoots are then inserted
in the rooting medium, and subjected to a continuous water-mist during the
day. A nutrient solution containing urea, potassium nitrate and ammonium
phosphate is applied through the mist nozzles at nightfall, to replace the
minerals which have been leached from the rooting medium by the continual
application of water throughout the day. Rooting commences ca. 6 weeks
after the insertion of the cuttings. The cuttings are then potted and hardened
off by gradually reducing the duration of the water-mist. It is found that
cuttings tend to develop lateral roots rather than a single tap-root. The
proportion of successful strikes obtained by this method depends upon the
clone from which the cuttings were obtained. It seems that the tendency of
a cutting to strike roots is not increased by taking a 'heel' with the cutting.
Alternative methods of vegetative propagation in which some interest has
been shown recently include tissue culture and protoplast culture. In outline,
the procedure for tissue culture is to separate anthers from male flowers
under sterile conditions, culture them on a medium supplemented with
18 Natural latices
appropriate hormones so that calluses form on the anther walls, and then
transfer the callus cells to a medium which induces the development of
embryoids which then develop into complete clonal plants, each with its own
tap-root. Protoplasts are naked cells from which the cellulosic cell walls have
been removed by enzymatic digestion. It is possible to culture such cells to
form complete plants, and to form hybrid protoplasts which contain genetic
material from different sources. It is said that haploid (half-chromosome)
Hevea brasiliensis trees have been grown by vegetative propagation from
gametic cells, and that these trees appear to be normal. Normal Hevea has
36 chromosomes; the anther cells used for tissue culture have this number,
being somatic.
9.2.4 Tapping
By tapping is meant the process by which latex is obtained from the latex
vessels of the Hevea brasiliensis tree. In principle, because the latex is under
considerable osmotic pressure in the latex vessels, it is merely necessary to
sever the vessels for the latex to exude spontaneously. As the latex exudes
and the vessels empty, so the osmotic pressure decreases and the flow
gradually subsides until the latex flows so slowly that it coagulates by
evaporation before it has run very far. The coagulum which forms effectively
plugs the severed vessels, so that the flow soon ceases altogether. The vessels
now gradually fill up again with latex, until the equilibrium osmotic pressure
is once again attained. If the original cut is re-opened, or a fresh incision
made, then more latex flows and the cycle of events is repeated. During the
period of flow associated with one incision, the flow is rapid at first but soon
declines to a more or less steady rate which persists throughout the greater
part of the flow period.
9.2.4.2 Early tapping systems

The earliest known tapping system was that employed by the natives of Brazil
to extract latex from wild rubber trees. It consisted in the making of numerous
small incisions in the bark of the tree, using a small axe on a long handle.
At each tapping, the incisions were roughly equally spaced in the form of a
ring surrounding the trunk of the tree. Each ring was made below the previous
one, from the highest level attainable right down to the base of the tree. A
second series of incisions was then made midway between the rings of the
first series, and so on. The cuts themselves were 2.5-5 cm long, the horizontal
spacing in the rings 10-40 cm, and the vertical separation between consecutive
rings 5-30cm. The trees were tapped for approximately 6 months of each
year, during the dry season. They were rested during the rainy season until
the next dry season. The cuts usually penetrated the sapwood, and the trees
often reacted by developing calluses where they had been wounded. After a
few years, the part of the trunk accessible to a tapper standing on the ground
became so bumpy that either the tree had to be abandoned or else it had to
be tapped from a ladder. The early plantation tapping methods were
developed from those used by the Brazilian natives. Thus we are told that,
in the early days of the Ceylonese plantations, it was the practice to make
systematic gashes in the bark of the tree, using either a knife or a hammer
and chisel. Gradually, the custom developed of making small V-shaped
incisions in the trunk.
The principal early Malayan systems are illustrated in Figs 9.6(a)-(e).
Figure 9.6(a) shows the half-herringbone system favoured by Ridley ca. 1889.
(a) (b) (c)
"- '-
"- '-.
"
'<::::<"
{ f
(0) {e} (f)
Figure 9.6 Illustrating various tapping systems for Hevea brasiliensis: (a) half-herringbone
system favoured by Ridley; (b) full-herringbone system; (c) system comprising V-cuts on a half
circumference; (d) full spiral system in which one cut gradually winds around the trunk several
times; (e) system of multiple spirals winding their separate courses up the tree; (f) a modern
tapping panel, showing the way in which it develops.
20 Natural latices
A narrow, flat chisel was used to remove shavings from the lower side of the
oblique cuts, and thus to re-open the cuts after the initial latex flow had
ceased. Re-opening of the cuts ceased when a strip of the bark ca. 1 cm wide
had been removed. A new oblique cut of the herringbone was then opened
on virgin bark. Other early systems are illustrated in Figs 9.6(b)-(e). These
are, respectively, the full-herringbone system, a system comprising a series of
V-cuts on a half circumference separated by a vertical distance of ca. 30 em,
a full spiral in which one cut gradually winds around the trunk several times,
and a system of multiple spirals winding their separate courses up the tree.
9.2.4.3 Modern tapping systems

All the modern systems of tapping Hevea brasiliensis trees depend upon the
continual re-opening of a single cut in the bark of the trunk. A characteristic
tapping panel gradually develops as the bark is continually shaved away from
the lower edge of the cut. The development of a tapping panel is illustrated
in Fig. 9.6(f). As the panel develops downwards, so the bark tends to be
spontaneously regenerated from the top of the panel. The discovery of the
principle of continually re-opening a single incision is attributed to Ridley.
It has been noted in section 9.2.4.2 above that initially the oblique incisions
of the half-herringbone system were re-opened until the cut was ca. I cm
wide. It was believed that the tree would not tolerate wider cuts. Gradually,
the number of re-openings, and hence the width of the cut, was increased,
until it became apparent that the most satisfactory tapping procedure was
to re-open the same cut indefinitely.
Four distinct types of incision are recognized in modern tapping practice.
By far the most widely employed is the single spiral cut. Logically this should
be left-handed, i.e., it should run from high left to low right. By this means,
the number oflatex vessels severed per unit length of cut should be maximized
because (see section 9.2.1.2 above) the latex vessels tend to form a right-handed
spiral as they progress up the tree. However, tapping panels which run from
low left to high right are also used. V-cuts are sometimes preferred to spiral
cuts for larger trees, but are less efficient than the latter. Also used are
extended V-cuts, in which the left-hand arm of the V is extended, and full
circumferential cuts of various types. The length of the tapping cut is specified
as a fraction of the circumference of the trunk at that elevation. A common
modern tapping system comprises a half-circumference left-handed spiral
which is re-opened on alternate days. A healthy tree is able to withstand this
treatment almost indefinitely. A somewhat less intensive system is the same
spiral re-opened every third day. This is the preferred system for certain
clones, e.g., Glenshiel I, which have a predisposition to a physiological
derangement of the bark known as brown bast. This condition is encouraged
by intensive tapping.
Much, if not all, the rubber tapping which has been practised is carried
out by hand. This remains the case at present. Typically, a single operative
can tap ca. 600 trees. As regards tapping implements, whereas carpenters'
chisels and gouges were commonly used at one time, a variety of
specially-designed tapping knives is now available. An essential feature of
these is that the depth of the cut can readily be controlled. Many kinds of
receptacle are used for the collection of the exudate from the tree; these
include brown earthenware cups of capacity 350 to 700cm 3 glazed on the
inside, aluminium cups, glass cups, and coconut shells. To direct the flow of
latex into the cup, a small galvanized iron spout is usually worked into the
bark by hand pressure.
9.2.4.4 Factors which affect the yield of latex and rubber

The yield of latex which is obtained as a result of a single incision depends
upon a number of factors, in addition to the age, variety and state of health
of the tree, or whether or not artificial means (considered in section 9.2.5
below) are used to stimulate the production of latex in the tree.
The number of times the cut has been opened previously has a profound
effect upon both the character and the yield of the exudate. Initially, the latex
is viscous and contains ca. 60% m/m of rubber. Very little latex is obtained
at this stage. As the cut continues to be re-opened, a larger volume of a less
viscous exudate containing less rubber is obtained. After several re-openings,
the rubber content falls to ca. 30% m/m, and is maintained at this level. The
yield of dry rubber per re-opening increases somewhat, but not as markedly
as the volume of latex, because of the fall in the rubber content of the latex.
The phenomena associated with the repeated re-opening of a cut are known
collectively as the wound response or dilution reaction. They are mainly a
consequence of the influx of water into the latex vessels from the surrounding
parenchymatous cells.
The rate of bark consumption also affects the yield of latex per re-opening.
If insufficient bark is cut away, then the small rubber plugs which have sealed
the vessels at the cessation of the previous flow are not completely removed.
On the other hand, if the rate of bark consumption is too high, then the
available bark is used up too rapidly. Hence a compromise is necessary. The
rate of bark consumption for a cut which is re-opened on alternate days is
typically ca. 2.5 cm per month. The depth to which a cut is made is also
important, because the latex vessels are most numerous near the cambium.
Furthermore, owing to the scarcity of radial anastamoses, the latex will not
drain from the inner mantles of vessels through incisions in the outer mantles.
For maximum yield, it is therefore necessary to cut in as closely as possible
to the cambial layer without actually damaging the latter. The angle and
direction of the cut can also influence the yield. The cut is made in an oblique
direction in order to encourage the latex to flow into the receptacle. The
preferred angle is ca. 30· to the horizontal. As noted in section 9.2.4.3 above,
22 Natural latices
the preferred direction is upper left to lower right, because more latex vessels
per unit length are severed than if the cut is in the opposite direction. The
yield per cut decreases as the height of the cut above ground level increases.
This is because the density of latex vessels in the trunk is a maximum at
ground level, and decreases as the tree is ascended. The rate at which the
yield diminishes with height from the ground is said to be ca. 15-20% m- 1 .
The time of day at which the cut is opened is an important variable in
determining the yield of latex from the cut: the earlier the cut is opened, the
greater the yield. There are two main reasons why this is so:
1. The turgor of the latex vessels and of the surrounding parenchymatous
cells is greatest in the early morning when the humidity is high and the
temperature is low. As the morning progresses, so the rate of transpiration
increases and turgor falls.
2. The latex dries out more easily as the temperature rises, and so the flow
tends to cease more rapidly as the severed ends of the latex vessels become
plugged with coagulated rubber.
9.2.4.5 Other methods of tapping

The conventional method used for collecting natural rubber latex from the
Hevea brasiliensis tree is tedious and labour-intensive. It is therefore not
surprising that alternative methods of tapping have been sought. One of the
more promising of these is known as puncture tapping, in which the
conventional tapping panel is replaced by a series of holes punched in the
bark of the tree to a controlled depth. Interest has been expressed from time
to time in the possibility of automating the collection of latex from the tree,
but no feasible processes have yet been developed. The collection of natural
rubber latex from the Hevea brasiliensis tree remains a labour-intensive,
rather primitive operation which requires some degree of skill to enable the
tapping panel to be efficiently re-opened continually with minimum damage
to the tree.
9.2.5 Improvement of latex yield
9.2.5.1 By selective breeding

The yields of latex and rubber which are obtained from un selected Hevea
brasiliensis seedlings are generally low. A plantation industry based upon
such trees could survive only whilst very cheap labour was available. Early
yields of rubber were ca. 100 kg acre - 1 year - 1; ca. 90 trees had to be tapped
for a year to obtain sufficient rubber for one bale. Gradually yields have
been increased by selective breeding, first to ca. 250 kg acre - 1 year -1, then
to ca. 500 kg acre - 1 year - 1, and now to yields in the range 1500-2000
kg acre - 1 year - 1 or more. The process of selection is, however, slow. I t has
been estimated that a period of some 20 years is required for the selection,
development and intensive planting of a new variety. This period may be

compared with the few months required for the construction and commissioning
of additional plant capacity for the production of synthetic rubber.
In one important respect the plantation industry has been fortunate.
Wickham's original selection of Hevea brasiliensis seeds was favourable, in
that it contained sufficient potential genetic heterogeneity which succeeding
generations of plant breeders have been able to explore to great effect.
Taxonomists are said now to recognize at least nine distinct Hevea species,
of which Hevea brasiliensis is but one, all of which have the same number
of chromoses, and all of which hybridize freely. Furthermore, these species
show wide variations in characteristics such as habit, leaves, flowers, etc., as
well as in the quality and quantity of the latex which they produce. It is
probable that, propitious though Wickham's original selection was, it
represents but a small proportion of the genetic potential available from the
seeds of the South American Hevea hybrids as a whole.
Of prime importance in the selection of seedlings is, of course, the yield
of rubber which the variety promises. To obtain an early indication of the
potential of a particular cross in this respect, various test-tapping procedures
for young trees have been developed. It has become increasingly evident,
however, that high yield is not the only important characteristic to be sought
in a clone. Others include resistance to damage of the tree by wind, resistance
to disease, and, in some cases, the ability to produce a latex which is suitable
for concentration and utilization as such.
9.2.5.2 By yield stimulation

The technique of applying to the tree substances which will stimulate the
flow of latex has a long history which has been outlined by Blackman [13].
A more recent review is available in a paper by Abraham et at. [14]. By the
use of such stimulants, the plantations are to some extent able to regulate
the supply of natural rubber and natural rubber latex to the market, thereby
responding one way or the other to fluctuations in demand and price. The
precise effect of any given yield stimulator depends upon the clone to which
the tree belongs. Thus, for example, members of the clone RRIM 605 are
said to be much more susceptible to the effects of these substances than are
members of the clone RRIM 501.
Two particular techniques for stimulating the yield of latex are noted here.
In the first, which has been described by Compagnon and Tixier [15], a few
grams of copper sulphate are injected into the tree as an aqueous solution.
This causes an initial sharp increase in yield, which gradually falls over a
period of 6 months. The initial response may be restored by a further injection
of copper sulphate. Although it has been claimed that the copper content of
the resultant latex is not thereby significantly increased, the fear of such
increase is an important factor which has discouraged commercial exploitation
24 Natural latices
of this procedure. The basis of this fear is that traces of certain transition-metal
compounds, notably those of copper, cobalt and manganese, are able very
effectively to catalyse oxidative degradation of natural rubber hydrocarbon,
both before and after vulcanization.
The second technique makes use of certain derivatives of phenoxyacetic
acid (Structure I). These substances are well-known as activators of plant
growth. Especially effective as stimulators of latex yield are 2,4-dichloro-
phenoxyacetic acid (Structure II) and 2,4,5-trichlorophenoxyacetic acid
(Structure III). These substances can be applied either as aqueous solutions
CI CI
CI CI
of the sodium salts containing some emulsified palm oil, or as the butyl esters
dissolved in palm oil. They are applied to the bark in the region of the
tapping panel, or to the tapping panel itself, immediately above the tapping
cut. For the first few days after the application of phenoxyacetic acid
derivatives, the main effect is to increase the initial rate of exudation of the
latex. By contrast, the longer-term effect is to prolong the period for which
the latex flows, rather than to increase the initial flow rate. It is possible to
envisage several ways in which these substances could increase the yield of
latex. The available evidence indicates that one important factor is an increase
in the region of the tree from which the latex drains. The phenoxyacetic acid
derivatives also appear to reduce the viscosity of the latex, and so encourage
flow. The dilution reaction seems to be little affected by these substances.
Likewise, neither the colloid stability of the latex nor the osmotic pressure
within the latex vessels is affected significantly.
9.3 PRESERVATION OF NATURAL RUBBER LATEX
9.3.1 Preliminary considerations
9.3.1.1 Why it is necessary to preserve natural rubber latex

It has been noted above that natural rubber latex is the contents (sap) of a
specialized type of cell in the Hevea brasiliensis tree, namely, the latex vessels.
Preservation of natural rubber latex 25
As such, it is a complex biochemical system. It is therefore not surprising

that obvious chemical changes occur shortly after the latex leaves the tree.
The first of these changes is that the latex coagulates within a few hours.
This process is known as spontaneous coagulation. The time required for
coagulation depends upon the ambient temperature and upon the colloid
stability of the latex; but always the result is that the latex separates into
clots of rubber and a clear aqueous phase. Spontaneous coagulation is to
be distinguished from the second of the obvious changes. This is putrefaction,
which sets in at a later stage, with the development of bad odours. It is to
prevent both these processes occurring that preservation is necessary. A
distinction is made between long- and short-term preservation. The former
implies preservation during the period of shipment to, and transportation
and storage within, the consumer country. The latter implies merely preservation
for a sufficient time to ensure that the latex remains in a liquid condition
for a few hours or a few days before being processed into the various forms
of dry rubber. Short-term preservatives are commonly known as anticoagulants.
9.3.1.2 Mechanism of spontaneous coagulation of natural rubber latex

Two theories have been proposed to explain the phenomenon of spontaneous
coagulation. The first postulates that the process is primarily a consequence
of the development of acidity through the interaction of microorganisms with
the various non-rubber constituents of the latex. The second attributes the
effect to the liberation of fatty-acid anions through the hydrolysis of various
lipid substances present in the latex. Such anions are then thought to be
adsorbed at the surfaces of the rubber particles, possibly partially displacing
adsorbed proteins. They then interact with divalent metal ions, such as those
of calcium and magnesium, which are either present in the latex initially or
else are gradually released from complexes by the action of enzymes.
The more important of the facts relating to the spontaneous coagulation
of fresh natural rubber latex, in the light of which these two theories are to
be assessed, can be summarized as follows:
1. The pH of the latex does not usually fall below 6.3 -6.0 during spontaneous
coagulation, whereas the pH offresh natural rubber latex has to be reduced
to below ca. 5 for coagulation to be effected by acidification. Furthermore,
spontaneous coagulation is not prevented by maintaining the pH at ca.
7 by the addition of, say, dilute alkali.
2. The process of spontaneous coagulation can be accelerated by the addition
of small amounts of carboxylate soaps. However, whereas small additions
sensitize the latex to spontaneous coagulation, larger additions cause
retardation of the process. The range of soap addition over which the
coagulation time is a minimum depends upon the calcium and magnesium
content of the latex.
26 Natural latices
3. Spontaneous coagulation, and its sensitization by carboxylate soaps,

depends upon the presence of divalent metal ions, such as those of calcium
and magnesium. If these are absent, then unpreserved latex can be kept
in a fluid condition for a longer period until putrefaction commences.
4. The proportion of ether-soluble substances in the rubber coagulum which
is obtained by precipitation with alcohol decreases with time after tapping.
Furthermore, with the passage of time, the saponification value of the
ether extract falls to a shallow minimum and then increases again.
Clearly, these observations accord with the second of the theories outlined
above, rather than with the first. In particular, the fourth of these observations
is interpreted as implying the conversion of ether-soluble lecithins to various
derivatives, some of which (e.g. inorganic phosphates) are not soluble in ether.
It is, however, perhaps significant that de Haan-Homans [16] failed to detect
the presence of fat-splitting enzymes (lipases) in the fresh latex of Hevea
brasiliensis. The presence of esterase was established. This enzyme is capable
of hydrolysing rapidly, and to a high degree, the esters of the lower fatty
acids, notably those of acetic acid.
9.3.2 The ideal preservative for natural rubber latex
9.3.2.1 Efficacy as a preservative

The primary requirement of a preservative is that it should effectively preserve
the latex against spontaneous coagulation and putrefaction. Obvious though
this requirement is, at least three ways in which a preservative should be
effective can be distinguished:
1. It should destroy latex microorganisms, or at least suppress their activity

and growth in the latex.
2. It should enhance the colloid stability of the latex by such means as:
(a) increasing the density of bound electric charge and the electrokinetic
potential at the interface between the rubber particles and the latex
aqueous phase;
(b) increasing the ability of surface-bound hydrophilic macromolecules to
confer steric stabilization on the latex;
(c) increasing the degree of hydration of the particle surface.
3. It should de-activate trace metal-ions, especially those of multivalent
metals, either by sequestration or by precipitation as insoluble salts. Such
deactivation is desirable for two unconnected reasons:
(a) some of these metal ions are essential for the well-being of the
microorganisms which cause spontaneous coagulation and putrefaction
in natural rubber latex;
(b) multivalent metal ions tend to reduce the colloid stability of the latex,
the principal culprit in natural rubber latex in this respect being
magnesium ions.
9.3.2.2 Ancillary requirements

The ideal preservative also satisfies a number of ancillary requirements,
besides being effective as a preservative. It should be harmless both to people
and to rubber. It should not discolour the latex, or films of rubber obtained
from the latex. It should not impart an offensive odour to the latex. It should
not interfere with established latex processes. It should be cheap and
convenient to handle.
9.3.3 Ammonia as a preservative
9.3.3.1 Historical aspects

The preservative effect which ammonia has upon natural rubber latex was
first recorded in patents to Johnson [17] and to Norris [18] published in
1853. The Johnson patent gives the following brief description of the
preservation of natural rubber latex by ammonia and other compounds:
Shortly after the milk or juice is collected it is strained, and has then added
to it a quantity of the concentrated liquor of ammonia, or other ammoniacal
matter, or any combination of nitrogen and carbon. The mixture is then
well mixed, when it will remain in a white fluid state, capable of
transportation and use as a preserved material, ifkept in air-tight receptacles.
The preferred preservative is concentrated aqueous ammonia solution, and
the preferred proportion given as ca. 1 fluid ounce (ca. 30cm 3 ) of ammonia
solution to every pound mass (453 g) oflatex. Assuming that the concentration
of the ammonia solution used by Johnson was ca. 18 mol dm - 3, it appears
that Johnson's preferred concentration of ammonia was ca. 2% mjm on the
whole latex; this would be regarded as excessive today (see section 9.3.3.2
below). It will be evident from the above quotation from the Johnson patent
that, in effect, he claimed a very large number of nitrogen-containing
compounds as preservatives for natural rubber latex. It is hardly surprising
that a disclaimer to the Johnson patent was filed in 1856 (on 18th February,
by Armstrong, assignee of the Norris patent). The effect of this disclaimer
was to restrict the claim to concentrated aqueous ammonia solution.
Since the middle of the 19th century, ammonia has come to be regarded
as the standard preservative for natural rubber latex, notwithstanding the
disadvantages attending its use which are noted below. It will become evident
from what follows that, even some 140 years after the publication of the
Johnson and Norris patents, the majority of preservative systems for natural
rubber latex contain some ammonia.
28 Natural latices
9.3.3.2 Ammoniation of natural rubber latex

0.2% m/m of ammonia on the whole latex is sufficient for short-term
preservation; ca. 0.7% m/m is required for long-term preservation in the
absence of ancillary preservatives. Assuming that the ammonia is confined
entirely to the aqueous phase of the latex, 0.7% m/m ofammonia in 60% m/m
latex corresponds to a molar concentration of almost exactly 1 mol dm - 3 in
the aqueous phase. It has long been the practice to add a small amount (ca.
0.01 % m/m) to the latex as an 'anticoagulant' as soon as possible after the
latex leaves the tree, i.e., whilst it is still in the collecting cups and buckets.
However, this may be bad practice, because it may lead to the more rapid
multiplication of microorganisms than would occur in the absence of
ammonia altogether (see further section 9.3.3.3.1 below).
The ammonia is supplied to the latex from anhydrous liquid in cylinders,
rather than as concentrated aqueous ammonia solution. There are two
reasons for preferring the anhydrous liquid to concentrated aqueous solution.
Firstly, the concentrated aqueous solution is more hazardous to handle under
tropical conditions than is the anhydrous liquid in cylinders. Secondly, there
is a tendency for the concentrated aqueous solution to cause local coagulation
of the latex. For long-term ammoniation, equipment of the type illustrated
in Fig. 9.7 is commonly employed for ammoniation. The cylinder of ammonia
rests on a scale, so that the amount of ammonia passed to the latex can be
inferred from the loss in weight of the cylinder. A limit to the rate at which
ammonia can be drawn from the cylinder is set by the insulating effect of
Third valve to
----prevent suck-back
Rubber hose
Secondary
/ Drilled with holes
valve ____
Spreader-to be
immersed in latex
/
Main valve
Ammonia cylinder
Figure 9.7 Typical equipment used for long-term ammoniation of natural rubber latex.
the layer of ice which builds up around the cylinder by condensation from
the atmosphere of high humidity surrounding the cylinder. The ammonia
passes into the latex through a distributor. This comprises four perforated
iron pipes set at right angles to each other in a horizontal plane, and connected
by means of a five-way union to the vertical main feed pipe.
9.3.3.3 Ammonia and the requirements for the ideal preservative
9.3.3.3.1 Ammonia as a bactericide

We consider now the extent to which ammonia meets the first of the exacting
requirements for the ideal preservative outlined in section 9.3.2 above, namely,
efficacy as a bactericide. It appears that ammonia is very effective as a
bactericide if the level exceeds 0.35% mlm on the whole latex. Indeed, work
by Lowe [19] has indicated that the effectiveness of preservation is dependent
upon just how soon ammoniation to 0.35% takes place after the latex has
left the tree. Figure 9.8, which has been constructed from results reported by
Lowe, shows the effect of ammonia content upon the bacterial count of
Initial bacterial count

before ammoniation
104~----~----~-----L----~----~
o 0.1 0.2 0.3 0.4 0.5
Initial ammonia content / % mlm on whole latex
Figure 9.8 Effect of ammonia content upon the bacterial count of natural rubber latex, drawn
from results reported by Lowe [19].
30 Natural latices
natural rubber latex. It is seen that small amounts of ammonia (up to ca.
0.05%) encourage the growth of bacteria, presumably because the pH of the
latex has increased from the value of ca. 6.5 for the latex in the tree to ca.
8.0. The latter value is approximately that for the multiplication of many
common forms of bacteria. Furthermore, the level of ammoniation must be
at least 0.1 % before the mortality rate is equal to the rate of multiplication,
so that the population of viable bacteria remains constant. These observations
cast doubt upon the wisdom of adding small amounts of ammonia to the
tapping receptacles (see section 9.3.3.2 above).
A further risk which may arise from initial small additions of ammonia to
the latex is that the multiplication of strains of microorganisms which have
some degree of resistance to ammonia may thereby be encouraged. Even if
sufficient ammonia is added at the cup and bucket stage to suppress a
considerable proportion of the bacterial population, a small population which
contains ammonia-resistant varieties can be more troublesome to eliminate
ultimately than the much larger population of ammonia-sensitive bacteria
which would have been present had there been no initial ammoniation.
9.3.3.3.2 Ammonia as an alkali

Ammonia imparts alkalinity to the aqueous phase of natural rubber latex.
The effect of ammoniation upon the long-term pH of 60% mlm natural
rubber latex concentrate is shown in Fig. 9.9. Two curves are given: curve
A shows the variation of pH with ammonia content expressed as % mlm on
12
11 c
B
pH
60~------~O~.5~------~I~.O~-------I~.5~-------2~.O
Ammonia content /% mm
Figure 9.9 Effect of ammoniation upon the long-term pH of 60"10 m/m natural rubber latex
concentrate. Curve A shows pH as a function of ammonia content expressed as % m/m on the
whole latex; curve B shows pH as a function of ammonia content expressed as % m/m on the
aqueous phase of the latex. Also shown (curve q is the theoretical prediction for the variation
of pH with concentration for aqueous ammonia solutions at 15'C, taking Kb =
1.6982 x 10- 5 mol dm - 3 and K" = 4.5186 x 10- 15 mol dm - 3.
the whole latex; curve B shows the variation of pH with ammonia content
expressed as % mlm on the aqueous phase of the latex. Also shown (curve
C) is the theoretical variation of pH with concentration for aqueous ammonia
solutions. Comparing the latter curve with curve B shows that, for a given
ammonia concentration, the pH of the aqueous phase is always lower than
that of an aqueous solution. The difference is ca. 1 pH unit in the region of
industrial interest. It is a consequence of chemical interactions with various
non-rubber constituents in the latex.
The increase in the initial pH of the latex has several consequences. These
fall under two broad headings as follows:
1. The density of bound electric charge and the electrokinetic potential at

the interface between the rubber particles and the aqueous phase are
increased, thereby increasing the initial colloid stability of the latex. The
increase in the electrokinetic potential is believed to be a consequence of
interaction between the ammonia and proteins adsorbed at the surface of
the rubber particles. The ability of surface-bound hydrophilic macromolecules
(mainly proteins) to confer steric stabilization on the latex may also be
increased; so also may be the degree of hydration of the particle surface.
However, little appears to be known about these aspects of the effect of
ammoniation upon fresh natural rubber latex.
2. The hydrolysis of several of the non-rubber constituents of the latex is
accelerated. This matter is considered further in section 9.5.5 below. It is
sufficient to note here that the chemical changes which occur as a result
of these hydrolysis reactions also tend to increase the colloid stability of
the latex markedly. Thus, apart from being effective as a bactericide,
ammoniation increases the colloid stability of natural rubber latex both
short-term and long-term.
9.3.3.3.3 Ammonia and multivalent metal ions

Ammonia is able to de-activate some multivalent metal ions through
sequestration by the formation of ammine ions. Other multivalent metal ions
are effectively de-activated through removal from the latex aqueous phase
by precipitation as sparingly-soluble compounds. Some metal ions may be
precipitated as hydroxides. More important for the colloid stability of
ammonia-preserved natural rubber latex is the interaction of magnesium ions
with ammonia and any orthophosphate anions which may be present in the
aqueous phase, to precipitate the very sparingly-soluble magnesium ammonium
phosphate according to the following reaction:
Mg2+ + NH3 + HPO;- --+ MgNH 4 P0 4 !

The precipitated magnesium ammonium phosphate sediments from the latex
together with any sand, dirt or soil which may be present. Gradual
crystallization of this compound after the latex has been concentrated has
32 Natural latices
periodically caused problems; the deposits which form have been termed
'magnesium crystallites'.
For the precipitation of magnesium ammonium phosphate to occur, it is
necessary that sufficient orthophosphate anions be present in the latex aqueous
phase. These anions are principally derived by alkaline hydrolysis of phospholipids
present in the fresh latex (see further section 9.5.5 below). Some natural rubber
latices do not contain sufficient phosphate to enable the magnesium to be
completely precipitated in this way. The magnesium/phosphate ratio is to some
extent a clonal characteristic, although there can be large variations within a
given clone. Some clones, notably Glenshiel I, give a latex for which this ratio
is high. The ammonia-preserved latex concentrate tends to be of low and
variable colloid stability because of the relatively high concentration of residual
magnesium ions. The mechanical stability ofthe concentrate is low, its viscosity
at a given solids content is high, and its gelling pH is high. These abnormalities,
which are accentuated during the wintering period, are unfortunate because
the Glenshiel I clone is particularly high-yielding.
Adverse magnesium/phosphate ratios are usually the consequence of an
excess of magnesium, rather than of a deficiency of phosphate. Indeed, it is
significant that the abnormal latices are not deficient in the phospholipids
which are believed to be the principal source of orthophosphate anions.
Madge, Collier and Peel [20] have shown that small additions of water-soluble
phosphates, such as diammonium hydrogen phosphate or sodium pyrophos-
phate, improve the mechanical stability of Glenshiel I latex, up to the point
where the added phosphate is roughly equivalent to the residual magnesium.
Further additions of water-soluble phosphate cause the mechanical stability
to decrease again, presumably because the ionic strength of the aqueous
phase is increased. The addition of water-soluble phosphates to a normal
natural rubber latex concentrate brings about an immediate reduction in
mechanical stability. Figure 9.10 shows results reported by Madge, Collier
and Peel illustrating these effects. Addition of water-soluble phosphates to
Glenshiel I latex also reduces the gelling pH to a normal value. Resing [21]
has described a procedure for the treatment of natural rubber latices which
have a high magnesium/phosphate ratio. A sample of the field latex is
ammoniated to 0.6% m/m, allowed to stand for 3 hours, and then centrifuged
for 30 minutes at a rotational frequency of 2000 minute - 1. A 1-g sample of
the centrifugate, free of sludge, is diluted with lOOcm 3 of water, and the
residual magnesium content estimated by titration with disodium dihydrogen
ethylenediaminetetraacetate in the presence of a buffer, using Erichrome
Black T as indicator. Water-soluble phosphate is then added to the remainder
of the latex in slight excess of the requirement evidenced by this titration.
9.3.3.3.4 Ammonia and the ancillary requirements

Ammonia is not especially harmful to people if reasonable precautions are
taken, unless it is inhaled in excessive quantities, or unless a concentrated
600
500
-E
III
CD
:;:;
2 4 6 8 10
Level of added diammonium hydrogen
orthophosphate I g (dm3 latexr 1
Figure 9.10 Effect of addition of water-soluble phosphates upon mechanical stability of

Glenshiel I (curve A) and normal natural rubber latex concentrate (curve 8) (Madge. Collier and
Peel [20]).
aqueous solution comes into contact with the skin. It appears to have no
effect upon natural rubber, although it does cause important changes to some
of the non-rubber constituents of natural rubber latex (see section 9.5.5 below).
It imparts no colour to the latex, or to deposits obtained from the latex,
from which it is eliminated during drying, but it does impart a strong odour.
Ammonia has to be handled either as cylinders of the anhydrous liquid, or
as concentrated aqueous solutions. Neither alternative is very convenient.
Because of the handling difficulties, ammonia is not particularly cheap. A
major disadvantage of ammonia as a preservative for natural rubber latex
is that it interferes with some established latex processes. Thus ammonia-
preserved natural rubber latex undergoes a marked and progressive thickening
when compounded with sparingly-soluble zinc compounds, notably zinc
oxide. After a sufficient time has elapsed, the latex may gel. The rate of this
thickening, and of the subsequent gelation, increase markedly with increasing
temperature. These effects, and possible mechanistic explanations, are discussed
in section 5.3.2 of Chapter 5 (Volume 1). Ammonia also interferes with the
34 Natural latices
gelation of natural rubber latex by silicofluorides (see section 5.4.2 of Chapter

5 (Volume 1». This is because the ammonia provides a source of hydroxide
ions which have to be removed by reaction with the hydrogen ions generated
by the hydrolysis of silicofluoride anions before the pH can be reduced to a
value at which gelation of the latex can occur. It may therefore be necessary
for the ammonia content of ammonia-preserved natural rubber latex to be
reduced to a low level prior to the use of silicofluorides as delayed-action
coacervants.
It is appropriate to refer at this point to the methods which are used for
partially de-ammoniating ammonia-preserved natural rubber latex. The most
usual method is aeration. A stream of warm air is passed over the surface of
the latex, the bulk of which is gently agitated so that fresh latex is continually
being brought to the surface. A type of vessel used for this purpose is illustrated
schematically in Fig. 9.11. Gradually the ammonia is swept away by the air
stream. Partial de-ammoniation can also be effected chemically. Of course, the
ammonia can be removed by addition of acid. However, this is impractical,
being accompanied by the two-fold disadvantage of converting the ammonia
to ammonium ions and increasing the ionic strength of the aqueous phase, as
well as with the risk of localized coagulation occurring. A more satisfactory
chemical method makes use of the quantitative reaction between ammonia and
formaldehyde to form hexamethylenetetramine according to the following reaction:
4NH3 + 6H.CHO--+(CH2)6N4 + 6H 20
The amount of formaldehyde required to remove the requisite amount of
ammonia is calculated according to this stoichiometry. The reaction is not
Warmair--...J
Latex
Figure 9.11 Typical equipment for partial de-ammoniation of ammonia-preserved natural

rubber latex by aeration.
instantaneous, taking a few minutes to go to effective completion. Excessive

amounts of formaldehyde should be avoided, as they cause thickening and
gelation of the latex. Hexamethylenetetramine is un-ionized and neutral in
reaction. Although it is an accelerator of the sulphur-vulcanization of natural
rubber at high temperatures, it has but a slight activation effect upon the
sulphur-vulcanization of deposits from natural rubber latex at the temperatures
normally used for such vulcanization. This method of de-ammoniating
natural rubber latex is not extensively used industrially, partly because of
health hazards associated with formaldehyde and with hexamethylene-
tetramine. It is useful in laboratory work. The heat generated by the reaction
between formaldehyde and ammonia has been used to warm large quantities
of natural rubber latex during cold weather.
9.3.4 Low-ammonia preservation systems
9.3.4.1 Rationale of low-ammonia preservation systems

In view of the several disadvantages which accompany the use of ammonia
as a preservative for natural rubber latex, a continual search has been made
for satisfactory preservatives which are not subject to these disadvantages
or any others. Many alternatives to ammonia have been considered, but the
prevailing view remains that, notwithstanding its disadvantages, ammonia
is so generally satisfactory that it is unlikely to be entirely replaced in the
foreseeable future. Thus McGavack [22], writing in t 959, has estimated that
more than 95% of all imported natural rubber latex contains some ammonia.
This probably remains essentially so more than 30 years later. The trend
since the end of World War II has been to reduce the level of ammoniation
rather than to eliminate ammonia altogether, and to supplement its efficacy
by incorporating a secondary preservative. Several secondary preservatives
are used for this purpose. Their main function is to supplement the bactericidal
effect of ammonia. Natural rubber latices preserved in this way are known
as low-ammonia (LA) preservates, to distinguish them from conventional
ammonia-preserved latices, which are now known as high-ammonia (HA)
preservates. The principal secondary preservatives which have found application
in natural rubber latex include pentachlorophenates, zinc dialkyldithiocar-
bamates, thiuram sulphides, aminophenols and boric acid. Typically, the
ammonia content of the latex is reduced to ca. 0.2% mjm on the whole latex,
and the amount of secondary preservative is also ca. 0.2% mjm on the whole
latex.
9.3.4.2 Pentachlorophenate latices

Pentachlorophenol (Structure IV) was first proposed as a preservative for
natural rubber latex in t 938. It and its alkali-metal salts, notably sodium
36 Natural latices
pentachlorophenate, have found some application as secondary preservatives

in conjunction with ammonia. They are excellent bactericides. A typical
combination comprises 0.2% m/m ammonia and 0.2% m/m sodium penta-
chlorophenate on the whole latex. Although pentachlorophenate latices
have found considerable use in the past, they are little used now because
of the unpleasant and hazardous nature of pentachlorophenol and its salts.
Not only are they poisonous; their dusts produce unpleasant effects if
inhaled. Furthermore, they may cause dermatitis with sensitive skins.
Because of their toxicity to other life forms, they also present problems of
effluent disposal.
OH
CI CI
CI CI
CI
IV
Pentachlorophenate latices tend to be rather stable colloidally. This is

evidenced by, for example, high mechanical stability. The enhanced colloid
stability may in part be a consequence of the additional alkali which is added
with the pentachlorophenate, and in part to adsorption of pentachlorophenate
anions at the surface of the rubber particles, thereby enhancing the surface
electric charge density. Howev~r, the second of these explanations does not
accord with the observation of Bowler [23] that sodium pentachlorophenate
has no effect upon the relationship between electrophoretic mobility and pH
for natural rubber latices which contain ammonia.
Deposits dried down from pentachlorophenate latices discolour to some
extent, especially on ageing. Deposits from latices which contain the sodium
salt are said to be more prone to discoloration than are those from latices
which contain the ammonium salt. The explanation is presumably that, in
the former case, the pentachlorophenate anion remains as such in the dried
deposit, whereas in the latter case ammonia gradually volatilizes from the
deposit leaving the pentachlorophenate in the form offree pentachlorophenol.
It is well known that phenate anions tend to discolour more readily on
exposure to oxygen than do the corresponding phenols.
It is likely that the pentachlorophenate anion has a slight sequestering
capacity for multivalent metal ions. It has been proposed to enhance this
attribute of the ammonia-pentachlorophenate preservation system by the
incorpration of additional sequestering agents. A typical combination of this
type comprises 0.1 % m/m ammonia together with 0.1 % m/m sodium penta-
chlorophenate and 0.1 % m/m ethylenediaminetetraacetic acid. It is possible
that, in the proportions used for the preservation of natural rubber latex,
pentachlorophenol and its salts may have a slight adverse effect upon the
ageing of deposits from the latex. Many polychloro aromatic compounds do
have an adverse effect in this respect.
9.3.4.3 Zinc dialkyldithiocarbamate latices

Zinc dialkyldithiocarbamates have the general structure shown below as V,
in which R represents an alkyl group. The nature of the bonding between
the zinc atom and the dialkyldithiocarbamate moieties is uncertain, but
probably involves the formation of chelate rings, so that these compounds
are probably not ionic. The ability of zinc dimethyldithiocarbamate to
preserve natural rubber latex was first described by Townsend [24] in 1958.
A typical combination comprises 0.2% mlm ammonia together with
0.1-0.2% mlm zinc dialkyldithiocarbamate and 0.2% mlm lauric acid. The
R, /R
N-C-S-Zn-S-C-N
R/ II II 'R
S S
v
addition of lauric acid is optional. It forms laurate anions in the alkaline

aqueous phase, and these anions enhance the colloid stability to a level
similar to that for a normal ammonia-preserved natural rubber latex. The
sulphur-vulcanization behaviour of this type of latex, and of deposits dried
down from the latex, is only slightly affected by the presence of the
dithiocarbamate, notwithstanding that zinc dialkyldithiocarbamates are
powerful accelerators of the sulphur-vulcanization of diene rubbers (see
section 16.3.2 of Chapter 16 (Volume 3».
9.3 .4.4 Aminophenol-ethylenediaminetetraacetate latices

In this type of low-ammonia natural rubber latex, an amino phenol, such as
tri(dimethylaminoethyl)phenol, is used as the principal bactericide in
combination with ca. 0.2% mlm ammonia. Disodium dihydrogen
ethylenediaminetetraacetate is added to sequester multivalent metal ions.
Little interest has been shown in this method of preserving natural rubber
latex. Deposits from this type oflatex discolour very badly on ageing, although
it has been claimed that discoloration can be prevented by compounding
with zinc oxide.
9.3.4.5 Boric acid latices

Boric acid was first proposed as a secondary preservative for natural rubber
latex by Sekaran and Morris in a patent published in 1960 [25]. A typical
38 Natural latices
combination comprises 0.2% mjm ammonia together with 0.2% mjm boric
acid and 0.05% mjm lauric acid. As in zinc dialkyldithiocarbamate latices,
the latter forms laurate anions in the alkaline aqueous phase, and these
anions enhance the colloid stability to a level similar to that for a normal
ammonia-preserved natural rubber latex. In particular, the laurate anions
counteract any slight destabilizing effect which the boric acid may have upon
the latex.
Of the various low-ammonia preservation systems for natural rubber latex
which have been described so far, that using boric acid as a secondary
preservative would appear to be the most promising. Deposits from latices
preserved in this way are remarkably free from discoloration on ageing:
indeed, there is some evidence that boric acid may suppress some types of
discoloration. Boric acid is an excellent bactericide. It has little, if any, effect
upon vulcanized natural rubber, although there have been cases reported
where unvulcanized deposits from boric acid latices have softened rapidly.
This effect is not understood, but substances such as thiourea have been
found to be effective as preventatives. Boric acid is cheap and convenient to
handle, being a crystalline solid. It might be thought that boric acid would
have an excellent record in respect of harmlessness toward people. However,
this appears not to be so, and this is the reason why boric acid latices have
not been as widely used as might otherwise have been expected.
9.3.4.6 Tetraalkylthiuram sulphide latices

The tetraalkylthiuram sulphides have the general structure shown below as
VI, in which R represents an alkyl group and n can have an integral value
from t upwards. n is normally t, 2 or 4, the compound then being a thiuram
monosulphide, disulphide or tetrasulphide respectively. It has been known
VI
for many years that tetramethylthiuram disulphide (R = CH 3' n = 2) (TMTD)

can preserve natural rubber latex if small amounts of ammonia are also
present. A combination of this compound and less than 0.3% mjm ammonia
is described by Ten Broeck in a patent published in t 959 [26]. Since that
time, a preservation system comprising 0.2% mjm ammonia together with
less than 0.05% mjm tetramethylthiuram disulphide and not more than
0.02% mjm zinc oxide has been developed. Natural rubber latex concentrates
preserved by this type of combination are known as LAZN or LATZ
concentrates. Regular production and shipment of this type of latex from
Malaysia commenced in 1976. Since that time, LAZN natural rubber latex
has become established as the principal type of low-ammonia natural rubber
latex. The properties of LAZN natural rubber latex have been described by
Calvert, Sundaram and Tan [27]. It appears that the presence of the small
amount of zinc oxide in this type of latex is essential for effective preservation;
by themselves, tetramethylthiuram disulphide and a low level of ammonia
do not provide adequate preservation for industrial purposes. The presence
of zinc ions from the zinc oxide inevitably has a slight detrimental effect
upon colloid stability, but this is not serious. The zinc oxide cannot
satisfactorily be replaced by zinc sulphide, probably because of the lower
solubility of the latter in aqueous ammonia solutions.
Calvert et at. [28] have described variants of LAZN natural rubber latex
which are known as ultra low-ammonia latices: The objective in developing
these latices was to reduce the ammonia level to the lowest possible value
consistent with adequate preservation. Two representative types of ultra
low-ammonia latex are described. The first, known as ULAZN latex, contains
0.10% mlm ammonia together with less than 0.08% mlm tetramethylthiuram
disulphide and 0.02% mlm zinc. In effect, the level of thiuram sulphide has
been increased to compensate for the reduced ammonia content. In the second
type, known as KLAZN latex, the ammonia content is reduced to 0.05% mlm,
the tetramethylthiuram disulphide level is now less than 0.10% mlm, the zinc
content remains at 0.02% mlm, and potassium hydroxide is present at a level
of 0.10% m/m. The latter is required to supplement the alkalinity of the latex
in order to ensure that the phospholipids in the fresh latex are hydrolysed
with the formation of carboxylate soaps to enhance mechanical stability, and
also of orthophosphate anions to precipitate magnesium ions as magnesium
ammonium phosphate. It appears that it is not possible at present to produce
a satisfactory natural rubber latex preservate of this type which is entirely
free of ammonia. A small amount of ammonia is necessary to effect
preservation of the latex in the field, and also to precipitate magnesium ions.
9.3.4.7 Other low-ammonia preservation systems

The results of an extensive investigation of various other low-ammonia
preservation systems for natural rubber latex has been reported by Angove
and Pillai [29-33]. The compounds investigated as secondary preservatives
included certain metal and halogen derivatives of 8-hydroxyquinoline, and
various organic compounds of the types used as accelerators of the sulphur-
vulcanization of diene rubbers. Some of these compounds were also investigated
as tertiary preservatives used at low levels in conjunction with ammonia as
primary preservative and zinc oxide as secondary preservative. In a further
paper, Angove [34] has given an assessment offour promising potentially-cheap
preservation systems as indicated by this investigation. Many other publications
40 Natural latices
provide information concerning low-ammonia preservation systems for

natural rubber latex. These include papers by Clough and Glymph [35], by
John et al. [36], by Smith et al. [37], and by John et al. [38].
9.3.5 Other preservatives for natural rubber latex
9.3.5.1 Formaldehyde and ammonia

Formaldehyde is an excellent bactericide, being superior to ammonia in this
respect. A combination of a formaldehyde pre-treatment to sterilize the latex,
followed by ammoniation, is the basis of the so-called McGavack system of
preservation [22]. Typically, fresh natural rubber latex is treated with
0.15-0.2% mlm formaldehyde at the collecting station, as soon as possible
after tapping. This quickly kills all the bacteria. The latex is then allowed to
stand for a period of 15-30 hours at the factory, during which time the
formaldehyde interacts with the proteins in the latex, and a consequent fall
in pH occurs. Ammonia is added to bring the pH up to a little above 6. The
latex may be concentrated by centrifugation at this stage if desired. Finally,
the ammonia level is brought up to 2% mlm on the aqueous phase of the
latex. The product of this process is claimed to have enhanced colloid stability
and resistance to freezing, because of the reaction which takes place between
the latex proteins and the formaldehyde. It is not clear why this should be
so. Further information concerning the use of formaldehyde as a preservative
for natural rubber latex concentrate is available in a paper by Nadarajah et
al. [39].
Formaldehyde has also been used as an anticoagulant at the cup and
bucket stage. The usual amount used for this purpose is ca. 0.02% mlm added
as 1% mlv aqueous solution. It has been claimed that formaldehyde is
unsatisfactory as a field preservative for natural rubber latex which is to be
concentrated for subsequent use in certain processes for the manufacture of
latex foam rubber. It appears that the latex proteins become denatured by
the formaldehyde, thereby impairing the foaming characteristics of the
latex.
9.3.5.2 Sodium sulphite

Sodium sulphite is used as an anticoagulant at the cup and bucket stage,
especially when the latex is to be used for the production of a form of dry
natural rubber known as pale crepe. As the name of this product implies, it
is essential to keep discoloration of this type of natural rubber to a minimum.
In this connection, it may be noted that sodium bisulphite is used as an
enzyme inhibitor in the manufacture of pale crepe, the objective again being
to minimize discoloration. Ca. 0.05% mlm of sodium sulphite on the whole
latex is recommended as an anticoagulant, added as a 3% m/v aqueous

solution. Of this amount, a few drops are added in the cup at the time of
tapping, and the balance is added to the buckets during collection.
9.3.5.3 Potassium hydroxide

The only practically-important alternative to ammonia as a sole preservative
for natural rubber latex is potassium hydroxide. This is used for the
preservation of 75% m/m evaporated natural rubber latex concentrate (see
section 9.4.2 below). It is an effective bactericide by virtue of its high alkalinity.
For the same reason, it gives latex which is colloidally very stable. The
problem of zinc oxide thickening is entirely eliminated.
9.3.5.4 Other proposed preservatives for natural rubber latex

Many other compounds have been investigated as possible preservatives for
natural rubber latex over the years. Thus some interest has been shown in
the use of certain antibiotics as preservatives for natural rubber latex, either
with or without ammonia. Thus the use of certain thiobis halogenophenols,
notably 2,2'-thiobis-(4,6-dichlorophenol) (Structure VII) as a preservative for
natural rubber latex has been patented by Graham and Taysum [40]. At
least 0.001% m/m on the whole latex is added, either in the form of an
aqueous emulsion or dispersion, or preferably as an aqueous solution of an
alkali-metal or ammonium salt. These compounds are claimed to show
excellent anticoagulant action, and not to cause discoloration of either the
latex or solid deposits obtained from the latex. Amongst the more exotic
possibilities which have been seriously considered for this application in
recent years is cis-l-(3-chloroallyl)-3,5, 7-triaza-l-azoniaadamantane chloride
(Structure VIII). It has been investigated both with and without ammonia
as primary preservative. Although an excellent microbicide, it appears to be
toxic to an insufficiently wide range of microorganisms for this application.
Information concerning the use of this substance as a secondary preservative
with ammonia is available in a paper by John et al. [41].
CI CI
CI CI
VII VIII
42 Natural latices
9.4 CONCENTRATION OF NATURAL RUBBER LATEX
9.4.1 Preliminary considerations
9.4.1.1 Why concentration is necessary

Natural rubber latex exudes from the Hevea brasiliensis tree having a rubber
content of ca. 30-40% mlm, the average being ca. 33%. The remainder of
the latex is mainly water. There are four principal reasons why it has become
common practice to concentrate natural rubber latex before transportation:
I. It is uneconomical to transport preserved field natural rubber latex over
large distances, because this would necessitate transporting a mass of
water equal to approximately twice the mass of the dry rubber.
2. Several of the industrial processes which use natural rubber latex require
it in a more concentrated form than ca. 33%. In particular, this is true of
processes for the manufacture of latex foam rubber.
3. Natural rubber latex concentrates tend to be more uniform in quality
than the field latices from which they were obtained. This is mainly a
consequence of the bulking and concomitant blending which precede and
follow the concentration process. Differences between batches offield latex
therefore tend to be mitigated.
4. The ratio of non-aqueous non-rubber substances to dry rubber is reduced
in the course of some of the concentration processes, because the
non-rubber substances are preferentially eliminated (see section 9.4.1.2
below). The product is in consequence a closer approximation to a colloidal
dispersion of natural rubber in water than is preserved field natural rubber
latex. As such, it more closely approximates to what is perceived to be
ideally required in several industrial processes.
The normal practice is to concentrate the latex so that the dry rubber
content is ca. 60% mlm or above. Natural rubber in latex form is then
substantially more economically attractive for transportation, and the
concentrate is more suitable for industrial processes than is preserved field
latex, for the reasons set out above. Because of these considerations, there
is now very little industrial interest in the bulking and shipment of preserved
field natural rubber latex.
9.4.1.2 Outline of principal methods for concentrating natural rubber latex

Of the various methods which have been proposed for concentrating natural
rubber latex, four have received serious attention as practicable processes.
These are evaporation, creaming, centrifugation and electrodecantation. The
first of these methods involves the removal of water only. Hence the ratio
of non-aqueous non-rubber substances to dry rubber remains unaffected.
The particle-size distribution of the latex is also virtually unaffected. By
Concentration of natural rubber latex 43
contrast, the other three methods reduce the ratio of non-aqueous non-rubber
substances to dry rubber and also affect the particle-size distribution of the
latex because a proportion of the smaller particles is eliminated during these
processes. The preferential elimination of non-aqueous non-rubber substances
is partly a consequence of elimination of part of the aqueous phase, and
partly a consequence of elimination of some of the smaller rubber particles.
The specific surface area of the particles increases as the particle size decreases
(see section 2.2.2.3 of Chapter 2 (Volume 1)). Thus, for a given extent of
surface coverage, the ratio of adsorbed substances to rubber increases as the
particle size decreases. The preferential elimination of the smaller particles
therefore results in the preferential elimination of adsorbed non-aqueous
non-rubber substances, additional to the partial elimination of such substances
as may be dissolved in the aqueous phase of the latex.
9.4.2 Concentration by evaporation

The earliest process for concentrating natural rubber latex by evaporation
appears to have been that patented by the KDP Ltd [42-44]. One particular
way of carrying out this process is illustrated schematically in Fig. 9.12. The
evaporation takes place within the inner vessel of a horizontal double-walled
cylindrical drum which revolves about its own axis. The latex is heated by
means of hot water circulating in the space between the two walls. The water
is heated externally, e.g., by an oil-fired furnace, and is recycled to the boiler
for further use. A hollow iron roller, with its axis parallel to that of the
cylinder, is placed inside the inner vessel and allowed to rotate freely as the
drum revolves. Its function is to homogenize the latex, prevent skin formation,
destroy froth, and increase the surface area of latex from which evaporation
can occur. During the evaporation, a current of cold air is drawn through
the cylinder in order to remove water vapour. Evaporation takes place directly
from the bulk of the latex and also from the thin films of latex on the walls
of the vessel and on the roller as vessel and roller rotate. An essential feature
JJr;:================~",~( wate, 1-'

Air-stream
Figure 9.12 Schematic illustration of typical equipment for concentrating natural rubber latex
by evaporation, as envisaged by British Patent No. 244,727 to KDP Ltd
44 Natural latices
of the process is that fresh portions of latex are continuously brought to the
surfaces at which evaporation is occurring, thereby minimizing skin and
coagulum formation.
In a process developed by Metallgesellschaft Aktiengesellschaft [45], the
latex is heated and sprayed into a partially-evacuated chamber, so that the
ambient pressure is always less than the saturated vapour pressure of water
at the temperature to which the latex has been heated. In consequence, rapid
evaporation of moisture occurs as the latex leaves the spray. The operation
can be repeated if the required degree of concentration is not attained in one
operation. Little or no frothing occurs during evaporation. The latex may
be colloidally stabilized with hydrocolloids prior to eva~oration, and
preservatives may also be added prior to evaporation. In a typical example
of this process, the latex to be concentrated is heated to 90°C and introduced
through suitable nozzles into an evaporation chamber in which the air
pressure has been reduced to ca. 10 kPa. The corresponding boiling-point of
water at such a pressure is ca. 46°C. At the moment of entering the evaporator,
a rapid fall in temperature of about 44°C occurs, and at the same time the
water content of the latex is reduced by ca. 7%. In this case, the operation
must be repeated ca. 11 times if it is required to increase the total solids
content of the latex from 35% to 75% m/m. A suitable design of nozzle is
one which comprises convergent and divergent sections. The narrowest part
is constructed to so throttle the flow of latex that no evaporation commences
in the convergent parts of the nozzle. The divergent part of the nozzle is
designed to facilitate a gradual reduction in pressure as the latex enters the
evaporator.
The Metallgesellschaft Aktiengesellschaft process [45] suffers from a
serious disadvantage, in that the vacuum which prevails in the evaporator
chamber must be released every time the finished concentrate is withdrawn.
In an improvement of the method, described by Quasebarth [46], a plurality
of reservoir tanks is employed in conjunction with a single supply of latex
to be fed to the evaporator, and as a reservoir to receive concentrate returning
from the evaporator. Circulation of the latex can be effected either by pump
or by differential air pressure.
In recent years, much evaporated natural rubber latex has been produced
using turbulent-film evaporators of the Luwa type. This type of evaporator
is described in section 10.4.3.2 of Chapter lOin connection with the
concentration of synthetic latices by evaporation.
9.4.3 Concentration by creaming
9.4.3.1 Elementary theory of creaming

Particles dispersed in a fluid medium and subjected to a gravitational field
have a tendency to move relative to the dispersion medium if the density of
the particles, p, differs from that of the dispersion medium, (1. If p is less than
(1,as it is for natural rubber latex, then the particles tend to rise to the surface
of the dispersion medium. This process is known as creaming, and the
concentrate as cream, by analogy with the process by which the fat globules
in milk concentrate at the surface to form cream. The dilute latex which
forms as the lower layer when a latex creams is known as skim. If p is greater
than (1, then the particles tend to fall to the bottom of the dispersion medium.
This process is known as sedimentation.
The processes of creaming and sedimentation are opposed by the thermal
dispersive motion imposed upon the particles by random collisions with the
molecules of the dispersion medium. The effects of these collisions are evident
as the Brownian motion of the dispersed particl~s. Eventually, an equilibrium
variation of particle concentration with height is established between the
opposing tendencies of separation and dispersion. The nature of this
equilibrium variation can be predicted as a special case of the Boltzmann
distribution. This distribution gives the ratio, c2 1c l ' of the equilibrium
concentrations of a given particulate species at two locations I and 2 as
e- W /kT , where W is the work which has to be done to move a particle
reversibly from location t to location 2, T is the absolute temperature, and
k is the Boltzmann constant, there being a proviso that the particles are free
to move between the two locations. For a colloidal dispersion which creams
or sediments under the influence of gravity alone, W for locations at zero
height and height h is equal to mgh, where m is the effective mass of the
particle. For a spherical particle of diameter x and density p immersed in a
dispersion medium of density (1, m = 1tX 3 (p - (1)/6. The equilibrium particle
concentration, c, at height h is then related to that, co' at zero height by the
equation
_ [ (1)9h]
1tX 3(p -
(9.1)
c - coexp - 6kT
This equation shows that the extent of the sedimentation or creaming, as

quantified by the ratio clco at any particular height, depends upon the
parameter x 3 (p - (1). Thus if h > 0, then c < Co if p > (1 (i.e., the particles
sediment), and c < Co if, as for natural rubber latex, p < (1 (i.e., the particles
cream). Furthermore, the ratio clco at any particular height increases sharply
with particle size. Figure 9.13 shows predictions given by equation (9.1) for
the variation of the ratio clco with h at 25°C for a dispersion for which
p = 0.950Mgm- 3 and (1 = t.OOOMgm- 3 . These density values are similar
to those for the particles and dispersion medium of natural rubber latex.
Curves are shown for particles of three diameters: 50 nm, tOO nm and t 50 nm.
However, it should be noted that equation (9. t) is derived assuming that the
only contribution to W is that from the difference in the gravitational
potentials at the two locations. In practice, as c increases, so the particles
46 Natural latices
--rf
u
°O~----~-----2~----~3~--~4
hlcm
Figure 9.13 Predictions given by equation (9.1) for variation of the ratio clco with h at 2S'C
for dispersions of monodisperse spherical particles havinp diameters 50nm (curve A), l00nm
(curve B) and 150nm (curve q, taking p = 0.950 Mgm- and (J = 1.000Mgm- 1 .
become increasingly crowded, the interparticle separation decreases, and

contributions to W from those factors which confer colloid stability upon
the dispersion become increasingly important. The increase of c with h for
latices having large particle concentrations is therefore expected to be less
than that predicted by equation (9.1).
The considerations of the preceding paragraph relate only to the variation
of equilibrium particle concentration with height in a creaming process. Of
equal importance for industrial concentration processes is the rate at which
creaming occurs. The average steady speed at which a particle moves through
the dispersion medium depends upon the balance between the gravitational
force which tends to accelerate motion, and the force of viscous drag which
tends to retard motion. For a spherical particle of diameter x, the gravitational
force is 1tX 3 (p - u)g/6. The magnitude of the force of viscous drag is
approximately 31t"x(dh/dt), where" is the viscosity of the dispersion medium,
dh/dt being the speed at which the particle is moving upward through the
dispersion medium. Thus the steady value of dh/dt is such that
1tX 3 (p - u)g = -31tnx dh
6 'r dt (9.2)
This gives
dh (p - u)g 2
-=- x (9.3)
dt 18"
This equation shows that the steady speed of creaming is directly proportional
to the density difference (p - 0). The sign of this difference determines the
direction of movement of the particle in the dispersion medium; dh/dt is
negative or positive according as p is greater or less than u. For particles of
natural rubber in an aqueous phase, p is less than u, and so dh/dt is positive,
confirming that the particles move upwards through the aqueous phase.
Equation (9.3) also shows that the speed at which the particle moves through
the dispersion medium is directly proportional to the square of the particle
diameter, and inversely proportional to the viscosity of the dispersion
medium. Thus rapid creaming is favoured by large particles, low viscosity
of the dispersion medium, and a large density difference between particles
and dispersion medium. For an aqueous dispersion for which,., = 1 mPa s,
p = 0.950Mgm- 3 and u = 1.000 Mgm- 3 , the factor (p - u)g/18,., is ca.
- 2.72 x 104 m -1 s - 1. The predicted steady speeds at which particles having
diameters 50nm, l00nm and 150nm rise through the aqueous phase are
respectively ca. 6.8 x 10- 11 ms- I , 2.7 x 1O- iO ms- I , and 6.1 x 1O- iO ms- I ,
i.e., ca. 2.5 x 1O- 4 mmhour- I , 1.0 x 1O- 3 mmhour- I and 2.2 x 10- 3
mm hour -1. The creaming process is thus predicted to be slow, at least for
these dispersions.
It follows from the considerations of the preceding paragraph that the
duration of creaming can affect the particle-size distribution in both the
cream and the skim fractions. The following approximate analysis indicates
the nature of the effects to be expected. Suppose that a time t has elapsed
since creaming commenced. For simplicity, assume that throughout this time
all the particles rise through the aqueous phase at a steady speed dh/dt which
depends only upon the particle diameter, and which is given by equation
(9.3) above. Then a particle which creams at a particular speed dh/dt can
have moved, at the most, a vertical distance equal to t(dh/dt). Consider now
the distribution of particle diameters which will be present at the time t in
the region of thickness bh between layers of the latex at heights hand h + bh
above the bottom of the container, these layers being at a location which is
intermediate between the bottom of the container and the upper surface of
the latex. The distribution will be determined by the numbers of the various
particles which have moved upwards into this region during the time t, and
the numbers which haved moved upwards out of that region in the same
time. The distribution of diameters for those particles having diameters such
that t(dh/dt) ~ h will be identical to that for the initial latex, because for
every particle of this type which is lost from the region by upward motion
from the region, a particle of identical size will have been gained by similar
upward motion into the region. But particles having diameters such that
t(dh/dt) > h will be completely absent from the region ofthickness bh, because
all particles of this type which were initially below this region will have
moved above it during the time t. Let X h denote the diameter of the largest
particle which remains in the region of thickness bh at the level h after
creaming has continued for time t. Then the region above the level t, including
48 Natural latices
that which is regarded as comprising cream, is relatively rich in particles

having diameters which exceed X h • The value of X h is that for which
t(dhldt) = h. It follows from this, and from equation (9.3), that X h is given
by
18'1h
(9.4)
(0" - p)gt
Thus X h is predicted to be directly proportional to h l/2 and inversely

proportional to t 1/2 • These considerations lead to the conclusion that the
cream layer will contain a wide distribution of particle sizes, whereas the
particle-size distribution for the skim will have a sharp cut-off above a certain
particle diameter which depends upon the length of time for which creaming
has occurred.
9.4.3.2 Creaming in practice

In practice, use is made of the observation that the rate of creaming of natural
rubber latex can be markedly increased by small additions of water-soluble
hydrocolloids to the latex. Substances used in this way are known as creaming
agents. A typical example is ammonium alginate. In addition to hydrocolloids,
small amounts of carboxylate soaps may be added further to promote
creaming. In a typical process, field latex is bulked in reception tanks and
is ammoniated to ca. 1% m/m. The creaming agent is then added as an
aqueous solution to the extent of ca. 0.1 % mlm on the latex aqueous phase.
The latex is then stirred for ca. 1 hour, after which it is left to stand for ca.
40 hours, during which time creaming proceeds to near completion. The skim
is then run off, and the cream ammoniated to 0.8% mlm on the whole latex.
The cream is next transferred to bulking tanks which hold several days'
supply. Creaming is completed in these tanks. The uniformity of the product
is improved by blending together several batches. After the removal of any
further skim which has separated, the concentrate is gently stirred for 1-2
days to ensure homogeneity. It is then ready for transportation.
Concentration by creaming has several important advantages over other
methods. The requisite equipment is simple and easy to operate; power
requirements are negligible; and the proportion of rubber lost in the skim is
small. The disadvantage of the method is that it is slow, it is sensitive to
variations in the field latex, and after-creaming inevitably occurs during
storage and transportation. Furthermore, the presence of the residual
creaming agent in the latex may affect both the colloidal properties of the
latex and the properties of films obtained from the latex. For these reasons,
the market preference for many years has been for natural rubber latex
concentrate produced by centrifugation rather than by creaming.
9.4.3.3 Mode of action of creaming agents
9.4.3.3.1 The facts to be explained

There has been controversy concerning the mechanism by which creaming
agents accelerate the creaming process. Although the precise way in which
these substances are effective remains obscure, it is easier to understand how
they work now than in former times, because of our greatly increased
understanding of the manner in which lyophilic macromolecules adsorb at
the surface of the particles in lyophobic sols, and of the likely effects of such
adsorbed macromolecules upon the interactions between the particles.
The facts to be explained by a satisfactory theory of the action of creaming
agents can be summarized as follows.
1. When ammonia-preserved field natural rubber latex which contains a
creaming agent is allowed to stand, a slowly-increasing volume of skim
separates in the form of a lower layer. This process has the appearance
that the skim is gradually exuding from the bulk of the latex, i.e., that the
skim has been forced to separate from the latex because of interactions
between the latex particles, thereby causing the formation of a concentrated
cream. The skim contains a small but definite and constant concentration
of rubber particles, which depends upon the amount of creaming agent
present. The boundary separating the skim and cream gradually moves
upwards, so that the volume of cream decreases progressively whilst its
rubber content increases progressively, again confirming the impression
that the skim is being gradually expelled from the latex. The separated
skim shows no tendency to cream further, although the cream fraction
may continue to exude skim for several days. These characteristics contrast
with those observed when certain other lyophobic colloid systems are
allowed to cream. Thus in the case of milk, the volume of cream which
collects as an upper layer gradually increases, and the boundary between
skim and cream gradually moves downwards. In this case, it is the cream
rather than the skim which appears to exude during the creaming process.
2. The progressive separation of skim follows a fairly well-defined pattern.
This is illustrated by the results reported by Twiss and Carpenter [47],
reproduced here as Fig. 9.14(a). These workers proposed the following
relationship for the variation of the fractional volume of skim, 4>, with the
time of creaming, t:
(9.5)
In this equation, to is an induction period, A. and b are constants which

depend upon the creaming agent, a is a characteristic constant for the
latex, and ~ is the concentration of creaming agent in the aqueous phase
of the latex. It is claimed that this equation satisfactorily describes the
observed variation of 4> with t. In particular, it correctly predicts the
50 Natural latices
60
A
B
~ 40
#. C
.......
CD
E
::I 0
g
E 20
32
en
00
20 40 60 80 120
(a) Time/hour
14
~
:.e
• 12
~
E
0
....... :.e
10 .::...
E
IU E
!!! 32
f/l
t>
'0 8 '0
'E 'E
Sc: 6 ~0
0 t>
t>
Qj Q;
.c 4 .c
.c .c
2 2
~ ~
0 10 2 0
00 0
0.1 0.2 0.3 0.4 0.5 0.6
Concentration of creaming agent on aqueous phase / % m/v
(b)
Figure 9.14 (a) Progressive separation of skim during creaming of natural rubber latex (Twiss
and Carpenter [47]). The creaming agent was methylcellulose, the concentrations (m/m on
aqueous phase) being 0.120"10 (curve A), 0.366% (curve B), 0.458% (curve C) and 0.715% (curve
D). (b) Effect of concentration of creaming agent upon rubber content of cream and skim formed
by creaming of natural rubber latex (Hauser and Dewey [48]). The creaming agents were
ammonium alginate and locust bean gum. Significance of points for ammonium alginate:
• (cream), 0 (skim); locust bean gum: ... (cream), c. (skim).
ultimate volume fraction of skim, cP oo ' which the actual volume fraction
approaches asymptotically. The prediction for cP oo is obtained as 1 - a - b~
by setting t = 00 in the above equation.
3. For each creaming agent there is a minimum nominal aqueous-phase

concentration which, if exceeded, results in the separation of a skim
containing very little rubber. These minimum concentrations vary little
with variables such as the initial rubber content of the latex, the level of
ammoniation, and the pH. Table 9.1 gives the values reported by Twiss
and Carpenter [47] for this minimum concentration for various creaming
agents.
4. As the overall concentration of creaming agent increases, so the rubber
content of the skim falls progressively, whereas the final concentration of
rubber in the cream passes through a shallow maximum. These effects are
illustrated in Fig. 9.l4(b), which reproduces results reported by Hauser
and Dewey [48] for ammonium alginate and locust bean gum as the
creaming agents.
5. There is a rough correlation between the tendency of a water-soluble
hydrocolloid to accelerate creaming and its ability to increase the viscosity
of aqueous solutions. However, the viscosity of the aqueous phase of the
skim which separates is rather less than that which would be expected if
all the added creaming agent were entirely present in the aqueous phase.
It appears that a considerable proportion of the added hydrocolloid is
adsorbed at the surface of the rubber particles; the results of Twiss and
Carpenter [47] indicate that the proportion is approximately one half. It
also appears that the tendency to accelerate creaming is closely related
to the tendency of the molecules of hydrocolloid to adsorb at the interface
between the rubber particles and the aqueous phase.
6. The induction period before creaming commences is significantly reduced
by prior violent agitation of the latex. The effect of agitation is probably
two-fold: to ensure thorough homogenization of the system, and to induce
a degree of micro-flocculation of the latex particles. As regards the second
of these effects, it is relevant that recent investigation of the mechanism
whereby the mechanical destabilization of natural rubber latex occurs has
indicated that partial loose agglomeration of the rubber particles may
occur during the initial stages of the process prior to the onset of
catastrophic coagulation (see section 4.6.2.2 of Chapter 4 (Volume I)).
Table 9.1 Minimum nominal aqueous-phase concentrations of various creaming

agents which, if exceeded, result in the separation of skim containing very little rubber
(Twiss and Carpenter [47])
Creaming agent Minimum nominal concentration / % m/m on aqueous phase

Sodium alginate 0.211
Methyl cellulose 0.292
Locust bean gum 0.135
Gum tragacanth 0.107
Carageenin 0.091
52 Natural latices
7. It is possible to re-dilute a creamed natural rubber latex and then subject

it to a second creaming process. However, it is observed with some
creaming agents that each time this is done the creaming proceeds more
slowly than on the previous occasion, until a stage is eventually reached
when added creaming agents bring about almost no separation into two
layers. In practice, the number of creaming operations to which a given
sample of natural rubber latex can be subjected is limited to about three
or four. This observation suggests that some of the constituents of the
latex which are dissolved in the aqueous phase have a promoting effect
in combination with creaming agents. This view is confirmed by the
observation of Vester [49] that the addition of skim to the latex generally
facilitates creaming. However, whereas some creaming agents appear to
depend for their effectiveness upon the presence of certain hydrophilic
non-rubber constituents, it appears that others do not, and that unlimited
repeated creaming is possible with creaming agents in this latter group.
9.4.3.3.2 Theories of the mode of action of creaming agents
How are these observations to be interpreted? Several theories of the creaming
process in natural rubber latex have been suggested. All propose an increase
in the effective size of the latex particles. The theoretical considerations given
in section 9.4.3.1 above indicate that this will be accompanied by two separate
effects:
1. The rate of creaming will increase because, for a given density difference,
the buoyancy force on a particle increases as the third power of the particle
diameter, whereas the viscous drag on the particle increases as only the
first power of the diameter.
2. The variation of the ultimate equilibrium particle concentration with
vertical location in the latex will become more pronounced because, for
a given density difference, the work required to move a particle a given
distance down through the aqueous phase increases as the third power
of the particle diameter.
Closely related to the second of these effects is the consideration that any
increase in particle size is inevitably accompanied by a reduction in the speed
of the Brownian motion of the particles. Because Brownian motion opposes
the creaming tendency, any restriction of Brownian motion must favour
creaming. Some theories also appear to propose that some further suppression
of the Brownian motion of the particles may occur, additional to that which
inevitably accompanies an increase in particle size.
The two effects which have been noted in the preceding paragraph as
accompanying any increase in the effective size of the dispersed latex particles
may have quite distinct consequences for the creaming oflatex systems which
are such that the increase in effective particle size is reversible. The variation
of the ultimate equilibrium particle concentration with vertical location in
the latex may then be little affected by the increase in effective particle size,
because this variation will be mainly determined by the initial distribution

of particle sizes in the latex. What will be significantly affected is the rate at
which that variation of ultimate equilibrium particle concentration is attained,
because the temporary particle aggregates rise rapidly through the aqueous
phase. On the other hand, if irreversible flocculation of the latex particles
has occurred, then this will be expected to affect the variation of the ultimate
equilibrium particle concentration with vertical location, as well as the rate
at which that ultimate variation is attained.
One theory of the mechanism of creaming in natural rubber latex supposes
that the effective size of the latex particles is enhanced by adsorption of the
heavily hydrated hydrocolloid macromolecules which are added as creaming
agents, and that the viscous drag on the additional hydration layer thereby
acquired by the particle effectively suppresses the Brownian motion. This
theory has to be rejected on two grounds. Firstly, it requires the particles to
be hydrated to an improbable extent. Secondly, the proposed extensive
hydration would reduce the effective density difference between the particles
and the aqueous phase, thereby seriously offsetting the advantage gained
through the apparent increase in particle size.
A second theory proposes that the adsorption of the macromolecules of
creaming agent promotes reversible agglomeration of the latex particles by
reducing the effective density of electric charge at the particle surface. The
particles are thereby allowed to approach each other and to cohere loosely
without actually coalescing; indeed, it is suggested that the presence of an
adsorbed layer of heavily-hydrated hydrocolloid macromolecules effectively
prevents interparticle coalescence. From the investigation reported by Baker
[50], it is clear that such reversible agglomeration probably does occur. He
examined by microscope dilute latices which contained concentrations of an
added creaming agent (gum tragacanth) similar to those employed in practice.
He observed a gradual aggregation of the particles into clusters, with a
concomitant slowing of the Brownian motion. Aggregation was complete in
a few minutes, and the Brownian motion ceased altogether. The clusters were
still present after the latex had creamed, but dispersed again when the latex
was diluted. Redispersion was accompanied by recommencement of the
Brownian motion. Aggregation was also observed to occur when latex, not
otherwise containing creaming agent, was diluted with the skim from latex
creamed in the normal way. By this means, Baker simulated as nearly as
possible the conditions which prevail in a normal creaming operation.
Although the occurrence of particle clustering seems established, this mechanism
is to be rejected because the water-soluble hydrocolloids commonly used as
creaming agents, such as the algi nates, are anionic. Thus the adsorption of
these macromolecules at the particle surface would not be expected to reduce
the effective surface electric charge density. Furthermore, measurements of
electrophoretic mobility provide no evidence for any reduction of surface
electric charge density, but indicate the expected increase.
54 Natural latices
A third theory proposes the formation of a loose network between the

macromolecules of creaming agent which have become adsorbed at the
particle surfaces and those which remain dissolved in the aqueous phase.
The particles are envisaged as being entrapped in such a network, and the
Brownian motion suppressed in consequence. However, because the networks
are not necessarily distributed uniformly throughout the latex, the Brownian
motion is not everywhere suppressed. A particle is envisaged as moving until
it becomes entrapped in a localized network which is sufficiently strong to
hold it. As successive particles become entrapped in a given localized network,
so a cluster of particles is formed. These clusters are envisaged as growing
by the continual entrapping of particles until their buoyancy is sufficient to
cause them to break free from the network and move upwards through the
aqueous phase. In the upper layers of the latex, the clusters may become
more compact as a consequence of the compressive effect of those rising from
below, thereby causing expulsion of the aqueous phase. The induction period
for creaming is interpreted as the time required for the initial clusters to
develop to a size sufficient for their buoyancies to overcome the restraining
effect of the network of macromolecules of creaming agent. The promoting
effect of certain hydrophilic non-rubber constituents in the latex is understandable
as the promotion by those substances of the formation of networks of
macromolecules of creaming agent. On this view, it is understandable that
the observed effects in this respect should be specific to the particular creaming
agent used.
Although obscure in several of its details, the third of the above theories
gives the most plausible explanation for the effect of certain water-soluble
hydrocolloid macromolecules in accelerating latex creaming. It accords with
modern views concerning the behaviour of soluble lyophilic macromolecules
which are present in lyophobic sols. Such macromolecules are expected to
be adsorbed at the particle surfaces, and some interparticle bridging is
expected to occur if the polymer molecular mass is sufficiently high. Indeed,
perhaps the most satisfactory mechanism for the mode of action of creaming
agents available at present is a modification of the third theory in which the
proposed networks of macromolecules of creaming agent form by reversible
adsorption of those macromolecules at the particle surfaces, followed by
reversible interparticle bridging. It may be that, through the mechanism of
reversible interparticle bridging, the creaming agent macromolecules cause
the reversible formation of particle aggregates which rise rapidly through
the aqueous phase because of their increased size. This latter view explains
most, if not all, the features of the creaming process listed above. The
importance of the ability of the creaming agent macromolecules to adsorb
at the particle surface is self-evident. The correlation between ability to
accelerate creaming and the ability to enhance the viscosity of aqueous
solutions can be understood because the higher the molecular mass of the
creaming agent macromolecule, the more readily can it adsorb at the particle
surface and bring about interparticle bridging, and the more effective it is in
enhancing solution viscosity. The apparent exudation of aqueous phase
during creaming may be a consequence of the rapid rising of large temporary
particle agglomerates, perhaps augmented by the elimination of part of the
interparticle aqueous phase as the particle agglomerates form. The induction
period can be interpreted as the time required for the formation of particle
agglomerates of sufficient size to rise sharply. That the concentration of
rubber in the skim should fall sharply with increasing concentration of
creaming agent is readily understandable. That there should be a minimum
concentration of creaming agent for the formation of an effectively rubber-free
skim is understandable in relation to the specific surface area of the particles,
especially as the particles which are most resistant to creaming are those
which have the greatest specific surface area. The effect of prior agitation
can be understood as augmenting the formation of particle agglomerates.
The presence of certain hydrophilic non-rubber latex constituents may be
essential for some creaming agents to be effective for various reasons, such
as facilitation of adsorption.
9.4.3.4 Other methods of accelerating the creaming of natural rubber latex

The creaming of natural rubber latex can also be accelerated in ways other
than by the addition of water-soluble hydrocolloids. Any influence which
tends to increase the effective size of the particles, either reversibly or
irreversibly, is expected to promote creaming. Thus excessive dilution of the
latex with aqueous solutions of non-coagulating electrolytes is effective. Such
dilution probably results in the desorption of some of the natural latex colloid
stabilizers with accompanying irreversible micro-flocculation of the latex
particles under the influence of the increased ionic strength of the aqueous
phase. Creaming is also accelerated by the addition of dilute solutions of
strong alkalis. In this case, the effect is probably a consequence of partial
hydrolysis of proteinaceous colloid stabilizers, leading to desorption of those
stabilizers from the particle surface, again causing irreversible micro-flocculation
of the latex particles under the influence of the increased ionic strength of
the aqueous phase. Rapid creaming can also be induced by controlled
reduction in the pH of the latex, or by the careful addition of quaternary
ammonium surface-active substances. In both cases, the density of bound
electric charge at the particle surface is reduced, allowing irreversible
micro-flocculation of the latex particles to occur.
9.4.4 Concentration by centrifugation
Centrifugation is by far the most important of the methods currently used
for the concentration of natural rubber latex. Very roughly, ca. 90% of the
56 Natural latices
natural rubber latex concentrate used industrially is produced by centrifugation,

most of it by subjecting the latex to a single stage of centrifugation. The
remaining 10% or so ofthe natural rubber latex used industrially is produced
either by creaming or by evaporation, the market proportions for these two
types of concentrate being approximately equal. Small amounts of concentrate
produced by repeated centrifugation with intermediate dilution have been
produced for special purposes, as have concentrates produced by a combination
of creaming and centrifugation.
Centrifugation is in effect a type of accelerated creaming process, in which
the motion of the rubber particles relative to the aqueous phase is effected
by means of a centrifugal field rather than a gravitational field. Like creaming,
successful concentration by centrifugation depends upon there being a
significant difference between the density of the rubber particles and that of
the aqueous phase. Neither method is applicable if the difference between
the densities is very small. By analogy with the products obtained when
natural rubber latex is concentrated by creaming, the centrifuge concentrate
is also known as cream, and the dilute latex obtained as a by-product is
known as skim.
9.4.4.2 Elementary theory of centrifugation

To a first approximation, the motion of a centrifuging latex particle relative
to the dispersion medium in which it is suspended is confined to the radial
direction. For practical purposes, the so-called Coriolis motion, i.e., the lateral
movement of the particle relative to the dispersion medium, can be neglected.
The theory of centrifugation is analogous to that of creaming, the acceleration
due to gravity, g, being replaced by the centrifugal acceleration associated
with the circular path which any given particle is constrained to follow. The
theory of centrifugation is somewhat more complex than that of creaming
because, whereas gravitational acceleration is independent of particle location
within the latex, this is not so for centrifugal acceleration. If the radius of
the particle path is R, then the centrifugal acceleration is 00 2 R, where 00 is
the angular speed of the particle. Thus the centrifugal acceleration is directly
proportional to the distance ofthe latex particle from the axis of the centrifuge.
Like creaming, the process of centrifugation is opposed by the Brownian
motion of the dispersed particles. Eventually, an equilibrium variation of
particle concentration with distance from the centrifuge axis is established
between the opposing tendencies of separation and dispersion. The nature
of this equilibrium variation can again be predicted as a special case of the
Boltzmann distribution. The reversible work which has to be done against
the centrifugal force to move a particle from a circular path of radius R to
a concentric circular path of radius R + oR is -moo 2R. oR, where m is the
effective mass of the particle. Integration of this expression between limits
R = Ro and R = R gives V2 moo2(R~ - R2) as the reversible work which has
to be done against the centrifugal force in order to move a particle from a

circular path of radius Ro to a concentric circular path of radius R( < Ro).
As for creaming, m = nx 3 (p - a)/6 for a spherical particle of diameter x and
density p immersed in a dispersion medium of density a. The Boltzmann
distribution then gives the variation of equilibrium particle concentration
with distance R from the axis of the centrifuge as
(9.6)
Thus the extent to which the particle concentration at distance R differs from
that at distance Ro depends upon (J)2 and upon the parameter x 3 (p - a).
The direction, as well as the extent, of the centrifugal separation is determined
by the density difference (p - a). If, as for the particles in natural rubber
latex, p is less than a, then c> co' i.e. the particles centrifuge towards the
axis of the centrifuge. Equation (9.6) also shows that the extent of the
centrifugal separation is very dependent upon particle diameter. Figure 9.15
shows predictions given by equation (9.6) for the variation of clco with
(R2 - R~) at 25°C for a dispersion for which p = O.950Mgm- 3 and
a = 1.000 Mg m - 3, and for rotational frequency 6000 minute - 1. Curves are
shown for particles of diameter 50nm, lOOnm and 150nm. As for creaming,
very marked dependence of concentration variation upon particle size is
predicted. However, these predictions are derived assuming that the only
6
0
u
U
4
2 A
00 4 8 12 16 20
(R~-R2) x 103/cm 2
Figure 9.15 Predictions given by equation (9.6) for variation of clc o with (R2 - R~) forrotational
frequency 6000 minute - I at 25'C for dispersions of monodisperse spherical particles having
diameters 5Onm(curve A), 100 nm (curve B) and 150nm (curve C), takingp = 0.950 Mgm- 3 and
(J = l.OOOMgm- 3 •
58 Natural latices
contribution to W is that from the difference in the centrifugal potentials at

the two locations. In practice, as for creaming, as c increases, so the particles
become increasingly crowded, the interparticle separation decreases, and
contributions to W from those factors which confer colloid stability upon
the dispersion become increasingly important. The increase of c with
(R2 - R~)for latices having large particle concentrations is therefore expected
to be less than that predicted by equation (9.6).
It is also necessary to consider the rate at which particle separation occurs.
The average steady speed at which a particle moves through the dispersion
medium depends upon the balance between the centrifugal force which tends
to accelerate motion and the viscous drag which tends to retard motion. For
a spherical particle of diameter x at a distance R from the axis of the centrifuge,
the centrifugal force is nx 3 (p - (J)w 2 R/6, and the magnitude of the force of
viscous drag is ca. 3n"xldR/dtl, where" is the viscosity of the dispersion
medium. The direction of the viscous drag is opposed to that of the centrifugal
force. Thus the steady value of dR/dt is such that
nx 3 (p - (J)w 2 R 3 dR
= n"x- (9.7)
6 dt
i.e.,
dR (p - (J)w 2 R 2
-= x (9.8)
dt 18"
Equation (9.8) shows that the steady speed of movement of a particle in a
centrifugal field is directly proportional to the density difference (p - (J). The
sign of this difference determines the direction of movement of the particle
in the dispersion medium. dR/dt is positive or negative according as p is
greater or less than (J; for particles of natural rubber in an aqueous phase,
p is less than (J, and so dR/dt is negative, confirming that the particles move
inwards through the aqueous phase towards the axis of the centrifuge.
Equation (9.8) also shows that dR/dt is directly proportional to the square
of the particle diameter and inversely proportional to the viscosity of the
dispersion medium. Thus, like creaming, rapid centrifugation is favoured by
large particles, low viscosity of the dispersion medium, and a large density
difference between particles and dispersion medium. Equation (9.8) also shows
that dR/dt is directly proportional to the distance of the particle from the
centrifuge axis. Thus the particles in a dispersion for which p > (J move away
from the centrifuge axis at a speed which increases with time, whereas the
particles in a dispersion like natural rubber latex, for which p < (J, move
towards the centrifuge axis at a speed which decreases with time. The
maximum radial displacement of a particle of diameter x which was initially
at a distance Ro from the centrifuge axis, and which has been subjected to
centrifugation for a time t, can be obtained by integrating equation (9.8). The
result for R is
R = Roexp [
(p - u)w 2 X2
18'7
tJ (9.9)
Thus the radial displacement is
AR = R - Ro = Ro { exp [
(p - u)w 2 X2
18'7
tJ - 1} (9.10)
The effect of centrifugation upon the particle-size distribution of the latex

is evident from the above considerations. The following approximate analysis
is similar to that given in section 9.4.3.1 above for the effect of creaming
upon particle-size distribution. Consider a particle which initially moves in
a path of radius R o , but after a time t moves in a path of radius R. Equation
(9.9) shows that the diameter of the particle must be
18'7 ln~ (9.11 )

(p - U)W2 t Ro
Now suppose that t is the time which has elapsed since centrifugation
commenced, that a path of radius R or less must be attained if a particle is
to emerge from the centrifuge in the cream and not in the skim, and that Ro
is the radius of the path which is most distant from the centrifuge axis. Then
no particle of diameter greater than X R can remain in the skim; for, according
to equations (9.10) and (9.11), even the most distant of particles of this size
or greater will have had time to move the requisite distance (Ro - R). On
the other hand, there is no reason why particles having diameters less than
X R should not be present in the cream, because many of these will be
sufficiently close to, and indeed within, the critical radius R. Thus all sizes
less than X R are predicted to be present in both cream and skim, whereas all
particles of diameter x R or greater are absent from the skim. These conclusions
accord generally with what is observed in practice. They are illustrated in
Figs 9.16(a)~(c) for aqueous dispersions for which p = 0.950 Mg m - 3,
U = 1.000 Mg m - 3 and '7 = 1 mPa s, and for rotational frequency 6000 minute - 1.
Figures 9.16(a)~(c) show respectively the variation of X R with R/Ro for various
times of centrifuging, the variation of X R with time of centrifuging for various
values of the ratio R/R o , and the variation with R/Ro of time of centrifuging
required for X R to have specified values.
9.4.4.3 Centrifugation in practice

Factories for the concentration of natural rubber latex by centrifugation are
usually sited on the side of a hill. The field latex is delivered to bulking and
storage tanks at the highest elevation. Gravity is then used to drive the latex
downwards through the centrifuges to bulking and storage tanks at the lowest
60 Natural latices
1.0
0.8 0.8
E 0.6 E
0.6
.2- .2-
IX:
)( 0.4 )(IX: 0.4
0.2 0.2 A
00 00 400 800 1200 1600 2000

(a) R/Ro (b) tis
1000
800
600
---
VI
400
200
00
0.2 0.4 0.6
R/Ro
(c)
Figure 9.16 Predictions given by equation (9.11) for effect of variables upon X R for a dispersion
forwhichp = 0.950 Mgm~' andu = 1.000 Mgm~', forrotationaI frequency6000minute~ 1 with
" = I mPas: (a) variation ofxR with R/Ro for times of centrifuging 5OOs(curve A), lOOOs(curve B)
and 1500s (curve q; (b) variation of XR with time of centrifuging (I) for R/Ro = 0.90 (curve A),
R/Ro = 0.75 (curve B), R/Ro = 0.50 (curve q and R/Ro = 0.25 (curve D); (c) variation of X R with
R/Ro of time of centrifuging (I) required for X Rto have values 1.5 J.1m (curve A), 10 J.1m(curve B) and
0.75 J.1m (curve q.
elevation, from which tanks the concentrate is transported onwards. In this

way, the need for pumps is eliminated.
Of the various types of centrifuge which are used for the concentration of
natural rubber latex, the best-known is the de Laval. A cross-section of the
centrifuge bowl, i.e., that part of the machine in which the latex is separated
into cream and skim, is shown schematically in Fig. 9.l7(a). The mass of
latex being concentrated is broken up into a number of thin conical shells
within a bowl which rotates at high speed. The maximum distance which a
latex feed from

constant-head device
Regulating screw
Distributor
Bowl casing
Separator discs
Outside of
centrifuge (static)
{a}
Cream
i Axis of rotation
Skim incoming
(0) latex
Figure 9.17 Illustrating the de Laval centrifuge: (a) schematic cross-section of bowl, indicating
path of latex through bowl; (b) illustrating effect of separator discs in facilitating separation.
particle has to traverse in order to pass from the skim into the cream is
small. As is illustrated in Fig. 9.l7(a), the rotating latex is broken up into
62 Natural latices
thin layers by means of a number of metal discs. The effect of these discs in
facilitating separation is illustrated in Fig. 9.17(b). Because the distance of
traverse required to effect separation is so small in relation to the distance
from the centrifuge axis, it follows that the ratio Roj R of section 9.4.4.2 above
is nearly unity, and becomes more nearly so as the skim flows outwards
down the shell. The considerations discussed in section 9.4.4.2 above lead to
the conclusion that the critical particle diameter, X R , particles larger than
which are absent from the skim, is then very small. Little rubber remains in
the skim, and the efficiency of separation, defined as the fraction of the total
rubber entering the machine which effluxes in the form of latex concentrate,
is high.
The degree of ammoniation of the latex prior to centrifuging depends upon
the time which has elapsed since reception. It ranges from ca. 0.25% m/m if
centrifuging is immediate, to ca. 0.8% m/m if the delay is 2 days. The latex
is fed to the centrifuge from a constant-head device, so as to ensure a constant
flow through the machine. The latex enters the centrifuge bowl through a
central feed tube, and thence passes through a distributor to the bottom of
the bowl, where it flows through holes into a stationary gully. The skim flows
outwards away from the centrifuge axis, and leaves the bowl through orifices
which are controlled by regulating screws. The positioning of these screws
is shown in Fig. 9.17(a); their function is described in section 9.4.4.4 below.
The skim passes out of the centrifuge through a second gully. Typically, the
bowl rotates at a frequency of ca. 6000minute- 1 ; at this particular rotational
frequency, the centrifugal acceleration at a distance t 5 cm from the centrifuge
axis is ca. 6000 g.
When the centrifuge has settled down to a steady condition, there is a
progressive change in the density of the latex, and hence in the rubber content
of the latex, from the periphery of the centrifuge bowl to its innermost surface.
The design of the machine is such that the incoming latex enters at a point
where the latex density is as near as possible to that which corresponds to
its rubber content. It is desirable that the initial separation into skim and
cream should be as rapid as possible, and that the cream should be led
quickly away from the region of intensive separation, in order to avoid
over-concentration and concomitant decrease in the fluidity and colloid
stability of the latex. Thus the concentrate should flow rapidly to a region
where the centrifugal force is less intense, and thence by an easy route out
of the machine. On the other hand, to ensure maximum recovery of rubber
as latex concentrate, the skim should flow out of the machine by as devious
a route as possible, passing through regions of increasingly intense centrifugal
force. The manner in which these requirements are implemented in the de
Laval centrifuge should be evident from Fig. 9. t 7(a).
The parts of the machine which come into contact with the latex should
be constructed from materials which are resistant to corrosion by dilute
aqueous ammonia solutions, and which do not contaminate the latex. To
meet these requirements, the bowl, distributor and collecting gullies are
usually made of tinned steel, and the separator discs and orifices of stainless steel.
9.4.4.4 Factors which affect composition of cream in relation to that of the

incoming latex
Amongst the more important factors which affect the composition of
centrifuge cream in relation to that of the incoming latex are the feed rate,
the rotational frequency of the centrifuge, and the length of the regulating screw(s).
As regards the effect of feed rate, a reduction in the rate of supply of latex
to the centrifuge causes an increase in the efficiency of the separation process,
and in the rubber content of the concentrate obtained. The latter may increase
to such an extent that the cream is no longer sufficiently fluid to flow through
the centrifuge. Increasing the rotational frequency of the centrifuge increases
the efficiency of separation by increasing the centrifugal field to which the
latex particles are subjected. The effect of centrifuge speed upon the efficiency
of separation is marked because, as demonstrated in section 9.4.4.2, the
centrifugal acceleration is proportional to the square of the angular speed
of the centrifuge.
The length of the regulating screws can be varied by inserting screws of
selected length into the skim-discharge orifice. In this way, fine control can
be exercised over the equilibrium difference between the density (and hence
rubber content) of the discharging skim and that of the cream. As a
consequence, fine control can be exercised over the rubber content of the
cream relative to that of the skim. The free liquid surfaces inside the centrifuge
bowl are approximately cylindrical, because the effect of gravity is negligible
in comparison with the centrifugal acceleration. In the steady-state, there
must be approximate equality of pressure at the points of discharge of skim
and cream. The position of the orifice through which the cream effluxes is
determined by the design of the machine, but the effective position of the
orifice through which the skim effluxes can be varied within limits by varying
the length of the threaded portion of the regulating screw which projects
into the centrifuge bowl (see Fig. 9.17(a». A shorter screw length increases
the horizontal distance between the skim and cream orifices, and thus
encourages a greater difference in density between skim and cream; hence
the cream tends to have increased rubber content, and the proportion by
volume of the input which effluxes as skim also increases. Longer screws
have the opposite effects.
9.4.4.5 Efficiency of latex concentration by centrifugation

As stated in section 9.4.4.3 above, the efficiency of a centrifuge latex-
concentration process is defined as the fraction of the total rubber entering
the machine which effluxes as concentrate. There is a simple relationship
between the efficiency defined in this way and the dry rubber contents of the
64 Natural latices
initial field latex, the cream and the skim. This relationship can be derived
as follows: If c is the mass of cream of dry rubber content C% m/m which
effluxes in a given time, and f is the mass of field latex of dry rubber content
F% m/m which enters during the same time, then, from the above definition,
the efficiency, E, of the process is given by
Cc
E=- (9.12)
Ff
If s is the mass of skim of dry rubber content S% m/m which effluxes during
the same time, then the mass balance for the influxing and effluxing latex
gives (c + s) = J, and the mass balance for the rubber contents of the latices
gives cC + sS = fF. Eliminating s from these two relationships gives
elf = (F - S)/(C - S). Substitution in equation (9.12) then gives
E = C(F - S) (9.13)
F(C - S)
Typical values for F, C and S are 30%, 60% and 5% respectively. These
values give E as ca. 0.91 and e/f as ca. 0.45. Thus a centrifuge concentration
process which is operating at 90% efficiency produces approximately equal
volumes of cream of 60% m/m dry rubber content and skim of 5% m/m dry
rubber content from field latex of dry rubber content 30% m/m. The
industry-wide average for the fraction of rubber recovered as cream appears
to be rather less than 0.90; a more realistic estimate seems to be ca. 0.85.
9.4.4.6 Utilization of centrifuge skim

Large volumes of skim latex are produced as a by-product ofthe concentration
of natural rubber latex by centrifugation. It is evident from the discussion
given in section 9.4.4.5 above that some 10-15% of the rubber which enters
the centrifuge effluxes as skim latex. The dry rubber content of the skim
varies between 2.5 and 10% m/m. The ratio of non-rubber solids to rubber
is very high, being ca. 1/1 in some cases. The non-rubber constituents are
mainly proteinaceous and other nitrogenous substances. There are two
reasons for the high content of non-rubber substances in the skim:
1. The ratio of aqueous phase to rubber is much higher in the skim than in
either the initial field latex or the cream. Thus the ratio of water-soluble
non-rubber substances to rubber in the skim is much higher than in either
the initial field latex or the cream.
2. The rubber in the skim is present in the form of very small particles of
high specific surface area. The ratio of adsorbed non-rubber substances
to rubber in the rubber phase of the latex is therefore higher in the skim
than in either the initial field latex or the cream.
Most skim latex is coagulated to give a product which is known as skim

rubber. Direct coagulation of skim latex is not easy, because the colloid
stability is enhanced by the high content of ammonia and proteinaceous
substances, and by the small size of the dispersed rubber particles. One
method is to add sulphuric acid. Calcium chloride and formaldehyde have
also been used as coagulants. The skim may also be left to coagulate
spontaneously. The coagulum is processed by conventional means, and is
thoroughly washed in order to improve the appearance of the product.
Even the effluent from skim-coagulation processes can contain significant
amounts of rubber. Not only is it economically worthwhile to recover this
rubber; it is also environmentally desirable to remove the rubber from the
effluent before the latter is discharged into streams and rivers, in order to
reduce the level of pollution. One method of dealing with this problem is to
allow the effluent to reside for a period in a large pond in which the conditions
are anaerobic. The rubber particles gradually coagulate and rise to the surface
of the pond to form a thick layer of rubber admixed with various non-
rubber substances. This layer of crude rubber is removed manually by cutting
with knives, and is then subjected to further processing. The conditions
in the pond are then allowed to become aerobic before the effluent is
discharged.
Some interest has been shown in processes for the production of improved
grades of natural rubber from skim latex. One method has been described
by Morris [51]. The skim latex is subjected to enzymatic deproteinization
to ensure that the rubber which is subsequently precipitated is essentially
free of proteinaceous substances. In a typical process, a mixture of skim latex
and a suitable enzyme extract is pumped into a wax-lined tank and allowed
to digest for 24 hours. The best-known enzyme for this purpose is trypsin,
which is available in the form of pancreatic extract. The recommended level
of addition is ca. 0.1 kg of extract per m 3 of skim latex containing 6% mlm
rubber. The extract is added as an 8% mlm aqueous slurry. After digestion,
the skim latex is de-ammoniated to 0.1 % mlm by aeration and agitation. It
is then run into coagulation tanks to which have been added a small amount
of either sodium thiosulphate or sodium metabisulphite. The rubber is
precipitated by addition of dilute formic acid. When precipitation is complete,
the aqueous phase should be clear and amber-coloured. The coagulum is
cut into strips and thoroughly leached in clean water before being creped,
washed and dried. The success of this process depends upon the proteinaceous
substances in the latex being hydrolysed to such an extent under the catalytic
influence of the enzyme that they are not co-precipitated with the rubber.
An alternative process for the treatment of skim latex, developed by the
Firestone Plantations Company, has been described by Verhaar [52].
A certain amount of skim latex has been used in the past for the production
of 50% mlm cyclized rubber masterbatches, the rubber for the cyclized-rubber
component being derived from skim (see further section 14.5.2 of Chapter 14).
66 Natural latices
9.4.5 Concentration by electrodecantation
The concentration of natural rubber latex by electrodecantation is seldom,

if ever, practised now. The method is based upon an observation that, when
purifying lyophobic sols of substances such as silicic acid by electrodialysis
between vertical semi-permeable membranes, a separation of the sol into
horizontal strata of different concentrations sometimes occurs. The more
concentrated layers form uppermost if the disperse phase is less dense than
the dispersion medium, and at the bottom of the vessel if the disperse phase
is the more dense. The effect is probably a consequence of the combined
influences of the electrical and gravitational forces acting upon the dispersed
particles. It can be demonstrated by means of the apparatus illustrated in
Fig. 9.l8(a).
The principle of the concentration of natural rubber latex by electro-
decantation is illustrated in Fig. 9.18(b). An electrical potential difference is
applied between the two electrodes. Being negatively-charged, the rubber
particles tend to move towards the anode. The moving particles are
intercepted by a semi-permeable membrane which allows the passage of
aqueous phase but not of rubber particles. Provided that the potential
gradient is not too great, the rubber particles form reversible agglomerates
in the vicinity of the membrane surface. These agglomerates cream rapidly,
and thus a layer of latex concentrate is formed at the surface of the latex. It
is advantageous to reverse the polarity of the electrodes periodically, in order
to prevent blockage of the semi-permeable membrane by the formation of
irreversible agglomerates of rubber particles. The greater the potential
gradient used, the more frequently should the polarity be reversed. Reversal
periods of 2-10 minutes have been found to be satisfactory. The inertia of
the system is too great to allow the reversing potential difference to be
imposed by means of a commercial alternating-current supply; instead, a
direct current supply with mechanical (or other) means for periodic reversal
must be used. The effect of reversing the polarity of the electrodes is to cause
reversible agglomerates of rubber particles to form in the vicinity of the other
membrane, and for the residual agglomerates in the vicinity of the first
membrane to move away from that membrane.
Equipment for the industrial exploitation ofthe electrodecantation principle
for latex concentration has been described by Murphy [53], and is illustrated
schematically in Fig. 9.l8(c). The rate of separation of latex into cream and
skim is increased by introducing a number of semi-permeable membranes
between the two electrode compartments, analogously to the use of a number
of conical shells in a centrifuge bowl to break up the mass of the latex into
several regions of lesser thickness. The electrodecantation effect is found to
take place simultaneously at each membrane surface, without substantially
increasing the current consumption. In this way, the effective area of
membrane surface can be considerably enlarged. In the apparatus described
(ai
+
Cream
/ - ---..
..........
" \
\
\
\
Latex
,
\
\
\
Electrolyte Semi-permeabJe membranes
(oj
Electrodes
Skim outlet Membranes

(c)
Figure 9.18 Illustrating concentration of natural rubber latex by electrodecantation: (a) typical
laboratory demonstration apparatus; (b) illustrating principle of method; (c) schematic illustration
of equipment for the industrial exploitation of the e1ectrodecantation principle for latex
concentration (Murphy [53]).
68 Natural latices
by Murphy, up to 150 membranes are used. It has been found that, if the
layers of skim and cream are continuously led away, and the volume of latex
in the bath maintained constant by the continuous addition of fresh latex,
then the separation can be continued for several days without interruption.
Continuous removal of cream also has the advantage of improving the
uniformity of the product. The input rate must be carefully controlled so as
to avoid turbulence. The most suitable position for introducing the fresh
latex is through apertures in the side wall of the bath situated at a level
corresponding to one-third of the height of the membranes from their lower
edges. The level of the cream is ca. 2.5 cm above the tops of the membranes.
The cream is removed by a slowly-reciprocating skimming member which
conducts the cream over weirs at the ends of the bath. Skim is drawn off
from the base of the bath, its rate of flow being controlled by an adjustable
restriction. The electrodes are of stainless steel, and are contained in electrode
compartments at the ends of the bath. Provision is made for the venting of
evolved gases, and for the entry of the current leads. The membranes are of
'Cellophane', and are suspended from supports which are held in slots along
the edge of the tank. They are kept in vertical position by means of horizontal
glass rods inserted in hems in their lower edges. The bath itself is made of
wood which has been impregnated with wax.
The current consumed during latex concentration by electrodecantation
is small. According to Stevens [54], the energy required to produce one
kilogram of 60% m/m concentrate from 35% m/m latex is ca. 40-120 W h.
Voltage gradients are ca. 1 V cm -I. Murphy [53] states that a small unit
with an effective membrane area of ca. 10m 2 has an output of just over 9 dm 3
of concentrate per hour. Steve~s states that units with outputs in excess of
30dm 3 per hour have been constructed.
Temperature control of the latex bath is important, because the dissipation
of electrical energy as the process proceeds tends to cause the temperature
to rise. Increasing temperature has the effect of increasing the electrical
conductance of the bath, thereby causing unnecessary consumption of
electricity, and of promoting the formation of irreversible deposits upon the
membrane surfaces. It is therefore desirable to pre-cool the ingoing latex if
the ambient temperature is high. The efficiency of the process can be improved
by adding small amounts of anionic surface-active substances to the bath.
These substances adsorb at the surface of the particles, thereby increasing
the surface charge density of the particles and facilitating their movement in
the electric field.
A feature of the electrodecantation concentration process is the very small
amount of rubber which is present in the skim. The dry rubber content of
the skim rarely exceeds 1.5% m/m, contrasting favourably with an average
of ca. 7% m/m for centrifuge skim. Thus the electrodecantation method is
especially suitable for the production of natural rubber latex concentrate of
low non-rubber content by repeated re-dilution and re-concentration. Such
latices have been of some interest for the coating of wires with rubber deposits
of maximum electrical insulation resistance. A corollary of the low rubber
content of the skim latex is that the distribution of particle sizes in the cream
more nearly corresponds to that in field latex than does that of centrifuge
cream. The disadvantages of the electrodecantation process are that it is
slow, and it yields a product which is not as suitable for the manufacture of
latex foam rubber as is that produced by centrifugation.
9.4.6 Properties of natural rubber latex concentrates
9.4.6.1 Once-concentrated natural rubber latices

Typical properties of once-concentrated natural rubber latex concentrates
are summarized in Table 9.2. Natural rubber latex concentrates produced
by evaporation have a high content of non-rubber substances, many of which
are hydrophilic. These substances can amount to up to ca. 8% m/m of the
latex. Because many of the non-rubber substances are hydrophilic in nature,
deposits from evaporated concentrates tend to absorb more water than do
deposits from other types of concentrate. Deposits readily absorb water, and
may even partially redisperse in water if they have not been fully dried out.
Natural rubber latex concentrates produced by evaporation have high colloid
stability. The principal reason is the high content of non-rubber substances,
many of which function as colloid stabilizers. One expected consequence of
Table 9.2 Typical properties of once-concentrated natural rubber latices
Type of concentrate
Evaporated
KOH- KOH/NH 3 - NH 3 - Electro-

Property preserved preserved preserved Creamed Centrifuged decanted
Total solids content

/%mjrn 72-75 67-69 61-63 61-62 61-62 61-62
Dry rubber content
/%mjrn 65-68 60-62 55-57 59.5-60.5 59.5-60.5 59.5-60.5
Alkalinity / % mjrn on
latex concentrate 0.8-0.9 0.8-0.9 0.6-0.7 0.6-0.7 0.6-0.7· 0.6-0.7
Ash / % mjrn on latex
concentrate :;'1 :;'1 :;'0.5 0.6 0.5 0.5
Nitrogen / % mjrn on
latex concentrate 0.5 0.5 0.4 0.2 0.2 0.2
Viscosity / Pas ca. 6 ca. 1 0.08-0.15 0.01-0.05 0.01-0.05 0.01-0.05
• High-ammonia type centrifugate.

70 Natural latices
the high colloid stability is enhanced tolerance of evaporated concentrates

to compounding with high levels of inorganic fillers. As noted in section
9.4.1.2 above, concentrates produced by evaporation have wider particle-size
distributions than do those produced by the other methods; this is advantageous
in certain applications. The resistance to the deteriorative influences of heat
and oxygen of films from evaporated concentrates is said to be superior to
that of films from the other types of concentrate, this superiority being
attributed to the retention of natural antioxidants during concentration by
evaporation.
Three types of evaporated natural rubber latex are currently produced.
One type has a total solids content in the range 72-75% m/m, and is preserved
with potassium hydroxide and stabilized during production by the addition
of a small quantity of a potassium soap. A second type, which was developed
from the first and is currently the most important of the three, has a total
solids content of ca. 68% m/m. It is preserved with a mixture of potassium
hydroxide and ammonia. The third type is an ammonia-preserved concentrate
which has a total solids content of ca. 62% m/m. The first type of concentrate
is marketed commercially as Standard Revertex; it is a very thick pasty
material. The second is marketed as LCS Revertex, and the third as
T Revertex. The latter is considerably more fluid than the 73% m/m
concentrate; it resembles centrifuged concentrate in appearance. It is colloidally
less stable than the 73% concentrate, but is more stable than the concentrates
produced by the other concentration processes. It is claimed by its producers,
however, that, in resistance to alternate cycles of freezing and thawing, it is
superior to the 73% concentrate, having been found to be unaffected by a
series of freeze-thaw cycles even to temperatures as low as - 40°C. Furthermore,
it is said to be undamaged by a single freezing to temperatures down to ca.
- 70°C. The name Revertex appears to have been invented by Hauser, who
was closely concerned with the early development of evaporation processes
for natural rubber latex. It is intended to signify that, because only water
has been removed during concentration, the latex concentrate can, by careful
blending with water, be reverted to a condition similar to that which pertained
prior to evaporation.
The concentrates which are obtained by creaming, centrifugation and
electrodecantation are all very similar in appearance and properties. The
product of creaming is less stable than the other two types, in the sense that
it is subject to after-creaming during storage. The results of van den Tempel
[55] for the effects of creaming and centrifugation upon the particle-size
distribution of natural rubber latex are shown in Fig. 9.19; these results are
presented as cumulative distributions for particle number with respect to
particle diameter. The range of particle diameters of the two types of
concentrate is virtually the same as that in the latex before concentration,
but, as expected, the proportion of particles smaller than any given size is
always less in the concentrates than in the latex before concentration. The
Q)
N
0.8
.~
Q)
"0
c:
j
Qj
.c 0.6
Ej
c:
~ 0.4
'0
c:
0
U
!!!
u. 0.2
o~----~----~----~----~----~
o 0.2 0.4 0.6 0.8 1.0
Particle diameter / IJm
Figure 9.19 Effect of creaming and centrifugation upon cumulative distribution for particle
nwnber with respect to particle diameter for natural rubber latex (van den Tempel [55]). Curve A
shows the distribution for unconcentrated latex, curve B for centrifuged concentrate, and curve C
for creamed concentrate.
distributions of particle numbers by particle diameters in the skim latices

are similar to those in the latices before concentration, except that, as predicted
by the theoretical considerations given in sections 9.4.3.1 and 9.4.4.2 above,
no particles larger than a certain size are found in the skims. This diameter
for centrifuge skim is ca. 450 nm. Nevertheless, the proportion of smaller
particles in the cream is only slightly diminished by creaming or by centrifugation.
Various specifications for ammonia-preserved natural rubber latex concentrate
have evolved over the years in an attempt to provide a measure of quality
assurance for a commodity which is of natural origin. Being of natural origin,
natural rubber latex is subject to considerable inherent variation, as well as
to variations arising from handling and processing. The most important of
these specifications is that issued by the International Organization for
Standardization (ISO) [56]. This covers natural rubber latices which have
been produced by once-centrifugation and once-creaming processes. The
concentrates produced by centrifugation are denoted as Types HA, LA and
XA without further qualification, the Type XA being in effect a low-ammonia
concentrate with a somewhat greater ammonia content than normal.
Concentrates produced by creaming are distinguished by the word creamed
following the type designation. The requirements of the ISO specification
are summarized in Table 9.3. A common requirement for all latices conforming
72 Natural latices
Table 9.3 Requirements specified in ISO Standard [56] for ammonia-preserved

natural rubber latex concentrates produced by centrifugation and by creaming
Requirement for type indicated
Type H A Type LA Test

Property Type HA Type LA Type XA creamed creamed method
Total solids
content· / % m/m /61.5 /61.5 /61.5 /66.0 /66.0 ISO 124
Dry rubber
content· / % m/m /60.0 /60.0 /60.0 /64.0 /64.0 ISO 126
Non-rubber
solidst / % m/m i- 2.0 i- 2.0 i- 2.0 i- 2.0 i- 2.0
Alkalinity (as
ammonia)/% m/m
on latex
concentrate / 0.60 i- 0.29 / 0.30 / 0.55 i- 0.35 ISO 125
Added fixed alkali None None None None None
Mechanical
stability timetfs /650 /650 /650 1:650 1:650 ISO 35
Coagulum
content / % m/m i- 0.05 i- 0.05 i- 0.05 i- 0.05 i- 0.05 ISO 706
Copper content
/ mg (kg total
solids)-I i- 8 i- 8 i- 8 i- 8 i- 8 ISO 8053
Manganese
content / mg (kg
total solids) - I i- 8 i- 8 i- 8 i- 8 i- 8 ISO 7780
Sludge content
/%m/m i- 0.\0 i- 0.10 i- 0.10 i- 0.10 i- 0.\0 ISO 2005
Volatile fatty acid
(VF A) number As agreed by the interested parties. but not to exceed 0.20 ISO 506
KOH number' As agreed by the interested parties, but not to exceed 1.0 ISO 127
Colour No pronounced blue or grey Visual
inspection
Odour No pronounced odour of putrefaction Neutralize
with boric
acid
• The requirement is ror either total solids content or dry rubber content.
t The difference between total solids content and dry rubber content.
~ A minimum mechanical stability time may be required which is greater than the minimum
value specified.
§ If the latex contains boric acid, the KOH number may exceed the specified value by an amount
equivalent to the boric acid content as determined according to the method specified in ISO 1802.
to this specification is that there shall have been no addition of fixed alkali
at any stage of the production of the latices. Thus, for example, the KLAZN
type of ultra-low-ammonia latex described in section 9.3.4.6 above does not
conform to this specification because of the addition of a small amount of
potassium hydroxide.
The various test methods which are specified to be used to determine the
properties listed in Table 9.3 are described in Chapter 8 (Volume 1). As
regards the specification requirements, the following points should be noted:
1. Centrifuged concentrates are to contain a minimum of 60.0% mlm rubber.

2. There is also a requirement that the total solids content shall be at least
61.5% mlm, but this must not be so high that the difference between the
total solids content and the dry rubber content (this difference representing
the non-rubber solids content of the latex) exceeds 2.0% m/m.
3. For creamed latices, the minimum total solids and dry rubber contents
are respectively 66.0% and 64.0% m/m. These minima reflect the higher
solids contents and rubber contents which are attainable by creaming
relative to those attainable by centrifugation. In creaming processes, the
latex is subject to a gravitational field for much longer than the latex is
subject to the centrifugal field in centrifugation processes, notwithstanding
that the latter is much stronger than the former. Again, the difference
between the total solids content and the dry rubber content must not
exceed 2.0% m/m.
4. The ammonia content of the HA-type concentrates must not be less than
a specified minimum value, whereas that of the LA-type concentrates must
not exceed a certain maximum value.
5. Maxima are placed upon the copper and manganese contents of the latices
because these metals in appropriate forms are able to catalyse rapid
oxidative degradation of natural rubber, both unvulcanized and vulcanized.
6. A maximum is placed upon the VF A number in order to ensure that no
excessive microbiological degradation has occurred during the processing,
storage and transportation of the latex.
7. A maximum is placed upon the KOH number in order to ensure that
excessive amounts of soaps have not been added to the latex, e.g., to offset
the adverse effects of inadequate preservation upon the colloid stability
of the latex.
8. A pronounced blue or grey colour usually indicates that the latex has
become contaminated with iron, probably through improper storage.
A second important specification for ammonia-preserved natural rubber

latex concentrates is that issued as an American Society for Testing and
Materials (ASTM) Standard [57]. Three types of concentrate prepared
specifically from Hevea brasiliensis latex are recognized:
74 Natural latices
1. Type 1: centrifuged natural rubber latex preserved with ammonia only,

or by formaldehyde followed by ammonia;
2. Type 2: creamed natural rubber latex preserved with ammonia only, or
by formaldehyde followed by ammonia;
3. Type 3: centrifuged natural rubber latex preserved with a low level of
ammonia, together with other preservatives as necessary.
ASTM Type 1 latex is similar to ISO Type HA; the difference is that ASTM
permits initial preservation with formaldehyde, whereas ISO specifies that
ammonia shall be the sole preservative. ASTM Type 2 latex is similar to ISO
Type HA creamed, but with the same difference. ASTM Type 3 latex is
effectively the same as ISO Type LA. The property requirements for the three
ASTM types are the same as for the corresponding ISO types, with two
exceptions:
1. The limit for KOH number is 0.80, not 1.0. Again it is recognized that
the KOH numbers of latices containing boric acid may be higher than
normal by an amount equivalent to the boric acid in the latex.
2. No maximum for volatile fatty acid number is specified.
Extensive results for the composition and properties of industrial natural
rubber latex concentrates produced by centrifugation are available in papers
by Jurado and Mayhan [58], by Gorton and Pendle [59, 60], and by Gorton [61].
Natural rubber latex concentrates produced by evaporation are the subject
of a further ISO Standard [62]. Three types are recognized:
1. Type HA evaporated, which is preserved with ammonia only or with
ammonia and other preservatives, no fixed alkali having been added at
any stage in its production, and the alkalinity being at least 0.60% mlm
on the whole latex;
2. Type KHS evaporated, which is preserved with potassium hydroxide and
has a nominal total solids content of 73% m/m;
3. Type KlS evaporated, which is preserved with potassium hydroxide and
has a nominal total solids content of 68% m/m.
The requirements for these types are summarized in Table 9.4.
9.4.6.2 Multiply-concentrated natural rubber latices

By the term multiple concentration as applied to natural rubber latex is meant
a process in which a latex concentrate is diluted with ammoniated water and
then re-concentrated at least once. The principal effect of multiple concentration
is to reduce further the ratio of non-rubber substances to rubber in the latex.
This is what is usually meant by the 'purification' of natural rubber latex. A
secondary effect is the progressive elimination of the smaller rubber particles.
The colloid stability of the latex also falls with repeated dilution and
re-concentration, because the natural colloid stabilizers are progressively lost.
Table 9.4 Requirements specified in ISO Standard [62] for natural rubber latex
concentrates produced by evaporation
Requirement for type indicated
Type HA TypeKHS Type KLS Test

Property evaporated evaporated evaporated method

/%m/m -/:61.5 -/:72.0 -/:67.0 ISO 124
Non-rubber solids·
/%m/m 1 5.5 1 s.o 1 7.5
Alkalinity (as ammonia)
/ % m/m on latex
concentrate -/: 0.60 ISO 125
Alkalinity (as potassium
hydroxide) / % m/m on
latex concentrate -/: 0.75 -/: O.SO ISO 125
Mechanical stability
time / s -/:540 ISO 35
Coagulum content / % m/m 1 0.05 1 0.05 1 0.05 ISO 706
Copper content / mg (kg
total solids) - I 1S 1S 1S ISO S053
Manganese content / mg
(kg total solids)-I 1S 1S 1S ISO 77S0
Sludge content / % m/m 1 0.40 1 0.40 1 0.40 ISO 2005
Volatile fatty acid As agreed by the interested parties,
(VF A) number but not to exceed 0.20 ISO 506
Colour No pronounced blue or grey Visual
inspection
Odour No pronounced odour of putrefaction Neutralize
with boric
acid
• The difference between total solids content and dry rubber content, the dry rubber content
to be determined by the test method specified in ISO 126.
However, this effect can be offset by diluting the successive concentrates with
ammoniated water containing a suitable colloid stabilizer, such as a carboxylate
soap or a surface-active ethoxylate, instead of with plain ammoniated water.
Table 9.5 shows results reported by Murphy [53] for the effects of subjecting
a typical field natural rubber latex to one and two processes of centrifugation
and electrodecantation. Table 9.6 shows the results reported by Tan [63] for
the effect of one, two, three and four processes of centrifuging upon the
properties of an LAZN-type concentrate. All the differences between the
76 Natural latices
once-processed concentrates and the products of multiple concentration

shown in Tables 9.5 and 9.6 are consequences of repeated replacement of the
aqueous phase of the latex with water. In the context of synthetic latices,
this process has become known as serum replacement; more colloquially, a
latex which has been subjected to this treatment is said to have been
'laundered'. Referring to the results shown in Table 9.6, the following effects
of repeated dilution and centrifugation may be noted:
1. The non-rubber solids content of the latex, as quantified by the difference
between the total solids content and the dry rubber content, decreases
progressively, but does not become zero.
2. The KOH number decreases as the concentrations of the anions which
give rise to this property decrease, but again does not become zero.
3. The volatile fatty acid (VF A) number decreases, becoming virtually
immeasurable after two centrifugations.
4. The higher fatty acid (HF A) number decreases, but does not become zero.
5. The protein content (inferred from the nitrogen content) is reduced, but
again does not become zero.
6. The electrical conductivity decreases. This is attributed to removal of
water-soluble electrolytes from the aqueous phase of the latex.
Properties such as non-rubber solids, KOH number, HFA number and
nitrogen content which do not become zero with repeated dilutions and
centrifugations indicate the extent to which various non-rubber substances
are tightly bound upon or within the rubber particles. In particular, it is
virtually impossible to eliminate entirely nitrogen-containing compounds
Table 9.5 Effect of multiple concentration upon properties of natural rubber latex
concentrated by centrifugation and by electrodecantation (Murphy [53])
Latices concentrated Latices concentrated

by centrifugation by electrodecantation
Once Twice Once Twice
Total solids content / % m/m 61.3 61.7* 61.4 61.8

Dry rubber content / % m/m 59.7 61.5* 59.8 61.5
Alkalinity / % m/m ammonia
on latex concentrate 0.63 0.67 0.59 0.66
Ash / % m/m on latex
concentrate 0.45 0.14 0.40 0.14
Nitrogen / % m/m on latex
concentrate 0.22 0.23
* The difference between these two values is unusually low for twice-centrifuged natural rubber
latex. The difference between the corresponding values for twice-centrifuged latex shown in
Table 9.6 is more typical.
Table 9.6 Properties of multiply-centrifuged natural rubber latices (LAZN type)

(Tan [63])
Number of times latex was centrifuged
Property 1 2 3 4
Total solids contentI % mlm 61.8 60.8 61.3 60.3

Dry rubber content I % mlm 60.4 60.1 60.7 59.9
Non-rubber solids· I % mlm 1.4 0.7 0.6 0.4
Alkalinity I % mlm ammonia on latex
aqueous phaset 0.21 0.19 0.19 0.19
Mechanical stability Is 940
Volatile fatty acid (VF A) number 0.03 0.02
KOH number 0.61 0.24 0.15 0.13
Higher fatty acid (HF A) number 0.20 0.16 0.13 0.11
% m/mt
Protein in total solids fraction / 1.72 1.04 0.81 0.74
Protein in dry rubber fraction / % m/mt 1.32 0.87 0.75 0.66
pH 9.90 10.32 10.51 10.59
Electrical conductivity at 55% mlm total
solids content ISm-I 0.260 0.123 0.075 0.072
• The difference between total solids content and dry rubber content.
t These are the units as reported by Tan; but the figures suggest that the units may in fact have
been mass percentages on the latex concentrate, and not on the latex aqueous phase.
t As determined by the semi-micro Kjeldahl method as specified in BS 1673: Part 2: 1967, using
the factor 6.25.
from natural rubber latex by multiple concentration. Thus Kemp [64] found
that, even after nine centrifugations, natural rubber latex cream still contained
ca. 0.056% mlm of nitrogen, notwithstanding that the ash content was almost
zero.
9.4.7 Redispersible natural rubber from natural rubber latex

It is appropriate to conclude this discussion of methods of concentrating
natural rubber latex by brief reference to the production of what may be
regarded as the ultimate natural rubber latex concentrate. This is a solid
material from natural rubber latex which can be reconstituted to a colloidal
dispersion resembling natural rubber latex by the addition of water. A product
of this type would be the natural rubber latex analogue of powdered milk,
and would be very attractive commercially. It would reduce transportation
and storage costs to a minimum. The user could reconstitute as much or as
little as is required for immediate factory needs. Unfortunately, it has so far
78 Natural latices
not proved practicable to provide redispersible materials of this type.

Furthermore, the motivation for developing redispersible natural rubber latex
powders has been reduced in recent years because of the increasing amounts
of natural rubber which are used as latex in the rubber-producing countries.
One attempt to produce a redispersible rubber from natural rubber latex has
been described by Chin and Lau [65]. The method is said to be also applicable
to synthetic rubber latices. Water-soluble hydrophilic substances, such as
urea, a glycoside or a polyol, are added to the latex with or without an
anionic surface-active substance. The latex is then dried to a water content
of 3-5% m/m. It is important to avoid further dehydration of the mixture
if the resulting material is to be redispersible when water is added.
9.5 CONSTITUTION OF FRESH AND AMMONIA-PRESERVED

NATURAL RUBBER LATEX
9.5.1 Introduction
A clear distinction must be made between the constitution of fresh natural
rubber latex and that of ammonia-preserved natural rubber latex, whether
or not the latex has been subjected to a concentration process. It has been
pointed out in section 9.2.1.2 above that fresh natural rubber latex is the sap
of the latex vessels of the Hevea brasiliensis tree. As such, it is prone to
chemical change when it leaves the tree and when it is ammoniated. The
chemical changes which occur when the latex is ammoniated and then stored
are summarized in section 9.5.5 below, together with the consequences for
the colIoid stability of the latex. However, because prior to section 9.5.5 some
reference is made to ammonia-preserved natural rubber latex as welI as to
fresh natural rubber latex in considering the nature of the various phases
present in the latex, it is noted here that the principal chemical effect of
ammoniation is to promote alkaline hydrolysis of some of the non-rubber
substances present in fresh natural rubber latex. If the latex has also been
subjected to concentration by creaming, centrifugation or electrodecantation,
then this will have implications for the particle-size distribution and for the
ratio of non-rubber substances to rubber.
Freshly-tapped natural rubber latex is a whitish fluid having a density in
the range 0.975-0.980 Mg m - 3, pH 6.0-7.0, and surface free energy
40-45 mJ m - 2. Its viscosity is variable. Progressive dilution of fresh natural
rubber latex with water causes the viscosity to increase at first, reach a
maximum, and then suddenly decrease. These changes are attributed to the
swelling and bursting of bodies within the latex known as lutoids (see further
section 9.5.4 below, and also section 6.3.4 of Chapter 6 (Volume 1».
Being a natural product, the composition of fresh natural rubber latex
varies between wide limits. The composition given in Table 9.7 is typical.
Constitution of fresh and ammonia-preserved natural rubber latex 79
The substances present in fresh natural rubber latex are distributed between
the following three principal phases:
1. the rubber particles, which account for ca. 35% mlm of the latex;
2. the aqueous phase, which accounts for ca. 55% mlm of the latex; and
3. the lutoid phase, which accounts for most of the remaining ca. 10% mlm
of the latex.
There are also several minor phases present in the latex, note of which is
taken in section 9.5.4 below. The earliest of these other phases to be described
are the so-called Frey-Wyssling particles.
The rubber content of fresh natural rubber latex, excepting the latex which
exudes during the initial stages of wound response (see section 9.2.4 above),
varies over the range 25-35% m/m; the figure of 33% in Table 9.7 is an
average. The difference between the total solids content and the dry rubber
content of fresh latex is ca. 3% m/m; this compares with ca. 1.5% mlm for
the concentrate obtained by centrifuging once. The aggregate amount of non-
aqueous, non-rubber constituents is ca. 5% mlm of the whole latex. Of the
content of proteinaceous substances, about half is dissolved in the aqueous
phase, quarter is adsorbed on the surface of the rubber particles, and the
remaining quarter is associated with the larger particulate bodies such as
the lutoids.
9.5.2 The rubber phase
9.5.2.1 Shape, size, size distribution and physical structure of the rubber
particles in natural rubber latex
The first important microscope study of natural rubber latex to be reported
appears to have been that of Hauser [66]. He came to the conclusion that
the rubber particles are predominantly pear-shaped rather than spherical,
and that they consist of a tough, hard elastic shell which encloses a viscous
liquid. He reached this latter conclusion from a series of elegant studies using
Table 9.7 Typical composition of fresh natural rubber latex
Constituent Proportion j % mjm on whole latex
Total solids 36
Dry rubber 33
Proteinaceous substances 1-1.5
Resinous substances 1-2.5
Ash Up to 1
Sugars 1
Water ad 100
80 Natural latices
a micro-manipulator, by means of which the particles were pierced with a

fine needle. The Hauser two-phase model for the particles in natural rubber
latex accorded well with the then widely-accepted view that solid natural
rubber comprises two separate fractions, a sol fraction and a gel fraction.
Whilst the existence of sol and gel fractions in solid natural rubber is still
recognized (see further section 9.5.2.2.2 below), these fractions are now
regarded as being mixed on the molecular scale, rather than separated. Most
other observers have found the rubber particles in natural rubber latex to
be predominantly spherical in shape, especially in the latex from young trees.
There is general agreement that the shape varies to some extent with the age
and type of tree from which the latex is obtained.
The size of the particles in fresh natural rubber latex varies over wide
limits, the range 20-5000nm being not uncommon. Results have been
reported by van den Tempel [55] for the distribution, obtained from electron
micrographs of particles hardened by bromination, of particle sizes in
unconcentrated natural rubber latex. The cumulative particle-size distributions
shown in Figs 9.20(a) have been derived from these results making various
assumptions. These distributions are for (i) aggregate particle number, (ii)
aggregate particle surface area, and (iii) aggregate particle volume, all with
respect to particle diameter. The corresponding differential distributions have
also been derived; they are shown in Figs 9.20(b) (i), (ii) and (iii) respectively.
Although these distributions undoubtedly lack accuracy, they probably give
correct impressions. The distribution of particle sizes in fresh natural rubber
latex is so wide that most of the volume of the dispersed rubber in the latex
is present in the form of a small minority of large particles. Thus less than
4% of the particles have diameters larger than 400 nm; this is roughly the
fraction which is visible through an optical microscope. This minority of
particles accounts for ca. 85% v/v of the total dispersed rubber. It is possible
that this minority of very large particles is formed by the agglomeration and
coalescence of smaller particles.
The distribution of particle sizes in fresh natural rubber latex has also
been reported by Schoon and van der Bie [67]. Again the particles were
hardened by bromination to facilitate examination by electron microscope.
These workers also observed a highly asymmetric distribution of particle
sizes, which they interpreted as arising from the superimposition of several
relatively narrow Gaussian distributions. The observed particle diameters
were considered to be integer multiples of either 58 or 69 nm. From this,
they inferred (although the argument is unclear) that the larger particles may
have been formed by the clustering of smaller primary particles, without any
actual fusing of the particles occurring. Further investigation by Schoon and
Phoa [68], using the technique of phase-contrast microscopy applied to
brominated particles, has provided circumstantial evidence for such a
composite structure. Spherical, oval and pear-shaped particles were reported,
and their appearance was such as to lead these workers to conclude that
1.0
(i) (ii) (iii)
"0
J!lGl
.~.~ 0.8
0(1)
gj"O
al § 0.6
]ill)
0..!!2
-0
"0 'E 0.4
c:al
00.
~j 0.2
u.
oo .
0.4
..
0.8 1.2
. 1.6
. ..
2.0 o 0.4 0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0
(a) PARTICLE NUMBER AGGREGATE PARTICLE AGGREGATE PARTICLE

SURFACE AREA VOLUME
[l) = 0.20 of total in all cases
(i) (ii) (iii)
.~
o
..
0.4 0.8 1.2 1.6 2.0 o
.~ ..
0.4 0.8 1.2 1.6 2.0
o 0.4 0.8 1.2 1.6 2.0
(b) Particle diameter film
Figure 9.20 Particle-size distributions for unconcentrated natural rubber latex derived from results :-eported by van den Tempel [55]: (a) cumulative
distributions for (i) particle number, (ii) aggregate particle surface area, and (iii) aggregate particle volume with respect to particle diameter; (b) corresponding
differential distributions.
82 Natural latices
they were formed from much smaller primary particles. The addition of dilute
strong alkali to the latex was found to enhance the clustered appearance.
Van den Tempel [55] has also reported results for the cumulative
distribution of particle sizes in ammonia-preserved natural rubber latex
concentrates produced by centrifugation and by creaming (see Fig. 9.19
above). The results of a more extensive investigation of the distribution of
particle sizes in ammonia-preserved natural rubber latex concentrates has
been reported recently by Pendle and Swinyard [69]. The principal technique
used was photon correlation spectroscopy. Average particle diameters
calculated from the results obtained by this technique are said to be
'hydrodynamic' average particle diameters, i.e., they are the diameter of the
particle which, if it replaced all the particles in the actual sample, would give
hydrodynamic behaviour equivlent to that of the actual sample. The average
particle diameter calculated in this way is said to approximate to the
volume-average particle diameter. Samples taken from some 40 shipments
of centrifuged concentrates were investigated using this technique. The
average particle diameters obtained from the high-ammonia concentrates
were in the range 510-600 nm, whereas those for the low-ammonia concentrates
(LATZ type) tended to be somewhat lower, being in the range 480-550nm.
Analysis of Variance indicated that this apparent difference in average particle
diameter between high- and low-ammonia concentrates is real, and not a
consequence of random fluctuations. No comment is offered as to possible
causes; there is no obvious reason why concentrates produced by the same
concentration process but preserved with different levels of ammonia should
tend to have significantly different average particle diameters. It may be that
the difference is associated with the presence of the secondary preservatives
in the LATZ latices, but, if so, it might be expected that the average particle
sizes would be larger in the low-ammonia latices, rather than smaller, because
of slight colloidal destabilization by the secondary preservatives. A sample of
latex concentrate produced by creaming gave an average particle diameter of
494 nm; this is at the lower end of the range for centrifuged concentrates. As
expected. field latex gave an appreciably smaller average particle diameter (313 nm).
Pend Ie and Swinyard [69] have also presented information concerning
the distribution of particle sizes in the two types of ammonia-preserved
natural rubber latex concentrate. Typical distributions are shown in Figs
9.21 (a) and (b) for high-ammonia and LA TZ low-ammonia latices respectively.
An interesting feature of the distributions obtained by these workers is that
most of them are clearly bimodal. The peak diameters vary somewhat, the
ranges being 200- 300 nm for the lower peak, and 700-1500 nm for the higher
peak. Bimodality is also reported for a sample of field natural rubber latex,
although, as expected, the peaks are observed at lower particle diameters.
Pend Ie and Swinyard state that bimodality in natural rubber latices has been
confirmed by measurements made by the du Pont Company using the
technique of sedimentation field flow fractionation.
1000
Diameter / nm
(a)
500 1000
Diameter / nm
(b)
Figure 9.21 Typical particle-size distributions for (a) high-ammonia and (b) LATZ low-
ammonia latices (Pendle and Swinyard [69]).
9.5.2.2 Chemical constitution and structure ofthe particles in natural rubber latex
9.5.2.2.1 Overall composition of the rubber particles
A typical composition for the rubber phase in fresh natural rubber latex is
shown in Table 9.8. Trace metals, notably magnesium, potassium and copper,
are also associated with the rubber particles to an aggregate extent of ca.
0.05% m/m. The density of the rubber particles is ca. 0.920 Mg m - 3, this
being determined principally by the density of the rubber hydrocarbon.
9.5.2.2.2 The natural rubber hydrocarbon

The rubber hydrocarbon in natural rubber latex is predominantly linear
cis-l,4-polyisoprene (Structure IX). The following provisos should be noted
concerning this chemical description of the natural rubber hydrocarbon:
I. Description as a polyisoprene implies no more than that the structure is
such as would be produced by the addition polymerization of isoprene
Table 9.8 Typical composition of rubber particles in fresh natural rubber latex
Constituent Proportion I % mlm on

whole latex
Rubber hydrocarbon 86
Water (possibly dispersed in the rubber hydrocarbon) 10
Proteinaceous substances 1
Lipid substances 3
84 Natural latices
(2-methyl-l,3-butadiene) (CH 2 :C(CH 3 ).CH:CH 2 ) using an appropriate

initiation system. There is no implication that the natural rubber hydrocarbon
macromolecule is actually formed in this way.
2. The 1,4 refers to the mode in which the isoprene monomer units would
have to combine by addition polymerization in order to produce a
macromolecule like natural rubber hydrocarbon.
3. The adjective cis implies that all the carbon-carbon double bonds in the
polyisoprene macromolecule have the cis configuration (as opposed to the
trans configuration), i.e., that the disposition of the substituent groups
about each successive carbon-carbon double bond is such that the
polymeric chain substituent enters and leaves the double bond on the
same side (as opposed to the opposite side) as that on which it entered.
In describing natural rubber hydrocarbon as a linear polyisoprene, it is

implied that it is not branched. To the extent that some natural rubber
hydrocarbon macromolecules are or become crosslinked, the crosslinking is
light and is probably not an essential feature of the macromolecule as
produced in the tree.
Only an outline of the evidence upon which this conclusion concerning
the structure of the natural rubber hydrocarbon is based can be given here.
Briefly, chemical analysis shows the overall composition to correspond to
CsHs. Bromination and other tests diagnostic of olefinic unsaturation reveal
the presence of one carbon-carbon double bond for each CsHs unit.
Ozonolytic degradation to levulinic aldehyde and acid reveals that the
molecule of natural rubber is based upon the chemical unit 2-methyl-
2,3-butene. Physical properties indicate that natural rubber is a polymer of
high molecular mass. The average molecular mass of the soluble fraction can
be determined by techniques such as osmometry. Results of X-ray diffraction
studies of crystallized natural rubber are consistent with a repeat distance
corresponding to the cis configuration about the carbon-carbon double
bond, rather than with the trans configuration. The detailed microstructure of
the natural rubber macromolecule has been the subject of some controversy.
Infrared spectroscopy suggests the configurations about the double bond to
be ca. 97.8% cis-l,4 and 2.2% 3,4. This last figure is, however, in doubt,
because it is based upon an absorption band at 11.25 11m, the origin of which
is unknown, but which has been attributed to 3,4 microstructures. It may
be that the natural rubber hydrocarbon has exclusively the cis-l,4 microstructure.
IX
It is well known that dry solid natural rubber hydrocarbon comprises two
fractions which differ as regards behaviour towards solvents for polyisoprene:
a sol fraction, which is soluble in such solvents; and a gel fraction, which is
insoluble in, but highly swollen by, the same solvents. It has been known for
many years that the experimentally-determined proportion of gel in solid
natural rubber hydrocarbon depends upon the solvent which was used to
extract the sol. Results reported by Allen and Bristow [70] demonstrate this
phenomenon for a particular grade of pale crepe natural rubber. From these
results, shown in Table 9.9, it is evident that the gel content can vary over
a wide range, depending upon the solvent. Table 9.9 also gives values for
the thermodynamic polymer-solvent interaction parameters (Xl) for natural
rubber and the various liquids, and also for the diffusion coefficients for the
diffusion of the various liquids through natural rubber. It is evident that the
gel content does not correlate with the solvent power of the solvent for
natural rubber as quantified by Xl; it will be recalled that Xl increases in
magnitude as the liquid becomes a poorer solvent for the polymer. The gel
content does, however, correlate with the diffusion coefficient of the liquid
through rubber: the gel content decreases as the diffusion coefficient increases.
Although equilibrium between natural rubber hydrocarbon and the solvent
phase is reached only slowly, the variation of apparent gel content with the
nature of the solvent is not thought to be primarily a consequence of failure
to reach equilibrium. That the behaviour of the natural rubber hydrocarbon
towards solvents is unusual is confirmed by results reported by Allen and
Bristow for the gel contents for an artificial 25/75 m/m mixture of lightly
crosslinked and uncrosslinked synthetic cis-l,4-polyisoprene. The initial
polymer was gel-free; crosslinking was effected by heating with a peroxide.
The gel content of this mixture was determined by the same procedure, and
Table 9.9 Apparent gel contents of natural rubber (pale crepe grade) as determined
using various solvents (Allen and Bristow [70])
Diffusion
coefficient for
Experimentally- diffusion of
determined Rubber-solvent solvent through
gel content interaction rubber at 25°C
Solvent /%m/m parameter / XI / cm 2 S-I
Carbon tetrachloride 29 0.334 2.16 x 10- 6

Chloroform 22 0.383 3.90 x 10- 6
Toluene 18 0.391 3.28 x 10- 6
Cyclohexane 46 0.399 2.05 x 10- 6
Tetrahydrofuran 16 0.452 3.85 x 10- 6
2,2,4-trimethylpentane 40 0.513 2.34 x 10- 6
n-butyl acetate 30 0.561 2.50 x 10- 6
n-propyl acetate 68 0.649 2.05 x 10- 6
86 Natural latices
using the same solvents, as for natural rubber hydrocarbon. The gel content
was found to be essentially independent of the solvent used, being ca.
25% m/m in all cases. This is typical of the behaviour normally encountered
when mixtures of sol and gel fractions of the same polymer are extracted
with solvents for the sol fraction. Other unusual features of the gel content
of natural rubber hydrocarbon are that it has associated with it proportionately
more of the nitrogenous substances from the latex than does the sol material,
the addition of small amounts of polar liquids such as aliphatic alcohols to
the extraction liquid can greatly reduce the apparent gel content, and the
apparent gel content is also greatly reduced when the latex is enzymatically
deproteinized. Thus, concerning the latter phenomenon, Shiibashi [71] has
reported results which show that the apparent gel content of natural rubber
could be reduced from ca. 40% m/m to almost zero by treating the latex
with a deproteinizing enzyme for ca. 3 weeks.
These observations concerning the gel content of natural rubber provoke
at least three questions:
I. What is the sol/gel ratio in the particles of natural rubber latex when
freshly tapped?
2. What is the sol/gel ratio in the particles of natural rubber latex concentrate
as used industrially?
3. What is the nature of the crosslinks which cause a variable proportion of
the rubber hydrocarbon to be insoluble in organic liquids?
As regards the first of these questions, it seems to be generally agreed that

the gel content of the rubber hydrocarbon in very fresh natural rubber latex
is small but variable. For very fresh latex from trees in regular tapping, it
may be effectively zero. It also seems to be generally agreed that the gel
content of the rubber particles increases as the latex ages, whether the ageing
takes place in the tree or outside of it. It appears that a crosslinking reaction
commences immediately after tapping. Allen and Bristow [70] postulate that
the rate of crosslinking is dependent upon the entry of some chemical species
into the latex particles from the aqueous phase, so that the level of crosslinking
is greatest in the smallest particles. There is certainly much experimental
evidence that the gel in natural rubber hydrocarbon is predominantly present
in the form of very small particles. Thus Freeman [72] has reported that the
size of the gel particles in the rubber from fresh natural rubber latex is ca.
lOOnm, whereas the average size of the particles is ca. lOOOnm. If the
solvent-insoluble fraction ofthe natural rubber hydrocarbon is predominantly
present as very small particles, then this insoluble fraction is appropriately
referred to as microgel. The latter term is used to denote the gel fraction of
latex polymers which comprises aggregates of crosslinked polymer molecules
which together are of a size comparable to that of the particles in the latex
from which the polymer was obtained. As regards crosslink concentration,
Allen and Bristow picture natural rubber latex as comprising a mixture of

particles for which the crosslink concentration decreases as the particle size
increases, but in anyone particle increases as the latex ages. According to
this picture, the natural rubber latex concentrates used industrially contain
small highly-crosslinked particles, large particles which contain few if any
crosslinks, and particles of intermediate size which contain intermediate
concentrations of crosslinks. The manner in which the crosslinks are
distributed within and between the particles in any given sample of natural
rubber latex will depend upon the age and history of the latex.
It remains to consider the third of the above questions, namely, the nature
of the crosslinks which cause a variable proportion ofthe rubber hydrocarbon
to be insoluble. These crosslinks clearly differ in chemical nature from those
which are normally present in crosslinked synthetic polymers, typified by the
carbon -carbon crosslinks which form when a diene rubber is crosslinked by
heating with a peroxide. It has long been recognized that the natural rubber
molecule does not consist exclusively of carbon and hydrogen atoms, but
also contains minor amounts of oxygen, probably in the form of carbonyl
and ester groups. The present view is that a substantial proportion of the
microgel in natural rubber forms by polar interaction between these groups
attached to the rubber hydrocarbon macromolecules and the proteinaceous
macromolecules. Allen and Bristow [70] have suggested that the microgel
particles in solid dry natural rubber are themselves bound together in a loose
network by association with proteinaceous substances. This view of the nature
of the crosslinks in natural rubber gel appears to be consistent with the
observations. The disproportionate presence of nitrogenous substances in
the gel is then readily explained. The dependence of the apparent gel content
upon the nature of the extracting solvent is explicable as a consequence of
differing abilities to disrupt the polar associations between the rubber and
proteinaceous macromolecules. The inverse correlation between apparent gel
content and diffusion coefficient in rubber indicates a requirement that a
solvent molecule should be able to penetrate the macromolecular network
intimately if it is to be effective in disrupting these polar associations. The
addition of polar liquids such as alcohols reduces the apparent gel content
because these liquids are particularly effective in disrupting these polar
associations. Enzymatic deproteinization reduces the apparent gel content
to very low levels because the proteinaceous crosslinks between the rubber
macromolecules are destroyed.
The molecular mass of the gel fraction of the natural rubber hydrocarbon
is, of course, infinite. The molecular masses of the macromolecules which
comprise the sol fraction are distributed over a wide range. Figure 9.22(a)
shows a molecular-size distribution for the sol fraction of natural rubber
hydrocarbon reported by Shiibashi [71]. It is difficult to draw conclusions
as to the range of molecular masses from these results, except that it is very
wide. The distribution is unimodal, having a peak at a molecular mass of
88 Natural latices
30 10 7
-- c:
*
0
:l "'iii
Q).,=, 20 106~
c: c: I/)
.- :l I/)
c:
o >- Ol
~
:;:; Ol E
~ .~
C
Q)
-eOl 10 10 5 ~(J
(J Q)
c: (5
0
l) ~
0 104
60 30 103 104 105 106 107
(a) GPC count Molecular mass / dalton
(b)
c: 2
~
--
'f
""0
><
I/)
I/)
Ol
E
tti
:5
(J
Q)
(5
~ 00 10 20 30
(c) Time /day
Figure 9.22 (a) Molecular-size distribution for the sol fraction of natural rubber hydrocarbon
(Shiibashi [71 ]). (b) Cumulative molecular-size distribution for the rubber hydrocarbon in natural
rubber latex concentrate (Gazcley and Mente [73]). (c) Effect of enzymatic deproteinization
upon the mass (weight)-average molecular mass (points e) and number-average molecular mass
(points A) for the sol fraction of natural rubber hydrocarbon (Shiibashi [71]).
ca. 106 daltons. Shiibashi reports results which confirm the generally-accepted
view that the number-average molecular mass for the sol fraction of natural
rubber hydrocarbon is ca. 3 x 10 5 daltons. This corresponds to a macromolecule
which contains ca. 5 x 10 3 isoprene units. The value Shiibashi reports for
the mass (weight)-average relative molecular mass is much higher, being ca.
1.8 x 106 daltons. Thus the ratio MwlMn is ca. 6. This is high, and confirms
that the natural rubber hydrocarbon macromolecules are very polydisperse.
The cumulative molecular-size distribution shown in Fig. 9.22(b) for the
rubber hydrocarbon in natural rubber latex concentrate has been published
by GazeJey and Mente [73]; it shows the cumulative mass fraction of
macromolecules as a function of molecular mass. The curve representing this
distribution terminates at a mass fraction which is far short of unity, indicating
that the rubber contained gel. Gazeley and Mente state that the fractional
gel content was ca. 0.4 using tetrahydrofuran as the solvent. There is inevitably
some uncertainty as to the nature of the distribution in the region of molecular
mass where the transition from sol to gel is occurring. The distributions
published by both Shiibashi and Gazeley and Mente show that the molecular-
size distribution for the sol fraction of the rubber hydrocarbon in natural
rubber latex effectively terminates when the molecular mass exceeds ca.
5 x 106 daltons, corresponding to a degree of polymerization of ca. 105 (i.e.
some 20 times the number-average degree of polymerization for the sol
fraction), and that most macromolecules of higher molecular mass have
become incorporated in the gel fraction of the polymer.
Shiibashi [71J has also reported the interesting observation that the mass
(weight)-average molecular mass for the sol fraction of natural rubber
hydrocarbon (but not the number-average molecular mass) is greatly reduced
when natural rubber latex is subjected to enzymatic deproteinization, as well
as the gel content of the rubber. The results are reproduced here as Fig.
9.22(c). Over a period of ca. 3 weeks treatment, Mw was reduced to a value
of ca. 6 x 10 5 daltons, which remained approximately constant thereafter.
The polydispersity, as quantified by the ratio Mw/Mn' was reduced to ca. 3.
The implication of these observations is that the natural rubber hydrocarbon
macromolecules of apparently very high molecular mass are formed by
interaction between smaller hydrocarbon macromolecules and certain
proteinaceous substances in the latex. These interactions are essentially the
same as those which have been invoked to explain the formation of most of
the gel fraction of the natural rubber hydrocarbon (see above).
Another consequence of the presence of a minority of polar groups attached
to the natural rubber hydrocarbon is that further crosslin king of the rubber
hydrocarbon macromolecules gradually occurs in both the latex and dry-
rubber states, possibly by interaction between the polar groups and active
methylene groups on adjacent polymer chains. This crosslinking contributes
significantly to the irreversible hardening which occurs in natural rubber
during prolonged storage. It has been shown by Sekhar [74J and by Wren
[75J that hardening arising from this cause can be inhibited by the addition
of monofunctional carbonyl reagents, such as hydroxylamine, or of twice the
equivalent of difunctional carbonyl reagents, such as 5,5-dimethyl-l,3-
cyclohexanedione (dimedone). If less than two equivalents of the latter are
added, then, as expected, the hardening is enhanced by crosslinking through
the added reagent. A further cause of irreversible hardening in natural rubber
arises from the possible presence in natural rubber latex of hydroperoxides
which form redox systems with some of the other non-rubber constituents.
The consequences can be both crosslin king and chain scission. These redox
systems may be activated by the addition ofpolyamines, such as tetraethylene-
pentamine, and inhibited by the addition of formaldehyde. The actual degree
of hardening which occurs during storage is therefore a balance between the
crosslinking which is induced by the redox systems, supplemented by that
which occurs through the carbonyl groups, and the chain scission which
results from the effects of the peroxides.
90 Natural latices
9.5.2.2.3 Nature of the particle surface

The outer regions of the surface of the particles in fresh natural rubber latex
are believed to be proteinaceous in nature. Indeed, the protein component
of the natural rubber latex particle is believed to be exclusively associated
with the particle surface as an adsorbed layer. The principal component of
the adsorbed protein layer is probably identical with, and presumably in
equilibrium with, the (X-globulin which is dissolved in the aqueous phase (see
section 9.5.3.2 below). It is this adsorbed layer which determines the electric
charge carried by the particles, the electrophoretic behaviour of the particles,
and the behaviour of the latex when it is colloidally destabilized. Indeed, the
inference that the adsorbed protein is similar to the dissolved protein has
been drawn from the similarity between the electrophoretic behaviour of the
particles and that of the dissolved proteins.
Electrophoresis has been an important technique for investigating the
nature of the surface of the particles in natural rubber latex. Important
investigations include those of Kemp [76], of Kemp and Twiss [77], of
Hauser and Bender [78], of Blow [79], of Bowler [80], and of Ho and Ng
[81]. Hauser and Bender have reported an electrophoretic mobility of
7.7 Jlm s - 1 (V cm - 1) - 1 for the particles in ammoniated field latex at pH 9.9
and rubber content 0.095% m/m. From this, they calculated the electrokinetic
potential at the rubber-aqueous phase interface to be - 97 mV. As is to be
expected, the pH of the aqueous phase has a profound effect upon the
electrophoretic mobility of the particles in natural rubber latex. Progressive
reduction of the pH causes the electrophoretic mobility to decrease numerically
to zero, and then to increase in the opposite direction. Thus the surface of
the particles is amphoteric in nature, and the particles themselves have an
isoelectric point as regards electrophoretic behaviour. Results reported by
Bowler are reproduced as Fig. 5.2(b) of Chapter 5 (Volume 1), from which it
is evident that the isoelectric point for the particles is at pH ca. 4.1. Hauser
and Bender have found that, at very high pH, the electrophoretic mobility of
the particles decreases again numerically; this they attribute to the adsorption
of counterions (in this case cations) on to the particle surface, with consequent
reduction in the electrokinetic potential. Hauser and Bender have also shown
that the addition of an inert electrolyte, such as sodium chloride, to natural
rubber latex causes a drastic reduction in both the electrophoretic mobility of
the particles and the electrokinetic potential at the particle surface.
Ho and Ng [81] have reported the results of an investigation of the
electrophoretic behaviour of the particles in ammonia-preserved natural
rubber latex which had been concentrated by centrifugation, allowed to stand
undisturbed for 6 months, and then subjected to extensive 'purification' by
dialysis against dilute ammonia solution at pH 10.5. Dialysis was continued
until the electrophoretic mobility of the particles became constant. This
required some 25-30 changes of dialysate once every 24 hours; the dialysate/latex
ratio was ca. 15/1 v/v. The use of dilute ammonia solution rather than pure
water for dialysis ensured that weak-acid groups, in particular carboxylate

groups, remained ionized during the dialysis process. Aqueous ammonia
solution of pH 1O.S is comparable in this respect to the aqueous phase of a
high-ammonia natural rubber latex concentrate. Although precise estimates
are not possible because Ho and Ng do not specify the temperature at which
the dialysis was carried out, it is clear that there would have been a large
reserve of un dissociated ammonia available for reaction with acidic substances.
The latex investigated by these workers was therefore such that most of the
changes which occur in natural rubber latex when it is ammoniated (see
section 9.S.S below) are likely to have occurred, that most of the water-soluble
non-rubber constituents had been removed from the aqueous phase, and that
most of the desorbable water-soluble non-rubber constituents adsorbed at
the surface of the particles had been desorbed into the aqueous phase and
subsequently removed from the latex. Thus in assessing the significance of
the results reported by these workers, it must be borne in mind that they
relate not to the particles in fresh natural rubber latex, nor to the particles
in ammonia-preserved natural rubber latex as normally encountered, but to
the particles in ammonia-preserved natural rubber latex which has been
extensively 'laundered' to remove all labile water-soluble substances. These
results may therefore be regarded as characterizing the surface ofthe 'inherent'
particle in ammonia-preserved natural rubber latex, in the sense that this is
the particle from which nothing else can be removed by processes which are
commonly used to remove extraneous substances from lyophobic sols. The
effect of extensive dialysis was to reduce the ash content of the latex by ca.
80% m/m, but the nitrogen content by only ca. 33% m/m, thus indicating
that nitrogenous substances are tenaciously bound to or within the particles
in natural rubber latex. The particle diameters in the latex samples used
ranged from ca. SOO nm to ca. 2000 nm, with xn = 1036 nm.
The results reported by Ho and Ng [81] can be summarized as follows:
I. There was some variation of electrophoretic mobility amongst the particles,

the distribution at pH 7.2 being as shown in Fig. 9.23(a). The distribution
is unimodal, and such as to imply that the population of particles was
homogeneous as regards electrophoretic mobility. This in turn implies
that, notwithstanding the wide range of particle sizes in ammonia-preserved
natural rubber latex concentrate, the particle surfaces are inherently uniform.
2. The average value for electrophoretic mobility at pH 7.2 was found to be
3.Sllms- l (Vcm-l)-l.
3. The electrophoretic mobility was essentially independent ofthe concentration
of rubber particles down to ca. 2.3 x 10-4% m/m dry rubber content.
4. The variation of electrophoretic mobility with pH was as shown in Fig.
9.23(b). These results confirm the amphoteric nature of the 'inherent' rubber
particle in ammonia-preserved natural rubber latex. They are consistent
with the presence of proteinaceous substances at the particle surface,
92 Natural latices
-4
-2.7 -3.1 -3.5 -3.9 -4.3 22~---'----..L6---..L8---.j10

Electrophoretic mobility / 11m S·1(V cm·\1 Surface pH
(a) (b)
Figure 9.23 (a) Distribution of electrophoretic mobility at pH 7.2 amongst the particles of
dialysed natural rubber latex (Ho and Ng [81]). (b) Effect of pH upon electrophoretic mobility
of particles in dialysed natural rubber latex (Ho and Ng [81 ]).
notwithstanding that the latex had been subject to alkaline hydrolysis at

normal ambient temperature for some 6 months. However, these results
give the isoelectric pH of the particles as 3.8. This is lower than that
reported by other workers, and outside the lower limit (pH 4.0) usually
observed for proteins. Ho and Ng suggest the reason as tightly adsorbed
long-chain carboxylate anions. They have reported an infra-red spectrum
of films from the dialysed latex which is consistent with this hypothesis.
5. The electrophoretic mobility reached its maximum value when the pH
was ca. 6 (Fig. 9.23(b», and did not increase significantly thereafter. Ho
and Ng calculated a value of 4.22 for the pK g of the surface ionogenic
groups. This is consistent with those groups being of the carboxylic-acid
type. The conclusion is drawn that the surface was unlikely to have
contained ionogenic phosphate groups, because a value of pK g in the
range 1.0-2.0 would then have been expected.
6. An electrokinetic potential of - 44.5 m V was calculated from the maximum
and constant electrophoretic mobility observed above pH 6. From this,
a surface electric charge density of 1.43 x 10 - 2 C m - 2 was inferred. This
corresponds to an area of ca. 11.2 nm 2 per electronic charge, and to an
average charge separation of ca. 3.35 nm. It must be noted that this surface
charge density refers to ammonia-preserved natural rubber latex which
had been extensively dialysed; it is not possible to infer the charge density
for the particles in normal ammonia-preserved natural rubber latex from
these results.
The overall conclusion to be drawn from this investigation is that the particles
in ammonia-preserved natural rubber latex certainly have a low surface

concentration of strongly-bound ionogenic groups which are amphoteric in
nature. These groups are probably associated with the presence of surCace-
bound nitrogenous substances which are probably proteinaceous in nature.
Ho [82] has also investigated the effect of dissolved metal cations upon
the electrophoretic mobility ofthe rubber particles present in dialysed ammonia-
preserved natural rubber latex. The effects of non-hydrolysable cations
(sodium, calcium, magnesium and barium) were consistent with mere
enhancement of the ionic strength of the aqueous phase. The effects of
hydrolysable cations, such as copper(II) and cerium(III), were more complex:
charge-reversal, interpreted as a consequence of specific ion adsorption, was
observed. In a further paper, Ho [83] has reported the effects of surface
chemical modifications upon the electrophoretic properties of the particles
in the dialysed latex.
The lipids associated with the rubber particles in fresh natural rubber latex
comprise sterols and sterol esters (ca. 0.4% m/m), fats and waxes, such as
eicosyl alcohol (ca. 0.6% m/m), and phospholipids (ca. 2.0% m/m). The sterols,
sterol esters, fats and waxes occur predominantly within the bulk of the
particles, being probably dissolved in the rubber hydrocarbon. The phospholipids
are strongly adsorbed at the surface of the rubber particles. It has been
suggested that they function as intermediaries by which the proteins are
strongly anchored to the particle surface. Although the proteins themselves
would adsorb at a hydrocarbon-water interface, they would not be expected
to be as strongly bound to the surface as they appear to be at the surface
of the particles in fresh natural rubber latex.
The principal phospholipids associated with the particles in fresh natural
rubber latex are of the lecithin type. The chemical structure of this type is
shown as X below, in which the R groups are long-chain hydrocarbon moieties
o
CH 2- O - t-O-CH2CH2.~(CH3)3
I
OH HO-
x
such as n-heptadecyl (n-C I7 H 3S ) or n-heptadecenyl (n-CI7H33)' The presence
of these long-chain hydrocarbon moieties in a molecule which also contains
a distant polar and permanently-ionized site makes the molecule strongly
surface-active. This is the reason why the molecule is strongly adsorbed at
94 Natural latices
the rubber-aqueous phase interface. The hydrocarbon moieties are anchored

to the rubber surface; the polar and permanently-ionized sites tend to mix
with the aqueous phase. The strong tendency for the adsorbed lecithin to
associate with proteins and so bind the latter to the particle surface may be
a consequence ofthe lecithin molecule carrying a slight overall positive charge
at the pH of fresh latex (ca. 6.5), whereas the proteins, being on the alkaline
side of their isoelectric points; are negatively charged; thus ionic association
occurs between the two types of molecule. Figure 9.24(a) demonstrates that
this is theoretically possible if appropriate values are chosen for the acid-
and base-dissociation constants of the two types of molecule. Predictions are
shown for the effect of pH change upon the overall degree of ionization of
a lecithin and of a protein for which the ratio of free carboxylic-acid
groups to free amino groups is Ill, taking pKa = 4 and pKb = 00 for the
lecithin, and pKa = 3 and pKb = 8 for the protein (implying an isoelectric
point of ca. pH 5). In acid medium, both molecules are strongly positively
charged. As the pH increases, so the overall charge on the molecules decreases,
but the decrease is more rapid for the protein than for the lecithin. If the pH
0.8
0.6
c: Rubber hydrocarbon
0 0.4
~
'c
.Q
"0 0.2
QI
I!!
CI
QI
"C 8
~II)
~ -0.2
QI
CI
!!! Phospholipid
~ -0.4 Protein
C(
(b)
-0.6
<8> -0.8
Figure 9.24 (a) Theoretical predictions for effect of pH change upon state of ionization of a
lecithin, taking pK. = 4, pK. = 00 (curve A) and of a protein for which the ratio of free
carboxylic-acid groups to free amino groups is 1/ I, taking pK. = 3, pK. = 8 and pKw = 14.38
(curve B). (b) Representation of rubber particle in fresh natural rubber latex as comprising
mainly rubber hydrocarbon, surrounded by an inner shell of phospholipid and an outer shell
of protein.
is increased sufficiently, the polarity of the ionized protein can be reversed

by suppression of the ionization of its amino groups. Reversal of polarity is
not possible with a lecithin (except for a very slight effect at high pH) because
its basic functionality is a permanently-ionized quaternary ammonium group
(pKa = 4 and pKb = (0). The ionization behaviour of proteins in acid and
alkaline media is discussed in section 16.8.3.1 of Chapter 16 (Volume 3),
where the relevant reactions are given. The behaviour of lecithins is probably
to be interpreted in terms of the following reactions, in which MOH represents
an alkali and HX an acid:
CH 2 -O.CO.R CH 2 -O.CO.R
I I
CH 2 -O.CO.R CH 2 -O.CO.R
I '+ +
CH 2 -O- P-O-CH 2 CH 2 ·N(CH 3 )2
I '+ +
CH 2 - O - P-O-CH 2 CH 2 ·N(CH 3 )2
I I
OH X- O-M+ HO-
Acidic medium Alkaline medium
Figure 9.24(b) shows a crude representation of the rubber particle in fresh

natural rubber latex as comprising mainly rubber hydrocarbon, surrounded
by an inner shell of phospholipid and an outer shell of protein. It is more
likely that the phospholipid and protein molecules are much more intimately
mixed together over the particle surface than is suggested by this picture,
but nevertheless this picture does emphasize the function of the phospholipid
as an adhesive which bonds the protein to the rubber surface. Other
phospholipids which are found associated with the particles of fresh natural
rubber latex include metal phosphatides (Structure XI) and cephalins
(Structure XII). It is commonly believed that the electric charge carried at
the surface of the particles in fresh natural rubber latex is mainly attributable
to the presence of bound carboxylate anions derived from adsorbed
proteinaceous substances.
XI XII
96 Natural latices
9.5.3 The aqueous phase
9.5.3.1 Carbohydrates
The aqueous phase of fresh natural rubber latex is a dilute aqueous solution
of density ca. 1.020 Mg m - 3 which contains many different chemical species,
of which the principal types are carbohydrates, electrolytes, proteins and
amino acids.
The principal carbohydrate present in the aqueous phase of fresh natural
rubber latex is a substance known as I-methyl inositol or quebrachitol
(Structure XIII). It occurs to the extent of ca. 1% m/m on the whole latex.
The chief point of interest about it is that, so far, no use has been found for
it. Large amounts of this substance are thrown away annually. Other
carbohydrates which are present in small amounts include galactose, sucrose,
glucose, fructose, and various other inositols. The biological significance of
these carbohydrates (and of quebrachitol in particular) is obscure. They have
little influence upon the properties of the latex or of the rubber which it
contains. In the absence of adequate preservation, the carbohydrates become
microbiologically oxidized to the so-called volatile fatty acids, which comprise
mainly formic, acetic and propionic acids. This matter is considered in more
detail in section 9.5.5.3 below.
/CH~H3
CHOH CHOH
I
CHOH
I
CHOH
~/
CHOH
XIII
9.5.3.2 Proteins and amino acids

The aqueous phase of fresh natural rubber latex contains several proteins of
differing isoelectric points. The principal proteins are known as a-globulin
and hevein. a-globulin can be obtained by precipitation from a dialysed
solution of freeze-dried aqueous phase in sodium citrate buffer solution of
pH 4.5 and ionic strength 4 x 1O-2 moldm- 3 . To prevent denaturation of
the protein, it is necessary to work at o·e and to avoid keeping the protein
at low pH for longer than is essential. a-globulin is a surface-active protein
of molecular mass ca. 2 x lOs daltons. It is readily adsorbed at air-aqueous
phase and oil-aqueous phase interfaces, with concomitant lowering of the
respective interfacial free energies. It is insoluble in distilled water, but is
soluble in neutral salt solutions, acid solutions and alkaline solutions. Its
isoelectric point is at pH 4.8; this is close to that for the particles in fresh
natural rubber latex. Ammoniation increases the electrophoretic mobility of
both ex-globulin and latex particles. Fresh natural rubber latex most readily
undergoes gross colloidal destabilization under the conditions of pH at which
ex-globulin is least soluble in aqueous media. The similarities between the
electrophoretic behaviour of dissolved ex-globulin and that of the particles in
fresh natural rubber latex, and between the colloid behaviour of dissolved
ex-globulin and that of the particles in fresh natural rubber latex, provide the
principal evidence for believing that this protein is an important component
of the protein layer which is bound to the surface of the rubber particles.
ex-globulin is denatured by heating and by storage in the dry state.
Hevein has been isolated by ammonium-sulphate fractionation of the
freeze-dried solids which are derived from the so-called bottom fraction of
natural rubber latex (see section 9.5.4 below). After reprecipitation and the
removal of salts by electrodialysis, the hevein can be freeze-dried and
crystallized from cold water. It is a crystallized protein of abnormally low
molecular mass (ca. 1 x 104 daltons). Its isoelectric point is at pH 4.5. It
contains ca. 5% of sulphur as cystine-type linkages. Hevein displays little
surface-activity, is soluble in water at all pH values, and is not precipitated
from water by boiling. It is unlikely to affect the colloidal properties of natural
rubber latex significantly.
Several other proteins are present in the aqueous phase of fresh natural
rubber latex besides ex-globulin and hevein. Thus the presence ofthree protein
fractions was demonstrated by Bishop [84] and confirmed by Kemp and
Straitiff [85]. Roe and Ewart [86] have reported an electrophoretic analysis
of fresh natural rubber latex aqueous phase; this showed the presence of at
least seven distinct protein components having various isoelectric points.
This finding has been confirmed by Moir and Tata [87], who used the
technique of paper electrophoresis. These workers have also investigated the
proteins which are contained in the bottom fraction of fresh natural rubber
latex. Especial interest attaches to the proteins of high isoelectric point,
because these may be cationic under the conditions of pH which prevail in
the aqueous phase of fresh natural rubber latex. They will therefore tend to
reduce the colloid stability of the latex.
Archer and McMullen [88] have shown that at least one nucleoprotein,
known as hevea ribonucleoprotein, is present in the aqueous phase of fresh
natural rubber latex. It can be extracted by neutral salt solution from fresh
latex serum solids after acid-soluble nucleotides and other substances of low
molecular mass have been removed. It can be purified by repeated precipitation
in 66% v/v ethanol,and also by treatment with acid at pH 3.5-4.0. Extraneous
contaminating proteins can be removed by precipitation with cetyltrimethyl-
ammonium bromide. The final product is homogeneous in the ultracentrifuge,
and comprises ca. 70% m/m of a polynucleotide of molecular mass ca.
98 Natural latices
2.8 x 104 daltons and ca. 30% mlm of a protein of molecular mass ca.
11.4 x 103 daltons. The base composition is an unusual combination of
aden ide, uracil, cytosine and guanine, with no fifth component.
A multiplicity of polypeptides and simple amino acids has been observed
in the aqueous phase of fresh natural rubber latex, but whether these are the
precursors or the degradation products of the latex proteins (or possibly
both) is not clear. A useful review of this subject, with some extension, was
provided by Ng [89] in 1960, from which paper Tables 9.10 and 9.11 have
been reproduced in modified form. Table 9.10 summarizes the amino acids
which had been identified in the aqueous phase of fresh natural rubber latex
up to that time, and Table 9.11 the amino acids which had been identified
in the hydrolysis products of latex proteins. The free amino acids comprise
together ca. 0.1 % mlm of the mass of the whole latex.
There has been increasing concern in recent years regarding human
allergenic reactions to certain proteinaceous substances which are present in
films obtained from natural rubber latex. The most serious reaction attributed
to contact with latex proteins is that known as anaphylactic shock. It is
characterized by a severe reduction in blood pressure, breathing difficulties,
increased heart-beat rate and unconsciousness. It can result in death. Other
reactions include contact urticaria, which is a temporary eczematous reaction
in which weals or flares form at the contact site, with itching or stinging.
These reactions are quite distinct from those attributable to rubber compounding
Table 9.10 Free amino acids which have been identified in the aqueous phase of
fresh natural rubber latex (Ng [89])
McGavack Whitby
and and
Rumbold Greenberg
Amino acid [90] Altman [91] [92] Drake [93] Ng [89]
Glycine No Yes No Yes Yes

Alanine Yes Yes No Yes Yes
Valine No Yes Yes No Trace
Leucine No Yes Yes Yes Yes
Isoleucine No Yes Yes Yes Yes
Phenylalanine Yes Yes Yes No Yes
Tyrosine No Yes Yes No Yes
Aspartic acid No No Yes Yes Yes
Glutamic acid No Yes No Yes Yes
Arginine No No Yes No Trace
Proline No Yes Yes No Trace
Lysine No No No No Yes
Cystine No Yes No No Yes
Serine No No No Yes Trace
Tryptophan No No No No Yes
Threonine No No No No Trace
Table 9.11 Amino acids which have been identified in the hydrolysis products of
natural rubber latex proteins (Ng [89])
Midgely, Whitby
Henne and
and Green-
Amino Belgrave Recoil Altman Tristram berg Drake
acid [94] [95] [96] [97] [92] [93]
Glycine Yes No No Yes Yes Yes

Alanine No No Yes No No Yes
Valine No No Yes No No Yes
Leucine No Yes Yes Yes No Yes
Isoleucine No No No No No Yes
Phenyl-
alanine No No No No No Yes
Dihydroxy-
phenyl-
alanine No No Yes No No No
Tyrosine No Yes Yes No Yes Yes
Aspartic
acid No Yes Yes Yes Yes Yes
Glutamic
acid No Yes Yes No Yes Yes
Arginine Yes Yes Yes Yes Yes Yes
Histidine Yes Yes Yes Yes Yes Yes
Lysine Yes No No Yes Yes Yes
Proline Yes No Yes Yes Yes Yes
Hydroxy-
proline No No Yes No No No
Tryptophan No Yes No No No No
Cystine Yes Yes No No No No
Ornithine No No No No No Yes
ingredients which may have been incorporated during the manufacture of

the product (see further sections 16.3 and 16.3.7 of Chapter 16 (Volume 3)).
They are, as it were, inherent to natural rubber latex itself, being attributed
to the presences of water-soluble proteinaceous substances in the final
product. Thus extensive leaching of vulcanized latex films considerably
alleviates the problem. Natural rubber articles manufactured from dry rubber
do not appear to give rise to these problems. Several possible reasons are
recognized for the recent increased perception of this problem in respect of
products manufactured from natural rubber in latex form. Greater use of
films from natural rubber latex in the form of gloves and condoms for
protection of humans against life-threatening microorganisms has increased
exposure to allergenic substances in the films. Greater demand for protective
products may have led to more rapid manufacture, with less-thorough
leaching. Increasingly, such products are being manufactured in countries
100 Natural latices
where natural rubber latex is produced. Exposure of proteinaceous substances

to the ammonia preservative is therefore reduced. It is possible that prolonged
exposure to ammonia may render such substances less allergenic. Variations
in latex composition caused by clonal, seasonal and environmental changes
may also be a factor, as may improved skill in diagnosis of allergenic reactions.
There is now an extensive literature relating to this matter. The reader is
referred specifically to a review by Levy et al. [98] for information concerning
medical aspects, and to a paper by Dalrymple and Audley [99] for information
concerning the factors which influence the level of extractable proteinaceous
substances present in products manufactured from natural rubber latex.
9.5.3.3 Other substances present in the aqueous phase offresh natural rubber latex
Other substances present in the aqueous phase of fresh natural rubber latex
include free nitrogenous bases such as choline and methylamine, organic
acids (additional to amino acids), inorganic anions (principally phosphate
and carbonate), and various metal ions (including those of potassium,
magnesium, iron, sodium and copper). Fresh natural rubber latex is also said
to contain a small amount (ca. 0.25% m/m) of cyanide, which appears to be
present as free hydrocyanic acid. It has been suggested that the cyanide
suppresses aerobic respiration.
There appear to be at least two thiols present in the aqueous phase of
fresh natural rubber latex. These are cysteine and reduced glutathione, both
of which may function as co-factors in some of the reactions which lead to
the formation of the rubber hydrocarbon macromolecule. Both are readily
oxidized to the corresponding disulphides, cystine and glutathione, which
have also been detected in the aqueous phase of fresh natural rubber latex.
The ratio of reduced to oxidized states is usually ca. 10/1, but this is subject
to considerable variation. There is also some evidence that this ratio can be
significantly altered in vivo by the application of certain growth regulators
and stimulants to the bark and leaves of young Hevea brasiliensis seedlings.
A wide range of enzymes is also present in fresh natural rubber latex.
Presumably these are located mainly in the latex aqueous phase. It appears
that these substances, originally present in the latex vessels, efflux from the
tree with the latex. In particular, fresh natural rubber latex has the
characteristic of being able to assimilate various simple inorganic substances
and convert them into rubber hydrocarbon (see further section 9.7.3 below).
This suggests that the complex enzyme systems which are responsible for
the biosynthesis of the natural rubber hydrocarbon macromolecule are
amongst those which efflux with the latex. Information concerning the
enzymes which are present in freshly-tapped natural rubber latex, and the
related matter of oxygen absorption by the latex, is available in a paper by
Hsia [100]. References to earlier work are also given.
9.5.4 Lutoids and other particulate phases
Lutoids (or viscoids, as they have also been called) were first reported to be
present in fresh natural rubber latex by Homans and van Gils [101]. These
workers centrifuged undiluted fresh latex at low speeds, thereby causing the
latex to separate into two fractions, a white fraction which contained most
ofthe rubber particles, and a bottom yellow fraction. At the boundary between
the two fractions was a thin vivid yellow layer. The volume of the yellow
fraction comprised ca. 20-30% of the whole. Examination of this fraction
by optical microscope revealed the presence of ill-defined aggregates, quite
distinct in character to the rubber particles. They were called lutoids because
it was thought that they were responsible for the yellow colour of the bottom
fraction. The presence of more well-defined spherical particles, interpreted
as being unaggregated lutoids, was subsequently reported by Ruinen [102].
These are rather larger than typical rubber particles, and comprise a fluid
substance bounded by a membrane. Lutoids are sometimes seen to contain
inclusions which exhibit Brownian motion. The structure of lutoids has been
described in detail by Dickenson [103]. The boundary membrane has a
single-layer structure. Inside is an aqueous solution which contains dissolved
substances. Fibrils of a proteinaceous nature are suspended in this solution.
The appearance of these varies, being sometimes coarse and sometimes fine.
The fibrillar contents seem to be fairly specific to the latex obtained from
young vessels. It is therefore inferred that they gradually disappear as the
particles age. The fibrils are seen against a background network which is
probably also fibrillar. They are seen to be arranged in a regular hexagonal
pattern when viewed in cross-section. The individual fibrils are probably
hollow, with a double-helix structure. They vary in length. Some are
considerably longer than the particles which contain them. The apparent
absence of observable fibrils in the lutoids of the latex of the species Hevea
benthamiana is consistent with the view that they are not involved in the
biosynthesis of the rubber hydrocarbon macromolecule.
Chemically, the lutoids appear to be mainly water. They contain, in
addition, small quantities of soluble protein (ca. 3% m/m), insoluble protein
(ca. 2% m/m) and phospholipid (ca. 0.5% m/m). They are very labile,
disappearing when the latex is ammoniated. They are not therefore found
in ammonia-preserved natural rubber latex. Furthermore, after tapping they
rapidly decompose at tropical temperatures. The interior Brownian motion
slows down and ceases as the inclusions adhere to the enclosing membrane.
Then the membrane ruptures as the fluid contents of the particles discharge
into the aqueous phase. The membranes are left as apparently amorphous
material which readily aggregates, entrapping other particles in the process.
The chemical characteristics of the fibrillar inclusions have been described
by Audley [104]. These inclusions are thought to be entirely proteinaceous
in nature. They contain ca. 15% mlm of nitrogen, ca. 0.6% mlm of sulphur,
102 Natural latices
and less than 0.01 % m/m of phosphorus. They mayor may not have
associated with them small amounts of carbohydrates. The low sulphur
content is inconsistent with the otherwise attractive hypothesis that the fibrils
are crystallized or agglomerated hevein macromolecules. The fibrillar protein
has an isoelectric point of pH ca. 4.0. It is readily coagulated by heat. Analysis
of the amino acids produced by hydrolysis reveals nothing unusual. An
interesting feature is the ease with which the fibrils can be broken up into
fragments, sometimes reversibly and sometimes irreversibly. It was initially
thought that the fragments might be held together by magnesium atoms, but
this hypothesis is inconsistent with the observation that dialysis against a
solution of a powerful sequestering agent for magnesium ions (ethylene-
diaminetetraacetic acid) does not cause rapid fragmentation of the fibrils.
The intact lutoids are osmotically sensitive, in that the addition of distilled
water to the bottom fraction causes them to swell and then burst. It is these
processes which seem to be responsible for the marked viscosity changes
which occur when fresh natural rubber latex is diluted with water, noted in
section 9.5.1 above (see also section 6.3.4 of Chapter 6 (Volume 1)). It was
because of this effect that the alternative name visco ids was proposed for
these particles. The dilution reaction which occurs in the latex vessels of the
tree during tapping may likewise cause the lutoids to swell; the resultant
increase of viscosity may be one of the factors which causes the latex flow to
cease after a time. Dilution of the bottom fraction with isotonic salt solution does
not cause swelling ofthe lutoids. The lutoids and their residues coagulate more
readily than do any of the other particulate phases present in fresh natural
rubber latex. Natural rubber latex from which the lutoids have been removed is
less susceptible to spontaneous coagulation than is whole latex.
The bottom fraction has a slightly lower pH (5.7 to 5.8) than that of fresh
natural rubber latex. It discolours upon exposure to air, and when acid is
added. Such discolorations do not occur in the fresh latex. It is therefore
possible that oxidizing enzymes separate with the bottom fraction, whereas
inhibitors of oxidation separate with the rubber fraction.
The yellow colour of the bottom fraction, and, indeed, of some whole
latices, is primarily a consequence of entrapped Frey-Wyssling particles, and
not to the colour of the lutoids themselves. The Frey-Wyssling particles are
the chief component of the vivid yellow layer which separates the bottom
and top fractions when fresh natural rubber latex is centrifuged. They were
first observed in 1929 by the worker from whom their name derives. They
are spherical, non-rubber particles, often bright yellow in colour. Their colour
is caused by the presence of carotenoid pigments. They are somewhat larger
and slightly more dense than are the rubber particles, and are of high refractive
index. They often appear as clusters of two or three, with an associated
vacuole. Frey-Wyssling 'complexes', as they have been termed, are of
comparatively rare occurrence, although their frequency does depend
significantly upon the clone from which the latex was derived. The yellow
colour of some natural rubber latices is attributed to a relatively high

concentration of these bodies. The Frey-Wyssling particles themselves are
mainly lipid in character. Associated with the vacuolar component of the
Frey-Wyssling 'complexes' are ramified tubular structures characterized by
a spiral thickening. These complexes are not observed in ammonia-preserved
concentrated natural rubber latex. Not only do they tend to be eliminated
from the cream as the latex is concentrated by centrifugation; they also tend
to dissolve in the aqueous phase of the latex when the latex is ammoniated.
The technique of centrifugation, under conditions of both low and high
centrifugal accelerations, has been used as a method of separating fresh
natural rubber latex into its various constituent fractions. Centrifugation at
accelerations of ca. 6 x 104 g brings about separ~tion into four main fractions.
The innermost is the white fraction, or rubber cream, which amounts to ca.
60% v/v of the whole. The next fraction is the bright yellow layer to which
reference has been made above. This is of negligible volume fraction. Then
comes a clear aqueous phase which constitutes ca. 25% v/v of the whole.
The final ca. 15% v/v is what is now regarded as comprising the bottom
fraction. Most of the magnesium content of fresh natural rubber latex is
associated with this bottom fraction. A pioneering investigation of these
various fractions was reported by Moir [105]. A combination of ultracen-
trifugation and staining techniques was used. The results show that both the
sedimentable and the non-sedimentable constituents of fresh natural rubber
latex are far more complex than had hitherto been supposed. In retrospect,
the complexity is not surprising, given that fresh natural rubber latex is the
sap of a plant cell. The latex investigated by Moir was from the clone P8186.
It was collected within 30 minutes oftapping, and was kept as cool as possible.
It was then centrifuged at a maximum acceleration of 53620g for 40 minutes,
the temperature being kept in the range O-SOc. Further delineation of the
components of the latex, together with some indication of their cytological
significance, was achieved by treating the latex with small quantities of
appropriate stains (Janus Green B or Neutral Red) prior to centrifuging.
Eleven distinct zones were distinguished after centrifugation under these
conditions. The details of these zones are illustrated in Fig. 9.25, which shows
approximate plane projections of the pattern which was visible from the
outside of the centrifuge tube. The zones themselves were distinguished by
colour, or by texture, or by both.
Sou thorn [106J has described eight types of particle which occur in fresh
natural rubber latex, although it must be recognized that some of these
particles may have originated from the contents of other cells besides the
latex vessels. These types are as follows:
1. large rubber particles of various shapes, occasionally having long tails,

which are unlikely to have been formed by the agglomeration of smaller
particles;
104 Natural latices
Gas space
r--'--~-,
L-------8------------~
Figure 9.25 Fractionation offresh natural rubber latex into eleven distinct zones by centrifugation
and staining (Moir [105]).
2. small spherical rubber particles having diameters less than 1()()() nm;
3. Frey-Wyssling particles, which are typically yellow spheres of glistening
appearance, often occur in groups, and are variable in their reactions to stains;
4. yellow particles which resemble the Frey-Wyssling particles, but which
centrifuge to a different zone and so must be of different density;
5. grey, spherical vacuolar bodies, which are often associated with one or
more Frey-Wyssling particles;
6. semi-transparent vacuolar particles which are again associated with the
Frey-Wyssling particles;
7. solid spherical particles which are dull-grey in appearance, and which
have diameters in the range 1-5 ~m;
8. lutoids, which SouthorD describes as spheres having diameters in the range
2-10 ~m when fresh, which have included within them many small moving
particles.
Threadlike structures have also been observed throughout centrifuged

natural rubber latex. It has been suggested that these may be the remnants
of an extensive reticulum of hollow threadlike processes which originally
existed within the latex vessels, to which the various particulate phases of
the latex were initially connected. Possibly such a reticulum could be the site
for the biosynthesis of the natural rubber hydrocarbon. Figure 9.26, which
is due to SouthorD [107], illustrates tentative ideas which have been proposed
Figure 9.26 Illustrating possible relationship between the reticulum and the various particulate
phases in unperturbed latex vessels (Southorn [107]). Key: R = rubber particle; L = lutoid;
FW = Frey-Wyssling particle.
concerning the relationship between the reticulum and the various particulate
phases.
9.5.5 Changes in the chemical constitution of natural rubber latex as a

consequence of ammoniation and storage
9.5.5.1 Resume of changes

The chemical constitution of ammonia-preserved natural rubber latex after
ammoniation and storage differs considerably from that offresh, unammoniated
natural rubber latex, quite apart from the presence of the ammonia itself.
The quantities of both proteins and resinous substances tend to decrease as
they are converted to polypeptides, amino acids, soaps, glycerol, organic
bases and inorganic phosphate. The overall concentration of ions present in
the aqueous phase tends to increase, particularly the concentrations of
carboxylate anions of high and low molecular masses, phosphate anions and
amino-acid anions. The equivalent of cations is derived mainly from the
ammonia which has been added to the latex as a preservative. The
concentration of ionic magnesium in the aqueous phase is greatly reduced,
except in the case of certain clonal latices, such as Glenshiel I, which give
colloidally-unstable preservates (see section 9.3.3.3.3 above). These chemical
106 Natural latices
changes have significant consequences for the colloidal properties of natural

rubber latex, and for the processing characteristics, but not, it appears, for
the technological properties of films of vulcanized natural rubber obtained
from the latex. In what follows, it is convenient to consider first the chemical
effects of ammoniation upon natural rubber latex, then the problem of the
formation of the anions of the so-called volatile fatty acids (VF As), and then
the effect of prolonged storage upon the colloid stability of ammonia-preserved
natural rubber latex. These matters are of considerable industrial importance,
because several months may elapse between the production of preserved and
concentrated natural rubber latex and its eventual use. During this time, it
will have been transported and stored under various conditions oftemperature
and aeration, and it will also have been subjected to mechanical agitation
and aeration during transfer from one storage container to another. Papers
relevant to the matters under consideration here include those of McGavack
and Bevilacqua [lOS], of Collier [109], ofVerhaar [52], of Lowe [110,111,19],
of Pillai [112], of Sinclair and Janssen [113], and of Chin, Singh and Loke
[114]. Experimental investigations ofthe effects of ammoniation and subsequent
storage upon the chemical constitution of natural rubber latex have been
concerned principally with changes in three properties: mechanical stability,
concentration of volatile fatty acid anions, and total concentration of cations,
as assessed by the so-called KOH number (see section S.S of Chapter 8 (Volume I)).
9.5.5.2 Chemical effects of ammoniation

The principal chemical effect of the ammoniation of fresh natural rubber
latex is to promote hydrolysis of the proteins and lipids, these reactions being
a concomitant of the alkalinity which ammonia imparts to aqueous phases.
An increase in the pH from 6.5 to 10.5, which is typical for the ammoniation
of natural rubber latex, implies an increase in the concentration of hydroxyl
ions by a factor of 104 . Thus any reactions in which these ions participate
are expected to be greatly accelerated by ammoniation.
Hydrolysis of the proteins leads to the formation of polypeptides and
amino acids. Hydrolysis of the phospholipids leads to the formation of various
substances such as glycerol, long-chain carboxylate anions, phosphate anions,
and organic bases. In the case of an ex-lecithin, the complete alkaline-hydrolysis
reaction can be represented as follows:
The liberated long-chain carboxylate anions are adsorbed at the particle

surfaces (presumably by remaining in situ as the lipid molecule undergoes
hydrolysis), thereby enhancing the colloid stability of the latex. This is believed
to be the principal reason why the colloid stability of ammonia-preserved
natural rubber latex tends to increase during storage. Furthermore, it is
usually assumed that the electric charge carried at the surface of the particles
in alkali-preserved natural rubber latex is mainly a consequence of the
presence of adsorbed long-chain carboxylate anions derived from phospholipids
by hydrolysis. However, there is some experimental evidence which indicates
that much of the surface-bound electric charge on the particles in ammonia-
preserved natural rubber latex may be a consequence of adsorbed proteinaceous
substances, the balance arising from adsorbed phosphate anions derived from
phospholipids by partial hydrolysis in which only the choline moiety is
removed. The lipid-hydrolysis reactions appears to be independent of the
conditions of storage (i.e., whether aerobic or anaerobic), or of the extent to
which the anions of the volatile fatty acids are formed simultaneously. On
the other hand, the difference between the KOH number and the VF A number
(see sections 8.8, 8.9 and 8.10.4 of Chapter 8 (Volume I)) is greater when a
latex is stored under aerobic conditions than when it is stored under anaerobic
conditions. Thus possibly non-volatile fatty and other acids are formed by
additional processes under aerobic conditions. Some enzymes are also able
to bring about the rapid splitting of lipids to give various products, but these
enzymes tend to be de-activated by ammonia.
It has been questioned whether the above account of the hydrolysis of
lipids in dilute ammonia solutions adequately represents what really happens.
In addition to hydroxide ions, there are ammonium ions, free ammonia and
water molecules present in such solutions. The ammonium ions and water
molecules are very weak bases, and so do not participate significantly in the
alkaline hydrolysis of the lipids. Although the hydroxide ions are more
strongly basic than are the ammonia molecules, they are present in far lower
concentration. Thus a I-mol dm - 3 aqueous ammonia solution at 15°C has
a pH of ca. 12, and a hydroxide-ion concentration of ca. 4 x 10 - 3 mol dm - 3.
It is well-known that esters react with ammonia to form amides. It may be
that this is the initial reaction which occurs when the lipids in fresh natural
rubber latex undergo change in the presence of ammonia. Carboxylate anions
may then be formed by the subsequent alkali-induced hydrolysis of these ami des.
The lutoids, Frey-Wyssling particles, and also many of the other particulate
phases, dissolve in the aqueous phase of natural rubber latex when the latex
is ammoniated. Dissolution is assisted by slow hydrolysis of their proteins
and lipids.
Thus it is that ammonia-preserved natural rubber latex is predominantly
a two-phase colloid system, comprising rubber particles and aqueous phase
only. The particles themselves are colloidally stabilized by adsorbed fatty-acid
anions, and by other surface-bound anions derived from the proteins, instead
108 Natural latices
of by proteins only, as is the case with the fresh latex. This accounts for some
of the observed differences in colloidal and processing behaviour between
fresh and ammonia-preserved natural rubber latex. Thus, for example, the
addition of dilute mineral acids to fresh natural rubber latex does not cause
immediate evident colloidal destabilization, but slow gelation instead. Use
is made of this fact to separate dry natural rubber from the latex. By contrast,
the addition of dilute acid to ammonia-preserved natural rubber latex causes
the immediate formation of coagulum.
9.5.5.3 Formation of the anions of volatile fatty acids

The volatile fatty acids, the anions of which gradually form in natural rubber
latex, are so-called because they can be steam-distilled; indeed, this is the
basis for the method of estimating their concentration in natural rubber latex
(see section 8.9 of Chapter 8 (Volume 1». The acids of this type which are
of importance for natural rubber latex are formic, acetic and propionic. The
anions are formed by the action of microorganisms, especially bacteria, upon
certain of the carbohydrates which are present in the aqueous phase of the
latex. Ammonia and other bactericidal preservatives retard the formation of
these anions. For this reason, the volatile fatty acid content of an
ammonia-preserved natural rubber latex gives an indication of the extent of
microorganic activity which has occurred during the time since the latex
exuded from the tree. There is good evidence that the volatile fatty acid
content of the latex in the vessels of the tree is almost zero.
The principal facts relating to the formation of the anions of the volatile
fatty acids, and to the microorganisms responsible for their formation, appear
to be as follows:
1. The natural rubber latex in the latex vessels is believed to be substantially
free from bacteria.
2. Bacterial infection of fresh natural rubber latex can arise from many
sources, such as the atmosphere, the bark of the tree, and the utensils used
for the tapping process. An annulus of bacteria and coagulated rubber
particles may form around the top of the severed latex vessels, causing
infection of any latex which subsequently flows from the tree.
3. The manner in which the volatile fatty acid anions develop in fresh natural
rubber latex is illustrated by results reported by Lowe [110], shown here
as Fig. 9.27. After a short induction period, the concentration of these
anions increases rapidly over a period of a few days, until a stage is reached
after which very little further increase occurs. These characteristics suggest
an enzyme-catalysed process. It is presumed that the enzymes are of
bacterial origin. The eventual decline of the rate of formation of the volatile
fatty acid anions to effectively zero implies either that the substrate for
the process has become exhausted, or that the enzymes have become
de-activated.
0.8
0.6
Q;
.c
E
:::>
c:
-c
.~
0.4
~
]!
J1!
~
~
0.2
10 20 30 40
Time/day
Figure 9.27 Development of volatile fatty acid anions in fresh (points -) and ammonia-preserved
(points .) natural rubber latex (Lowe [110]).
4. The manner in which the volatile fatty acid anions form in ammonia-
preserved natural rubber latex is also shown in Fig. 9.27. It is qualitatively
similar to that in which they develop in the fresh latex. However, the
induction period is longer, and the increase in the concentration of these
anions during the period of rapid formation is greatly reduced.
5. Temperature has a considerable effect upon the extent of formation of the
volatile fatty acid anions, as is evident from results reported by Lowe
[110] and reproduced here as Table 9.12. The rate of formation initially
increases with increasing temperature, reaching a maximum at ca. 35°C.
As the temperature is further increased, the rate of formation decreases
sharply. This behaviour is typical of processes catalyzed by enzymes. The
rate at which volatile fatty acid anions form is presumably a reflection of
110 Natural latices
Table 9.12 Effect of temperature upon formation of anions of the volatile fatty acids
(Lowe [Ito])
V FA number after time indicated
Temperature
j"C 44 hours 52 hours 70 hours 94 hours 118 hours
16 0.042 0.042 0.045 0.045 0.087

25 0.073 0.097 0.275 0.420 0.560
30 0.356 0.445 0.550
36 0.350 0.435 0.505 0.560
46 0.042 0.045 0.045 0.050 0.056
55 0.042 0.047 0.049 0.048 0.048
the activity of the enzymes responsible for the formation process, the rapid
decline at temperatures above ca. 3SoC being a consequence of thermal
de-activation of the enzymes.
6. The substrate for the formation of the volatile fatty acid anions is believed
to be the various carbohydrates which are present in the aqueous phase
of the latex. Additional anions are formed if sugars, notably glucose, are
added to the latex. However, Lowe [Ill] has stated that quebrachitol is
not a substrate. Amino acids stimulate the formation of the volatile fatty
acid anions, probably by activating the sugars, rather than by acting as
substrates themselves. The initial stage of the process whereby volatile
fatty acid anions are formed may be the formation of a sugar-amino acid
complex. Citric 2-hydroxy-l,2,3-propanetricarboxylic and malic
(l-hydroxy-l,2-ethanedicarboxylic) acids also have a stimulating effect.
Proteins do not appear to be substrates in themselves, but, by partial
hydrolysis to amino acids, may bring about activation of other substrates.
7. The formation of the volatile fatty acid anions is progressively inhibited
by increasing concentrations of ammonia. This is illustrated by results
reported by Lowe [110,19], shown here as Fig. 9.28.
8. The extent of formation of the volatile fatty acid anions appears to be
dependent upon the redox potential of the latex. The redox potential of
fresh natural rubber latex is ca. 100m V, corresponding to an oxidative
condition. Formation of these anions under alkaline conditions (pH ca.
II) appears to commence when the redox potential has fallen to ca.
- 130 mV. The principal factor which influences the redox potential is the
extent of aeration of the latex. It is well-established that, in the presence
of excess oxygen, the formation of the volatile fatty acid anions occurs to
a much smaller extent than under essentially anaerobic conditions. Rather
paradoxically, reducing conditions can also be created as a consequence
of the limited aeration which occurs when natural rubber latex is pumped
from one tank to another; the metabolism of aerobic bacteria is thereby
stimulated, with consequent rapid consumption of the limited supply of
0.4
0.3
~
.c
E
:::I
c:
-c
·0
; 0.2
J!!
j1
~
(5
> 0.1
OL-------~------~------~------~----~
0.2 0.3 0.4 0.5 0.6 0.7
Ammonia concentration I % m/m on latex
Figure 9.28 Effect of increasing concentration of ammonia in inhibiting formation of volatile

fatty acid anions in natural rubber latex (Lowe [110, 19]). Significance of points: • volatile fatty
acid number after 24 hours; .t. volatile fatty acid number after 32 hours; • volatile fatty acid
number after 44 hours; ~ volatile fatty acid number after 60 hours.
oxygen. However, in sealed containers which are essentially air-free and

completely full, the stimulation of aerobic bacteria does not occur, and
the redox potential does not fall to the minimum required for formation
of these anions.
9. Boric acid appears to be particularly effective as an inhibitor of the
formation of the volatile fatty acid anions. The low concentration of these
anions in ammonia-preserved natural rubber latex which contains boric
acid as a secondary preservative has been attributed to a number of causes.
These include:
(a) inhibition of the enzymes which catalyze the formation of the volatile
fatty acid anions;
(b) de-activation ofthe microorganisms which produce the enzymes which
catalyse the formation of the volatile fatty acid anions;
(c) sequestration, by chelate complex formation, of the carbohydrates
which function as substrates for the formation of the volatile fatty acid
anions, thereby preventing the formation of the requisite sugar-amino
acid complexes;
(d) a tendency for boric acid to maintain the redox potential of the latex
in the oxidative condition;
(e) the possible ability of borates to inhibit a switch from the Krebs cycle
to a cycle which results in the formation of the volatile fatty acid anions.
112 Natural latices
As regards explanation (c), it is well-known that boric acid can interact

with the hydroxyl groups attached to adjacent carbon atoms in carbohydrate
molecules as follows:
'C-OH HO 'c-o
,/ I + )B-OH - - . . ::: I )B-OH + 2H 2 0
)C-OH HO ,/C-O
As regards explanation (e), this carries the implication that the anions of
non-stearn-volatile fatty acids of higher molecular mass than the volatile
fatty acid anions may be formed instead of the latter.
9.5.5.4 Other chemical changes during storage

Chin, Singh and Loke [114] have reported that the electrical conductivity
of ammonia-preserved natural rubber latex increases significantly during
storage. This indicates a progressive increase in the ionic strength of the
aqueous phase as additional anions are formed together with equivalent
numbers of ammonium ions. The enhancement of electrical conductivity
increases with increasing temperature, but is little affected by the degree of
aeration of the latex. These workers also report that the KOH number
increases during storage, again indicating enhancement of the concentration
of anions in the latex aqueous phase. The concentration of carbon dioxide
in the latex increases during storage. Storage temperature has little effect,
but aeration gives slightly higher concentrations that do anaerobic conditions.
The content of anions of carboxylic acids of higher molecular mass increases
initially, but then decreases with prolonged storage.
9.5.5.5 Effect of storage upon the colloid stability of ammonia-preserved

natural rubber latex
The volatile fatty acids which occur in ammonia-preserved natural rubber
latex are present as their ammonium salts. It is to be expected that the
principal effect of these salts will be to reduce the colloid stability of the
latex, and that this will be manifest as reduction of the mechanical and
chemical stability of the latex. However, Chin, Singh and Loke [114] have
questioned whether the mechanical stability of ammonia-preserved natural
rubber latex is significantly affected by the development of volatile fatty acid
anions, because they observed large changes in mechanical stability (decreases
as well as increases) during storage, notwithstanding that only small amounts
of volatile fatty acid anions were formed. Reduction of mechanical stability
is certainly observed when ammonium acetate is added to ammonia-preserved
3500
...
3000
--:§'"
Ql
2500
1800
~
:c<IS
Ui 2000
--
'" 1600
Ql
E
+= 1400
ro0
·c
<IS
r. 1500
~ 0
:c ~
Ql
$ 1200
ro'"
()
1000 •
.~ 1000
r.
()
Ql 500
~ 800
6000~--~0~.1~--~0.~2--~0~.3'-~0~.4 °0~--~----~2~--~3~--~4·
Level of added ammonium acetate expressed (April) (May) (June) (July) (August)
as equivalent increment of VFA number Storage time / month
(a) (b) (Month)
Figure 9.29 (a) Effect of addition of ammonium acetate upon mechanical stability of 62% mlm
anunonia-preserved natural rubber latex concentrate (Collier [109]). (b) Effect of storage
conditions upon mechanical stability of 58% mlm anunonia-preserved natural rubber latex
concentrate (Collier [109]). Significance of points: elatex stored in full container; • latex stored
in presence of excess air.
natural rubber latex. This is demonstrated by results reported by Collier

[109], shown here as Fig. 9.29(a). The reduction in colloid stability is
presumably mainly a consequence of increase of the ionic strength of the
aqueous phase of the latex, the principal mechanism being that discussed in
section 5.2.3.1 of Chapter 5 (Volume 1). However, in addition to electrostatic
effects arising from compression of the electrical double layer surrounding
the latex particles, and to possible effects upon counterion adsorption, steric
and hydration stabilization associated with the presence of surface-bound
proteinaceous substances may also be adversely affected. In any event, it
seems likely that the reduction in colloid stability is not primarily attributable
to the presence of the volatile fatty acid anions per se, but to the asscociated
counterions.
The mechanical and chemical stability of ammonia-preserved natural
rubber latex which has been stored for an appreciable length of time is thus
the result of several factors, apart from the initial stability of the latex. The
most important of these factors are believed to be:
114 Natural latices
1. the formation of volatile fatty acid anions, which tend to reduce colloid
stability because of the presence of the associated cations;
2. the formation of the anions of long-chain carboxylic acids by hydrolysis
of lipids, these anions tending to increase colloid stability.
The final colloid stability of the latex is mainly determined by the balance
between these two factors. However, as noted above, Chin, Singh and Loke
[114] have questioned whether the mechanical stability of ammonia-
preserved natural rubber latex is significantly affected by the development
of volatile fatty acid anions. Results reported by Collier [109], reproduced
here as Fig. 9.29(b), show that ammonia-preserved natural rubber latex which
has been stored under the nominally air-free conditions which prevail in an
almost-full sealed container can undergo progressive reduction in stability,
whereas the stability of ammonia-preserved natural rubber latex which has
been stored under aerobic conditions in half-full containers increases markedly
and passes through a maximum with increasing storage time. These observations
are explicable in a general way if it is accepted that the formation of volatile
fatty acid anions is accompanied by reduction in colloid stability, because,
as stated above, reducing conditions favour the formation of these anions.
Results reported by Chin, Singh and Loke indicate that, whereas prolonged
storage at lower temperatures has little adverse effect upon the mechanical
stability of ammonia-preserved natural rubber latex, storage at 40'C causes
rapid reduction in mechanical stability after an initial increase. On the basis
of these results, these workers recommend that temperatures above 30'C
during storage and transportation be avoided, or at least that exposure to
such temperatures be minimized.
9.5.5.6 Note on clonal characteristics

It has been noted that natural rubber latices from different clones have
different characteristics, and that, in particular, the latex from the clone
Glenshiel I has very low mechanical stability. The latex is therefore unsuitable
for concentration by centrifugation. Latices from the clones PB 86 and
RRIM 573 are white in colour and of high mechanical stability. They are
therefore suitable for concentration by centrifugation. The latex from the
clone Tj 1 is yellowish in colour, and the contained rubber is usually rather
hard. PR 107 is another clone which produces latex oflow mechanical stability
which is unsuitable for concentration by centrifugation. RRIM 501 latex is
of interest in that it has rather low mechanical stability when first concentrated,
but the stability subsequently rises to a high level during storage over a
period of ca. 3 months. For a discussion of the influence of clonal
characteristics upon the properties of foam rubber manufactured from
natural rubber latex, the reader is referred to a paper by Gyss, Woo and
Chan [115].
Other naturally-occurring latices 115
9.6 OTHER NATURALLY-OCCURRING LATICES
9.6.1 Summary of characteristics of other naturally-occurring latices

The principal characteristics of natural latices from various sources are
summarized in Table 9.13. This table is based upon information reported by
Stavely et al. [116]. The latex of Hevea brasiliensis is included for the purpose
of comparison. All the latices listed in Table 9.13 contain polyisoprene. The
microstructure found in the polyisoprene of Hevea brasiliensis (cis-l,4)
predominates. The frequency with which this polyisoprene microstructure
occurs in natural latices suggests that the polymers are formed by similar
biosynthetic mechanisms and from similar precursors (see further section 9.7
below). The marked differences in crystallinity between the various polyisoprenes
are, however, notable. Important exceptions to the Hevea brasiliensis
polyisoprene microstructure include the polyisoprenes contained in balata
and gutta percha latices, both of which are predominantly trans-l ,4, and that
contained in chicle latex, which is a mixture of cis- and trans-l,4 isomers.
9.6.2 Sources of other naturally-occurring latices

It has been estimated that the number of known polyisoprene-producing
plant species is ca. 2000, and that rubbers of sorts have been collected from
about 500 of these. The size of these species varies considerably, as does the
manner in which the latex is distributed within the plant. In consequence,
widely differing techniques have been used for the harvesting of the latex.
Gutta percha latex is obtained chiefly from trees of the SapotaclP family,
e.g., Palaquium gutta and Palaquium oblongifolia. These trees grow to a height
of up to 30m, and have trunk diameter 0.6-1 m when mature. They are
propagated either from seed or from cuttings, and are confined almost entirely
to the Malaysian peninsular. Ralata latex occurs in trees of the species
Mimusops balata. These trees are native to the West Indies and to South
America, occurring in great abundance in Guyana. The substance known as
chicle is obtained from the species Archras zapote, also known as Saprota
achros, or the saprodilla tree. These trees are found in tropical America,
notably in Mexico, Guatemala and British Honduras. Ficus elastica is a
species which is cultivated in European countries as an ornamental rubber
plant. The species Castilla elastica, which occurs in Central America and
Brazil, yields a product known as caucho rubber. Also found in Brazil is the
species Manihot glaziovii, from which is obtained ceara rubber. The substances
known as guayule rubber and jelutong are obtained respectively from the
species Parthenium argentatum and Dyera costulata. The former species is
said to produce up to 16% of its own mass of rubber. This is a greater
proportion than any known plant. However, this plant does not possess any
definite system of latex vessels. Much interest has been shown in the former
Table 9.13 Characteristics of naturally-occurring latices from various sources (Stavely et al. [116])
% microstructures Crystallinity on
(by infra-red analysis) stretching (X-rays) Gel
Intrinsic content
Particle Degree viscosity of
size Modal of Exten- of dispersed
range size cis- trans- crys- sion dispersed polymert
Source /nm /nm 1,4 1,4 3,4 tal/inity /% polymer* /% m/m
H evea brasiliensis 100-1500 120 97.8 0.0 2.2; Very strong 600 8 O§
Alstonia scholaris 130-530 330 97.8 0.0 2.2 Not extensible
Castilla elastica 54-1740 97.6 0.0 2.4 None 600 1.9 33
Castilla ulei 97.0 0.8 2.2 Very weak 770
Clitandra orientalis 97.7 0.0 2.3 None 675 3.0 31.5
Cryptostegia
grandiflora 80-960 200 97.7 0.0 2.3 Strong 370
Ficus elastica 43-2560 173 and 97.6 0.0 2.4 Trace 600 3.1 14
1040
Funtumia elastica 86-500 250 97.6 0.0 2.4 Very strong 900 1.7 24.3
Hancornia speciosa 97.6 0.0 2.4 Trace 775 0.3 11.5
Landolphia species 65-2000 162 97.6 0.0 2.4 Strong 900 3.2 12.5
M anihot glaziovii 60-180 120 97.6 0.0 2.3 Very strong 600
M ascarenhasia
elastica 20-1000 20 and 140 97.7 0.0 2.3 Very strong 600
Parthenium
argentatum 97.8 0.0 2.2 Strong 300 1.3 23.8
Achras zapote 22.2 76.5 1.3 Very strong 0 0.6 0
Balata and gutta
percha 0.0 98.7 1.3; Very strong 0
Palaquium
philippinensis 2560 0.0 100.0 0.0
* The units of intrinsic viscosity are not specified, but are assumed to be the usual units of l00cm J g- '. Neither is the solvent stated.
t Solvent not stated.
t Controversial: There is some evidence indicating that the Hevea and balata macromolecules contain 100"10 1,4 microstructures, and that the double bonds
are at least 99"10 cis and trans respectively.
§ A possible reason for the discrepancy between this figure and those recorded elsewhere is that this fi~e refers to the hydrocarbon in latex immediately
after tapping, whereas others refer to the hydrocarbon in latex which has been exposed to air. The so vent (unspecified) will also have affected the result.
Other naturally-occurring latices 117
Soviet Union in the past in a rubber-producing plant known as Taraxacum

kok-saghyz. This plant has been utilized by harvesting and processing the
roots, for it is in this part of the plant that the rubber occurs in rather variable
amount. The plant is harvested once a year. The process is said to be expensive,
and the yields low, being ca. 15-100 kg of rubber per acre per year.
9.6.3 Comments concerning some other naturally-occurring latices

9.6.3.1 Gutta percha latex
Gutta percha latex is a very viscous, greyish fluid which flows with difficulty
from cuts in the tree. It discolours readily when exposed to light, rapidly
becoming brown. The quality of the latex is deemed to be diminished by this
discoloration. For this reason, it has been the custom to obtain the latex by
felling the tree and crushing its stems, although some latex can be obtained
by merely crushing the leaves. Mature trees which have been felled and
crushed are said to yield ca. 1-1.5 kg of gutta percha. Neither gutta percha
latex, nor the polymer which it contains, is of any industrial importance today.
9.6.3.2 Halata latex

Some balata latices are reddish in colour, whilst others are white. The red
latex is more viscous than the white, although the tota! solids content of the
latter is higher than that of the former. The contained particles are spherical
in shape. The small particles exhibit a lively Brownian motion, whereas the
motion of the larger particles is very slight. Balata latices are notable for the
large amounts of resinous substances which they contain relative to the balata
hydrocarbon content. For various reasons (see Kemp [117]), it is thought
that the resinous substances are dissolved within the hydrocarbon particles,
and are not present as a separate phase. It appears that the resins function
as a plasticizer for the balata hydrocarbon. Typical chemical analyses for
the white and red latices, as given by Kemp, are reproduced here as Table
9.14. The principal difference between the two is the considerably higher ratio
of resin to hydrocarbon in the red type.
9.6.3.3 Chicle latex

Chicle latex is now little more than a curiosity, notable chiefly in that it
contains a mixture of cis-I,4 and trans-I,4-polyisoprenes in the ratio ca. 1/3.
According to Schlesinger and Leeper [118], it was the practice for itinerant
tappers to tap the trees during the rainy season (October to February) by
making zig-zag cuts in the bark ca. 45 cm apart. The exuded latex was boiled
in open vessels with constant stirring, in order to reduce the water content
to below 40% m/m. The concentrated latex was then poured into moulds
to form, on cooling, blocks weighing ca. 10 kg each. It is in this form that
the chicle was exported.
118 Natural latices
Table 9.14 Typical chemical analyses for white and red balata latices (Kemp [117])
White latex Red latex
based based based based

upon upon upon upon
whole solids whole solids
Property latex content latex content
Total solids content / % m/m 55.4 45.9

Volatiles at loo·C I % mlm 44.6 54.1
Resinous substances (by
petroleum ether extraction)
1% mlm 20.6 37.1 20.0 43.6
Hydrocarbon (by solvent
extraction / % m/m 29.5 53.3 19.6 42.8
Water-solubles and dirt
1% mlm 5.3 9.6 6.3 13.6
Ash / % m/m 0.43 0.77 0.49 1.08
Nitrogen content (by
Kjeldahl) / % m/m 0.27 0.49 0.39 0.85
Acidity I mg-equivs of alkali
per 100 cm 3 latex 9.0 10.0
Hydrocarbon/resin ratio 1.43 0.98
9.7 BIOGENESIS OF POLYISOPRENES
9.7.1 Biological function of the polyisoprenes

Almost nothing is known for certain concerning the part which the polyisoprenes
play in the metabolism and life cycle of the plants which produce them.
Various suggestions have been made. According to one view, natural rubber
latex in particular provides protection for the tree against infection when the
bark is damaged. When the tree is wounded, it is argued, the latex exudes
and dries to form a thin film which effectively seals the damaged area from
infection. One difficulty with this view is that the dried latex film itself is
able to support microorganisms of many kinds.
A second view regards the polyisoprenes which are contained within
naturally-occurring latices as reserve sources of nutrient. This is an attractive
hypothesis, because hydrocarbons are known to be very efficient storers of
energy. There is, however, no definite evidence which supports this view. A
further difficulty is that hydrocarbons are chemically rather inert, and
biochemical mechanisms by which their stored energy can be realized are rare.
A third view is that the polyisoprenes are the waste products of metabolic
Biogenesis of polyisoprenes 119
processes which are peculiar to the plant species which generate them. The
difficulty with this view is that the polyisoprenes accumulate in the plant
until the plant is damaged, and then any loss of polyisoprene is subsequently
made good by further biosynthesis. This observation does not accord well
with the polymer being an incidental by-product of metabolic processes which
have other primary purposes. Nevertheless, this seems to be the favoured
view. It has been known for some time that the respiration oflatex is essentially
anaerobic, and that plants which produce polyisoprenes often thrive under
arid conditions, where the capacity for gas and vapour interchange with the
atmosphere is limited. Furthermore, polyisoprenes are formed during periods
of intensive growth, when rapid anaerobic respiration is known to occur. It
has been suggested that, for some reason, the waste product of such anaerobic
respiration is polyisoprene, and not the ethyl alcohol which is the normal
waste product in these cases. A detailed mechanism has been proposed by
Bealing [119]. A defect of this mechanism is that it fails to reveal any purpose
for anaerobic respiration; for instance, no net amount of energy is made
available to the plant by its operation.
A fourth view is that the naturally-occurring latices are specialized
transport media or reserve water supplies. On this view, the system of latex
vessels and their contents serve to carry substances necessary to the life of
the plant from one place to another within the plant. It is, however, difficult
to see any function for the polyisoprenes which are present in these latices.
A fifth view is that the function of the polyisoprene is to provide a large
area for the adsorption of certain (unspecified) substances. There appears to
be no evidence to support this view.
9.7.2 Outline of polyisoprene biosynthesis
Although the biological function of the polyisoprenes remains obscure, much
is now known about the way in which these substances are synthesized by
plants. Indeed, the gradual unravelling of the biosynthetic pathways for these
substances, and increasing appreciation of the complexity of these pathways,
has emphasized the problem of why living organisms should produce these
substances at all. There is an extensive literature on polyisoprene biosynthesis
in plants. Archer, Audley and Bealing [120] have given a brief summary of
modern views on the biosynthetic mechanism. A comprehensive review of
work up until ca. 1960 has been provided by Fournier and Tuong-Chi-Cuong
[121]. The papers presented at a symposium on rubber biosynthesis [122],
held in 1964, provide a useful source of information. The chapter by Archer
et al. [1], and a paper by Bonner [123], provide authoritative summaries of
understanding of polyisoprene biosynthesis up to ca. 1960; a chapter by
Audley and Archer [124] provides a more recent summary. A useful summary
is also available in a paper by Whitby [125].
Isoprene itself is rejected as the precursor of the naturally-occurring
polyisoprenes. There are several reasons for this. The presence of isoprene
120 Natural latices
in polyisoprene-producing plants has never been demonstrated; nor have

any substances been found from which isoprene might reasonably be formed
in a simple manner. Again, whatever might be the precise mechanism by
which polyisoprene could be produced from isoprene, the overall reaction
would necessarily be an addition polymerization. Whereas biochemical
condensation polymerization reactions are common, addition polymerizations
are not. Furthermore, it has not been possible convincingly to demonstrate
the polymerization of isoprene in natural rubber latex without introducing
essentially 'non-biological' reaction conditions. And, until some 40 years ago,
none of the polyisoprenes which had been produced synthetically were linear
or stereoregular, in marked contrast to natural polyisoprenes.
Acetate anion is the simplest presursor of the natural polyisoprenes. This
is utilized in the form of an acetyl complex with a coenzyme. From there,
the main steps of the biosynthesis are believed to be as follows:
CH 3
+CH3. CO- +CH3.CO- I
CH3.CO- • CH 3.CO.CH 2 .CO- • H0 2 C.CH 2 ·C.CH 2 ·CO-
acetyl acetoacetyl I
OH
3-hydroxy-3-methyl-
glutaryl
CH 3
Itt
0 ° phosphorylation
CH 3
I
H0 2 C. CH 2 ·C.CH 2 CH 2 0.P.O.P.OH H0 2 C. CH 2 • C. CH 2 CH 2 0H
I I I I
OH OH OH OH
~yrophosphomevalonic acid mevalonic acid
o 0 o 0
CH 3 , t t CH 3 , t t
.,C.CH 2 ·CH 2 ·O.P.O.P.OH -,C=CH.CH 2 ·O.P.O.P.OH
CH ,. I I CH 3.... I I
2 OH OH OH OH
isopentenyl pyrophosphate dimethylallyl pyrophosphate
/'
Polyisoprene
Biogenesis of poiyisoprenes 121
Three acetyl moieties combine to form a 3-hydroxy-3-methylglutaryl moiety

via an acetoacetyl moiety as an intermediate. A reduction step then leads to
the formation of mevalonic acid, and this in turn is phosphorylated to
5-pyrophosphomevalonic acid. Dehydration and decarboxylation of this
molecule lead to the formation of isopentenyl pyrophosphate (2-methyl-
I-butenyl-4-pyrophosphate). This is tautomeric with dimethylallyl pyrophos-
phate (2-methyl-2-butenyl-4-pyrophosphate), being an example of the
three-carbon tautomeric system. The polyisoprene is then formed by
condensation polymerization of the isopentenyl pyrophosphate with elimination
of pyrophosphate groups, the dimethylallyl pyrophosphate probably functioning
in effect as a polymerization initiator.
9.7.3 More detailed consideration of mechanism of polyisoprene biosynthesis
9.7.3.1 Acetate as a precursor

In common with other terpenoid biosyntheses, acetate appears to be a
precursor for the biosynthesis of natural rubber. Natural rubber latex itself
appears to contain a combination of enzymes which are able to assimilate
acetate anions to give cis-l,4-polyisoprene. The primary evidence for this
belief is that 14C-labelled polyisoprene is obtained when 14C-labelled acetate
is added to fresh natural rubber latex or fed to rubber-producing plants. This
is also true of many of the other intermediates which are postulated as
comprising the biochemical pathway from acetate to polyisoprene.
It has also been shown using radio-tracer methods that the methyl and
carboxylic-acid acetate carbon atom are not incorporated in equal numbers.
For every two carbon atoms derived from carboxylic-acid groups, three are
derived from methyl groups. Furthermore, it has been demonstrated that the
various carbon atoms of each CsHs unit ofthe eventual polyisoprene sequence
have originated as shown in Structure XIV, in which C m and C c denote
XIV
carbon atoms of methylic and carboxylic origins respectively. Evidence for

this latter conclusion has been presented by Rabinowitz and Teas [126].
Rubber-producing plants, and also fresh natural rubber latex, were fed with
14CH3.12CO; in one instance, and with 12CH3.14CO; in another. The
polyisoprene obtained in each case was then subjected to an elegant
degradation, enabling the radioactive carbon atoms of the resultant polyisoprene
to be traced for each in turn of the radioactive precursors. Two schemes for
122 Natural latices
polyisoprene degradation are shown below. The simpler effects a resolution

of the original isoprene units into compounds or mixtures which contain
respectively its C 1 + C 4 , C 2 + C 3 , and C s carbon atoms; the more complex
of the two schemes resolves all five carbon atoms into separable com-
pounds.
CSH. CSH.
. I . I .
. . .1. C4H.-C'=C'H-C'H.+C4H.-C'=C'H-C' H. +...
i Isoprene Unit Ii Isoprene Unit I
ozonolysIs
CH 3 C 5H3
I I
••• -CH.-C C'H-C'H.-C'H.-C' CH-CH.- •••
AA ~ ~
levulinic aldehyde
lH.O.
Ho.C'.C'H•.C4H.. C'O.CsH3
levulinic acid
halolorm r~ ~H.NH.
~ ~
Ho.C'.C'H•. C4 H•. C30.H + ICsHIJ Ho.C'.C'H •. C'H··t C5H3
succinic acid
lNaN31H.so4 lAVHQ
N.NH .•
IC'o.l+IH.N.C'H•.C4H•. NH.I+lc3o.1 HO.C'.C'H2.C'H'·fH.C5H3

NH2
4-aminopentanoic acid
P
1
(i)A9
(ii) CH31
Ho.C'.C'H•.C4H•. C3H: CSH. ....._ _..:.K~O:;.H:..;a::.:t..:;35;:.;O:;...·_ __
l
Ho.C2.C'H2.C4H"r"H.CSH3
rearrangement +N(CH 3)3
betaine of 4-aminopentanoic acid
Ho.C'.C'H: C4H.c"H •. C5H3
1 spontaneous degradation
Ho.C'.C1H3 + HO.C'.C3H2.C sH3 NaN.iH 2SO, IC'o.l + H2N.C3H3 . C5H 3
acetic acid propionic acid ~KMn04
1
Ho.C 3 .C5H 3
NaN3/H.SO,
1 NaN 3 /H.SO,
IC2o.l+ IC'H3.NH.1
I C3o.HcsH3.NH.1
9.7.3.2 Biochemical formation of the acetyl moiety, and its immediate

utilization
Acetyl or acetate moieties are metabolized in the form of a thioester with a
complex substance known as coenzyme A (Structure XV). This substance
consists of adenine, ribose phosphate pyrophosphate, pantothenic acid and
thioethanolamine groups combined together in a linear sequence in that
order. For convenience, it is represented in chemical reactions as HS.CoA.
NH2
°°
-IH---+--+-'CH2-0-r-0·f·o.CH2C(CHa)2CHOH.CO.NH.CH2CH2.CO.NH.CH2CH 2·SH
OHOH
xv
In the biosynthesis of natural rubber, the acetate precursor is obtained
either by the glycolysis of glucides via pyruvic acid, or by the oxidation of
fatty acid. Restricting attention to the former mechanism, pyruvic acid
appears to be the immediate acetate precursor. The overall reaction for the
conversion is as follows:
CH 3.CO.C0 2H + HS.CoA --_a CH3.CO.S.CoA + CO 2 + 2[H]
pyruvic acid
However, many complex compounds are required for the expedition of the
overall reaction. The following scheme summarizes the likely sequence of
reactions [121]:
pyruvate
rl
key:
DPT: diphosphothiamine
FAD: flavine adenine dinucleotide, oxidized form
FADH 2 : flavine adenine dinucleotide, reduced form
DPN: diphosphopyridine nucleotide, oxidized form
DPNH,: diphosphopyridine nucleotide, reduced form
OPT OPT-pyruvate complex CoA: coenzyme A
oxidized lipoic a c i d X A D H X D P N
acetyl lipoic acid

CoA
reduced lipoic acid FAD DPNH 2
acetyl CoA
124 Natural latices
The acetylation of coenzyme A, which forms the last stage of these reactions,
may take place by means of a simple redox transfer reaction between acetyl
lipoic acid and HS.CoA as follows:
+ CH3CO.S.CoA
,
reduced lipoic acid
/
acetyl CoA
acetyl lipoic acid /

\ /'
, /
\\
\ /
,
/
\ /
\ I
\ I
\ I
\ I
""CH 2,
CH 2 CH(CH 2 )4 C0 2 H
I I
S S
oxidized lipoic acid
9.7.3.3 Formation of acetoacetyl coenzyme A

Acetoacetyl coenzyme A is believed to be formed by an enzymatic auto-
condensation of acetyl coenzyme A, the overall reaction being as follows:
~-oxothiolase
• CH 3CO.CH 2 .CO.S.CoA + HS.CoA
9.7.3.4 Formation of 3-hydroxy-3-methylglutaryl coenzyme A

3-hydroxy-3-methylglutaryl coenzyme A is believed to be produced by the
condensation of a further molecule of acetyl coenzyme A with acetoacetyl
coenzyme A. A possible mechanism, consistent with the observation that the
coenzyme A which is lost during the final hydrolysis step comes from the
acetyl coenzyme A and not from the acetoacetyl coenzyme A, is as follows:
o OH OH
II I I
CH -C CH 3 -C 0+ .. ---0-CH 2-CO.S.CoA CH 3 -C-CH 2·CO.S.CoA
3 I II I I
CH2 ~ CHo----ll+H _ CH 2
I I I
CO co co
I I I
S.CoA S.CoA S.CoA
acetoacetyl CoA
OH
I
CH a-C-CH 2C0 2H
I + HS.CoA
CH 2
I
CO.S.CoA
3-hydroxy-3-methylglutaryl
coenzyme A
9.7.3.5 Formation of mevalonic acid

There is much evidence to indicate that mevalonic acid is a direct precursor
of all terpenoid substances. In particular, this acid can be converted into
cis-! ,4-polyisoprene by means of the enzyme system which is present in fresh
natural rubber latex. The reactions by which mevalonic acid is formed from
3-hydroxy-3-methylglutaryl coenzyme A may involve a reduction via mevaldic
acid, during which coenzyme A is eliminated. If so, then in outline this
particular step of the biosynthetic pathway is as follows:
HO CHa
\ /
C
/ \ [H]
CH 2 CH2 -
I I
co C02H
I
H + HS.CoA
3-hydroxy- mevaldic mevalonic acid

3-methylglutaryl acid (showing numbering
coenzyme A of carbon atoms)
126 Natural latices
9.7.3.6 Formation of 5-pyrophosphomevalonic acid

The phosphorylation of mevalonic acid is achieved by means of the adenosine
triphosphate (ATP)/adenosine diphosphate (ADP) interconversion. 5-phospho-
mevalonic acid is formed as an intermediary. As well as appropriate enzymes,
the presence of divalent metal ions, such as magnesium, appears to be essential.
In outline, the sequence of reactions is as follows:
mevalonic 5-phosphomevalonic 5-pyrophosphomevalonic

acid acid acid
9.7.3.7 Formation of isopentenyl pyrophosphate

It is well-established that isopentenyl pyrophosphate is an important
intermediary in the biosynthesis of terpenoid substances generally. In
particular, it has been found that the enzyme system present in fresh natural
rubber latex is able to convert isopentenyl pyrophosphate into cis-
l,4-polyisoprene rapidly and quantitatively. Not only is this true of fresh
natural rubber latex itself; it is also true of latex aqueous phase which is
substantially free from particulate material. Indeed, isopentenyl pyrophosphate
has been described as 'active isoprene', i.e., the form in which the isoprene
moiety is presented for condensation to form polyisoprene. The mechanism
shown below has been suggested for the conversion of 5-pyrophosphomevalonic
acid into isopentenyl pyrophosphate. According to this mechanism, an
unstable intermediate triphosphate, 3-phospho-5-pyrophosphomevalonic acid,
is produced by further phosphorylation (at the 3 position) with adenosine
triphosphate. The triphosphate moiety may then disproportionate spontaneously
in the manner indicated.
OH
I ,.~
o-p-o,J CH
I ,,-X 3
ATP .. OH
/~
C I
°t 0
+
CH 2 \ CH 2
I >·1
HO-P-O-p-O-CH2 : c=o
I I ')j ,
OH OH :.OtH
'-'
5-pyrophosphomevalonic 3-phospho-5-pyrophospho-
/
acid mevalonic acid
CH 3
/
C
/~
o 0 CH 2 CH 2 + CO 2 + H+ + H2PO",
Ho-t-o-t-o-6H 2
I I
OH OH
isopentenyl pyrophosphate
9.7.3.8 Isomerization of isopentenyl pyrophosphate to dimethylallyl

pyrophosphate
It is believed that dimethylallyl pyrophosphate is an 'initiator' for the
subsequent condensation polymerization of isopentenyl pyrophosphate to
form polyisoprenes. This compound is probably formed by the following
three-carbon tautomeric isomerization reaction:
CH 3 0 0 CH 3 0 0
CH 2=6-CH 2CH 2.0JOJOH === CH3-6=CHCH2.0JO.~OH

I I I I
OHOH OHOH
isopentenyl pyrophosphate dimethylallyl pyrophosphate

128 Natural latices
This equilibrium is strongly (ca. 90%) in favour of the right-hand side. This
isomerization requires the presence of an enzyme called isopentenyl pyrophosphate
isomerase, together with traces of magnesium ions. The enzyme contains a
thiol group. The mechanism of the isomerization seems to involve an addition
of this thiol (denoted by En.SH) across the carbon-carbon double bond as
follows:
CH3
__ I
HS-En - CH 3 -C-CH 2 -CH2 - •••
thiol I
enzyme S-En
/
isopentenyl moiety
dimethylallyl moiety
9.7.3.9 Condensation polymerization of isopentenyl pyrophosphate to

give polyisoprenes
Two distinct types of polyisoprene occur naturally:
1. those which contain head-to-tail adducts, as in the geranyl and farnesyl

compounds, and in the polyisoprenes of high molecular mass; and
2. those which contain head-to-tail adducts, as in squalene.
The concern here is exclusively with the addition reaction which leads to the
first of these types. The initial stage in the formation of the polyisoprenes is
believed to be the condensation of one molecule of isopentenyl pyrophosphate
with one molecule of dimethylallyl pyrophosphate to form either neryl or
geranyl pyrophosphate. These latter are referred to collectively as C1o-prenyl
pyrophosphates. Neryl pyrophosphate is the cis isomer, and thus might be
supposed to be the precursor of the natural rubber hydrocarbon; geranyl
pyrophosphate is the trans isomer. Suitable enzymes, and also the
presence of magnesium ions, are required. A possible mechanism is as
follows:
°t °t
, 0. p.o. P.OH
(-" °t °•
y H3 .\_-IIOHOH
I + C~=y-CH2- CH2-O' y O' OH y
CH 3-C=CH-CH 2 CH 3 OHOH
dimethylallyl pyrophosphate isopentenyl pyrophosphate
J
3
I
CH3
+
CH -C=CH-CH 2 +
&- &+
CH 2=C-CH2-CH2-O.P.O.P.OH
°t °t
I I I
+ H3 P207 CH3 OH OH
carbonium ion
C~
I t t
I °°
CH3-C=CH-CH2-CH2-r=CH-CH2-o,yo'1.oH + H+
CH3 OH OH
C1o-prenyl pyrophosphate
The C1o-prenyl pyrophosphate may then condense with a further molecule

of isopentenyl pyrophosphate to form a C1s-prenyl pyrophosphate. The
following may be the mechanism, analogous to that given above for the
formation of a ClO-prenyl pyrophosphate:
130 Natural latices
C,o-prenyl
pyrophosphate
,s
C -prenyl
pyrophosphate
Further condensation with successive isopentenyl pyrophosphate molecules

by similar mechanism leads to the formation of C 20 -, C 25 -, ••• prenyl
pyrophosphate molecules. Ultimately a polyisoprenyl pyrophosphate of
indefinite chain length is formed. According to the above reaction scheme,
the dimethylallyl pyrophosphate does indeed function as the initiator of the
isopentenyl pyrophosphate condensation polymerization, even though it is
present in large excess of the effective monomer.
An alternative mechanism for the condensation polymerization of isopentenyl
pyrophosphate has been proposed by Archer [127]. This mechanism does
not require the prior isomerization of isopentenyl pyrophosphate to dimethylallyl
pyrophosphate. It envisages the agency of a thiol enzyme, and seeks to explain
the predominance of the cis configuration about the eventual carbon-carbon
double bonds in the polyisoprene chain as resulting from the formation of
a cyclic transition-state complex as shown below. Clearly, a mechanism such
as this would, if repeated indefinitely, lead to the formation of a 1,4-polyisoprene
having the cis configuration around each carbon-carbon double bond.
H H
\/
c~/\
C CH20P206H3
+ H5-En
II
CH 2
thiol
enzyme
isopentenyl
pyrophosphate
+ isopentenyl
pyrophosphate
If the definition of an emulsion polymerization reaction given in section

10.1.1 of Chapter lOis accepted, then the biosynthesis of polyisoprenes, such
as the natural rubber hydrocarbon, must be regarded as an example of a
biochemical emulsion polymerization. The question then arises as to the
mechanism of particle nucleation. Considerable interest has been shown in
this matter. According to a view put forward by Hager et al. [128], particle
nucleation occurs when isoprenoid oligomers of low molecular mass separate
from the aqueous phase as droplets, possibly with detachment of the enzyme
responsible for polymerization. These nuclei then grow in size, possibly by
adsorbing other propagating polyisoprenoid species. This mechanism appears
to be roughly analogous to the homogeneous nucleation which is believed
by many to be the origin of the particles in at least some emulsion
polymerization reaction systems (see section 10.2.2.2.2 of Chapter 10). Archer
and Audley [129], on the other hand, have proposed a mechanism which
appears to be broadly analogous to the micellar mechanism for nucleation
in emulsion polymerizations (see section 10.2.2.2.1 of Chapter 10). According
to their suggestion, the polyisoprene macromolecule begins to develop within
a micelle which forms from molecules of a C 2o -prenyl pyrophosphate
combined with either an enzyme molecule or perhaps a molecule of some
other non-rubber substance. The micelle may be either free in the cell aqueous
phase or attached to some surface.
132 Natural latices
REFERENCES
l. Archer, B. L., Barnard, D., Cockbain, E. G., Dickenson, P. B. and McMullen,
A. I. (1963) Structure, composition and biochemistry of Hevea latex, in The
Chemistry and Physics of Rubber-like Substances (ed. L. Bateman), Maclaren,
London, Chapter 3.
2. Cockbain, E. G. and Philpott, M. W. (1963) Colloidal properties oflatex, in The
Chemistry and Physics of Rubber-like Substances (ed. L. Bateman), Maclaren,
London, Chapter 4.
3. Gazeley, K. F., Gorton, A. D. T. and Pendle, T. D. (1988) Latex concentrates:
properties and composition, in Natural Rubber Science and Technology (ed.
A. D. Roberts), Oxford University Press, Oxford, Chapter 3.
4. Calvert, K. O. (1982) Natural rubber latices, in Polymer Latices and their
Applications (ed. K. O. Calvert), Applied Science, London, Chapter 2.
5. Polhamus, L. G. (1962) Rubber: Botany, Production and Utilization, Interscience,
New York.
6. Bobiliolf, W. (1923) Anatomy of Physiology of Hevea Brasiliensis, Zurich.
7. Dijkman, M. J. (1951) Hevea, University of Miami Press, Florida.
8. D'Auzac, J., Jacob, J.-L. and Chrestin, H. (eds) (1989) Physiology of Rubber Tree
Latex, CRC Press, Florida.
9. Edgar, A. T. (ed.) (1958) Manual of Rubber Planting (Malaya), Incorporated
Society of Planters, Kuala Lumpur.
10. Bud-grafting Techniques (1961) Rubber Research Institute of Malaya, Kuala
Lumpur. (This booklet is in concurrent Chinese and Malay. Note that both text
and diagrams proceed from the back to the front of the book.)
1l. McIndoe, K. G. (1961) Rubber Chemistry and Technology, 34, 413.
12. Tinley, G. H. (1960) Proceedings of the Natural Rubber Research Conference,
Kuala Lumpur, 1960, p. 409.
13. Blackman, G. E. (1960) Proceedings of the Natural Rubber Research Conference,
14. Abraham, P. D., Blencowe, J. W., Chua, S. E., Gomez, J. B., Moir, G. F. J.,
Pakianathan, S. W., Sekar, B: C., Sou thorn, W. A. and Wycherley, P. R. (1971)
Journal of the Rubber Research Institute of Malaya, 23, 85.
15. Compagnon, P. and Tixier, P. (1950) Revue Generale de Caoutchouc, 27, 525.
16. de Haan-Homans, L. N. S. (1951) Industrial and Engineering Chemistry, 43, 403.
17. Johnson, W., British Patent No. 467, 18 August, 1953.
18. Norris, H. L. and Armstrong, S. T./ United States Patent No. 9,891, 26 July, 1853.
19. Lowe, J.S. (1960) Transactions of the Institution of the Rubber Industry, 36,225.
20. Madge, E. W., Collier, H. M. and Peel, J. D. (1950) Transactions of the Institution
of the Rubber Industry, 26, 305.
2l. Resing, W. L. (1954) Proceedings of the Third Rubber Technology Conference,
London, 1954, p. 50.
22. McGavack, J. (1959) Rubber Chemistry and Technology, 32, 1660.
23. Bowler, W. W. (1953) Industrial and Engineering Chemistry, 45, 1790.
24. Townsend, H. B./General Latex and Chemical Corporation, British Patent No.
788,580, 2 January 1958.
25. Sekaran, K. C. and Morris, J. E./The British Rubber Producers' Research
Association, British Patent No. 852,280, 26 October 1960.
26. Ten Broeck, W. T. L./The Goodyear Tire & Rubber Company, United States
Patent No.2, 888, 504, 26 May 1959.
27. Calvert, K. 0., Sundaram, P. and Tan, T. Y. (1978) Preprints of Plastics and
Rubber Institute International Polymer Latex Conference, London, 1978, Paper No.8.
References 133
28. Calvert, K. 0., Tan, T. Y., Prichards, R. J. and Sundaram, P. (1983) NR

Technology, 14, 45.
29. Angove, S. N. (1964) Transactions of the Institution of the Rubber Industry,
40, 71.
30. Angove, S. N. and Pillai, N. M. (1964) Transactions of the Institution of the
Rubber Industry, 40, 257.
33. Angove, S. N. and Pill ai, N. M. (1965) Transactions of the Institution of the
34. Angove, S. N. (1965) Transactions ofthe Institution ofthe Rubber Industry, 41,143.
35. Clough, J. T. M. and Glymph, E. M. (1975) Proceedings of International Rubber
Conference, Kuala Lumpur, 1975, Volume V, p. 103.
36. John, C. K., Nadarajah, M., Rama Rao, P. S., Lau, C. M. and Ng, C. S. (1975)
Proceedings of International Rubber Conference, Kuala Lumpur, 1975, Volume
IV, p. 339.
37. Smith, M. G., Robert, C. F., John, C. K., Rama Rao and Ng, C. S. (1977)
Elastomerics, November, p. 35.
38. John, C. K., Wong, N. P., Chin, H. c., Latiff, A. and Lim, H. S. (1986) Proceedings
of Rubber Research Institute of Malaysia Rubber Growers' Conference, Kuala
Lumpur, 1986, p. 320.
39. Nadarajah, M., Coomarasamy, A., de Silva, K. N. P., Weeraratne, W. G. and
Munasinghe, B. K. S. (1975) Proceedings of International Rubber Conference,
Kuala Lumpur, 1975, Volume V, p. 183.
40. Graham, D. J. and Taysum, D. H./The Natural Rubber Producers' Association
(sic), British Patent No. 908,283, 17 October 1962.
41. John, C.K., Wong, N. P., Chin, H. c., Rama Rao, P. S. and Latiff, A. (1985)
Proceedings of International Rubber Conference, Kuala Lumpur, 1985, Volume
II, p. 451.
42. Stutchbury, M. S./KDP Ltd., British Patent No. 214,583, 11 September 1924.
43. KDP Ltd., British Patent No. 243,016, 2 September 1926.
44. KDP Ltd., British Patent No. 244,727, 11 November 1926.
45. Metallgesellschaft Aktiengesellschaft, British Patent No. 382,235,20 October 1932.
46. Quasebarth, K. H., United States Patent No.2, 360, 445, 17 October 1944.
47. Twiss, D. F. and Carpenter, A. S. (1938) Proceedings of the Rubber Technology
Conference, London, 1938, p. 81.
48. Hauser, E. A. and Dewey, B. (1941) Industrial and Engineering Chemistry, 33, 127.
49. Vester, C. F. (1938) Proceedings of the Rubber Technology Conference, London,
1938, p. 126.
50. Baker, H. C. (1937) Transactions of the Institution of the Rubber Industry, 13,70.
5!. Morris, J. E. (1954) Proceedings of the Third Rubber Technology Conference,
London, June 22-25, 1954, p. 13.
52. Verhaar, G. (1959) Proceedings of International Rubber Conference, Washington,
1959, p. 272.
53. Murphy, E. A. (1942) Transactions ofthe Institution ofthe Rubber Industry, 18, 173.
54. Stevens, H. P. (1948) Rubber Developments, I, 1.
55. van den Tempel, M. (1942) Transactions of the Institution of the Rubber Industry,
18, 173.
56. ISO Standard 2004- 1988: Natural rubber latex concentrate - Centrifuged or
creamed, ammonia-preserved types - Specification.
134 Natural latices
57. ASTM Standard D 1076-88: Standard Specification for Rubber - Concentrated,

Ammonia Preserved, Creamed, and Centrifuged Natural Latex.
58. Jurado, C. W. and Mayhan, K. G. (1986) Rubber Chemistry and Technology, 59, 84.
59. Gorton, A. D. T. and Pendle, T. D. (1985) Proceedings of International Rubber
Conference, Kuala Lumpur, 1985, Volume II, p. 468.
60. Gorton, A. D. T. and Pendle, T. D. (1986) Journal ofNatural Rubber Research, 1, 122.
61. Gorton, A. D. T. (1972) Rubber Chemistry and Technology, 45, 1202.
62. ISO Standard 2027-1990: Natural rubber latex concentrate, evaporated, preserved
- Specification.
63. Tan, K. Y. (1982) Mechanical Destabilisation of Natural Rubber Latices, Ph.D.
thesis, United Kingdom Council for National Academic Awards.
64. Kemp, A. R. (1937) Rubber Age, New York, 40, 366.
65. Chin, P. S. and Lau, C. M. (1975) Proceedings ofInternational Rubber Conference,
Kuala Lumpur, 1975, Volume V, p. 79.
66. Hauser, E. A. (1926) Industrial and Engineering Chemistry, 18, 1146.
67. Schoon, Th. G. F. and van der Bie, G. J. (1955) Journal of Polymer Science, 16,63.
68. Schoon, Th. G. F. and Phoa, K. L. (1955) Journal of Polymer Science, 10, 226.
69. Pendle, T. D. and Swinyard, P. E. (1991) Journal of Natural Rubber Research, 6, 1.
70. Allen, P. W. and Bristow, G. M. (1963) Journal of Applied Polymer Science, 7, 603.
71. Shiibashi, T. (1987) International Polymer Science and Technology, 14(12), T33.
72. Freeman, R. (1954) Proceedings of the Third Rubber Technology Conference,
London, 1954, p. 3.
73. Gazeley, K. F. and Mente, P. G. (1985) Pressure-sensitive adhesives from modified
natural rubber latex. Paper presented at Adhesives, Sealants and Encapsulants
Conference, London, 1985.
74. Sekhar, B. C. (1960) Proceedings of the Natural Rubber Research Conference,
75. Wren, W. G. (1961) Rubber Chemistry and Technology, 34,378.
76. Kemp, I. (1935) Transactions of the Faraday Society, 31, 1347.
77. Kemp, I. and Twiss, D. F. (1936) Transactions of the Faraday Society, 32, 890.
78. Hauser, E. A. and Bender, M. (1938) Proceedings of the Rubber Technology
Conference, London, 1938, p. 101.
79. Blow, C. M. (1938) Proceedings of the Rubber Technology Conference, London,
1938, p. 101.
80. Bowler, W. W. (1953) Industrial and Engineering Chemistry, 45, 1790.
81. Ho, C. C. and Ng, W. L. (1979) Colloid and Polymer Science, 257, 406.
82. Ho, C. C. (1988) Colloid and Polymer Science, 266, 70.
83. Ho, C. C. (1989) Colloid and Polymer Science, 267, 643.
84. Bishop, R. O. (1927) Malayan Agricultural Journal, 15,27.
85. Kemp, A. R. and Straitiff, W. G. (1940) Journal of Physical Chemistry, 44, 788.
86. Roe, C. P. and Ewart, R. H. (1942) Journal ofthe American Chemical Society, 64, 262.
87. Moir, G. F. J. and Tata, S. J. (1960) Journal of the Rubber Research Institute of
Malaya, 16, 155.
88. Archer, B. L. and McMullen, A. I. (1960) Proceedings of the Natural Rubber
Research Conference, Kuala Lumpur, 1960, p. 787.
89. Ng, T. S. (1960) Proceedings of the Natural Rubber Research Conference, Kuala
Lumpur, 1960, p. 809.
90. McGavack, J. and Rumbold, J. S. (1934) Industrial and Engineering Chemistry,
26,789.
91. Altman, R. F. (1940) Archiefvoor die Rubbercultur in Nederlandsch-Indie, 24, 647.
92. Whitby, G. S. and Greenberg, H. (1941) Biochemical Journal, 35, 640.
93. Cited by Ng [89] as 'Drake (1951), without reference.
References 135
94. Belgrave, W. N. C. (1925) Malayan Agricultural Journal, 13, 154, 367.

95. Midgley, T., Henne, A. L. and Renoll, M. W. (1937) Journal of the American
Chemical Society, 59, 2501.
96. Altman, R. F. (1939) Archiefvoor die Rubbercultuur in N ederlandsch-Indie, 23, 239.
97. Tristram, G. R. (1940) Biochemical Journal, 34, 301.
98. Levy, D. A., Charpin, D., Pecquet, c., Leynadier, F. and Vervloet, D. (1992)
Allergy, 47, 579.
99. Dalrymple, S. J. and Audley, B. G. (1992) Rubber Developments, 45, 51.
100. Hsia, R. C. H. (1958) Transactions ofthe Institution ofthe Rubber Industry, 34,267.
101. Homans, L. N. S. and van Gils, G. E. (1948) Proceedings of the Second Rubber
Technology Conference, London, 1948, p. 292.
102. Ruinen, J. (1950) Archief voor die Rubbercultuur, 27, 255.
103. Dickenson, P. B. (1965) Symposium on Rubber Biosynthesis, in Proceedings of
the Natural Rubber Producers' Research Association Jubilee Conference, Cambridge
1964 (ed. L. Mullins), Maclaren, London, for Natural Rubber Producers'
Research Association, p. 52.
104. Audley, B. G. (1965) Symposium on Rubber Biosynthesis, in Proceedings of the
Natural Rubber Producers' Research Association Jubilee Conference Cambridge
105. Moir, G. F. J. (1959) Nature, 184, 1626.
106. Southorn, W. A. (1960) Proceedings of the Natural Rubber Research Conference,
107. Southorn, W. A. (1961) Rubber Developments, 14,2.
108. McGavack,J. and Bevilacqua, E. M. (1951)Jndustrial and Engineering Chemistry,
43,475.
109. Collier. H. M. (1955) Transactions ofthe Institution ofthe Rubber Industry, 31,166.
110. Lowe, J. S. (1959) Transactions o{the Institution of the Rubber Industry, 35, 10.
111. Lowe, J. S. (1960) Transactions of the Institution of the Rubber Industry, 36, 202.
112. Pillai, N. M. (1968) Journal of the Rubber Research Institute of Malaysia, 20,152.
113. Sinclair, E. A. and Janssen, F. W. (1972) Polymer Journal, 4, 93.
114. Chin, H. c., Singh, M. M. and Loke, S. E. (1979) Plastics and Rubber: Materials
and Applications, 4, 164.
115. Gyss, P. R., Woo, C. E. C. and Chan, S. P. (1960) Proceedings of the Natural
Rubber Research Conference, Kuala Lumpur, 1960, p. 710.
116. Stavely, F. W., Biddison, P. H., Forster, M. J., Dawson, H. G. and Binder, J. L.
(1961) Rubber Chemistry and Technology, 34, 423.
117. Kemp, A. R. (1938) Proceedings of the Rubber Technology Conference, London,
1938, p. 68.
118. Schlesinger, W. and Leeper, H. M. (1951) Industrial and Engineering Chemistry,
43,398.
119. Bealing, F. J. (1965) Symposium on Rubber Biosynthesis, in Proceedings of the
Natural Rubber Producers' Research Association Jubilee Conference, Cambridge
1964 (ed. L. Mullins). Maclaren, London, for Natural Rubber Producers'
120. Archer, B. L., Audley, B. G. and Bealing, F. J. (1982) Plastics and Rubber
International, 7, 109.
121. Fournier, P. and Tuong-Chi-Cuong (1961) Rubber Chemistry and Technology,
34, 1229.
122. (1965) Symposium on Rubber Biosynthesis, in Proceedings of the Natural Rubber
Producers' Research Association Jubilee Conference, Cambridge 1964, (ed. L.
Mullins), Maclaren, London, for Natural Rubber Producers' Research Association.
136 Natural latices
123. Bonner, J. (1960) Proceedings of the Natural Rubber Research Conference, Kuala
Lumpur, 1960, p. 11.
124. Audley, B. G. and Archer, B. L. (1988) Biosynthesis of rubber, in Natural Rubber
Science and Technology (ed. A. D. Roberts), Oxford University Press, Oxford,
Chapter 2.
125. Whitby, G. S. (1962) Transactions ofthe Institution ofthe Rubber Industry, 38, 116.
126. Rabinowitz, J. L. and Teas, H. 1. (1960) Proceedings of the Natural Rubber
Research Conference, Kuala Lumpur, 1960, p. 777.
127. Archer, B. L. (1965) Symposium on Rubber Biosynthesis, in Proceedings of the
Natural Rubber Producers' Research Association Jubilee Conference, Cambridge
128. Hager, T., MacArthur, A., McIntyre, D. and Seeger, R. (1979) Rubber Chemistry
and Technology, 52, 693.
129. Archer, B. L. and Audley, B. G. (1987) Botanical Journal of the Linnean Society,
94, 181.
10
Synthetic latices: general

principles of production
10.1 INTRODUCTION
10.1.1 Definition of a synthetic latex

This chapter is concerned with the general principles relating to the
production of those colloidal systems which have become known as synthetic
latices. The subsequent chapter will discuss individual types of synthetic latex.
The main concern in the subsequent chapter will, of course, be those synthetic
latices which have attained industrial importance to a greater or lesser degree.
However, note will also be taken of certain other interesting types which so
far have either been of very limited specialist application interest, or have
remained of essentially academic interest only.
For many years, the term synthetic latex has been understood to mean an
aqueous dispersion of a polymer which is produced by the chemical reaction
known as emulsion polymerization. The term has been applied both to the
immediate product of this type of reaction and to the latices which are
obtained by subjecting the immediate product of emulsion polymerization
to various after-treatments such as partial particle agglomeration, concentration,
and chemical modification.
However, in recent years, matters have become rather more complicated
for at least two reasons. Firstly, polymerization reactions analogous to
emulsion polymerization, but carried out in non-aqueous media, are now
well known. The products of such reactions are effectively non-aqueous
synthetic latices, and are widely recognized as such. Secondly, and more
importantly, it is now far from easy to give an entirely satisfactory definition

138 Synthetic latices: general principles of production
of an emulsion polymerization which, on the one hand, embraces all the

varied types of polymerization reaction which workers in the field generally
recognize as being emulsion polymerizations, and, on the other, clearly
distinguishes emulsion polymerizations from other types of polymerization
reaction which are superficially similar.
In the light of developments which have taken place in the last two decades,
a reasonable definition of an emulsion polymerization reaction is a polymerization
reaction which produces polymer in the form ofa stable lyophobic sol. According
to this definition, an emulsion polymerization reaction is recognized by the
nature of the end-product, rather than by the reaction mechanism which has
led to the formation of that end-product. This definition has the advantage
over other possible alternatives that it embraces not only 'conventional'
aqueous emulsion polymerization reactions, but also such reactions as
non-aqueous emulsion polymerizations, reaction systems which are initially
homogeneous, and micro-suspension polymerization reactions. However,
with regard to the science and technology of latices, it has the disadvantage
of being circular when taken in conjunction with the usual definition of a
synthetic latex: the latter is defined as the product of an emulsion polymerization
reaction, and an emulsion polymerization is defined as a reaction which leads
to the formation of a stable sol of a polymeric substance, i.e., of a latex! But
in practice this apparent circularity has caused little confusion; the term
synthetic latex is now used loosely to denote any lyophobic colloidal
dispersion ofa polymer in a liquid phase which is produced by a polymerization
reaction ex vivo; and the term emulsion polymerization is understood to mean
any ex vivo polymerization reaction which produces such a polymer
colloid.
Most emulsion polymerization reactions of industrial importance are free-
radical addition polymerizations of monomers which are olefinically-
unsaturated carbon compounds containing one or more carbon-carbon
double bonds. Most, if not all, industrial synthetic latices are produced by
this type of reaction. When the term emulsion polymerization is used without
qualification in what follows, it is this type of reaction which is meant.
However, in principle, the terms synthetic latex and emulsion polymerization
are of wider meaning, embracing respectively all types of polymer lyophobic
sol produced by polymerization reactions carried out ex vivo, and the
reactions by which they are produced. The concern in this and the subsequent
chapter will be almost exclusively with the free-radical emulsion polymerizaton
of olefinic monomers, and the latices produced by this type of reaction. From
what follows, it will become evident that the feature which is believed to
distinguish this type of free-radical addition polymerization from other types
is that most of the monomer-polymer conversion occurs by way of
propagation reactions which are compartmentalized into a large number of
isolated reaction loci, the average number of propagating radicals per reaction
locus being small, usually of the order of 1.
Introduction 139
10.1.2 Distinguishing features of synthetic latices relative to natural

latices
10.1.2.1 Particle shape, size and size distribution
Provided that attention is confined to the immediate products of conventional

aqueous emulsion polymerization reactions, it is possible to enumerate several
characteristics which generally distinguish synthetic latices as a class from
naturally-occurring latices, in particular, from natural rubber latex. The more
important of these characteristics are noted in what follows. In considering
these characteristics, it must be borne in mind that they do not necessarily
pertain, or pertain to the same extent, to latices which are obtained by
subjecting the immediate product of a conventional emulsion polymerization
reaction to various after-treatments. Indeed, in the case of some synthetic
rubber latices, the principal purpose of the after-treatment is to obtain a
product which resembles ammonia-preserved natural rubber latex more
closely than did the immediate product of the emulsion polymerization
reaction.
The most obvious of the differences relate to particle shape, size and size
distribution. The particles dispersed in the immediate product of a conventional
emulsion polymerization reaction are usually almost exactly spherical in
shape. The more complex geometrical forms which are sometimes observed
in natural rubber latex are absent from synthetic latices. Furthermore, as far
as is known, the particles of synthetic latices lack the complex structure which
the particles of fresh natural rubber latex possess, except in those instances
where deliberate attempts have been made to produce latex particles of
complex structure by emulsion polymerization.
The particles in synthetic latices are considerably smaller than are many
in ammonia-preserved natural rubber latex, and the spread of particle sizes
is much less than in that latex. In particular, the long 'tail' which is
characteristic of the particle-size distribution in natural rubber latex is absent
from the immediate product of the typical conventional emulsion polymerization
reaction. Typically, the number-average particle diameter for this product is
the range 50-100 nm; and because the particle-size distribution is usually
narrow, the other average particle diameters are similar. By contrast, particles
as large as 5J.1m occur in ammonia-preserved natural rubber latex (see section
9.5.2.1 of Chapter 9), and the volume-average particle diameter is
accordingly large, even though the number-average diameter is considerably
smaller. However, these comparisons refer specifically to the immediate
products of typical conventional emulsion polymerization reactions. It is
possible to increase the average particle size considerably in such latices, and
also to broaden the particle-size distribution, by, for instance, using the
particles in the immediate product as 'seeds' from which much larger
synthetic latex particles can be grown, or by partially agglomerating the particles.
10.1.2.2 Specific surface area

One consequence of the small size of the particles in synthetic latices is that
the specific surface area of the polymer-aqueous phase interface is high (see
section 2.2.2.3 of Chapter 2 (Volume 1)). For example, a ten-fold reduction
in particle diameter is accompanied by a ten-fold increase in specific surface
area, if the geometrical shape of the particles remains unaltered. The increase
in specific surface area is especially sharp for particles with diameters below
ca. tOO nm. Many of the distinctive characteristics of the immediate product
of emulsion polymerization are consequences of this.
10.1.2.3 Ratio of colloid stabilizers to polymer

One consequence of the greater specific surface area of synthetic latices as
compared with ammonia-preserved natural rubber latex is that the ratio of
hydrophilic colloid stabilizers to polymer in the former tends to be significantly
higher than in the latter. Based on the whole latex, the content of such
substances in ammonia-preserved natural rubber latex is typically ca.
1% m/m. For the immediate product of emulsion polymerization, the content
may be as high as 5% m/m. Even with such high levels of adsorbable colloid
stabilizers, the polymer particles in the immediate product of emulsion
polymerization are seldom saturated with adsorbate. For such latices
stabilized by surface-active substances, the coverage of the particles is often
equivalent to only ca. 30-50% of the total surface area of the particles. Only
very small amounts of surface-active substances are usually present in the
aqueous phase of such latices, the concentrations being usually well below
the critical micelle concentration. The surface free energy of such latices is
therefore often much higher than would be expected on the basis of the
overall concentration of surface-active substances in the latex.
A further consequence of incomplete coverage of the particles in these
latices is the great affinity which the particles have for added surface-active
substances. Thus care must be exercised when blending synthetic latices with
other aqueous lyophobic sols, including, inter alia, ammonia-preserved
natural rubber latex. There is often a marked tendency for the synthetic latex
to 'rob' an otherwise stable colloidal dispersion of its stabilizers, and thereby
to cause micro-flocculation. This phenomenon accounts, in part at least, for
the large viscosity changes which are sometimes observed when ammonia-
preserved natural rubber latex is blended with synthetic rubber latices (see
further section 15.1 of Chapter 15 (Volume 3)).
10.1.2.4 Wet-gel strength of latex deposits

The term wet-gel strength, as applied to latex films, has been defined in section
1.8.6 of Chapter 1 (Volume 1). It refers to the mechanical strength of gelled
latex films before they have completely dried out. It is an important property
Introduction 141
for processes in which the product has to be manipulated before it is properly

dry, notably in processes in which moulded products have to be stripped
from their moulds whilst still in a wet condition. An important example of
such processes is the manufacture of moulded latex foam rubber (see Chapter
18 (Volume 3». Deposits from synthetic rubber latices usually have low
wet-gel strength relative to those from natural rubber latex, especially if the
deposits are hot. This is inevitably disadvantageous to synthetic rubber latices
in processes for which wet-gel strength is an important property. The factors
which determine the strength of wet latex deposits are not well understood.
The low wet-gel strength of deposits from synthetic rubber latices is probably
a consequence of several factors, the principal of which are the high ratio of
hydrophilic colloid stabilizers to polymer, the ~mall size of the particles, the
high molecular mass of the polymer molecules, and the molecular complexity
of the polymer molecules. High polymer molecular mass and molecular
complexity are a consequence of the polymer having been produced by
free-radical addition polymerization under conditions of high polymer/monomer
ratio (see section 10.2.2 below). Polymerization under this condition leads to
the formation of highly-branched polymer of high molecular mass which
eventually becomes crosslinked to form polymer of high gel content. All these
molecular factors tend to reduce the mobility of the segments of the polymer
molecule, and this reduces the tendency to interparticle integration. The
presence of a high ratio of colloid stabilizers to polymer provides further
barriers to interparticle integration, and thus aggravates the effect of the
molecular factors. In these respects, synthetic rubber latices contrast with
ammonia-preserved natural rubber latex: the ratio of colloid stabilizers to
polymer in the latter is lower, the particles are larger, and the polymer is
essentially linear and not extensively crosslinked.
10.1.2.5 Water-resistance of films dried from latices

Synthetic latices dry down to give deposits which tend readily to absorb
larger quantities of water again from aqueous media than do deposits from
ammonia-preserved centrifuged, creamed or electrodecanted natural rubber
latex. This characteristic is primarily a consequence of the high ratio of
hydrophilic colloid stabilizers to polymer. A further factor is that, insofar as
those stabilizers are anionic, the cation is usually derived from a non-volatile
'fixed' alkali, and not from a base, such as ammonia, which is capable of
volatilizing from the latex deposit as it dries.
At least two methods are available for reducing the water-sensitivity of
deposits from synthetic latices. Both approaches involve modifications in the
production of the latex. They are as follows:
1. Reduce the content of colloid stabilizers in the latex as far as possible, by
reducing the level of them in the emulsion polymerization reaction system
used to prepare the latex. Although simple in principle, this method can
cause problems in practice. Both the rate of polymerization and the colloid
stability of the resultant latex tend to be reduced if the amount of colloid
stabilizer in the reaction system is reduced. In any case, to the extent that
the colloid stabilizers are anionic, the deposits from the latex will still
contain some cations derived from a 'fixed' alkali.
2. Instead of reducing the level of colloid stabilizer in the emulsion polymerization
reaction system, replace any anionic colloid stabilizer which has a cation
derived from a non-volatile 'fixed' alkali by the equivalent colloid stabilizer
having a cation derived from a base which is capable of volatilizing from
the latex deposit as it dries. Substances of this type are sometimes known
as volatile base colloid stabilizers (see section 16.7.3.2.2 of Chapter 16
(Volume 3». Because they gradually lose their alkali by volatilization, the
anionic colloid stabilizer becomes converted to the acid form. The latter
is not significantly ionized if, as is often the case, the acid is a weak
carboxylic acid. The gradual conversion of the ionized form to the
essentially un-ionized form in the latex deposit brings about a significant
reduction in the water-sensitivity of the deposit. This approach can lead
to fewer problems during the emulsion polymerization reaction than the
alternative.
Deposits from synthetic latices tend to dry more slowly than do those
from ammonia-preserved centrifuged, creamed or electrodecanted natural
rubber latex. This is again a consequence of the high ratio of hydrophilic
colloid stabilizer to polymer, causing the deposits to become somewhat
hygroscopic during the latter stages of drying. Furthermore, there is a
pronounced tendency for deposits from synthetic latices to crack during the
final stages of drying. This phenomenon, sometimes referred to as mud-cracking,
is attributed to the low wet-gel strength of the deposit: the deposit is unable
to support the tensile stresses which develop as the deposit dries. If such
cracking is to be avoided, then the deposit must be dried slowly.
10.1.2.6 Mechanical stability of synthetic latices

It is a curious characteristic ofthe immediate product of emulsion polymerization
reactions that, on the one hand, they are remarkably resistant to mechanical
colloidal destabilization under conditions of high-speed stirring such as are
employed in the conventional test for the mechanical stability of ammonia-
preserved natural rubber latex (see section 8.7.1 of Chapter 8 (Volume 1»,
but, on the other hand, are very sensitive to the destabilizative effect of
mechanical pressure. Even light hand-rubbing is often sufficient to cause rapid
colloidal destabilization. The resistance to high-shear mechanical destabiliz-
ation is attributed to the small size of the particles in synthetic latices.
The reason for the susceptibility to colloidal destabilization by pressure is less
clear.
Introduction 143
10.1.2.7 Chemical composition

As is evident from section 9.5 of Chapter 9, the chemical constitution of
natural rubber latex is exceedingly complex. The fresh latex is the contents
of living cells. As such, it has the biochemical complexity of living organisms.
The industrial alkali-preserved concentrate is somewhat simpler in composition,
especially if produced by centrifugation, creaming or electrodecantation, for
the reasons which have been discussed in Chapter 9. Nevertheless the
concentrate contains many other chemical species besides polyisoprene and
water. Furthermore, the producer of preserved natural rubber latex concentrate
has little control over the detailed chemical composition of his product; he
has to accept it as produced by the Hevea brasiliensis tree, and modified by
the processes of preservation and concentration. By contrast, the chemical
composition of synthetic latices is relatively simple, and is in principle largely
controllable by the producer. The number of different chemical species present
in synthetic latices is small. The particles contained in synthetic latices are
usually of one type only, and are of simple structure unless deliberate attempts
have been made to produce particles of complex structure. One conse-
quence of the comparative chemical simplicity of synthetic latices is that
they are in principle inherently less variable than are natural latices,
because a greater degree of control is possible during their production.
However, the expected reduced variability in the case of synthetic latices is
not always realized in practice, for reasons which are sometimes not under-
stood.
The absence from synthetic latices of proteinaceous substances and other
natural products means that spontaneous coagulation and putrefaction do
not normally occur. Furthermore, objectionable odours seldom develop as
a result of the activity of microorganisms. However, synthetic latices do not
necessarily remain sterile. They are frequently able to support the life of
various microorganisms. Residual monomers, initiators, and their decomposition
products are frequently insufficiently toxic to prevent infection. For this
reason, it is common practice to add small amounts of microbicides to
synthetic latices to keep them sterile and thereby to prevent the formation
of by-products of microorganic activity, such as slimy deposits. Formaldehyde
is a typical microbicide which has been used at low conentration to inhibit
the activity of microorganisms in synthetic latices.
10.1.3 Literature pertaining to emulsion polymerization, and to synthetic

latices in general
There is now a very extensive technical literature pertaining to the subject
of emulsion polymerization, and to the latices produced by this reaction.
There are many original papers in the scientific and technical journals, review
articles and books, as well as an extensive patent literature. All that is possible
here is to indicate some of the many reviews which are presently available.
These will direct the reader to the very extensive body of original literature
which now exists pertaining to the various aspects of the subject.
Much of the early work relating to emulsion polymerization is described
in a book by Bovey et al. [1] published in 1955. The main concern is with
emulsion polymerization reactions of the type used to produce general-purpose
synthetic rubbers. This book summarizes in convenient form much of the
work carried out during World War II in pursuance of the United States of
America Synthetic Rubber Programme (see section 1.6.2 of Chapter 1 (Volume
1)), the primary objective of which was to produce general-purpose synthetic
rubbers which were usable as replacements for natural rubber in the
manufacture of items, such as tyres, essential to the war effort. Some of the
early work is also reviewed in a book edited by Whitby, Davis and Dunbrook
[2], published in 1954, which has synthetic rubbers as its main subject.
A comprehensive review of the theory and practice of emulsion polymerization
by the present author [3] was published in 1975. This has been supplemented
by three Specialist Periodical Reviews [4-6] by the same author. Useful
reviews by Duck [7] and by Gardon [8] have appeared in successive editions
of the Encyclopedia of Polymer Science and Technology. A more recent review
by Poehlein [9] has appeared in the Encyclopedia of Polymer Science and
Engineering. Other important reviews include those of Ugelstad and Hansen
[10], of Ugelstad, Hansen and Kaggerud [11], of Gardon [12, 13], and of
Dunn [14]. Theoretical aspects of the subject are dealt with extensively in
a very recent book by Gilbert [15], and the more practical aspects in a very
recent book edited by Lovell and El-Aasser [16]. An English translation of
a book by Eliseeva et al. [17] provides a convenient summary of the subject
as viewed from the Russian perspective. A series of important review articles
on various aspects of the subject can be found in a book edited by Piirma
[18]. Useful collections of papers are available in books edited by Piirma
and Gardon [19], by Bassett and Hamielec [20], by Fitch [21, 22], and by
Daniels, Sudol and El-Aasser [23]. The standard work on non-aqueous
synthetic latices, and the reactions by which they are produced, is the book
edited by Barrett [24].
In view of the extensive developments which have been made in recent
years in the subject area under review in this chapter, it is impossible to give
comprehensive coverage here, or to refer comprehensively to the relevant
literature. As regards emulsion polymerization, a brief resume of the subject
is given which should be sufficient for an understanding of the principles
which underlie the production of synthetic latices. No references to the
original scientific and technological literature are cited, other than those
which are necessary to validate specific matters. Readers who desire further
information are referred in the first instance to the general references cited
in the preceding paragraph.
Emulsion polymerization reactions 145
10.2 EMULSION POLYMERIZATION REACTIONS
10.2.1 Introduction
10.2.1.1 Outline description of a typical simple aqueous emulsion

polymerization reaction
The emulsion polymerization of styrene in water will be described in outline
as an example of a typical simple aqueous emulsion polymerization reaction.
This particular emulsion polymerization reaction is not one which has ever
been of great industrial importance. However, it is one which has been very
extensively studied for various reasons. The monomer is readily available in
a reasonably pure state. It is easy to handle in the laboratory. It readily
polymerizes in aqueous emulsion systems to give stable latices which contain
hard spherical particles. Furthermore, the free-radical addition polymerization
of styrene in bulk and in solution has been extensively studied; thus a
substantial corpus of information is available concerning the polymerization
behaviour of this monomer, quite separate from the complications which are
introduced when the reaction is carried out in the emulsion mode. A
consequence of these considerations is that many of our ideas concerning
the mechanism of emulsion polymerization have been developed using this
system, notwithstanding the comparative industrial unimportance of this
particular reaction.
There are four essential components in a simple conventional reaction
system for the aqueous emulsion polymerization of styrene: the monomer,
water, a colloid stabilizer, and an initiator. The colloid stabilizer is commonly
believed to fulfil two functions: it assists with the formation of a large number
of separate reaction loci in which most of the conversion of monomer to
polymer takes place; and it stabilizes the latex which gradually forms as the
reaction proceeds. Ionic surface-active substances such as sodium n-dodecyl
sulphate are particularly effective as colloid stabilizers for the emulsion
polymerization of non-polar monomers of low water-solubility, such as
styrene. The function of the initiator is to generate free radicals, thereby
initiating the polymerization reaction. In the simplest conventional emulsion
polymerization reaction systems, the initiator is water-soluble, the free
radicals being generated in the aqueous phase. The best-known water-soluble
initiators are the peroxodisulphates (persulphates) of monovalent cations,
notably potassium, ammonium and sodium. Because the free radicals are
generated from water-soluble initiators, those radicals are inevitably hydrophilic
in nature. With peroxodisulphate salts, the radicals are generated by thermal
decompositon of the peroxodisulphate ions into two sulphate radical-ions in
aqueous solution as follows:
S20~- --+ 2·0S0;
The sulphate radical-ions may then subseqently react with water molecules
to form bisulphate ions and hydroxide radicals as follows:
Thus initiation of polymerization may occur by reaction of a monomer

molecule either with a sulphate radical-ion or with a hydroxide radical. In
either case, the initiating radical is hydrophilic, and the propagating polymer
chain has a hydrophilic moiety at the end at which it was initiated.
Thus a simple conventional reaction system for the aqueous emulsion
polymerization of styrene comprises styrene, water, sodium n-dodecyl sulphate
and potassium peroxodisulphate. Typical proportions for the latter three
components are respectively 250, 5 and 0.5 parts by mass per hundred parts
by mass of monomer (pphm). The sodium n-dodecyl sulphate and potassium
peroxodisulphate are dissolved in the water at normal ambient temperature.
The styrene is then added with stirring. There is no requirement that the
styrene should be emulsified in the aqueous phase. The system is then purged
with nitrogen, and the temperature raised to ca. 60·C, continual gentle stirring
being maintained. Rapid polymerization soon commences. It proceeds
smoothly almost to completion after a period of ca. 2 hours. The product is
a stable latex which contains almost monodisperse spherical particles of
polystyrene of high molecular mass, having diameters in the region of 70 nm.
As will appear subsequently, the conversion-time curve for simple aqueous
emulsion polymerizations of this type is S-shaped. Initially, the rate of
polymerization increases with increasing time, the conversion-time curve in
this region giving the impression that the reaction is auto-catalysed. The rate
of polymerization soon reaches a maximum. It remains at this maximum
value for a substantial proportion of the monomer-polymer conversion
reaction. The conversion-time curve over this region is linear. The rate of
polymerization then declines, eventually becoming zero. The stage at which
the rate of polymerization becomes effectively zero does not necessarily
coincide with complete conversion of monomer to polymer, although it does
in the simplest types of reaction.
10.2.1.2 Comparisons and contrasts between emulsion polymerization and

other modes of effecting free-radical polymerization
10.2.1.2.1 Mechanism of free-radical addition polymerization in outline

Emulsion polymerization is one of several modes of carrying out the
free-radical addition polymerizaton of olefinically-unsaturated compounds.
In all these modes, the polymerization takes place by a free-radical chain
reaction which involves the following three principal types of reaction:
1. an initiation reaction, in which free radicals are formed;

2. a series of propagation reactions in which many monomer molecules are
successively added to a free-radical centre, thereby producing an essentially

linear molecule of large size; and
3. a termination reaction, in which the activity of the free radical is lost, and
a stabilized non-propagating macromolecule remains.
The overall polymerization reaction may be further complicated by the

occurrence of one or more types of transfer reaction. In these, the activity
of a propagating chain is transferred to another molecule, such as monomer,
solvent, polymer, initiator or substance deliberately added to promote chain
transfer. The products of transfer reactions are a de-activated non-propagating
macromolecule and a free-radical site on another molecular species. Further
propagation may then occur from the latter by the addition of successive
monomer molecules. Transfer reactions to species of low molecular mass,
such as monomer or solvent, have in common that they reduce the average
number of monomer units which become incorporated in the polymer
molecules being produced. This number is known as the chain length or
degree of polymerization; the molecular mass of the polymer formed is directly
proportional to this number. Transfer reactions to polymer molecules lead
to the formation of branched and crosslinked macromolecules.
Addition polymerizations of olefinically-unsaturated compounds can also
occur by way of propagating active centres other than free radicals, notably
anionic and cationic centres. However, almost without exception, addition
polymerizations of olefinically-unsaturated compounds which are effected in
the emulsion mode take place by way of free-radical propagating centres.
10.2.1.2.2 Bulk polymerization

The simplest method of effecting free-radical addition polymerization is by
what is known as bulk polymerization. This is a process in which the monomer
is polymerized in bulk form (usually liquid, but occasionally gaseous) with
merely the addition of initiating substances and (possibly) the application of
heat, ultraviolet light, or high-energy electromagnetic radiation. This method
is notable chiefly for the absence of added substances. There is thus no
problem of separating the polymer from solvents or dispersion media. There
are no solvents present during the polymerization reaction which may act
as chain-transfer agents, thereby reducing the degree of polymerization of
the polymer which is formed. In many cases, the polymer is soluble in the
monomer from which it was derived. The reaction system then remains
homogeneous as polymerization proceeds, gradually changing from liquid
monomer through viscous and elastoviscous liquid to solid polymer. In some
cases, however, the polymer is not soluble in the monomer from which it
was derived. The reaction system then becomes inhomogeneous as
polymerization proceeds, the reaction system comprising a slurry of polymer
in monomer at intermediate states of conversion between pure monomer and
pure polymer.
Although in principle bulk polymerization is a very simple method of

effecting free-radical addition copolymerization, it is beset with several
difficulties. The chief of these are that the system becomes intractably viscous
as the reaction proceeds, the dissipation of the heat of polymerization becomes
increasingly difficult as the viscosity increases, the product is likely to be of
variable structure because of local increases in temperature which may occur
during the reaction, and the polymer produced in the latter stages of the
reaction is often highly branched or crosslinked. The maximum rate of
polymerization in a bulk reaction is limited by two main factors: the problem
of heat dissipation, and the degree of polymerization which it is desired to
achieve by means of the reaction. As regards the second of these factors,
whereas the rate of propagation is kinetically of first order with respect to
the concentration of propagating free radicals, the rate of termination is
usually second order with respect to this concentration. This is because the
usual mode oftermination is bimolecular interaction between two propagating
free radicals, with consequent de-activation of those propagating centres.
Any attempt to increase the rate of polymerization by increasing the
concentration of propagating free radicals (e.g., by increasing the concentration
of initiator) is accompanied by a larger increase in the rate of termination.
In the absence of transfer reactions, the degree of polymerization of the
polymer formed is determined by the ratio of the rate of chain propagation
to that of chain termination. Thus any increase in the rate of polymerization
is inevitably accompanied by a decrease in the degree of polymerization of
the polymer which forms. The effect of transfer-to-monomer reactions is to
reduce further the attainable degree of polymerization.
/0.2.1.2.3 Solution polymerization

As its name implies, solution polymerization is polymerization which is carried
out in the presence of a solvent for the monomer. The solvent may also
dissolve the polymer. If so, then the end-product of the reaction is a solution
of polymer in solvent. If the solvent does not dissolve the polymer, then the
polymer precipitates from the reaction system as it forms. The result is a
slurry of polymer in a mixture of solvent and unpolymerized monomer. The
reaction system is consequently, in part at least, heterogeneous. Under
favourable circumstances, the polymerization system will remain fluid over
the whole course of the reaction, if sufficient solvent is present.
The principal reason for using a solvent is to prevent the polymerizing
system becoming intractably viscous as polymerization proceeds. The problem
of heat dissipation is thereby mitigated to some extent. Furthermore, the rate
of generation of heat of polymerization per unit volume of reaction system
is reduced, partly because the system has been diluted by non-polymerizing
solvent, and partly because the rate of polymerization is reduced as a
consequence of the dilution. Offsetting these advantages, there are a number
of disadvantages. Solution polymerization requires the use of expensive, toxic
and often flammable solvents. Separation of polymer from solvent is necessary

at the end of the reaction, unless, of course, the polymer is required for use
as a solution in the solvent or as a slurry in what has become, in effect, a
dispersion medium. Separation from the solvent can be effected by filtration
if the polymer has precipitated during polymerization. If not, then a more
complex operation, such as solvent distillation, is necessary. A further
disadvantage of solution polymerization relative to bulk polymerization is
that the degree of polymerization of the polymer formed may be considerably
reduced by the reaction of transfer to solvent.
10.2.1.2.4 Suspension polymerization

Suspension polymerization is a process in which the monomer is dispersed in
a medium (usually aqueous) in which it is insoluble, and conditions established
such that polymerization occurs within the individual monomer droplets.
Small amounts of colloid stabilizers (so-called protective colloids) are usually
added to stabilize the dispersion to some extent. Monomer-soluble initiators
are used. This type of reaction is appropriately regarded as a 'micro-bulk'
polymerization, i.e., a bulk polymerization in which the polymerizing
monomer has been subdivided into a large number of droplets. The kinetics
and mechanism of the reaction are essentially those for the bulk polymerization
under the same conditions of temperature and initiator concentration in the
monomer phase. Droplet size may have a slight effect upon the polymerization
rate, especially when the droplets are very small.
Suspension polymerization offers several advantages over bulk and solution
polymerizations. Ifcarried out properly, the reaction system remains completely
fluid throughout the reaction. Heat dissipation is a minor problem only,
partly because the system has been diluted by the dispersion medium, but
principally because the heat of reaction can be absorbed and conducted away
by the dispersion medium. The polymer produced by this type ofpolymerizaton
is less pure than that obtained by bulk polymerization, but is not contaminated
to the same extent as that produced by a conventional emulsion polymerization
reaction. If water is used as the dispersion medium, then problems of expense,
toxicity and flammability are minimized. It is necessary to separate the
polymer from the dispersion medium, but this is not usually difficult. The
main practical problem associated with this process is the prevention of
droplet coalescence in the intermediate stages of the reaction when the
droplets are very tacky, consisting of what is essentially monomer-
plasticized polymer. In the early stages of the reaction, the droplets are
sufficiently fluid to break up continuously as they coalesce. In the latter
stages, they are too hard to coalesce. In the intermediate stage, they coalesce
readily but re-disperse with difficulty. Careful attention to stirring is necessary
if the reaction is to be carried through to completion satisfactorily.
The end-product of a solution polymerization process is a dispersion of
spherical polymer particles in a medium which is usually aqueous. The
product might therefore be regarded as a type of latex. That these products

are not normally so regarded is mainly a consequence of the very large size
ofthe particles relative to those present in conventional latices. It is customary
to refer to the polymer spheres formed during solution polymerization as
beads rather than as particles.
10.2.1.2.5 Emulsion polymerization

From the description of a conventional emulsion polymerization reaction
system which has been given in section to.2.l.l above, this reaction bears a
superficial resemblance to suspension emulsion. The polymerization is again
carried out in a system comprising monomer and a dispersion medium
(usually aqueous) in which the monomer is only sparingly soluble. Suitable
colloid stabilizers must be present, but, in contrast to suspension polymerization,
the initiator is usually water-soluble. Again the system remains fluid over
the entire extent of the reaction. Heat dissipation is no problem, again partly
because of the dilution of the polymerizing system and partly because the
heat of polymerization is conducted away by the dispersion medium. Provided
that the reaction system has been correctly formulated, there is little tendency
for the particles to coalesce as the reaction proceeds. The product is obtained
in the form of a latex; this is an obvious advantage if it is desired to use the
polymer in this form. Apart from these important advantages, the dominant
advantage of the emulsion polymerization process is that, in contrast to the
other methods of effecting free-radical addition polymerization, it is possible
to obtain polymer of high degree of polymerization at relatively fast rates.
High rate of polymerization is advantageous because it leads to reduced
production costs. A product of high degree of polymerization is especially
important when producing synthetic rubbers, because the develoment of
satisfactory elastomeric properties requires high degree of polymerization of
the primary polymer. However, the degree of polymerization of polymers
produced by emulsion polymerization can be too high for practical applications;
for this reason, it is sometimes necessary to include within the reaction system
a chain-transfer agent (or modifier) in order deliberately to reduce the degree
of polymerization and molecular complexity of the product.
The emulsion polymerization process has certain disadvantages attending
its use. The chief of these is that, if the polymer is required in bulk form,
then it has to be separated from the latex. The separation is not always
simple. Furthermore, the polymer which is isolated from the polymer latex
has associated with it a relatively large proportion of non-polymer constituents,
notably, insoluble derivatives of the colloid stabilizers. Although for certain
polymer applications, the presence of these derivatives may be advantageous,
emulsion polymers are effectively excluded from those applications which
require a polymer of high purity. A further disadvantage of emulsion
polymerization is its variability. Whilst this may not be a serious problem
when producing, say, conventional styrene-butadiene copolymers required
for processing as general-purpose synthetic rubbers, it can be a greater

problem for the production of special-purpose grades of synthetic latices.
These problems are aggravated as increasingly stringent restrictions are
imposed upon the levels and types of ancillary recipe ingredients.
Two further general, and to some extent related, comments concerning
emulsion polymerization reactions should be noted at this stage:
1. It has become increasingly realized in recent years that the essential feature
which distinguishes free-radical addition polymerization reactions carried
out in the emulsion mode from reactions carried out in other modes is
the manner in which the various steps of initiation, propagation and
termination are distributed physically within the reaction system. In bulk,
solution and suspension polymerizations, these steps all occur homogeneously
within a single reaction system (which is, of course, subdivided in the case
of suspension polymerization). By contrast, in an 'ideal' emulsion
polymerization reaction, the propagation step is isolated from the initiation
step, and the individual propagation reactions are also isolated from each
other. For this reason, the individual propagation reactions are said to
be compartmentalized in an emulsion polymerization reaction. Termination
is retarded relative to propagation, because an additional free-radical has
to become available to a propagating chain before mutual termination
becomes a possibility. It is for this reason that it is possible by emulsion
polymerization to produce a polymer of high degree of polymerization at
high rates of polymerization. In the past, the idea of compartmentalization
ofthe propagation step in emulsion polymerization reactions has sometimes
been expressed by saying that emulsion polymerization proceeds by way
of a micellar mechanism. This is because, for many years, it was thought
that the individual loci into which the propagation reaction is compart-
mentalized are principally formed from the micelles into which the
molecules of colloid stabilizers are associated in aqueous solution. This
is still regarded by many workers in the field as an important mechanism
by which loci are formed in many emulsion polymerization reaction
systems (see section 10.2.2.2.1 below). It is not, however, the only
mechanism. This is one of the reasons why a distinction is now made
between, on the one hand, the undisputed fact of compartmentalization
of the propagation step into a large number of separated loci, and, on the
other, speCUlation as to the origin of those loci.
2. The industrial development of synthetic latices for use as such was
undoubtedly assisted by the possibility of producing by emulsion
polymerization polymers of high degree of polymerization at high rates
of polymerization. The principal incentive to developing free-radical
emulsion polymerization processes came from the requirement to produce
general-purpose synthetic rubbers during World War II: as noted above,
a high degree of polymerization is an essential prerequisite for the
development of satisfactory elastomeric properties in a polymer. To a

lesser extent, incentive to develop emulsion polymerization processes was
also provided by the desire in the years immediately preceding World
War II to produce special-purpose synthetic rubbers which had properties
not attainable in vulcanizates from natural rubber. The most important
of such properties is resistance to swelling in hydrocarbon oils. For these
reasons, manufacturers of synthetic rubbers found themselves producing
and processing the rubber in latex form as an intermediary to the bulk
rubber. It was therefore inevitable that, in the years immediately following
World War II, those manufacturers should seek to develop synthetic latices
which could be used as such rather than as mere intermediaries to polymers
in bulk form.
10.2.1.2.6 Reasons for believing the propagation step in emulsion

polymerization reactions to be compartmentalized
We conclude this discussion of the comparisons and contrasts between the
various methods of carrying out free-radical addition polymerization by
summarizing the reasons for believing that emulsion polymerization contrasts
with the other modes of effecting free-radical addition polymerization in that
the progagation step is compartmentalized, so that most of the monomer is
converted to polymer in a large number of separated reaction loci. Four
types of site can be envisaged for the principal locus of the reaction:
I. the monomer droplets;
2. the interface between the monomer droplets and the aqueous phase;
3. the aqueous phase;
4. reaction loci which have formed within the aqueous phase, but are not
part of the aqueous phase.
The monomer droplets are the most obvious principal locus for the
polymerization. However, there are cogent reasons for rejecting this hypothesis.
There are also cogent reasons for rejecting the second and third of the above
possibilities as principal loci. By elimination. the conclusion is reached that
most of the polymerization occurs in reaction loci which have formed within
the aqueous phase. This hypothesis is found to be consistent with most of
the characteristics of the reaction.
The principal reasons for rejecting the monomer droplets as the principal
locus for the polymerization are as follows:
I. The emulsified monomer droplets are usually very much larger than are
the particles of the latex which is formed by the reaction.
2. The initial rate of polymerization, expressed as mass of monomer converted
to polymer in unit time, is independent of the volume of monomer phase,
provided that the volume and composition of the aqueous phase are kept
constant.
3. The rate of polymerization, expressed as in (2), increases as the volume
of the aqueous phase increases, provided that the composition of the

aqueous phase is kept constant.
4. The rate of polymerization depends upon the amount of colloid stabilizer
present in the reaction system. However, in many cases, after the reaction
has reached a certain stage, addition of neither further colloid stabilizer
nor initiator causes any significant increase in the rate of polymerization.
5. Typical emulsion polymerization reactions are often most easily initiated
by water-soluble initiators, rather than by monomer-soluble initiators.
6. The kinetics of emulsion polymerization reactions, including, inter alia,
the relationship between degree of polymerization and the rate of
polymerization, differ markedly from those for bulk polymerization. The
typical emulsion polymerization is not a 'micro-bulk' polymerization.
7. Although it is usual to carry out emulsion polymerization reactions in
systems in which the monomer has been at least crudely emulsified in the
aqueous phase, prior emulsification of the monomer is not a prerequisite
for a typical conventional emulsion polymerization reaction to occur. A
product similar to that obtained from a typical emulsion polymerization
reaction can be formed even if the monomer is present as a separate phase
in the system. Even more telling is the observation that there need not
even be direct physical contact between the liquid monomer phase and
the aqueous phase: a product similar to that obtained from a typical
emulsion polymerization reaction can be formed provided that the
molecules of monomer can reach the aqueous phase via the vapour phase.
These observations provide conclusive demonstration that there is no
relationship between the size of the particles in the eventual latex which
is formed, and the state of subdivision of the initial monomer phase.
Considerations such as these exclude the emulsified monomer droplets as
the principal locus of the reaction. The seventh in the above list is particularly
persuasive. These considerations also effectively exclude the interface between
the monomer and the aqueous phase as the principal locus; for, on this view,
the reaction would commence in the interface, and then the propagating
radicals would penetrate the monomer droplets. Whilst this view would
accommodate the fact that the typical emulsion polymerization reaction is
initiated by water-soluble substances, it is inconsistent with the other
characteristics listed above.
The true aqueous phase is rejected as the principal locus on two main grounds:
1. Emulsion polymerization can proceed quite smoothly and rapidly with
many monomers of varying water solubilities, even with monomers which
are almost insoluble in water.
2. If two monomers are emulsion copolymerized, it is the less water-soluble
of the two which becomes incorporated in the polymer to a greater extent
than in a bulk polymerizaton employing the same overall monomer feed
composition.
Although these observations effectively exclude the aqueous phase as the

principal locus for polymerization, it should be noted that there are some
emulsion polymerization reactions, notably those which use the more
water-soluble monomers, for which the initial reaction almost certainly does
occur in the aqueous phase. An example of considerable industrial importance
is the emulsion polymerization of vinyl acetate.
It will be apparent from the evidence which has been summarized above
that the case for a mechanism in which the principal monomer-polymer
conversion reaction is compartmentalized in reaction loci dispersed within
the aqueous phase is essentially negative. Historically, this type of mechanism
has been accepted largely because all the plausible alternatives appeared to
be unsatisfactory. Direct evidence for this mechanism is, however, hard to
find; this has not gone unnoticed in recent years. This type of mechanism is
appropriately known as the compartmentalization mechanism; it generally
gives a reasonable explanation of many of the observed phenomena.
In view of the conclusion that the monomer droplets are not the principal
locus of the polymerization, it might be thought that they playa subsidiary
role in the reaction. The view commonly held at present is that, insofar as
some of the unreacted monomer remains in the reaction system as a separate
droplet phase, those droplets function as reservoirs of monomer which
maintain the concentration of monomer in the other phases of the reaction
system at essentially the equilibrium value. The driving force underlying the
maintenance of an equilibrium concentration of monomer in the other phases
derives from the thermodynamic requirement that at equilibrium the chemical
potential of monomer molecules must have the same value in all phases of
the reaction system. As long as excess monomer is present as a separate
droplet phase, the equilibrium condition will be such that the chemical
potential of monomer has the same value in the other phases (including the
aqueous phase) as it has in the monomer droplets. There has been considerable
controversy in the past as to whether equilibrium with respect to monomer
concentration is or is not rapidly attained, and hence maintained, as
polymerization proceeds. The present view is that equilibrium is rapidly attained.
10.2.2 Mechanism and kinetics of emulsion polymerization reactions
10.2.2.1 The three 'intervals' ofa conventional emulsion polymerization reaction

It has been customary for many years to regard the course of a conventional
emulsion polymerization reaction as being divided into three more-or-less
distinct stages or intervals. The criteria by which these intervals are delineated
are defined from the S-shaped conversion-time curve for the reaction, to
which reference has been made in section 10.2.1.1 above. A typical
conversion-time curve is illustrated schematically in Fig. lO.l(a). The criteria
which delineate the intervals are also evident from the variation of rate of
c c
.Q
~
.!::!
1ii
N
"5j
~
E E ~
.l!!
~
0
c.
~
0
c.
.5 =
'0 '0 ~
'E 'E .l!!
.5
;n
Q)
~
w
(8) Time of reaction (b) Time of reaction
Figure 10.1 (a) Schematic conversion-time curve for typical emulsion polymerization reaction,
showing delineation of the three intervals. (b) Schematic illustration of corresponding variation
of rate of polymerization with time of reaction, showing again the delineation ofthe three intervals.
polymerization with time of polymerization, illustrated schematically in Fig.

10.1 (b). These three intervals are as follows:
1. Interval I: This is the stage of the reaction in which are formed the loci
in which most of the monomer-polymer conversion will take place
subsequently. It is the stage of the reaction in which the entities are formed
which will subsequently develop into the eventual latex particles. This
stage is often referred to as partial nucleation or locus nucleation. In relation
to the conversion-time curve for the reaction, this interval is identified as
that stage of the reaction in which the rate of polymerization increases
with increasing time of reaction. Typically, the proportion of monomer
which has been polymerized by the end of Interval I is very small.
2. Interval II: This is the stage of the reaction during which polymerization
occurs within the loci formed during Interval I, and in which excess
monomer is present as a separate droplet phase. The number of reaction
loci during this stage is believed to be essentially fixed at the number
which had become formed at the end of Interval I. The monomer/polymer
ratio in the reaction loci during this interval is believed to be constant.
In relation to the conversion-time curve for the reaction, this interval is
identified as that stage of the reaction in which the rate of polymerization
is constant. Thus in this interval, the polymerization reaction appears to
be kinetically of zero order with respect to overall monomer concentration
in the reaction system. Typically, ca. 50% of the monomer has been
polymerized by the end of Interval II.
3. Interval III: This is the final stage of the reaction. During this interval,
polymerization continues within the loci which were formed during
Interval I, and which persisted and grew during Interval II. However,
excess monomer is no longer present as droplets. In the ideal case, the
number of reaction loci during this stage is believed to be essentially fixed

at the number which had become formed at the end of Interval I. The
monomer/polymer ratio in the reaction loci during this interval decreases
continually. In relation to the conversion-time curve for the reaction, this
interval is identified as that stage of the reaction in which the rate of
polymerization gradually diminishes to zero with increasing time of
polymerization. Ideally, it would be expected that the polymerization
reaction would appear to be kinetically of first order with respect to overall
monomer concentraton in the reaction system. However, the kinetics
during Interval III are usually more complex than would be indicated by
this simple expectation.
10.2.2.2 Theories of particle nucleation in emulsion polymerization reactions
10.2.2.2.1 Micellar nucleation

Many mechanisms have been proposed to explain the origin of the reaction
loci in emulsion polymerization reactions, and hence the origin of the particles
in the latex which is eventually formed. Attention here will be principally
confined to the two most important of these mechanisms, but brief mention
will be made of two others which have been proposed. The first of the two
mechanisms to be considered in some detail is known as micellar nucleation.
For many years, this was regarded as the mechanism for particle nucleation,
although it has long been recognized that it is incapable of explaining
satisfactorily why particles become nucleated in certain types of emulsion
polymerization reaction system.
According to the micellar mechanism for particle nucleation, the particles
are formed by polymerization of monomer molecules which have been
absorbed into the hydrophobic interior of the aggregates of colloid stabilizer
molecules, known as micelles, which often form in aqueous solutions of these
substances. Micelles typically contain ca. 50 primary colloid stabilizer
molecules. The colloid stabilizer molecules in the micelles are orientated such
that their hydrophilic moieties are in contact with the water and their
hydrophobic moieties are as far removed from the water molecules as possible,
thus forming the interior of the micelle. The interior of the micelles is
consequently hydrophobic in nature, and provides an environment into which
other hydrophobic molecules can accumulate preferentially to the aqueous
phase. The process by which molecules of hydrophobic susbstances, such as
monomers, become preferentially absorbed into the interior of colloid
stabilizer micelles in aqueous solution is known as solubilization. The effect
of solubilization is to increase the apparent (but not the true) solubility of
the hydrophobic substance in the aqueous phase. According to the theory
of micellar nucleation, a small proportion ofthe micelles containing solubilized
monomer molecules acquire a free radical from the surrounding aqueous
phase. Polymerization of the solubilized monomer is then initiated. As
monomer is consumed by polymerization, so more monomer is absorbed

into the interior of the micelle, the motivation being the requirement that at
equilibrium the chemical potential of monomer molecules must be maintained
at the same value in all phases of the reaction system. As long as excess
monomer is present as a separate droplet phase (which is certainly so during
Interval I of most emulsion polymerization reactions), the equilibrium
condition will be such that the chemical potential of monomer has the same
value in all the micelles (nucleated and non-nucleated) as it has in the
monomer droplets. The micelles in which polymerization has been initiated
thus grow in size as a consequence of further absorption of monomer, and
of continued polymerization of the increased amount of monomer now
present in the interior. The surface area of these micelles increases, and the
concentration of colloid stabilizer molecules at the surface decreases. This
leads to adsorption of colloid stabilizer molecules from the aqueous phase,
and hence to dissociation of colloid stabilizer micelles in which polymerization
has not commenced, the motivation being preservation of equilibrium
throughout the reaction system in respect of individual colloid stabilizer
molecules which are present in various forms. Eventually the supply of colloid
stabilizer molecules from non-nucleated micelles is exhausted. Further
nucleation of micelles is now impossible. Particle nucleation ceases. The
disappearance of non-nucleated micelles from the reaction system marks the
end of Interval I and the commencement of Interval II. The essential features
of the process of micellar nucleation can be summarized as follows:
I Dissociation
,
Micelles of colloid stabilizer
~ + radical
Individual molecules of Particle containing propagating
colloid stabilizer dissolve polymer chain - as monomer depleted,
in aqueous phase so further monomer diffuses from
monomer droplets, and particles
A_d_so~r~Pt-io-n~o~n~t-o----g-ro-w-i-n~COnsequ~
L -_ _ _ _ _ _ _ _ _
growing particles Monomer from

droplets
The nucleated micelles which remain at the end oflnterval I are appropriately
described as being latent latex particles, being the entities which become the
polymer particles in the latex which eventually forms. Comparison between
the concentration of micelles in the initial aqueous phase of a typical emulsion
polymerization reaction system and the concentration of particles in the final
latex shows that the fraction of micelles which become nucleated to form
latent latex particles is very small, being typically ca. 0.001. The majority of
the micelles are sacrificed to provide colloid stabilizer for the particles as
they grow in size after having been nucleated.
The theory of micellar nucleation was first proposed by Harkins [25-28]
and subsequently developed quantitatively by Smith and Ewart [29]. The
theory has been re-examined by Gardon as part of a comprehensive
consideration ofthe quantitative aspects of emulsion polymerization reactions
in which nucleation occurs by the micellar mechanism; for further details
and references, the reader is referred to two summarizing papers by Gardon
[30,31]. The theory of micellar nucleation was developed primarily to explain
the formation of particles in the type of emulsion polymerization reaction
system which was used for the production of general-purpose styrene-
butadiene rubbers during World War II. This type of system contains
hydrocarbon monomers of low water solubility, and relatively high
concentrations of micelle-forming surface-active substances, usually carboxylate
soaps.
Central to any theory of particle nucleation in emulsion polymerization
reactions is the ability satisfactorily to predict the concentration, N, of
particles present in the reaction system at the end of the nucleation process,
or at least the dependence of N upon certain obvious parameters which may
be relevant to the nucleation process. The latter include the concentrations
of colloid stabilizer and initiator in the aqueous phase of the reaction system.
The quantitative theory of Smith and Ewart [29] predicts that N is jointly
proportional to the %th power of the concentration of micelle-forming colloid
stabilizer and the %th power of the initiator concentration (see below), the
latter prediction being subject to a proviso which will become clear
subsequently. Both concentrations are with reference to the aqueous phase.
Exact prediction of N is difficult. Smith and Ewart solved the problem for
two idealized situations, one of which should give more particles, and the
other fewer particles, than are generated in the actual reaction system. The
results for the two idealized situations are remarkably similar. Smith and
Ewart then argue that the value of N for the real system will lie between
these two extremes. The calculation giving too many particles is made on
the assumption that, as long as micelles are present, they capture all the
radicals available for capture. This assumption predicts the formation of too
many particles, because some of the radicals will be captured by particles
which have been nucleated by a previous radical. The calculation giving too
few particles assumes that a given interfacial area always has the same
effectiveness in capturing free radicals, regardless of the size (and therefore
curvature) of the particle upon which it is situated. This assumption predicts
the formation of too few particles, because it is known from classical diffusion
theory that the flux of diffusing matter across unit area of interface is inversely
proportional to the radius of curvature of the interface.
The calculation giving too few particles proceeds as follows: The stated
assumption gives the value of N as pt*, where t* is the duration of Interval I
and p is the rate at which free radicals are generated in unit volume of the
aqueous phase of the reaction system. Let v(r, t) denote the volume at time
t of a particle in which polymerization was initiated at time -r. Then
v(-r, t) = Vo + Jl(t - -r) (10.1)
where Vo is the initial volume of the micelle, and Jl, = dvldt, is the rate of
increase of volume of the nucleated particle due to polymerization and
monomer absorption. In what follows, t, and hence -r, are measured from the
commencement of Interval I. Jl is taken to be constant, provided that, as
assumed, a nucleated reaction locus contains one, and only one, propagating
radical throughout Interval I. Vo is assumed to be negligible, so that
v(-r, t) ~ Jl(t - -r). The surface area, s(T, t), at time t of a particle which was
nucleated at time -r is then calculated from v(-r, t) by noting that s = 41ta 2
and v = 41ta 3 /3, where a is the particle radius at time t. Eliminating a from
these two relationships gives s = (361tV 2 )1/3. Thus
(10.2)
This can be written as
s(r, t) = lJ(t - -r)2/3 (10.3)
where () = -Y361tJl2. The procedure is then to calculate the variation of

aggregate surface area of nucleated particles with time t < t·, and from this
to infer the amount of colloid stabilizer which is adsorbed on this surface.
The nucleation of new particles is deemed to cease, and hence Interval I to
end, when this amount of colloid stabilizer is equal to the total amount of
colloid stabilizer in the reaction system. This enables t· to be calculated as
a function of the total amount of colloid stabilizer in the system. N is then
obtained as pt·. Additional assumptions made in the course of the argument
include the following:
1. the amount of colloid stabilizer present dissolved in the aqueous in
non-micellar form is negligible relative to the total amount of colloid
stabilizer in the system;
2. the surface of the nucleated particles is always saturated with adsorbed
colloid stabilizer molecules right up to the end of Interval I;
3. the surface area, ro; occupied by a single colloid stabilizer molecule when
adsorbed at the surface of a nucleated particle is constant throughout
Interval I.
The aggregate surface area at time t of all the particles in unit volume of
aqueous phase which have been nucleated up to time t < t·, denoted by
Sp(t), is calculated by noting that the contribution made to Sp(t) by the
particles which were nucleated in the short interval of time between rand
-r + ~-r is ps(T, t)~-r. Sp(t) is then the sum of all such contributions from particles
which were nucleated over the period from = 0 to T = t. Thus
I I
T
Sp(t) = pS(T, t) dT = pO(t - T)2/3 dT (10.4)
Evaluation of the integral gives

Sp(t) = %pOt S/3 (10.5)
Given the assumptions which have been stated above, the time t* must be
such that Sp(t*) = zuO, where 0 is the total number of molecules of colloid
stabilizer initially present in unit volume of the aqueous phase of the reaction
system. 0 is, of course, directly proportional to the mass of colloid stabilizer
initially present in unit volume ofthe aqueous phase ofthe reaction system. Thus
%pOt*S/3 = zuQ (10.6)
Solving for t*, and putting N = pt*, 0 = ~36nJ.l2 gives
N = (%)3/S(36n)-1/SJ.l- 2/Sw/ Sp2/S03f5 ~ O.S3J.l-2!5W/ Sp2/S03/S (10.7)
The calculation for too few particles is more complicated. Perhaps rather
surprisingly, it gives the same functional dependence of N on J.l, l1T, P and 0
as does the calculation for too many particles. The only difference is that
the numerical pre-factor is 0.37 instead of 0.53. Thus the Smith-Ewart theory
of micellar nucleation predicts the concentration of particles formed during
Interval I to be directly proportional to the factor J.l- 2/S W /S p2!5 0 3 / 5 . It is
not possible from this theory to predict the absolute value of N, not merely
because of the uncertainty concerning the numerical pre-factor, which could
reasonably be taken as, say, 0.45. In addition to this numerical pre-factor, it
is necessary to know J.l, l1T, P and O. 0 can be calculated from the formulation
for the reaction system and the molecular mass of the colloid stabilizer. zu
is either known or can be reasonably inferred, given the type of colloid
stabilizer in the reaction system. p can be calculated from the concentration
of the initiator in the reaction system and the decomposition kinetics of the
initiator, if information is available concerning the latter. The problem is the
value to be assigned to the quantity J.l, the growth rate for individual particles.
However, although in the absence of a value for J.l, the theory cannot be used
to predict absolute values of N, it does nevertheless predict that N is directly
proportional to 0 3 / 5 and p2/S. The latter prediction implies that N is directly
proportional to the %th power of the level of initiator in the reaction system
if, as is usually the case, the initiator decomposes to give free radicals at a
rate which is kinetically of first order in initiator concentration. There is
good experimental evidence that these predictions are fulfilled in practice by
certain emulsion polymerization reaction systems.
In concluding this brief discussion of the Smith-Ewart theory of micellar
nucleation, attention is drawn to the following additional points:
1. The quantities p and 0 in the theory refer to Interval I of the reaction,
and not to any subsequent values which may differ from those during
Interval I because further additions of colloid stabilizer and/or initiator
have been made to the reaction system.
2. The exponents of p and Q in the prediction for N arise as a consequence
of the relationship between the surface area and the volume of geometrically-
similar particles, and not as a consequence of the kinetics of the
polymerization reaction itself.
3. It is not fortuitous that the sum of the exponents of a and p in the
expression for N is 1. This is a consequence of t· being proportional to
(Q/p)« (equation (10.6), where in this case a = Ys), and of N then being
given by pt·. These two relationships lead to N ex:: all pi-ll.
10.2.2.2.2 Homogeneous nucleation

Although micellar nucleation was for many years widely regarded as the
principal mechanism of particle nucleation in emulsion polymerization
reactions, it has also long been realized that this theory is incapable of
explaining why particle nucleation can occur quite readily in reaction systems
which initially at least contain no micelles derived from colloid stabilizers.
It is well-known that more water-soluble olefinic monomers, such as vinyl
acetate, are able to emulsion polymerize in reaction systems which contain non-
micellizing colloid stabilizers, and, indeed, in systems which do not contain
any added colloid stabilizer at all. Such reactions are often described as
emulsifier-free emulsion polymerizations. It has also been found possible to
emulsion polymerize monomers of low water-solubility, notably styrene, in
emulsifier-free reaction systems, and also in reaction systems which contain
a micellizing colloid stabilizer but at such a concentration that micelles would
not have been present. These observations have demonstrated that particles
can be nucleated by other than the micellar mechanism. The further question
then arises as to the importance of micellar nucleation for those reaction
systems which do contain a micellizing colloid stabilizer in sufficient concentration
for micelles to form.
The best-known alternative theory for particle nucleation is that of
homogeneous nucleation. According to this theory, polymerization commences
amongst monomer molecules dissolved in the aqueous phase, and not
amongst monomer molecules solubilized within any micelles of colloid
stabilizer which may be present. The reaction which leads to particle
nucleation is a solution polymerization which takes place in a homogeneous
phase in which are being generated the free radicals which initiate the
polymerization; this is why the mechanism is distinguished by the adjective
homogeneous. Assuming that the polymerization is initiated by sulphate
radical-ions, the initial solution polymerization can be represented as:
M + ,O.SO; --+ . M.O.SO;
M + .MMn'OSO~ --+ . MMn+ l'O.SO~

in which M denotes a monomer molecule. A particle is nucleated when a

propagating oligomer in the aqueous phase attains a sufficient degree of
polymerization for it to become insoluble in the aqueous phase. It then
collapses upon itself, thereby providing within itself a hydrophobic domain
into which monomer molecules can be absorbed. Polymerization then
continues in the same way as is postulated for the micellar mechanism.
Indeed, the collapsed propagating oligomer can be regarded as being
analogous to a colloid stabilizer micelle. The oligomer is to some extent
surface-active, because it incorporates both a hydrophilic moiety derived
from the water-soluble initiator and hydrophobic units derived from the
monomer molecules. According to the theory of homogeneous nucleation,
the function of any colloid stabilizer which may be present in the reaction
system is to stabilize colloidally the collapsed oligomers as they precipitate
from the aqueous phase. If micelles of colloid stabilizer are present, then
these are regarded as reservoirs of stabilizer molecules, dissociating as
required to provide stabilizer for new and growing particles. To this extent,
the presence and concentration of micelles has been said by some proponents
of this theory to be of trivial importance.
An early quantitative theory of homogeneous nucleation is that of Roe
[32]. According to this theory, the critical event which determines the end
of Interval I is not the disappearance of micellar colloid stabilizer but the
depletion ofthe concentration of non-micellar colloid stablizer in the aqueous
phase to a level below which new collapsing oligomers cannot achieve colloid
stability by adsorption of adequate stabilizer. Although collapsed oligomers
continue to precipitate from the aqueous phase, they are now so colloidally
unstable that they coalesce with. the particles already formed. Thus the particle
number remains constant, notwithstanding that new latent particles continue
to form. Apart from the re-interpretation of t*, the duration of Interval I, as
the time required for the concentration of non-micellar colloid stabilizer to
fall to the critical value, the mathematics of the Roe theory is identical to
that of the Smith~Ewart theory for micellar nucleation. Again it is predicted
that the concentration of particles nucleated is directly proportional to the
%th power of the overall concentration of colloid stabilizer (micellar and
non-micellar) and the %th power of the initiator concentration, the latter
prediction assuming that the initiator-decomposition reaction is kinetically
of first order in initiator concentration. Clearly, it is impossible to discriminate
between the two theories on the basis of experimental results for the
dependence of particle concentration upon the concentrations of colloid
stabilizer and initiator.
The theory of homogeneous nucleation in emulsion polymerization reactions
has been considerably developed in recent years by Fitch and his coworkers
[33~35], and by others. A convenient review is available in a paper by Fitch
published in 1973 [36]. It is now recognized that at least two types of process
may occur which can lead to the precipitation of oligomers propagating in
the aqueous phase, and thus to the creation of new potential reaction
loci:
1. The first type of process is that which has been described above. The
oligomers propagate until they attain a critical chain length beyond which
they are insoluble in the aqueous phase. At this critical chain length, the
various enthalpic interactions between the hydrophobic monomer units
and the water molecules are such as to outweigh the entropic tendency
for polymer units to remain mixed with water molecules. There is now
an overall thermodynamic tendency for the polymer units to de-mix
from the water molecules. The oligomers collapse upon themselves and
become effectively a separate phase. This type of process provides a
potential reaction locus which has within it an active propagating free
radical. Polymerization can therefore continue if monomer molecules are
available.
2. In the second type of process, a propagating oligomer mutually terminates
a second propagating oligomer by combination, thereby attaining a degree
of polymerization which exceeds the critical value for precipitation. This
type of process provides a potential reaction locus which does not have
within it an active propagating free radical. The continuance of polymerization
therefore requires the acquisition of a further free radical, as well as of
monomer molecules.
In addition to these processes, a third type can occur in which the propagating
oligomer is captured by an existing particle. A matter of some controversy
has been the mechanism of capture; the possibilities are capture by diffusion
and capture by collision. Notwithstanding this uncertainty, the main features
of the theory are clear. Initially the reaction system contains no particles; all
the propagating oligomers produce potential reaction loci by processes such
as have been outlined above. These latent particles grow by absorption of
monomer and polymerization of that monomer. As the concentration of
particles increases, so the probability increases that two further types of event
will occur. These are:
1. that a propagating oligomer will be captured by an existing particle before

it can attain by one means or another the critical degree of polymerization
for precipitation;
2. that flocculation will occur between pairs of particles.
Eventually a stage is reached at which the number of reaction loci remains

constant because the rate of formation of new oligomers is equal to the rate
of loss of actual and potential reaction loci by oligomer capture and particle
flocculation. According to this theory, the presence of colloid stabilizer is
incidental; its function is to alter the probabilities of the various outcomes
of oligomer initiation in the homogeneous aqueous phase.
10.2.2.2.3 Other proposed mechanisms fo,.. particle nucleation

A so-called coagulative nucleation mechanism has been proposed by Feeney,
Napper and Gilbert [37]. This theory was developed to explain observations
relating to the particle-size distribution during Interval I. Expressed as the
distribution of particle number with respect to particle volume, this distribution
has been found to be positively skewed. This is difficult to explain on the
basis of a single-stage nucleation process; such processes predict that the
distribution should be negatively skewed. Positive skew implies that the rate
of particle nucleation increases with increasing time. The hypothesis underlying
the coagulative nucleation mechanism is that 'mature' primary latex particles
are formed from 'precursor' latex particles by coagulation. The precursor
particles are probably formed by homogeneous nucleation. They are thought
to differ from the mature particles in at least two important respects: they
are colloidally less stable than the mature particle; and monomer within
them polymerizes more slowly than it does within the mature particles. Both
these differences are thought to be attributable to the smaller size, and hence
greater surface curvature, ofthe precursor particles. The rate of polymerization
is reduced partly because the monomer concentration is reduced as a
consequence of the high surface curvature reducing the equilibrium swelling
of the polymer by monomer (see section 5.2.4.2 of Chapter 5 (Volume 1»,
and partly because the smaller the particles, the more easily can free radicals
exit from the particles.
Nucleation by hydrophobic association between the growing oligomers in
the aqueous phase and individual colloid stabilizer molecules (i.e., stabilizer
in non-micellar form) has been proposed by Chen and Piirma [38, 39]. The
idea is that, when the growing oligomer has attained a certain critical degree
of polymerization, it can form what is effectively a 'mixed micelle' by
association with dissolved individual colloid stabilizer molecules. The two
molecular species are envisaged as being held together by association between
the hydrophobic moiety of the colloid stabilizer molecules and the monomer
units in the oligomer. The mixed micelle then provides a locus for further
polymerization. The locus will already contain a propagating free radical,
and so polymerization can continue as soon as a few monomer molecules
have been absorbed into the locus.
Finally, brief mention must be made of the possibility of particle nucleation
by the entry of free radicals into emulsified monomer droplets. It must be
emphasized that, even if the particles in an emulsion polymerization reaction
system have been nucleated in this way, the system still differs conceptually
from a conventional suspension emulsion. In the latter, the free radicals are
generated within the emulsified monomer droplets, whereas particle nucleation
within monomer droplets occurs by the entry of a radical from a contiguous
external phase. For many years, the monomer droplets were discounted as
a locus of significant polymerization in emulsion reactions, for the reasons
which have been set out in section 10.2.1.2.6 above. Those reasons remain
valid. Initiation of polymerization in monomer droplets is oflittle importance,

provided that the monomer has not been emulsified into very small droplets.
If the monomer has been so emulsified, then there will be a very large number
of monomer droplets, and most of the colloid stabilizer in the system will
be adsorbed at the monomer-aqueous phase interface, so that little, if any,
remains in the aqueous phase to facilitate nucleation in that phase. One
consequence of the large number of particles, and of the large surface area
available for adsorption of colloid stabilizer molecules, is that the monomer
droplets may provide an aggregate surface area which is similar in magnitude
to, or perhaps greater than, the aggregate surface area of any colloid stabilizer
micelles which remain in the system. Nucleation by entry of radicals into
emulsified monomer droplets has been found to be important when the
monomer has been emulsified as very fine droplets using a mixed emulsifier
system such as an anionic surfactant and a long-chain fatty alcohol. For
further information concerning nucleation in such systems, the reader is
referred to papers by Ugelstad, EI-Aasser and Vanderhoff [40], and by
Ugelstad, Hansen and Lange [41]. In recent years some interest has been
shown in latices nucleated by this technique which contain very small particles
(typically ca. 20 nm in diameter). These polymer colloids are sometimes known
as microlatices.
10.2.2.2.4 Overall conclusions conerning mechanism of particle nucleation

The general view at the time of writing is that particle nucleation in reaction
systems which contain relatively water-soluble monQllners occurs mainly by
homogeneous nucleation, especially if the concentration of colloid stabilizer
in the aqueous phase is low or zero. Particle nucleation in systems which
contain monomers of low water-solubility and high concentrations of
micelle-forming colloid stabilizers is generally assumed to be mainly by
micellar mechanism, although there may still be some contribution from
homogeneous nucleation. The latter becomes increasingly important as the
concentration of micelle-forming colloid stabilizer is reduced. Even with
monomers of low water-solubility, particle nucleation is generally considered
to be by the homogeneous mechanism if the concentration of colloid stabilizer
is below its critical micelle concentration, or if added colloid stabilizer is
absent. The conditions under which nucleation by entry of radicals into
emulsified monomer droplets becomes significant are well understood; they
are not encountered in most emulsion polymerization reaction systems. The
importance or otherwise of nucleation by processes such as hydrophobic
association and the coagulative mechanism is unclear at the time of writing.
Morrison et al. [42] have recently provided a critical appraisal of various
particle-nucleation mechanisms in the light of the available experimental
evidence.
10.2.2.3 Mechanism and kinetics of polymerization during Interval II
10.2.2.3.1 Outline of mechanism of polymerization in Interval II

Polymerization during Interval II ideally takes place in a fixed number of
reaction loci, that number being the aggregate number which had been
nucleated by the time Interval I had ended. Colloid stabilizer micelles are
absent from the reaction system, even if they were present initially. Unreacted
monomer is present in three forms: as unpolymerized monomer in the reaction
loci; as monomer droplets; and as monomer molecules dissolved in the
aqueous phase. Most of the unreacted monomer is present in the first and
second of these forms; the proportion dissolved in the aqueous phase is small.
As monomer is polymerized in the reaction loci, the monomer/polymer ratio
is maintained at essentially its equilibrium value by absorption of monomer
molecules from the aqueous phase. To maintain the concentration of
monomer in the aqueous phase, monomer molecules dissolve from the
droplets. In this way, it is believed, the monomer droplets function as
reservoirs which supply monomer to maintain the monomer/polymer ratio
in the reaction loci at a constant, and essentially equilibrium, value. Transfer
of monomer from droplets to reaction loci as a consequence of collisions
between droplets and reaction loci is not thought to be significant. Free
radicals enter the reaction loci in a random fashion throughout Interval II.
In principle, four types of process involving the free radicals occur within
the reaction loci:
I. propagation reactions;
2. transfer reactions;
3. exit processes, in wl1fch radicals diffuse from a reaction locus back into
the aqueous phase;
4. bimolecular termination reactions between pairs of radicals.
Neither propagation reactions nor transfer reactions alter the number of
radicals present in a reaction locus, although the occurrence of transfer
reactions may affect the ease with which radicals are lost from reaction loci
by exit to the aqueous phase. Each exit process reduces the number ofradicals
in a reaction locus by 1. Each bimolecular termination process reduces the
number by 2. The overall effect of these processes in conjunction with random
acquisition of radicals from the aqueous phase is to create a system which
comprises a very large number of reaction loci which are such that the number
of propagating radicals, i, in each one at any instant can be 0 or can have
an integer value from I upwards. The system is dynamic rather than static,
in that the number of radicals present in anyone reaction locus will vary
from one instant to another as the various processes of radical acquisition
and loss occur. However, because the total number of reaction loci is very
large, a steady-state distribution of number of reaction loci with respect to
number of radicals is rapidly achieved. As will appear shortly, particular
significance attaches to the arithmetic mean of this distribution, i.e. to the

al'erage number of propagating radicals per reaction locus, denoted by l.
10.2.2.3.2 Kinetics of polymerization in Interval II

The fundamental equation which is commonly believed to govern the rate
of polymerization during Interval II is
dM
dt = Nikp[M] (10.8)
where N is the number of reaction loci per unit volume of aqueous phase,
[M] is the concentration of monomer in the reaction locus, kp is the rate
coefficient for the propagation reaction, and dM/dt is the rate of conversion
of monomer to polymer expressed in molecules (unit time)-l (unit volume
of aqueous phase) -1. The usual units for these quantities are mol dm - 3 for
[M], mol- 1 dm 3 s -1 for kp, and molecules s -1 (unit volume of aqueous
phase)-l for dM/dt. This equation is almost the only one of the very large
number which have been derived in the course of developing the theory of
emulsion polymerization about which there has been no controversy.
However, there has been considerable discussion as to the numerical values
which should be assigned to the quantities f, kp and [M] in particular
instances. Equation (10.8) can be derived from the equation for the rate of
propagation in a bulk free-radical polymerization as follows: The latter
equation is
d[M] = -k [M][R.] (10.9)

dt p
where [R·] denotes the total concentration of propagating radicals in the

system, in units of mol (unit volume) - 1. Consider, now, a single reaction
locus of volume v in an emulsion polymerization reaction system. The rate
of polymerization in molecules (unit time) - 1 in that single locus, denoted by
dm/dt, is
dm = -Lv d[M] = Lvk [M][R·] = ik [M] (10.10)

~ ~ p p
where L is the Avogadro constant. The latter step follows because L[R ·]v = i.
Thus for a volume of reaction system which contains N reaction loci, the
value of dM/dt will be the sum of all the individual values of dm/dt, i.e.,
(10.11)
where the summations extend over all N individual reaction loci. It is assumed
that kp and [M] have the same value in each reaction locus. The latter step
follows because I = 'E.i/ N.
Certain conclusions follow immediately from equation (10.8):
1. The rate of polymerization in Interval II is directly proportional to the

number of reaction loci in the reaction system.
2. The rate of polymerization in Interval II is independent of the total volume
of the reaction system in which those loci are dispersed. This explains why
the rate of polymerization in Interval II is independent of the volumes of
the monomer and aqueous phases, provided that the composition of the
latter remains constant.
3. Insofar as N, f, kp and [M] remain constant throughout Interval II, the
rate of polymerization also remains constant. There has been some
controversy concerning the constancy or otherwise of these quantities
during Interval II. The general view is that they do remain effectively
constant for most of the interval, and this explains why the rate of
polymerization is constant. In particular, the constancy of [M] during
Interval II, which is ensured by the presence of the monomer droplets,
explains why the polymerization reaction appears to be kinetically of zero
order in monomer concentration in the reaction system as a whole. The
reaction is, like all conventional free-radical propagation polymerization
reactions, in reality first order in monomer concentration. However,
because the monomer concentration is maintained at a constant value by
transfer from the monomer droplets through the aqueous phase, the rate
of polymerization is constant.
To obtain a prediction for the rate of polymerization in Interval II, or at
least a prediction of the effect of particular system variables upon that rate,
it is necessary to substitute the appropriate values for N and I in equation
(10.8). As regards N, if the value given by the Smith-Ewart theory [29] of
micellar nucleation or the Roe theory [32] of homogeneous nucleation is
used, and if the initiator-decomposition reaction is first order in initiator
concentration, then the resulting equation predicts that the rate of polymerization
in Interval II is jointly proportional to the %th power of the concentration
of colloid stabilizer and the %th power of the initiator concentration, both
concentrations referring to the aqueous phase of the reaction system. As
regards these predictions, it should be noted that:
1. because the quantities p and n in the expression for N refer to Interval
I of the reaction, and not to any subsequent values which may differ from
those during Interval I, it follows that the rate of polymerization in Interval
II also depends upon the values of p and n in Interval I, provided that
the conditions during Interval II are such that no new reaction loci are
nucleated;
2. because the exponents of p and n in the expression for N arise as a
consequence of the relationship between the surface area and the volume
of geometrically-similar particles, and not as a consequence of the kinetics
of the polymerization reaction itself, the same is true for the rate of
polymerization in Interval II. It is well-known that the rate of bulk, solution
and suspension polymerization is of order 'l'2 in initiator concentration,

if the initiator decomposition is first order in initiator concentration and
termination is bimolecular. Because order % with respect to initiator
concentration is very similar to, and difficult to distinguish experimentally
from, order 'l'2' it might be supposed that the dependence of the Interval
II rate upon initiator concentration is a consequence of bimolecular
termination of the propagating chains. This is not so. To the extent that
the Interval II rate is approximately proportional to the square root of
the initiator concentration, this is believed to be a consequence of the
effect upon the number of reaction loci nucleated in Interval I, broadly
in accordance with the theories of Smith and Ewart and of Roe.
It remains to discuss the value which is to be assigned to l, the average
number of propagating radicals per reaction locus. The value of l, and also
the distribution of numbers of locus populations with respect to i, depends
upon the balance between that rate at which radicals enter reaction loci from
the aqueous phase and the rates at which radicals are lost from reaction loci
by exit to the aqueous phase, and by termination. In the second part of their
paper, Smith and Ewart [29] discuss quantitatively the factors which affect
the value of i. They distinguish three so-called 'cases', of which Case 2 has
been of the greatest interest because it has been regarded as the 'ideal' case
to which real emulsion polymerization systems approximate to a lesser or
greater degree. The conditions which must be fulfilled for Case 2 are as follows:
1. The radicals must enter reaction loci from the aqueous phase at random.
2. Polymerization must commence immediately a radical becomes present
in a reaction locus.
3. Bimolecular termination of radicals must occur instantaneously whenever
two radicals become present in a particular reaction locus.
4. There must be no mechanism, such as exit of radicals from the reaction
loci, whereby the activity of the free radicals can be transferred from the
reaction loci back to the aqueous phase.
If these conditions are fulfilled, then the variation with time of the rate of
polymerization in a given reaction locus is as shown schematically in Fig.
lO.2(aXi). The intervals between the entry of successive radicals into the
reaction locus are random. Polymerization commences immediately the first
radical enters. It continues at a constant rate until the second radical enters,
whereupon it ceases immediately because bimolecular termination occurs
instantaneously upon the entry of the second radical. The reaction locus is
now quiescent until the third radical enters, whereupon polymerization
immediately commences at the same rate as previously. It continues until
the entry of the fourth radical, when it ceases again. This pattern continues
until the reaction system changes in such a way that one or more of the
above conditions is invalidated. As long as those conditions are fulfilled,
(i)
(i)
r-- - I"""" r-- -
•
2 3 4 5 6 7 8 9 10
Entry of successive radicals into reaction locus
2 45 6 7
Entry of successive radicals into reaction loci
Time
1.0
°0~----~----~2------~3---
(b) Number of radicals in reaction locus
Figure 10.2 (a) Variation with time of: (i) rate of polymerization at a given reaction locus; (ii)
extent of polymerization at that reaction locus. for reaction loci which confonn to Case 2.
(b) Distribution of relative locus populations between the two states of radical occupancy for
reaction loci which confonn to Case 2. (c) Effect of slight retardation of tennination relative to
radical entry upon variation with time of: (i) rate of polymerization at a given reaction locus;
(ii) extent of polymerization at that reaction locus.
polymerization commences immediately at the entry of every odd-numbered

radical, and continues at a constant rate until the entry of the next
even-numbered radical, when it ceases immediately. The variation of extent
of polymerization with time for a single reaction locus consists of a series of
ramps and levels, as shown schematically in Fig. lO.2(a)(ii).
Several conclusions follow as regards polymerization during Interval II

for a reaction system which conforms to Case 2:
1. Only two states of radical occupancy are possible for the reaction loci.
i = 0 for one, and i = 1 for the other. All reaction loci in the system are
either devoid of propagating radicals or contain at most one.
2. Although the entry of successive radicals into a given reaction locus is
random, and so the intervals between the entries of successive radicals
vary randomly, in the long run any given reaction locus is active for half
the time and quiescent for the other half of the time.
3. Thus at anyone instant, almost exactly half the reaction loci in the system
are active and the other half quiescent. This is true at all instants
throughout Interval II, but it is a dynamically-maintained constancy, not
a static constancy, in that anyone particular reaction locus is continually
interchanging between the two states of radical occupancy. The distribution
of relative locus populations between the two states of radical occupancy
has the very simple form shown in Fig. 1O.2(b).
4. The average value of i is '12' and thus equation (10.8) becomes
(10.12)
5. There is complete suppression of the type of bimolecular termination

reaction which occurs in bulk, solution and suspension free-radical
addition polymerization reactions, because two propagating radicals can
never be present in the same reaction locus.
Deviations from Case 2 can be expected if the exit of radicals from reaction
loci to the aqueous phase becomes possible, and if termination by an entering
radical is retarded so that two or more propagating radicals are able to
coexist within a single reaction locus for a non-zero time. Facilitation of
radical exit will reduce the value of I; retardation of termination will increase
it; the actual value of 1 will be determined by the balance of these factors. In
any event, the value of 1 is likely to be of order of magnitude 1 or less, so
that the propagation step of the polymerization is always effectively
compartmentalized into a large number of separate reaction loci, and
termination is retarded relative to propagation. Thus the explanation which
has been given above for the unusual features of emulsion polymerization
compared with the other modes of free-radical addition polymerization
remains valid, even if the reaction deviates significantly from Case 2.
The effects of retarded termination upon the variation with time of the
value of i and the rate of polymerization for a single reaction locus are shown
in Figs 1O.2(c)(i) and (ii) respectively. It is assumed that termination is only
slightly retarded relative to radical entry, so that the number of propagating
radicals in the locus seldom, if ever, exceeds 2. Three states of radical
occupancy are now possible for each reaction loci, namely, those for which
i = 0, 1 and 2. The distribution of relative locus populations between the
three states of radical occupancy (not shown) is slightly more complex than
that shown in Fig. 1O.2(b). The principal factor which might be expected to
retard bimolecular termination in Interval II is the increasing size of the
reaction locus as polymerization proceeds. As locus size increases, so the
second entering radical has increasing difficulty finding the radical which is
already present; or, what amounts to the same thing, the radical concentration
decreases with increasing locus size. Furthermore, as locus size increases, the
exit of a radical to the aqueous phase is retarded for the same reason. Thus,
as locus size increases, so the effect of an exit mechanism in reducing I below
the 'ideal' value of Y2 is mitigated to some extent.
Any variation of I during Interval II will have a proportionate effect upon
the rate of polymerization, provided that N, kp and [M] remain constant.
In view of the various possibilities noted in the two preceding paragraphs,
some surprise has been expressed in the past that the rate of polymerization
does remain effectively constant so long as some unreacted monomer is
present as a separate droplet phase. That this is so in many cases is presumably
a consequence of the very small size of the reaction loci.
Although the other cases, Case 1 and Case 3, delineated by Smith and
Ewart [29] do not concern us, it is appropriate to note their distinguishing
characteristics. Case 1 corresponds to a reaction system which is such that
the rate at which radicals exit from the reaction loci back into the aqueous
phase is great compared with the rate at which radicals enter reaction loci
from the aqueous phase. Under these conditions, the average number of
propagating radicals per locus is small compared with I. Case 3 corresponds
to a reaction system in which radicals enter the loci from the aqueous phase
much more rapidly than they are lost from loci, either by bimolecular
termination or by exit back to the aqueous phase. Under these conditions,
the average number of propagating radicals per locus is large compared with
I. The distinction between near-Case 2 reaction systems, of the type
considered in the preceding paragraphs, and reaction systems which conform
to either Case 1 or Case 3 is one of degree rather than of kind, and is arbitrary.
Near-Case 2 systems are those for which I is approximately Y2; Case 1 and
Case 3 systems are those for which I is respectively much less than and much
greater than Y2'
10.2.2.3.3 Mechanism of entry of radicals into reaction loci

Much of the experimental work on aqueous emulsion polymerization
reactions has been carried out using very hydrophobic monomers, notably
styrene, peroxydisulphates as initiators, and anionic micelle-forming colloid
stabilizers such as sodium n-dodecyl sulphate. It used to be supposed that
the radicals which enter the reaction loci at random during Interval II are
sulphate radical-ions. This hypothesis presents two serious difficulties. Firstly,
it is difficult to understand why a hydrophilic radical such as the sulphate

radical-ion should leave the aqueous environment and enter the hydrophobic
environment of the reaction loci. Secondly, there is an energy barrier to such
entry. The reaction loci are colloidally stabilized by adsorbed anions. The
approaching initiator anions will therefore be repelled from the surface of
the reaction loci. In order to enter the loci, they would have to penetrate the
layer of adsorbed anions. One suggestion which overcomes the second of
these objections is that it is not anionic radicals such as sulphate radical-ions
which enter the reaction loci, but electrically-neutral hydroxide radicals
formed by interaction between the radical-anions and water molecules (see
section 10.2.1.1 above). However, the entering radical is still hydrophilic. The
generally-accepted view today is that the hydr<?philic radicals (whether ionic
or not) which form in the aqueous phase react with monomer molecules
dissolved in the aqueous phase, thereby initiating solution polymerization
in the manner hypothesized by the theory of homogeneous nucleation. These
oligomeric radicals are surface-active, and carry the free radical in the
hydrophobic moiety of the molecule. As the molecule grows in chain length,
there is an increasing tendency for it to adsorb at the surface of a reaction
locus. Furthermore, it becomes increasingly possible for the radical end of
the molecule to penetrate the interior of the reaction locus without necessitating
simultaneous penetration by the hydrophilic, possibly ionized, end.
10.2.2.3.4 Mechanism of exit of radicals from reaction loci

Propagating radicals are expected to exit from reaction loci to the aqueous
phase occasionally if there is any tendency for them to partition between the
reaction loci and the aqueous phase. The two principal factors which are
expected to reduce this tendency are high intrinsic hydrophobicity and large
molecular size; both these factors will tend to reduce the solubility of a
propagating radical in an aqueous medium. Hence there are two principal
factors which are thought to facilitate exit of radicals from reaction loci, and
thus invalidate one of the requirements for Case 2. The first of these is
increasing hydrophilicity of the monomer, and hence of oligomers derived
from the monomer. The second is the tendency for propagating radicals to
transfer to monomer molecules, thereby reducng the average molecular size
of the propagating radical. In consequence, a rough index of the tendency
of radicals to exit from reaction loci to the aqueous phase is provided by
the product of the solubility of the monomer in water and a parameter which
quantifies the tendency for transfer-to-monomer to occur relative to propagation.
The latter parameter is usually taken as the ratio km/kp, where km is the rate
coefficient for the transfer-to-monomer reaction. Monomers which are
relatively hydrophilic, and hence relatively water-soluble, and which are also
prone to transfer-to-monomer reactions as they polymerize by free-radical
mechanism, tend to give emulsion polymerization reactions which conform
to Case 1 rather than to Case 2.
10.2.2.4 Kinetics of polymerization during Interval III

The transition between Intervals II and III occurs when the monomer droplets
disappear from the reaction system, all the unreacted monomer now having
become absorbed within the reaction loci. The point in the reaction at which
this occurs is determined by those factors which influence the extent to which
polymer particles absorb a lyophobic liquid when an excess of the latter is
available as a separate phase. These factors have been considered in section
5.2.4.2 of Chapter 5 (Volume 1). In many emulsion polymerization reactions,
the transition from Interval II to Interval III occurs when approximately
half the monomer has been converted to polymer. This implies that the
monomer/polymer ratio in the reaction loci at the transition point is ca.
1/1 m/m. After the transition, the polymerization continues in the loci, but
replenishment of reacted monomer from the droplets is not now possible.
Consequently, the monomer concentration in the reaction loci falls progressively.
Ideally, it is expected that the polymerization would now appear to be
kinetically of first order in monomer concentration in the reaction loci, and
therefore in the reaction system as a whole. Let Moo denote the number of
monomer molecules initially present in the reaction system per unit volume
of aqueous phase. Moo is the total number of monomer molecules available
for polymerization in the system per unit volume of aqueous phase. Let M
be the number of monomer molecules which have polymerized after time
t, > to, where to is the time at which Interval III commenced. Insofar as N,
Tand kp remain constant during Interval III, and the monomer disappearance
is first order, dM/dt is directly proportional to (Moo - M). Denoting the
proportionality constant by tX, and integrating the resultant equation between
time to and t, the variation of extent of polymerization with time is obtained as
(10.13)
where M 0 is the number of monomer molecules per unit volume of aqueous

phase which have polymerized at time to, i.e., at the end of Interval II.
There are several factors which cause the simple analysis of the preceding
paragraph to be at best a very crude approximation in many systems. The
most important of these is that the polymer/monomer ratio in the reaction
loci increases as polymerization proceeds, and hence the viscosity of the
interior of the reaction loci also increases. This has several consequences for
the kinetics of the polymerization reaction. The more important of these are
as follows:
1. kp decreases because the diffusion of monomer molecules within the
reaction loci is hindered.
2. Bimolecular termination of radicals is retarded even more than propagation,
for the same reason.
3. Any tendency for radicals to exit to the aqueous phase is reduced.
4. The number of reaction loci may tend to decrease because flocculation occurs.
One overall effect of these consequences is that the value ofi tends to increase.
An additional complication is that complete conversion of monomer to
polymer may be prevented if the polymer/monomer mixture in the reaction
loci becomes glassy as polymerization proceeds and the polymer/monomer
ratio increases. This is known as the limiting conversion phenomenon. It is
likely to be encountered if the polymerization temperature is below the
glass-transition temperature of the pure polymer. Then, as long as sufficient
unreacted monomer is present to plasticize the polymer so that its glass-
transition temperature is below the polymerization temperature, the mixture
will contain sufficient free volume for the monomer molecules to diffuse to
polymerization sites. But when the mixture vitrifies as polymerization
proceeds, the monomer molecules are no longer able to diffuse to the
polymerization sites, and the rate of polymerization falls to a low value.
The overall effect of factors such as these upon the rate of polymerization
as the reaction proceeds through Interval III is difficult to predict. In some
cases, the rate of polymerization may increase somewhat as the transition
from Interval II to Interval III occurs, because the value of I increases.
However, the rate of polymerization always decreases as Interval III
progresses, because monomer depletion inevitably occurs as polymerization
proceeds. It should be noted that increase in the size of reaction loci during
Interval III due to polymerization (as opposed to flocculation) does not
occur, because of the absence of further monomer to be absorbed into the
loci. Such published experimental information as is available indicates that
the effects of reaction variables upon the limiting conversion phenomenon
are as expected. Raising the polymerization temperature for a reaction system
in which this phenomenon occurs increases the attainable extent of conversion,
because the system can now accommodate a higher glass-transition temperature
and therefore attain a higher polymer/monomer ratio. The attainable extent
of conversion can also be increased by incorporating in the reaction system
a comonomer which reduces the glass-transition temperature of the polymer.
A further effect of the increasing polymer/monomer ratio in the reaction
loci as Interval III proceeds is that there is an increasing tendency for
branching and crosslinking of the polymer to occur as the concentration of
polymer molecules in the reaction loci increases. The tendency for
transfer-to-polymer reactions (noted in section 10.2.1.2.1 above) to occur will
always increase with increasing polymer concentration, and the tendency will
further increase if the concentration of propagating radicals increases.
Transfer of radical activity to polymer molecules leads first to branching and
then possibly to crosslinking. A further possibility arises if one of the
monomers from which the polymer is forming contains more than one olefinic
double bond. Butadiene is the most common example of such a monomer.
The polymer chain which forms then contains olefinic double bonds. In
principle, propagation of another growing polymer chain can proceed
through such double bonds. Whenever this happens, a tetrafunctional branch
point is introduced into a macromolecule. The presence of such branch points

will soon cause a crosslinked network of indefinite size to form. The tendency
for propagation to proceed through olefinic double bonds in other polymer
chains will also increase as the polymer concentration increases. Furthermore,
the presence of olefinic unsaturation in the polymer chain may also facilitate
transfer-to-polymer reactions at adjacent sites through activation of hydrogen
or other atoms. These possibilities are further discussed in section 14.3.1 of
Chapter 14 in connection with the use of these reactions for the preparation
of graft copolymers by polymerization reactions in latices; schematic reactions
leading to the formation of graft copolymers are shown there. The overall
effect of transfer-to-polymer and propagation through un saturation in other
chains is that the polymer obtained at the end ofthe reaction is more complex
in structure than that which is produced by, say, solution polymerization.
In the case of polymers produced from monomers which contain more than
one olefinic double bond, the polymer produced if the polymerization is
taken to near completion usually contains much polymer gel, some of which
may be very tightly crosslinked. For this reason, it can be a disadvantage of
the emulsion mode of free-radical addition polymerization, compared with
some other ways of effecting the polymerization reaction, that much of the
polymer is produced under conditions of high polymer/monomer ratio.
10.2.3 Principal components of practical aqueous emulsion polymerization

reaction systems, other than monomers and water
10.2.3. J Initiators
10.2.3.1.1 Classification of initiator systems for emulsion polymerization

reactions
Initiators for emulsion polymerization reactions fall into two broad classes:
1. dissociative initiators, for which the primary radicals are generated by

thermal dissociation of the initiator molecule without interaction with any
other molecule;
2. redox initiators, which comprise at least two types of molecule, a reducing
agent and an oxidizing agent, and for which the primary radicals are
generated by redox interaction between these molecules.
As has been indicated above, the primary radicals generated by an initiation

system may undergo further reactions before their radical activity becomes
transferred to a polymerization locus to form a latent propagating chain.
10.2.3.1.2 Dissociative initiators

The peroxodisulphates, to which reference has been made in section to.2.1.1
above, are the best-known initiators for aqueous emulsion polymerization
reactions. They are normally believed to be of the dissociative type (but see
also section 10.2.3.3 below). Because these initiators rely upon thermal
dissociation for the production of radicals, and because the initiator must
be stable at normal ambient temperatures, it is inevitable that emulsion
polymerization reactions using these initiators require the temperature to be
raised, typically to 50-70°C. It is not generally possible to carry out emulsion
polymerizations at ambient or sub-ambient temperatures using dissociative
initiators. Although peroxodisulphate salts are by far the most widely-used
types of dissociative initiator for emulsion polymerization reactions, various
other types of compound can be used. These include water-soluble analogues
ofthe well-known initiator, azobisisobutyronitrile (Structure I). This is widely
used for initiating bulk, solution and suspension free-radical addition
polymerizations, and can also be used for initiating emulsion polymerizations.
Water-soluble analogues include 4,4'-azobis-4-cyanopentanoic acid (Structure
II) and its alkali-metal salts, and disodium 2,2'-azobis-2-cyanopropane-
I-sulphonate (Structure III). These compounds probably do always function
as true dissociative initiators. Other dissociative initiators which have been
used in emulsion polymerization reactions include aromatic diazoamino
compounds, aromatic diazothioethers, and alkali-metal aryl diazates.
I
CH3
I
CH 3 yH 3 yH 3
CH3-C-N=N-C-CH3 H02C.CH2CH2-C-N=N-C-CH2CH2·C02H
I I I I
CN CN CN CN
II
10.2.3.1.3 Redox initiators

Redox initiators have the advantage of being able to initiate polymerization
at much lower temperatures than dissociative initiators. The components of
a redox initiation system are unreactive at normal ambient temperatures,
provided that they are kept separate from each other. The activity of a redox
system depends upon the ability of the components to react to form free
radicals when they are brought together in such a way that they can interact.
Redox initiators are commonly used for the initiation of emulsion polymerization
reactions at near-ambient temperatures, and at sub-ambient temperatures
such as 5°C.
The best-known redox initiators are based upon combinations of iron(II)

salts and peroxides. The initial reaction between iron (II) ions and hydrogen
peroxide is conventionally represented as follows:
H 20 2 + Fe2+ __ HO· + HO- + Fe3+
It is generally believed that this combination initiates free-radical polymerization
by way of the hydroxide radicals which are formed in this reaction. However,
initially, hydroxide radicals are generated too rapidly; only a small proportion
of these radicals initiates polymerization. Subsequently the rate of generation
decreases considerably as the components are consumed. It is therefore redox
systems comprising oil-soluble organic hydroperoxides and sparingly-soluble
iron(II) salts or complexes which have been most commonly used. In this
way, the concentrations of reactants in the aqueous phase are limited, and
the generation of radicals proceeds at a uniform rate. Predominantly alkoxy
radicals are produced, the reaction being as follows:
ROOH + Fe2+ __ RO· + HO- + Fe3+
where R is an organic moiety. It is these radicals which are commonly
supposed to bring about initiation of polymerization. The iron-hydroperoxide
initiation system is capable of great diversification because of the wide range
of organic hydroperoxides which is available, the possibility of controlling
the availablity of iron(II) ions by using pre-formed complexes or else very
insoluble iron(II) salts, and the possibility of adding other components to
the system, such as reducing sugars. The hydroperoxide usually partitions
mainly in the non-aqueous phase, whereas the iron compound is confined
to the aqueous phase. This feature provides additional scope for control over
the rate at which initiating radicals are produced, through limitation of the
rate at which two components of the redox couple are brought together.
The iron(II) component is conveniently added to the system in the form
of a complex with either pyrophosphate or ethylenediaminetetraacetic acid
anions. It may also be added as a sparingly-soluble iron(II) compound such
as the sulphide. Reducing sugars, such as glucose or dextrose, may be added
to convert the iron(III) which is formed back to iron(II). Such sugars should
not react directly with the hydroperoxide. The interactions between the
various components of the hydroperoxide-iron(II)-sugar-monomer com-
bination can be conveniently summarized as shown in Fig. 10.3.
Redox systems based upon iron(II)-hydroperoxide combinations have
been widely used for initiating the emulsion copolymerization of styrene and
butadiene at low temperatures. The effectiveness of these combinations in
initiating this emulsion copolymerization has been found to depend upon
the way in which the suspension of iron(II) complex is formed. A typical
procedure involves heating a mixture of dextrose and potassium pyrophosphate
for a short time, allowing it to cool partially, adding the requisite amount
of iron(II) sulphate solution, agitating, ageing, and finally cooling. The
Fe 2+ from complex or
sparingly-soluble salt
Oxidized form +
Fe2+ ROOH ROM·
of sugar
Reduced form
of sugar RO· M
+
HO-
Figure 10.3 Interactions between various components of the hydroperoxide-iron(II)-sugar-

monomer initiation system.
temperatures and times for the heating, ageing and cooling are critical for
the effectiveness of the initiator under any given conditions, as also is the
extent of the turbulence during mixing.
Of the many suitable oil-soluble hydroperoxides which are avail-
able, important examples include tert-butyl hydroperoxide (Structure IV),
p-methane hydroperoxide (Structure V), p-methylisopropylbenzene
(p-cymene) hydroperoxide (Structure VI), and isopropylbenzene (cumene)
hydroperoxide (Structure VII). The order of decreasing reactivity is
IV > V > VI > VII.
CH,Oc(CH,),.OOH
VII
Comparative investigations of the effectiveness of various hydroperoxides

as the oxidants in iron(II) redox initiation systems for the emulsion
copolymerization of styrene and butadiene have been reported by Wicklatz,
Kennedy and Reynolds [43] and by Fryling and Follett [44]. No clear-cut
correlation between chemical structure and reactivity emerges from either of
these investigations, although it is evident that the effectiveness of a given
hydroperoxide in low-temperature redox systems depends very much upon
its structure. An important factor may be the solubility of the hydroperoxide
in the aqueous phase of the reaction system. However, contrary to expectation,
it appears that the less soluble the hydroperoxide, the more effective it is as
an initiator. Wicklatz, Kennedy and Reynolds investigated the performance
of a number of ring-substituted isopropyl benzene hydroperoxides in an
iron(I1) sulphate-potassium pyrophosphate-dextrose combination when
used to initiate the emulsion copolymerization of styrene and butadiene at
5°C. They found the order of decreasing reactivity to be:
chi oro > tert-butyl > isopropyl > sec-dodecyl > methyl
> sec-amyl > hydrogen
The position of the substituent in the ring was found to have little effect
upon the reactivity of the hydroperoxide. As might be expected, the optimum
ratio by mass of hydroperoxide to iron(I1) salt was found to increase as the
molecular mass of the hydroperoxide increased. The optimum ratio is ca.
1/1 mol/mol.
Iron-free hydroperoxide redox initiators for effecting emulsion polymerization
at low temperatures have been of some interest for producing latices and
polymers uncontaminated with iron. Such polymers should be less prone to
staining, discoloration and oxidative degradation than those which contain
traces of iron. Of the iron-free hydroperoxidic redox initiators, combinations
of hydroperoxides and polyamines are probably the best known. The
interaction between the two components can be represented as follows:
ROOH + R'NH2 ---+ RO' + R'NH + H 20
To be effective, the polyamine should contain at least two primary amine
groups and one or more secondary amine groups. A typical combination
comprises cumene hydroperoxide and diethylenetriamine (H 2NCH 2CH 2
NHCH 2CH 2NH 2). There is, however, some doubt as to whether in practice
hydroperoxide-polyamine systems are truly iron-free; it has been reported
that rigorous purification of the components from all traces of iron results
in a reduction ofthe effectiveness of the combination. It may be that in practice it
is minute traces of an iron-polyamine complex which interact with the
hydroperoxide. If so, this could explain why these combinations are rather
sensitive to changes of pH; the cause could then be variable hydrolysis of
the iron-polyamine complex.
Other well-known redox combinations which are used for the initiation
of emulsion polymerization reactions include the peroxodisulphate- bisulphite
system and the chlorate- bisulphite system. The former combination generates
free radicals by a reaction such as
S20~- + HSO; ---+ SO;- + ·OSO; + 'OS02H
If polymerization is initiated by both the bisulphite radicals and sulphate
radical-ions, then polymers will be formed which contain both sulphonate
and sulphate end-groups. The reactions by which the chlorate-bisulphite
combination initiates emulsion polymerization appear to be complex.
10.2.3.1.4 Effect ofadventitious impurities upon emulsion polymerization initiation

systems
A matter of some importance in practice is the extent to which the various
initiation systems are subject to retardation or inhibition by oxygen or other
adventitious impurities. Oxygen is a powerful inhibitor of many emulsion
polymerization reactions, notably that of styrene and butadiene. Low-
temperature redox systems, in particular, those which contain no reducing
sugars, are generally more sensitive than are those initiated with peroxodisul-
phates. If oxygen is not effectively eliminated from such systems, then long
and variable induction periods are experienced, followed by slow and variable
rates of polymerization.
The addition of small amounts of metal-ion sequestering agents to the
aqueous phase of the reaction system is sometimes found to increase the rate
of polymerization. The effect can be such as to permit a reduction in the
amount of initiator used. In any case, process and quality control are rendered
easier because day-to-day variations in rate of polymerization are reduced,
and the product is more uniform. It seems that traces of metal ions, such as
calcium and magnesium, which may always be present in water however
carefully it is purified on an industrial scale, are able to retard emulsion
polymerization.
10.2.3.2 Colloid stabilizers

In the context of emulsion polymerization reactions, colloid stabilizers are
also known as emulsifiers. This name is misleading, because it implies that
their primary function is to facilitate emulsification of the monomer droplets
in the aqueous phase. They may indeed facilitate such emulsification, but
this is an incidental function in most emulsion polymerization reaction
systems. The primary functions of these substances are two-fold: to facilitate
particle nucleation, and to stabilize the latex particles against flocculation as
the polymerization proceeds. If the colloid stabilizers are capable of forming
micelles, and if the concentration is such that micelles are present, then, as
stated in section 10.2.2.2.4 above, it is generally believed that particle
nucleation occurs principally by the micellar mechanism, especially if the
monomer is of low water-solubility. Unless steps are taken to so post-treat
the latex obtained from an emulsion polymerization reaction that the initial
colloid stabilizer is displaced or swamped by other stabilizers, many of the
colloidal properties of the final latex will be determined by the stabilizer
which was used to facilitate particle nucleation. In particular, the polarity of
any bound electric charge carried by the latex particles will be determined
by that colloid stabilizer.
A wide range of substances can be used as colloid stabilizers for emulsion
polymerization reactions. The best-known are surface-active carboxylates,
sulphates and sulphonates. The sodium and potassium soaps of the saturated
long-chain n-alkanoic acids have been found to be particularly suitable for
the emulsion copolymerization of styrene and butadiene. So also have oleates.
Pure potassium palmitate and potassium and sodium oleates have been found
to result in substantially the same extent of polymerization after 12 hours
reaction at 50·C when used in a standard reaction system for styrene-butadiene
emulsion copolymerization. For reaction systems containing n-alkanoates of
lesser alkyl chain length than n-hexadecanoate, the extent of polymerization
has been found to decrease progressively as the chain length is reduced. Soaps
derived from fatty acids, such as linoleic (heptadeca-8, II-diene-l-carboxylic)
acid, which contain multiple olefinic unsaturation, exert a retarding effect
upon the polymerization. They should therefore be avoided. Naturally-occurring
fatty substances should be partially hydrogenated prior to saponification to
convert acids such as linoleic and linolenic (heptadeca-8, 11, 14-triene-
I-carboxylic) to less unsaturated analogues.
Considerations of cost and availability have led to other surface-active
carboxylates besides those derived from fatty acids being investigated for the
emulsion copolymerization of styrene and butadiene. In particular, the soaps
derived from dis proportionated rosin acids have been found to be excellent
for this purpose, especially for low-temperature polymerizations. They may
be used either alone, or in combination with fatty-acid soaps, e.g., as tall-oil soap.
For detailed investigations of the comparative behaviour of various
fatty-acid and rosin-acid soaps in the peroxodisulphate-initiated emulsion
copolymerizations of styrene and butadiene, the reader is referred to papers
by Carr et al. [45,46]. Information concerning the effects of level of soap
addition upon the rate of polymerization is available in a paper by Kolthoff,
Meehan and Carr [47].
Many synthetic latices are produced using either surface-active sulphates
or surface-active suI phonates as colloid stabilizer. Sodium n-dodecyl (lauryl)
sulphate is the best-known surface-active sulphate used in emulsion
polymerization reactions. The sodium alkylbenzenesulphonates, especially
sodium n-dodecylbenzenesulphonates, are the best-known surface-active
sulphonates for this purpose. A major advantage of these colloid stabilizers
in relation to the carboxylates is that they can be used to prepare anionic
latices by emulsion polymerization under acidic conditions. This advantage
is of crucial importance for the preparation of various types of functionalized
synthetic latices (see further section 10.3.3 below). Sulphonates have an
advantage over sulphates in this application in that they are less susceptible
to hydrolysis under acidic conditions. The hydrolysis reaction (given in section
10.3.3.2 below) separates the polar hydrophilic moiety from the hydrophobic
moiety by cleavage of the C-O-S bond in the sulphates, thereby destroying
the capability to function as an efficient colloid stabilizer. The equivalent
bond in the sui phonates is C-S; this is not so susceptible to acid hydrolysis.
The results of an extensive investigation of the potentialities of many
surface-active substances in styrene-butadiene emulsion copolymerizations

have been reported by Helin et al. [48]. The reaction system used in
this investigation comprised styrene-butadiene-water in the ratio 29171
180 m/m/m. Polymerization temperatures were 5°C and 50°C. A wide range
of surface-active substances at levels of ca. 5 pphm was compared with
fatty-acid and rosin-acid soaps as controls. The conclusion was reached that,
as a class, the anionic sulphates and sulphonates generally give good
polymerization rates, and that the latices obtained have adequate colloid
stability. Furthermore, the results confirm that, as expected, the rates of
polymerization are essentially independent of the pH of the aqueous phase.
Cationic surface-active substances of the quaternary-ammonium-salt and
amine-salt types were found to give poor polymerization rates at 5°C, but
high rates at 50°C. The colloid stability ofthe latices produced was inadequate.
As a class, surface-active ethoxylates (mostly non-ionogenic under the
conditions used) were found to produce latices of rather large particle size,
and to be unsatisfactory because polymerization rates were low and the
colloid stability of the latices produced was variable. Of this type of
surface-active substance, only the amide ethoxylates were found to approximate
in effectiveness to the fatty-acid soap control in respect of both polymerization
rate and the colloid stability of the latex produced. Even so, the performance
of such adducts was observed to be rather sensitive to pH. Ethoxylates formed
from alcohols, fatty acids and amines were found to give only moderate
polymerization rates, and the latices produced varied considerably in colloid
stability. It is suggested by these workers, however, that the excellent stability
which these ethoxylates display towards chemical and physical deteriorative
influences should commend their use in applications where resistance to these
influences is important, notably in latices which are intended for use as bases
for surface-coatings formulations, and in latices intended for incorporation
in paper-coating compositions. It is doubtful, however, if this characteristic
has been exploited to any extent in this way. As illustrative of the wide
variations in colloid stability which were observed in the latices obtained
using non-ionogenic surface-active substances, it may be noted that those
latices in which no micro-coagulum developed during polymerization were
so extremely stable colloidally as to require special procedures for their
subsequent coacervation. Helin et al. have observed interesting differences
between the behaviour of various surface-active substances when used for
styrene-butadiene emulsion copolymerizations at 5°C and when used for
polymerizations at 50°C. These differences may be associated with differences
in solubility at the two temperatures. Thus the effectiveness of the amine
ethoxylates was found to be very sensitive to pH at both temperatures, but
the nature of this sensitivity at the two temperatures differed. At SOC, they
gave low polymerization rates at low pH, where they are presumably cationic,
and moderately high rates of polymerization at high pH. At 50°C and high
pH, however, they gave higher polymerization rates than even the fatty-acid
soap controls, whereas at this temperature and low pH only moderate

polymerization rates were observed.
Various water-soluble hydrocolloids are used as colloid stabilizers in
emulsion polymerization reaction systems, either alone or in combination
with micelle-forming surface-active substances. The best-known of these
hydrocolloids is the substance conventionally known as 'polyvinyl alcohol'.
This is partially-hydrolysed polyvinyl acetate. It is effectively a copolymer
of vinyl alcohol and vinyl acetate units ofrather uncertain chemical structure.
It is widely used in reaction systems for the emulsion polymerization and
copolymerization of vinyl acetate. Other hydrocolloids used as colloid
stabilizers in emulsion polymerizations include water-soluble cellulose derivatives
and water-soluble acrylate polymers.
Other more specialized types of colloid stabilizer which have been of some
interest for use in emulsion polymerization reaction systems include synthetic
oligomeric anionic surface-active substances which give latices of high surface
free energy, and polymerizable surface-active substances. Colloid stabilizers
which give colloidally-stable latices of high surface free energy are of interest
industrially because of certain advantages which can accrue, such as low
tendency to foam. One class of synthetic oligomeric anionic surfactant which
can form latices of this type has been described by Roe [49,50]. These
oligomers are copolymers of acrylic acid and acrylonitrile terminated by an
alkyl group. They are produced by free-radical solution polymerization. The
molecular mass ofthe oligomers is kept low by carrying out the polymerization
in the presence of a mercaptan such as n-octyl mercaptan. In functioning as
a chain-transfer agent, the mercaptan also introduces a hydrophobic moiety
into most of the polymer molecules (see section 10.2.3.3 below). The
hydrophilic moiety is provided by the copolymer of acrylic acid and
acrylonitrile. The structure of the oligomers obtained using n-octyl mercaptan
as the chain-transfer agent is shown as Structure VIII, where the sequence
CsH 17 S+CH 2 CH.C0 2 H-)m(-CH 2 CH.CN)nH

VIII
of acrylic acid and acrylonitrile units is random. In principle, the properties

of these surfactants can be varied over a wide range by varying the
composition. Even if the mercaptan is fixed, it is possible to vary the values
of m and n, and hence the ratio of acrylic acid plus acrylonitrile to mercaptan,
and also the ratio of acrylic acid to acrylonitrile. Molecular masses vary from
a few hundred up to a few thousand daltons. The alkali-metal salts of these
oligomers are soluble in water. They are able to stabilize aqueous latices and
function as colloid stabilizers for emulsion polymerization reactions. The
unusual properties ofthese oligomers seems to arise from different dependencies
upon molecular size of adsorption at the two types of interface presented by
a polymer latex. The reader is referred to the original papers for detailed
discussion of these matters. It appears that the oligomers of lower molecular

mass are more strongly adsorbed at the surface of the particles than are those
of higher molecular mass, but are less effective as colloid stabilizers for the
particles. On balance, the oligomers of lower molecular mass appear to be
more effective as colloid stabilizers than are those of higher molecular mass,
if the comparison is made on a basis of equal masses. Roe also found that
the surface activity at the air-water interface decreases with increasing
oligomer size, and that the maximum adsorption at the air-water interface
is almost independent of oligomer size. Thus the particles tend to adsorb
preferentially those oligomer molecules which show most activity at the
air-water interface. To this characteristic is attributed the high surface free
energy and low foaming tendency of latices prepared using these oligomers.
Whatever may be the true explanation of this phenomenon, by using these
oligomers it is possible to prepare colloidally-stable latices having surface
free energies which are almost as high as that of pure water. The concentration
of surface-active molecules in the aqueous phases of such latices must therefore
be very low.
There is considerable interest in polymerizable colloid stabilizer for use in
emulsion polymerization reaction systems. Typical of such stabilizers are the
sodium salts of9- and lO-acrylamido-n-octadecanoic (stearic) acid (Structures
IXa and IXb). These have been the subject of three papers by Greene and
CH 3 (CH z)s-CH-{CH zh.CO; .Na + CH 3 (CH zh-CH-{CH z)s.CO; .Na +

I I
NH.CO.CH:CH z NH. CO. CH: CH z
IXa IXb
her co-workers [51-53], who have investigated the use of these compounds
as colloid stabilizers for the preparation of 60/40 m/m styrene-butadiene
copolymer latices. Comparisons were made with latices which contained
equivalent amounts of the monomeric surface-active substances added after
polymerization. Greene et at. found that colloid stabilizer which had been
added to the reaction system before polymerization, and which therefore had
the opportunity to polymerize in situ, conferred higher mechanical stability
upon the resultant latex than did stabilizer added after polymerization,
provided that the surface coverage exceeded 20%. The greater the surface
coverage, the more marked was the difference between the two types of latex.
Colloid stabilizer added after polymerization would have remained in the
latex as unpolymerized molecules, whereas some at least of that added before
polymerization was presumably copolymerized with the main monomers,
and was therefore covalently bound near the particle surface. It was also
observed that the stability of the latex towards added electrolyte was greater
at high surface coverage if the stabilizer was added before polymerization.
10.2.3.3 Modifiers
In the context of emulsion polymerization reactions, modifiers are very
efficient chain-transfer agents which serve to reduce the degree of polymerization
of the polymer which forms. In this context, these substances are also known
as regulators. Modifiers have been especially widely used in reaction systems
for the emulsion copolymerization of styrene and butadiene, because the
molecular mass and gel content of polymers produced without modifiers can
be too high for satisfactory processing of the polymers as solid rubbers, or
for satisfactory application of the polymers as latices. The best-known
modifiers are mercaptans. The reactions by which they effect reduction in
degree of polymerization can be represented as follows:
Mn· + RSH --+ MnH + RS·
RS·+M--+RSM·
As the amount of mercaptan in reaction systems for the emulsion
copolymerization of styrene and butadiene is increased, the polymers
produced become softer, the gel content decreases, the solid polymer becomes
more easily processable, and the particles of the latex integrate more easily
during film formation. The chemistry of mercaptans, with particular reference
to their behaviour in emulsion polymerization systems, has been discussed
in papers by Kharasch, Nudenberg and Mantell [54] and by Kharasch,
Nudenberg and Kawahara [55].
When first added to the reaction system, a mercaptan modifier is present
mainly in solution in the monomer-droplet phase. A small amount is also
dissolved in the aqueous phase, partly as un-ionized mercaptan, and, because
mercaptans are weak acids, partly as mercaptide anions formed by interaction
between the mercaptan molecules and hydroxide ions according to the
equilibrium
RSH + HO- --+ RS- + H 2 0
It seems likely that the transfer of mercaptan from the monomer phase to
the reaction loci occurs in part at least by way of mercaptide anions which
subsequently become re-converted to un-ionized mercaptan. It certainly
appears to be the case that, for mercaptans for which the rate of consumption
in the reaction is limited by diffusion through the aqueous phase, that rate
increases as the pH of the system is increased, the effect being especially
marked when the system is strongly alkaline (pH> 11).
Mercaptans present in emulsion polymerization systems can also participate
in reactions other than chain transfer during polymerization. Whether or not
such reactions occur depends upon the monomers in the system. Addition
across the olefinic double bond of the monomer can occur. If peroxides are
absent, the addition is in accordance with the Markownikov rule:
R'CH:CH 2 + RSH --+ R'CH(SR).CH 3
If peroxides are present, then addition occurs in the manner contrary to that
predicted by the Markownikov rule, i.e. the product is R/CH 2 • CH 2 SR. The
rate of addition of mercaptans to styrene has been found to increase in the order
tert-mercaptan < sec-mercaptan < primary mercaptan
A further reaction, an analogue of the well-known Michael reaction,
can occur between mercaptans and monomers if the latter contains one or
more electron-attracting groups adjacent to the olefinic double bond, and
the reaction system is alkaline. It results in addition of the mercaptan across
the olefinic double bond. As a consequence, the monomer becomes unavailable
for polymerization, and the mercaptan unavailable for transfer. An important
example of a monomer which undergoes this reaction is acrylonitrile, for
which the reaction mechanism is probably as follows:
ASH + HO-
In peroxodisulphate-initiated styrene-butadiene emulsion copolymerizations,

mercaptans also function as initiator promoters. Whereas the emulsion
polymerization of styrene can be initiated by peroxodisulphates alone, this
is not so for the emulsion copolymerization with butadiene. The presence of
a small quantity of a mercaptan, normally added primarily as a modifier, is
necessary if the polymerization is to proceed at a useful rate. At low levels
of mercaptan addition, the rate of polymerization increases sharply as the
amount of mercaptan increases, but above ca. o. t pphm the rate of polymerization
becomes virtually independent of further addition. This phenomenon has
never been satisfactorily explained. The conventional view is that mercaptans
increase the rate of initiation in such systems by forming a redox couple with
the peroxodisulphate ions. The reactions by which initiating radicals are
generated are usually written as follows:

S20~- -+ 2·0S0;
·OSO; + RSH - + HSO; + RS·
RS· + M - + RSM·
However, although these reactions may seem plausible, they do not account
for the observed difference between styrene emulsion homopolymerizations
and styrene-butadiene emulsion copolymerizations.
Much of the published information concerning the effects of mercaptans
as modifiers for emulsion polymerization reactions relates to their behaviour
in systems for the emulsion copolymerization of styrene and butadiene. Of
the wide range of mercaptans available, alkyl mercaptans containing ca. 12
carbon atoms are preferred as modifiers for this copolymerization. The lower
homologues tend to interfere unduly with the course of the polymerization;
the higher homologues are relatively ineffective because the rate of consumption
is limited by restricted diffusion through the aqueous phase from monomer
droplets to the reaction loci. Tertiary mercaptans are preferred to the
corresponding straight-chain isomers, because they transfer to micelles and
reaction loci more readily. Papers which give much useful detailed information
concerning the effects of various mercaptans upon the emulsion copolymerization
of styrene and butadiene include those by Kolthoff and Harris [56,57], by
Harris and Kolthoff [58,59], by Starkweather et al. [60], by Frank et al.
[61], by Meehan, Kolthoff and Sinha [62], and by Uraneck and Burleigh
[63,64]. These papers deal with two distinct but interrelated aspects of the
behaviour of mercaptans in such reactions: the rate of consumption of
mercaptan, and the effect of the mercaptan upon the polymer produced. It
is usually desirable that the mercaptan/monomer ratio change as little as
possible as polymerization proceeds. The rate of reaction of the mercaptan
at the reaction locus is kinetically first order in monomer concentration. The
rate of polymerization of the monomer is also first order in monomer
concentration. Thus ifthe mercaptan/monomer ratio were to remain constant
throughout the reaction, the modifying effect would be uniform as reaction
proceeds. It is for this reason that a knowledge of the rate of consumption
of mercaptan relative to monomer is important, in order to know what steps
should be taken to maintain the mercaptan/monomer ratio approximately
constant as polymerization proceeds. Uraneck and Burleigh [65] have
reported the results of an extensive investigation of the behaviour of various
mercaptans as modifiers for acrylonitrile-styrene and acrylonitrile- butadiene
emulsion copolymerizations. A useful review by Krause [66] of the behaviour
of mercaptans as modifiers for emulsion polymerization reactions was
published in 1954. This paper gives a comprehensive list of references up to
the early part of the decade in which it was published.
Many other types of substance have been proposed and used as modifiers
in emulsion polymerization reactions. These include various mercaptan

derivatives, dialkyl xanthogen disulphides, di-aroyl disulphides, carbon
tetrachloride, thiuram disulphides, hydrazines, amines, substituted phosphines,
Schiffs bases, nitroso compounds, diazoamino compounds, sulphur and
selenium. Sulphur has been widely used as a modifier for the emulsion
polymerization of chloroprene to give polychloroprene synthetic rubber. This
application is unusual in that the degree of polymerization of the polymer
is reduced not during the polymerization but subsequent to polymerization
(see further section t 1.4. t of Chapter 11).
10.2.3.4 Electrolytes
It is often the practice to include small amounts of simple inorganic
electrolytes in emulsion polymerization reaction systems. The principal effect
of such electrolytes is to increase the ionic strength of the aqueous phase,
rather than to bring about specific chemical interactions with other components
of the system. At least for styrene-butadiene emulsion copolymerizations
using micelle-forming carboxylate soaps as colloid stabilizer, such electrolytes
appear to perform two functions:
1. In low concentration, they increase the rate of polymerization somewhat.
The probable reason for this is that they lower the concentration at which
the soap first begins to form micelles. Thus the concentration of micellar
colloid stabilizer is increased somewhat, as also is the concentration of
micelles if the aggregation number of the micelles is unaffected by the
presence of the electrolyte. The increase in the concentration of micellar
colloid stabilizer is in turn thought to increase the number of loci which
are nucleated. This latter hydrothesis is consistent with the Smith-Ewart
[29] theory of particle nucleation (see section 10.2.2.2.1 above). According
to that theory, the number of particles nucleated at the end of Interval I
depends only upon the total amount of micellar colloid stabilizer initially
present in the system; it is independent of the number of micelles amongst
which the micellar colloid stabilizer is distributed. The effect of electrolytes
in increasing the amount of micellar colloid stabilizer in the aqueous phase
is particularly marked if the concentration of such stabilizer in the absence
of the electrolyte is low or zero. This will be so if the concentration of
colloid stabilizer is low relative to the minimum concentration necessary
for micelle formation.
2. Although large additions of electrolytes to emulsion polymerization
systems cause irreversible colloidal destabilization of the latex as it forms,
small additions have been found to prevent reversible gelation which ma}
occur during the polymerization of certain electrolyte-free systems.
The second of these aspects of the effects of simple inorganic electrolytes

upon emulsion polymerization reaction systems has been investigated by
Fryling and Gindler [67] in respect of the preparation of latices of

styrene-butadiene copolymers. The basic reported facts are as follows:
1. A certain minimum concentration of electrolyte is always required to
prevent gelation of the latex as it forms.
2. The addition of neutral electrolytes to a gelled electrolyte-free latex causes
a reduction in pH as well as liquefaction; but the latter effect cannot be
attributed to the former.
3. The greatest overall reduction in pH accompanies the addition of sufficient
electrolyte to cause complete liquefaction of the gel.
4. The addition of water-soluble hydrocolloids tends to narrow the range of
electrolyte concentration over which it is possible to obtain fluid latices
which are also free from coagulum.
5. Mere addition of water to a gelled latex does not effect liquefaction.
Electron-microscope studies of latices which have been obtained from
electrolyte-free reaction systems have revealed a tendency for the particles
to align and cluster without coalescence occurring. For this reason, the cause
of the clustering has been suggested to be association, possibly by hydrogen
bonding, between un-ionized carboxylic-acid groups which are present on
the surfaces of adjacent particles. Whatever the cause, the strength of the
interparticle bonding is evident from observation (5) of the above list.
Presumably, addition of water to a gelled latex is not of itself sufficient to
destroy the interparticle bonding. The peptizing effect of alkalis is attributed
to ionization of the carboxylic-acid groups to give carboxylate anions at the
surface of the particles. The electrical double layer which is thereby established
at the particle surface then gives rise to repulsive forces between the particles.
It is clear that this mechanism is unable satisfactorily to explain peptization
by neutral electrolytes, such as sodium and potassium chlorides, because
these electrolytes of themselves have little effect upon the pH of aqueous
media. Fryling and Gindler [67] have offered a partial explanation. They
suppose that latices obtained from electrolyte-free reaction systems contain
particles the surfaces of which are covered with ionized but substantially
undissociated ion-pairs derived from the soap initially added to the system.
Such particles experience little mutual repulsion, because the electrical double
layers surrounding them are almost completely collapsed, being effectively
a Helmholtz double layer, for which the theoretical electrokinetic potential
is zero. These particles may therefore be capable of reversibly associating
into clusters through hydrogen bonds formed with layers of water molecules
bound to their surfaces. It is then postulated that the addition of small
amounts of electrolytes causes dissociation of the ion-pairs, with the
consequent formation of an electrical double layer of the diffuse type and
the establishment of a non-zero electrokinetic potential. Repulsive forces are
now operative between the particles, which separate under the influence of
these forces. At higher electrolyte concentrations, the diffuse double layer is
compressed in the usual way, with consequent reduction in the electrokinetic

potential and the repulsive forces between the particles. The electrokinetic
potential which is established at any particular electrolyte concentration is
therefore determined by the balance between these two effects. The clustering
of particles which is caused by the addition of excess electrolyte is essentially
irreversible, possibly because the added electrolyte competes with the
ion-pairs for water molecules, so that the hydration sheath which hitherto
surrounded the particles is now severely disrupted. The variation of electrokinetic
potential with concentration of added electrolyte, as envisaged by Fryling
and Gindler, is shown schematically in Fig. 10.4. Two cases are illustrated:
one for which the electrokinetic potential in the absence of electrolyte is
exactly zero, and the other for which the electrokinetic potential in the absence
of electrolyte is small but non-zero. As the electrolyte concentration is
increased, so the electrokinetic potential is envisaged as passing through a
maximum. This maximum corresponds to the condition of maximum latex
colloid stability.
The cause of the reduction of pH which accompanied the addition of
neutral electrolytes to the latices investigated by Fryling and Gindler [67]
is obscure. It may be that, in some way, hydrolysis of the soap anions was
partially suppressed. As regards the observation (4) above, it is well-known
that water-soluble hydrocolloids are able to sensitize the response of
lyophobic sols to electrolyte additions. In particular, the phenomenon may
be attributable to two effects: enhancement of association between particles
in the absence of electrolyte, due to interparticle bridging by the hydrocolloid,
and possibly also to the encouragement of hydrogen bonding; and competition
for the water of hydration in the presence of excess electrolyte.
Reversible gelation Coacervation

Electrolyte concentration
Figure 10.4 Schematic variation of electrokinetic potential with concentration of added

electrolyte, as envisaged by Fryling and Gindler [67]. Two cases are illustrated: A, for which
the electrokinetic potential in the absence of electrolyte is exactly zero; and D, for which the
electrokinetic potential in the absence of electrolyte is smaIl but non-zero.
10.2.3.5 Other components ofaqueous emulsion polymerization reaction systems
10.2.3.5.1 Crosslinker and graftlinker components

Monomers as the providers of the repeat units of the polymer molecule which
forms have been excluded from this review of the principal components of
emulsion polymerization reaction systems. It is nevertheless appropriate to
refer briefly to two special related types of comonomer which are sometimes
incorporated in reaction systems. They are incorporated not because of the
chemical nature of the units which are thereby incorporated in the polymer,
but because they give rise respectively to potential crosslinking or grafting
sites in the polymer. The distinction between these two types of comonomer
is not sharp, but is useful in practice. Both types contain two or more olefinic
double bonds which are either not conjugated or only weakly conjugated.
The difference between the two types lies in the relative reactivities of the
various double bonds in the molecule towards free-radical addition
polymerization. The difference is therefore one of degree rather than of kind.
Confining attention to comonomers which contain two olefinic double bonds,
the distinguishing characteristics of the two types can be summarized as follows:
1. Crosslinker comonomers are such that both double bonds have similar
reactivities, those reactivities being similar to that of the main monomers
in the reaction system.
2. Graftlinker comonomers are such that one double bond is much less
reactive than the other, the more reactive of the double bonds having
similar reactivity to that of the main monomers in the reaction system.
Incorporation of either type of comonomer in the polymer by polymerization

through one of the double bonds provides a unit which retains the other
double bond of the monomer as a potentially-reactive pendant functionality.
Propagation of polymerization through the residual double bonds leads to
the formation of either a potential crosslinking site or a grafting site. A further
possibility can arise with both types of monomer, analogous to that which
can arise if a main monomer contains more than one olefinic double bond
(see section 10.2.2.4 above). This is that transfer-to-polymer reactions may
be facilitated at adjacent sites through activation of hydrogen or other atoms.
Mainly crosslinking occurs if propagation through the residual pendant
double bond occurs more or less simultaneously with the continuing
propagation of the main polymer chain. Mainly grafting occurs if, because
of the lesser reactivity of the residual pendant double bond, propagation
through it is retarded until a later stage of the reaction. Grafting is also
encouraged if transfer-to-polymer reactions are facilitated by the presence of
the second olefinic double bond. The functions ofthe two types of comonomer
overlap to some extent: graftlinker comonomers bring about some crosslinking,
and crosslinker comonomers bring about some grafting. In principle, it is
possible to envisage comonomers which contain three or more olefinic double

bonds such that two at least have reactivities similar to that of the main
monomers, and at least one has lesser reactivity. Such co monomers would
function both as a crosslinker and a graftlinker. However, complex comonomers
of this type are rarely, if ever, encountered. If both crosslinker and graftlinker
functions are required in an emulsion polymerization reaction system, then
it is more usual to incorporate two separate comonomers.
Common examples of crosslinker comonomers include p-divinylbenzene
(Structure X), ethyleneglycol diacrylate (1,2-ethylene diacrylate) (Structure
XI) and ethyleneglycol dimethacrylate (1,2-ethylene dimethacrylate) (Structure
XII). Crosslinker monomers such as these, which contain two double bonds
CH 2 =CH.CO CH 2 = C(CH 3 ).yO

I
(y°H
I °I
2)2' (yH 2 )2
o o
CH=CH 2 I I
CH2 CH.CO CH 2=C(CH3)·CO
X XI XII
lead to the formation of tetrafunctional crosslink sites. It is usually desirable

that the chemical structure of the crosslinker monomer be as similar as
possible to that of one or other of the main monomers. This encourages
reasonably random insertion of crosslink sites as polymerization proceeds.
Thus p-divinylbenzene is particularly suitable for inserting tetrafunctional
crosslink sites in polymers which contain a substantial proportion of repeat
units derived from styrene, and ethyleneglycol diacrylate and dimethacrylate
are particularly suitable for inserting such crosslink sites in polymers which
contain substantial proportions of repeat units derived from acrylates and
methacrylates respectively. The amounts of such monomers which are
incorporated are small, being typically ca. 0.5 pphm.
Typical graftlinker comonomers include allyl methacrylate (Structure XIII)
and diallyl maleate (Structure XIV). Both monomers contain olefinic double
bonds of widely differing reactivities in free-radical addition polymerization
reactions. Taking allyl methacrylate as the example, the double bond of the
methacrylate moiety readily polymerizes, the comonomer molecule thereby
becoming incorporated in the polymer. The allyl moiety does not polymerize
nearly so readily as does the methacrylate moiety, and so initially remains
largely unreacted as an unsaturated pendant group. The allyl moiety does,
however, copolymerize to some extent during subsequent polymerization,
thereby leading to the formation of graft copolymer. The hydrogen atom in
the a-position to the allylic double bond also has increased susceptibility to
abstraction by radical transfer; thus transfer-to-polymer at this site is
facilitated. The amounts of such monomer-s which are incorporated are again
small, being again typically ca. 0.5 pphm.
CH2 =C(CH 3 )·CO

I CH 2 =CH.CH 2 ·O.CO.CH=CH.CO.O.CH2 ·CH=CH 2
CH 2 =CH.CH 2 ·O cis
XIII XIV
Conceptually, the effect of both crosslinker and graftlinker comonomers

upon polymer structure is the converse of that of modifiers. Especially is this
true of the crosslinker comonomers. However, the effects of modifiers on the
one hand, and of the comonomers under consideration here on the other,
are not exactly converse. For this reason, it is sometimes the practice to
include both modifier and crosslinker or graftlinker comonomer in an
emulsion polymerization reaction system.
10.2.3.5.2 Freezing-point depressants for aqueous dispersion media

For reasons discussed in section 11.2 of Chapter 11, there has been
considerable interest in the past in reducing the polymerization temperature
at which rubbery emulsion copolymers of styrene and butadiene are produced.
Published evidence indicates that, for several reasons, it is advantageous to
reduce the polymerization temperature as low as possible. There has therefore
been some interest in developing practicable emulsion polymerization reaction
systems which are essentially aqueous, but which can nevertheless be operated
at sub-zero temperatures. In order to prevent freezing of the dispersion
medium, it is necessary to replace part of the water with a miscible liquid of
lower freezing point which does not seriously interfere with the polymerization
reaction. Methanol is one such liquid which has been proposed. One potential
problem is that, if particle nucleation occurs by the micellar mechanism, and
if the critical micelle concentration of the colloid stabilizer is raised significantly
by the presence of the freezing-point depressant, then fewer reaction loci will
be nucleated and the rate of polymerization will be correspondingly reduced.
Howland et al. [68] have reported information relating to the emulsion
copolymerization of styrene and butadiene at ca. -18°C using mixtures of
fatty-acid and rosin-acid soaps as colloid stabilizer. They found that, using
a 75/25 mlm water Imethanol mixture as dispersion medium, up to approximately
one half of the fatty-acid soap could be replaced by rosin-acid soap without
there being much effect upon the rate of polymerization. However, replacement
of more than half the fatty-acid soap led to increasingly severe retardation
of the polymerization. This effect is illustrated in Fig. lO.5(a). The effects of
methanol concentration and of temperature upon the critical micelle
concentrations of fatty-acid and rosin-acid soaps are shown in Figs 1O.5(b)
and (c) respectively. In general, it is found that decrease of temperature and
increase of methanol concentration both cause the property changes by which
'?E 8
"C
(5
E
..,......
c:
0 6
of! ...
o
)(
~ 5 c:
c: ;I
o
0
0 !
:.e. 4 1:
b
U) g
.9 3 8
CD .l!!
,§ 2
~
~ E
fii ~
~ ;::
8 0o~----~----~----~
o o1,----....,2..,,o-4~0-6~0,....-8~0,....-1'"'-00 10 20 30
Proportion of rosin acid soap in mixed Methanol concentration 1% mlm
rosin acid-fatty acid soap I % mlm
(b)
(a)
20
O~--~~--~~--~----~
-20 0 20 40 60
Temperature 1°c
(e)
Figure 10.5 (a) Effect of methanol concentration upon time required for 60"1. conversion in a
styrene-butadiene emulsion copolymerization at ca. -18°C (Howland et at. [68]). (b) Effect of
methanol concentration in methanol-water mixtures upon critical micelle concentrations of
fatty-acid soap (points e) and rosin-acid soap (points .) at 2,C (Howland et al. [68]). (c) Effect
of temperature upon the critical micelle concentrations oHatty-acid soap (points e) and rosin-acid
soaps (points .) in 25% (% m/m) methanol/water mixture (Howland et at. [68]).
critical micelle concentration is identified to become less sharp. This indicates

that the transition between dissolved monomolecular soap and micellar soap
has become more gradual. It appears that the number of micelles formed by
rosin-acid soaps dissolved in water-methanol mixtures is relatively small,
even though the overall concentration of soap is high. This may be the reason
why the rate of polymerization is relatively low. Inorganic electrolytes appear
to have little effect upon the critical micelle concentrations of either fatty-acid
or rosin-acid soaps in water-methanol mixtures at low temperatures.
Other non-electrolyte freezing-point depressants which have been considered
for use in styrene-butadiene emulsion copolymerization reaction systems
include ethanol, isopropanol, acetone, acetonitrile, dioxan, ethyleneglycol,
glycerol and formamide. Inorganic electrolytes have also been considered as
freezing-point depressants. Advantages which are claimed over the non-
electrolyte organic freezing-point depressants include cheapness and a
tendency to encourage, rather than discourage, micelle formation with those
colloid stabilizers which micellize in aqueous solution.
10.2.3.5.3 Sequestering agents

Quite apart from any metal-ion sequestering agents which may be added to
an emulsion polymerization reaction system to complex the iron component
of a redox initiation system, it is also common practice to add very small
amounts of such substances to increase the rate of polymerization, to reduce
the variability of the polymerization behaviour, and to increase the colloid
stability of the latex which is produced. Ethylenediaminetetraacetic acid and
its alkali-metal salts have been widely used for this purpose. The usual
explanation given for the effects of these additives is that they de-activate
traces of calcium and magnesium ions which may be present in any component
of the reaction system, especially the water. Contamination with traces of
metal ions can also occur through contact with defective reactor linings.
10.2.3.5.4 Shortstoppers
Shortstoppers are substances which can be added at any stage of the
polymerization in order to inhibit further polymerization. They are of especial
importance in connection with the emulsion copolymerization of styrene and
butadiene to give rubbery polymers, because the copolymer which is produced
in the later stages of the reaction is inferior in properties to that produced
in the earlier stages. As discussed in section 10.2.2.4 above, the main reason
for this inferiority is increased tendency for polymer branching and crosslinking
during the latter stages of polymerization.
The main requirements of the ideal shortstopper are as follows:
1. It should swiftly halt the polymerization reaction when added to the

reaction system in small quantity.
2. It should discourage further chemical modification of the polymer (e.g.,

by degradation or crosslinking) once polymerization has ceased. Such
chemical modification might tend to occur, for instance, during the stage
at which unreacted monomer is stripped from the latex.
3. It should not adversely affect the colloid stability of the latex.
4. It should not adversely affect the physical or chemical properties of the
polymer obtained from the latex.
5. It should not tend to be left behind in reactors after the shortstopped
latex has been removed; otherwise subsequent polymerizations in that
reactor will be severely retarded.
6. It should not cause discoloration of the polymer.
7. It should be cheap, readily available, and have no hazards associated with
its use.
8. For convenience, it should be readily soluble in water, and should be
capable of being stored as an aqueous solution over long periods.
9. It should be of wide application, and should display all the above desirable
attributes when used with a given monomer reaction system initiated by
various methods.
The primary purpose in adding a shortstopper is to halt the propagation

step of the polymerization abruptly. This is normally brought about either
by a transfer reaction with the formation of a relatively unreactive free radical,
or by a copolymerization reaction, again with the formation of a relatively
unreactive free radical. In either case, the free radical which is formed may
itself be effective in terminating a second propagating free radical by mutual
combination or disproportionation. A shortstopper ideally also has a
secondary effect, in that it should react with, and destroy, the radical-initiation
system. This is one way in which the requirement of no post-polymerization
modification of the polymer can be met. One consequence ofthis consideration
is that the effectiveness of a given shortstopper may depend to some extent
upon the initiation system with which it is used. The requirements for
shortstoppers to be used with low-temperature emulsion polymerizations
tend to be more stringent than those for shortstoppers to be used with high-
temperature reactions, partly because the average lifetime of free radicals is
longer at low temperatures than at high temperatures.
A wide range of substances has been used for the shortstopping of
styrene-butadiene emulsion copolymerizations. These include compounds
containing, or capable of forming, quinonoid structures, nitro and nitroso
compounds, oxygen, aromatic polyhydroxy compounds such as catechol and
pyrogallol (which can be classified as a compound capable of forming a
quinonoid structure), water-soluble dialkyldithiocarbamates, sulphur, and
2,4-dinitrochlorobenzene and certain of its derivatives. The water-soluble
dialkyldithiocarbamate salts are particularly suitable for shortstopping
styrene-butadiene emulsion copolymerizations. They are added to the extent
of ca. 0.2 pphm. The potassium salts have been preferred to the sodium,
because, although more expensive, they have the advantage of not promoting
the formation of microcoagulum. Thus the potassium salts may be added as
concentrated solutions; this is important if it is desired to avoid undue dilution,
as is the case if the latex is subsequently to be concentrated. It might be
thought that the use of dialkyldithiocarbamates as shortstoppers for the
latices of rubbery butadiene copolymers might lead to erratic vulcanization
behaviour in the subsequent dry rubbers. However, the available evidence
indicates that the rate of sulphur-vulcanization of the solid rubbers is not
significantly affected by the use of these shortstoppers, provided that large
excesses are avoided. It appears that residual dithiocarbamate tends either
to be decomposed or to be extracted during the coagulation and washing
stages of the production ofthe dry rubber from the latex. Other shortstoppers
which have found extensive use include hydroquinone (ca. 0.1 pphm) and
di-tert-butylhydroquinone (ca. 0.2 pphm).
Extensive surveys of the various types of substance which can be used as
shortstoppers for emulsion polymerization reactions have been published by
Wakefield and Bebb [69] and by Antlfinger and Lufter [70]. The use of
water-soluble dialkyldithiocarbamates has been described by Howland et al. [68].
10.2.3.5.5 Post-reaction additives

A variety of substances can be added to the immediate product of an emulsion
polymerization reaction to induce desirable effects. Post-reaction stabilizers
are sometimes added to enhance the colloid stability of the latex, either as
a preliminary to processes such as removal of unreacted monomer or latex
concentration, or to render the latex more suitable for particular applications.
Adjustment of the latex pH may also be necessary. If adjustment is required,
it is usually that the pH is to be raised. Especially is this so when functionalized
latices have been produced by emulsion polymerization under acidic conditions,
but are required to be alkaline for application (see further section 10.3.3.6
below). Raising of pH is accomplished by addition of a suitable alkali. The
choice depends upon several factors, including effect upon colloid stability
and possible pollution problems arising from disposal of effluent from
processes in which the latex is subsequently used.
10.2.3.6 Emulsifier-free emulsion polymerizations

It is now recognized that the minimum number of components essential for
the creation of an aqueous emulsion polymerization reaction system is three,
rather than four, namely, monomer(s), water and initiator. The presence of
colloid stabilizer is unnecessary. If the reaction conditions are appropriate,
latices can be formed by emulsion polymerization in reaction systems which
do not contain added colloid stabilizers, micelle-forming or otherwise,
although the latices may have low colloid stability. Such reaction systems
are known as emulsifier-free emulsion polymerizations, and the products

formed as emulsifier-free latices. As for most emulsion polymerization
systems, the initiators used for the preparation of emulsifier-free latices
generate radical-anions; thus some at least of the polymer chains which are
produced have anionic end-groups. The colloid stability of emulsifier-free
latices so produced is believed to be conferred principally by the presence of
anionic end-groups at the surface of the polymer particles. Latices of this
type have been principally of interest as model colloids which are used to
investigate some of the more fundamental aspects of the behaviour of
lyophobic sols. These latices have the advantage of being relatively 'clean',
in that, especially after dialysis against water, they approach very closely
dispersions of polymer particles in water, colloidally stabilized by the
minimum number of hydrophilic moieties chemically bound at the particle
surface. Emulsifier-free latices, and latices prepared using very small amounts
of colloid stabilizer, may also be of interest in certain specialized applications
where it is essential to reduce to the minimum the concentration of hydrophilic
substances in the latex.
Emulsion polymerization reaction systems which contain no colloid
stabilizers as such, but do contain small amounts of polymerizable surface-
active substances of the type noted in section 10.2.3.2 above, can also be
regarded as being emulsifier-free, because initially the system does not contain
any conventional colloid stabilizer. Colloid stability is now conferred upon
the latex principally by those copolymerized units derived from the polyrnerizable
stabilizer which are situated at the particle surface.
10.2.4 Methods of carrying out emulsion polymerization in practice
10.2.4.1 Resume of methods

Methods of carrying out emulsion polymerization reactions in practice can
be classified broadly by two distinct types of criteria. These are respectively
the way in which the reaction loci are formed, and the way in which the
Interval II and Interval III reactions are effected.
Classification according to the way in which the reaction loci are formed
gives two types of reaction system:
1. Those in which no reaction loci are initially present. The reaction loci are
generated during the early stages of the reaction by mechanisms such as
those described in section 10.2.2.2 above. Reactions of this type are
sometimes known as ab initio emulsion polymerizations.
2. Those in which reaction loci are present in the initial system, those loci
having been formed by a separate reaction. Reactions of this type are
usually known as seeded emulsion polymerizations.
Of course, it is possible to design reaction systems in which some of the

reaction loci are formed ab initio and some have been derived from seed
particles. Industrially it is sometimes the practice to have present in a reaction
system a small proportion, often known as a heel, of a similar reaction system
which has previously been polymerized.
When classified according to the way in which the Interval-II and
Interval-III reactions are effected, emulsion polymerization reaction systems
are of three types as follows:
1. batch processes, in which all the monomers and most of the other
ingredients are placed in a reactor initially;
2. semi-continuous processes, in which all the reaction system is of such a
size that it can be accommodated in the reactor, but not all the monomers,
and possibly not all the other ingredients, are placed in that reactor initially;
3. continuous processes, in which the reaction system as a whole is of indefinite
size, and is continuously fed to, and removed from, a suitable reactor at
rates such that the total volume of system undergoing reaction at any
instant is constant.
Batch and semi-continuous emulsion polymerizations are procedures which
are adopted when the amount of latex to be produced is relatively small,
and when it is desired to manufacture different grades of latex according to
demand. Continuous emulsion polymerization is appropriate for the large-
scale manufacture of a single grade of latex for which there is an assured
demand. Once the start-up period for a continuous process has been
completed, the product obtained is usually more uniform than that produced
by batch and semi-continuous processes. The foliowing sub-sections give
brief accounts of these three ways in which the Interval II and Interval III
reactions can be effected. It should be borne in mind that in each case the
reaction loci can in principle be formed ab initio or by the seeding technique,
or by a combination of both methods.
10.2.4.2 Batch emulsion polymerization

The batch-type process is the simplest method for effecting emulsion
polymerization. It is the type of process which has been described in outline
in section 10.2.1.1 above. In its simplest form, all the reactants are placed in
a reactor initially, the system is stirred, and the required temperature-time
profile for the reaction is established. The reaction is taken to the requisite
extent of conversion, shortstopper is added if necessary, and the product is
removed from the reactor. In more complex variants of the batch process,
although all the monomer and water are added to the reactor initially, some
of the other ingredients are withheld and then added subsequently, possibly
incrementally or even continuously. In particular, it may be advantageous
to withhold some of the modifier and colloid stabilizer: the modifier to
minimize variations in the modifier/monomer ratio as the polymerization

proceeds, and the colloid stabilizer to enhance colloid stability as the
aggregate surface area of monomer-swollen polymer particles increases as
polymerization proceeds. If the colloid stabilizer is added incrementally, then
the rate of addition relative to the rate of polymerization is adjusted so that
new reaction loci either are or are not nucleated, according to intention. If
new reaction loci are allowed to nucleate after Interval I is completed, then
it is likely that the particle-size distribution of the final latex will be broader
than if new reaction loci are not nucleated. Some of the initiator may also
be withheld initially. A situation where this is advantageous is that in which
it is desired to polymerize the monomers as completely as possible, so that
the content of unreacted monomers in the latex is as low as possible. The
procedure may then be adopted of withholding a proportion of the initiator
until the final stage of the reaction. After the remaining initiator has been
added, the temperature of the reaction system may be raised somewhat.
Although simple in principle, the batch procedure has the important
disadvantages that no control is exercised over either the monomer/polymer
ratio in the reaction loci, or over the comonomer ratio if, as is common in
industrial practice, it is an emulsion copolymerization rather than an emulsion
homopolymerization which is being carried out. In a batch process, the way
in which both these ratios will vary as the polymerization proceeds is
determined entirely by the nature of the reaction system as a whole.
10.2.4.3 Semi-continuous emulsion polymerization
In semi-continuous emulsion polymerizations, an attempt is made to overcome
the disavantages of the batch process. Control is exercised over both the
monomer/polymer ratio in the reaction loci, and, in the case of copolymerizations,
over the comonomer ratio, by imposing an appropriate monomer-
addition regimen. The desirability of exercising control over the mono-
mer/polymer ratio is often interrelated to the matter of exercising control
over the comonomer ratio at the reaction loci as polymerization proceeds.
For this reason, both these matters are considered in more detail when
emulsion copolymerization is discussed in section 10.2.4.5 below. It is
sufficient to note here that, if the monomer-addition regimen is such that
monomers are added at the same rate as they are polymerizing, then the
concentration of unreacted monomer in the reaction system as a whole
remains essentially constant. A special case of this situation occurs if the
monomer-addition regimen is such that the concentration of unreacted
monomer is always low. Emulsion polymerizations carried out under this
condition are described as being monomer-starved. A wide range of
monomer-addition regimens in semi-continuous emulsion polymerization is
possible in principle. Addition may be continuous or incremental. Superimposed
upon the monomer-addition regimen can be regimens for the addition of the
other ingredients of the system.
10.2.4.4 Continuous emulsion polymerization

As noted above, in continuous emulsion polymerization processes, the
reaction system is continuously fed to, and removed from, a suitable reactor
at rates such that the total volume of system undergoing reaction at any
instant is constant. In principle, continuous additions can be made at more
than one point in the reactor, and therefore at different extents of polymerization.
Four types of reactor are used for effecting continuous emulsion polymerization:
1. single continuous stirred-tank reactors, which, as the name implies, are

single reactors into which reaction system is fed continuously, and from
which it is removed continuously at the same rate, the mixture in the tank
being agitated to such an extent that it remains effectively homogeneous
at all stages of the process;
2. trains of continuous stirred-tank reactors in series;
3. tubular reactors, which are such that the reaction system moves at a
uniform rate through a long tube;
4. loop reactors, which are effectively tubular reactors in which the tube has
the form of a continuous loop, reaction system being fed into the loop at
one point and removed from another.
These four types of continuous reactor are illustrated schematically in Figs

10.6(a)-(d) respectively.
An important matter relating to continuous reactors is that of residence
time, i.e. the time for which the reaction system is exposed to the reaction
conditions imposed by the reactor. This variable is important because it
determines the extent of polymerization which is achieved in the reaction
system. There is always a distribution of residence times, because the actual
residence time of any given small portion of reaction system is determined
by random factors, as well as by the imposed flow rate. This implies that
there is a distribution of extents of polymerization within the emergent
reaction system as a whole, and that therefore polymer has been produced
under a distribution of conditions as regards variables such as monomer/polymer
ratio, modifier/monomer ratio, and comonomer ratio. Insofar as a single
residence time is quoted, this must always be an average value. Known as
the mean residence time, and denoted here by t, it is defined as the ratio of
the total volume of the reactor to the overall volumetric flow rate of reaction
system through the reactor. Residence times are also sometimes scaled relative
to the mean residence time by forming the ratio of the actual residence time
to t. The resultant dimensionless parameter is known as the reduced residence
time, denoted by fJ. The mean reduced residence time, IJ, is, of course, 1. As
will be seen subsequently, the distribution of residence times about the mean
residence time varies widely according to the type of reactor. This has
implications for the distribution of reaction conditions under which the
polymer is produced. For some types of reactor, the distribution is so narrow
(a) (b)
(c)
(d)
Figure 10.6 Schematic illustration of four types of continuous reactor for effecting emulsion
polymerization: (a) a single continuous stirred-tank reactor; (b) a train of continuous stirred-tank
reactors in series; (c) a tubular reactor; (d) a loop reactor.
that the residence times for all parts of the reaction system are effectively
identical, so that all the polymer has been produced under effectively identical
conditions.
The distribution of residence times is broad for a single continuous
stirred-tank reactor. This is obvious qualitatively, because some small
portions of the reaction system always pass through the tank in times much
shorter than t, whilst other small portions are retained for times much longer
than t. Quantitatively the distribution of residence times is simple, being such
that, if q,(t)c5t is the probability that a small element of the reaction system
has resided in the reactor for a time within the range t ± Y2 c5t, then the residence-
time distribution function, q,(t), is given by
(10.14)
By a small element of the reaction system is meant a portion of the reaction

system which is sufficiently small to have passed through the tank (or series
of tanks in the sequel) without losing its identity.
Expressed in terms of the reduced residence time, the distribution function
for residence times in a single continuous stirred-tank reactor is even simpler
than q,(t). If t/I(O)c50 is the probability that a small element of the reaction
system has resided for a time such that () is within the range () ± 1'2 ~(), then
the residence-time distribution function expressed in terms of () is
(10.15)
Thus the distribution of residence times in a single continuous stirred-tank

reactor is exponentional and therefore broad. As the number of continuous
stirred-tank reactors in series is progressively increased, so the distribution
of residence times about the mean narrows. This is again qualitatively obvious,
because, for a given element of reaction system passing through the series of
tanks, statistically there is a tendency for deviations from the mean residence
time in one reactor to be partially offset by deviations from the mean residence
time in the opposite direction in another reactor. In considering the matter
quantitatively, it is assumed for simplicity that the individual mean residence
times for all the reactors have the same value t, implying that all the tanks
contain the same volume of reaction mixture. Then in terms of overall reduced
residence time relative to the total mean residence time, i.e., in terms of
() = Tint where n is the number of tanks in series and T is the overall residence
time in the train of reactors, the distribution function for () is given by
(10.16)
This result includes the result for a single continuous stirred-tank reactor as
the special case where n = 1. Curves for I/I.(() as a function of () for n = I,
2, 3 and 4 are shown in Fig. 10.7. There is a qualitative difference in the
shapes of the curves between that for which n = 1 and those for which n > 1.
For the former, I/I(() decreases monotonically, whereas for the latter I/I(()
has a maximum at a non-zero value of (). Thus the distribution of residence
times is very broad for a single continuous stirred-tank reactor. Much of the
reaction system passes through the reactor in a time which is considerably
shorter than the mean residence time; this is offset by smaller proportions
of the reaction system residing in the reactor for considerably longer than
the mean residence time. The distribution of residence times is narrower for
two continuous stirred-tank reactors in series than for each individually. The
distribution narrows further as the number of reactors in series increases
further; however, the effect is not great.
Equation (10.14) can be derived as follows. Let V denote the volume of
reaction mixture in the single reactor, and let p denote the rate of volumetric
flow ofreaction mixture through the reactor. Then t must be such that pt = V,
i.e., t = Vip. Because the contents of the reactor are continuously agitated
sufficiently to maintain homogeneity at all stages of the addition of reaction
mixture, the probability that a given reaction locus selected at random will
efflux from the reactor in a given time interval, bt, is independent of the
3 4
Figure 10.7 Residence-time distribution function, "'.(8~ as a function of 8 for n = I, 2, 3 and

4, for n continuous stirred-tank reactors in series, as predicted by equation (10.16).
length of time the reaction locus has resided in the reactor. That probability
is equal to the ratio of the number of reaction loci which efflux from the
reactor in the time interval c>t to the total number of reaction loci in the
reactors. Once a steady state has been attained, these numbers are directly
proportional to the respective volumes of reaction mixture. Hence the
probability is equal to the ratio p{JtjV, i.e., to {Jtlf. The probability that a
single reaction locus will reside in the reactor for a time t and then efflux
during the subsequent time interval c>t (so that effectively the locus effluxes
during the time interval t ± Y2 c>t if (Jt is very small) is then obtained by
dividing the time t into a large number, v, of small time increments {Jt, so
that t = vc>t and t + c>t = (v + l){Jt, and calculating the probability that any
particular reaction locus will remain in the reactor for the first v of these
time increments and efflux during the (v + 1)th time increment. The probability
that the reaction locus will efflux during anyone of these time increments
is {Jt/i. The probability that it will not efflux during the same time increment
is (1 - c>t/f). Hence the probability, c/J(t )c>t, that a single reaction locus will
reside in the reactor for a time t, and then efflux during the following time
interval c>t, is the probability that it remains in the reactor for the first v of
the time increments c>t and effluxes during the (v + l)th time increment. Thus,
noting that c>t = tlv, c/J(t)c>t is given by
c/J(t)c>t = ( 1- -{Jt)V {Jt

- = ( 1- -t )V c>t- (10.17)
t [ v[ [
i.e.,
t)Y
q,(t) = -1 ( 1 - - (10.18)
f vf
As v becomes very large, and hence (jt very small, the second factor on the
right-hand side of equation (10.18) approaches e- I / r. Thus in the limit where
the time interval (jt is infinitesimally small, q,(t) has the form given in equation
(10.14). Changing the variate from t to () = tlf, so that the increment of
variate becomes (j() = (jtlf, immediately gives equation (10.15) for 1/1 «().
Equation (10.16) for the distribution function for the overall reduced
residence time in a train of n similar continuous stirred-tank reactors in series
can be derived from equation (10.15) by noting that:
1. the reduced residence time in anyone reactor is a gamma variate of
parameter 1;
2. the residence time for any given reaction locus in the train of reactors is
the sum of the residence times in the individual reactors; and
3. the distribution function for the sum of n independent gamma variates
having parameters 11' 12 , ••• , In is a gamma variate of parameter 11 +
12 + ... + In'
Thus the overall residence time in a train of n similar continuous stirred-tank
reactors in series, reduced relative to the mean residence time in a single
reactor, is a gamma variate of paremeter n. It follows immediately that the
distribution function for () = Tlnf, is as shown in equation (10.16).
A tubular reactor is, in effect, equivalent to an infinite number of
infinitesimally small continuous stirred-tank reactors in series. The distribution
of residence times is consequently very sharp. Almost all the reaction system
passes through the reactor in periods of time which differ little from the mean
residence time. Thus all the polymer produced is formed under virtually
identical reaction conditions. The distribution of residence times in a loop
reactor is more complex. It depends upon the average number of times the
reactants circulate within the loop before effluxing from the loop.
10.2.4.5 Emulsion copolymerization
10.2.4.5.1 General considerations

The principles of free-radical addition copolymerization in the bulk, solution
and suspension modes are well-understood. In what follows, it will be seen
that the application ofthese principles to free-radical emulsion copolymerizations
introduces some additional complexities and possibilities. Attention will be
confined to polymerization systems which contain only two monomers (i.e.,
to binary copolymerizations), denoted by M 1 and M 2' The principles which
apply to these systems also apply to systems containing more than two
monomers, but the outworking is more complex.
Copolymerization occurs during free-radical addition polymerization if a

propagating chain terminating in a radical unit derived from MI can react
with M2 as well as possibly with M I , and likewise a propagating chain
terminating in a radical unit derived from M 2 can react with M I as well as
possibly with M 2 • Usually both these conditions are fulfilled. In the simplest
theoretical treatment of free-radical addition copolymerization, it is assumed
that the reactivity of a radical at the end of a propagating chain is independent
of the nature of the penultimate unit of the chain, or of that of any units
further back in the chain. Thus it is usually assumed that account has to be
taken of only four separate types of propagation reaction. The composition
of the copolymer which forms at any instant is determined by the relative
rates at which these four propagation reactions occur. These relative rates
are determined partly by the concentration of the two monomers, and of the
two types of radical-end of the propagating chains. They are also partly
determined by the inherent reactivities of the two types of radical-end with
the two types of monomer. These inherent reactivities are quantified by the
rate coefficients kll' k 12 , k21' k22 for the four bimolecular propagation steps.
The convention is that the rate coefficient kij refers to the reaction between
monomer M j and a propagating chain terminating in a radicalized unit
derived from monomer Mi'
The composition of the copolymer which forms at any particular instant
in a copolymerization reaction is determined by the instantaneous relative
rates of entry of the monomers into the copolymer. For binary copolymerizations,
the relative rates of entry of the two monomers into the copolymer are given
approximately by the equation [71,72]
(10.19)
where 'I and '2' known as the monomer reactivity ratios, are defined as
'I = kll/k12 and '2 = k22/k21' and [MJ is the molar concentration of
monomer Mi' Thus in each case 'i is the ratio of the rate coefficient for an
Mi radical-end reacting with an Mi molecule to the rate coefficient for that
radical-end reacting with an Mi molecule. The larger is 'i' the more likely is
homo-propagation of M j to occur relative to cross-propagation under any
given conditions of radical and monomer concentrations. The larger is 'I
relative to '2' the richer is the copolymer likely to be in MI relative to M 2 .
Furthermore, the larger is the value of " the more likely is the monomer to
which it refers to enter the copolymer as long blocks of units rather than as
short isolated blocks ofa very few units. Equation (10.19) is commonly known
as the copolymer composition equation. Figure 10.8 shows the variation of
instantaneous copolymer composition with composition of the monomer
feed mixture, as predicted by equation (10.19) for various pairs of values of
' I and '2' Feed compositions are expressed as mole fraction of Mp and
1.0
lii
~
(50.8
Do
8
.5
J!l 0.6
·c
::s
~
'15 0.4
c:
o
~
Q)
0.2
(5
::!:
°o~~~--~--~--~~~
0.2 0.4 0.6 0.8 1.0
Mole fraction of M t in feed
Figure 10.8 Variation of instantaneous copolymer composition with composition of monomer

feed mixture, as predicted by equation (10.19), for pairs of values of ' 1 and'2 as follows: curve
'I 'I 'I
A: = I, '2 = I; curve B: = 4, '2 = 4; curve C: = ~, r2 = 14; curve 0: '1 = 4, '2 = ~;
curve E: r 1 = ~,r2 = 4. Feed compositions are expressed as mole fraction of M h and copolymer
compositions as mole fraction of MI units.
copolymer compositions as mole fraction of MI units. If r l = r 2 = 1, then

the copolymer always has the same composition as the monomer feed mixture.
As r 1 and r 2 deviate from t, so in general the copolymer composition deviates
from the composition of the monomer mixture from which it was formed.
However, it can be shown that, if (rl - l)(r2 - I) > 0, there is a particular
monomer feed composition, called the azeotropic monomer feed composition,
for which the copolymer has the same composition as the monomer feed
from which it forms. If an azeotropic composition exists, it will be revealed
in a diagram such as Fig. 10.8 as the point at which the curve for the variation
of copolymer composition with monomer feed composition crosses the line
for which r l = r 2 = 1 (see curves Band C of Fig. 10.8).
The considerations of the preceding paragraph refer to the composition
of the incremental amount of copolymer which forms at any instant from a
monomer feed mixture of a particular composition. Except in the special
case where r l = r 2 = t, and also of particular azeotropic mixtures if they
exist, the composition of the monomer feed mixture changes as polymerization
proceeds, because the copolymer which forms is relatively richer in one
monomer than the other. This phenomenon is known as composition drift.
Except for azeotropic mixtures, it occurs wherever r l # r 2 and the reaction is
taken to an appreciable extent of conversion. To predict the variation in the
overall composition ofthe copolymer as polymerization proceeds, and also the
drift in the composition of the incremental copolymer which forms at any
particular stage of the reaction, it is necessary to integrate equation (10.19).
The assumption is made that '1 and '2 remain constant as polymerization
proceeds. Even so the integration is complex. One particularly useful form
ofthe integrated copolymer composition equation has been derived by Meyer
and Lowry [73]. Their result is:
(10.20)
In this equation, M denotes the total number of moles of unpolymerized

monomers in the system after an arbitrary extent of polymerization has
occurred, (Jll and (Jl2 are the mole fractions of Ml and M2 respectively after
that extent of polymerization, the corresponding symbols with superscript
'0' against them denote the corresponding initial values, and a, p, y and b
are four parameters related to the monomer reactivity ratios as follows:
a = '2
-- p=_'_I_
1 -'2 1 -'1
(10.21)
Equation (10.20) requires amendment in the special cases where = 1, or '1

'2 = 1, or '1 + '2 = 2 exactly. Such cases are unlikely to occur in practice,
but may arise in theoretical analyses. Figure 10.9 shows the predicted effects
of conversion upon overall copolymer composition and instantaneous
1.0
-a~ O.B
a.
o(.)
.5
.l!l 0.6
'c
:::J
~
~ 0.4 1...--==:::::::::..----
o
tl
~ 0.2 ~_ _--==:::::::._;:_---
(5
~
0.2 0.4 0.6 O.B 1.0

Fraction of monomer mixture copolymerizes
Figure 10.9 Predicted effects of conversion upon overall copolymer composition (curves A and
8) and instantaneous copolymer composition (curves A' and 8') given by equations (10.20) and
(10.19) respectively for a monomer pair for which '. = ·h and '2 = 2, and for which the initial
monomer mole ratio is M./M2 = 25;75 (curves A and A') and M./M2 = SO/SO (curves 8 and 8').
copolymer composition given by equations (10.20) and (10.19) respectively

for a monomer pair for which r 1 = 0.5 and r 2 = 2.0; these monomer reactivity
ratios provide a rough model for the copolymerization of styrene and
butadiene [74]. Curves are shown for reaction systems for which <D? = 0.25
(i.e. MdM2 = 25/75) and for which <D? = 0.50 (i.e. MdM2 = 50/50). The
phenomenon of composition drift is evident from these predictions. It is also
evident that the overall copolymer produced from this copolymerization
becomes increasingly heterogeneous at the molecular level as polymerization
proceeds.
The above predictions apply in the first instance to free-radical binary
copolymerizations occurring in homogeneous phases. For free-radical emulsion
copolymerization reactions, the monomer concentrations [M and [M 2]a
refer to the reaction loci, and not to the reaction system as a whole. In order
to make predictions of instantaneous and overall copolymer compositions
from the overall ratio of the monomers in the reaction system as a whole,
it is necessary to know the relative concentrations of the two monomers in
the reaction loci. This in turn requires a knowledge of the relative partitioning
behaviours of the monomers in the various phases of the reaction system.
These are complications which are usually absent from bulk, solution and
suspension free-radical copolymerizations. An additional complication arises
with emulsion copolymerization reactions if the monomers differ markedly
in water solubility. The relationship between the composition ofthe copolymer
produced and that of the monomer feed mixture is now further confounded
by those differences. The more water-soluble oftwo monomers is disadvantaged
as regards copolymerization behaviour relative to the less water-soluble
monomer, because more of it is dissolved in the aqueous phase, and therefore
is not available for copolymerization in the reaction loci. Notwithstanding
these complications, the view generally held is that, in most cases, if the
overall monomer ratio in the reaction system as a whole is adjusted to allow
for differential partitioning, so that the actual comonomer ratio in the reaction
loci can be estimated, then copolymer compositions can be predicted with
reasonable confidence using the concepts which have been developed for
bulk copolymerization and the monomer reactivity ratios obtained from
experiments using bulk-copolymerization reactions.
It should also be noted that the foregoing discussion is concerned only
with the composition of the copolymer which is formed in free-radical addition
copolymerizations, and not with the rate of copolymerization. It is true that
the rate of entry of one monomer into a copolymer relative to the rate of
entry of the other is equal to the ratio of the absolute time rates of
polymerization of the two monomers. However, as the equations (10.19) and
(10.20) show, it is not necessary to know the absolute rates in order to predict
the relative rates of entry, and hence the copolymer composition. The rate
at which a free-radical binary copolymerization reaction occurs, whether in
bulk, solution, suspension or emulsion, is usually approximately the mean
of the two corresponding homopolymerizations, weighted according to the

relative amounts of the two monomers in the mixture, unless some perturbing
feature, such as enhanced biomolecular termination, is present to reduce the rate.
10.2.4.5.2 Possibility of preparing structured latex particles by emulsion

polymerization
By the term structured latex particle in the context of emulsion polymerization
is meant a synthetic latex particle which has a structure other than that which
it would have if it had been formed in a batch emulsion polymerization
reaction. This possibility arises naturally in connection with emulsion
copolymerization, because the matter of particle structure is usually of
concern only when the polymer contains more than one type of repeat unit.
The usual way of endeavouring to persuade structured latex particles to form
in an emulsion polymerization reaction system is to impose an appropriate
monomer-addition regimen. This is most conveniently achieved when a
semi-continuous polymerization process is being used.
Because of the possibility of imposing structure upon the polymer particles
produced in emulsion copolymerization reactions, it has been rightly said
that latices and polymers produced by this process are appropriately called
products by process. This is because the details of the process can exert a
powerful influence upon the properties of the polymer particles formed, and
hence of final products, such as polymer films, formed from those particles.
Consequently, a copolymer (and to a lesser extent a homopolymer) produced
by emulsion polymerization is not defined solely by its overall composition.
A corollary of these considerations is that emulsion polymerization, more
than any other mode of effecting free-radical addition polymerization, offers
scope for combining together common monomers to form products which
may have special properties, and therefore which may be useful in special
applications. The possibility of so exploiting the potentialities inherent in
unusual combinations of common monomers is becoming important as special-
purpose monomers become increasingly costly to develop and produce.
Three examples of attempts to produce structured particles by the adoption
of appropriate monomer-addition regimens will be discussed. These are:
1. the production of latex particles in which the variation of copolymer
composition with extent of polymerization has been varied by tapered
emulsion copolymerization;
2. the production of nominally core-shell latex particles by sequential, or
multi-stage, emulsion copolymerization;
3. the production of nominally gradient copolymer latex particles by
power-feed emulsion copolymerization.
The adjective nominally is included in examples (2) and (3) because the
structures of the particles which actually form may differ considerably from
those which the monomer-addition regimen was designed to form. The factors
which are believed to determine the structure of the particles which actually
form in an emulsion copolymerization reaction are discussed in section
10.2.4.5.3 below.
By a tapered emulsion copolymerization is meant an emulsion copolymerization
reaction in which the monomer-addition regimen is such that the variation
of instantaneous copolymer composition with extent of polymerization differs
from that which would pertain if all the comonomers were blended initially
and allowed to copolymerize. For simplicity, attention is confined to binary
emulsion copolymerization systems; the principles are the same if there are
more than two monomers, but the implementation is again more complex.
Usually, the desire is to offset the composition drift which otherwise occurs
if the monomer reactivity ratios for the comonomers differ significantly from
1 (see section 10.2.4.5.1 above). The copolymer which forms throughout the
reaction then has less compositional variation than it would otherwise. The
procedure is to adopt a monomer-addition regimen such that the polymerization
commences with a monomer feed mixture which is appropriately richer in
the less reactive monomer than is the copolymer which it is desired to produce.
The monomer feed mixture is then 'tapered' in composition so that it becomes
progressively richer in the more reactive monomer as the polymerization
proceeds. If the polymerization is being taken to completion, then the overall
ratio of the two monomers added over the whole reaction period must be
equal to the overall comonomer ratio in the copolymer which it is desired
to produce. If the polymerization is being stopped short of complete
conversion, then the monomer-addition regimen must be such that the ratio
of the amounts of the two monomers which have been polymerized at the
point of shortstopping is that desired in the final copolymer. In any event,
the tapering of the monomer feed composition should be such as to effect
as closely as possible the desired variation (or non-variation) of overall
copolymer composition with conversion.
The simplest type of core-shell latex particle is one which comprises an
inner core of a polymer of one composition and an outer concentric shell of
a polymer of a different composition. The two polymers can be either
copolymers or homo polymers. Even if both are homopolymers, the latex
particle as a whole is still appropriately described as a copolymer particle,
although it is of an extreme form of copolymer in which the two monomers
have polymerized effectively independently of each other. An obvious
extension of the concept of a core-shell latex particle is to particles which
comprise more than two concentric regions, each of which differs in polymer
composition from the contiguous concentric regions. Nominally core-shell
latex particles are made by so-called multi-stage or sequential emulsion
copolymerization. The monomer mixture which will give the desired composition
for the core is first fed to a reactor and allowed to polymerize to virtual
completion. The monomer mixture which will give the desired composition
for the shell is then fed to the reactor and polymerized under conditions such
that no new reaction loci are nucleated. If further shells are required, then
successive monomer mixtures of appropriate compositions are fed to the
reactor in stages. Minor amounts of crosslinking comonomers of the type
described in section 10.2.3.5.1 above may be included in order to introduce
covalent crosslinks within the various regions of the particle. The motive is
to stabilize those regions against subsequent distortion and disruption. Minor
amounts of graftlinking co monomers of the type described in section 10.2.3.5.1
above may also be included in some of the monomer mixtures from which
the various regions of the latex particle are to be formed. The motive is to
improve the adhesion between the various regions by promoting graft
copolymerization between the polymer which has formed during one stage
and that which forms during the subsequent stage. The graft copolymers
which it is intended should be formed comprise polymer chains of the first
stage as the backbone, and polymer chains formed during the second stage
as the grafts. Graftlinking co monomers used for this purpose are sometimes
known as bonding comonomers, because it is hoped that some covalent
bonding is effected between the polymer which formed first and that which
forms subsequently.
Many multi-stage emulsion copolymerization processes have been described
for the production of structured copolymer latex particles. Typical is that
described by Rohm and Haas [75]. The objective of the process is to produce
by multi-stage emulsion polymerization structured latex particles having
diameters in the range lOO-300nm, each of which comprises a rigid
non-elastomeric polymeric core surrounded by alternating elastomeric and
non-elastomeric concentric shells, the outermost shell being always non-
elastomeric. A large number of monomer combinations can be used to provide
the successive layers of the particles. It is not necessary that all the non-
elastomeric and all the elastomeric regions should have the same chemical
compositions. Typical of the monomers which can be used to provide the
non-elastomeric regions are styrene and methyl methacrylate; typical of the
monomers which can be used to provide the elastomeric regions are butadiene
and n-butyl acrylate. The minimum number of distinct regions within each
particle is 3. In the simplest, and the preferred, implementation ofthe process,
the particles produced contain only three distinct regions:
l. an inner core of a non-elastomeric polymer which has a glass-transition

temperature greater than at least 25·C and preferably greater than 6O·C,
and which preferably constitutes 20-40% m/m of the particle;
2. an intermediate shell of an elastomeric polymer which has a glass-transition
temperature not exceeding at most 2S·C and preferably not exceeding
-lO·C, and which preferably constitutes 30-S0% m/m of the particle;
3. an outer shell of a non-elastomeric polymer which has a glass-transition
temperature greater than at least 2S·C and preferably greater than SO·C,
and which preferably constitutes IS-S0% m/m of the particle.
The composite latex particles are to be prepared in aqueous dispersion or

emulsion by a process in which 'successive monomer charges are polymerized
onto, or in the presence of, a preformed latex prepared by the polymerization
of one or more prior monomer charges and/or stages'. The polymer particles
described by Rohm and Haas are to be made 'by a process in which the
total particle content is kept substantially constant after completion of the
first-stage polymerization, i.e., new additional and distinct particles are
avoided after formation of the first-stage particles'. Crosslinker comonomers
can be included in the monomer mixture for the formation of any of the
regions within the composite latex particles; it is especially desirable that the
elastomeric region should be crosslinked. Graftlinker comonomers may also
be incorporated as bonding comonomers to promote adhesion between the
various regions of the composite latex particles. Some covalent bonding
between the regions occurs even if graftlinker and crosslinker comonomers
are absent, the mechanism by which such covalent bonding occurs being
mainly graft copolymerization. However, the degree of covalent bonding in
the absence of graftlinker comonomers is not usually sufficient. The preferred
degree of attachment of the third stage to the elastomeric shell is at least
20%, and is preferably within the range 75- 100%. The compositions of the
various monomer mixtures are to be adjusted to achieve this degree of
covalent bonding.
The so-called power-feed emulsion copolymerization process is an example
of a procedure which has been developed for the production of gradient
copolymer latex particles, i.e., latex particles the copolymer composition of
which varies in a controlled manner from the centre of the particle to its
surface. Processes of this type have been developed principally to overcome
the disadvantage inherent in sequential emulsion copolymerization reactions
that, whatever may be the detailed morphology of the particles produced,
those particles will almost certainly contain regions in which the composition
of the copolymer changes abruptly because the monomer feed composition
changes abruptly as one stage of monomer addition succeeds another.
Processes of which power-feed emulsion copolymerization process is an
example attempt to overcome this problem by continuously varying the
composition of the monomer feed as the reaction proceeds.
Detailed information concerning the power-feed emulsion copolymerization
process is available in papers by Bassett and Hoy [76] and by Johnston,
Bassett and MacRury [77]. For simplicity, attention is confined to binary
emulsion copolymerization. Similar principles can be applied to emulsion
copolymerizations in which there are more than two comonomers; however,
the implementation is again inevitably more complex. In the following
exposition, attention will also be confined to the simplest experimental
arrangement for effecting power-feed emulsion copolymerization. This
arrangement, which is illustrated schematically in Fig. to. to, utilizes one
reactor which is supplied with monomers from two feed tanks in series. These
Far feed Near feed Polymerization

tank tank reactor
Figure 10.10 Schematic ilustration of simple arrangement for carrying out power-feed emulsion
copolymerization utilizing one reactor supplied with monomers from two feed tanks in series.
are designated the near and far feed tanks respectively, according to their
distance from the reactor. Quantities relating to these two tanks are
distinguished by subscripts A and B respectively. More complex arrangements
are possible, in which more than two feed tanks in series are used, but the
principle of the method is adequately demonstrated by the two-tank
arrangement. The two feed tanks contain monomer mixtures of differing
compositions. The following conditions must then be fulfilled:
I. The rate of transfer of monomer from the far feed tank to the near feed
is to be constant.
2. The rate of transfer of monomer from the near feed tank to the reactor
is to be constant, but this rate will in general differ from the rate of transfer
of monomer from the far feed tank to the near feed tank.
3. Transfer of monomer from the far feed tank to the near feed tank
commences at the same time as transfer of monomer from the near feed
tank to the reactor.
4. Both feed tanks empty simultaneously.
5. The near feed tank is agitated sufficiently vigorously to ensure that the
contents are always effectively homogeneous.
6. The emulsion copolymerization takes place under monomer-starved
conditions.
7. In the simplest case, which is the one considered here, the reactor is either
initially devoid of monomer or polymer, or, if some polymer is initially
present as seed particles, the amount so present is negligible relative to
the total amount of monomer to be polymerized.
Condition (7) ensures that the overall composition of the copolymer particles
produced, and its variation with extent of polymerization, can be inferred
from the monomer-addition regimen. Condition (6) ensures that the composition
of the copolymer which forms at any instant is effectively that of the monomer
mixture from which it was formed; thus the composition of the copolymer
is not determined by the monomer reactivities, but by the imposed monomer-
addition regimen. The composition of the copolymer which forms initially
will be effectively that of the initial monomer mixture in the near feed tank;
that of the copolymer which forms at the end of the reaction will be that of
the monomer mixture in the far feed tank; and that of the copolymer which
forms at any intermediate stage of the reaction will be that of the monomer
mixture in the near feed tank at that particular instant. Because the ways in
which the monomers flow into and out of the near feed tank are known
(conditions (1)-(4», and because the contents ofthe near feed tank are always
well stirred (condition (5», it is possible to predict approximately how the
instantaneous copolymer composition will vary with extent of polymerization.
Let IX denote the fraction of the total initial quantity of monomers which has
beed fed to the reactor after some arbitrary time, and hence, because the
polymerization is taking place under monomer-starved conditions, effectively
the fraction of the total initial quantity of monomers which has been
polymerized. Let C A be the fraction of one of the monomers (say, M \) in the
near feed tank (and therefore in the monomer feed mixture), when the fraction
IX of the total quantity of monomers has been fed, C~ the initial value of C A'
C~ the fraction of the same monomer in the far feed tank (which is constant
throughout the reaction), and K the ratio of the initial quantity of monomer
mixture in the far feed tank to the initial quantity in the near feed tank. Then
the prediction is
CA = C~ - (C~ - C~)(1 - IX)" (10.22)
Before indicating how equation (10.22) is derived, it is necessary to clarify
the units in which C~, CA and C~ are expressed, and what is meant by
quantity ofmonomers in the preceding paragraph, and to note the consequences
which follow from these clarifications. In principle, the composition of binary
copolymers can be defined in various ways, including mass ratios, mass
fractions, mole ratios, mole fractions, etc., of the two types of monomer unit
in the copolymer. In developing the theory of power-feed emulsion
copolymerization, it is convenient to define copolymer composition as the
mass ratio or mass fraction of the two monomers. Thus, in what follows, C~,
CA and C~ are the mass fractions of monomer M \ in the various mixtures,
and K is the ratio ~/W~, where W~ and ~ are respectively the initial
masses of monomer mixtures in the near and far feed tanks. This in turn
implies, as will be clear from what follows, that the rates of transfer of
monomers from the near feed tank to the reactor, denoted here by PA' and
the rate of transfer of monomers from the far feed tank to the near feed tank,
denoted by PH' must be expressed in units of mass (unit time) -1. Both these
rates are to be constant, although in general different, throughout the
monomer addition. Thus they must be maintained constant when expressed
as mass (unit time) -1. To the extent that the monomers differ in density, the
rate PA expressed as volume (unit time)-1 would have to be adjusted
appropriately as monomer addition proceeds and the composition of the
monomer mixture in the near feed tank changes. However, if, as is usually
the case, the densities of monomers are similar, this refinement is unnecessary.
Suppose that monomer addition commences at time t = 0 and is complete
at time t = !. It is possible to set down a number of simple relationships
between the various quantities which have been introduced. Firstly, the rates
PA and PH must be such that
PA =
w2+~ an d PH=-
~ (to.23)
! !
because both tanks empty simultaneously at the time t = !, and the initial
contents of the near tank have been supplemented by the complete contents
of the far feed tank by the time the monomer addition is complete. Thus
inevitably PA > PH. Equations (10.23) lead to several other simple relationships,
such as
K + 1 and PH = K (10.24)
K PA - PH
These relationships are all consequences of the requirement that the transfer
from the two feed tanks should commence simul~aneously and should
continue at constant rates until both tanks empty simultaneously. The fraction
of the total initial mass of monomers which has been fed to the reactor at time t is
IX= PAt _PA-PH t

(10.25)
w2+~- ~
provided that 0 ~ t ~ !. The second of these equations follows from the first
using the relationships shown as equations (10.24). Thus there is a simple
linear relationship between the variables IX and t. There is also a simple linear
relationship for the mass of monomer mixture in the near tank at time t,
denoted by WA , namely
(10.26)
provided that 0 ~ t ~ !. The second of the relationships shown as equations
(10.25) enables equation (10.26) to be transformed to
WA = w2(t - IX) (10.27)
Equation (10.22) can now be derived by considering the mass balance for
monomer M 1 in the near tank over the time interval t ± V2 Jt, during which
+ Y2 (joc. This mass balance gives

oc increases from oc - Y2 (joc to oc
C~PB(jt - CAPA& = (j(CAWA ) = CA(jWA + WA(jC A (10.28)
The first term on the left-hand side is the mass of M 1 which has flowed into
the near feed tank from the far feed tank in the interval &. The second term
on the left-hand side is the mass of Ml which has flowed from the near feed
tank into the reactor during that interval. The product CA WA is the mass of
M 1 which is present in the near feed tank when the mass fraction of M 1 in
that tank is CA and the mass of monomers in that tank is WA • Thus the
centre term is the change in the mass of M 1 in the near feed tank during
that interval. Substituting for WA from equation (10.27) into the right-hand
side of equation (to.28), changing the variable from t to oc in the left-hand
side, rearranging, and noting that K = PB/(PA - PB) (equation (to.24», gives
l)C A {joc
o =K-- (10.29)
CB-C A l-oc
Integrating this equation between the limits C A = C~, oc = 0 (corresponding
to t = 0) and CA = CA, oc = oc (corresponding to t = t) gives
C~ - C A
In C~ _ C~ =K In(1 - oc) (10.30)
i.e.
C~-CA=(I_ )"
cg - C~ oc
(10.31)
Equation (to.22) then follows. The process is known as power1eed emulsion

copolymerization because the composition of the monomer feed mixture at
any instant is a power function of the extent of monomer addition at that instant.
It is also possible to predict how the overall composition of the copolymer
which has formed varies with the extent of monomer addition, if it is assumed
that the monomer-starved condition prevails throughout the monomer feed,
and that therefore almost complete conversion of monomer to polymer has
occurred by the time the monomer addition is complete. If F 1 (oc) denotes the
overall mass fraction of monomer M 1 in the copolymer which has formed
when the fraction of monomers fed is oc, then F 1 (0) = C~. The overall mass
fraction of monomer M 1 in the final copolymer, F 1 (1), is easily obtained as
the ratio of the total mass of M 1 initially present in the two feed tanks to
the total mass of monomers initially present in the two feed tanks, i.e.
F (1) _ C~ W~ + C~ ~ C~ + KC~ _ CO C~ - C~ (10.32)

1 - W~+~ K+l - B- K+l
F 1 (oc) at any intermediate oc can be calculated by noting that the total mass
of monomers added is oc( W~ + ~), and that the mass of M 1 added at this
stage is the sum of all the infinitesimal increments which have been added
from t = 0, ct = 0 to t = t, ct = ct. The incremental mass of M 1 which is added
as ct increases from ct - J'2bct to ct + J'2bct is CA(W: + W~)bct. Thus F 1 (ct) is
given by
F 1(ct) = ~ f
'"
o
CAdct (10.33)
Thus the area under the curve for c1

as a function of ct between the ordinates
ct = 0 and ct = ct is equal to ctF 1 (ct). This area, and the integral by which it
can be evaluated, is equal to the ratio of the total mass of monomer M 1
which has been added to the reactor when the fraction ct of monomers has
been fed, to the total mass of monomers to be added over the whole course
of the addition. Substituting for CAin equation (10.33) from equation (1 0.22) gives
'"
F 1 (ct) = a1 f [CB - 0
(C 0B - CA)(I
0
- ct)"Jdct (10.34)
o
from which
F (ct) = CO _ C~ - c1 [1 _ (1 _ ct)K+ 1 J (10.35)

1 B ct(K + 1)
Equation (10.35) gives as particular cases the above values for F 1 (0) and F 1 (I).
Figure 1O.II(a) shows the variation of C A with ct for c1
= I, C~ = 0 and
various values of K, as predicted by equation (10.22). When K = I, C~ is
always a linear function of ct. As K deviates from I, so the curve departs
increasingly from linearity. For K > I, the curve is concave to the ct-axis; for
K < I, it is convex. It is possible to introduce points of inflexion in the curve
if there are more than two feed tanks. Figure 10.11 (b) shows the corresponding
curves for the variation of F 1 (ct) with ct, as predicted by equation (10.35).
Figure 10.11 (c) compares the variation of C Awith ct for three monomer-addition
regimens for which the overall mass ratio MdM2 is 80/20: uniform monomer
feed (curve A), stepwise monomer feed (curve B), and power monomer feed
(curve C). The case chosen for the latter is that for which ~ = 1, ~ = 0,
K =~. This diagram illustrates the relationship between the power-feed
monomer-addition regimen and the two extremes of uniform addition of a
single monomer mixture and stepwise addition of the two.
The concern so far has been with the variations of C A and F 1 (ct) with ct.
No regard has been had to what happens to the monomers in the reactor,
other than that they copolymerize rapidly, thus maintaining monomer-starved
conditions. If it is assumed that no new latex particles are nucleated during
the course of the monomer addition, and that the particles are uniform in
size at all stages of the monomer addition, then the size of the particles will
1.0
0.8 0.8
0.6 0.6
CA ~ (a)
0.4 0.4
0.2 0.2
00 0.2 0.6 0.8 1.0 00 0.2 0.4 0.6 0.8 1.0

a a
(a) (b)
1.0
0.8
0.6
CA
0.4
0.2
00 0.2 0.4 0.6 0.8 1.0

(c) a
Figure 10.11 (a) Variation of C A with ex for various values of K as indicated, as predicted by
~uation (10.22), for a simple power-feed emulsion copolymerization reaction for which C~ = I,
CD = o. (b) Corresponding variation of F ,(ex) with ex, as predicted by equation (10.35). (c) Variation
of C A with ex for three monomer-addition regimens for which the overall mass ratio of the two
monomers is M,iM2 = 80;20; uniform monomer feed (curve A), stepwise monomer feed (curve
8), and power monomer feed with C~ = I, C~ = 0, K = '/4 (curve q.
be related to the value of IX. If the particles are uniform spheres of radius a,
and changes of density with composition are neglected, then the relationship
between IX and a is a is IX = (a/a"Y, where a", is the radius of the final latex
particles. This substitution enables equations (10.22) and (10.35) to be
re-written to give C A and F 1 (IX) (now written as F 1(a» as functions of a thus:
(10.36)
and
(10.37)
The relationship IX = (a/aocY, and the relationships embodied in equations

(10.36) and (10.37), make no assumption concerning the structure of the
particles which are forming in the latex. Although the intention of the
power-feed monomer-addition regimen may be that the copolymer composition
should grade uniformly from the centre of the particle to its exterior, the
relationship ex = (a/aaY between particle size and IX is true whether or not
this actually happens, provided that no new particles nucleate. If the
copolymer composition grades uniformly from the centre of the particle to
its exterior in such a way that the copolymer which is formed at the radius
a has the composition equal to the value of CA at the extent of monomer
addition IX, then curves for the variation of C A with a show the way in which
copolymer composition changes with location within the latex particle.
Likewise, curves for the variation of F 1 (a) with a show the dependence upon
a of the overall composition of the copolymer which is contained within the
interior of the latex particle as far out as the radius a. Figures 1O.12(a), (b)
and (c) show curves for the variation of CA and F 1(a) with a, corresponding
to those shown in Figs 10.1 1(a), (b) and (c) for the variation of CA and F 1 (ex) with IX.
10.2.4.5.3 Factors which affect the structure 0/ the latex particles which/orm
in emulsion copolymerization reaction systems
We turn now to consider such evidence as is available concerning the actual
structure of synthetic latex particles produced by various monomer-addition
regimens. It is reasonable to suppose that the monomer-addition regimen
affects the particle structure. However, other factors may also be operative,
so that the structure which is most obviously indicated by the monomer-addition
regimen may not be that which actually forms. There is often a difference in
properties between solid copolymers of the same overall composition which
have been prepared by emulsion copolymerization reactions using different
monomer-addition regimens. This has been shown clearly by Bassett and
Hoy [76] for the dynamic mechanical properties of solid 50/50 m/m styrene / ethyl
acrylate copolymers separated from latices made in various ways. Their
results are reproduced here as Figs 10.13(a)-(d). They are for the temperature-
variation of the in-phase and quadrature components of dynamic modulus
(G' and G" respectively) at a frequency of 11 Hz and a heating rate of
I·e minute - 1 for copolymers produced in the following ways: (a) by the
continuous feed of a comonomer mixture of uniform composition; (b) by a
two-stage monomer-addition regimen in which ethyl acrylate was added at
the first stage and styrene at the second; (c) by power-feed addition with the
near feed tank initially containing pure ethyl acrylate and the far feed tank
pure styrene, and (d) by power-feed addition with the near feed tank initially
containing pure styrene and the far feed tank pure ethyl acrylate. All the
copolymers were produced under monomer-starved conditions. For the
power-feed additions, the parameter" was t in both cases; thus the variation
of monomer feed composition with ex was linear. As expected, the results for
0.8
0.6
F, (a)
0.4
0.2
0 00
0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0
(a) a/a oo (b) a/a oo
1.0 B
0.8
A ~
\
CA
0.6
0.4
'\
0.2
00 0.2 0.4 0.6 0.8 1.0

(c) a/a oo
Figure 10.12 (a) Variation of CA with a/a", for various values of /( as indicated, as predicted
by equation (10.36), for a simple power-feed emulsion copolymerization reaction for which
C~ = I, C~ = O. (b) Corresponding variation of F ,(a) with a/a,,,, as predicted by equation (10.37).
(c) Variation of CA with a/a~ for three monomer-addition regimens for which the overall mass
ratio of the two monomers is M,iM2 = 80;20: uniform monomer feed (curve A), stepwise
monomer feed (curve B), and power monomer feed for C~ = I, C~ = 0, /( = '4 (curve q.
the copolymer prepared from a comonomer mixture of uniform composition

indicate a single sharp glass-transition as the temperature is raised or lowered.
This observation is consistent with the copolymer having fairly uniform
composition. The copolymer produced by the two-stage monomer-addition
regimen shows two well-defined transitions, the one (the lower) being
essentially that of an ethyl acrylate homopolymer, and the other (the upper)
being essentially that of a styrene homopolymer. The copolymers produced
by the power-feed monomer-addition regimen show behaviour quite different
from that shown by the uniform-feed and two-stage-feed copolymers. The
transitions now cover a wide range of temperature. This behaviour is
G'
G"
1~~30 -90 -50 -10 30 70 110 150 1~~30 -90 -50 -10 30 70 110 150
Temperature I °C Temperature I °C
(a) (b)
8!. 105
--'cib
..I<:
104 G"
102~~~~__~~~~~~~
1(~~ 30 -90 -50 -10 30 70 11 0 150 -130 -90 -50 -10 30 70 110 150
Temperature I °C Temperature I °C
(e) (d)
Figure 10.13 Temperature-variation of in-phase and quadrature components of dynamic

modulus (G' and G" respectively) at frequency 11 Hz and heating rate IT minute-· for the
following SO/SOmjrn styrene/ethyl acrylate copolymers produred under monomer-starved
conditions (Bassett and Hoy [76]): (a) that produced by continuous feed of a comonomer mixture
of unifonn composition; (b) that produred by a two-stage monomer-addition regimen in which
ethyl acrylate was added at the first stage and styrene at the second; (c) that produced by
power-feed addition with near feed tank initially containing pure ethyl acrylate, far feed tank
pure styrene, and K = 1; (d) that produced by power-feed addition with near feed tank initially
containing pure styrene, far feed tank pure ethyl acrylate, and K = 1.
consistent with the copolymers containing macromolecules having a wide

range of compositions, such as would be expected to be produced by a
monomer-addition regimen of the power-feed type. An expected corollary of
this is that mixing of the segments of the various copolymer macromolecules
at the molecular level should be facilitated.
Although observations such as those described in the preceding paragraph

provide evidence that solid copolymers derived from latices produced using
different monomer-addition regimens may differ in properties, they provide
no evidence concerning the structure of the particles in the latex itself. A
theoretical discussion of some of the factors which might affect the morphology
of latex particles produced by two-stage emulsion binary copolymerization,
together with some supporting experimental evidence, has been given recently
by Waters [78, 79]. Let the two monomers be designated M land M 2. Suppose
M2 to be more hydrophilic than M l , and there to be no tendency for the
two homopolymers to mix at the molecular level. Thus, insofar as both types
of homopolymer are present in a single latex particle, they are present as
separate phases. For simplicity, in what follows attention will be confined
to systems for which the dispersion medium is aqueous. It is a simple matter
to generalize the discussion to systems for which the dispersion medium is
non-aqueous; it is then merely necessary to designate Ml and M2 as being
relatively lyophobic and lyophilic respectively, instead of as being relatively
hydrophobic and hydrophilic respectively.
It is possible to envisage many different morphologies for the latex particles
produced by a two-stage monomer-addition regimen in which the monomers
M land M2 are added to the reaction system sequentially. The more obvious
of these morphologies are illustrated schematically in Figs 1O.14(a)-(e). The
expected core-shell morphology is shown at (a): a core of the less hydrophilic
polymer derived from M 1 is encapsulated by a shell of the more hydrophilic
polymer derived from M 2. The inverted core-shell morphology is shown at
(b): a core of the more hydrophilic polymer derived from M2 is encapsulated
by a shell of the less hydrophilic polymer derived from MI' The morphology
shown at (c) is such that the particle is composite, but neither polymeric
region encapsulates the other. The morphology, shown at (d) is intermediate
between (a) and (c), the less hydrophilic polymeric phase having emerged
from the shell of more hydrophilic polymer to such an extent that it has just
ceased to be completely encapsulated by the shell: this might be termed the
emergent morphology. The morphology shown at (e) is such that the two
homopolymers are present as separate particles: this might be termed the
separate-particle morphology.
The theory which has been developed by Waters [78,79] attempts to
predict the morphology which will be adopted by the particles. The
assumption is that the preferred morphology will be that which minimizes
the total interfacial free energy of the system, allowing for contributions from
all the interfaces present. The criterion is therefore thermodynamic; it assumes
that equilibrium has been attained. No account is taken of the possibility
that the particles may be forced to adopt a non-equilibrium morphology
because constraints have been imposed, e.g. by the insertion of crosslinks in
one or both the polymer phases. Making these assumptions, we now consider
theoretical answers to the following questions:
(8)
(c) (d)
(e)
~ Polymer derived from less hydrophilic monomer, Mj
Polymer derived trom more hydrophilic monomer, Mz
Figure 10.14 Some possible morphologies for latex particles produced by two-stage monomer-
addition regimen in which monomers Ml Oess hydrophilic) and M2 (more hydrophilic) are
added to the reaction system sequentially: (a) expected core-shell morphology in which a core
of the less hydrophilic polymer derived from M 1 is encapsulated by a shell of the more hydrophilic
polymer derived from M 2 ; (b) inverted core-shell morphology in which a core of the more
hydrophilic polymer derived from M2 is encapsulated by a shell of the less hydrophilic polymer
derived from M 1; (c) the particle is composite but neither polymeric region encapsulates the
other; (d) emergent morphology. intermediate between (a) and (c). in which the less hydrophilic
polymeric phase has emerged from the shell of more hydrophilic polymer to such an extent
that it has just ceased to be completely encapsulated by the shell; (e) separate-particle morphology,
in which the two homopolymers are present as separate particles.
1. What condition must be fulfilled if one polymer phase is to encapsulate

the other, rather than the two phases forming as separate particles?
2. Is transition possible between the expected and inverted core-shell
morphologies as the volume ratio of the two polymer phases is varied? If
so, what conditions must be fulfilled for such transition to occur?
In what follows, y denotes interfacial free energies, with appropriate

subscripts to indicate the two phases to which the particular symbol refers.
The subscripts are 1 for the polymer derived from M l ' 2 for the polymer
derived from M 2 , and 3 for the aqueous dispersion medium. These interfacial
free energies refer to the actual interfaces which are present in the latex. They
may therefore be affected by the absorption of water into the polymer phases.
It is also assumed that the interfacial free energies of the various interfaces
are independent of the morphology of the particle(s) in which these interfaces
occur. All the polymer phases are assumed to be spherical, whether as separate
particles or as part of a core-shell particle. Particle radii are denoted by a,

with subscript c to denote core phase, and subscript c + s to denote the
complete core-shell particle. A prime is used to distinguish radii pertaining
to an inverted core-shell particle from the equivalent radii pertaining to an
expected core-shell particle. No prime is necessary with a.+ s , because this
has the same value for both types of particle. The total volume of the two
polymer phases comprising a single particle, or a single pair of particles if
separate particles form, is denoted by V. The volumes of the individual
polymer phases derived from Ml and M2 are denoted by VI and V 2
respectively. The fractions of the volume V occupied by the two copolymer
phases are denoted by 4>1 and 4>2 respectively. Thus always VI + V 2 = V and
4>1 + 4>2 = 1, but the ratio vtlv 2, = 4>tl4>2' will change as the ratio of the
volumes of the two polymer phases is varied. These notations are illustrated
in Fig. 10.15. The following simple relationships exist between the various
Expected Inverted
Separated
Figure 10.15 Illustrated notation for exposition of Waters [78, 79] theory of factors which
influence morphology of latex particles produced during emulsion copolymerization of two
monomers.
volumes, volume fractions and radii:
(10.38)
Thus
ac3 +,3
ac = a c3 + s (10.39)
Also
(10.40)
and so
(10.41)
i.e.,
(10.42)
and
a'c .1.. 1/3
'Y2
V l/3
2
~=~=vm
(10.43)
c 'YI I
Turning now to the first of the above questions, the thermodynamic

criterion shows that the expected core-shell morphology will be preferred
to the separate-particle morphology if
(10.44)
i.e., if
(10.45)
i.e., if
(10.46)
By substituting for the ratios ac+s/a c and a~/ac from equations (10.42) and
(10.43), the condition to be fulfilled can be expressed in the form
113 - 112 1 -l/J~/3 1 -l/J~/3
(10.47)
123 > l/Ji /3 ,= (1 - l/J2)2 /3
This is the condition which must be fulfilled by the various interfacial free
energies and the volume fractions of the two polymer phases if the expected
core-shell morphology is to be preferred to the separate-particle morphology.

A similar condition must be fulfilled if the inverted core-shell morphology
is to be preferred to the separate-particle morphology. The inequality (10.45)
is now replaced by the inequality
(10.48)
This then gives the following inequality, analogous to the inequality (10.47):
Y23 _ Y12 1- ,1,.2/3 1_ ,1,.2/3
'1'1 '1'1
(10.49)
Yl3 > <p~/3 ,= (1 - <PI)2/3
Two general comments can be made concerning these criteria which must
be met if a core-shell morphology, either expected or inverted, is to be
preferred to the separate-particle morphology. Firstly, they are independent
of the overall size of the particle, but are dependent upon the volume ratio
of the two polymer phases. Secondly, in both cases, the quantity which the
left-hand side of the inequality has to exceed decreases uniformly from 1 to
o as the value of the appropriate <P increases from 0 to I. Furthermore, it
can change quite rapidly with variation in <p. Figure 10.16(a) shows the way
1.0
1.0 0.8
0.6
0.4
~ 0.6
~ ~
I
-&--&-
I I ~
0.2
.,.... SO.4 ~
.,....I
00 0.2 0.8 1.0
I
0.2 ~
-&- -0.2
00 0.2 0.4 0.6 0.8 -0.4

(a) ip
-0.6
-0.8
(b) -1.0
Figure 10.16 (a) Variation of function (I - cp2 /3lJ(1 - cp)2 /3 with cp over range 0.;; cp .;; 1.
cp2 /3 - (I - cp)2 /3 with cp over range 0 .;; cp
(b) Variation of function .;; 1.
in which the function (l - cP 2/3)!(1 - cP)2/3 varies with cP over the range
o ~ cP ~ 1. As the volume fraction cP decreases, so the quantity which the
left-hand side of the inequality has to exceed increases.
Consider now, firstly, the criterion to be met if the expected core-shell
morphology is to be preferred to the separate-particle morphology. The cP
to be inserted in the function (1 - cP2/3)/(l - cP)2/3 is cPz. Then:
1. If Y23 is small relative to (Y13 - YI2)' (Y13 - YI2)!Y23 will exceed 1, and
thus the expected morphology will always be preferred over the separate-
particle morphology regardless of the value of cPz. In particular, the
expected morphology will be preferred even if the volume fraction of
hydrophilic shell is very small, so that the shell is very thin relative to the
overall size of the particle. The condition that Y23 should be small relative
to (Y13 - Y12) will probably be met if the polymer derived from the
monomer M2 is very hydrophilic, especially if much water has been
absorbed by the polymer from M 2 •
2. The condition that Y23 should be small relative to Y13 - YJ2 will certainly
be met if Y13 (the interfacial free energy between the polymer from M J
and the aqueous phase) greatly exceeds YJ2 (the interfacial free energy
between the two polymers).
3. If the values of the interfacial free energies are such that the expression
(Y13 - YJ2)!Y23 is less than 1, then there will be a lower limit to the value
of cP2 (and hence an upper limit to the value of cPI) for which the expected
morphology is preferred to the separate-particle morphology. This lower
limit to the value of cP2 arises because, if cP2 is below this limit, then the
reduction of interfacial free energy which accompanies replacement of the
1-3 interface of the separate-particle morphology by the 1-2 interface of
the expected morphology is insufficient to offset the increase in interfacial
free energy which accompanies the concomitant expansion of the 2-3
interface.
Consider, secondly, the criterion to be met if the inverted core-shell

morphology is to be preferred to the separate-particle morphology. The cP
to be inserted in the function (1 - cP2/3)/(l - cP)2/3 is now cPl. Then:
1. It will probably be the case that YI3 is greater than Y23 because the polymer
derived from M 2 is more hydrophilic than that derived from M I. Thus
(Y23 - Y12)/Y13 is likely to be less than 1, and hence there will probably
always be a lower limit to the value of cPI (and hence an upper limit to
the value of cP2) for which the inverted morphology is preferred to the
separate-particle morphology. This lower limit to the value of cPI arises
because, if cPI is below this limit, then any reduction of interfacial free
energy which accompanies replacement of the 2-3 interface ofthe separate-
particle morphology by the 1-2 interface of the inverted morphology is
insufficient to offset the increase in interfacial free energy which accompanies
the concomitant expansion of the 1-3 interface. This situation is clearly

likely to arise frequently in practice. Thus the inverted core-shell
morphology is likely to be preferred to the separate-particle morphology
only if the volume fraction of hydrophobic shell is large.
2. There is in principle no reason why 112 should not be greater than 123
for some pairs of polymers, bearing in mind that the polymer derived from
M2 is the more hydrophilic ofthe two. (123 - 112)/113 will then be negative,
and the criterion for the inverted core-shell morphology to be preferred
over the separate-particle morphology can never be met, however thick
is the shell, and however small is the core, relative to the overall size of
the particle. In this case, it is never possible for the increase in interfacial
free energy which accompanies the concomitant expansion ofthe 1-3 inter-
face to be offset by the replacement of2-3 interface by 1-2 interface, because
the latter is always accompanied by an increase in interfacial free energy.
We now consider the second of the above questions, namely, whether a
transition is possible between the expected and inverted core-shell morphologies
as the ratio of the volumes of the two polymer phases is varied, and, if so,
what conditions must be fulfilled for such a transition to occur. A transition
between the expected and inverted core-shell morphologies is possible only
if two conditions are fulfilled:
1. the inverted core-shell morphology must be a possible morphology
relative to the separated-particle morphology; and
2. as 4JI is increased, there must come a point at which the total interfacial
free energy of expected core-shell particle exceeds that of the inverted
core-shell particle.
The second of these conditions implies that
(10.50)
i.e.
e - de > 113 - Y23
a 2 2
-=----.,..-___=__
(10.51 )
a~+s 112
Using equations (10.42), this inequality can be expressed as
4J~/3 - 4J~/3, = 4J~/3 - (1 - 4JI)2 /3, > Y13 - 123 (10.52)

112
As 4J varies from 0 to 1, so 4J2 /3 - (1 - 4J)2 /3 increases from -1 to + 1. The
nature of the variation is shown in Fig. 1O.l6(b). Thus a transition between
the expected and inverted core-shell morphologies as 4JI increases is predicted
to occur when 4JI is such that the function 4Ji /3 - (1 - 4JI)2/3 just exceeds
the ratio (113 - 123)/112. This ratio is likely always to be positive, because
113 is likely to be greater than 123. However, the transition can occur only
if this ratio is < 1; this requires 112 > (113 - Y23).
The applicability of the Waters [78,79] theory to latex particles formed

by the two-stage emulsion copolymerization of methyl methacrylate and
styrene has been tested experimentally by Lee and Rudin [80]. This paper
also provides a useful list of further references to investigations of the
morphology of latex particles formed by two-stage emulsion copolymeriz-
ation.
Lee [81] has reported the results of an investigation of the morphology
of the latex particles which are formed by the two-stage emulsion polymerization
of styrene and mixtures of styrene and butadiene. The variables investigated
were the styrene/(styrene + butadiene) ratio in the mixture, the molecular
mass of the polymer (varied by adding modifier), polymer compatibility
(varied by varying the styrene/butadiene ratio in the styrene-butadiene
copolymerization mixture), polymerization sequence (styrene first and
styrene + butadiene second, or vice versa), and polymerization conditions
(monomer-starved, monomer-excess, polymerization temperature, etc.). Although
several reaction systems contained a small amount of acrylic acid (lor
2 pphm), both main monomers were hydrophobic, and neither polymer would
be expected to absorb much water from the dispersion medium. The
observations can be summarized as follows:
1. When styrene-butadiene mixtures were polymerized in the presence of

polystyrene particles, and the ratio of first stage to second was 40/60 mlm,
the second-stage styrene-butadiene copolymer phase formed as a separate
phase as microdomains dispersed within the first-stage polystyrene phase.
2. When styrene-butadiene mixtures were polymerized in the presence of
polystyrene particles, and the ratio of the first stage to the second was
reduced below 40/60m/m, the second-stage styrene-butadiene copolymer
always formed a separate phase from the first-stage polystyrene phase,
but phase inversion occurred at a stage ratio somewhere between
30/70 m/m and 20/80 m/m. Thus phase inversion may have been a
consequence of the styrene-butadiene copolymer domains becoming
close-packed.
3. Reducing the molecular mass of the polymers led to the formation of
particles which consisted of two distinct separate polymers, probably
because increased molecular mobility facilitated phase separation. If the
stage ratio was 50/50 m/m, then particles comprising two hemispherical
domains tended to form. As the ratio of styrene to styrene-butadiene was
reduced, so the styrene-butadiene copolymer began to encapsulate the
first-stage polystyrene, resulting in the formation of particles which had
the structure of an off-centred polystyrene phase embedded in a continuous
styrene-butadiene copolymer phase.
4. As expected, increasing the styrene/butadiene ratio in the styrene-butadiene
mixture reduced the incompatability of the two phases. When the ratio
had increased to 95/5 m/m, no phase separation was observed.
5. As regards the effect of changing the order of addition of the two stages,
it appears that, at least if the stage ratio was 50/50 m/m, the particles
which formed comprised microdomains of the second-stage polymer
dispersed in a continuous matrix of the first-stage polymer.
6. The morphology of the particles produced by two-stage emulsion
polymerization was found to be greatly affected by the polymerization
conditions. Two-phase structures tended to rearrange towards a more
stable structure if the polymerization conditions were such as to encourage
the mobility of the polymer segments, e.g. the polymerization temperature
was high and excess monomer was present. Conversely, polymerization
conditions which tended to discourage the mobility of the polymer
segments, such as low temperature and monomer starvation, tended to
immobilize two-phase particle structures.
10.2.4.6 Preparation of large-particle and 'monodisperse' synthetic latices

The best-known method for preparing large-particle synthetic latices is seeded
emulsion polymerization, in which the particles produced by each successive
stage of the reaction are used to provide the reaction loci for the next stage.
It is generally necessary to add sufficient additional colloid stabilizer at each
stage just to ensure adequate colloid stability; but the amount added must
be insufficient to permit the nucleation of further reaction loci. It may also
be necessary to add further initiator. It so happens that the distribution of
particle sizes tends to narrow at each successive stage of seed polymerization.
The particle-size distribution in the latex obtained from the first stage is
usually quite narrow, and thus the seeding technique produces synthetic
latices which not only contain large particles but are also of very narrow
particle-size distribution and thus effectively monodisperse.
The process of seeded emulsion polymerization has been discussed by
Vanderhoff et al. [82] and by Bradford, Vanderhoff and Alfrey [83]. Let v
and a denote respectively the volume and radius of the polymer particle
which would be formed from a reaction locus at time t if all the unpolymerized
monomer were removed. Assume that the rate ofincrease of v is an exponential
function of a. Then
dv = ka" (10.53)
dt
where k and n are constants which may depend upon the reaction system.
If the particles are spherical, then v = 4na 3 /3 and hence dv/dt = 4na 2 • da/dt.
Equating these two expressions for dv/dt gives
1 da k 11-3
--=-a (10.54)
a dt 4n
Thus the relative rate of growth of particle radius is directly proportional to

an- 3. If n = 3, then this relative rate is independent of particle size. da/dt is
then directly proportional to particle radi us. Th us if particle radi us (or, indeed,
any linear dimension) is taken as the measure of particle size, then the nature
of the distribution of particle sizes remains unaltered throughout the growth
period. If n < 3, then the radii of the smaller particles increases relatively
faster than do those of the larger particles, and so tend to catch them up in
size. The distribution of sizes then narrows as growth proceeds. If n > 3, then
the distribution broadens as growth proceeds. If, as is assumed in the particle-
nucleation theory of Smith and Ewart [29], n = 0 because dv/dt is assumed
to be constant, then the particle-size distribution should sharpen very rapidly
as growth proceeds. Experimental results reported by Vanderhoff et al. for
polymerizations using polystyrene latex are consistent with n ~ 2.5. This
value indicates that the particles in seeded emulsion polymerizations should
narrow in particle-size distribution, but not as rapidly as predicted by the
Smith-Ewart theory. The experimental evidence upon which the value
n ~ 2.5 is based becomes evident if equation (10.54) is integrated. Writing k
as k(t) to emphasize possible dependence upon time ofreaction, this equation
gives
a I
f a 2 - nda = 41n f k(t)dt (10.55)

o
where ao and a are the values of a at t = 0 and t = t respectively. Thus
I
ea3-n _ a~-n = 3 - n fk(t)dt (10.56)

4n
o
Suppose, now, that the emulsion polymerization of excess monomer is allowed
to occur in the presence of two monodisperse seed latices which contain
particles of initial unswollen radii a o and a~, the conditions of concentration,
temperature, etc., being the same for both. Whatever may be the actual value
of k(t) at any instant in the reaction, its value for both of the growing latices
should be the same. Thus, if the growth period is the same for the two latices,
the integral on the right-hand side of equation (10.56) should have the same
value for both. So if a and a' are the respective unswollen particle radii at
the end of the growth period, then
a 3-n - a o3-n = a13-n - a13-n
o (10.57)
so that a 3 - n - a~ - n has the same value for both species of particle at all
stages of particle growth. Putting Q( = a~/ao and P = a/a o gives
~
I
= _1 (Q(3-n + p3-n _ 1)1/(3-n) (10.58)
a p
Thus the value of n can be estimated from an experiment in which an emulsion

polymerization reaction is seeded with monodisperse particles of two different
initial sizes, and these particles are then allowed to compete with each other
for excess monomer and free radicals in the same reaction system. As particle
growth proceeds, the ratio a'ia is determined by electron microscopy, and
its variation with p (which is a measure of the extent of polymerization)
ascertained. It is then possible to compare the observed variation of a'ia with
pfor given (X with that predicted assuming various values of n. Having carried
out an experiment of this type, Vanderhoff et al. concluded that their results
were reasonably consistent with n ~ 2.5. No allowance appears to have been
made for the effect of particle size upon the distribution of absorbed monomer
within the reaction system, and for possible implications for the relative
growth rates of the two sizes of particles.
Serafini and Bobalek [84] have investigated the use of lithium carboxylate
soaps as colloid stabilizers in reaction systems in which control can be
exercised over the particle size of styrene-butadiene copolymer latices
produced by emulsion polymerization. They showed that, although lithium
carboxylate soaps form micelles in aqueous solution, the solubility only just
exceeds the critical micelle concentration, especially in the case of lithium
stearate. A consequence of this is that lithium carboxylate soaps have limited
capacity for effecting micellar nucleation of reaction loci. The solubility
properties of lithium stearate in particular appear to be such as to provide
a self-regulating mechanism for control of micellar nucleation. Serafini and
Bobalek have reported that uniform latex particles can be produced in
emulsion polymerization reaction systems if the nucleation of reaction loci
during Interval I is restricted by the use of lithium carboxylates as colloid
stabilizer, even if the quantities used are sufficient to confer adequate colloid
stability upon the resultant latex. The number of soap micelles can be
controlled by varying the concentration of added lithium ions, and hence so
too can the number of reaction loci nucleated, the rate of polymerization,
and the ultimate particle size of the latex formed.
10.2.4.7 Problem of formation of non-latex polymer during emulsion
polymerization
There are various forms in which polymer can be present in emulsion
polymerization reaction systems, other than as colloidally-stable latex
particles. Polymer present in such forms is collectively known as non-latex
polymer; the term coagulum is also commonly used to denote such polymer.
The forms in which such polymer can be present are usually known by terms
which describe them appropriately, such as filterable solids, buttons, reactor
fouling, sediment, silt, grit, sand, etc. The formation of non-latex polymer can
be a serious industrial problem for several reasons. The yield of useful polymer
is decreased. Cleaning of the reactor between successive batches of reaction
system may be necessitated; as well as incurring additional expense in itself,
the cycle time for the reaction is increased and the output of the reactor
correspondingly decreased. The variability of the latex between batches may
be increased. In severe cases, damage may be caused to the reactor. Other
problems include the necessity of disposing of any non-latex polymer which
forms, e.g., by incineration, and the probability that the non-latex polymer
will absorb some residual unreacted monomer. The latter can present health
hazards in respect of the handling and disposal of the non-latex polymer.
Notwithstanding the industrial importance of the formation of non-latex
polymer during emulsion polymerization reactions, there appear to be
comparatively few references to this matter in the technical literature.
Vanderhoff [85] has given a general discussion of the problem. It appears
that the mechanisms whereby non-latex polymer can be formed in emulsion-
polymerization reactors are understood in only the most general terms.
Vanderhoff distinguishes three main types of non-latex polymer as follows:
1. polymer which is formed during the polymerization reaction, and which
can be removed from the eventual latex afterwards by sedimentation/creaming
or by filtration;
2. polymer which is deposited upon the various internal reactor surfaces
during polymerization, the principal of such surfaces being those of the
reactor walls, roof and bottom, those of inserts such as cooling coils,
baffles and thermometer wells, and those of the reactor agitator(s);
3. non-latex polymer which separates from the latex subsequent to
polymerization, during storage and transportation.
The concern here is with types (1) and (2) non-polymer latex. Vanderhoff
also distinguishes two fundamentally different types of mechanism by which
non-latex polymer can form in emulsion polymerization reaction systems.
These are formation by colloidal destabilization of latex particles which have
formed by emulsion polymerization, and formation by polymerization
processes other than emulsion polymerization.
The colloid stability of the latex particles at the various stages of an
emulsion polymerization reaction is determined by the balance between
various of the factors which are discussed in Chapters 3, 4 and 5 (Volume
1). The important factors which reduce the colloid stability of the latex
particles in these reaction systems as polymerization proceeds can be
summarized as follows:
1. increase in the aggregate area of interface between particles and dispersion
medium, such increase being partly a consequence of polymerization and
partly of the concomitant absorption of unreacted monomer;
2. formation of lyophilic polymers in the dispersion medium of the system,
such polymers then reducing colloid stability by interparticle bridging;
3. mechanical destabilization brought about by the agitation which is
necessary to maintain the reaction system homogeneous and to facilitate
dissipation of the heat of polymerization.
The second of these factors is likely to be important only in reaction systems

for the production of copolymers for which at least one of the comonomers
is lyophilic. The most common reaction systems of this type are aqueous
emulsion polymerizations for the production of the latices of hydrophobic
polymers which are being functionalized by the inclusion of minor proportions
of hydrophilic comonomers (see further section to.3 below). The primary
driving force promoting colloidal destabilization during the early stages of
emulsion polymerizations when the particles are small is thought to be
Brownian encounters between particles. As the size of the particles increases,
so agitation-induced motion of the particles become increasingly important.
Furthermore, as particles flocculate, they become increasingly prone to further
colloidal destabilization under the influence of agitation-induced motion. For
this reason, any inherent colloid instability in a latex as it forms by emulsion
polymerization can lead to a stage of auto accelerated colloidal destabilization
as polymerization proceeds.
Important factors which enhance the colloid stability of a latex as it forms
during emulsion polymerization include:
t. the amount and type of colloid stabilizer which was present in the reaction
system initially, and the amounts which are added incrementally as
polymerization proceeds;
2. the increasing presence of lyophilic groups bound at the surface of latex
particles as they grow.
As regards the second of these factors, the lyophilic groups may be derived
either from the polymerization initiator or from any lyophilic comonomers
which are present in the reaction system. Again, these effects are best
illustrated by aqueous emulsion polymerizations for the production of the
latices of hydrophobic polymers which are being functionalized by the
inclusion of minor proportions of hydrophilic comonomers. The initiator for
such reactions is usually hydrophilic in nature. Incorporation in the growing
polymer molecules usually leads to some at least of the combined initiator
fragments becoming located at the particle surface, thereby increasing the
hydrophilicity of that surface. If, as is usual, the initiator is anionic as well
as being hydrophilic, then colloid stability is also enhanced by the concomitant
increase ofthe density of surface-bound electric charge. Combined hydrophilic
comonomers can enhance colloid stability in at least two ways. Firstly, such
comonomer units which have become located at the particle surface enhance
colloid stability in the same manner as do combined initiator fragments.
Secondly, they may enhance colloid stability if they have become combined
in the form of hydrophilic macromolecules which have a tendency to adsorb
at the particle surface, thereby conferring steric stabilization upon the
particles. Enhancement of colloid stability by both these mechanisms is likely
to be increased if, as is usual, the hydrophilic comonomer is ionogenic and
is at least partially ionized in the reaction system. It follows from the discussion
in this and the preceding paragraphs that the effects of lyophilic comonomers
upon the colloid stability of a latex as it forms during emulsion polymerization
can be conflicting: insofar as the lyophilic groups become combined at the
particle surface and in the form of adsorbable lyophilic macromolecules which
confer colloid stability upon the particles, colloid stability is enhanced; but
insofar as the lyophilic groups become combined in the form of adsorbable
lyophilic macromolecules capable of promoting interparticle bridging, colloid
stability is reduced.
An experimental investigation of the factors which affect the formation of
non-latex polymer during the semi-continuous emulsion polymerization of
acrylic monomers has been reported by Snuparek [86]. The colloid stability
of the reaction system was greatly affected by the way in which the colloid
stabilizer was distributed between the initial charge to the reactor and the
monomer emulsion which was added subsequently. This investigation also
demonstrates clearly the effect of copolymerizing with ionogenic monomers,
as does a second investigation reported by Snuparek and Tutalkova [87].
The most obvious mechanism, alternative to emulsion polymerization, by
which polymer can form in an emulsion polymerization reaction system is
bulk polymerization within monomer droplets, or within a separated
monomer layer. Polymerization within small emulsified monomer droplets,
initiated by the occasional free radicals which may find their way into those
droplets, is the most likely origin of the gritty particles of non-latex polymer
which are sometimes found in synthetic latices. The non-latex polymer which
becomes deposited upon the various surfaces within the reactor may be
formed by polymerization within large monomer droplets, or within layers
of monomer which are coating these surfaces. According to Vanderhoff [85],
polymerization of monomer in the vapour phase above the reaction system
can also lead to the formation of non-latex polymer, in particular, on the
internal surface of the reactor roof. The tendency for non-latex polymer to
form on reactor surfaces depends upon the smoothness of those surfaces: the
smoother the surface, the less is the tendency for polymer to form on it. For
this reason, glass-lined reactors are found to be less prone to the formation
of surface polymer than are reactors for which the inner surface is stainless
steel, especially if the latter is not highly polished. The formation of surface
polymer during emulsion polymerization is facilitated by surface imperfections,
such as fractures in glass linings and scratches in stainless-steel surfaces.
The above considerations provide very general guides to certain obvious
steps which can be taken in attempting to minimize or eliminate the formation
of non-latex polymer during emulsion polymerization. Attention should
clearly be paid to the amount of colloid stabilizer present in the reaction
system, and also to the manner in which it is added. The rate of agitation
should at all stages of the reaction be the minimum necessary to ensure that
the reaction system remains homogeneous, and that heat is efficiently
transferred from the reaction system. If possible, the rate of agitation should
be reduced as polymerization proceeds and the particles increase in size. The

design ofthe reactor, and of the agitation and baffle system within the reactor,
should be such as to minimize conditions of shear and turbulence which
might tend to colloidal destabilization. It is also recommended that the
temperature-time profile within the reaction system be controlled as rigorously
as possible: in particular, rapid rises in temperature associated with reaction
exotherms should be avoided. There are several variations in polymerization
technique which can be explored in attempting to minimize the formation
of non-latex polymer. A seeded reaction system can be substituted for an ab
initio system. The mode of addition of monomer can be varied. Incremental
or continuous additions of monomer can be made below the surface of the
reaction system, instead of by precipitation through the vapour space above
the reaction system. The process can be changed from batch to semi-continuous,
or from semi-continuous to continuous. However, Vanderhoff [85J has
emphasized that there are no general rules concerning these matters. Each
reaction system has to be considered individually. In the long run, the most
fruitful approach to this problem is likely to derive from a detailed
understanding of the particular reaction system, and of the ways in which
non-latex polymer can form in that reaction system.
There is a curious phenomenon which sometimes arises in connection with
the presence of non-latex polymer particles in a synthetic latex, and which,
as far as is known has never been satisfactorily explained. This is that the
removal of such particles from an apparently otherwise colloidally-stable
latex by filtration sometimes produces a latex from which further non-latex
polymer particles tend to separate during storage. It might be thought that
such polymer as was going to separate from the latex in this form would
have done so during production and during the initial post-polymerization
storage period. It might also be thought that the colloid stability of the latex
after filtration would be greater than that of the latex as it formed, because
the ratio of colloid stabilizer to polymer was increased to the extent that
polymer had been removed. Neither of these assumptions appears always to
be borne out in practice.
10.3 PREPARATION OF FUNCTIONALIZED LATICES BY EMULSION

COPOLYMERIZATION
10.3.1 Introduction
As has been noted in section 1.6.2 of Chapter 1.5.4 (Volume 1), functionalized
latices are latices which contain polymers which have been functionalized by
the incorporation of minor amounts of specific reactive chemical groupings.
The intention is usually that the functional groups attached to the polymers
dispersed in these latices are distributed in essentially random fashion along
Preparation offunctionalized latices 239
the polymer chains, although the degree of functionalization may differ

amongst the various polymer molecules in a given latex particle. Random
distribution offunctional groups within a given polymer molecule is achieved
by adopting reaction conditions and procedures such that the units which
bear the functional groups become copolymerized essentially at random with
the other monomer units in the polymer. However, this is not the case if the
emulsion copolymerization procedure is such that the monomer units bearing
the functional groups are constrained to enter the polymer in a markedly
non-random manner, e.g., because the functionalizing monomer is withheld
from the reaction system until near to the end of the polymerization reaction,
or because the reaction conditions are such that the functionalizing comonomer
partitions strongly in favour of the dispersion medium rather than the reaction
loci, and so tends to homopolymerize in the dispersion medium if it
polymerizes at all. The latter possibility arises specifically in connection with
the preparation of aqueous carboxylated latices if the emulsion polymerization
is carried out under alkaline conditions; further reference is made to this
matter in section 10.3.3.1 below. Cognate to the matter of the distribution
of functionalizing groups within individual polymer molecules is that of the
distribution ofthese groups within individual latex particles. If the functionalizing
comonomer is markedly more lyophilic than are the main monomers, as is
so for aqueous carboxylated latices of essentially non-polar polymers, then
there is a thermodynamic tendency for the functional groups to concentrate
at or near the particle surface. If this is desirable, then the polymerization
procedure may be designed to encourage this tendency by withholding the
functionalizing comonomer until the later stages of the reaction. As noted
above, in extreme cases where the functionalizing comonomer is withheld
until virtually the end of the polymerization, the distribution of the functional
groups amongst the other monomer units is expected to be very markedly
non-random. The important matter of the distribution of carboxylic-acid
groups in aqueous carboxylated latices of non-polar polymers is considered
further in section 10.3.4 below.
Although, in principle, non-aqueous synthetic latices can be prepared in
functionalized variants, industrial interest has been primarily in aqueous
functionalized latices. The following discussion is therefore restricted to this
type. Furthermore, by far the most important chemical grouping which is
used to functionalize the polymers contained in synthetic latices is the carboxylic-
acid moiety. Latices which contain polymers modified by the inclusion of a
small proportion of pendant carboxylic-acid groups are known as carboxylated
latices. The concern here will be almost exclusively with this type of
functionalized latex. Other types of chemical grouping which are incorporated
in the polymers contained in functionalized latices include sulphonate,
hydroxyl and amino moieties. The proportion of copolymerized functionalized
comonomer (not the functionalizing group itself) is frequently in the range
O.5-5.0pphm, but larger amounts can be incorporated. As also noted in
section 1.6 of Chapter 1.6.2 (Volume 1), the development of functionalized

latices had its origin in the realization that benefits could in principle accrue
both to the polymer dispersed in the latex and to the latex itself if small
numbers of reactive chemical groupings were chemically-combined with the
polymer. Functionalized latices are at present invariably produced by emulsion
polymerization. Thus they inevitably fall into the class of synthetic latices.
Functionalized variants of many types of synthetic latices, containing both
rubbery and glassy polymers, have been, and are currently, produced industrially.
Of particular industrial importance are carboxylated styrene-butadiene rubber
latices; much of what follows is directed specifically at this type.
The most important of the many comonomers which are available for
introducing carboxylic-acid groups into latex polymers are acrylic acid
(Structure XV), methacrylic acid (Structure XVI), maleic (cis-I,2-ethene-
dicarboxylic) acid (Structure XVII), fumaric (trans-I,2-ethenedicarboxylic)
acid (Structure XVIII), and itaconic (2,3-propenedicarboxylic) acid (Structure
XIX). The half-esters of dicarboxylic acids with mono- and dihydric alcohols
are also used; examples include monoethyl maleate (Structure XX) and
monopropyleneglycol maleate (Structure XXI) respectively. Carboxylating
acids of this last type give polymers which contain hydroxyl groups as well
as carboxylic-acid groups. Comonomers for introducing respectively sulphonate,
hydroxyl and amino groups include 2-sulphoethyl methacrylate (Structure
XXII), hydroxyethyl methacrylate (Structure XXIII), and 2-aminoethyl
methacrylate (Structure XXIV). There is an extensive technical literature
concerning the production and properties of functionalized aqueous synthetic
latices, especially carboxylated synthetic latices. A review of the production
of carboxylated latices has been given by Blackley [88]; much of what follows
is based upon that review.
CH 2 =CH.C0 2 H CH 2=C(CH 3). C0 2 H
XV XVI
H0 2e. CH=CH. C0 2 H H0 2 e. CH=CH. C0 2H
cis trans
XVII XVIII
CH 2 =C(C0 2 H)CH 2 • C0 2 H
XIX
H0 2 e. CH=CH. C0 2 CH 2 CH(OH)CH 3
cis cis
XX XXI
CH 2=C(CH 3). C0 2 CH 2CH 2 • S03H CH 2=C(CH 3). C0 2 CH 2 CH 20H
XXII XXIII
CH 2=C(CH 3)· C0 2CH 2CH 2 NH 2
XXIV
Preparation ojJunctionalized latices 241
10.3.2 Advantages and disadvantages which accrue from the functionalization

of synthetic latices
As noted in section 10.3. t above, the industrial advantages which accrue
from the functionalization of synthetic latices fall under two broad headings:
those which pertain to the latex itself, and those which accrue from
modification of the properties of the polymer which is dispersed in the latex.
The following discussion of these advantages refers specifically to carboxylated
latices.
The principal advantages which pertain to the latex itself are attributable
to enhanced colloid stability. This is partly a consequence of the types of
colloid stabilizer which are used in emulsion polymerization reaction systems
for the production of these latices, and partly of the presence of combined
functionalizing groups at the surface of the latex particles. The likely effects
of combined hydrophilic comonomers upon the colloid stability of the
resultant latex have been outlined in section 10.2.4.7 above. It has been noted
in section 4.6.2.3. t of Chapter 4 (Volume t), and also in section 10.2.3.2 of
this chapter, that carboxylate anions which are covalently bonded to the
surface of latex particles are more effective in conferring colloid stability
upon a latex than are the same number of carboxylate anions anchored to
the surface by adsorption. The same is probably true for other types of ionized
functionalizing group. Whatever may be the reason for enhancement of colloid
stability by functionalizing comonomers, it is important industrially. It is
manifest by attributes such as increased mechanical and chemical stability,
and increased tolerance to the addition of large quantities of mineral fillers.
It is reasonable to suppose that, if enhancement of colloid stability by the
presence of covalently-bonded carboxylic-acid groups at the particle surface
is desirable, then it is advantageous that the carboxylic-acid monomers and
the polymerization procedure be such that the concentration of combined
carboxylic-acid groups in the surface regions of the particle should considerably
exceed the average concentration of those groups in the particle as a whole.
Insofar as experimental evidence is available, this supposition appears to be
borne out in practice. An extreme manifestation of enhanced colloid stability
is to be found in synthetic latices which are claimed to dry to deposits which
are capable of being re-dispersed in water, the resultant dispersions having
essentially the same particle size as did the initial latices. Synthetic latices of
this type are described in section 11.9.5 of Chapter 11.
Under the second of the above headings is included advantages such as
susceptibility to reaction with additional reagents, improved adhesion to
polar substrates such as textile fibres, paper fibres and metals, enhanced
polymer tensile strength, and, to a lesser extent, improved resistance to
swelling in hydrocarbon oils. Amongst the additional reagents which are
available for reaction with the polymers contained in carboxylated latices
are metal oxides, polyamines, and epoxides. These reactions are discussed in
section 11.8.2.2 of Chapter 11. They are used to effect crosslinking of the
polymers. Reaction of the carboxylic-acid groups with reagents dissolved in
the aqueous phase of the latex is facilitated by location of the carboxylic-acid
groups near the surface of the particles. Improved adhesion of the particles,
as distinct from the bulk polymer, to fibrous substrates should also be
facilitated by location of the carboxylic-acid groups near the particle surface.
If such enhancement of adhesion is desirable, it is again advantageous that
the surface concentration of functional groups should be markedly greater
than in the particle as a whole.
The carboxylic-acid groups in carboxylated latices are also, of course,
susceptible to neutralization with any strong alkali, the corresponding
carboxylate anion being thereby formed. Use is made of this effect in a special
type of carboxylated latex which contains polymer carboxylated to a much
greater extent than is normal for the common applications. These latices
have the property that they undergo pronounced increase in viscosity when
the pH is raised above ca. 8. They are therefore known as alkali-thickenable
latices. They are considered further in section 11.8.3 of Chapter 11. The
property of marked alkali-thickenability is attributed to pronounced swelling
of the latex particles by the water of the aqueous phase, and also to possible
dissolution of the more heavily-carboxylated polymer molecules in the
aqueous phase, both effects being consequences of ionization of the
carboxylic-acid groups of the polymer.
The principal disadvantage which accrues from carboxylating non-polar
polymers in latex form is that the intrinsic water-sensitivity of the polymers
themselves is increased. Whether or not this is reflected in an increased
tendency for films derived from the latices to absorb water depends upon
other factors, such as the levels and types of other hydrophilic substances
present in the latex. If films from a carboxylated latex do show an increased
tendency to absorb water, as compared with films from similar non-carboxylated
latices, then this is partly a consequence of the presence of carboxylic-acid
groups in the polymer, and partly of the levels and types of colloid stabilizer
which are used for the production of the latices (see section 10.3.3.2 below).
If significant amounts of un polymerized carboxylic-acid monomer are present
in the latex, then the water absorption of films derived from the latex will
be expected to be further enhanced; however, it is not usually expected that
significant amounts of unpolymerized acid will be present. Of these factors,
the most important for latices for which the level of carboxylation is low is
probably the level and type of colloid stabilizer. As expected, water-absorption
is further increased if the carboxylic-acid groups are present in the polymer
as carboxylate anions rather than as un-ionized carboxylic-acid groups. Thus
the presence of alkalinity in the water tends to encourage water-absorption.
In some circumstances, the amounts of water absorbed into films derived
from carboxylated latices can be very large. It is probable that films from
most other types of functionalized aqueous latex tend to show increased
sensitivity to water, especially if the functionalizing comonomer is polar, and

especially if it is ionized.
The levels of functionalization commonly introduced into functionalized
latices of industrial interest are not usually sufficient to affect greatly the
glass-transition temperature of the dispersed polymer. In particular, the
glass-transition temperature of the polymers present in carboxylated
styrene-butadiene rubber latices is usually essentially that determined by
the styrene/butadiene ratio.
10.3.3 Production of carboxylated latices by emulsion polymerization
10.3.3.1 General procedure

There are two particular general features of emulsion polymerization reaction
systems for the production of carboxylated rubber latices which call for
comment. The first is that the polymerization usually takes place under acidic
conditions (ca. pH 3-4), and not under the alkaline conditions which are
usual for the production of many types of non-functionalized synthetic latices.
Polymerization is carried out under acidic conditions to encourage the
carboxylic-acid monomer to become copolymerized in the main polymer
molecules which are forming. If the reaction is carried out under alkaline
conditions, then the carboxylic-acid monomer is present mainly as the
carboxylate salt. As such, it partitions strongly in favour of the aqueous
phase. If it polymerizes at all under these conditions, polymerization occurs
mainly in the aqueous phase, and the polymer molecules in which it becomes
incorporated are far more hydrophilic than are the majority of the polymer
molecules being formed in the reaction system. For these reasons, it is
desirable that the pH of the reaction system should be such that most of the
carboxylic-acid monomer is present as carboxylic acid, and not as carboxylate
saIto This requires that the pH should be at least one unit less than the pKa
of the acid. The value of Ka for ionization of the (first if there is more than
one) carboxylic-acid group of the common carboxylic-acid monomers is ca.
10- 5 mol dm - 3. For this reason, these emulsion copolymerizations are carried
out at pH ca. 3-4 or below. There is usually some increase in pH as the
carboxylic-acid monomer becomes incorporated in the polymer. This is partly
because carboxylic-acid groups are removed from the aqueous phase into
the polymer particles, and partly because carboxylic-acid groups attached
to macromolecules have lower acid strength than do carboxylic acids of low
molecular mass. Typically the pH of the reaction system at the end of the
reaction is ca. 4.5-5.5.
The second general feature is that, in contrast to the production of, say,
solid styrene- butadiene rubber by emulsion copolymerization, where it is
common practice to short-stop the reaction at ca. 65% conversion in order
to prevent undue branching and crosslinking of the polymer, reaction systems

for the production of carboxylated latices by emulsion polymerization are
usually taken as near to complete conversion as possible. An inevitable
consequence, for carboxylated styrene-butadiene rubber latices at least, is
that the polymer dispersed in the resultant latex is highly branched and
crosslinked. Notwithstanding that modifiers may be used to mitigate this
problem somewhat, the gel content of such polymers can be virtually 100%,
and the gel itself very tightly crosslinked.
10.3.3.2 Colloid stabilizers

Surface-active substances of medium molecular mass are commonly used as
colloid stabilizers for the production of carboxylated latices by emulsion
polymerization. Regardless of the type which is used, the requirement that
the emulsion polymerization reaction should take place under acid conditions
imposes constraints upon the choice of colloid stabilizer, in that it has to be
stable to acid conditions. In practice, this implies that the colloid stabilizer
should be either non-ionogenic, e.g. of the ethoxylate type, or anionic, the
anionic moiety being derived from a strong acid. The most important colloid
stabilizers for the production of carboxylated latices fall within this latter
category. The most widely-used type comprises the alkylarylsulphonates;
probably the most widely-used is sodium dodecylbenzenesulphonate. Surface-
active sulphates, such as sodium n-dodecyl sulphate are also used, but, as
noted in section 10.2.3.2 above, suffer from the disadvantage of being more
prone to slow hydrolysis under acid conditions than are the alkylaryl-
sui phonates. The hydrolysis reaction can be represented as follows:
RO.SO; + H 20 ------> ROH + HO.SO;

The consequence of this reaction is that the surface-activity of the sulphate
anion is lost. An analogous hydrolysis reaction does not occur so readily
with the sui phonates. For this reason, acidic emulsion polymerization
reactions which use sulphates as the principal colloid stabilizer tend to be
more variable in behaviour than do similar reactions using sui phonates.
Regardless of the nature of the principal colloid stabilizer, it is common
practice to add small amounts of secondary surface-active substances, such
as sodium methylene-bis(naphthalenesulphonate), the purpose being to
improve the colloid stability of the latex produced.
There is much evidence to indicate that the performance of nominally-
identical alkylarylsulphonates as colloid stabilizers for the production of
carboxylated latices by emulsion polymerization can be considerably affected
by the nature and distribution of alkyl chain lengths in the hydrocarbon
moiety. Sodium dodecylbenzenesulphonate has been available over the years
in at least two different forms, depending upon the origin of the alkyl
substituent. Formerly, the alkyl substituent was derived from so-called

'tetrapropylene'. The latter is a highly-branched olefin fraction obtained by
catalytic oligomerization of propylene, the product being fractionated so that
the average number of carbon atoms per molecule is 12. More recently, the
alkyl substituent has been straight-chain, in order that the alkylaryl-
sulphonate molecule shall be biodegradable. Even so, many variations in
molecular structure are possible. Karsa [89J has reported the effects of certain
structural variables in respect of the emulsion homopolymerization of styrene
and the emulsion polymerization of styrene and butadiene using various
industrial grades of straight-chain sodium dodecylbenzenesulphonate as the
colloid stabilizer. It is not clear whether or not the reaction systems for the
emulsion copolymerization of styrene and butadiene contained carboxylic-
acid monomer. However, regardless of whether they did or did not, it seems
likely that the observations reported by Karsa apply in a general way to
styrene-butadiene emulsion copolymerization reaction systems in which
carboxylic-acid monomers are present; in any event, the principal motivation
for carrying out this investigation derived from the widespread use of
alkylarylsulphonates for the production of carboxylated styrene-butadiene
copolymer latices. Industrial grades of so-called sodium dodecylbenzenesul-
phonate of all types inevitably contain minor, but significant, amounts of
impurities, such as unsulphonated alkylbenzene, sodium sulphate, and
sulphonated derivatives of other hydrocarbons. However, Karsa considers
that, at least for the straight-chain sulphonates, such impurities have less
effect upon performance in emulsion polymerization reactions than do
variations which arise from differences in the overall number of carbon atoms
in the alkyl substituent (which, of course, affects the molecular mass),
differences in the position in the alkyl moiety at which the phenyl group is
substituted (which affects the isomer distribution of alkyl substituents in the
benzene ring for a given molecular mass of the alkylarylsulphonate), and
differences in the ratio of 0- to p-substitution of the sulphonate group in the
molecule. As regards average molecular mass, this ranged from 235 to
256 daltons for the benzene alkylates from which the sulphonates investigated
by Karsa were produced; the value for the alkylate corresponding to pure
sodium dodecylbenzenesulphonate is ca. 246 daltons. As regards the o-/p-ratio,
this is affected by the size and number of the alkyl substituents at the I-position
of the benzene ring. It is therefore affected by the position at which the phenyl
group is substituted in the alkyl chain. 2- and 3-phenylalkanes, which are
termed end isomers, favour an increased o-/p-ratio relative to 4-, 5- and
6-phenylalkanes, which are termed internal or middle isomers. The reason is
steric hindrance at the o-position when more than one large alkyl group is
attached to the carbon atom at which the phenyl group is substituted. The
important conclusions reached by Karsa as regards performance in
styrene-butadiene emulsion copolymerization reaction systems can be
summarized as follows:
1. The rate of polymerization decreases as the average molecular mass of

the sodium dodecylbenzenesulphonate increases.
2. The amount of small-particle flocculum decreases as the average molecular
mass of the sodium dodecylbenzenesulphonate increases.
3. The amount of tOO-mesh coagulum is very high if the average molecular
mass of the sodium dodecylbenzenesulphonate is high, but there is no
trend if the average molecular mass is lower.
4. The amount of large-particle coagulum (> 10 Jlm) decreases as the average
molecular mass of the sodium dodecylbenzenesulphonate increases.
5. Isomer distribution affects performance as a colloid stabilizer in this
reaction system, but the trends are not clear.
The overall conclusion from this investigation is that industrial grades of

sodium dodecylbenzenesulphonate are complex mixtures of different molecular
species, the composition of which can significantly affect its performance as
a colloid stabilizer in emulsion polymerization reactions. It is therefore hardly
surprising that nominally-identical products from different suppliers are not
always interchangeable.
Dialkylsulphosuccinates and N-alkylsulphosuccinamates are also used as
colloid stabilizers in emulsion polymerization reaction systems for the
production of carboxylated latices, particularly when special properties are
desired. The chemical structures of these types of compound are given in
section t 6.7.3.3.4 of Chapter t 6 (Volume 3). The N-alkylsulphosuccinamates
are effectively mixed sulphonate-carboxylates.
The levels of colloid stabilizer used in emulsion polymerization reaction
systems for the production of carboxylated latices vary widely. The upper
limit is ca. 5 pphm. Over the years, the tendency has been to reduce the level
progressively; in some reaction systems, the level of the principal stabilizer
can be as low as 0.5 pphm. Important reasons for reducing the level of colloid
stabilizer include a desire to reduce the foaming tendency of the latex, and
to reduce the water-sensitivity of films derived from the latex.
10.3.3.3 Initiators
Potassium, ammonium and sodium peroxodisulphates are the initiators
normally used for the production of carboxylated styrene- butadiene copolymer
latices. Polymerization is usually effected at 'high' temperatures, e.g. SO-80°e.
It is not usual to effect polymerization at 'low' temperatures, e.g. 5°C, using
redox initiators.
10.3.3.4 Modifiers
The compounds which have been most commonly used as modifiers for the
production of carboxylated styrene- butadiene rubber latices are the straight-
and branched-chain alkyl mercaptans of average alkyl chain length ca. 12,
although many other compounds have been investigated for this purpose.
Levels are typically ca. 0.5 pphm. The mode of addition of the modifier is
important, because it is usually desirable that the extent of modification
should be as uniform as possible throughout the polymerization. It is therefore
desirable to add the modifier incrementally or continuously. Some compromise
is necessary as regards the level of polymer modification. High polymer
molecular mass, with some degree of crosslinking, is desirable for the
attainment of good mechanical properties such as strength; over-modification
leads to the formation of polymers which are weak and soft. On the other
hand, high molecular mass and degree of branching and crosslin king are
detrimental to integration of the particles as the latex dries. In some reaction
systems, it appears that a useful compromise is achieved by having present
a relatively high level of modifier together with a small amount of a
crosslinking monomer such as divinylbenzene.
10.3.3.5 Electrolytes
Inorganic electrolytes are usually added to emulsion polymerization reaction
systems for the production of carboxylated latices. The cations of these
electrolytes are almost invariably monovalent, being usually either sodium
or potassium. The anions are more varied. They include chloride, sulphate,
phosphate, acetate, carbonate and bicarbonate. If the added electrolyte is
alkaline, it reacts with part at least ofthe carboxylic-acid monomer to produce
the corresponding salt. The amounts of alkaline electrolytes added must
therefore not be such as to raise the pH of the aqueous phase of the reaction
system above ca. 4. The total level of added electrolytes is usually ca. t pphm.
The effect of the added electrolyte is to increase the ionic strength of the
aqueous phase of the reaction system, and thereby to bring about an increase
in the particle size and particle-size distribution of the latex which is produced.
One consequence of this is that the rate of polymerization is reduced to some
extent; another is that the product of the reaction has a more advantageous
viscosity-concentration relationship during the subsequent latex-concentration
step than if the electrolyte is omitted (see further section to.4.2.2.2 below).
The precise effects which are observed in these respects depend to some extent
upon the particular electrolytes which are added. Correlations between the
nature of an electrolyte and its effect in these reaction systems are not well
understood at present; the determination of suitable electrolyte additions is
essentially a matter for empirical experimentation.
10.3.3.6 Post-polymerization additives

It is usual to increase the pH of carboxylated latices from the typical
post-reaction value of ca. 5 to ca. 9 before supplying the latex to the user.
Raising the pH has several desirable effects: the colloid stability of the latex
is enhanced, making the latex more suitable for subsequent processing, and
the tendency to corrode metals with which the latex may come into contact
is reduced. Some increase in the viscosity of the latex usually occurs when
the pH is raised. This effect is attributed to ionization of the carboxylic-acid
groups of the contained polymer, with consequent swelling of the particles
and possible partial dissolution of the polymer in the latex aqueous phase
if the level of carboxylation is sufficiently high. As the pH of a carboxylated
latex is progressively raised, the viscosity usually passes through a maximum
in the region of pH 9.5-10.0. This matter is discussed further in section 11.8.3
of Chapter 11, and also in section 6.3.3 and 6.3.4 of Chapter 6 (Volume 1).
As regards the choice of alkali to be used for raising the pH, several factors
have to be taken into account. Not all common alkalis are technologically
equivalent. Economic and environmental factors also influence the choice.
It may also be desirable to enhance the colloid stability ofthe latex by adding
a post-reaction stabilizer. In particular, non-ionogenic surfactants, such as
surface-active ethoxylates, have been used for this purpose.
10.3.4 Distribution of carboxylic-acid groups in carboxylated latices

An important aspect of the science and technology of functionalized latices
is the manner in which the functionalizing groups are distributed within the
latex particles and also, possibly, within the dispersion medium of the latex.
A considerable technical literature now exists concerning the distribution of
carboxylic-acid groups in the particles contained in aqueous carboxylated
latices of non-polar polymers. It is recognized that, in principle, the carboxylic-
acid groups in such latices can occur principally in the following distinct
locations within the latex:
1. as combined acid groups buried in the interiors of the particles;
2. as combined acid groups present at or near the surface of the particles;
3. as combined acid groups present in the form of surface-active polymers
adsorbed at the surface of the particles;
4. as combined acid groups present in the form of water-soluble polymers
dissolved in the aqueous phase of the latex;
5. as acid groups present as unpoly.merized carboxylic acid, predominantly
in solution in the aqueous phase of the latex, especially if the pH of the
latex is above ca. 8.
Functionalizing comonomers which remain unpolymerized are, of course,
wasted as regards the primary motivations for functionalizing the latex;
indeed, there appear to be no applications of functionalized latices for which
the presence in the latex of unpolymerized functionalizing comonomer
(location (5» is advantageous. It is evident from section 10.3.2 above that it
is generally desirable that a substantial proportion ofthe combined carboxylic-
acid groups be located at or near the surface of the particles (locations (2)
and (3», and that, more specifically, they should be covalently bonded at the
surface (location (2» rather than bound to the surface merely by adsorption
(location (3)). Carboxylic-acid groups which have become buried in the
interior of the latex particles (location (1)) are effectively wasted, at least until
the particles have become integrated to form solid polymer. Combined
carboxylic-acid groups present as water-soluble and possibly surface-active
polymers (locations (4) and (3» may have important, and possibly undesirable,
influences upon the colloid stability and rheological properties of the latex.
The more important of the factors which affect the distribution of
carboxylic-acid groups in the carboxylated latex which forms in any particular
reaction system are as follows:
1. the inherent reactivities of the monomers;
2. the pH of the reaction system;
3. the hydrophilicity of the carboxylic-acid monomer(s);
4. various procedural aspects of the reaction, such as the way in which the
carboxylic-acid monomer is added to the reaction system;
5. any tendency for the carboxylic-acid monomer to react with other
substances in the reaction system, thereby forming products of modified
hydrophilicity or modified polymerization behaviour.
Little comment is necessary concerning factors (1) and (2). As regards factor
(1), if the carboxylic-acid monomer is reluctant to copolymerize with the
main monomers, but does tend to homopolymerize, then it will tend to be
present in the final latex as hydrophilic polymers which are either dissolved
in the aqueous phase or adsorbed at the surface of the particles. If the
carboxylic-acid monomer is not only reluctant to copolymerize with the main
monomers but also to homopolymerize, then it will tend to be present in
the final latex mainly as unreacted monomer dissolved in the aqueous phase.
As regards factor (2), it has already been noted that high pH favours
partitioning of the carboxylic-acid monomer in the aqueous phase as the
corresponding carboxylate salt. If the carboxylate anion homo polymerizes
in aqueous solution, the carboxylic-acid monomer will be present in the final
latex mainly as dissolved or adsorbed polymers which are rich in units derived
from the carboxylic-acid monomer; but if the carboxylate anion is reluctant
to homo polymerize in aqueous solution, then the carboxylic-acid monomer
will tend to be present in the final latex in the form of unpolymerized
carboxylate anions dissolved in the aqueous phase.
Several experimental investigations have been reported concerning the
effect offactor (3), namely, the hydrophilicity of the carboxylic-acid monomer.
The order of increasing hydrophilicity of the five commonly-used carboxylic-
acid monomers listed in section 10.3.1 above is as follows:
methacrylic < acrylic < itaconic, maleic, fumaric
The half-esters of dicarboxylic acid with mono- and dihydric alcohols are
less hydrophilic than are the dicarboxylic acids themselves; the actual
hydrophilicity depends upon the alcohol used to form the half-ester. It would
be expected that the less hydrophilic the carboxylic-acid monomer, the greater
is the tendency for it to become buried in the interior of the particles;
conversely, the more hydrophilic the carboxylic-acid monomer, the greater
is the tendency for it to become located near or within the aqueous phase.
These expectations have been confirmed by experimental investigations of
the location of carboxylic-acid groups in carboxylated latices. Thus methacrylic
acid units tend to become buried in the particles, whereas acrylic and itaconic
acids tend to become incorporated at the surface of the particles or in
water-soluble and surface-active polymers. Typical results, reported by
Greene [90,91], for the effects of level of carboxylic-acid monomer and pH
upon the distribution of combined carboxylic-acid groups in styrene-butadiene
copolymer latices carboxylated with acrylic acid and with methacrylic acid,
are shown in Tables 10. t and 10.2 respectively. The latices were prepared by
peroxodisulphate-initiated emulsion polymerization at 90·C, using sodium
n-dodecyl sulphate as the colloid stabilizer, a 20-nm seed latex, and continuous
monomer addition. The styrene/butadiene/carboxylic-acid monomer ratio in
the two series of latices investigated was (59-X)/4t/X m/m/m. The latices
obtained were analysed for:
1. the total amount of combined carboxylic-acid monomer associated with
the polymer phase of the latex as a whole (by neutron-activation analysis); and
Table 10.1 Effect of level of carboxylic-acid monomer and pH upon distribution of

combined carboxylic-acid groups in (59-X)/41/X m/m/m styrene/butadiene/
carboxylic acid copolymer latices carboxylated with acrylic acid (Greene [90])
Ratios of
combined
carboxylic-acid
Proportion Amounts of combined monomer
of acid carboxylic-acid monomer present in
Level of neutralized present in various forms / pphm various forms
acrylic with sodium interior/
acid,X hydroxide Latex Particle Particle Aqueous surface/
/ pphm /% pH interior surface phase aq.ph
1.25 25 4.6 0.21 0.89 0.15 17/71/12
2.50 25 4.9 0.45 1.20 0.85 18/48/34
3.75 25 5.4 0.66 2.04 1.05 18/54/28
5.00 25 5.2 0.88 2.62 1.50 18/52/30
2.50 0 2.0 0.40 1.70 0.40 16/68/16

2.50 25 4.9 0.45 1.20 0.85 18/48/34
2.50 50 5.4 0.14 t.ll 1.25 6/44/50
2.50 75 6.1 0.02 0.96 1.52 1/38/61
2.50 100 6.7 0.09 0.58 1.83 4/23/73
Table 10.2 Effect of level of carboxylic-acid monomer and pH upon distribution of

combined carboxylic-acid groups in (59-X)/41/X m/m/m styrene/butadiene/
carboxylic acid copolymer latices carboxylated with methacrylic acid (Greene [91])
Ratios of
combined
carboxylic-acid
Proportion Amounts of combined monomer
of acid carboxylic-acid monomer present in
Level of neutralized present in various forms / pphm various forms
acrylic with sodium interior/
acid,X hydroxide Latex Particle Particle Aqueous surface/
/pphm /% pH interior surface phase aq.ph
1.50 25 5.75 0.66 0.76 0.08 44/51/5

3.00 25 6.70 1.45 1.40 0.15 48/47/5
4.50 25 6.45 2.18 2.10 0.22 48/47/5
6.00 25 6.70 2.95 2.58 0.47 49/43/8
3.00 0 2.20 1.54 1.43 0.03 51/48/1

3.00 25 6.70 1.45 1.40 0.15 48/47/5
3.00 50 6.60 1.19 1.28 0.53 40/42/18
3.00 75 8.00 0.97 1.24 0.79 32/42/26
3.00 100 11.65 0.72 1.19 1.09 24/40/36
2. the amount of combined carboxylic-acid monomer associated with the

surface of the latex particles (by turbidimetric titration of the purified latex
with a cationic surface-active substance).
The amount of combined carboxylic-acid monomer buried in the latex
particles was calculated as the difference between (I) and (2). The amount of
combined carboxylic-acid monomer present in the aqueous phase was
calculated as the difference between the total amount of carboxylic-acid
monomer charged to the reactor and the amount of combined acid associated
with the polymer, it being assumed that all the carboxylic-acid monomer
had been polymerized, and that acid units which were not associated with
the polymer phase of the latex were present in the aqueous phase. These
latter quantities as calculated from the values of the quantities (1) and (2)
reported by Greene are shown in Tables 10.1 and 10.2, together with the
published values of the quantity (2). Also shown are the ratios of the amounts
of combined acid present in the location's aqueous phase, particle surface
and particle interior respectively, as calculated from the amounts derived
from the results reported by Greene. The principal conclusions to be drawn
from the results shown in Tables 10.1 and 10.2 are as follows:
1. For reaction systems which contained acrylic acid neutralized to an extent
of 25%, the amount of combined acid present in the interior of the particle
was directly proportional to the total amount of acrylic acid in the reaction
system, being in each case ca. 17%, regardless of the level of acrylic acid
in the reaction system. The partitioning of the remainder of the combined
acrylic acid between the aqueous phase and the interior of the particles
was somewhat more variable, but was usually such that the ratio of amount
of combined acrylic acid in the aqueous phase to amount at the particle
surface to amount in the particle interior was ca. 2/3/1 m/m/m.
2. For reaction systems which contained methacrylic acid neutralized to an
extent of 25%, the amount of combined acid present in the interior of the
particle was again directly proportional to the total amount of acrylic
acid in the reaction system, but in each case was ca. 50%, regardless of
level of acid in the reaction system. The partitioning of the remainder of
the combined methacrylic acid between the aqueous phase and the interior
of the particles was again somewhat more variable, but was always such
that the ratio of amount of combined methacrylic acid in the aqueous
phase to amount at the particle surface to amount in the particle interior
was ca. 1/10/10 m/m/m. Comparison between these results and those for
acrylic acid at 25% neutralization shows clearly the effect ofthe hydrophilicity
of the carboxylic-acid monomer upon the distribution of combined acid
groups in the final latex.
3. As the pH, and hence the proportion of acid present as carboxylate salt,
of the reaction system containing acrylic acid was increased keeping the
level of acrylic acid constant, the amount of combined acid present in the
aqueous phase increased sharply, and the amounts present in both the
interior of the particles and at their surface decreased accordingly.
However, the amount present in the interior decreased proportionately
more sharply than did the amount at the surface, becoming almost zero
when the acrylic acid was completely neutralized.
4. As the pH, and hence the proportion of acid present as carboxylate salt,
of the reaction system containing methacrylic acid was increased keeping
the level of methacrylic acid constant, the amount of acid combined at
the surface ofthe particles remained almost constant in the range 40- 50%.
The proportion of combined acid groups present in the aqueous phase
increased from almost zero when the acid was present entirely in the acid
form to ca. 36% when present entirely as the carboxylate salt. The
proportion present in the interior of the particles decreased accordingly.
Comparison between these results and those for acrylic acid again shows
clearly the effect of the hydrophilicity of the carboxylic-acid monomer
upon the distribution of combined acid groups in the final latex.
5. In summary, these results indicate that the proportion of acrylic or
methacrylic acid combined in the interior of the particles was generally
roughly the same for both acids, being ca. 50% regardless of the level of
acid or the pH of the reaction medium, and hence of the proportion
present as carboxylate salt. The difference between the two acids in respect
of the location of the acid groups in the final latex is that methacrylic
Preparation of functionalized latices 253
acid always gave a higher proportion of acid groups in the interior of the
particles and a lower proportion in the aqueous phase.
These comparisons between acrylic and methacrylic acids are confirmed by

results reported by Vanderhoff, van den Hul and Hamburg [92]. Vijayendran
[93] has compared the behaviours of acrylic, methacrylic and itaconic acids
in a reaction system for the preparation of carboxylated polystyrene latex.
The results are consistent with the order of increasing hydrophilicity amongst
these three carboxylic-acid monomers indicated above.
Table 10.3 shows results reported by Greene, Nelson and Keskey [94] for
the carboxylation of a styrene-butadiene copolymer latex with half-esters
formed from maleic acid and various monohydric alcohols. The latices were
again prepared by peroxodisulphate-initiated emulsion polymerization at
90·C, but using a sodium alkylarylsulphonate as the colloid stabilizer and a
conventional emulsion-polymerization procedure. The styrene/butadiene/
carboxylic-acid monomer ratio was 45/50/5 m/m/m for these latices. Like the
results shown in Tables 10.1 and 10.2, these results illustrate that reducing
the hydrophiIicity of a carboxylic-acid monomer reduces the amount which
becomes present in the final latex at the surface of the particles. It appears
from these results that the surface concentration of carboxylic-acid groups
is reduced by a factor in excess of ten if the methyl half-ester is replaced by
the n-butyl half-ester. However, the true effect of the replacement in respect
of comonomer hydrophilicity is obscured by the tendency for these half-esters
to hydrolyse under acid conditions, and therefore for the effective carboxylic-acid
monomer to become more hydrophilic. The tendency to hydrolyse decreases
in the order
methyl> ethyl> n-propyl > n-butyl
Table 10.3 Properties of latices of 45/50/5 m/m/m styrene/butadiene/monoalkyl

maleate copolymers (Greene, Nelson and Keskey [94])
Distribution of carboxylic-acid groups
Concentration of
Average surface
Alkyl group particle carboxylic-acid Ratio surface/total·
of monoalkyl Latex diameter groups / mg-equiv. carboxylic-acid
maleate pH /nm (g polymer)-l groups
Methyl 2.8 205 0.502 0.72/1

Ethyl 3.1 236 0.392 0.70/1
n-propyl 3.8 248 0.219 0.68/1
n-butyl 4.0 255 0.047 0.71/1
• Total carboxylic-acid groups as found by potentiometric titration.

The methyl half-ester was found to be completely hydrolysed under the

polymerization conditions employed in this investigation. The extents of
hydrolysis of the ethyl, n-propyl and n-butyl half-esters were found to be ca.
82%, 45% and 9% respectively. Thus it is not clear to what extent the results
for carboxylic-acid-group distribution are a consequence of decreasing
comonomer hydrophilicity as the chain length of the alkyl group is progressively
increased, and to what extent they are a consequence of different tendencies
to hydrolyse under the polymerization conditions. A further minor factor
which complicates comparisons between the various half-esters is that,
whereas they were added in fixed amounts by mass, and therefore the molar
level of addition decreased as the alkyl chain length increased, the surface
concentration of bound acid was in mg-equivalent of acid groups per unit
mass of polymer.
The polymerization procedure (factor (4) in the above list) can have an
important effect upon the way in which the combined carboxylic-acid groups
are distributed in the final latex. Slow continuous addition of a uniform
monomer mixture at a rate which is less than the rate of polymerization, so
that the polymerization proceeds under monomer-starved conditions, encourages
a uniform distribution of carboxylic-acid groups throughout the latex
particles. On the other hand, if all the monomers are added initially, then
the combined carboxylic-acid groups tend to become preferentially present
at the surface of the particles. Withholding some or all of the carboxylic-acid
monomer from the reaction system until the later stages of the reaction
encourages the combined acid to be present at the particle surface or in the
aqueous phase. If the polymerization is carried out by continuous monomer
addition under monomer-starved conditions, and the composition of the
monomer mixture with respect to the carboxylic-acid monomer is varied,
e.g., according to a power-feed regimen (see section 10.2.4.5.2 above), then
the distribution of carboxylic-acid groups within the latex particle is
influenced accordingly. An example of an emulsion-polymerization procedure
for the production of carboxylated latices is the two-stage process described
by Ceska [95] for the production of carboxylated styrene-butadiene latices
which are free from added surface-active substances. In the first stage, the
reaction mixture comprises all the ingredients, except that much ofthe styrene
and butadiene is withheld, so that the solids content is relatively low at the
end of this stage (ca. 25% m/m). The first-stage polymerization is carried out
at low pH (typically 2.5), so that the carboxylic-acid monomer is present in
the acid form. The product is a carboxylated seed latex. The pH is then
raised (typically to 9.5), thereby causing the copolymerized carboxylic-acid
groups to ionize to carboxylate anions. The highly-charged seed latex particles
thereby formed then function as reaction loci for the formation of a stable,
high-solids latex during the second stage of the polymerization, when the
remaining styrene and butadiene are added. It is claimed that ionization of
the carboxylic-acid groups of the seed latex before adding the remaining
monomers induces these anions to remain near the surface of the latex
particles during the second stage of the polymerization, thereby utilizing to
the full their capacity to stabilize the latex. The final solids content is typically
ca. 40% m/m. Particle-size analyses of the latices formed at the end of the
two stages indicate that the first stage is essentially a particle-nucleation
stage, whereas the second is predominantly a particle-growth stage.
Reaction between the carboxylic-acid monomer and other substances in
the reaction system (factor (5) in the above list) will almost certainly lead to
the formation of products having hydrophilicity and polymerization behaviour
which differ from those of the initial carboxylic-acid monomer. Indeed, it
may be that the product does not polymerize at all under the normal
conditions of free-radical emulsion polymerization. One example of such a
reaction is that between maleic acid and butadiene to form 1,2,3,6-
tetrahydrophthalic acid (Structure XXV) by a Diels-Alder reaction:
xxv
10.3.5 Effects of carboxylic-acid monomers upon rate of polymerization and

particle nucleation
Emulsion polymerization reaction systems which contain carboxylic-acid
monomers usually polymerize more rapidly than do similar reaction systems
which do not contain such acids. This effect is illustrated by results reported
by Ceska [96] for the peroxodisulphate-initiated emulsion polymerization
of styrene at 70·C in the presence of various carboxylic-acid monomers and
in the absence of added surface-active substances. Some of these results are
reproduced here as Figs lO.l7(a) and (b). Figure lO.l7(a) shows the effect of
increasing level of methacrylic acid upon the conversion-time relationship,
as evidenced by the effect of reaction time upon the solids content of the
latex formed. Figure lO.l7(b) compares the effects of equal molar amounts
of acrylic, methacrylic and itaconic acids upon rate of polymerization. In
assessing the significance of Figs 1O.17(a) and (b), it should be noted that the
theoretical solids content of the latices produced by complete polymerization
of the monomers in the reaction systems used range from ca. 21.4% to ca.
22.5%, depending upon the level of carboxylic-acid monomer, i.e., 20% solids
content corresponds to almost complete polymerization. These results show
that the rate of polymerization increases uniformly with increasing level of
acid, at least over the range of levels investigated. They also show that the
30 30
5;E ~E
§E20 §E20
0,* o'#.
(1)- (1)-
;g~ ;g~
0_
0_
~..!!! 10 ~..!!!10
.mo So
t2 t2
4 6 8 4 6 8
Time/hour Time/hour
(a) (b)
~ 1016
~
~
'"E
~
Q;
-
.c
E
~
I:
1015
I:
:8
e!
C
CD
U
I:
0
U
CD
~ 1014 1
~ 10- 10° 101
Level of itaconic acid / pphm
(c)
Figure 10.17 (a) Effect of increasing level of carboxylic-acid monomer upon conversion-time
relationship for peroxodisulphate-initiated emulsion copolymeri7.ation of styrene and methacrylic
acid at 70 C in absence of added colloid stabilizer (Ceska [95]). Significance of points: • 0, •
D
I, • 2, • 3, • 4 parts by mass of methacrylic acid per 60 parts by mass of styrene. (b) Effects
of equimolar amounts of itaconic acid (points .~ acrylic acid (points .) and methacrylic acid
(points .) upon oonversion-time relationship for peroxodisulphatc-initiated emulsion polymeri7.ation
of styrene at 70 C in absence of added colloid stabilizer (Ceska [95]). The level of acid was ca.
D
0.028 mole per 60 g of styrene. (c) Effect of level of carboxylic-acid monomer upon concentf"dtion
of particles nucleated in peroxodisulphate-initiated emulsion copolymerization of styrene,
butadiene and itaconic acid at 70 C in absence of added colloid stabili7.er (Ceska [95]).
D
ability of a given molar level of acid to increase the rate of polymerization

decreases in the order
methacrylic acid > acrylic acid> itaconic acid
i.e. this ability decreases with increasing hydrophilicity of the carboxylic-acid
monomer. Figure 1O.17(c) shows the variation of the particle concentration
(presumably in the final latices) with level of carboxylic-acid monomer for
the peroxodisulphate-initiated emulsion copolymerization of styrene and

butadiene at 70°C in the presence of itaconic acid and absence of added
surface-active substances. The styrene/butadiene ratio was 59/41 m/m. These
results show that the relationship between logarithm of the particle concentration
and logarithm of level of carboxylic-acid monomer is linear. The results
shown in these diagrams refer to emulsion polymerization reaction systems
which were free from added surface-active substances. Similar effects are
observed in reaction systems to which colloid stabilizer was added initially.
However, the presence of the added colloid stabilizer may then mask that
of the carboxylic-acid monomer to some extent.
Ceska [96] postulates that carboxylic-acid monomers have two important
effects when incorporated in reaction systems for the emulsion polymerization
of hydrophobic monomers in the absence of added surface-active substances:
particle nucleation is facilitated, so that the concentration of reaction loci in
the system, and hence the rate of polymerization, are increased; and the
polymer particles which form are colloidally stabilized. These postulates seem
to be generally valid, although the rate of polymerization for a given main
monomer may be influenced by the inherent reactivity of the carboxylic-
acid monomer in free-radical polymerization, as well as by its hydrophilicity.
Further comments concerning the results of Ceska [96] shown in Figs
1O.17(a)-(c) are as follows:
1. The conversion-time curves for the emulsion copolymerization of styrene
and methacrylic acid are linear up to high conversion, as is the conversion -
time curve for the emulsion homopolymerization of styrene. There is no
evidence of the curves having the S-shape which is characteristic of curves
for conventional emulsion polymerizations when colloid stabilizer is added
initially. The results for the emulsion copolymerization of styrene and
methacrylic acid in the absence of added surface-active substances provide
no basis for dividing the reaction into Intervals I, II, and III, and thus no
evidence of a distinctive particle-nucleation stage. The curves for the
emulsion copolymerization of styrene and acrylic acid, and of styrene and
itaconic acid, are non-linear, but again are not S-shaped, and so a similar
comment applies.
2. The effect of the hydrophilicity of the carboxylic-acid monomer upon the
rate of emulsion polymerization is interpreted by Ceska as implying that
the hydrophilic-hydrophobic balance of the acid monomer should be
such that
(a) the monomer is present in the aqueous phase and so be able to react
with the initiator radicals, but also such that oligomers have a strong
tendency to become hydrophobic as they grow, and so either collapse
to form new reaction loci, or become absorbed by pre-existing reaction
loci; and
(b) the reaction loci created by collapse of the oligomers are sufficiently
hydrophobic to absorb styrene strongly.
3. The slope of the line for the relationship between the logarithms of the
particle concentration and of the level of carboxylic-acid monomer is ca.
0.6, implying that particle concentration is directly proportional to the
0.6 power of the carboxylic-acid monomer level. This is identical to the
%th-power dependence of particle concentration upon the concentration
of colloid stabilizer present during Interval I of a conventional emulsion
polymerization reaction as predicted by the Smith-Ewart [29] micellar
theory and the Roe [32] homogeneous theory for particle nucleation (see
section t 0.2.2.2. t above). Just how close is this relationship between particle
concentration and level of itaconic acid is evident when values of particle
concentration (N) and level of itaconic acid (A), read off the double
logarithmic plot reported by Ceska, are plotted as N against AO. 6 • The
data are well-represented by a straight line through the origin; this confirms
direct proportionality between N and the 0.6th power of the itaconic acid
level. The equation for the linear regression of N upon AO. 6 is
N = 3.91 X 1013 + 9.74 X 10 14 AO. 6 (10.59)
assuming the units of A to be pphm on the styrene + butadiene in the

reaction system. It is unclear whether or not the similarity between this
exponent and that predicted by the Smith-Ewart and Roe theories is
other than fortuitous.
The results reported by Ceska [96], discussed above, are for emulsion
polymerization reaction systems which contain no added surface-active
substances. Many of the reaction systems which are used industrially for the
production of carboxylated latices contain some added colloid stabilizer. It
seems likely that, in reaction systems which contain both micellar colloid
stabilizer and carboxylic-acid monomer, particle nucleation occurs by
concurrent micellar and non-micellar mechanisms, and that, as the ratio of
colloid stabilizer to carboxylic-acid monomer decreases, so non-micellar
nucleation becomes increasingly important. Sakota and Okaya [97] have
reported results for the effect of level of sodium n-dodecyl sulphate upon the
concentration of particles nucleated during the emulsion polymerization of
styrene in the presence of a fixed amount of carboxylic-acid monomer. The
monomers comprised 90.9(sic)/5/2.5 m/m/m styrene/acrylic acid/fumaric acid.
The degree of neutralization of the acids was 0.85, and the polymerization
temperature 70·C. The results, reproduced here as Fig. 1O.l8(a), indicate that
two concurrent nucleation mechanisms were operative. Further results
reported by Sakota and Okaya, shown in Fig. 10. t 8(b), demonstrate that the
degree of neutralization of the carboxylic-acid monomer has an important
effect upon the concentration of particles nucleated, at least in reaction
systems which are essentially free of added surface-active substances. The
general trend is for the particle concentration to decrease as the carboxylic-
acid monomer is converted to carboxylate anions. This effect is explainable
3
•
•
Qi
.a
§c: 5
"4/
00 0.02 0.04 0.06 0.08 0.10
b
~ 3 .,
~E 2 ,.~~---
2l
8
Q)
"0
~ °o~--~~--~~--~~---L----~----~--~
~ 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Level of sodium n-dodecylsulphate /
(a) parts by mass per 98.4 parts of monomers
~20
E
~
o
a.
.2l
... 15
.8
E
::l
c:
..,
b 10
)(
c:
o
~
5j 5
oc:
o
o
Q)
"0
~
~ 00 0.2 0.4 0.6 0.8 1.0
(b) Degree of neutralization
Figure 10.18 Effect of (a) level of sodium n-dodecyl sulphate, (b) degree of neutralization of
the carboxylic-acid monomers, upon concentration of particles nucleated during the emulsion
copolymerization of90.9/5;2.5 m/m/m styrene/acrylic acidft"umaric acid mixture at 70·C (Sakota
and Okaya [97]). The degree of neutralization of the carboxylic acids was 0.85 for (a).
as being a consequence of increasing hydrophilicity of the oligomeric radicals

which form in the aqueous phase ofthe reaction system, as the carboxylic-acid
monomer is increasingly neutralized. There is, however, an interesting
subsidiary sharp maximum in the region of 0.85 degree of neutralization, for

which no satisfactory explanation can be offered. The effect of degree of
neutralization upon the rate of polymerization broadly parallels the effect
upon particle concentration.
10.4 AGGLOMERATION AND CONCENTRATION OF SYNTHETIC

LATICES
10.4.1 Some problems associated with the production of high -solids synthetic latices
The product of a conventional emulsion polymerization reaction is often a
latex oflow total solids content (typically not exceeding ca. 30% m/m), which
contains particles of relatively small average diameter (typically in the range
40~80 nm). The distribution of sizes about the average is narrow. This is
especially so for the products of emulsion polymerization reaction systems
for general-purpose synthetic rubbers, and may also be so for the products
of reaction systems for other types of synthetic latex as well. Thus a typical
reaction system for the production of a low-solids styrene~butadiene rubber
latex would contain initially 5 pphm of soap and 200 pphm of water. The
polymerization would be shortstopped in the range 60~ 70% conversion, to
prevent the formation of unduly tightly crosslinked rubber. The total solids
content after the removal of the unreacted monomers by stripping is therefore
ca. 26% mlm and the soap content ca. 7.5% mlm on the polymer. The average
particle diameter is ca. 60 nm. For many applications, however, synthetic
latices of total solids contents not less than 60% mlm are required. Attempts
to increase the total solids confent of a small-particle synthetic latex by, say,
simple evaporation results in a sharp increase in viscosity at total solids
contents much below the acceptable minimum of 50% mlm, the latex
undergoing a transition from a free-flowing liquid to a pasty, intractable
mass. Thus, in the case of the typical low-solids styrene-butadiene rubber
latex referred to above, the evaporation rate decreases markedly when the
total solids content exceeds 45% mlm, and, at a total solids content of ca.
50% mlm, the latex becomes too viscous to handle.
It is pointed out in section 1.8.3 of Chapter I (Volume I) that the
viscosity~concentration relationship for a latex depends upon the average
size of the particles in the latex, and also upon the distribution of sizes about
the average. The larger the average particle size and the broader the
distribution, the more fluid is the latex at any given concentration. The
reasons are that
1. as the average particle size increases, so the average distance of separation
between the surfaces of the neighbouring particles increases (see section
2.2.2.3 of Chapter 2 (Volume I)), so that the viscous forces opposing the
motion of a particle relative to its neighbours are reduced; and
Agglomeration and concentration of synthetic latices 261
2. as the particle-size distribution broadens, so the particles are able to pack

together more efficiently, again increasing the average distance of separation
between the surfaces of neighbouring particles.
A corollary of increased fluidity at any given latex concentration is that the

total solids content at which the transition from a free-flowing liquid to a
pasty mass occurs is increased. For these reasons, it has become general
practice partially to agglomerate some types of synthetic latex either before
or during the concentration process. In this way, the desired increase in
average particle size and broadening of the particle-size distribution are
achieved. In this context, the word partial when used to qualify the word
agglomeration implies that the size of most or all of the polymer particles in
the latex is increased somewhat, and not that part of the contained polymer
is converted to some form of massive coacervate.
The effect of partial agglomeration upon the viscosity-concentration curve
for a typical styrene-butadiene rubber latex is illustrated by results reported
by Daniels, Watson and White [98], reproduced here as Fig. 10.19. Johnson
and Kelsey [99] have reported results which show the importance of average
particle size and particle-size distribution upon the viscosity of synthetic
latices. Reference is also made to these results in section 6.3.2 of Chapter 6
(Volume 1). 30/70m/m styrene/butadiene copolymer latices were used.
Concentration was effected by evaporation, the procedure being to direct an
infrared lamp at the surface of the latex whilst the mass of latex was agitated
II)
A
~3
--
~
II)
0
l;l 2
:>
40 50 70
Total solids content /%mlm
Figure 10.19 Effect of partial agglomeration upon the viscosity-concentration curve for a
typical styrene-butadiene rubber latex (Daniels. Watson and White [98]). Viscosities were
detennined by Brookfield viscometer using spindle # 3 at rotational frequency 30 minute - I.
Curve A is for the initial latex. which had average particle diameter 60 om; curve B is for the
same latex after agglomeration to average particle diameter 160nm.
continuously. A stream of air was blown across the latex surface to facilitate
rapid removal of water vapour. By this means, skinning and thermal
coagulation were minimized. The results shown in Table 10.4 confirm that
the greater the average particle diameter, the lower the viscosity at any given
total solids content, and also the greater the total solids content at which
the viscosity attains any given value. Johnson and Kelsey have also reported
curves which show the effect of concentration upon latex viscosity. These
curves, shown as Fig. 6.5(a) of Chapter 6 (Volume 1), are broadly similar to
those obtained by other workers. They confirm that an increase in average
particle diameter is accompanied by a considerable decrease in viscosity at
any given total solids content. Particle-size distributions for these latices
indicate that the effects upon the viscosity-concentration relationship are
primarily a consequence of change in average particle diameter, and not of
differences in particle-size distribution, the distributions being similar for the
various latices. That particle-size distribution can have an important effect
upon the viscosity at any given total solids content was demonstrated by
determining the viscosity of blends of 95-nm and 325-nm latices (see Fig.
6.5(b) of Chapter 6 (Volume I». The viscosity of the blend was always
considerably less than that of the 95-nm latex, and slightly less than that of
the 325-nm latex, at any given total solids content. That this effect was not
a consequence of particle agglomeration during blending and/or concentration
was confirmed by determining the particle-size distribution of the blends. In
a second paper, Kelsey and Johnson [100] discuss theoretical implications
of their observations.
Qualitatively, it is found that a polydisperse styrene-butadiene rubber
latex with an average particle diameter of 250 nm is still sufficiently fluid for
practical application, even though its total solids content is increased to
70% m/m. For application in the manufacture of latex foam rubber, it is
generally required that a synthetic latex should have a total solids content
of at least 60% m/m and a viscosity of not more than 1 Pa s. For
styrene-butadiene rubber latices made by any given process, there is a definite
relationship between the average particle diameter and the total solids content
Table 10.4 Effect of average particle diameter upon viscosity-total solids content
relationship for a styrene-butadiene copolymer latex (Johnson and Kelsey [99])
Modal particle diameter of latex
Property 95nm 17Jnm 325nm
Brookfield viscosity· j Pa s at 60% mjm

total solids content 8.00 1.80 0.25
Total solids content j % mjm at which
Brookfield· viscosity was 1 Pa s 55 59 63
* Model LVF, rotational frequency 60 min - I, various spindles.

at which the Brookfield viscosity is 1 Pa s. In practice, it appears that an

average particle diameter of at least 160nm is required if these minimum
requirements are to be attained. Indeed, according to Jones [101], a
styrene-butadiene rubber latex which has been agglomerated to an average
particle diameter of ca. 160nm can be evaporated to a total solids content
in excess of 52% m/m whilst still retaining high fluidity.
It may be enquired why the preferred practice is to agglomerate and
concentrate low-solids, small-particle synthetic rubber latices, rather than to
produce large-particle latices directly by emulsion polymerization. This
alternative route to high-solids synthetic latices has been thoroughly investigated,
and has been rejected on two principal grounds: the colloid stability of
high-solids emulsion polymerization reaction ~ystems tends to be low, and
thus unacceptably high levels of coagulum are often produced; and the rates
of polymerization in such reaction systems are usually unacceptably low. In
general, it is found that large particles are formed during emulsion polymerization
only under conditions of low colloid stability. This is because the large
particles form as a result of coalescence of smaller particles formed initially.
Amongst the factors which favour the formation of large particles are low
concentration of colloid stabilizer, low pH, high ionic strength, surface-active
colloid stabilizers of low molecular mass, and high ratio of monomer to
water. It has also been found that the chemical nature of the stabilizer has
an influence; thus sulphates and sulphonates tend to give larger particles
than do carboxylates. The effect of low concentration of colloid stabilizer is
two-fold: it lowers the overall colloid stability of the reaction system, thereby
encouraging particle coalescence; and the concentration of reaction loci
formed during Interval I is reduced, so that a given amount of monomer is
polymerized in fewer reaction loci. The latter effect is the reason initial rates
of polymerization are lower than in conventional emulsion polymerization
reaction systems. The rate of polymerization is further reduced as particle
coalescence occurs, usually at ca. 30% conversion. It has been estimated that
the effect of particle coalescence may be to reduce the number of reaction
loci to as few as 3% of the number initially present. A further problem which
may be encountered in the production of high-solids synthetic latices by
direct emulsion polymerization is that of heat transfer. Although the rate of
polymerization may be considerably reduced relative to that for a conventional
emulsion polymerization reaction system, heat transfer can still be a problem
because the viscosity of the reaction system increases as polymerization
proceeds. Papers dealing with various aspects of the production of high-solids
synthetic latices directly by emulsion polymerization include those of
Howland and Nisonoff [102], of Brown and Howland [103, 104], of Morton
and Gibbs [105], of Rupar and McLeod [106], and of Levine, Lindlaw and
Vona [107]. The last of these papers discusses problems which attend the
direct production by emulsion polymerization of high-solid latices containing
polymers based upon vinyl acetate. Examples of patented processes for the
production of large-particle and high-solids synthetic latices directly by

emulsion polymerization include those to McCracken and Betts [108],
Gauslaa [109], The General Tire and Rubber Company [110], and Celanese
Corporation [111]. The General Tire and Rubber patent provides a good
example of the complex emulsion polymerization reaction systems which
have been proposed for the direct production oflarge-particle synthetic latices.
The claim is that such latices can be simply and directly produced at rapid
rates of polymerization in a single-stage emulsion polymerization process
using an aqueous medium which contains the following:
1. a colloidal dispersion of a polymer or mixture of polymers possessing

either anionic or cationic hydrophilic groups, and which swells in the
conditions prevailing in the emulsion polymerization system to an extent
in the range 50- 500% vIv;
2. a hydrophilic surface-active substance (referred to as a 'bihydrophilic
surfactant') which contains two different types of hydrophilic group, one
of which is an ionic group of opposite polarity to that present in the
polymer, or a mixture of such substances; and
3. either one or more monofunctional bases in a quantity sufficient to give
a pH in the range 8-11 when the polymer or mixture of polymers has
anionic groups, or one or more monofunctional acids in a quantity
sufficient to give a pH in the range 6-3 when the polymer or mixture of
polymers has cationic groups.
10.4.2 Methods of partially agglomerating synthetic latices
10.4.2.1 Preliminary considerations; resume of methods

We turn now to consider the various methods which have been proposed
for partially agglomerating the latices which are obtained as the immediate
products of emulsion polymerization. Most of the published literature
concerning the partial agglomeration of synthetic latices relates to latices of
rubbery styrene-butadiene copolymers. This is not surprising, in view of the
industrial importance of high-solids versions of such latices. An important
factor which has led to the development of agglomeration processes has been
that the concentrated synthetic latices, the production of which are thereby
enabled, have found wide application as competitors for ammonia-preserved
natural rubber latex concentrates; it will be recalled (see section 9.4.6 of
Chapter 9) that the latter are available at total solids contents in excess of
60% m/m. However, although the methods which are described in the
following sub-sections have all been developed primarily with a view to
application to styrene-butadiene rubber latices, they are in principle applicable
to synthetic latices of all types.
All the methods which have been developed for the partial agglomeration
of synthetic latices involve reduction of the colloid stability of the latex in a
controlled manner. Such reduction is not of itself sufficient if the process is
to be satisfactory. It is also necessary that the colloid destabilization should
be self-terminating. This implies that the particles should become colloidally
stable after the particle size has increased to a certain rather small extent. If
this condition is not fulfilled, then particle coalescence continues, and the
polymer becomes converted to some form of massive coacervate. Thus the
development of processes for the partial agglomeration of synthetic latices
has provided an interesting practical exercise in colloid science, in that colloid
stability has to be reduced in such a way that, although particle coalescence
occurs initially, stability is soon recovered sufficient to prevent further particle
coalescence. It is generally observed that particle agglomeration is accompanied
by a significant reduction in the surface free energy of the latex. The reason
for this is that agglomeration increases the chemical potential of the colloid
stabilizers in the latex, so that a higher surface concentration can adsorb at
the interface between air and the aqueous phase. Indeed, the amount of
colloid stabilizer in an agglomerated latex is often more than sufficient to
saturate the latex particles. One consequence of the decrease in surface free
energy is that latices often foam more readily after agglomeration than they
do before. For latices stabilized by carboxylate soaps, the pH may increase
slightly during agglomeration. This phenomenon is also probably a consequence
of the release of colloid stabilizer into the latex aqueous phase.
The following is a convenient classification of the principal methods which
have been proposed for the partial agglomeration of synthetic latices:
1. Chemical methods, which subdivide into:
(a) agglomeration by addition of solvents, both polymer- and water-
miscible;
(b) agglomeration by addition of electrolytes;
(c) agglomeration by partial destruction of surface-activity of colloid
stabilizers;
(d) agglomeration by addition of water-soluble hydrocolloids;
(e) agglomeration by addition of surface-active substances.
2. Physical methods, which subdivide into:
(a) agglomeration by subjecting the latex to freezing and thawing;
(b) agglomeration by subjecting the latex to mechanical agitation.
Each of these methods is described in the following sub-sections. Not only
do they provide interesting examples of the applied colloid technology of
latices; also the development of successful methods of achieving partial
agglomeration has had an important influence upon the synthetic rubber
latex industry in the period since the end of World War II. Some of the
processes which have been proposed use combinations of two or more of
the above methods, making for difficulty in classifying them under the above
headings. In such processes, the intensity of each separate type of treatment

can usually be reduced, because the different types of treatment are often
additive in effect, and may be synergistic.
In principle, all the agglomeration methods to be described in the following
sub-sections can be used to co-agglomerate mixtures of dissimilar latices.
Co-agglomeration may be defined as a process in which the particles of two
or more dissimilar latices are agglomerated to form heterogeneous composite
particles in which the particles of one type of latex have become embedded
in the particles of another, but otherwise retain their identity. Co-agglomeration
has been applied particularly to mixtures of synthetic latices of rubbery
polymers and glassy polymers. The objective is to produce latices which
contain composite particles comprising both rigid domains and rubbery
domains. Films dried down from such latices comprise an intimate mixture
of the two types of particles, and in consequence exhibit some degree of
particulate reinforcement. Co-agglomerated synthetic rubber and resin latices
have been of considerable industrial interest for the manufacture of latex
foam rubber. In some respects, the films formed from the composite particles
present in such co-agglomerated latices are analogous to the intimate mixtures
of rubber and particulate fillers, such as carbon black, which are formed by
dispersing the filler in a rubber matrix by milling.
10.4.2.2 Agglomeration by chemical methods
10.4.2.2.1 Agglomeration by addition of solvents
Partial agglomeration can be effected by adding to the latex a polymer
solvent, and then subjecting the mixture to heating and agitation. The solvent
is subsequently removed from the latex by distillation. Solvents which can
be used with styrene- butadiene rubber latices include benzene, toluene and
styrene. Advantages claimed for this method of agglomeration as compared
with others include non-destruction of the colloid stabilizers of the latex,
thereby minimizing the tendency for coagulum to form during agglomeration,
and the absence of foreign substances in the latex after agglomeration and
removal of the solvent. The disadvantage of the method is that relatively
large volumes of polymer solvent are required to effect a useful degree of
agglomeration. In the case of styrene- butadiene rubber latices, it may be
necessary to add solvent equal in volume to ca. 80% of the volume of the
polymer in order to obtain an agglomerated latex which is capable of being
concentrated without undue increase in viscosity. The necessity for such large
volumes, and the associated cost of solvent recovery, as well as health and
fire hazards, have made this process industrially unattractive.
The effect of the added solvent is to swell the latex particles as the solvent
is absorbed into the particles. The theory of this phenomenon is discussed
in section 5.2.4.2 of Chapter 5 (Volume I). The colloid stability of the latex
becomes reduced principally because the area of polymer-aqueous phase
interface increases as the particles swell, bringing about a reduction in the
surface concentration of the colloid stabilizers present in the latex. If the

amount of colloid stabilizer present in the initial latex is inadequate, then
some bulk coagulation may occur during heating and agitating. Colloidal
destabilization is facilitated by reduction in the average distance of separation
between neighbouring particles as the particles swell. This reduction is
manifest by an increase in the viscosity of the latex. As agglomeration
proceeds, the viscosity then decreases as the average particle size increases,
and as the particle-size distribution broadens. Increase in the average distance
of separation between neighbouring particles as agglomeration occurs is
probably one factor which causes the agglomeration process to terminate.
Another factor is that the surface concentration of the colloid stabilizers
increases as the average particle size increases, thereby reducing the tendency
for further particle coalescence to occur. It has been claimed that, if the
process is carefully controlled, very little coagulum forms during solvent
agglomeration with conventional carboxylate-stabilized styrene-butadiene
rubber latices.
A typical solvent-agglomeration process as applied to styrene-butadiene
copolymer latices stabilized with carboxylate soaps is described in a patent
to Brown [112]. The results of an experimental investigation of the
agglomeration of such latices by polymer solvents has been reported by
Brown and Howland [103]. The principal solvent used was benzene. It was
demonstrated that the particle-size distribution broadens markedly as a
consequence of solvent agglomeration, as well as the average particle size
increasing. The broadening of the particle-size distribution appeared to be
a consequence of random merger of particles. The conclusions reached
concerning the effects of various variables upon the process can be summarized
as follows:
1. Amount of solvent Both the extent of agglomeration and the final average
particle size increased regularly as the amount of solvent increased. An
amount of solvent equivalent to 50- 70% of the polymer volume was
necessary to increase the particle size to within the range necessary for
subsequent concentration to high total solids content.
2. Solvent type Isoprene was somewhat less effective in promoting
agglomeration than was benzene. (Isoprene was used as a conveniently-
handleable model for butadiene in the laboratory; the interest was in the
possibility of using butadiene on the plant.) As regards other solvents,
hydrocarbons and their halogen derivatives gave smooth agglomeration,
whereas more polar solvents, such as diisobutyl carbinol and nonylphenol,
led to the formation of excessive coagulum. The general conclusion was
that suitability as an agglomerating agent correlates closely with solvent
power for the polymer.
3. Inclusion of crosslinking comonomer in the emulsion polymerization reaction
system used to produce the initial latex Extensive coagulation occurred
if the contained polymer was crosslinked by inclusion of a crosslinking

monomer such as divinylbenzene in the emulsion polymerization reaction
system by which the latex was produced. A possible reason for this
phenomenon is that, although the highly crosslinked particles are swollen
by the solvent, albeit to a reduced extent, and can aggregate together, they
are unable to coalesce properly. The area of interface between polymer
and aqueous phase is therefore not reduced to the same extent as when
the particles are uncrosslinked or only lightly crosslinked. Colloid stability
remains low, and the agglomeration process is no longer self-terminating.
4. Hydrocarbon/water ratio This variable had a considerable effect upon
the extent of agglomeration. As expected, optimum agglomeration efficiency
required as high a hydrocarbon/water ratio as possible.
5. Temperature The rate of agglomeration increased with increasing
temperature, but the effect upon the extent of agglomeration was slight.
6. Colloid stabilizer Addition of carboxylate soap had a slight effect upon
the extent of agglomeration. Addition of a dispersing agent of the naphthalene-
sulphonate type before agglomeration had virtually no effect upon the
agglomeration process, but did decrease the viscosity both during and
after treatment.
It may be noted that the high-solids emulsion polymerization reaction

systems discussed in section 10.4.1 above probably provide examples of
polymer-solvent-agglomerated systems. In such systems, the unreacted
monomer, which is present in large amount relative to the aggregate volume
of the reaction loci, can cause the latex particles which initially form to swell
and coalesce. It is significant that coalescence commences at ca. 25%
conversion, when sufficient polymer has formed for it readily to absorb a
large quantity of unreacted monomers. Because agglomeration is facilitated
by reduced distance of separation between the surfaces of neighbouring
growing latex particles, it is not surpising that the water content of the
reaction system is an important factor in determining the ultimate average
size of the particles which form.
Agglomeration using water-miscible solvents has also been proposed.
Agglomeration is brought about by addition of a liquid such as acetone,
which is subsequently removed by distillation prior to the latex being
concentrated. One such process is described in a patent to Rumbold [113].
Such processes depend upon the ability of water-miscible organic liquids to
reduce the colloid stability of aqueous latices (see section 5.2.5 of Chapter 5
(Volume 1».
Combinations of polymer-miscible and water-miscible solvents have also
been proposed as agglomerating agents for synthetic rubber latices. Waterman
and Le Heij [114] have described the so-called Vistex mixed-solvent process
for the production of high-solids styrene- butadiene rubber latices which are
suitable for the manufacture of latex foam rubber. The latex is agitated with
a mixture of a non-polar and polar solvents, and the solvents then removed
by steam-stripping. A suitable solvent mixture is 80/20v/v benzene/
isopropanol. Other suitable polar solvents include acetone, sec-butanol and
methylcellosolve. The average particle size and the particle-size distribution
of the final latex can be controlled by varying two principal factors: the pH
of the initial latex and the quantity of mixed solvent used. Waterman and
La Heij consider that agglomeration occurs through a combination of two
effects: the non-polar component of the solvent mixture is absorbed into the
latex particles, thereby swelling them; and the polar component reduces the
colloid stability of the latex by partially removing the stabilizer from the
surface of the particles. Further information on this type of process is available
in a patent to Shell Internationale Research ~aatschappij NV [115].
10.4.2.2.2 Agglomeration by addition of electrolytes
Processes for the partial agglomeration of synthetic latices by the addition
of electrolytes can be classified under three sub-headings:
1. processes which use non-volatile electrolytes containing univalent cations;
2. processes which use volatile electrolytes containing univalent cations;
3. processes which use electrolytes containing multivalent cations.
In processes which use electrolytes containing univalent cations, the requisite
partial colloidal destabilization is brought about principally by increase in
the ionic strength of the aqueous phase. In processes which use electrolytes
containing multivalent cations, interaction between the cations and stabilizer
anions bound at the particle surface probably contributes to the reduction
in colloid stability.
One of the earliest methods developed for the agglomeration of styrene-
butadiene rubber latices exploited partial colloidal destabilization by the
addition of a non-volatile inorganic electrolyte containing a univalent cation.
This was the basis of the German Stockpunkt (literally gel- or curdle-point)
process which was developed during World War II [116]. The German
workers found that, if their synthetic latices were cooled, they thickened
sharply to a pasty or butter-like consistency over a narrow temperature
range, usually in the range 5-15°C. The average temperature at which this
transition occurred was called the Stockpunkt. It depends primarily upon the
amounts of electrolytes present in the latex, being raised by addition of salts
such as sodium sulphate and sodium carbonate. It is to a lesser extent raised
by increasing the concentration of particles in the latex. The German process
for the concentration of synthetic latices based upon the Stockpunkt
phenomenon required the addition of sufficient sodium carbonate to raise
the Stockpunkt to within the range 1O-12°C, cooling the latex to a few
degrees below the Stockpunkt, and stirring or beating the paste which was
formed. Agitation was continued for ca. 4 hours, during which time the paste
broke up into small lumps. The agitation was then stopped, and the mass
allowed to stand, when the agglomerates of particles rose to the surface and
formed a creamy layer of concentrate above a substantially clear aqueous

phase. The aqueous phase was then drawn off, and the concentrate allowed
to warm to room temperature. Gentle stirring was sufficient to produce a
uniform stable latex concentrate. In this way, the total solids content could
be increased from ca. 25% to ca. 50% m/m. If the latex was held for too
long below the Stockpunkt, then slow coacervation occurred. According to
Maron and Moore [117], in a further development of the Stockpunkt process,
the latex was initially beaten into a foam before being cooled and sodium
carbonate added. The effect of the foaming was probably to initiate
agglomeration of the particles, and to encourage desorption of colloid
stabilizers from the surface of the particles by adsorption on to the increased
area of air-aqueous phase interface induced into the latex.
Immediately following World War II, the Stockpunkt process was investigated
and developed by American scientists and technologists as part of their
synthetic rubber program. It was significant that, unlike the German synthetic
latices, the American latices did not exhibit the Stockpunkt phenomenon
unless electrolytes were first added. This difference in behaviour was
eventually traced to the use by the German workers of napthalene-
sui phonates as colloid stabilizers for their emulsion polymerization reaction
systems, whereas the Americans used carboxylates (fatty-acid and rosin-acid
soaps). It was the high concentration of electrolyte impurities (mainly sodium
sulphate), invariably present in industrial grades of naphthalenesulphonates,
which caused the German latices to exhibit the Stockpunkt phenomenon
without addition of electrolytes.
The early American developments of the process have been described by
Maron et al. [lIS]. The initial foaming of the la-tex, and the subsequent
granulation of the paste, were both found to be unnecessary. The modified
process consisted in adjusting the colloid stability of the latex such that rapid
agglomeration would occur subsequently, adding sufficient sodium chloride
to cause gelation at normal ambient temperatures, cooling the gel until, after
an initial thickening, it became thin again, and then filtering off a clear
aqueous phase through paper under pressure. It is said that total solids
contents as high as 70% mlm could be attained by this method, but that
filtration was slow in the later stages. The soap content of the latex was
determined and adjusted, if necessary, to a level in the range 0.2S-0.29 mg-
equivalent (g polymer)-l by the addition of potassium rosin-acid soap. The
pH was adjusted to 10.5-11.5, filterability being poor if the pH was below
10. 3 mol dm - 3 aqueous sodium chloride solution was then slowly added to
the latex at room temperature, the amount being sufficient to bring the
concentration up to 0.24 mol dm - 3 on the final latex. As soon as possible
after addition of the salt, the mixture was cooled, the gelling mass being
stirred continuously to facilitate heat transfer and to prevent the formation
of a coherent gel. Cooling was continued until the latex thinned out; this
usually required cooling to 5-S°C. The latex was then allowed to warm under
mild agitation to ensure thorough homogenization. As an alternative to

concentration by filtration, the agglomerated latex could be centrifuged.
Further investigation has shown that a certain minimum critical concentration
of electrolyte must be present in the aqueous phase of the latex if agglomeration
is to proceed. The particles then grow in size at a definite rate which increases
as the difference between the actual and critical electrolyte concentrations
increases. Results reported by Schmidt and Kelsey [119] show that the surface
free energy of the latex falls progressively as the average particle diameter
increases, confirming that colloid stabilizers are released into the latex
aqueous phase as the specific surface area of the dispersed polymer decreases.
It has been found that particle agglomeration by electrolytes can be arrested
at any stage by diluting the latex to such an extent that the concentration
of electrolytes falls below the critical level. The existence of a critical electrolyte
level for particle agglomeration indicates that a prerequisite for electrolyte
agglomeration is that the colloid stability of the latex should be reduced to
a certain minimum level before the latex is cooled.
The particles which result from electrolyte agglomeration of the latices of
rubbery polymers have been shown to be spherical like the small particles
which were present in the initial latex. They do not appear to be merely
clusters of primary particles. Perfect coalescence appears to occur. That
substantial coalescence occurs is confirmed by the decreases in surface free
energy which accompanies the increase in average particle size.
Patents covering processes for the partial agglomeration of synthetic latices
by addition of non-volatile inorganic electrolytes include those to Peaker
[120], to Willson [121], and to Arundale [122]. The principal disadvantage
of agglomeration processes based upon the addition of electrolytes of this
type is the high non-polymer content of the total solids fraction of the resultant
latex. In particular, the high levels of electrolytes make the subsequent latex
concentrates unsuitable for the manufacture of latex foam rubber by processes,
such as the Dunlop process (see section 18.3 of Chapter 18 (Volume 3», which
were widely used at the time methods of agglomerating synthetic rubber
latices were being developed. Attempts were therefore made to mitigate this
disadvantage, if not to obviate it altogether, by using a 'volatile' electrolyte,
such as ammonium carbonate, instead of a non-volatile 'fixed' electrolyte,
such as sodium chloride. After inducing the desired degree of particle
agglomeration, the volatile electrolyte is removed by steam-distillation during
the subsequent concentration by evaporation. The principal disadvantage of
this method has been the high cost of the relatively large amounts of volatile
electrolytes which are required. A patent to Rumbold [123] claims a process
of this type. It has been found possible to promote electrolyte-induced particle
agglomeration by adding small amounts of various water-soluble hydrocolloids
and surface-active substances. These methods are described in more detail
subsequently (see sections 10.4.2.2.4 and 10.4.2.2.5 below) when the use of
these additives as partial agglomerants is considered.
Industrial interest has also been shown in the possibility of partially

agglomerating synthetic latices by the use of electrolytes which contain
multivalent cations. It is doubtful whether much use has ever been made of
such processes in practice. Schindler [124] has disclosed a process in which
the size of the particles in a perfluorocarbon polymer latex is increased by
addition of multivalent metal ions prior to emulsion polymerization. The
cations are to be at least divalent, must be inert towards the polymerization
reaction, and should preferably not discolour the polymer product. Typical
examples of suitable cations are zinc(II) and aluminium(lII). The concentration
range is 1O-5-10- 4 moldm- 3 , based upon the aqueous phase.
10.4.2.2.3 Agglomeration by partial destruction of the surface-activity of colloid

stabilizers
In most, if not all, agglomeration processes of the type considered here, partial
destruction of the surface-activity of the colloid stabilizers is achieved by
partial conversion of adsorbed carboxylate soaps to the un-ionized acid form.
This conversion is often referred to as partial soap neutralization. There are
several variants of this type of process. In one, a small-particle latex which
has been prepared using a carboxylate soap as colloid stabilizer is then further
stabilized by the addition of a small amount of a further anionic stabilizer
of the sulphate or sulphonate type. The carboxylate anions are then partially
converted to un-ionized acid by the gradual addition of acid. As the pH
decreases, so partial agglomeration of latex particles occurs. After the desired
degree of agglomeration has been achieved, the pH is raised again by addition
of alkali. Re-dispersion ofthe agglomerate~ does not occur. In another variant,
a salt of a volatile base and a fairly strong non-volatile acid is added to the
latex. Some of the base is then removed by heating, causing the pH to fall
with concomitant partial agglomeration. Finally, the pH is raised by addition
of a fixed alkali. A third possibility is that the initial emulsion polymerization
is carried out using a reaction system in which at least part of the colloid
stabilizer is a volatile-base carboxylate soap, such as ammonium oleate.
Subsequent to polymerization, the volatile base is again removed by heating,
so that the pH falls and agglomeration occurs. Finally, the pH is again raised
by addition of a fixed alkali. Yet another possibility is that the pH is reduced
by adding to the latex a compound which hydrolyses in contact with water
at an appropriate rate to produce an acidic compound. Whatever may be
the means by which the acidity is developed within the latex aqueous phase,
it is essential that the acidity should be stronger than the acid from which
the carboxylate soap was derived. In all these processes, the extent of soap
coverage of the latex particles is reduced by partial conversion to the
un-ionized acid. The average particle size corresponding to complete soap
coverage is thereby increased.
Patents covering agglomeration processes depending upon partial soap
neutralization by pH reduction include those to Peaker [125], to Svendsen
[126], to Rumbold [127], to Howland and Nisonoff [128], to Willson [129],

and to Monsanto [130]. Suitable volatile bases which can be used in forming
a volatile-base salt include ammonia, methylamine, dimethylamine, ethylamine
and triethylamine. Suitable acids for forming such salts include adipic
(1,4-butanedicarboxylic), benzoic, chloroacetic, citric 2-hydroxy-I,2,3-
propanecarboxylic, fumaric and hydrochloric. In the process covered by the
Monsanto patent, the pH is reduced by allowing a water-soluble acid
anhydride to hydrolyse in the latex aqueous phase. Suitable acid anhydrides
include acetic, propionic and maleic. The acid anhydride may with advantage
be dissolved in a water-soluble organic solvent, such as methanol, ethanol
or acetone, and then a small amount of water added prior to adding the
solution to the latex. Alternatively, the more water-soluble acid anhydrides
may be dissolved in a small amount of water without an organic solvent,
precautions being taken to avoid premature hydrolysis.
1004.2.204 Agglomeration by addition of water-soluble hydrocolloids

Certain water-soluble hydrocolloids are able, under appropriate conditions,
to induce partial agglomeration in synthetic latices. The best-known of such
hydrocolloids is polyvinylmethyl ether. In some processes, these hydrocolloids
are used as the primary agglomerant, possibly in the absence of other
deliberately-added destabilizative influences. In other processes, they are used
in conjunction with other added destabilizers, such as electrolytes and
pH-reducers, and may in effect function as sensitizers of agglomeration by
these other influences. It is sometimes difficult to judge whether in these latter
processes the hydrocolloid is acting as a promoter of colloidal destabilization
by the other additive, whether the other additive is sensitizing the latex to
agglomeration by the hydrocolloid, or whether the agglomeration is a truly
cooperative effect. In any effect, it seems certain that the hydrocolloid
contributes to reduction of the colloid stability of the latex; whether it does
so by the interparticle bridging mechanism (see section 5.2.7 of Chapter 5
(Volume 1», or by another mechanism such as adsorption of colloid
stabilizers, is an open question.
The best-known process in which a water-soluble hydrocolloid agglomerant
is used is that described by Howland et al. [131], for which the hydrocolloid
is a modified polyvinylmethyl ether. This process is unusual in comparison
with most other agglomeration processes, in that agglomeration occurs
during, and concurrently with, the concentration of the latex by evaporation;
unlike most other agglomeration processes, it is not effected prior to
concentration. An important advantage of this method is therefore that no
additional capital outlay for agglomeration equipment is required. The
process depends upon the observation that modified polyvinylmethyl ethers,
when present in certain synthetic latices in small amounts, are able to induce
extensive partial agglomeration during concentration by evaporation.
Unmodified polyvinylmethyl ethers have little or no ability to bring about

particle agglomeration, unless they are present at fairly high concentration
and an electrolyte has been added. No electrolytes or other substances need
to be added to supplement the action of polyvinylmethyl ethers which have
been suitably modified, although it may be expedient in practice to make
such additions. The extent to which agglomeration occurs is controlled by
the amount of polyvinylmethyl ether which is present. The agglomeration is
self-terminating, without the formation of macro-agglomerates or of large
amounts of coagulum, if appropriate conditions are established. The rate of
particle agglomeration can be greatly increased by increasing the initial total
solids content of the latex: this is interpreted as being a consequence of
reduced interparticle separation. The treatment by which the effectiveness of
polyvinylmethyl ether as an agglomerant is increased involves heating the
polymer either alone or with an acidic substance such as sulphuric acid (see
further below). It is believed that the treatment alters the molecular complexity
of the polyvinylmethyl ether, in particular, the extent to which it is crosslinked.
In modifying a polyvinylmethyl ether for this application, it is necessary to
ensure that the treatment is not carried to such an extent that the product
has insufficient water-solubility at normal ambient temperatures. A convenient
criterion of suitability for application as a latex agglomerant has been found
to be the heptane ratio of the polymer. This is defined as the number of grams
of n-heptane required to cause incipient precipitation (i.e. turbidity) when added
to 1 gram of a 1% m/v solution of the polyvinylmethyl ether in benzene at 25°C.
The heptane ratio is believed to be primarily an index of molecular structure.
Effectiveness as an agglomerant decreases as the heptane ratio increases.
In a typical process which makes use of the agglomerating effect of
polyvinylmethyl ether, a 35% m/m 60-nm styrene~butadiene rubber latex is
stripped of its unreacted monomers, and is transferred to a concentrator for
the concurrent removal of water and partial agglomeration of the particles.
At the commencement of the concentration operation, dilute aqueous
solutions of polyvinyl methyl ether and of an electrolyte such as potassium
sulphate are added to the latex. Howland et al. [131] report that, as
concentration proceeds, so the viscosity and surface free energy of the latex
undergo the changes indicated in Figs 1O.20(a) and (b) respectively. No further
increase of average particle size occurs after a total solids content of ca.
65% m/m has been attained. Concentration may be continued until the total
solids content is ca. 70% m/m. The curves of Figs 10.20(a) and (b) indicate
that little agglomeration occurs until the total solids content of the latex
reaches ca. 55% m/m. Rapid agglomeration then takes place and continues
until the total solids content reaches ca. 65% m/m. Presumably at this stage
the system attains a new colloidal equilibrium, as little further change in
surface free energy occurs as more water is removed. By selecting an
appropriate level of polyvinylmethyl ether, it is said to be possible to ensure
that agglomeration can be kept in advance of concentration, so that a
10-3~------~-------L------~------~------~
~ ~ W w ro w
Total solids content of latex / % m/m
(8)
70
'"'E
..., 60
E
.......
>-
Cl
CD
c
Q) 50
Q)
~
Q)
0
«I
't:
:J
en 40
3%·~O-------4~O--------~J-O------~~~O-------7~IO------~~O
Total solids content of latex / % m/m
(b)
Figure 10.20 Changes in: (a) viscosity, (b) surface free energy, during concentration of a
styrene-butadiene rubber latex in presence of modified polyvinylmethyl ether as agglomerant
(Howland el al. [131]). Significance of points: • 0.02,.t. 0.03,.0.04 pphls polyvinylmethyl ether.
relatively low viscosity is maintained throughout concentration. This is

advantageous, especially in relation to heat transfer.
Westerhoff, White and Gilbert [132] have reported the effects of certain
process variables upon the amount of coagulum produced in the agglomeration
and concentration of styrene-butadiene rubber latices in the presence of
polyvinylmethyl ether. Their conclusions can be summarized as follows:
1. The amount of coagulum increases both with increasing level of
polyvinyl methyl ether used and with decreasing plasticity of the polymer
dispersed in the latex. The effect of decreasing polymer plasticity is
presumably to reduce the ease of particle coalescence, thereby retarding
the mechanism by which the agglomeration becomes self-terminating.
2. The amount of coagulum produced increases sharply as the mass fraction
of polymer in the initial latex present as particles of diameter greater than
120 nm increases. This is interpreted as being a consequence of the
dependence of collision freq uency upon the sq uare of the particle diameter,
as well as upon the square of the particle concentration. Thus the
particle-size distribution of the initial latex should be fairly uniform if
excessive formation of coagulum is to be avoided.
3. Agitation increases both the rate of agglomeration and the amount of
coagulum which forms. Some agitation is required, but this should be
mild. However, just prior to agglomeration, agitation is necessary in order
to homogenize the mixture of latex and agglomerating agents.
4. Temperature has a significant effect upon coagulum formation, especially
for latices of small initial particle size. A small increase in temperature
during evaporation can be accompanied by an increase in the amount of
coagulum, indicating the need to maintain high vacuum during concentration.
White, Reynolds and Gilbert [133] have reported the results of an
investigation of the kinetics of particle agglomeration in the presence of
polyvinylmethyl ether. To simplify the interpretation of the results, the
concentration step was omitted. Agglomeration was investigated at 50·C
using latices of predetermined total solids contents, under conditions such
that the total solids content remained constant. Under these conditions, the
viscosity of the latex decreased as agglomeration proceeded. It is significant
that the surface free energy of the latex decreased concurrently with the
viscosity decrease, confirming that particle coalescence as well as particle
aggregation was occurring. The kinetics of particle aggregation were found
to differ from classical Smoluchowski kinetics, in that the relationship between
particle concentration, N, and time, t, was of the form
1 1
- =- + at - bt 2 (10.60)
N No
where No is the initial particle concentration and a and b are constants.
Rates of agglomeration were found to be ca. 10- 1 1 - 10- 12 of the theoretical
Smoluchowski rapid rate of flocculation, indicating high latex colloid stability

even in the presence of the polyvinyl methyl ether. The negative coefficient
for the t 2 term indicates that the latex colloid stability increased as
agglomeration proceeded, and that the latex reached a stable state after a
finite time, thereby confirming that the agglomeration was self-terminating.
The principal reason for self-termination was postulated to be saturation of
the particle surface with carboxylate soap anions as the specific surface area
of the particles decreased. A secondary effect may arise from increased ionic
strength of the aqueous phase as soap ions are released into that phase,
reducing the effectiveness of the polyvinyl methyl ether agglomerant by
decreasing its degree of molecular expansion, or even by salting it out of
solution. As expected, the extent of agglomeration at which colloidal
equilibrium was reached was found to increase with increasing level of
polyvinylmethyl ether and increasing latex total solids content, and to
decrease with increasing level of carboxylate soap. An unexpected effect is
that increasing the level of carboxylate soap was found to increase the rate
of agglomeration; a possible explanation is increase in the ionic strength of
the aqueous phase as the soap level increased. The conclusion reached was
that the mechanism of agglomeration may involve interparticle bridging.
There is still considerable doubt concerning the mechanism by which
particle agglomeration occurs in this process. The latex heat-sensitization
effect of polyvinylmethyl ether, described in section 5.3.3 of Chapter 5 (Volume
1), does not appear to be an essential feature of the mechanism, because
agglomeration occurs even if the evaporation is allowed to proceed slowly
at normal ambient temperatures. Furthermore, it appears that some
agglomeration can occur very slowly at normal ambient temperatures even
although no evaporation occurs. Whatever may be the precise mechanism
of agglomeration, it is undoubtedly significant that those polyvinylmethyl
ethers which are most effective as agglomerants are those which are on the
borderline between solubility and insolubility in water at normal ambient
temperatures. They are therefore probably present in the aqueous phase of
the latex in an incipiently insoluble form, possibly as networks. In this form,
they are expected to present a relatively large specific surface area which can
compete for the colloid stabilizers present in the system. The latex particles
tend to be partially denuded of adsorbed stabilizer molecules, and therefore
to be partially destabilized. Agglomeration then tends to occur, perhaps by
the mechanism of interparticle bridging, until a new equilibrium is attained
between the stabilizer adsorbed on the enlarged particles, that which is
adsorbed by the incipiently-insoluble hydrocolloid, and that which is
dissolved in the aqueous phase.
Various patents [e.g., 134-141] have been granted covering the use of
polyvinylmethyl ether as an agglomerant for synthetic latices, either alone
or, more usually, in combination with another destabilizative influence such
as an electrolyte addition or pH reduction. Clearly, a large number of
combinations of destabilizative influences is in principle possible. Of particular

interest is a patent [138] which describes the treatment of polyvinyl methyl
ether to enable it to effect agglomeration in the absence of other chemical
destabilizative influences, and greatly to enhance agglomeration in the
presence of electrolytes. In the preferred treatment, the polymer is heated
with or without acidic substances such as sulphuric acid, hydrochloric acid,
acetic acid or iron(III) chloride. The increased effectiveness as an agglomerant
is said to be a consequence of the crosslinking of the polymer; the concomitant
changes in molecular mass are not thought to be important. As far as is
known, processes which use polyvinyl methyl ether for partially agglomerating
synthetic rubber latices have never been widely used industrially, notwithstanding
their novelty and apparent attractiveness, and their intrinsic interest. The
problem seems to have been control ofthe process; in particular, the criticality
of the treatment of the polyvinylmethyl ether in processes which use the
modified polymer has presented serious problems in practice.
Several other water-soluble hydrocolloids have been used to promote
partial agglomeration in synthetic latices. These include polyvinyl alcohol,
polyethylene polyamines, polyvinyl pyridine, various water-soluble poly-
urethanes, ethylhydroxyethyIcellulose, and oxidized polyalkylene oxides. It
is doubtful whether any ofthese processes has ever been exploited industrially
to any significant extent.
10.4.2.2.5 Agglomeration by addition of surface-active substances

Various processes have been proposed in which partial agglomeration is
achieved by the addition of surface-active substances. Colloidal destabilization
is expected if surface-active ions of opposite polarity to the latex particles
are added to a latex. More unexpectedly, processes have been described in
which agglomeration is brought about by the addition of surface-active
substances which would be expected to increase the colloid stability of the
latex. Thus a patent to Polymer Corporation [142] claims partial agglomeration
of synthetic rubber latices by the addition of specified non-ionogenic ethylene
oxide adducts. The preferred compounds are obtained by reacting a polyethylene
glycol with the diepoxides obtained by reacting epichlorohydrin with a
polyhydric phenol. It is claimed that latices containing these additives are
mechanically stable, and can be concentrated to a higher total solids content
than is possible in their absence. It is not clear whether these additives should
be regarded as functioning as non-ionogenic surfactants or as water-soluble
hydrocolloids.
10.4.2.3 Agglomeration by physical methods

10.4.2.3.1 Agglomeration by subjecting the latex to freezing and thawing
It has been found that controlled partial agglomeration in synthetic latices
can be induced by subjecting the latex to freezing and then subsequent
thawing. This method has been very important industrially in the past as a
method of partially agglomerating styrene-butadiene rubber latices, but is

believed to have now been largely, if not entirely, superseded by methods
which depend upon subjecting the latex to mechanical shearing. The principal
reasons for the widespread adoption of this method were:
1. ease of control;
2. the absence of any additives which increase the non-polymer solids content
of the latex;
3. the rapid rate at which agglomeration occurs once the appropriate
destabilizative conditions have been established;
4. the high rate at which agglomerated latex can be produced relative to
rates of production using chemically-induced destabilization;
5. the absence of any tendency for agglomeration to proceed further during
the subsequent concentration process, or during storage.
Methods in which agglomeration is effected by mechanical agitation also
show these advantages. However, these latter methods have the additional
advantage that there is no requirement for the provision of a facility to enable
the latex to be rapidly frozen. This is the principal reason for the supersession
of freeze-thaw methods by methods which colloidally destabilize the latex
by mechanical agitation. In contrast to the chemical methods of achieving
partial agglomeration which have been discussed so far, agglomeration by
freezing and thawing does not necessarily involve any reduction in the
inherent colloid stability of the latex. Rather, it is primarily the restriction
of volume available to the latex particles, together with the pressures and
shear stresses induced into the latex by freezing, which are the primary
destabilizative influences which bring about partial agglomeration. These
effects are discussed in section 4.5 of Chapter 4 (Volume t). It is, of course,
possible in principle to supplement the destabilizative effect of freezing and
thawing by other destabilizative influences, such as solvent addition and pH
reduction.
A typical process for the partial agglomeration of styrene-butadiene rubber
latices by freezing and thawing is illustrated schematically in Fig. 10.21.
Initially the latex has a total solids content of ca. 28% mlm and a carboxylate
soap content equivalent to ca. 30-40% coverage of the particles. It is first
concentrated to ca. 40% mlm total solids content by evaporation, partly so
that, for a given throughput of polymer, the volume of latex to be frozen
and thawed is reduced, and partly so that the agglomeration shall be more
effective. The pH is adjusted after evaporation, the level being determined
by the degree of agglomeration which it is desired to achieve in relation to
the severity of the destabilizative conditions to be established subsequently.
The agglomeration itself is effected by allowing a thin film of latex to be
picked up by a rotating refrigerated drum. The film rapidly freezes. It is then
scraped off the drum by a blade as the drum revolves. The chips of frozen
latex fall into a container, where they thaw rapidly under gentle agitation.
Evaporator Evaporator
To 40"10 solids To 63"10 solids
28"10 63"10
solids solids
Storage agglomeration Thawing Settling pH Storage

tank for adjustment tank for
initial latex concentrate
Figure 10.21 Schematic illustration of typical process for partial agglomeration of styrene-
butadiene rubber latices by freezing and thawing.
The agglomerated latex is then left to stand, allowing any coagulum which
has formed to rise to the surface. The pH is adjusted again, and the latex
concentrated by evaporation to ca. 63% mlm total solids content.
The surface free energy of the latex decreases sharply as a consequence of
freezing and thawing. This is visually apparent because the thawed latex
foams freely, in contrast to the latex before freezing. It is indicative of the
liberation of hitherto adsorbed soap ions into the aqueous phase of the latex,
the average particle size of the agglomerated latex being such that the soap
content is now equivalent to ca. 115%. Other changes occur as a consequence
of freezing and thawing. Before freezing, the latex is somewhat translucent,
whereas at the moment of complete freezing it changes to an opaque white.
The whiteness and opacity are characteristic of latices which contain large
particles. This appearance persists after thawing. There is also an increase
in the pH of the latex of ca. 1. This is presumably also a consequence of the
liberation of soap ions into the aqueous phase.
The effects of various process variables upon agglomeration by freezing
and thawing have been reported by Talalay [143]. The following is a summary
of the conclusions which have been drawn concerning these matters:
1. Effect of soap coverage As expected, freeze-thaw agglomeration becomes

progressively less effective as the soap content of the initial latex is
increased. At soap contents equivalent to a slight excess of that required
for complete saturation of the particles (ca. 120%), there is no agglomeration
at all.
2. Effect of pH As the pH at which the latex is frozen is lowered, so some
of the carboxylate soap is converted to un-ionized free carboxylic acid,
thereby reducing the extent of soap coverage of the particles. However,
the pH cannot normally be reduced below ca. 8.5 without extensive

coagulation occurring when the latex is frozen. As would be expected, the
effect of pH reduction is more pronounced at high soap coverage than at
low. The pH of the initial latex is usually between 10 and 11. Typically
this is reduced to ca. 9 prior to freezing, the reduction being effected by
the addition of either acid or carbon dioxide. As the pH of the initial latex
is reduced, the process becomes increasingly one in which the destabilizative
effect offreezing and thawing is supplemented by that of partial destruction
of colloid stabilizer.
3. Effect of temperature and duration offreezing As would be expected, the
lower the temperature to which the latex is frozen, and the longer the
latex is maintained at that temperature, the more extensive is the
agglomeration. There is a tendency for very large particles to form as the
duration of freezing is increased.
4. Effect of chemical constitution of carboxylate soap There is a minimum
temperature, called the critical freeze temperature, to which a latex can
be frozen without appreciable coacervation being apparent when the latex
is subsequently thawed. The critical freeze temperature of a latex which
contains a given type of polymer depends primarily upon the chemical
constitution of the latex colloid stabilizer, and not upon the degree of
saturation of the particle surface. For fatty-acid soaps, the critical freeze
temperatures are found to increase in the same order as do the melting
points of the corresponding fatty acids. The critical freeze temperature for
latex stabilized with oleates is especially low (ca. - 50°C). The existence
of a critical freeze temperature may be associated with the reluctance of
fatty acids and their anions to spread at interfaces below a certain
temperature. It appears that the stabilization of the frozen latex requires
that the stabilizing carboxylate ions at the particle surface be mobile to
a certain extent. Latices which are stabilized with rosin-acid soaps
exclusively have been found to be rather difficult to agglomerate by the
freeze-thaw method. Those stabilized with mixtures of rosin-acid and
fatty-acid soaps can, however, be agglomerated satisfactorily by this
method.
5. Effect of polymer hardness Generally it is found that latices which contain
hard, non-rubbery polymers tend to give massive coagulum when subjected
to the freeze-thaw agglomeration process. A similar, though less pronounced,
effect is observed with latices which contain rubbery polymers of very
high viscosity. For the latter, the effect is not so severe; coagulum does
not necessarily form, but the agglomerated particles tend to be of irregular
shape and the agglomerated system tends to be colloidally rather unstable.
In both cases, the effects are almost certainly a consequence of imperfect
particle coalescence. The latex system is therefore unable to reduce the
specific surface area of polymer-aqueous phase interface to a value which
corresponds to complete saturation with the available soap.
The critical freeze temperature of styrene- butadiene rubber latices containing

a given type of carboxylate soap depends upon the composition of the
copolymer. Results reported by Mitchell [144], shown here as Fig. 10.22,
demonstrate the nature of this dependence. As the styrene content of the
copolymer is increased, so the critical freeze temperature decreases sharply
at first, until a minimum is reached when the styrene content of the dispersed
copolymer is ca. 35% m/m. Thereafter, the critical freeze temperature
increases rapidly with further increase of the styrene content of the dispersed
copolymer.
The above conclusions have been confirmed qualitatively by Daniels,
Watson and White [98]. Kazakevich et al. [145] have reported the results
of a Russian investigation of the freeze-thaw agglomeration of carboxylate-
stabilized synthetic styrene-butadiene copolymer latices. The potassium salts
of synthetic fatty acids were used as the colloid stabilizer. The extent of
agglomeration was assessed as the ratio of the degree of soap saturation of
the particles after agglomeration to the initial degree of soap saturation. This
ratio is a reasonable index of the extent of agglomeration provided that the
latex does not coagulate to any significant extent during the agglomeration
process. Experiments using a 30/70 m/m styrene/butadiene copolymer latex
showed that, whereas the extent of agglomeration increased somewhat with
the rate and duration of freezing, the temperature of freezing had a greater
effect. Optimum conditions of temperature and time of freezing could be
determined, these conditions being such as to maximize the extent of
agglomeration whilst minimizing the amount of coagulum formed. Table 10.5
20
~OO 20 40 60 80
Styrene content of copolymer / %m/m
Figure 10.22 Effect of polymer composition upon critical freezing temperature for styrene-
butadiene rubber latices stabilized with a given type of carboxylate soap (Mitchell [144]).
Table 10.5 Effect of pH and soap coverage upon freeze-thaw agglomeration of a

carboxylate-stabilized 50/50 m/m styrene/butadiene copolymer latex under optimum
conditions of freezing* (Kazakeyich el al. [145])
Properties after agglomeration

Optimum conditions
Initial offreezing Surface
soap free Soap Extent of
coverage Initial Temperature Time energy coverage agglomer-
/% pH j"C /min / mJm- 2 /% ationt
35.0 8.4 -16.0 to 41.3 78.0 2.23

35.0 9.0 -19.5 10 42.6 76.5 2.18
35.0 9.9 -10.0 15 52.1 61.5 1.75
35.0 11.0 -10.0 15 52.1 58.5 1.67
49.2 8.4 -24.5 to 35.8 96.0 2.00

49.2 9.0 -31.0 12 38.2 88.0 1.80
49.2 10.0 -13.0 12 48.5 65.0 1.32
49.2 11.0 -12.0 12 50.5 62.0 1.26
* See text for explanation.

t See text for method of assessment; the entries are those given in the Kazakeyich el al. paper.
shows their results for the effect of pH and soap coverage upon freeze-thaw
agglomeration of a 50/50 m/m styrene/butadiene copolymer latex under
optimum conditions of freezing. Table 10.6 shows results for the effect of
temperature and time of freezing upon freeze-thaw agglomeration of the
Table 10.6 Effect of temperature of freezing upon freeze-thaw agglomeration of a

carboxylate-stabilized 50/50 m/m styrene/butadiene copolymer latex (Kazakeyich et
al. [145])
Properties after agglomeration

Conditions
Initial of freezing Surface
soap free Soap Extent of
coverage Initial Temperature Time energy coverage agglomer-
/% pH j"C / min / mJm- 2 /% ationt
35.0 8.9 -10.0 15 49.0 68.3 1.95

35.0 8.9 -15.0 15 45.6 75.0 2.14
35.0 9.0 -18.5* to* 44.8 76.4 2.18
49.2 9.0 -15.0 15 46.3 77.5 1.60

49.2 8.9 -23.0 15 42.5 82.0 1.70
49.2 9.0 -31.0* 12* 40.6 86.0 1.80
* Optimum conditions of freezing; see text for definition.

t See text for method of assessment; the entries are those given in the Kazakeyich er al. paper.
same latex under optimum and non-optimum conditions of freezing. These

two sets of results show the expected trends.
The mechanism of the freeze-thaw agglomeration process probably in-
volves the pressing together of the latex particles by the ice crystals which
form and grow as freezing progresses. It has been calculated that internal
pressure of ca. 3.5 x 10 2 MPa may be established locally during the freezing
of water. Talalay [143] has investigated the mechanism of freeze-thaw
agglomeration using a large-particle, almost monodisperse styrene- butadiene
rubber latex obtained by double centrifugation of an industrial high-solids
latex. The particles were sufficiently large (ca. 600 nm) to be visible through
an optical microscope. Susceptibility to freeze-thaw agglomeration was
enhanced by allowing them to absorb benzene. This softened the polymer
and decreased the effective degree of soap coverage of the particles. It was
observed that, on freezing, the ice crystals which formed were virtually devoid
of latex particles. As the crystals advanced from the cold source across the
microscope field, so they entrapped groups of particles within the intercrystalline
regions. In the frozen state, the particles were seen to be lined up in reticulate
patterns at the boundary lines between the ice crystals. In those experiments
which resulted in coacervation of the latex particles, the chain-like structures
of the frozen state persisted after the latex had been thawed. However, when
no coagulum was apparent in the thawed latex, it was possible to see some
of the chains of particles remaining clustered together and subsequently
becoming spheroidal on warming.
There are various patents covering the freeze~ thaw agglomeration of
synthetic latices. A patent to Talalay [146] discloses the basic technology of
the process. Reynolds [147] claims a freeze-thaw agglomeration process in
which the latex contains unreacted monomer. The presence of the latter
supplements the agglomerating effect of freezing and thawing by acting as a
polymer solvent, and is said to confer two improvements:
I. the particle size after agglomeration is further increased; and
2. the colloid stability of the latex during the agglomeration process is
increased, as is evidenced by reduction in the amount of macroscopic
coagulum which forms.
Other patents [e.g., 148~ 153] disclose various other facets of the freeze~thaw
agglomeration processes for synthetic latices.
10.4.2.3.2 Agglomeration by subjecting the latex to mechanical agitation

In these methods, the primary destabilizative influence which brings about
partial agglomeration of the latex particles is mechanical agitation. This
increases the frequency and violence between the particles, as is discussed in
section 4.6.1 of Chapter 4 (Volume 1). Like freeze-thaw agglomeration, and
again in contrast to the chemical agglomeration methods, mechanical
agglomeration is not necessarily accompanied by any reduction in the
inherent colloid stability of the latex. Rather, it is primarily pressures and

shear stresses induced into the latex which bring about agglomeration. It is
possible in principle to supplement the destabilizative effect of mechanical
agitation by other destabilizative influences, such as solvent addition and
pH reduction. Furthermore, as is evident from some of the above discussion,
mechanical agitation is sometimes used as a supplementary destabilizative
influence in conjunction with chemical influences.
Latex agglomeration processes in which the principal destabilizative
influence is mechanical agitation can be subdivided into two types:
1. those in which the mechanical agitation is achieved by pumping the latex
through a confined space; and
2. those in which the mechanical agitation is achieved by ultrasonic means.
Of these, only the first type has achieved any industrial importance.
Mechanical agglomeration processes ofthe first of these types are collectively
known as pressure agglomeration. In the usual practical embodiment of this
type of process, the latex is forced under relatively low pressure through the
orifice of a homogenizer. Normally, such equipment is used to reduce the
size of the droplets in lyophobic systems such as oil-in-water emulsions.
However, in latex agglomeration processes use is made of the destabilizative
influence of the pressures and shear stresses which accompany the cavitation
induced in the latex as it is forced through the homogenizer orifice. By suitable
adjustments of the process variables, it is possible to produce latices which
have exceptionally favourable viscosity-total solids relationships. Indeed, it
has been claimed that it is possible by this method readily to obtain synthetic
rubber latices which have a viscosity of only ca. 2 Pa s at a total solids content
of 76% m/m. A second important advantage of this agglomeration method
is that only very small amounts of coagulum (typically less than 0.1 % m/m)
are usually formed.
Figure 10.23 illustrates schematically a typical procedure for industrial
implementation of the principle of agglomeration by passing the latex through
a homogenizer. The latex is first concentrated by evaporation to ca. 35% mlm
total solids content, to increase the throughput of polymer and to facilitate
particle agglomeration. The pH of the latex is then reduced from ca. 10.0 to
ca. 9.3, after which the latex is pumped through the homogenizer orifice at
a pressure of ca. 25 MPa. The latex is then held in surge tanks prior to being
concentrated by evaporation to a total solids content in the range 62-68% m/m.
The process is essentially continuous. It is claimed that the process is more
economical than freeze-thaw agglomeration, because there are no refrigeration
costs. During agglomeration by this method, the effective soap coverage of
the particles typically increases from ca. 40% to ca. 110%, and the pH
increases from ca. 9.3 to ca. 10.3.
The following is a summary of the effects of the more important process
variables upon agglomeration achieved by this method as they relate to
Evaporator Evaporator
Storage tank
for initial latex
1
Feed
tank
Homogeo••,
Surge
tank
Feed
tank
Storage tank
for concentrate
Figure 10.23 Schematic illustration of typical process for partial agglomeration of styrene-
butadiene rubber latices by passing the latex through a homogenizer (Jones [101]).
synthetic rubber latices stabilized by adsorbed carboxylate soaps, as reported

by Jones [101]:
I. Effect of pH As would be expected, the particle size after agglomeration
increases as the initial pH of the latex is reduced. The relationship between
the intial pH and the total solids content at which the viscosity of the
subsequent concentrate is 2 Pa s has been found to be linear and of negative
slope. These effects are presumably a consequence of partial conversion
of the carboxylate soap anions which stabilize the latex to un-ionized
carboxylic acid, thereby decreasing the effective initial soap coverage of
the latex particles. However, agglomeration by this method is not entirely
dependent upon the induction of controlled colloidal instability.
2. Effect of soap content For any given combination of process conditions,
the total solids content corresponding to a viscosity of 2 Pa s for the
subsequent concentrate varies linearly with the ratio of soap to polymer,
increasing as the latter ratio decreases. It appears that the agglomerating
effect of pressure and cavitation is very sensitive to this particular process
variable. Latices stabilized with rosin-acid soaps can readily be agglomerated
by this method; this is in marked contrast to freeze-thaw agglomeration
(see section 10.4.2.3.1 above).
3. Effect of other variables As expected, the extent of agglomeration is
generally found to increase with increasing initial total solids content of
the latex, increasing homogenizer pressure, and decreasing temperature
of the initial latex.
Results given by Jones for the effects of pressure agglomeration upon a
synthetic latex under typical process conditions are shown in Table to.7.
The basic technology of the pressure-agglomeration process is described
in a patent to Bennett [154]. Tables 10.8, 10.9, 10.10 and 10.11, which are
Table 10.7 Effect of pressure agglomeration under typical processing conditions

upon properties of a synthetic rubber latex (Jones [101])
Agglomeration temperature: normal ambient.
Agglomeration pressure: 24.1 MPa.
Total solids content of latex: 35% m/m.
Property Initial latex Agglomerated latex
Soap content I pphr 4.78 4.78

Average particle diameter I nm 60 175
Soap coverage I % 40 110
pH 9.3 10.3
Coagulum content I % o 0.01
Table 10.8 Effect of initial latex temperature at two pH levels upon pressure
agglomeration of a synthetic styrene-butadiene rubber latex (Bennett [154])
Conditions of agglomeration:
Agglomeration pressure: 31.0 MPa
Total solids content of latex: 46% m/m
Initial soap coverage: 41.0%
Initial average particle diameter: 67 nm
Initial Soap coverage Average particle

latex afier diameter afier Coagulum
temperature Initial agglomeration agglomeration content
rC pH 1% 1m 1%
40 8.75 75.0 122 0.02
24 8.75 111.0 180 0.05
40 8.00 85.0 138 0.03
26 8.00 \34.0 220 0.13
based upon results published in this patent, confirm, and indicate the
magnitudes of, the effects of some of the process variables upon the process.
The latex to which these results refer was a 30/70 m/m styrene/butadiene
rubber latex produced by low-temperature emulsion polymerization using a
redox initiator and potassium oleate as the colloid stabilizer. Three other
patents to Bennett [155-157] cover respectively pressure co-agglomeration
of synthetic rubber and resin latices, pressure agglomeration of blends of
styrene-butadiene rubber and natural rubber latices, and low-pressure
agglomeration. Halper, York and Moss [158] have described a process for
the pressure agglomeration of polystyrene latex. The particles are softened
by the presence in the latex of styrene monomer, thus enabling particle
agglomeration to take place under normal pressure-agglomeration conditions.
The process is effectively one of combined pressure and solvent agglomeration;
Table 10.9 Effect of initial pH upon pressure agglomeration of a synthetic

styrene-butadiene rubber latex (Bennett [154])
Initial soap coverage: 44.1 %
Initial latex temperature: normal ambient
Soap coverage alier Average particle diameter alier

agglomeration agglomeration
Initial pH 1% Inm
10.0 108.4 178

11.0 99.8 164
12.0 84.1 127
13.0 52.3 79
Table 10.10 Effect of agglomeration pressure upon pressure agglomeration of a

synthetic styrene- butadiene rubber latex (Bennett [154])
Initial soap coverage: 41.0%
Initial latex temperature: 12°C
A verage particle
Pump Soap coverage diameter a/ier Coagulum
pressure Initial alier agglomeration agglomeration content
I MPa pH 1% Inm 1%
13.8 9.5 93.0 150 0.04
20.7 9.5 \07.0 174 0.04
34.5 9.5 131.0 214 0.02
under normal pressure-agglomeration conditions, polystyrene particles are

too hard to fuse together.
Some interest has also been shown in a mechanical-agglomeration process
in which the necessary conditions are established by exposing the latex to
ultrasonic vibrations. Presumably the necessary agglomerative conditions of
pressure and shear stress are developed as a consequence of cavitation within
the latex. This process is described in a patent to Roberts [159]. One method
of carrying out the process is to pump the latex at a pressure of ca. 1 M Pa
through a thin elliptical jet so as to form the latex into a substantially flat
stream. The latex stream is then caused to impinge at high velocity on to a
thin blade of metal or other suitable material which is mounted in cantilever
form. This blade is excited into ultrasonic vibrations by the impingement of
Table 10.11 Effect of total solids content upon pressure agglomeration of a synthetic
styrene-butadiene rubber latex (Bennett [154])
Initial pH of latex: 8.0
Initial soap coverage: ca. 40%?
Initial average particle diameter: ca. 65 nm?
Initial latex temperature: normal ambient?
Total Average particle
solids Soap coverage diameter after Coagulum
content after agglomeration agglomeration content
/%m/m /% / nm /%
18.5 44.4 77 0
25.0 70.0 123 0
30.0 75.7 132 0
36.0 97.0 168 0
42.0 99.0 173 0
the latex stream, and may vibrate at its natural frequency, e.g., ca. 2 x 104 Hz.
The effectiveness of the process in agglomerating the latex depends upon the
process conditions, and upon the number of times the latex is submitted to
the treatment.
10.4.3 Concentration of synthetic latices
10.4.3.1 Concentration by creaming

It is difficult to concentrate synthetic rubber latices by centrifugation, whether
or not the latex has first been agglomerated. This is because of the small size
of the particles. Creaming has, however, been found to be feasible, especially
if the particles size has first been increased by agglomeration. The theory of
creaming is expounded in section 9.4.3.1 of Chapter 9, in connection with
the concentration of natural rubber latex. The theory leads to the conclusion
that particle size has a profound effect upon both the equilibrium distribution
of particles in a latex after creaming, and also the rate at which creaming
takes place. The smaller is the average size of the particles in the latex, the
greater is the concentration of creaming agent in the aqueous phase required
to effect a given degree of creaming. The larger the particle size, the more
depleted is the skim and the more concentrated is the cream produced by
any given level of creaming agent. The creaming behaviour of 40/60 m/m
styrene/butadiene rubber latex to which ammonium alginate had been added
as a creaming agent has been reported by Schmidt and Kelsey [119]. The
latex was produced by low-temperature emulsion polymerization using an
iron pyrophosphate-hydroperoxide-dextrose redox initiation system, and
the sodium soap of dis proportionated rosin acid as colloid stabilizer. The
particle size of the resultant latex was increased controllably by e1ectrolyte-
agglomeration using sodium chloride (see section 10.4.2.2.2 above). By
selecting suitable electrolyte concentrations and growth times, a series of
latices of different average particle sizes was obtained for the investigation
of creaming behaviour. The results of Schmidt and Kelsey for the effects of
latex particle size and of the aqueous-phase concentration of creaming agent
(ammonium alginate) upon the polymer concentrations in the cream and the
skim are shown in Figs 1O.24(a) and (b) respectively. The polymer content
of the initial latex was 5.15% m/m, and the sodium chloride concentration
0.0808 mol dm - 3, the latter being sufficiently low to ensure that no further
electrolyte-induced particle agglomeration occurred during creaming. After
addition of the ammonium alginate, the mixtures were stirred for 5 minutes
and allowed to cream for 3 days. The polymer content of the cream passed
through a maximum as the concentration of ammonium alginate increased.
The polymer content of the cream at any given concentration of ammonium
alginate depended greatly upon the particle size of the latex. At low
concentration of ammonium alginate, the polymer content of the cream
increased sharply with increasing latex particle size, but the trend reversed
at higher concentrations of ammonium alginate. The polymer content of the
skim decreased sharply as the concentration of ammonium alginate increased.
The concentration at which the sharp decrease occurred decreased as the
particle size of the latex increased. The larger the particle size of the latex,
the lower was found to be the polymer content of the skim at any given
concentration of ammonium alginate; but whether or not the polymer content
of the cream increased with .increasing particle size depended upon the
ammonium alginate concentration. The creaming behaviour of synthetic
rubber latices has been found to depend little upon either the nature or the
amount of soap stabilizer in the latex. Thus Schmidt and Kelsey also
investigated the creaming behaviour of a 30/70 m/m styrene/butadiene rubber
latex produced by high-temperature emulsion polymerization using potassium
peroxodisulphate as initiator and potassium castor-oil soap as colloid
stabilizer. Ammonium alginate was again the creaming agent. The behaviour
was found to be similar to that of the latex stabilized with rosin-acid soap.
The following procedure has been suggested for the creaming of a
styrene-butadiene rubber latex: 3 pphls of sodium sulphate and 0.8 pphls of
ammonium alginate are added successively to the latex. The sodium sulphate
is added as a 10% m/v aqueous solution, the pH of which has been adjusted
to 10.0 by addition of potassium hydroxide. The solution should be added
to the latex slowly with gentle agitation, and the pH of the latex then checked
to make sure that it exceeds 10. The ammonium alginate is then added slowly
as a 2% m/v aqueous solution. The stirring is continued for at least a further
30 minutes, after which the latex mixture is allowed to stand and cream.
Separation into two distinct layers should occur within 2 hours, but the
70
~ 60
~
~ 50
E
~
III
~ 40
'0
~ 30
'E
0
0
Q) 20
.0
.0
:::J
a:: 10
00 0.2 0.4 0.6

Ammonium alginate level/ % mlm on aqueous phase
(a)
6
E
E 5
--
~
0
E 4
32
rJl
'0 3
'E
~0
0
2
Q)
.0
.0
:::J
a::
00 0.2 0.4 0.6
Ammonium alginate level/ % mlm on aqueous phase
(b)
Figure 10.24 Effects of latex particle Si7.e and aqueous-phase concentration of creaIning agent
(ammonium alginate) upon polymer concentration in: (a) cream; (b) skim, of a 3O;70mjrn
styrene;butadiene rubber latex produced by high-temperature emulsion polymerization (Schmidt
and Kelsey [119]). Polymer content of the initial latex was 5.15%, and the sodium chloride
concentration 0.0808 mol dm - 3. Significance of points: • 79 nm mass-median-particle-diameter
latex; ... 85 nm mass-median-particle-diameter latex; _ 114 nm mass-median-particle-diameter
latex; ... 265 nm mass-median-particle-diameter latex; • 360 nm mass-median-particle-diameter latex.
concentration of the upper layer will continue to increase for up to 72 hours.

Any tendency to colloidal instability should be corrected by addition of up
to 1 pphls of stabilizer. In the laboratory, it is convenient to carry out the
creaming in a separating funnel, so that the skim can be run from the bottom.
By this method, it is said to be possible to produce a cream which contains

60-65% m/m total solids content.
10.4.3.2 Concentration by evaporation

The usual method for concentrating synthetic latices is by evaporation. The
concentration process which has been described by Howland et al. [131],
for use in connection with the simultaneous agglomeration and concentration
of styrene- butadiene rubber latices (see section 10.4.2.2.4 above), is illustrated
schematically in Fig. 10.25. The concentrator is approximately half-filled with
latex, so as to allow sufficient free volume to control foaming and entrainment.
A partial vacuum of ca. 100 kPa is established by means of a vacuum pump.
Uniform heating and evaporation rates are ensured by continuously pumping
the latex from the bottom of the vessel through a shell-and-tube heat
exchanger up to the top of the concentrator ready for recirculation. The
equilibrium temperature of the bulk of the latex in the concentrator is ca.
40·C, whereas the temperature at the point of discharge from the heat
exchanger is ca. 55·C. Water evaporates from the superheated latex as the
latex enters at the top of the concentrator. The water-vapour is condensed,
and passes out through a barometric seal.
An essentially similar piece of equipment for the concentration of synthetic
rubber latices has been described by Stell, Blackwell and Henry [160]. The
latex is continuously fed into the evaporator, and a proportion removed
Booster
condenser
Vacuum
pump
latex circulating
pump
Seal tank
Figure 10.25 Schematic illustration of typical process for concentrating synthetic latices by
evaporation (Howland el al. [131]).
Production of non-aqueous synthetic latices 293
continuously as concentrate at the point at which the latex enters the heat
exchanger. In this way, the evaporator can be worked continuously. It is
claimed that the product of concentration by continuous column evaporation
is more uniform in properties than is that which is produced by batchwise
concentration.
In processes where the agglomeration step precedes concentration by
evaporation, the Luwa thin-layer, or turbulent-film, evaporator has been
widely used for the concentration step. The principle of this concentrator is
illustrated in Fig. 10.26. Evaporation takes place from a thin film which is
kept mechanically in continuous turbulent motion, thereby encouraging rapid
and uniform concentration at relatively low temperatures. The concentrator
consists of two vertical cylindrical sections, one on top of the other, with the
feed inlet between them. In the lower section, evaporation of moisture takes
place, whereas in the upper section the separation of water vapour occurs.
A central rotor shaft, having longitudinal blades which project radially from
it, extends through both sections. The blades clear the heated wall in the
lower section by 0.75-1.50mm. Above the inlet is the separator section. This
has longitudinal fins fixed to the inside walls. The latex enters the lower
section, whereupon it is picked up by the rotor blades and whirled against
the heated wall. A thin turbulent layer of liquid is maintained over the entire
heat-transfer surface, because of the small distance between the edges of the
rotor blades and the wall surface. The latex descends in a spiral path down
towards the heated wall to the outlet at the bottom. Vapours ascend to the
separator along the rotor shaft in the centre of the cylinder, thereby being
removed from the evaporator. Any latex which becomes entrained in the
separator section is thrown against the stationary fins along the wall by the
centrifugal action of the rotor blades. This latex then coalesces and flows
back to the evaporator section. It is an important feature of this concentrator
that the latex is held on the evaporator wall for no longer than ca. 20 seconds,
and is therefore subjected to a minimum of heat treatment. The concentration
of styrene-butadiene rubber latices in a turbulent-film evaporator has been
described by Borg, Provost and Bawn [161]. Processes for the continuous
automatic concentration and stripping of synthetic latices are the subject of
a paper by Gudheim [162].
10.5 PRODUCTION OF NON-AQUEOUS SYNTHETIC LATICES

Non-aqueous synthetic latices are produced by what are virtually emulsion
polymerization reactions using, in place of the water of conventional emulsion
polymerization reaction systems, the appropriate non-aqueous dispersion
medium. Such reactions are often referred to as dispersion polymerizations.
Like conventional aqueous emulsion polymerization reactions, most of the
conversion of monomer to polymer takes place within a large number of
Rotor shaft
- - Water vapour outlet
Separator
Fins attached to
separator wall
Feed inlet - -
Ili:tt::t:~,... attached
Separator fins
to rotor
Evaporator /
Heating Heating medium inlets

medium
outlets~
Latex concentrate
Figure 10.26 Illustrating principle of the Luwa thin-layer evaporator.
separate reaction loci which are dispersed in the dispersion medium. The
average number of propagating radicals per reaction locus is again small,
and so termination of propagating chains is retarded relative to propagation.
Unless the reaction loci are provided as pre-formed seed particles, a stage
of locus nucleation, analogous to the Interval I of conventional aqueous
emulsion polymerizations, is necessary. However, non-aqueous emulsion
polymerization reactions typically differ from conventional aqueous emulsion
polymerization reactions using non-polar monomers in several respects, of
which the following are the more important:
Production of non-aqueous synthetic latices 295
1. The reaction system is initially homogeneous, because the monomer and

the dispersion medium are miscible.
2. The mechanism of locus nucleation is non-micellar, because non-aqueous
dispersion media do not encourage micellization of such colloid stabilizers
as are likely to be present. The commonly-held view is that locus nucleation
occurs by the homogeneous mechanism.
3. Because the monomer is not present as a separate droplet phase, there is
no transition analogous to that between Interval II and Interval III of a
conventional aqueous emulsion polymerization reaction. There is certainly
a gradual absorption of unreacted monomer into the reaction loci as
polymerization proceeds and the reaction loci grow in size. As with
conventional aqueous emulsion polymeriz~tion reactions, the extent of
the absorption at any given extent of polymerization is determined by the
condition that the chemical potential of the monomer molecules shall
have the same value in the various phases of the reaction system. What
is different in non-aqueous systems as compared with aqueous systems is
the absence in the earlier stages of the reaction of a separate monomer
droplet phase which maintains the chemical potential of monomer
molecules at a high and constant value for a substantial proportion of
the reaction. The absence of any mechanism for maintaining the chemical
potential of monomer molecules at a high and constant value has
implications for the kinetics of the reaction.
4. Because the dispersion medium is much less polar than water, and hence
of lower relative permittivity, electrostatic colloidal stabilization of the
reaction loci, and of the ultimate latex particles, is not possible. Colloid
stability is usally conferred primarily by steric stabilization, supplemented
possibly by solvation and depletion effects. The conditions of the reaction
are such that the steric stabilizer may become graft copolymerized at the
surface of the growing latex particles. If this happens, colloid stability is
usually enhanced.
No attempt is made here to describe in detail the production of non-

aqueous synthetic latices. For further information, the reader is referred to
a review paper by Walbridge [163], to the book edited by Barrett [24], and
to the references cited in that book. All that can be given here is a brief
description of a typical emulsion/dispersion polymerization reaction system
for the production of poly methyl methacrylate latices for which the dispersion
medium is a paraffin hydrocarbon of medium chain length. The initial reaction
system comprises a solution of methyl methacrylate in the hydrocarbon,
together with an oil-soluble initiator, such as azobisisobutyronitrile, and a
steric colloid stabilizer. A particularly suitable steric stabilizer is a comb-like
graft copolymer which comprises a backbone of polymethyl methacrylate
and essentially hydrocarbon grafts. The polymethyl methacrylate backbone
has a strong tendency to adsorb on to the surface of the polymethyl
methacrylate particles as they develop, whereas the hydrocarbon grafts tend

to mix with the dispersion medium. A steric stabilizer of this type can be
prepared by first forming poly-12-hydroxystearic acid by self-condensation
polymerization. The product should have number-average molecular mass
in the range 1500-2000 daltons, and be monofunctional with respect to
carboxylic-acid groups. Industrial grades of the monomeric acid contain
sufficient free long-chain carboxylic acid to limit the molecular mass of the
polymer. The polymer is converted to a macromolecular monomer by reacting
the free carboxylic-acid groups with glycidyl methacrylate. This monomer is
then copolymerized with methyl methacrylate to give a polymer having
average molecular mass in the range 1-2 x 104 daltons. Each polymer
molecule has ca. 5-10 long hydrocarbon chains derived from stearic acid
moieties pendant from the anchor backbone of polymethyl methacrylate.
Upon heating the mixture of monomer, initiator and colloid stabilizer,
solution polymerization of the monomer commences. Although the monomer
is soluble in the hydrocarbon dispersion medium, the polymer is not. Thus
the polymer precipitates from solution to form reaction loci which are
colloidally stabilized by the steric stabilizer. As polymerization proceeds, an
increasing proportion of the unreacted monomer becomes absorbed into the
reaction loci. Thus, as polymerization reaction proceeds, the nature of the
dispersion medium changes from being a solution of methyl methacrylate in
paraffin hydrocarbon to being almost pure paraffin hydrocarbon. This change
enhances the ability of the colloid stabilizer to function as such, because the
tendency of the dispersion medium to mix with the backbone of the stabilizer
decreases, whereas the tendency to mix with the grafts of the stabilizer
increases. If the polymerization is taken almost to completion, the end-product
of the reaction is a sterically-stabilized latex of polymethyl methacrylate
particles dispersed in the paraffin hydrocarbon.
REFERENCES
I. Bovey, F. A., Kolthoff, I. M., Medalia, A. I. and Meehan, E. 1. (\955) Emulsion

Polymerization, Interscience, New York.
2. Whitby, G. S., Davis, C. C. and Dunbrook, R. F. (eds) (1954) Synthetic Rubber,
John Wiley, New York.
3. Blackley, D. C. (1975) Emulsion Polymerization: Theory and Practice, Applied
Science, London.
4. Blackley, D. C. (1980) Emulsion Polymerization, in Macromolecular Chemistry:
A Specialist Periodical Report, Volume \ (eds A. D. Jenkins and J. F. Kennedy),
Royal Society of Chemistry, London, Chapter 2 Part 6.
A Specialist Periodical Report, Volume 2 (eds A. D. Jenkins and 1. F. Kennedy),
References 297
A Specialist Periodical Report, Volume 3 (eds A. D. Jenkins and J. F. Kennedy),

7. Duck, E. W. (1966) Emulsion Polymerization, in Encyclopedia of Polymer Science
and Technology, Volume 5, John Wiley, New York, p. 801.
8. Gardon, J. L. (1976) Emulsion Polymerization, Theory, in Encyclopedia of
Polymer Science and Technology, Supplement Volume 1, John Wiley, New York.
9. Poehlein, G. W. (1986) Emulsion Polymerization, in Encyclopedia of Polymer
Science and Engineering, Volume 6, John Wiley, New York, p. 1.
10. Ugelstad, J. and Hansen, F. K. (1976) Rubber Chemistry and Technology, 49,
536; (1977) Rubber Chemistry and Technology, SO, 639.
11. Ugelstad, J., Hansen, F. K. and Kaggerud, K. H. (1977) Faserforsch. u Textiltech.,
28,309.
12. Gardon, J. L. (1975) Mechanism of Emulsion Polymerization, in Applied Polymer
Science (eds J. K. Craven and R. W. Tess), Organic Coatings and Plastics
Chemistry Division, American Chemical Society, Washington, DC, p. 138.
13. Gardon, J. L. (1977) High Polymer, Polymer Processes, 29, 143.
14. Dunn, A. S. (1979) Emulsion Polymerisation, in Developments in Polymerisation
(ed. R. N. Haward), Applied Science, London, 2, 45.
15. Gilbert, R. G. (1995) Emulsion Polymerization: A Mechanistic Approach, Academic
Press, New York.
16. Lovell, P. A. and EI-Aasser, M. S. (eds) (1997) Emulsion Polymerization and
Emulsion Polymers, John Wiley, New York.
17. Eliseeva, V.I., Ivanchev, S. S., Kuchanov, S.1. and Lebedev, A. V. (1981) Emulsion
Polymerization and its Applications in Industry, translated from the Russian by
S. J. Teague, Consultants Bureau, New York.
18. Piirma, I. (ed.) (1982) Emulsion Polymerization, Academic Press, New York.
19. Piirma, I. and Gardon, J. L. (eds) (1976) Emulsion Polymerization, ACS
Symposium Series No. 24, American Chemical Society, Washington, DC.
20. Bassett, D. R. and Hamielec, A. E. (1981) Emulsion Polymers and Emulsion
Polymerization, ACS Symposium Series No. 165, American Chemical Society,
Washington, DC.
21. Fitch, R. M. (ed.) (1971) Polymer Colloids, Plenum Press, New York.
22. Fitch, R. M. (ed.) (1980) Polymer Colloids 1/, Plenum Press, New York.
23. Daniels, E. S., Sudol, E. D. and EI-Aasser, M. S. (eds) (1992) Polymer Latexes:
Preparation, Characterization and Applications, ACS Symposium Series No. 492,
American Chemical Society, Washington, DC.
24. Barrett, K. E. J. (ed.) (1975) Dispersion Polymerization in Organic Media, John
Wiley, London.
25. Harkins, W. D. (1945) Journal of Chemical Physics, 13, 381.
26. Harkins, W. D. (1946) Journal of Chemical Physics, 14,47.
27. Harkins, W. D. (1947) Journal of the American Chemical Society, 69, 1428.
28. Harkins, W. D. (1950) Journal of Polymer Science, 5, 217.
29. Smith, W. V. and Ewart, R. H. (1948) Journal of Chemical Physics, 16, 592.
30. Gardon, J. L. (1970) British Polymer Journal, 2, 1.
31. Gardon, J. L. (1970) Rubber Chemistry and Technology, 43, 74.
32. Roe, C. P. (1968) Industrial and Engineering Chemistry, 60(9), 20.
33. Fitch, R. M., Prenosil, M. B. and Sprick, K. J. (1969) Journal of Polymer Science
C, No. 27,95.
34. Fitch, R. M. and Tsai, C. H. (1970) Polymer Letters, 8, 703.
35. Fitch, R. M. and Tsai, C. H. (1971) Homogeneous nucleation of Polymer Colloids.
IV: The Role of Soluble Oligomeric Radicals, in Polymer Colloids (ed. R. M.
Fitch), Plenum Press, New York, p. 103.
36. Fitch, R. M. (1973) British Polymer Journal, 5, 467.

37. Feeney, P. J., Napper, D. H. and Gilbert, R. G. (1984) Macromolecules, 17,2520.
38. Chen, c.-Y. R. and Piirma, I. (1980) Organic Coatings and Plastics Chemistry,
Volume 43. Preprints of papers presented to Division of Organic Coatings and
Plastics Chemistry at the Second Chemical Congress of the North American
Continent (l80th National Meeting of the American Chemical Society), Las
Vegas, Nevada, 1980.
39. Chen, c.-Y. R. and Piirma, I. (1978) Journal of Polymer Science, Polymer
Symposia, 65, 55.
40. Ugelstad, J., EI-Aasser, M. S. and Vanderhoff, J. W. (1973) Journal of Polymer
Science: Polymer Letters Edition, II, 503.
41. Uge1stad, J., Hansen, F. K. and Lange, S. (1974) Die Makromolekulare Chemie,
175,507.
42. Morrison, B. R., Maxwell, I. A., Gilbert, R. G. and Napper, D. H. (1992) Testing
Nucleation Models for Emulsion-polymerization Systems, in Polymer Latexes:
Preparation, Characterization, and Applications (eds E. S. Daniels, E. D. Sudol
and M. S. El-Aasser), ACS Symposium Series No. 492, American Chemical
Society, Washington, DC, p. 28.
43. Wicklatz, J. E., Kennedy, T. J. and Reynolds, W. B. (1951) Journal of Polymer
Science, 6, 45.
44. Fryling, C. F. and Follett, A. E. (1951) Journal of Polymer Science, 6, 59.
45. Carr, C. W., Kolthoff, I. M., Meehan, E. J. and Stenberg, R. J. (1950) Journal
of Polymer Science, 5, 191.
46. Carr, C. W., Kolthoff, I. M., Meehan, E. J. and Williams, D. E. (1950) Journal
47. Kolthoff, I. M., Meehan, E.J. and Carr,C. W.(1951)Journalf?fPolymerScience,6, 73.
48. Helin, A. F., Gyenge, J. M., Beadell, D. A., Boyd, J. H., Mayhew, R. L. and
Hyatt, R. C. (1953) Industrial and Engineering Chemistry, 45, 1330.
49. Roe, C. P. (1971) Journal of Colloid and Interface Science, 37, 93.
50. Roe, C. P. (1971) Emulsion Polymerization with a Surface Active Polyelectrolyte as
the Emulsifier, in Polymer Colloids(ed. R. M. Fitch), Plenum Press, New York, p. 139.
51. Greene, B. W., Sheetz, D. P. and Filer, T. D. (1970) Journal of Colloid and
Interface Science, 32, 90.
52. Greene, B. W. and Sheetz, D. P. (1970) Journal ofCol/oid and Interface Science, 32, 96.
53. Greene, B. W. and Saunders, F. L. (1971) Journal of Colloid and Interface Science,
33,393.
54. Kharasch, M. S., Nudenberg, W. and Mantell, G. J. (1951) Journal ~fOrganic
Chemistry, 16, 524.
55. Karasch, M. S., Nudenberg, W. and Kawahara, F. (1951) Journal of Organic
Chemistry, 16, 1550.
56. Kolthoff, I. M. and Harris, W. E. (1947) Journal of Polymer Science, 2,41.
57. Kolthoff, I. M. and Harris, W. E. (1947) Journal of Polymer Science, 2, 49.
58. Harris, W. E. and Kolthoff, I. M. (1947) Journal of Polymer Science, 2,72.
59. Harris, W. E. and Kolthoff, I. M. (1947) Journal of Polymer Science, 2, 82.
60. Starkweather, H. W., Bare, P.O., Carter, A. S., Hill, F. B., Hurka, V. R., Mighton,
C. J., Saunders, P. A., Walker, H. W. and Youker, M. A. (1947) Industrial and
Engineering Chemistry, 39, 210.
61. Frank, R. L., Smith, P. V., Woodward, F. E., Reynolds, W. B. and Canterino,
P. J. (1948) Journal of Polymer Science, 3, 39.
62. Meehan, E. J., Kolthoff, I. M. and Sinha, P. R. (1955) Journal of Polymer Science,
16, 471.
63. Uraneck, C. A. and Burleigh, J. E. (1965) Journal of Applied Polymer Science, 9, 1273.
References 299
64. Uraneck, C. A. and Burleigh, 1. E. (1970) Journal of Applied Polymer Science, 14,267.
65. Uraneck, C. A. and Burleigh, J. E. (1968) Journal ofApplied Polymer Science, 12, 1075.
66. Krause, A. H. (1954) Rubber Age, New York, 75, 217.
67. Fryling, C. F. and Gindler, E. M. (1959) Proceedings of the International Rubber
Conference, Washington, 1959, p. 456.
68. Howland, L. H., Neklutin, V. c., Provost, R. L. and Mauger, F. A. (1953)
Industrial and Engineering Chemistry, 45, 1304.
69. Wakefield, L. B. and Bebb, R. L. (1950) Industrial and Engineering Chemistry, 42, 838.
70. Antlfinger, G. 1. and Lufter, C. H. (1953) Industrial and Engineering Chemistry, 45, 182.
71. Flory, P. J. (1953) Principles of Polymer Chemistry, Cornell University Press,
Ithaca, New York, Chapter V.
72. Ham, G. E. (ed.) (1964) Copolymerization, Interscience, John Wiley, New York.
73. Meyer, V. E. and Lowry, G. G. (1963) Journal of Polymer Science: Part A, 3, 2843.
74. Greenley, R. Z. (1989) in Polymer Handbook (eds J. Brandrup and E. H.
Immergut), 3rd Edition, John Wiley, New York, p. 11/217.
75. Rohm and Haas Company, British Patent No. 1,414,187, 19 November, 1975.
76. Bassett, D. R. and Hoy, K. L. (1981) Nonuniform Emulsion Polymers: Process
Description and Polymer Properties, in Emulsion Polymers and Emulsion
Polymerization (eds D. R. Bassett and A. E. Hamielec), ACS Symposium Series
No. 165, American Chemical Society, Washington, DC, p. 371.
77. Johnston, J. E., Bassett, D. R. and MacRury, T. B. (1981) Nonuniform Emulsion
Polymers: Carbon-13 NMR Spectroscopy, in Emulsion Polymers and Emulsion
Polymerization (eds D. R. Bassett and A. E. Hamielec), ACS Symposium Series
No. 165, American Chemical Society, Washington, DC, p. 389.
78. Waters, J. A./Imperial Chemical Industries PLC, United States Patent No.
4,997,864, 5 March, 1991.
79. Waters, J. A. (1994) Colloids and Surfaces A: Physicochemical and Engineering
Aspects, 83, 167.
80. Lee, S. and Rudin, A. (1992) Control of Core-Shell Latex Morphology, in Polymer
Latexes: Preparation, Characterization and Applications (eds E. S. Daniels,
E. D. Sudol and M. S. EI-Aasser), ACS Symposium Series No. 492, American
Chemical Society, Washington, DC, p. 234.
81. Lee, D. I. (1981) in Emulsion Polymers and Emulsion Polymerization (eds D. R.
Bassett and A. E. Hamielec), ACS Symposium Series No. 165, American Chemical
Society, Washington, DC, p. 405.
82. Vanderhoff, 1. W., Vitkuske, 1. F., Bradford, E. B. and Alfrey, T. (1956) Journal
83. Bradford, E. B., Vanderhoff, 1. W. and Alfrey, T. (1956) Journal of Colloid Science,
II, 135.
84. Serafini, T. T. and Bobalek, E. G. (1960) Official Digest of the Federation of Paint
and Varnish Producers' Clubs, 32, 1259.
85. Vanderhoff, J. W. (1981) The Formation of Coagulum in Emulsion Polymerization,
in Emulsion Polymers and Emulsion Polymerization (eds D. R. Bassett and
A. E. Hamielec), ACS Symposium Series No. 165, American Chemical Society,
Washington, DC, p. 199.
86. Sfluparek, 1. (1979) Journal of Applied Polymer Science, 24, 909.
87. Sfluparek, J. and Tu(alkova, A. (1979) Journal of Applied Polymer Science, 24, 915.
88. Blackley, D. C. (1983) Production of Carboxylated Latices by Emulsion
Polymerization, in Science and Technology of Polymer Colloids: Volume l:
Preparation and Reaction Engineering (eds G. W. Poehlein, R. H. Ottewill and
J. W. Goodwin), NATO ASI Series E: Applied Sciences No. 67, Martinus Nijhoff,
The Hague, p. 203.
89. Karsa, D. R. (1982) Preprints of Plastics and Rubber Institute Emulsion Polymers
Conference, London, 1982, Paper No.2.
90. Greene, B. W. (1973) Journal of Colloid and Interface Science, 43, 449.
91. Greene, B. W. (1973) Journal of Colloid and Interface Science, 43, 462.
92. Vanderhoff, J. W., van den Hul, H. J. and Hamburg, R. D. (1975) Paper presented
at American Chemical Society International Symposium on Emulsion Polymers,
Philadelphia, USA, 1975; Polymer Preprints, 16(1), 155.
93. Vijayendran, B. R. (1979) Journal of Applied Polymer Science, 23, 893.
94. Greene, B. W., Nelson, A. R. and Keskey, W. H. (1980) Journal of Physical
95. Ceska, G. W. (1974) Journal of Applied Polymer Science, 18, 2493.
96. Ceska, G. W. (1974) Journal of Applied Polymer Science, 18, 427.
97. Sakota, K. and Okaya, T. (1976) Journal of Applied Polymer Science, 20, 3255.
98. Daniels, T. E., Watson, W. H. and White, F. C. (1959) Rubber and Plastics Age,
40, 1057, 1061.
99. Johnson, P. H. and Kelsey, R. H. (1958) Rubber World, 138, 877.
100. Kelsey, R. H. and Johnson, P. H. (1958) Rubber World, 139, 227.
101. Jones, B. D. (1962) Proceedings of the Fourth Rubber Technology Conference,
London, 1962, p. 485.
102. Howland, L. H. and Nisonoff, A. (1954) Industrial and Engineering Chemistry,
46,2580.
103. Brown, R. W. and Howland, L. H. (1955) Rubber World, 132,471,486.
104. Howland, L. H. and Brown, R. W. (1961) Rubber Chemistry and Technology, 34, 1501.
105. Morton, M. and Gibbs, W. E. (1954) Rubber Age, New York, 74, 911.
106. Rupar, W. and McLeod, L. A. (1956) Rubher World, 134, 560.
107. Levine, E., Lindlaw, W. and Vona,J. A. (1969) Journal ofPaint Technology, 41, 531.
108. McCracken, E. A. and Betts, J. L./Standard Oil Development Company, United
States Patent No. 2,676,951, 27 April, 1954.
109. Gauslaa, K./Copolymer Rubber & Chemical Corporation, United States Patent
No. 3,318,831,9 May, 1967.
110. The General Tire and Rubber Company, British Patent No. 1,162,466, 27
August, 1967.
Ill. Celanese Corporation, British Patent No.1, 203, 524, 26 August, 1970.
112. Brown, R. W./United States Rubber Company, United States Patent No.
2,897,168,28 July, 1959.
113. Rumbold, J. S./United States Rubber Company, United States Patent No.
2,484,425, 11 October, 1949.
114. Waterman, J. A. and La Heij, G. E. (1968) Journal of the Institution of the Ruhher
Industry, 2, 94.
115. Shell Internationale Research Maatschappij NV, British Patent No. 988,301, 7
April, 1965.
116. BIOS Report No. 349 Item 22.
117. Maron, S. H. and Moore, C. (1947) India Rubber World, 116,789.
118. Maron, S. H., Moore, c., Kingston, J. G., Ulevitch, I. N., Trinistic, J. C. and
Borneman, E. H. (1949) Industrial and Engineering Chemistry, 41, 156.
119. Schmidt, E. and Kelsey, R. H. (1951) I ndustrial and Engineering Chemistry, 43, 406.
120. Peaker, C. R./United States Rubber Company, United States Patent No.
2,393,261,22 January, 1946.
121. Willson, E. A./The B. F. Goodrich Company, United States Patent No.
2,444,689,6 July, 1948.
122. Arundale, E./Standard Oil Development Company, United States Patent No.
2,444,801,6 July, 1948.
References 301
123. Rumbold, J. S./United States Patent No. 2,475,053, 5 July, 1949.

124. Schindler, D. L./E. I. du Pont de Nemours and Company, United States Patent
No. 3,526,614, 1 September, 1970.
125. Peaker, C. R./United States Rubber Company, United States Patent No.
2,446,101, 27 July, 1948.
126. Svendsen, E. C./United States Rubber Company, United States Patent No.
2,446,115,27 July, 1948.
127. Rumbold, J. S./United States Rubber Company, United States Patent No.
2,494,115,10 January, 1950.
128. Howland, L. H. and Nisonoff, A./United States Patent No.2, 839, 483, 17 June, 1958.
129. Willson, E. A.jThe B. F. Goodrich Company, United States Patent No.
2,357,861, 12 September, 1944.
130. Monsanto Company, British Patent No.1, 190,849,6 May, 1970.
131. Howland, L. H., Aleksa, E. J., Brown, R. W. and Borg, E. L. (1961) Rubber and
Plastics Age, 42, 868. .
132. Westerhoff, C. B., White, W. W. and Gilbert, R. D. (1963) Rubber Age, New
York, 94, 446.
133. White, W. W., Reynolds, J. A. and Gilbert, R. D. (1964) Journal of Applied
Polymer Science, 8, 2049.
134. United States Rubber Company, British Patent No. 856,337,14 December, 1960.
135. United States Rubber Company, British Patent No. 859,361,18 January, 1961.
136. United States Rubber Company, British Patent No. 859,699,25 January, 1961.
137. United States Rubber Company, British Patent No. 889,731,21 February, 1962.
138. United States Rubber Company, British Patent No. 889,732, 21 February,
1962.
139. Howland, L. H., Chambers, V. S. and Aleksa, E. J./United States Rubber
Company, United States Patent No. 3,049,500,14 August, 1962.
140. Howland, L. H. and Chambers, V. S. United States Rubber Company, United
States Patent No. 3,049,501,14 August, 1962.
141. Howland, L. H. and Chambers, V. S./United States Rubber Company, United
States Patent No. 3,049,502,14 August, 1962.
142. Polymer Corporation Limited, British Patent No. 991,394, 5 May, 1965.
143. Talalay, L. (1962) Proceedings of the Fourth Rubber Technology Conference,
London, 1962, p. 443.
144. Mitchell, 1. M. (1962) Rubber News, 1(11),24.
145. Kazakevich, T. S., Kuznetsov, V. L., Lebedev, A. V., Selivanovskii, S. A., Fermor,
N. A. and Yakovlev, Yu. M. (1967) Soviet Rubber Technology, 26(1), 6.
146. Talalay, L.jThe B. F. Goodrich Company, United States Patent No. 3,031,427,
24 April, 1962.
147. Reynolds, J. A./United States Rubber Company, United States Patent No.
3,081,275, 12 March, 1963.
148. Copolymer Rubber & Chemical Corporation, British Patent No. 881,362, 1
November, 1961.
149. Brewer, J. F.jCopolymer Rubber & Chemical Corporation, United States Patent
No. 3,032,524, 1 May, 1962.
150. McKenzie, 1. P. and Huddleston, G. R.jCopolymer Rubber & Chemical
Corporation, United States Patent No. 3,214,401, 26 October, 1965.
151. Sutherland, J. D./Copolymer Rubber & Chemical Corporation, United States
Patent No. 3,296,1278,3 January, 1967.
152. The B. F. Goodrich Company, British Patent No. 939,482, 16 October, 1963.
153. Talalay, L.jThe 8. F. Goodrich Company, United States Patent No.3, 284,396,
8 November, 1966.
154. Bennett, D. A.{fhe International Synthetic Rubber Company Ltd, British Patent
No. 976,212, 25 November, 1964.
155. Bennett, D. A./The International Synthetic Rubber Company Ltd, British Patent
No. 976,213, 25 November, 1964.
156. Bennett, D. A. and Burridge, K. G./The International Synthetic Rubber
Company Ltd, British Patent No. 976,214, 25 November, 1964.
157. Bennett, D. A./The International Synthetic Rubber Company Ltd, British Patent
No. 1,039,727,17 August, 1966.
158. Halper, W. M., York, R. F. and Moss, F. D. Shell Internationale Research
Maatschappij NV, British Patent No.1, 128, 114,25 September, 1968.
159. Roberts, D. Sto-Chem Ltd, British Patent No. 1,013,736,22 December, 1965.
160. Stell, R. c., Blackwell, 1. C. and Henry, P. J. (1957) Industrial and Engineering
161. Borg, E. L., Provost, R. L. and Bawn, C. V. (1955) Chemical Engineering Progress,
51(6),278.
162. Gudheim, A. R. (1960) Rubber Age, New York, 87, 466.
163. Walbridge, D. J. (1983) Preparation of Non-aqueous Polymer Latexes, in Science
and Technology of Polymer Colloids: Volume I: Preparation and Reaction
Engineering (eds G. W. Poehlein, R. H. Ottewill and J. W. Goodwin), NATO
ASI Series E: Applied Sciences No. 67, Martinus Nijhoff, The Hague, p. 40.
11
Synthetic latices: individual

types
11.1 INTRODUCTION
The previous chapter has dealt with the general principles which underlie
the production of synthetic latices by emulsion polymerization. This chapter
reviews the individual types of synthetic latex. The concern is exclusively
with aqueous synthetic latices. Non-aqueous synthetic latices are also used
industrially; they are produced by emulsion/dispersion polymerization in
non-aqueous media. However, aqueous synthetic latices far exceed non-
aqueous synthetic latices as regards multiplicity and industrial importance.
The main concern in this chapter is, of course, with synthetic latices which
have attained industrial importance to a greater or lesser extent. However,
brief reference is also made to certain other interesting types which so far
have either been of very limited specialist applicational interest only, or
remained of essentially academic interest. Almost all the synthetic latices of
industrial importance are produced by free-radical addition polymerizations
of olefinically-unsaturated monomers in aqueous dispersion media.
Even disregarding possible functionalization by copolymerization, the
majority of synthetic latices of industrial importance contain copolymers
rather than homopolymers. This is because polymer properties intermediate
between those of the respective homopolymers are usually required. This
raises the important matter of why it is preferred to produce copolymers,
rather than to blend appropriate homopolymer latices in the ratios required
to achieve the desired overall polymer composition. An important reason is
the inherent incompatibility of different types of homopolymers at the
molecular level. It is now recognized that complete mixing between the
segments of homo polymers of different types is rarely achieved. Homogeneity
at the molecular level is seldom achieved even if attempts are made to blend
the homopolymers by a technique such as dissolving the polymers in a mutual
solvent and then casting films by evaporation. Although the initial dilute

304 Synthetic latices: individual types
polymer solution may appear to be homogeneous, the different types of

homopolymer begin to phase-separate as the concentration of polymer in
the solution increases. The likelihood of obtaining homogeneity at the
molecular level is even less if the homopolymers are blended as latices, because
the homo polymers were phase-separated when the initial blend was formed,
and at no subsequent stage did they become mixed at the molecular level.
By whatever means blends of homo polymers may be produced, they are
usually heterogeneous at the molecular level, consisting of domains of
essentially the pure homopolymers. The physical properties of such
phase-separated blends often differ significantly from those of copolymers of
the same overall composition. Although good use is now being made of the
distinctive properties offered by polymer systems which comprise domains
of more than one type of polymer, copolymers in latex form usually offer a
more useful balance of properties than do homopolymer blends of the same
overall composition. In effect, emulsion copolymerization provides a way of
forcing dissimilar monomer repeat units to remain in close proximity in the
eventual polymer phase. Any thermodynamic tendency to phase separation
is thwarted by the absence of any physical mechanism by which the repeat
units of different types can become separated.
There is a second important reason for preferring copolymers to homopolymer
blends if it is desired subsequently to crosslink the entire polymer into a
single macromolecular network, and the crosslinking sites are derived from
only one of the monomers. Butadiene homopolymers and copolymers provide
a very important example of this situation.
The types of synthetic latices which have been, and are currently, produced
are many and varied. Sections 11.2 to 11.7 of this chapter deal with the
various families of non-functionalized synthetic latices which have found
widespread industrial application. The types considered are respectively the
styrene-butadiene copolymer latices, the acrylonitrile butadiene copolymer
latices, the polychloroprene latices, the latices of vinyl acetate polymers and
copolymers, the family of synthetic latices which are commonly known as
acrylic, and the vinyl chloride-vinylidene chloride copolymer latices. It
should be noted that it is not always evident into which of these broad classes
a particular synthetic latex should be placed, because of the multiplicity of
monomers which have been used in its manufacture. Functionalized synthetic
latices are dealt with as a single group in section 11.8. Although the
applications for which functionalized synthetic latices are used depend greatly
upon the main monomers which were used in the manufacture of the latex,
the principles underlying functionalization, and the advantages conferred,
are in general common to functionalized synthetic latices as a group. Thus,
for example, the additional chemical reactions available to functionalized
polymer molecules are effectively independent of the chemical nature of the
majority ofrepeat units. Under the heading offunctionalized synthetic latices,
it is appropriate to discuss the speciality latices known as alkali-thickenable
Latices of styrene-butadiene copolymers 305
synthetic latices, because these latices are produced by extensive functionalization

of latices with weak-acid groups. The final section of this chapter, section
11.9, deals with miscellaneous synthetic latices which, although interesting
as further examples of polymers in latex form, have so far achieved little
industrial significance as regards amounts produced and used. Latices which
contain particles comprising interpenetrating polymer networks, and which
have been formed from synthetic latices, are briefly considered in section
14.10 of Chapter 14 as examples of chemically-modified latices, together with
latices of this type derived from natural rubber latex.
11.2 LATICES OF STYRENE-BUTADIENE COPOLYMERS
11.2.1 General considerations

Styrene- butadiene copolymer latices comprise a large family of general-
purpose synthetic latices which are industrially very important. As a family,
they have found widespread application is several broad areas, of which the
following have been the most important:
I. as replacements for natural rubber latex, notably in the manufacture of

latex foam rubber;
2. in carpet-backing and other textile applications, for which purposes the
dispersed polymer is usually functionalized;
3. in paper-coating applications, for which purpose the dispersed polymer
is again usually functionalized;
4. as general-purpose adhesives for paper and other fibrous materials;
5. as rubber-to-textile bonding adhesives for tyre-cords and other fibres
to be used as reinforcement for mechanical rubber products.
Other important industrial uses for styrene-butadiene copolymer latices have

included application as binders for latex-based surface-coating compositions,
and as stiffening and reinforcing agents for films derived from latices of
rubbery polymers. Styrene-butadiene copolymer latices for use as tyre-cord
adhesives are described in section 11.2.5 below; functionalized synthetic latices
are considered as a class in section 11.8 below.
The physical character of styrene-butadiene copolymers changes progressively
from polybutadiene, which is flexible and rubbery, to polystyrene, which is
rigid and glassy, as the styrene content increases; this matter is considered
further in section 11.2.2 below. It is latices of copolymers of relatively low
styrene content which are used as alternatives to natural rubber latex. Such
latices have for many years been known as styrene-butadiene rubber (SBR)
latices. In the period immediately following World War II, they were known
as GR-S (Government Rubber-Styrene) latices. The copolymers in latices
intended for use as binders for latex-based paints have rather higher styrene
contents. They are stiffer than the copolymers used as rubbers, but are still
below their glass-transition temperatures at normal ambient temperatures.
The particles in latices which contain copolymers having very high styrene
contents, and so are rigid at ordinary temperatures, are intended for use as
stiffening and reinforcing agents for films derived from latices of rubbery
polymers.
As has been noted in section 1.6.2 of Chapter 1 (Volume 1), styrene- butadiene
rubber latices were developed during World War II, not primarily for use
as latices, but as intermediaries to the production of solid general-purpose
synthetic rubbers for use as alternatives to natural rubber. After World War
II ended, much effort was devoted, mainly in the United States of America,
to developing styrene-butadiene rubber latices suitable as replacements for
natural rubber latex, especially in the manufacture of latex foam rubber. The
result was the appearance in the early 1950s of styrene-butadiene rubber
latex concentrates which could be used partially to replace natural rubber
in this application. As improved grades of synthetic latex concentrate became
available, it became practicable to produce latex foam rubber from exclusively
styrene-butadiene rubber latex. The synthetic rubber latex then ceased to
be regarded as merely a troublesome diluent and price-stabilizer for natural
rubber latex; instead, it came to be accepted as a raw material in its own
right, preferred by some users to the product of natural origin.
After World War II ended, it was also realized that both the experience
which had been gained as a result of the war-time synthetic rubber
development programme, and the industrial capacity for carrying out
emulsion polymerization reactions which became available as a consequence
ofthe implementation ofthat programme, might be utilized for the production
of styrene-butadiene copolymer latices for other applications, notably as
binders for latex-based paints. Styrene butadiene copolymer latices suitable
for this latter application began to become available in North America during
the late t 940s. For several years, these latices were the preferred bases for
such paints in North America, probably for economic reasons. They have
subsequently been supplanted in this application by other types of synthetic
latex, notably latices of vinyl acetate polymers and copolymers, and acrylic
latices. In Europe, however, the production of styrene-butadiene rubber
latices for the manufacture of latex foam rubber began later than in North
America. Furthermore, little styrene-butadiene copolymer latex has been
used in Europe as bases for latex paints; latices of vinyl acetate polymers
and copolymers, and latterly acrylic latices, have been used for this application.
The chemical structures of styrene and butadiene (strictly t,3-butadiene)
are shown as I and II respectively. During free-radical polymerization in
general, and emulsion free-radical polymerization in particular, butadiene
molecules can become incorporated in the polymer chain by either 1,4-addition
or 1,2-addition, the incorporated units having the structures shown as III
and IV respectively. It has always been regarded as desirable that most, if
not all, the butadiene units should enter the polymer chain by t ,4-addition.
The consequence of such addition is the formation of a linear sequence of
units, each of which contains three carbon -carbon single bonds and one
olefinic double bond. It is the former which confer segmental flexibility upon
the polymer chain; the latter are rigid at normal temperatures. There are
further possible structural variations displayed by polymerized butadiene
units. The configuration of the substituents surrounding the olefinic double
bonds of the 1,4 units can be either cis (Structure Va) or trans (Structure Vb)
as regards the relative dispositions of the two parts of the polymer chain
CH 2 =CH-CH=CH 2
II
...-CH 2 -CH-...
I
... -CH 2 -CH=CH-CH 2- · · · CH
II
III CH 2
IV
... -CH 2 CH 2 - · · ·
'CH=CH/
H/ 'H
Va
which emanate from any particular double bond. The cis is the preferred of
the two configurations as regards facilitation of rotational flexibility around
the neighbouring carbon-carbon single bonds. A further complication is that
units formed by 1,2-addition and also the units formed by polymerization
of styrene contain a chiral (asymmetric) carbon atom. Thus any particular
1,2 unit can have either the (R) (formerly 0) or the (S) (formerly L)
configuration about its chiral carbon atom. Insofar as sequences of 1,2 units
are present, it is possible that all the units in a particular sequence have the
same configuration (either all (R) or all (S)), or that the (R) and (S)
configurations alternate, or that the distribution of configurations is random.
These three types of sequence are known as isotactic, syndiotactic and atactic
respectively. Each enchained styrene unit also contains a chiral carbon atom,
and so can have either the (R) or the (S) configuration. Again, insofar as
sequences of polystyrene units form, they can be isotactic, syndiotactic or
atactic. There are still further possible structural complications. Successive
pairs of styrene or 1,2-butadiene units may add head-to-head, head-to-tail
or tail-to-tail according to the relative dispositions of the two chiral carbon
atoms. Branching and crosslinking can occur through the residual olefinic
unsaturation of the butadiene units, however they have been enchained, and
also by reactions which involve other reactive groups associated with the
presence of such un saturation, in particular, IX-hydrogen (allylic) atoms, as
noted in section 10.2.2.4 of Chapter 10. Thus any particular butadiene
molecule and any particular styrene molecule can become incorporated in
the polymer chain in one of several different modes; and any particular
butadiene molecule may become a point of branching or crosslinking in the
final polymer. The properties of styrene- butadiene copolymers of given
overall composition depend upon the relative abundances of the various
types of polymerized units, especially of the various types of polymerized
butadiene units. The same is true for polybutadienes. For emulsion polymers,
these relative abundances depend to some extent upon the polymerization
conditions (see section 11.2.3 below). It is sometimes desirable deliberately
to introduce branches and crosslinks into the polymers contained in
styrene-butadiene copolymer latices, additional to those which become
present through interactions between propagating radicals and butadiene
units. This end is most readily achieved by including a small amount of a
tetrafunctional comonomer into the polymerization system. The best-known
example of such a monomer for use with styrene and butadiene is divinylbenzene,
the p-isomer of which is shown as Structure X in Chapter 10.
One consequence of the possibilities discussed in the preceding paragraph
is that a sequence of butadiene units in a polymer is unlikely to be
stereoregular, in that successive polymerized units may have different
molecular structures and configurations. In this respect, butadiene polymers
obtained by free-radical polymerization differ significantly from natural
rubber. Natural rubber hydrocarbon is an essentially stereoregular polymer,
being a linear cis-\ ,4-polyisoprene (see section 9.5.2.2.2 of Chapter 9). Being
stereoregular has the important practical consequence that the polymer
segments are able to crystallize on stretching. The material thus becomes
increasingly stiff as it is stretched, and this has the desirable consequence
that the tensile strength and tear strength of vulcanizates are high, even in
the absence of reinforcing aids. In contrast, because butadiene polymers
obtained by free-radical polymerization are non-stereoregular, they do not
crystallize on stretching. In consequence, vulcanizates are usually weak unless
effective reinforcing agents are present. The ability to strain-crystallize is
further reduced if the polybutadiene sequences in the polymer chain are
interrupted by copolymerized styrene units. Thus the tensile strength of a
typical filler-free ('gum') vulcanizate from dry natural rubber is ca. 30 MPa,
whereas that of a similar vulcanizate from an emulsion-polymerized
butadiene-based dry rubber is typically ca. 3 MPa. For these reasons, there
has been great interest in the past in producing stereo regular I,4-polybutadienes
(and also stereo regular \ ,4-polyisoprenes) by aqueous emulsion polymerization.
However, it has so far been impossible to produce useful rubbery materials
in this way, although such polymers can be, and are, produced by non-free-radical
solution polymerization. The production of stereo regular 1,4-polybutadienes

and 1,4-polyisoprenes by aqueous emulsion polymerization remains an
important challenge to the practitioners of emulsion polymerization.
11.2.2 Effect of styrene-butadiene ratio in copolymer

It is apparent from section 11.2.1 above that the ratio of bound styrene to
bound butadiene in the dispersed polymer is a very important variable for
non-functionalized styrene-butadiene copolymer latices. In principle, this
ratio can extend from O/I00m/m to 100/0m/m styrene/butadiene, the
polymers then being polybutadiene and polystyrene respectively. As the
styrene/butadiene ratio increases from the first of these extremes to the second,
so the polymer changes progressively from a rubbery material of low
glass-transition temperature (ca. - 85·C for the type of polybutadiene
produced by emulsion polymerization) to a glassy material of high
glass-transition temperature (ca. 100·C for atactic polystyrene). The
glass-transition temperature increases progressively as the styrene content of
the copolymer increases, the rate of increase increasing as the styrene content
increases. This is illustrated in Fig. 11.1, which shows results for the
relationship between the glass-transition temperature and composition for
styrene-butadiene copolymers, collected by Gordon and Taylor [1].
120
-1200~--2~O:---...J4~O---:6~O--~80-:---~100
Styrene content of copolymer I % m/m
Figure 11.1 Effect of styrene content upon glass-transition temperature of styrene-butadiene

copolymers (Gordon and Taylor [1]).
The styrenefbutadiene ratio of the dispersed polymer has at least five

important consequences for a styrene-butadiene copolymer latex and the
polymer itself, two of which are directly related to the effect upon the
glass-transition temperature: it determines whether or not the latex is
film-forming at normal ambient temperatures; it determines the physical
character of the polymer at normal ambient temperatures, and, indeed, at
other temperatures as well; it affects the chemical reactivity of the copolymer;
it affects the degree of branching and crosslinking which has developed in
the copolymer as it forms during emulsion polymerization; and it affects the
response ofthe latex to various destabilizative influences which tend to reduce
the colloid stability of the latex.
Provided that the styrene/butadiene ratio is less than ca. 70/30 m/m, the
latex readily forms coherent films when dried at normal ambient temperatures.
As the styrene content increases above this level, so the deposits become
increasingly powdery, unless the latex has first been compounded with a
polymer plasticizer. Unless the polymer is crosslinked, the films also become
coherent if heated to an appropriate temperature. For deposits formed at
normal ambient temperatures from un plasticized styrene- butadiene copolymer
latices containing copolymers of various compositions, a guide to the
appearance is as follows: a latex containing polystyrene dries to powdery
deposits; a latex of a 90/10m/m styrene/butadiene copolymer gives a film
which flakes readily; if the ratio is 80/20 m/m, the film is more coherent but
cracks readily; if the ratio is 70/30 m/m, the film is coherent, but hard and
tough; as the ratio falls to 60/40 m/m, so the film becomes rubbery; and with
ratios of 50/50 m/m and less, the film is coherent, soft and rubbery.
The effect of copolymer composition upon the physical character of the
polymers themselves will be evident from the comments in the preceding
paragraph concerning the appearance of films cast from the latices. As the
bound-styrene content increases, so the polymer becomes progressively
harder, less extensible and less rubbery. Some flexibility is retained at normal
ambient temperatures, provided that the styrene/butadiene ratio is less than
ca. 70/30 m/m. However, there is a tendency to regard all latex styrene- butadiene
copolymers which contain more than ca. 60% m/m of styrene as being
non-rubbery. The effect of the styrene content upon the tensile strength and
elongation at break of unvulcanized films obtained from styrene-butadiene
copolymer latices is illustrated in Fig. 11.2, which is based upon results
reported by D'Ianni, Hess and Mast [2]. It is evident that, in the region of
styrene content where the copolymer is ceasing to be rubbery at normal
ambient temperatures as the styrene content increases, the tensile strength
ofthe film increases sharply with increasing styrene content, and the extension
at break decreases sharply. The elastic modulus also increases sharply. In
addition to reduction in extensibility and increase in modulus and hardness,
increasing styrene content is accompanied by decrease in flexibility at low
temperatures, and by an increased tendency to creep and cold flow.
30 750
I
I
I
------ .....
500 ~
-'"
<Il
~
.c
<ii
<:
o
'iii
<:
250 ~
0~----~----~----~7-----~----
o 20 40 100
.. 0
Styrene content of copolymer / % mlm
Figure 11.2 Effect of styrene content of polymer upon tensile strength (points.) and extension
at break (points A) of films obtained from styrene-butadiene copolymer latices (based upon
results reported by D'Ianni, Hess and Mast [2]).
The styrene/butadiene ratio has consequences for the chemical properties

of styrene/butadiene copolymers. As the ratio increases, so the concentrations
of olefinic unsaturation and allylic (rx-methylenic) hydrogen atoms in the
polymer decreases. Reactions, such as sulphur-vulcanization, the rate of which
depend upon the concentration of olefinic double bonds, are retarded by
reducing this concentration. For the same reason, the resistance of the
copolymer to deterioration by ozone and oxygen is improved as the styrene
content increases. However, the general weathering resistance remains
relatively poor over the whole range of copolymer compositions, because
both types of repeat unit in the copolymer are susceptible to absorption of,
and activation by, ultraviolet light.
The styrene/butadiene monomer ratio in the emulsion polymerization
reaction system used to prepare the latex affects the degree of branching and
crosslinking which develops in the copolymer as it forms, because branching
and crosslinking are primarily consequences of the presence of residual
olefinic unsaturation and allylic hydrogen atoms in the polymerized butadiene
units (see section 10.2.2.4 of Chapter 10). The lower the styrene/butadiene
monomer ratio, the greater is the tendency for branching and crosslinking
to occur, other variables, notably extent of conversion of monomers to
polymers and modifier level, remain unchanged.
The colloid stability of styrene- butadiene copolymer latices prepared
under comparable conditions has been found to vary with the bound-styrene
content of the copolymer. The mechanical stability of the latex is found to
fall progressively as the styrene/butadiene ratio decreases from 100/0 to
50/50 m/m, whereas the colloid stability towards inorganic electrolytes, such
as sodium chloride, increases significantly. These differences are probably
consequences of the effect of bound-styrene content upon the ease of particle
coalescence. Hard, glassy particles are less likely to coalesce when they collide
as a result of mechanical agitation. However, such particles are unable to
agglomerate to form larger particles of reduced specific surface area in order
to adjust their colloid stability to the new environment which is established
by the addition of an inorganic electrolyte.
11.2.3 Production of styrene-butadiene copolymer latices

There are two important general matters relating to the production of
styrene-butadiene copolymer latices by emulsion polymerization which
require comment before proceeding to consider specific production procedures.
They are the effect of extent of monomer-to-polymer conversion upon the
polymer produced, and the effect of polymerization temperature upon the
polymer produced.
Economic considerations require that the conversion of monomers to
polymer should be as complete as possible, consistent with a satisfactory
product being obtained. The separation, recovery and recycling of unreacted
monomers is expensive, troublesome and time-consuming. With functionalized
styrene-butadiene rubber latices intended for application as such, emulsion
copolymerization is often taken almost to complete conversion, there being
left at the end of the reaction only minor amounts of unreacted monomers
to be removed by stripping. However, there are two disadvantages in allowing
styrene-butadiene emulsion copolymerizations to proceed to complete
conversion:
l. There is a problem of copolymer-composition drift as the polymerization

proceeds (see section 10.2.4.5.1 of Chapter 10). The copolymer which forms
initially is richer in butadiene thim was the monomer feed mixture from
which it was formed. The feed mixture therefore tends to become
progressively richer in styrene as the polymerization proceeds. The
styrene-rich copolymer molecules which form during the later stages of
the polymerization tend not to be completely miscible with the butadiene-rich
molecules which form during the early stages. The resultant heterogeneity
at the molecular level has a detrimental effect upon the mechanical
properties of vulcanizates obtained from the rubber.
2. Highly branched and crosslinked polymer molecules form during the later
stages of the polymerization, as discussed in section 10.2.2.4 of Chapter
10. The occurrence of extensive branching and crosslinking in the primary
rubber structure is again detrimental to the mechanical properties of

vulcaniza tes.
Although both these effects are detrimental, the second is the more serious
of the two. Furthermore, it is in principle possible to mitigate the first of
these effects by adopting an appropriate monomer-addition regimen which
minimizes copolymer-composition drift. Some mitigation of the second
effect is possible by including one or more modifiers in the emulsion
polymerization reaction system. However, complete mitigation is not
practicable; for this reason, it is customary to shortstop many emulsion
copolymerizations for the production of styrene-butadiene rubber latices
at conversions in the range 65-70%.
The temperature at which the emulsion copolymerization reaction is carried
out also has important effects upon the structure and properties of the polymer
produced. Reducing the polymerization temperature decreases the tendency
for branching and crosslinking to occur, thereby resulting in the formation
of a more linear polymer oflower gel content. The polymerization temperature
also affects the distribution of isomeric structures of the polymerized
butadiene units. Results reported by Hampton [3] for emulsion homo polymers
of butadiene are shown here in Fig. 11.3 (a). The presence of styrene in the
reaction system has little effect upon the distribution of isomeric structures
amongst the butadiene units. As the polymerization temperature is lowered,
so the proportion of trans-l ,4-butadiene units increases, mainly at the expense
of the cis-l,4 units. The proportion of 1,2 units is little affected. The effect
of reducing the polymerization temperature is thus to produce a polymer of
increased stereo regularity as regards the butadiene units. The mechanical
properties of dry-rubber vulcanizates obtained from styrene-butadiene
rubbers produced by polymerization at reduced temperatures (typically 5°C)
are superior to those produced at higher temperatures (typically 50°C). This
superiority is attributed partly to the increased stereoregularity of the
butadiene component, and partly to reduced branching and crosslinking in
the unvulcanized rubber. Other important contributions to improved mechanical
properties of dry-rubber vulcanizates are believed to accrue from reduction
in the proportion of polymer of relatively low molecular mass produced, and
from the formation of processable rubber of higher primary molecular mass
as the polymerization temperature is reduced. Broadly similar comparisons
apply to styrene-butadiene rubber vulcanizates derived directly from latices.
Figure 11.3(b) reproduces results reported by Smith et al. [4] for the effect
of polymerization temperature upon the tensile strength of sulphur-vulcanized
films from styrene-butadiene rubber latices. These results indicate that, by
emulsion polymerizing the monomers at sub-zero temperatures, it is possible
to obtain 'gum' vulcanized latex films having tensile strengths approaching
those of vulcanized films from natural rubber latex. This confirms the
importance of polymer structural factors such as stereoregularity, linearity
and high molecular mass in determining the tensile strength of vulcanized
80
Ul
!
~
~ 60
.!:!
E
Q)
r:::
~ 40
t
o
a.
~ 20
o
.~ ::t:J:7
••
o 40 80 120
Polymerization temperature / DC
(8)
30
~o -20 -10 0 10 20 30 40 50 60
Polymerization temperature / DC
(b)
Figure 11.3 (a) Effect of polymerization temperature upon proportions of isomeric structures
of polymerized butadiene units (Hampton [3]). Significance of points: • cis-I,4; • trans-I ,4; _ 1,2.
(b) Effect of polymerization temperature upon tensile strength of sulphur-vulcanized films from
styrene-butadiene rubber latices (Smith et al. [4]). Styrene;butadiene ratio was 29(11 m/m.
Formulation (parts by mass): rubber 100, zinc oxide 3, sulphur 2, zinc di-n-dibutyldithiocarbarnate
0.5, zinc mercaptobenzthiazolate 1.5.
films derived from latices of butadiene-based polymers; presumably these factors

affect the extent to which the rubber crystallizes on stretching. Vulcanizates
derived from styrene-butadiene rubber latices produced at conventional
polymerization temperatures are inevitably mechanically weak, partly because
the rubbers produced do not crystallize on stretching, and partly because it is
difficult to achieve significant particulate reinforcement of rubbers derived

directly from latices (see further section 16.6 of Chapter 16 (Volume 3».
Other mechanical properties which are improved by reducing the
polymerization temperature include the wet-gel strength of latex deposits,
and the low-temperature flexibility of vulcanizates. A further advantage which
accrues is reduction of odour when residual monomers have been removed.
This improvement, which can be important if the latex is to be used as such,
is attributed to reduced tendency for butadiene to dimerize at low temperatures.
Although the dimer distils in the same temperature range as does styrene, it
has a powerful odour, and so the presence of even small amounts is evident.
Table ILl gives the so-called standard GR-S formulation which was
mutually agreed during World War II for the production of styrene-butadiene
rubber latex from which was separated solid rubber to be used as a
replacement for natural rubber. Polymerization is effected at 50"e. The
reaction rate is such that ca. 75% of the monomers is polymerized after
12 hours reaction, although, for the reasons discussed above, it may be
desirable to shortstop the reaction before this conversion is attained. The
surface free energy of the aqueous phase increases sharply when the conversion is
ca. 13% . By ca. 50% conversion, all the unreacted monomer has disappeared as
a separate phase, having been absorbed into the growing polymer particles. At
the conclusion of the reaction, the excess butadiene is removed by reducing
the pressure above the system, thereby allowing the butadiene to flash off.
The unreacted styrene is then removed by steam stripping under reduced pressure.
Two formulations for the production of styrene-butadiene rubber latices
by emulsion polymerization at SOC are given in Table 11.2. Both use redox
initiation systems the oxidizing component of which is a hydroperoxide.
Formulation A is the so-called custom recipe. The reducing component in
this reaction system is an iron (II) salt. The iron (II) cations are complexed
with pyrophosphate anions before being added to the reaction system, so as
to ensure that they are present in the aqueous phase in an appropriate low
concentration. A reducing sugar, dextrose, is present to reduce iron (III) ions,
produced during the initiation reaction by oxidation of the iron (II) ions,
back to iron (II) ions. Direct interaction between the reducing sugar and the
Table 11.1 Standard GR-S formulation for production

of a styrene-butadiene rubber latex
Ingredient Parts by mass
Styrene 25
Butadiene 75
Water 180
Carboxylate soap 5
n-dodecyl mercaptan 0.5
Potassium peroxodisulphate 0.3
Table 11.2 Formulations for production of styrene-butadiene rubber latices by

low-temperature emulsion polymerization using hydro peroxide redox initiation systems
Parts by mass
Ingredient A B
Styrene 28 30
Butadiene 72 70
Water 180 180
Disproportionated rosin-acid soap 4.7
Potassium fatty-acid soap 4.5
Sodium alkylnaphthalenesulphonate 0.1
Mixed tert-mercaptans 0.24 0.20
Cumyl hydroperoxide 0.10 0.20
Iron(II) sulphate heptahydrate 0.14
Potassium pyrophosphate 0.177
Dextrose 1.0
Tetraethylenepentamine 0.20
Potassium chloride 0.50 0.50
Potassium hydroxide 0.10 0.10
hydroperoxide is not thought to contribute significantly to initiation. The

interactions between the various components of this initiation system are
believed to be essentially as illustrated in Fig. 10.3 of Chapter 10. An inorganic
electrolyte, potassium chloride, is present in the reaction system to ensure
that it remains fluid throughout the polymerization. A small amount of
potassium hydroxide is present in order that the reaction system, and the
latex which is produced, remain sufficiently alkaline to ensure adequate
colloid stability. Formulation Buses tetraethylenepentamine as the reducing
component of the redox initiation system. The rates of polymerization of
these two reaction systems at 5°C are such that System A gives ca. 60%
conversion after 13 hours, whereas System B polymerizes more rapidly, giving
ca. 60% conversion after 8 hours.
Various so-called volatile-base formulations for the production of styrene-
butadiene copolymer latices have been developed. The purpose of these
formulations is to produce latices which dry down to films having minimum
sensitivity to water. The underlying principle is that all the colloid stabilizers
should be anionic, and their cations be derived from a volatile base such as
ammonia or morpholine. Likewise, any simple electrolytes in the reaction
system should be derived from volatile acids and bases, and any additional
alkali present should be volatile. The intention is that, when a film dries
down from the latex, the base component of both colloid stabilizers and
electrolytes will gradually evaporate from the film, and so too will the acidic
component of the electrolyte. A typical formulation for a volatile-base latex
comprises ammonium laurate as colloid stabilizer, ammonium acetate as
electrolyte, ammonia as alkali, and a combination of a hydroperoxide and

diethylenetriamine as initiator.
Styrene-butadiene copolymer latices can be produced by either batch or
continuous emulsion polymerization. Batch emulsion polymerization tends
to be used when relatively small amounts of a range of latices is being
produced. It is particularly suited to the production of functionalized
styrene-butadiene copolymer latices. Continuous emulsion polymerization is
appropriate for the production oflarge quantities of standard non-functionalized
grades. An outline flow diagram of a typical process for the production of
styrene-butadiene rubber latex by continuous emulsion polymerization is
shown in Fig. 11.4. Butadiene, styrene, water, soap solution, initiators and
modifiers are blended and passed under pressure through a series of large
reactors, each of which can be by-passed if desired so as to reduce the total
reaction time. Each vessel is equipped with continuous stirring, and with
facility for heat exchange. The reaction mixture then passes to a series of
smaller reaction vessels, each again capable of being by-passed. Shortstopper
is injected into one of these reactors at the point at which the desired degree
of conversion has been attained. Unreacted butadiene is recovered by
releasing the pressure, whereupon it flashes off and is recycled. Unreacted
styrene is removed by steam-distillation, and is also recycled. If required to
be concentrated for use as latex, the product is then subjected to one of the
partial-agglomeration processes described in section 10.4.2 of Chapter 10.
11.2.4 Types of styrene-butadiene copolymer latex

Table 11.3 summarizes the characteristics of the more important of the early
grades of GR-S latex. Type I was that from which the standard solid GR-S
polymer was regularly obtained. It contained sodium stearate as the colloid
stabilizer. A staining antioxidant (usually phenyl-2-naphthylamine) at a level
of ca. 1.5- 2 pphr was added after polymerization. This latex usually contained
flocculated particles. Type II latex was similar to Type I, but contained no
antioxidant. Type III latex was prepared using the potassium soap derived
from crude rosin acid. It was taken to higher conversion than either Type I
or Type II, and contained no antioxidant. It was more viscous than either
Type I or II, had a higher total solids content and a lower surface free energy,
and contained no shortstopper. Type V was a large-particle, high-solids latex
produced directly by emulsion polymerization.
Many grades of high-solids styrene-butadiene rubber latex have become
available industrially in the years since the end of World War II. These latices
have been developed principally for application in the manufature of latex
foam rubber: The colloid stabilizer is usually, if not always, a carboxylate
soap. It may be either a fatty-acid soap, such as an oleate, or a mixture of
a fatty-acid and rosin-acid soap; the former is probably the most common.
Most industrially-successful grades have been agglomerated either by freezing
Styrene recovery and recycle
Styrene storage
I •
Butadiene recovery and recycle
Butadiene
storage
Butadiene
flash-off
Initiators
Soap solution
Modifier
Polymerization vessels
Water
Stripped latex to storage,

agglomeration,
concentation, etc.
Figure 11.4 Outline flow diagram of a typical process for the production of styrene-butadiene rubber latex by continuous emulsion polymerization.
Table 11.3 Early grades of GR-S latex
Total Average Styrene/ Temperature
solids particle butadiene of polymer- Type Mvt of pH range
content diameter ratio ization Conversion of contained of colloid
Type /% m/m /nm / m/m* j"C /% soapt polymer stability
Type I 24-28 60 25/75 NaFA 45-55 8.5-12.0

Type II 24-28 110 25/75 NaFA 45-55 8.5-12.0
Type III 35-40 110 50/50 High KRA 65-85 8.5-12.0
Type IV 39-42 110 50/50 RA 90 8.5-12.0
Type V >59 300 30/70 8.5-12.0
Type VI 60-63 200 45/55 60 95 RA 8.5-12.0
Type VII 60-63 200 20/80 60 FA+RA 8.5-12.0
Type VIII >50 150 50/50 60 8.5-12.0
* Mass ratio in monomer-feed mixture.

t FA = fatty-acid soap; RA = rosin-acid soap.
t Mooney viscosity at lOO'C.
or mechanically by application of pressure in a homogenizer, prior to

concentration. Typical properties of two typical grades of high-solids
styrene-butadiene rubber latex are given in Table 11.4. It is evident that the
total solids contents exceed somewhat those of typical natural rubber latex
concentrates obtained by centrifugation; indeed, styrene-butadiene rubber
latex concentrates have been produced having total solids contents of up to
ca. 70%. However, it must be borne in mind that there may be a greater
difference between total solids content and dry rubber content for the
synthetic rubber latex concentrates than for the natural rubber latex
concentrates with which they compete.
11.2.5 Styrene-vinylpyridine-butadiene terpolymer latices

Styrene-vinylpyridine-butadiene terpolymer latices have been developed
principally for application as rubber-to-textile adhesives. In particular, they
have found widespread application as tyre-cord adhesives. The third monomer
is either 2-vinylpyridine itself (Structure VI), or a derivative such as
2-methyl-5-vinylpyridine (Structure VII). The preparation of terpolymer
CH=CH 2
VI
latices of this type has been described by Tsailingold et al. [5]. The latices
can be produced by peroxodisulphate-initiated emulsion copolymerization
at 50-60·C, using reaction systems similar to those for the production of
styrene-butadiene rubber latices at these temperatures. Polymerization
Table 11.4 Typical properties of two typical grades of high-solids styrene- butadiene
rubber latex
Property Grade A Grade B
Total solids content / % m/m 66.0 65.0

Bound styrene content of polymer / % m/m on polymer 22.0 30.0
pH 10.0 10.5
Viscosity / Pa s (Brookfield LV Viscometer, # 3 spindle)
at rotational frequency 12 minute - I 1.5 2.0
at rotational frequency 30 minute-I 1.2 1.2
Free styrene / % m/m on latex 0.02 0.04
SO-mesh coagulum content / % m/m dry on latex 0.02 0.03
Formaldehyde preservative / g (kg latex)-I 1.0 1.0
occurs more rapidly than does the emulsion copolymer of styrene and
butadiene under similar conditions. The rate of polymerization depends to
some extent upon the vinylpyridine derivative which is used as the third
monomer. This is illustrated by results reported by Frank et al. [6], shown
here in Table 11.5, for the conversion attained during the peroxodisulphate-
initiated emulsion copolymerization of butadiene and various vinylpyridine
derivatives after 8 hours at 60°C. The conversion for a similar emulsion
copolymerization of styrene and butadiene is also included for comparison.
Typical latices of this type contain terpolymers of approximate composition
15/15/70 m/m/m styrene/vinyl pyridine/butadiene. The contained polymer can
be regarded as a 30/70 m/m styrene/butadiene rubber in which half the styrene
units have been replaced by the more polar vinyl pyridine units.
11.2.6 Production of masterbatches based upon styrene- butadiene rubbers for

use in dry-rubber technology
Styrene-butadiene rubber in latex form can be used to produce four types
of composite materials which are of importance in relation to the processing
of the bulk rubber, but not the processing of the rubber in latex form. These
types of composite material are as follows:
Table t 1.5 Results illustrating comparative rates of emulsion copolymerization of

butadiene with various vinylpyridine derivatives, together with comparative result
for emulsion copolymerization of styrene and butadiene (Frank et al. [6])
Reaction system (parts by mass):
Monomers 1()()
Water 180
Soap 5
n-dodecyl mercaptan 0.5
Reaction conditions: 8 hours at 50·C
Composition of monomer mixture Conversion

/ m/m /%
25/75 4-vinylpyridine/butadiene 91.0
12.5/12.5/754-vinylpyridine/2-vinylpyridine/butadiene 78.6
25/75 2-methyl-6-vinylpyridine/butadiene 78.5
12.5/12.5/752-vinylpyridine/3-vinylpyridine/butadiene 78.0
12.5/12.5/75 2-vinylpyridine/5-ethyl-2-vinylpyridine/butadiene 73.0
25/75 2-methyl-5-vinylpyridine/butadiene 69.5
25/75 5-ethyl-2-vinylpyridine/butadiene 66.5
25/75 styrene/butadiene 48.8

1. masterbatches of carbon blacks and styrene-butadiene rubbers;

2. oil-extended styrene- butadiene rubbers;
3. master batches of carbon blacks and oil-extended styrene- butadiene
rubbers; and
4. masterbatches of styrene-butadiene rubbers which contain particles of
styrene-butadiene copolymers of high styrene content.
Masterbatches of carbon black and styrene-butadiene rubber are produced

by the co-coagulation of blends of styrene-butadiene rubber latices and
appropriate aqueous dispersions of carbon black. Oil-extended
styrene-butadiene rubbers are produced by the co-coagulation of blends of
styrene-butadiene rubber latices and aqueous emulsions of suitable plasticizing
oils. Masterbatches of carbon black and oil-extended styrene-butadiene
rubber are produced by co-coagulation of blends of latices with aqueous
carbon black dispersions and aqueous oil emulsions. Masterbatches containing
high-styrene copolymers are produced by co-coagulation of appropriate latex
blends. Interest has also been shown in the production of similar composite
materials based upon natural rubber latex. An important reason for the
failure of such materials to be successfully exploited is the wide geographical
separation of the areas of the world in which the raw materials are produced.
Masterbatches of rubber and carbon black are unvulcanized rubbers
intimately mixed with carbon black in a considerably higher concentration
(typically in the range 35-60 pphr) than is normally present in rubber
compounds. By their use, the mixing of carbon black and rubber in the
required proportion can then be achieved by a process in which effectively
the master batch is diluted with additional rubber. By this means, the handling
of carbon black powder can be avoided. The advantages of using masterbatches
instead of carbon black itself include improved cleanliness, improved dispersion
of the carbon black in the rubber, reduced power requirement for mixing,
and convenience of storage, handling and weighing. Disadvantages include
restriction to those grades of carbon black which are available as masterbatches,
and possible interference with both the vulcanization process and the
attainment of maximum reinforcement by additives which were used to aid
the dispersion of the carbon black particles in water prior to addition to the
latex. Reduction in the ability of the carbon black to reinforce the rubber
vulcanizate is attributed to partial deactivation of the surface of the carbon
black particles by adsorbed dispersing agents.
The possibility, and economic desirability, of extending and cheapening
rubbers by the addition of plasticizing oils has long been recognized. The
polymer molecules of low molecular mass in a rubbery material do not
contribute significantly to the strength of the subsequent vulcanizate; indeed,
oligomers of very low molecular mass have a detrimental effect. These
molecules do, however, improve the processability ofthe unvulcanized rubber.
By oil-extending a rubber of high molecular mass, it is possible to obtain a
Latices of acrylonitrile-butadiene copolymers 323
product in which the relatively expensive low-molecular-mass fraction of a

non-oil-extended rubber is replaced by a much cheaper plasticizing oil. Thus
the concept of oil-extension is that the added oil should provide an effective
equivalent to the lower end of the molecular-size distribution of the rubber
more cheaply than would be the case if the plasticizing components of the
rubber were oligomers of low molecular mass formed during the production
of the rubber by polymerization of the monomers. Oil-extension of synthetic
rubbers first became feasible industrially in the early 1950s, following
successful production of processable styrene-butadiene rubber of high
molecular mass by low-temperature emulsion polymerization. Most, if not
all, current industrial grades of oil-extended styrene-butadiene rubber are
based upon rubbers produced by low-temperature emulsion polymerization.
The oil content is typically in the range 20-60 pphr. Oil-extended rubbers
of high oil content can be used in effect as high-concentration masterbatches
of oil and rubber, in which the oil component has been intimately mixed
with, and absorbed into, the rubber component. The oils used for oil-extending
styrene- butadiene rubbers are of the naphthenic, aromatic and highly-aromatic
types, but not of the paraffinic type. The latter give problems of limited
compatibility with the rubber.
Masterbatches of carbon black and oil-extended styrene- butadiene rubbers
typically have oil and carbon black contents in the ranges 15-65 pphr and
70-100 pphr respectively. The composition of masterbatches of styrene-
butadiene rubbers containing high-styrene styrene-butadiene copolymers
varies widely. The current range appears to be ca. 25-400 pphr.
11.3 LATICES OF ACRYLONITRILE-BUTADIENE COPOLYMERS

Acrylonitrile-butadiene latices are produced by the emulsion copolymerization
of acrylonitrile (Structure VIII) and butadiene. The same structural variations
are possible as for the styrene-butadiene copolymers (see section 11.2.1
above). The glass-transition temperature of atactic polyacrylonitrile is ca.
90·C, being similar to that of atactic polystyrene.
CH 2=CH.CN
VIII
Thus, qualitatively, the copolymerization of acrylonitrile with butadiene has

a similar effect to copolymerization with styrene upon the glass-transition
temperature and associated properties such as the tensile modulus of the
resultant copolymer, and the resilience, hysteresis loss and low-temperature
stiffening behaviour of vulcanizates. As with styrene-butadiene copolymer
latices, the composition of the copolymer dispersed in an acrylonitrile-butadiene
copolymer latex is a very important variable. The effect of acrylonitrile content
upon glass-transition temperature is shown in Fig. 11.5, which is taken from
-10
(;J -20
-...
l!!
:::J
'§ -30
Q)
0..
E
~ -40
c
o
:;:::;
'iii
c -50
jg
ch
l(l
a -60
-700~--~---L--~----~--~--~
10 20 30 40 50 60
Acrylonitrile content of copolymer / % m/m
Figure 11.5 Effect of acrylonitrile content upon glass-transition temperature of acrylonitrile-

butadiene copolymers (Hofmann [7]).
a review by Hofmann [7]. In contrast to the styrene- butadiene copolymers,

the acrylonitrile content of the acrylonitrile-butadiene copolymers which
are used in latex form is always such that the copolymers are rubbery at
normal ambient temperatures. This consideration places an upper limit upon
the acrylonitrile content of ca. 50% m/m.
Although the stiffening effect of copolymerized acrylonitrile in butadiene
copolymers is qualitatively similar to that of copolymerized styrene, there
are very significant differences between the two types of copolymer. It is these
differences which have provided the motivation for the production of
acrylonitrile-butadiene copolymer latices. Polymer repeat units derived from
acrylonitrile are far more polar than are those derived from styrene. Indeed,
the effect of copolymerized acrylonitrile units in raising the glass-transition
temperature of copolymers with butadiene is a consequence of the polarity
of the copolymerized units, rather than of steric hindrance; for styrene-
butadiene copolymers, the latter is the principal factor. The principal effects
which accrue from the polarity of the copolymerized acrylonitrile units in a
butadiene copolymer, other than increase of glass-transition temperature and
associated changes, are that adhesion to polar substrates is increased and
the extent of equilibrium swelling of vulcanizates in hydrocarbon liquids is
reduced. Figure 11.6, taken from a BP Chemicals publication [8], shows the
effect of the acrylonitrile content of acrylonitrile-butadiene rubbers upon
the swelling of vulcanizates in mineral oil, and also upon the low-temperature
stiffening temperature ofvulcanizates. The acrylonitrile- butadiene copolymers
which are available in latex form tend to be classified rather arbitrarily as
Latices of acrylonitrile-butadiene copolymers 325
v
--
I!!
0 30
:.>
~
--
:.!!
0
Q)
a. 20
E -10
Cl
:§
.l!l
Cl
c:
·c 1c:
~ -20 10 0
i3 Q)
Cl
c:
I!! ell
e -30 .s::.
:::>
(.)
Q) 0 Q)
a. E
E :::>
;g
~
...J -40 20
25 30 35 40
-10
Acrylonitrile content of copolymer I % mlm
Figure 11.6 Effect of the acrylonitrile content of acrylonitrile-butadiene rubbers upon low-
temperature stiffening temperature of vulcanizates (curve A) and swelling of vulcanizates in
mineral oil (curve B) (BP Chemicals [8]).
being of low, medium or high acrylonitrile content, according as the

acrylonitrile content of the copolymer is ca. 25%, ca. 33%, or ca. 45% m/m.
The emulsion copolymerization of acrylonitrile and butadiene can be
conveniently effected using systems similar to those used for the emulsion
copolymerization of styrene and butadiene. The same isomeric variations are
possible for the polymerized butadiene and acrylonitrile units as occur in
styrene- butadiene copolymers. The effects of polymerization temperature
upon the proportions ofthe various butadiene isomeric structures are similar.
However, polymerization temperature has less effect upon the mechanical
properties of vulcanizates from acrylonitrile-butadiene rubbers than it does
for styrene-butadiene rubbers. Consequently, polymerization temperature is
a less important variable. For this reason, it appears that the majority of
acrylonitrile-butadiene copolymer latices are produced by emulsion poly-
merization at ca. 50°C, rather than by low-temperature emulsion polymerization
at ca. soc. Problems can arise from the activating effect of traces of amines
which may be present in acrylonitrile as by-products of its manufacture, or
which may have been added to the acrylonitrile monomer as a polymerization
inhibitor. Ifpresent in variable amounts, such impurities may cause unexpected
variations of rate of polymerization. Low-temperature emulsion polymerization
reaction systems initiated by redox combinations are said to be rather less
sensitive to the influence ofthese impurities than are peroxodisulphate-initiated
systems. Reference has been made in section t 0.2.3.3 of Chapter tOto the
possibility of Michael addition between acrylonitrile and mercaptan modifiers.
11.4 POLYCHLOROPRENE RUBBER LATICES

Chloroprene (2-chloro-l,3-butadiene) (Structure IX) can readily be emulsion
polymerized and copolymerized to form stable latices. Provided that the
polychloroprene is not extensively branched and crosslinked (see further
below), the contained polymer is soft and elastomeric. It is more polar than
polybutadiene, because of the presence of the carbon-chlorine bonds, but
not so polar as acrylonitrile-butadiene copolymers of medium and high
acrylonitrile contents. The glass-transition temperature of polychloroprene
is ca. -4S°C. Polychloroprene rubber can be crosslinked to give strong
vulcanizates. This is usually achieved by heating with metal oxides, which
bring about crosslinking by interaction with the chlorine atoms, rather than
by reaction with reagents which utilize directly the reactivity of the
carbon-carbon double bonds and neighbouring sites. Polychloroprene
rubber vulcanizates have moderate resistance to swelling in hydrocarbon
oils, and good resistance to various chemical deteriorative influences such
as ozone. Adhesion to polar substrates is good. Apart from carboxylic-acid
monomers, the only comonomer commonly copolymerized with chloroprene
is 2,3-dichloro-l,3-butadiene (Structure X). By its use, it is possible to increase
the chlorine content of the polymer.
CH 2=CCI-CH=CH 2 CH 2=CI-CCI=CH 2
IX X
Like butadiene, individual chloroprene units can become incorporated in

a polymer chain in various ways, giving rise to polymerized chloroprene
units which have differing isomeric structures. However, matters are further
complicated with chloroprene because, unlike butadiene, the isomeric structure
which results from 3,4-addition differs from that which results from 1,2-
addition. Thus during free-radical polymerization in general, and emulsion
free-radical polymerization in particular, chloroprene molecules can become
incorporated in the polymer chain either by l,4-addition, or by 1,2-addition,
or by 3,4-addition. The incorporated units have structures XIa, XIb and XIc
... -CH 2-CCI=CH-CH 2-··· ... -CH2iCI- ... -9H- ...
. .. -CH 2
XIa
CH CCI
II II
CH 2 CH 2
XIb XIc
respectively. As with butadiene units, the disposition of the polymer chain
with respect to the olefinic double bond in anyone 1,4-unit can be either cis
or trans. Both 1,2 and 3,4 units contain chiral carbon atoms, and so may
have either the (R) or (S) configuration. Insofar as sequences of these units
Polychloroprene rubber latices 327
are present, they may be either isotactic, syndiotactic or atactic, and successive
pairs of units may add head-to-head, head-to-tail or tail-to-tail. The presence
of the chlorine substituent also introduces this latter complication for
sequences of 1,4 units. Thus any particular chloroprene molecule can become
polymerized in one of several modes. The predominant mode of addition in
free-radical polymerization is trans-1,4. Small proportions of units also
become polymerized in the cis-1,4, 1,2 and 3,4 modes. Units incorporated in
the 1,2 mode have important implications for the chemical properties of the
polymer, both in latex form and in bulk form (see section 11.4.3 below). The
relative proportions of the various isomeric chloroprene units depend to
some extent upon the poymerization temperature, but the effect is not so
marked as for the free-radical polymerization qf butadiene. Results reported
by Maynard and Mochel [9] for the isomeric proportions in polychloroprene
rubbers formed at various temperatures are shown in Table 11.6. It is seen
that the proportions of 1,2 and 3,4 units are small, are approximately equal,
and tend to increase with increasing polymerization temperature. The ratio
of trans-1,4 to cis-I,4 units decreases with increasing polymerization temperature,
but is always high, being ca. 20/1 and ca. 5/1 for polychloroprene rubbers
produced at - 40·C and lOO·C respectively. Branching and crosslin king occur
during the emulsion polymerization of chloroprene once significant amounts
of polymer are present.
The extent of the stereo regularity of the polychloroprene chain affects the
extent and rate of crystallization of the polymer. Unless the polymer contains
a high proportion of trans-I ,4 chloroprene units, the polymer crystallizes
only very slowly, or possibly not at all. The extent of crystallization affects
the physical properties of the polymer. As it increases, so the cohesive strength,
modulus, tensile strength, hardness, resistance to permanent set and resistance
to swelling in hydrocarbon liquids increase sharply, whereas the extension
at break decreases. The effects which accompany polymer crystallization are
reversed when the polymer is heated, but reappear on subsequent cooling.
Table 11.6 Effect of polymerization temperature upon proportions of various

isomeric structures resulting from free-radical polymerization of chloroprene (Maynard
and Mochel [9])
Polymerization Proportion of isomeric structure indicated / %

temperature
rC 1,2 3,4 cis-I,4 trans-I,4
-40 0.9 0.3 5 94

-10 7
10 1.1 1.0 9 84
40 1.6 1.0 10 86,81
50 11
80 2.0 2.1
100 2.4 2.4 13 71
When chloroprene is emulsion-polymerized in the absence of any substances

which either themselves reduce the chain length of the polymer, or make the
chain susceptible to cleavage by a subsequent reaction, the product which
forms, even at low conversions, is, and remains, a tough, insoluble, non-plastic
material. Emulsion polymerization reaction systems for the production of
polychloroprene rubber latices can be divided broadly into two types
according to the manner in which reduction of polymer chain length is
achieved. In the first type of system, sulphur to the extent of ca. 0.5-1.5 pphm
is incorporated as a latent modifier. The sulphur is thought to copolymerize
with the chloroprene in the manner indicated at Structure XII; the value of
... +CH 2-CCI=CH-CH 2-).-Sx+CH2-CCI=CH-CH2-).-Sx- ...

XII
x is in the range 2-6, and that of n is ca. l00x. The immediate polymeric
product of the emulsion polymerization reaction is still a tough, insoluble
material. If, however, the latex at alkaline pH is treated with a suitable
thiophilic reagent, such as tetraethylthiuram disulphide, the polymer becomes
converted to a soft, plastic material which is soluble in appropriate organic
solvents. Softening, or peptization, of the polymer is thought to be a
consequence of cleavage of the polysulphide links in the copolymer by
interaction with dithiocarbamate anions (R 2NCS.S -) derived from the
thiuram polysulphide (R 2NCS.S y.S.CSNR 2). Polychloroprene rubbers produced
by this method are described as sulphur-modified or (perhaps more appropriately)
as thiuram-modified; polychloroprene rubbers produced using conventional
modifiers, such as mercaptans, are described as nOR-sulphur-modified.
It will be evident from these considerations that, quite apart from any
variations in polymer structure and properties which may be brought about
by copolymerizing chloroprene with other monomers, a wide range of
polymer properties is in principle obtainable from chloroprene homopolymers.
The principal structural variations which give rise to property variations are
those which have been outlined above, namely, the degree of stereoregularity
of the polymer, the molecular size distribution of the polymer, the gel content
of the polymer, and the concentration of crosslinks in the polymer gel. The
degree of stereo regularity of the polymer is determined principally by the
polymerization temperature; the other molecular characteristics are determined
by polymerization temperature and the extent to which the chain length of
the polymer has been reduced by the use of modifiers.
11.4.2 Production of polychloroprene rubber latices

A simple emulsion-polymerization formulation for the production of a
carboxylate-stabilized latex of a sulphur-modifiable polychloroprene rubber
by peroxodisulphate-initiated batch emulsion polymerization at 40°C is
Table 11.7 Formulation for production of a sulphur-modifiable carboxylate-stabilized

polychloroprene rubber latex by batch emulsion polymerization at 40·C (Neal and
Mayo [10])
Ingredient Parts by mass
Chloroprene 100
N wood rosin 4 } dissolved in the
Sulphur 0.6 monomer
Water 150
Sodium alkylnaphthalenesulphonates 0.7
Sodium hydroxide 0.8
Potassium peroxodisulphate 0.2-1.0
shown in Table 11.7. The sulphur and rosin acid are first dissolved in the
chloroprene. This solution is then emulsified in water in which the sodium
hydroxide and the sodium alkylnaphthalenesulphonate have been dissolved.
The rosin acid and sodium hydroxide interact to form a carboxylate soap,
sodium resinate, in situ at the interface between the chloroprene droplets
and the aqueous phase. The soap anions then stabilize the chloroprene
droplets against coalescence. By this means, a stable emulsion of chloroprene
droplets in water is formed. Presumably some of the sodium resinate also
dissolves in the aqueous phase, thereby providing micelles for the nucleation
of reaction loci by the micellar mechanism. The function of the sodium
alkylnaphthalenesulphonate is partly to assist breaking up of the chloroprene
droplets initially, but primarily to enhance the colloid stability of the latex
which is subsequently produced. Emulsification is achieved by recirculating
the chloroprene-water mixture through a centrifugal pump. The monomer
emulsion contains ca. 40% mlm of chloroprene. After emulsification, the
emulsion is pumped into a jacketed, glass-lined kettle equipped with a
glass-coated agitator. An aqueous solution of potassium peroxodisulphate
is then added to initiate the polymerization. The temperature of the reaction
mixture is maintained at 40·C by circulating cooled brine through the jacket
of the kettle, and also by varying the agitator speed. The progress of the
polymerization can be followed by way of changes in the specific gravity of
the reaction mixture. At ca. 90% conversion, at which stage the specific
gravity is ca. 1.07, the polymerization is stopped by addition of tetraethylthiuram
disulphide as an aqueous emulsion of a solution in xylene. The latex is then
cooled to 20·C and allowed to stand for ca. 8 hours. During this time, the
thiuram disulphide cleaves some of the chloroprene-sulphur linkages, thereby
peptizing the polymer. The cleavage reaction can be arrested by acidifying
the hitherto alkaline latex to pH 5.5-5.8 by addition of 10% mlv acetic acid.
The thiuram disulphide also acts as a polymer stabilizer.
Some information concerning the peroxodisulphide-initiated emulsion
polymerization of chloroprene is available in papers by Morton, Cala and

Altier [11], and by Morton and Piirma [12]. Chloroprene is more reactive
in emulsion polymerization than are other common dienes. It appears that
propagating polymer chains can be initiated from the interface between the
monomer droplets and the aqueous phase, and also that reaction loci can
be nucleated in the aqueous phase by the homogeneous mechanism. However,
in conventional reaction systems, micellar nucleation is probably the dominant
mechanism whereby reaction loci are formed. The pH of the reaction system
tends to fall markedly during the reaction. This is partly because bisulphate
anions form from the peroxodisulphate initiator, but mainly because hydrochloric
acid is formed by hydrolysis of polychloroprene already formed. This
matter is discussed further below (see section 11.4.3). If carboxylate colloid
stabilizers are used to facilitate the reaction, it is desirable to buffer the
system with, say, sodium triphosphate. Somewhat variable induction periods
for the polymerization have been observed. Using sodium dodecylbenzene-
sulphonate as colloid stabilizer, a substantial rate of emulsion polymerization
can be observed even in the absence of peroxodisulphate initiator.
This phenomenon is attributed to the presence of peroxides in the colloid
stabilizer.
11.4.3 Hydrolysis of polychloroprene rubbers in latex form

The pH of polychloroprene rubber latices falls slowly during storage. The
principal reason is that hydrochloric acid is liberated by slow hydrolysis of
carbon-chlorine bonds of the dispersed polymer. Polymers which contain
chloroprene units provide one of the two common types of polymer which
in aqueous latex form are subject to interaction at appreciable rates with the
dispersion medium by hydrolysis; the other type comprises vinyl acetate
polymers and copolymers (see section 11.5.4 below). The fall in pH during
storage which occurs with polychloroprene rubber latices has important
adverse consequences both for the latex itself and for films obtained from
the latex. As regards the latex itself, the principal adverse effect is reduction
in colloid stability, which can lead to gross coagulation. The minority of
allylic carbon-chlorine bonds (C=C-C -Cl) present in the chloroprene units
which have polymerized by 1,2-addition are believed to be particularly
susceptible to hydrolysis; the vinylic carbon-chlorine bonds (C=C-Cl)
present in chloroprene units incorporated by 1,4- and 3,4- addition are
relatively unreactive as regards hydrolysis. Thus the proportion of the total
polychloroprene chlorine which can be liberated into the aqueous phase as
chloride ions corresponds almost exactly to the proportion of 1,2 chloroprene
units in the polymer, as determined by infrared spectroscopy. The hydrolysis
of the carbon-chlorine bond in a 1,2 chloroprene unit can be represented
as follows:
CI OH
I I
... -CHz-C- ... + HO- -----+ ... -CHz-C- ... + Cl-
I I
CH CH
II II
CH z CH z
The presence of the minority of allylic chlorine atoms is also thought to
make possible vulcanization of emulsion-polymerized polychloroprene rubber
by heating with metal oxides. If the allylic chlorine content of the polymer
is significantly reduced, e.g., by reaction with piperidine, it is impossible to
vulcanize the rubber properly by heating with metal oxides.
The rate at which chloride ions are liberated into the aqueous phase by
hydrolysis at 90°C is illustrated by results reported by Andersen and Kovacic
[13], reproduced here as Fig. 11.7. As expected, the rate decreases with
increasing extent of hydrolysis. The reaction kinetics fit neither the SN1 nor
the SN2 patterns. However, the hydrolysis mechanism within the latex
particles is probably of the SN2 type, the overall kinetics being complicated
by the heterogeneity of the reaction system and the presence in the polymer
of carbon-chlorine bonds of different reactivities towards hydrolytic reagents.
It has also been observed that the rate of liberation of chloride ions into the
aqueous phase is independent of the initial pH of the latex (and therefore of
the concentration of hydroxide ions in the aqueous phase), and of the
1.5
iD ~
E <1l
»..c
(50.
a. ~ 1.0
.!: 0
(]) (])
,S 5-
D <1l
:<:.,9
u .~
<UTI
'0,&0.5
- <1l
'0 iD
~~
10 20 30 40 50
Time/hour
Figure 11.7 Kinetics of liberation of chloride ions into the aqueous phase of a polychloroprene
latex by hydrolysis at 9O"C (Andersen and Kovacic [13]). Significance of points: • results
obtained by chemical analysis of aqueous phase; • results obtained by infrared analysis of
polymer phase.
concentration of any chloride ions which may be present in the aqueous

phase initially. As corollaries of this, the rate of consumption of alkali is in
practice substantially constant over normal storage periods, and the storage
life of a latex increases approximately linearly with its initial alkali content.
A consequence of the slow hydrolysis reaction is the necessity for
periodically estimating the residual alkali content of polychloroprene rubber
latices. A single pH measurement is not regarded as adequate. Being a
logarithmic scale as regards hydrogen-ion concentration, pH does not provide
a sufficiently sensitive measure of the alkali content of the aqueous phase at
the pH values normally encountered in these latices. In practice, advantage
is taken of the presence of two distinct inflexion points in the titration curve
for polychloroprene latices for the variation of pH with amount of acid added.
One is at pH 10.5, and is attributed to neutralization of unreacted alkali; the
other is at pH 4.5, and is attributed to reaction with the resinate soap anions.
Thus the residual available alkali is conveniently assayed from the amount
of acid required to reduce the pH from its initial value to 10.5. The following
procedure has been recommended by Cook and Fitch [14]; lOOg of
polychloroprene rubber latex is stabilized colloidally by addition of 20cm 3
of a 5% mjv aqueous non-ionogenic ethoxylate solution. The mixture is then
titrated with Y3 mol dm - 3 hydrochloric acid, the pH being followed during
the titration by a glass-electrode pH meter. The amount of acid required to reduce
the pH to 10.5 is noted. The result is expressed as mg-equivalents (l00 g
latex)- 1.
The adjustment of the alkali content of a polychloroprene rubber latex is
as important as its estimation. For many processes, the alkalinity should be
adjusted to a definite level if reproducible effects are to be observed. If the
latex is alkali-deficient, then it is merely necessary to add sufficient of a dilute
aqueous solution of an alkali such as sodium or potassium hydroxide. If too
much alkali is present, then the excess can be conveniently removed by
neutralization with glycine (aminoacetic acid), which also tends to buffer the
latex against further pH change. Glycine is a convenient substance to usc
for this purpose, because it is a strong enough acid to neutralize effectively
the excess alkali, and yet, being amphoteric, it does not cause colloidal
destabilization of the latex. Furthermore, it is readily soluble in water, has
no serious known associated health hazards, is odourless, does not cause
discoloration of films from the latex, and is cheap, readily available and
convenient to handle.
11.4.4 Effects of polymer hydrolysis during storage upon the properties of

polychloroprene rubber latices, and upon the properties of films derived
from the latices
Shepelev et al. [15] have reported the effects of accelerated ageing, in sealed
vessels for up to 5 days at 70·C, of polychloroprene rubber latex which
contained ammonia. They observed a slight fall in pH and alkalinity, a slight

increase in total solids content, due presumably to the fixation of volatile
alkali by the liberated acid, a reduction in surface free energy from 60 to
37 mJ m - 2, and an increase in the average particle size and extent of soap
coverage ofthe particles. These latter effects were attributed to partial particle
agglomeration, with the concurrent liberation of soap into the aqueous phase.
The tensile strength and extension at break of un vulcanized films derived
from the latices decreased with ageing, but there was an increase in modulus.
Similar trends were observed for vulcanized films. The processing characteristics
of the latices were found to deteriorate with ageing.
Increasing the alkali content of polychloroprene latex causes a significant
decrease in the initial viscosity of the latex, and ~lso in the increase of viscosity
which occurs as the latex is aged. However, an increase in alkali content
reduces the sensitivity of the latex to the types of coacervants which are used
in dipping processes and in processes for the manufacture of latex foam
rubber which depend upon gelation of the latex phase of the foam. There is
therefore a practical limit to the extent to which the viscosity of polychloroprene
latex should be stabilized by alkali addition. Alkali content can also have a
slight effect upon the physical properties of vulcanized films derived from
polychloroprene latex. There is a tendency for both modulus and tensile
strength to go through shallow maxima as the alkali content of the latex is
increased. On the other hand, the extension at break tends to increase slowly
but progressively.
Hydrolysis of the polymer in polychloroprene latex has appreciable effects
upon the properties of films derived from the latex. In particular, the tensile
strength of vulcanized films decreases significantly as hydrolysis progresses.
There are probably at least two reasons for this reduction:
t. Some crosslinking of the polymer probably occurs within the particles

during the ageing of the latex, perhaps by formation of ether linkages by
interaction between reactive chlorine atoms and the hydroxyl groups which
are produced by hydrolysis. Crosslinks formed during ageing of the latex
may reduce tensile strength because they interfere with the coalescence of
the latex particles as deposits form from the latex.
2. The number of reactive sites available for crosslinking by heating with
metal oxides is reduced by hydrolysis.
The essential correctness of these views is consistent with the observation

that, if (-thiocaproclactam (Structure XIII) is added to the latex prior to
ageing, then extensive crosslinking occurs within the latex particles, and a
drastic reduction of film tensile strength occurs. If the polymer has been sulphur-
modified, this reduction in tensile strength can be prevented by treating the
latex with, say, tetraethylthiuram disulphide, in which case it is possible that
cleavage offurther intrapolymer polysulphide linkages offsets to some extent

at least the effects of the formation of crosslinks during ageing.
XIII
11.4.5 Some types of polychloroprene rubber latex currently available

Information concerning the types of non-functionalized polychloroprene latex
manufactured and supplied by the du Pont Company at the time of writing
is summarized in Table 11.8. The principal monomer is, of course, chloroprene
in all cases. The polymers contained in two of the latices are copolymers of
chloroprene and 2,3-dichloro-I,3-butadiene. The copolymer contained in one
of these grades, Neoprene Latex Type 400, contains a relatively high
proportion of the comonomer. It crystallizes very rapidly. The copolymer
contained in the other grade, Neoprene Latex Type 750, contains less
comonomer, and crystallizes very slowly. In addition, a carboxylated
polychloroprene latex is currently available from the du Pont Company
under the designation Neoprene Latex Type 115. The carboxylating acid is
said to be methacrylic acid. Although the level of carboxylation is not stated,
the chlorine content of the solids fraction of this latex is only slightly less
than that ofthe solids fraction of a latex said to cont~in pure polychloroprene.
It is inferred from this that the level of carboxylation is low. The colloid
stabilizer is said to be polyvinyl alcohol. This is, of course, non-ionogenic.
The pH of the latex is typically 7, and thus it is unlikely that the particles
will carry many surface-bound anionic groups derived from combined
methacrylic acid. It therefore seems likely that the latex itself is essentially
non-ionic. The technology of polychloroprene latices manufactured and
supplied by the du Pont Company is described in detail in a book by Carl [16].
11.5 LATICES OF VINYL ACETATE POLYMERS AND COPOLYMERS
11.5.1 Types of vinyl acetate polymer and copolymer latices currently available
The homopolymer of vinyl acetate (Structure XIV) has a glass-transition
temperature of ca. 30·C. The minimum film-forming temperature for polyvinyl
acetate latices is somewhat lower, being typically ca. 20·C, because the
polymer in latex form is plasticized by absorbed water and also by substances
Table 11.8 Information relating to types of non -functionalized polychloroprene latices manufactured and supplied by the du Pont Company
Chlorine Brookfield viscosity

content Total at 25"C
Neoprene oJ latex Typical solids
Latex Comonomer. Colloid solids; pH content viscosity spindle v§ Distinguishing Jeatures
Type if any· stabilizert /% mfm at 25"C /% mfm /mPas oJ latex and polymer
60 Sulphur NaDRA 37.5 11.0 59 800 #2 6 Develops high wet-gel

350 #2 30 strength
400 2,3-dichloro- KDRA 48.0 12.5 50 9 #1 6 High chlorine content, rapid
1,3-butadiene 9 #1 30 crystallization, outstand-
ing ozone and weather
resistance
571 Sulphur NaRA 37.5 12.0 50 15 #1 6 High strength, low-
15 #1 30 permanent-set vulc. films
622 KDRA 38.0 12.5 61 65 #1 6 High solids, low-viscosity,
65 #1 30 high-gel polymer
654 KDRA 38.0 12.0 59 75 #1 6 High solids, low-viscosity,
55 #1 30 low-gel polymer
671A KDRA 38.0 12.5 59 60 #1 6 High solids, low-viscosity,
45 #1 30 medium-high-gel polymer
735A NaDRA 38.5 12.0 45 5 #1 6 For wet-end addition to
5 #1 30 fibrous slurries
750 2,3-dichloro- KDRA 40.0 12.5 50 10 #1 6 High wet-gel strength, low-
1,3-butadiene 10 #1 30 modulus films, very slow
crystallization
842A NaRA 37.5 12.0 50 IS #1 6 Medium-strength vulcan-
15 #1 30 ized films, slow
crystallization
• For the purposes or this table. the sulphur used to modiry the polymer is regarded as being a comonomer. However. sulphur-modified polychloroprenes are usually regarded
as being chloroprene homopol ymers.
t NaRA = sodium soap or rosin acids; NaDRA = sodium soap or disproponionated rosin acids; KDRA = potassium soap or disproponionated rosin acids.
! It is presumed that these entries are ror the chlorine content or the solids rraction or the latices. The theoretical chlorine content or pure polychloroprene is ca. 40.0% mjm.
§ v = rotational rrequency or viscometer (minute - I).
such as colloid stabilizers. These values of glass-transition temperature and

minimum film-forming temperature make polyvinyl acetate latices unsuitable
for many applications. The minimum film-forming temperature is too high
for the latex polymer to act as an effective binder for pigments and fillers at
normal ambient temperatures; it is also too high for it to be effective in
certain adhesive applications at these temperatures. On the other hand, the
glass-transition temperature is too low for the polymer to be useful as a rigid
plastic at normal ambient temperatures, because the polymer is too near its
softening temperature. To make the polymer more suitable for application
as a binder and an adhesive, it is usual to lower its glass-transition
temperature by plasticization. This can, of course, be achieved by the addition
of external plasticizers; this method was widely used in past years. However,
external plasticization can be attended by problems such as migration of the
plasticizer into substrates, with adverse consequences for both the polymer
and the substrate, and extraction of the plasticizer from the polymer by
organic liquids. The more usual method of lowering the glass-transition
temperature is by internal plasticization, achieved by copolymerizing the
vinyl acetate with one or more comonomers which give homopolymers
having glass-transition temperatures well below normal ambient temper-
atures. The comonomers commonly used for this purpose can be classified as
follows:
I. vinyl esters of n-alkanoic acids higher than acetic;

2. vinyl esters of certain synthetic branched-chain alkanoic acids;
3. esters of acrylic acid;
4. olefinic hydrocarbons.
Examples of vinyl esters of higher n-alkanoic acids include vinyl caprate

(n-decanoate) (Structure XV) and vinyl stearate (n-octadecanoate) (Structure
XVI). The best-known examples of vinyl esters of a synthetic branched-chain
alkanoic acid are those of the so-called versa tic acids. These are branched-
chain acids which have a quaternary carbon atom in the alkyl group. The
vinyl esters commonly copolymerized with vinyl acetate are derived from
acids which contain approximately nine carbon atoms in their alkyl moieties.
Such esters are now usually known as vinyl neodecanoates. Information
concerning the structure of the parent neodecanoic acids has been disclosed
by Shell Chemicals [17], from which it is evident that the structure of vinyl
neodecanoate is as shown at Structure XVII. In this structure, R 1 , R2 and
R3 are alkyl groups, at least one of which is methyl, and the total number
of carbon atoms in the three groups is 8. Examples of esters of acrylic acid
which are used as plasticizing comonomers for vinyl acetate include n-butyl
acrylate (Structure XVIII) and 2-ethylhexyl (isooctyl) acrylate (Structure XIX).
The only olefinic hydrocarbon commonly used for this purpose is ethylene
(Structure XX).
Latices of vinyl acetate polymers and copolymers 337
CH 2=CH.O.CO.CH 3
XIV
CH 2=CH.O.CO.(CH 2)sCH 3 CH 2=CH.O.CO.(CH 2)16 CH 3
XV XVI
CH 2=CH CH 2=CH.Co.O(CH 2hCH 3
I XVIII
o
I CH 2=CH.CO.O.CH 2 CH(CH 2CH 3 )(CH 2hCH 3
CO
I XIX
C xH 2x + l-C-C zH2z+ 1 CH 2=CH 2 CH 2=CHCl
I
C y H 2y + 1 XX XXI
x+y+z=8
XVII
It will be evident from the range of comonomers which is available for
emulsion copolymerization with vinyl acetate that a very large number of
types of copolymer can be prepared. Furthermore, several variations of the
structure of polymerized units and sequences of polymerized units are
possible: all the monomers shown as Structures XIV - XIX introduce a chiral
carbon atom into the polymer sequence, and head-to-head, head-to-tail and
tail-to-tail variations are possible for pairs of successive units. Superimposed
upon these structural complexities are further possibilities which arise from
the susceptibility of all the polymerized units, except ethylene, to variable
extents of hydrolysis. Thus vinyl acetate polymers and copolymers prepared
by emulsion polymerization almost invariably contain small proportions of
what are, in effect, other types of comonomer units derived from the main
monomer units by hydrolysis. Of these other types of comonomer unit, the
most prevalent is vinyl alcohol, formed by hydrolysis of vinyl acetate units.
The hydrolysis of vinyl acetate polymers and copolymers is important in
relation to the applications of these polymers; hydrolysis of these polymers
in latex form is considered in section 11.5.4 below.
The selection of the comonomer to be used to produce latices for any given
application depends principally upon the functional suitability of the comonomer
and its cost. In particular, considerable interest has been shown in recent
years in ethylene as a comonomer for vinyl acetate, mainly because of its
low cost and ready availability. It is remarkably efficient as a plasticizing
comonomer for vinyl acetate, as is discussed in section 11.5.3 below where
the plasticization of vinyl acetate polymers by copolymerization is considered.
However, if the proportion of ethylene in an ethylene- vinyl acetate copolymer
exceeds ca. 20% mlm, then the copolymer becomes too soft and weak for
use in applicational areas such as adhesives. To overcome this problem,
second stiffening comonomers have been used to offset the plasticizing effect
of the ethylene. One comonomer widely used for this purpose is vinyl chloride
(Structure XXI). Thus vinyl acetate-ethylene-vinyl chloride copolymer
latices provide yet another family of vinyl acetate copolymer latices. The
composition of such terpolymers can, of course, vary widely; a typical vinyl
acetate/ethylene/vinyl chloride ratio for a latex polymer suitable for application
as a pigment binder is 60/20/20 m/m/m.
11.5.2 Production of vinyl acetate polymer and copolymer latices

Much literature has been published concerning the emulsion polymerization
and copolymerization of vinyl acetate. A comprehensive review is available
in a book edited by EI-Aasser and Vanderhoff [18]. Although the principal
features of the reaction are understood, many of the details remain obscure.
The emulsion polymerization of vinyl acetate differs from that of styrene and
butadiene in several important respects. The greater water-solubility of the
monomer has the consequence that micellar nucleation is not the only
mechanism by which reaction loci can form. It has long been recognized that
vinyl acetate can be emulsion polymerized in aqueous media under conditions
such that conventional micellar nucleation is not possible, because no
micelle-forming surface-active substance is present; indeed, it was this
observation which primarily led to the development of the theory of
homogeneous nucleation. Another important respect in which the industrial
emulsion polymerization of vinyl acetate differs from that of styrene and
butadiene is that polymerization should be as complete as possible, because
it is not feasible to remove unpolymerized monomers. It is undesirable that
the residual monomer content should exceed 0.5 pphm, partly because of bad
odour, especially if the reaction system contains acrylic esters, and partly
because acetic acid is formed by the subsequent hydrolysis of residual vinyl
acetate during storage. The products of this hydrolysis are acetic acid and
acetaldehyde (which can subsequently oxidize to acetic acid), the latter being
formed by tautomeric rearrangement of the vinyl alcohol which is presumed
to be the initial product of the hydrolysis:
CH 2 =CH.O.CO.CH 3 + H 20 ----+ [CH 2 =CH.OH] + HO.CO.CH 3
vinyl alcohol
1
CH 3 ·CHO
acetaldehyde
Emulsion polymerization reaction systems for the production of vinyl acetate

polymer and copolymer latices usually contain some colloid stabilizer to
protect both the developing latex particles and the eventual latex against
colloidal destabilization. The colloid stabilizer may be entirely of the
hydrocolloid type, in which case particle nucleation probably occurs exclusively
by a non-micellar mechanism. Alternatively, a micelle-forming colloid stabilizer
may be used, either as the sole stabilizer or in combination with a

hydrocolloid, in which case particle nucleation by the micellar mechanism
can occur. Hydrocolloids used to confer colloid stability upon latices are
commonly known as protective colloids. One particular type of hydrocolloid
which has been widely used for this purpose is partially-hydrolysed polyvinyl
acetate, usually known as 'polyvinyl alcohol'. Being effectively a copolymer
of vinyl alcohol and vinyl acetate, it is well suited to this application. Quite
apart from any graft copolymerization which may occur between the
propagating polyvinyl acetate chains and the hydrocolloid (see further below),
association between the vinyl acetate units of the hydrocolloid and those of
the polymer which forms provide a means whereby the hydrocolloid
macromolecules can be firmly anchored to the particle surface. The vinyl
alcohol units provide the hydrophilicity which is essential if the hydrocolloid
macromolecule is to tend to mix with the aqueous phase and thereby impart
steric and hydration stabilization. Various free-radical generating systems
can be used for initiating vinyl acetate emulsion polymerizations and
copolymerizations. Peroxodisulphates have been widely used as thermal
dissociative initiators, typical reaction temperatures being in the range
60- 70·C. Redox systems are used to initiate polymerization at lower
temperatures. Thus Elgood and Gulbekian [19] have reported information
concerning the emulsion polymerization of vinyl acetate at 50·C using as
initiator a redox combination comprising ammonium peroxodisulphate, a
reducing agent such as sodium formaldehyde sulphoxylate, and traces of
metal ions such as iron (III).
Information concerning the emulsion polymerization of the vinyl esters of
longer-chain alkanoic acids has been reported by Port et al. [20]. The vinyl
esters with which they worked were those of caprylic (n-octanoic), capric
(n-decanoic), lauric (n-dodecanoic), myristic (n-tetradecanoic), palmitic
(n-hexadecanoic) and stearic (n-octadecanoic) acids. They encountered certain
difficulties in obtaining satisfactory latices, the principal one being that of
achieving satisfactory colloid stability both before and throughout the
polymerization reaction. Colloid stabilizers such as carboxylate soaps and
alkylarylsulphonates were unsatisfactory. Investigation of a large number of
possibilities led to the conclusion that the most satisfactory colloid stabilizers
for these monomers are mixtures of non-ionic substances such as the fatty-acid
esters of anhydrosorbitol and anhydromannitol, and ionic surface-active
sulphated branched-chain alcohols. A complicating factor in the emulsion
polymerization of the vinyl esters ofthe higher n-alkanoic acids is the decreasing
water-solubility of the esters as the alkyl chain length of the ester increases.
A typical procedure for the emulsion polymerization of vinyl acetate has
been described by Warson [21]. The polymerization kettles are cylindrical
in shape, and vary in size from ca. 2 x 102 to 4.5 X 104 dm 3 • They should be
lined either with stainless steel or with vitreous enamel, because of the
inhibiting effect of traces of iron and copper upon the polymerization reaction.
The vessels are double-walled to permit steam-heating and water-cooling.

Warson stresses particularly the need for efficient cooling. Heating may also
be by injection of live steam; this has the advantage of avoiding the possible
formation of a solid resinous deposit upon the walls of the reactor. Some
reactions are carried out under reflux conditions; in other cases, an autogenous
pressure is allowed to develop. A burst disc or safety valve should always
be present. In many cases, it is necessary to purge with nitrogen. The
monomers may with advantage be pre-emulsified, and may be added either
initially or else as a fine stream. In the latter case, the reaction can be carried
out under monomer-starved conditions, the amount of unreacted monomer
present at any instant being low. Continuous agitation must be maintained,
but the speed of stirring should be diminished as the size of the reaction
vessel increases, otherwise the increased peripheral stirrer speed may cause
mechanical destabilization of the system. When carrying out the reaction
under reflux conditions, the condenser should have an independent return
to the reaction vessel. Warson draws attention to the need to use softened,
completely de-ionized, or even distilled water, as the dispersion medium for
the emulsion polymerization of vinyl acetate. If this precaution is not
observed, the metal ions in the water may have an adverse effect upon the
rate of polymerization.
The emulsion copolymerization of vinyl acetate and ethylene has to be
carried out under pressure, because ethylene is a gas at normal temperatures
and pressures. For this reason, vinyl acetate-ethylene latex copolymers are
sometimes colloquially referred to as pressure polymers. When the process
is intended to produce copolymers which contain less than, say, 20% mlm
of ethylene, the pressure is usually less than ca. 10 MPa. Some information
concerning the factors which control the ethylene content of ethylene-vinyl
acetate copolymers produced by emulsion copolymerization is available in
a paper by L6hr [22]. The monomer reactivity ratios for the free-radical
copolymerization of ethylene and vinyl acetate are both almost 1; thus the
copolymerization behaviour of this pair is almost ideal, in that the composition
of the copolymer which forms at any instant is always near that of the
monomer feed mixture from which it was formed. L6hr found that the ethylene
content of the copolymer is affected not only by the ethylene pressure and
the reaction temperature, but also by the polymerization procedure. Thus
for a reaction at 60°C and 3 M Pa pressure, the ethylene content of the
copolymer can be increased from 10 to 20% mlm by adding the vinyl acetate
to the reaction system gradually, rather than adding the whole amount at
the commencement of the reaction. The increase in ethylene content of the
copolymer formed arising from this change of procedure is also influenced
by the nature of the initiator system and of the colloid stabilizers. This effect
has been explained qualitatively by L6hr by taking account of the small
solubility of ethylene in the copolymer.
The composition of a typical reaction system for the production of an
anionic latex of a 75/25 m/m vinyl acetate/vinyl neodecanoate copolymer

latex is given in Table t 1.9. A latex prepared from this reaction system is
said to have good electrolyte stability, good mechanical stability, good
freeze-thaw stability and good storage stability, and to be suitable as a
binder for latex-based surface coatings of high quality. The procedure for
carrying out the emulsion polymerization reaction, as given by Shell [23],
is as follows: The aqueous phase, together with t 0% of the monomer mixture,
are heated to 74- 77·C, and maintained at this temperature for ca. t 5 minutes
until the initial monomer charge has polymerized. The remainder of the
monomer mixture and the initiator are then added as two separate streams
over a period of ca. 2 hours, the temperature being allowed to rise to ca.
77-80·C. When the monomer addition has been completed, the reaction
temperature is maintained for a further period of t -2 hours. Initiator solution
is then added at such a rate that the addition is complete after the first 15-20
minutes ofthis latter period. The latex is then cooled off and, if necessary, filtered.
One particular aspect of the production of vinyl acetate polymer and
copolymer latices by emulsion polymerization which has been the subject of
some questioning concerns the tendency of any hydrocolloids present in the
reaction system during the polymerization to graft copolymerize with vinyl
acetate. Hartley [24] has presented evidence to show that polyvinyl
acetate-polyvinyl alcohol graft copolymers are formed when vinyl acetate
is emulsion copolymerized in an aqueous system comprising water, polyvinyl
alcohol, an alkali and a water-soluble peroxodisulphate initiator. The
presence of these graft copolymers has been established by turbidimetric
Table 11.9 Typical reaction system for the production of an anionic latex of a
75/25 m/m vinyl acetate/vinyl neodecanoate copolymer latex (Shell [23])
Amount
Ingredient / pphm
Aqueous phase
Demineralized water 70
Sodium dodecylbenzenesulphonate 85% active 0.3
20-mole nonylphenol ethoxylate 25% aqueous solution 8
Hydroxyethylcellulose 2
Borax 0.5
Glacial acetic acid 0.2
Monomer phase
Vinyl acetate 75
Vinyl neodecanoate 25
Initiator solution
Demineralized water 12
titration, and by various chromatographic techniques. It is possible that other

hydrocolloids, such as water-soluble cellulose derivatives and polyacrylates,
also form graft copolymers with vinyl acetate. Thus Bondy and Coleman
[25] cite hydroxyethyl cellulose and polyvinyl alcohol as examples of
water-soluble hydrocolloid colloid stabilizers which readily undergo graft
copolymerization in emulsion polymerization reactions, although they provide
no evidence. It appears to be generally recognized that vinyl acetate polymers
and copolymers produced by emulsion polymerization may contain a small
amount of gel as a result of grafting to the hydrocolloid. Bondy and Coleman
have expressed the view that graft-copolymerized hydrocolloids are likely to
be anchored firmly at the particle surface, and thus to have become an integral
part of the dispersed polymer phase. This certainly seems likely. If this is so,
then it is to be expected that there are consequences for films cast from the
latices. In particular, the presence of graft-copolymerized hydrocolloids at
the particle surface is expected to enhance the susceptibility of the latex films
to deterioration by moisture. .
11.5.3 Plasticization of vinyl acetate polymers by copolymerization

Much of the technology of vinyl acetate latex copolymers in recent years has
been concerned with internal plasticization of the polymer by copolymerization
in such a way as to produce, at the lowest cost, polymers which perform
satisfactorily in the applications for which they are intended. It is therefore
necessary to consider further the matter of the internal plasticization of vinyl
acetate copolymers by copolymerization with comonomers, and the
consequences of such plasticization for the mechanical properties of the
copolymers.
It has been claimed that ethylene plasticizes vinyl acetate-ethylene copolymers
by a different mechanism to that by which other common comonomers effect
plasticization. Thus Reynolds [26] has stated that 'Ethylene copolymerised
with vinyl acetate, under normal circumstances, enters the polymer backbone
only and provides neither bulky nor long-chain side groups such as are
characteristic of all other usual plasticising comonomers. Its primary contribution
to the plasticisation is by increasing intra-chain mobility, that is, making the
polymer backbone more flexible.' Reynolds further states that a consequence
of internal plasticisation by ethylene is that ' ... there is less inter-chain
adhesion leading, at equal degrees of plasticisation and polymerisation, to
lower cohesive and tensile strengths than in the case of conventional
comonomers.' A similar view has been expressed elsewhere by Reynolds [27J.
Referring first to the internal plasticization of vinyl acetate copolymers by
conventional comonomers, he says: ' ... the mechanism of internal plasticisation
occurs principally by the bulk of the side group separating the polymer chains
from one another in much the same way as occurs when an external plasticiser
is used. The efficiency of plasticisation is determined by the bulk or length
or branching ofthe side groups provided.' Referring to the internal plasticization

of vinyl acetate copolymers by copolymerized ethylene, he says ' ... Here there
are no side chain constituents, so that plasticisation cannot depend upon the
bulk or length or flexibility of side chains. Instead, it depends upon the greater
degree of flexibility conferred upon the backbone itself by the omission of
the side groups of vinyl acetate which impair free rotation about the
carbon -carbon back bone. In this case, the mechanism of internal plasticisation
is intramolecular, in contrast with systems which are much more familiar
where it is intermolecular.' Thus, according to these views, conventional
comonomers internally plasticize vinyl acetate copolymers not by increasing
the inherent flexibility of the polymer chain but by creating more space in
which those chains can move; conversely, ethylene internally plasticizes vinyl
acetate copolymers by increasing the inherent flexibility of the polymer chain
instead of by creating more space in which those chains can move.
The difference in plasticization mechanism has several consequences for
the properties of films from vinyl acetate-ethylene copolymer latices. The
more important ofthese are that both the minimum film-forming temperature
of these latices and the mechanical properties of films from them are very
sensitive to the ethylene content of the copolymer.
The first of these consequences is illustrated by the results shown in Fig.
11.8. This diagram, which is based upon results published by Edser and
25
()
-.. 20
0
l!!
:::l
~
CD
a. 15
E
~
Cl
c:
·E 10
~
;;::: 5
E
:::l
E
·2
0
~
-50 10 20 30
Comonomer content of copolymer / % m/m
Figure 11.8 Effect of copolymer composition upon minimum film-forming temperature of

latices which contain vinyl acetate-ethylene copolymers (points.). and of latices which contain
vinyl acetate-vinyl neodecanoate copolymers (points .) (Edser and Bulezuik [28] and Farmer
and Edser [29]). Point. is for vinyl acetate homopolymer.
Bulezuik [28] and by Farmer and Edser [29], compares the effect of
copolymer composition upon the minimum film-forming temperature of
latices which contain vinyl acetate-ethylene copolymers with that for the
latices which contain vinyl acetate-vinyl neodecanoate copolymers. The
copolymer compositions are expressed as % m/m of the second monomer.
These results demonstrate the sensitivity of the minimum film-forming
temperature of vinyl acetate-ethylene copolymer latices to the ethylene
content of the polymer. They also show that the minimum film-forming
temperatures oflatices of vinyl acetate-ethylene copolymers are considerably
lower than those of latices of vinyl acetate-vinyl neodecanoate copolymers
of similar vinyl acetate content. The minimum film-forming temperature of
a polymer latex is a measure of the average mobility of the segments which
make up the polymer molecule (see section 4.3.6 of Chapter 4 (Volume 1)).
The lower is the minimum film-forming temperature, the higher is the average
mobility of the polymer segments. Thus the results shown in Fig. 11.8 provide
evidence that the average mobility of the segments in a vinyl acetate-ethylene
copolymer molecule is very sensitive to the ethylene content of the copolymer,
and that the average mobility of the segments in a vinyl acetate-ethylene
copolymer is considerably greater than that of the segments in a vinyl
acetate- vinyl neodecanoate copolymer molecule of similar mass composition.
The efficiency of copolymerized ethylene as a plasticizer for vinyl acetate
copolymers, as indicated by reduction of the minimum film-forming temperature
of the latices, is also evident from a diagram published by Vinyl Products
[30], reproduced here as Fig. 11.9. Judged on a mass basis, ethylene is shown
as being far more effective in reducing the minimum film-forming temperature
of vinyl acetate copolymers than are either 2-ethylhexyl acrylate or vinyl
neodecanoate.
The sensitivity of the mechanical properties of films from vinyl acetate--
ethylene copolymer latices to ethylene content is illustrated by results reported
by Reynolds [27], shown here in Table 11.10. Increasing the ethylene content
ofthe copolymer markedly reduces both tensile strength and rigidity modulus.
It also increases the extension at break. These effects are attributed to
reduction of inter-chain bonding within the polymer structure. The effect of
ethylene content upon the tensile properties of vinyl acetate-ethylene
copolymers is also evident from results reported by Lee and Edser [31],
shown here as Fig. 11. to. This shows typical tensile stress-strain curves for
vinyl acetate-ethylene copolymers of various ethylene contents. The same
trends are apparent as are evident from Table 11.10.
Related to the matter ofthe internal plasticization of vinyl acetate polymers
by copolymerization with softening comonomers is the converse matter of
the stiffening of vinyl acetate copolymers by copolymerization with hardening
comonomers, to which reference has been made in section 11.5.1 above. The
effect of copolymer composition upon the minimum film-forming temperature
of vinyl acetate-vinyl chloride copolymer latices is illustrated by results
25
E
:::l
E
·2 5
~
°o~------~~-----L------~------~
10 20 30 40
Comonomer content of copolymer / % mlm
Figure 11.9 Efficiency of copolymerized ethylene (curve A), copolymerized 2-ethylhexyl acrylate
(curve B) and copolymerized vinyl neodecanoate (curve q as plasticizers for vinyl aoetate
copolymers, as indicated by reduction of the minimum film-forming temperature of copolymer
latioes (Vinyl Products [30]).
reported by Edser and Bulezuik [28], reproduced here as Fig. 11.11. Over
the composition range 0-90% mlm, the minimum film-forming temperature
increases linearly with composition, expressed as % mlm of vinyl chloride,
from ca. 20·C to ca. 80·C. Although of no interest in respect of copolymers
for application as adhesives and binders, the minimum film-forming temperature
then rises more sharply as the vinyl chloride content increases above
90% m/m; this effect is attributed to the development of crystallinity in the
vinyl chloride regions of the copolymer. Vinyl acetate-ethylene-vinyl
chloride copolymers in latex form are currently of considerable interest as
binders for latex-based surface coatings. Further reference to this matter is
made in section 20.2.2. t of Chapter 20 (Volume 3); Figure 20.2 of that chapter
indicates the composition limits for copolymers suitable for this application.
11.5.4 Hydrolysis of vinyl acetate polymers and copolymers in latex form

Amongst the common polymers which are available as synthetic latices, vinyl
acetate polymers and copolymers share with polychloroprenes a marked
tendency to undergo hydrolysis when in contact with aqueous media. Acrylic-
and methacrylic-ester polymers can also undergo hydrolysis, but the tendency
is less marked than for vinyl acetate polymers and copolymers. This is because
the direct attachment of the carbonyl group to the polymer chain in the case
10
500
Extension / %
Figure IUO Typical tensile stress-strain curves for vinyl acetate-ethylene copolymers
containing 13.5% mlm (points e). 18.3% mlm (points a). 19.0"/. mlm (points .) and 22.8% mlm
(points .) ethylene (Lee and Edscr [31]).
Table 11.10 Effect of ethylene content upon mechanical properties of vinyl

acetate-ethylene copolymers (Reynolds [27])
Extension at
Ethylene content Tensile strength break Rigidity modulus
1% m/m I MPa 1% I MPa
4.0 7.91 210 49.0
10.0 3.99 340 12.7
12.5 3.60 415 6.9
17.5 0.55 610 5.4
20.0 0.48 1220 2.0
of polyacrylates and polymethacrylates renders their ester groups less

susceptible to hydrolysis than are the ester groups in vinyl acetate polymers,
the carbonyl groups of which are one atom removed from the polymer chain.
In the case of the polymethacrylates, further steric protection is provided by
the a-methyl substituent of the main chain.
The possibility that vinyl acetate polymers and copolymers can be
hydrolysed has been known for many years. The principal reaction is the
100
,
,,
--
V
~
::J
T!!Q)
a.
E
.2!
OJ
c
'§
.2
~
E
::J
E
'2
~
00 20 40 60 80 100
Vinyl chloride content of copolymer / % m/m
Figure 11.11 Effect of copolymer composition upon minimum film-forming temperature of

vinyl acetate-vinyl chloride copolymer latices (Edser and Bulezuik [28]).
splitting off of the acetate moiety of the units derived from vinyl acetate,
leaving, in effect, a unit derived from vinyl alcohol in the polymer chain. The
acetate moiety enters the aqueous phase. The hydrolysis of vinyl acetate units
is subject to catalysis by both acids and alkalis. More interest has been shown
in alkali-catalysed hydrolysis than in acid-catalysed hydrolysis, principally
because these polymers are more likely to encounter alkaline conditions in
service (through direct or indirect contact with plaster, cement, etc.) than
acidic conditions. For hydrolysis under alkaline conditions, the eventual fate
of the acetate moiety is an acetate anion dissolved in the aqueous phase. The
overall hydrolysis reaction is then:
... -CH2-CH- ... + HO- - - ... -CH2-CH- ... + CH 3 CO;

I I
o OH
I
CO. CH 3
The tendency for vinyl acetate polymers and copolymers to hydrolyse is

especially marked if the polymers are in latex form, because the particles are
immersed in an aqueous medium, and the specific surface area of the polymer
is high. Thus chemical interaction between the polymer and the dispersion
medium is facilitated because the polymer has ready access to water molecules
and catalyst ions. It is possible that the entry of ions into the particles may
be hindered by the presence of bound electric charges of the same polarity
at the particle surface. However, if the surface-bound charge is of opposite

polarity to that of the catalysing ions, then the hydrolysis reaction may be
facilitated by the presence of that charge. Like the hydrolysis of polychloroprenes
in latex form, that of vinyl acetate polymers and copolymers in latex form
requires several separate steps. The reagents must diffuse into the particles,
the reaction must occur within the particles, and water-soluble products then
partition between the particles and the aqueous phase, presumably mainly
in favour of the latter.
Although little seems to be known at present concerning the mechanistic
details of the hydrolysis of vinyl acetate polymers and copolymers in latex
form, there is considerable empirical information available relating to the
overall kinetics of the reaction. Davies and Reynolds [32] have reported an
important investigation of the alkaline hydrolysis of a vinyl acetate homopolymer
and of various vinyl acetate copolymers in the form of anionic latices. The
latices were specially prepared for the investigation. A few acrylic-ester
copolymers which did not contain vinyl acetate were also investigated; they
were found to hydrolyse more slowly than the vinyl acetate polymers and
copolymers. The latices were prepared by conventional emulsion polymerization
using potassium peroxodisulphate as initiator, and a mixture of an anionic
surface-active substance (0.5 pphm), a non-ionogenic surface-active substance
(1.0 pphm) and hydroxyethylcellulose (2.5 pphm) as colloid stabilizers. The
experimental procedure used by Davies and Reynolds to effect and follow
polymer hydrolysis was as follows: To ca. 3.5 g of the latex (ca. 57% m/m
solids content) was added 25 cm 3 of distilled water. After neutralizing to the
first pink colour of phenolphthalein with 0.1 mol dm - 3 aqueous sodium
hydroxide solution, 50 cm 3 of I mol dm - 3 aqueous sodium hydroxide solution
was added with thorough mixing. The mixture of diluted latex and alkali
was allowed to stand at 25°C without agitation; it was under these conditions
that hydrolysis of the polymer occurred. After the desired time had elapsed,
the unreacted alkali was determined by titration with I mol dm - 3 aqueous
hydrochloric acid solution. The degree of hydrolysis was expressed as a
percentage of the maximum hydrolysis theoretically possible. It is unclear
whether, for vinyl acetate copolymers which contained hydrolysable monomer
units other than vinyl acetate, 100% hydrolysis refers to complete hydrolysis
of the copolymer or merely to complete hydrolysis of the vinyl acetate units.
The latter is probably the case; however, the detailed interpretation of this
aspect of the results does not affect the principal conclusions from this
investigation. That complete alkaline hydrolysis of the vinyl acetate polymers
and copolymers was possible was confirmed by refluxing the polymer with
alcoholic potassium hydroxide solution. The following features of the
hydrolysis conditions used in this investigation should be noted:
1. The temperature at which hydrolysis occurred was 25°C, i.e., only slightly
above normal ambient temperature.
2. The concentration of added sodium hydroxide was ca. 0.65 mol dm - 3 on

the aqueous phase, corresponding to pH ca. 13.8 at 25°C.
3. The amount of sodium hydroxide was ca. twice that required to hydrolyse
the polymer completely if it comprised vinyl acetate units only.
4. The latex particles almost certainly carried surface-bound negative electric
charges. Thus entry ofthe reagent responsible for catalysing the hydrolysis
reaction, namely, hydroxide anions, into the particles may have been
retarded somewhat. Thus polymer hydrolysis may have occurred more
slowly than it would have done in cationic or uncharged latex particles
of similar size and composition.
5. The maximum time of reaction was 360 hours, i.e., 15 days.
Figure 11.12(a) shows the course of the hydrolysis of a vinyl acetate

homopolymer and various vinyl acetate copolymers in latex form. Figure
11.12(b) shows the effect of copolymer composition, expressed as % mlm of
copolymer, upon extent of hydrolysis of various vinyl acetate copolymers
after 24 hours reaction. It is evident that the vinyl acetate homo polymers
and copolymers in latex form underwent significant hydrolysis under the
conditions used by Davies and Reynolds [32]. For polyvinyl acetate,
hydrolysis was almost complete after 2 days. Copolymerization with less
hydrolysable monomers could reduce the rate of hydrolysis markedly, the
effect depending upon the nature, as well as upon the amount, of the
comonomer. The curves for the variation of degree of hydrolysis with time
of reaction have the expected shape. The initial rate of hydrolysis is relatively
high, but diminishes progressively as hydrolysis proceeds. The principal
reasons for the progressive reduction in rate of hydrolysis are presumably
the consumption of alkali and the disappearance of hydrolysable groups as
the reaction proceeds. Although no firm conclusions can be drawn because
no detailed kinetic analyses of these results appear to have been carried out,
there are no indications of any tendency for the intrinsic rate of hydrolysis
to increase with increasing extent of hydrolysis. It might be expected that
this might happen, partly because the polymer becomes more hydrophilic as
hydrolysis progresses, and partly because the acetate moieties are replaced
by the smaller hydroxyl groups, thereby making adjacent unhydrolysed vinyl
acetate units more accessible to attack by the hydrolysing reagents. The
results shown in Fig. 11.12(b) demonstrate that the effect of the comonomer
is not merely to dilute the vinyl acetate units in the polymer available for
hydrolysis. The extent of hydrolysis after 24 hours reaction does not fall
linearly with copolymer composition; there is an initial rapid fall in extent
of hydrolysis as comonomer units are introduced into the polymer, the extent
of hyrolysis then declining more slowly as higher proportions of comonomer
units are introduced.
There are probably at least two reasons why comonomers are more effective
in reducing the extent of hydrolysis after a given period of reaction than
100
80
~
0
.......
.!!l
tJ)
~ 60
e
"C
>-
.r:
'0 40
C
~
20
00 100 200 300 400

(a)
Time/hour
100
--
~
0
.!!l
~ 60
e
"C
~
'0 40
C
~
20
00 20 40 60 80 100
Comonomer content of copolymer / % mlm
(b)
Figure 11.11 (a) Course of hydrolysis of a vinyl acetate homopolymer and various vinyl acetate
copolymers in latex form (Davies and Reynolds [32]). Significance of points: • vinyl acetate
homopolymer;.to 8O;20mjrn vinyl acetate!vinyl caprylate (n-octanoate)-vinyl caprate (n-decanoate)
copolymer; _ 8O;20mjrn vinyl acetatejvinyl neodecanoate; 11" 8O;20mjrn vinyl acetate!
n-butyl acrylate copolymer. (b) Effect of copolymer composition, expressed as % mjrn of
copolymer, upon extent of hydrolysis of various vinyl acetate copolymers after 24 hours reaction
(Davies and Reynolds [32]). Significance of points: • vinyl acetate homopolymer; .to vinyl
acetate-vinyl propionate copolymers; _ vinyl acetate-vinyl neodecanoate copolymers; 11" vinyl
acetate-n- butyl acrylate copolymers; • vinyl acetate- 2-ethylhexyl acrylate copolymers.
would be expected on the basis of mere dilution of the vinyl acetate units.
Firstly, all the comonomer units introduced by Davies and Reynolds [32]
are more hydrophobic than are units derived from vinyl acetate. Thus it is
to be expected that both the equilibrium water content of the particles and
the ease of entry of ions into the particles were reduced. Secondly, the presence
of non-hydrolysable units adjacent to hydrolysable vinyl acetate units may
have a protecting influence upon the latter, partly for steric reasons and
partly because local domains of relative hydrophobicity are created adjacent
to the hydrolysable units. In this connection, it should be noted that the
apparent comparative abilities of given types of comonomer unit to retard
hydrolysis can depend upon the way in which the respective copolymer
compositions are expressed. This is illustrated by the results for latices
containing, on the one hand, vinyl acetate-ethylene copolymers, and, on the
other, vinyl acetate-vinyl neodecanoate copolymers. Because the molecular
mass of vinyl neodecanoate is much greater than that of ethylene, vinyl
acetate-ethylene copolymers having similar compositions expressed in mass
units have very different compositions when expressed in molar units, and
vice versa. Consequently, conclusions drawn concerning the relative effectiveness
of the two types of comonomer unit in retarding polymer hydrolysis depend
upon the way in which copolymer composition is expressed. Figure 11.13(a)
shows the variation of extent of hydrolysis after 24 hours reaction with
copolymer compositions expressed as % m/m; Figure 11.13(b) shows the
same results with copolymer compositions expressed as % mol/mol. When
copolymer compositions are expressed on a mass basis, the effects of the two
100 100
--
.!II
~
III
60
E!
~
'0
'E
~ 20
(b)
00 20 40 60 80 100 00 20 40 60 80 100
Comonomer content of copolymer 1% mlm Comonomer content of copolymer 1% moVmol
(a) (b)
Figure 11.13 (a) Variation of extent of hydrolysis after 24 hours reaction with copolymer
compositions for vinyl acetate-ethylene copolymers (points .) and vinyl aretate-vinyl neodecanoate
copolymers (points A), copolymer compositions being expressed as % mfm of comonomer (Davies
and Reynolds [32]). (b) Same results with copolymer compositions expressed as % molfmol of
comonomer (Davies and Reynolds [32]).
types of comonomer unit in retarding hydrolysis under alkaline conditions

are effectively identical; both sets of points in Fig. 11.13(a) can be adequately
represented by a single curve. This indicates that, for this pair of vinyl acetate
comonomers, retardation of hydrolysis is determined primarily by the total
concentration of alkyl carbon atoms present as hydrophobic units, rather
than by the way in which those alkyl carbon atoms are distributed in the
copolymer. However, Fig. t t.t3(b) shows that, when the copolymer compositions
are expressed on a molar basis, the effects of the two types of comonomer
unit in retarding hydrolysis under alkaline conditions are very different. The
results presented in this way provide strong evidence that, on a molecular
basis, vinyl neodecanoate units are much more effective in retarding alkaline
hydrolysis than are ethylene units, and so presumably have a much greater
protective effect, presumably for steric reasons.
Davies and Reynolds [32] have also provided evidence that the incorporation
of small concentrations of copolymerized carboxylic-acid units in a vinyl
acetate copolymer sensitizes the polymer to alkaline hydrolysis. This effect
is illustrated by the results for 30/70 mlm vinyl acetate-vinyl neodecanoate
copolymer latices shown in Table t t.t 1. Although the particle sizes of the
two latices were the same, the inclusion of t pphm of acrylic acid in the
copolymer increased the extent of hydrolysis after 24 hours by a factor of
approximately 3. The probable reasons for this effect are that
t. the equilibrium water content of the particles was increased, thereby
increasing the ease of entry of ions into the particles;
2. such steric protection as was hitherto afforded to a minority of vinyl
acetate units by adjacent acetate moieties has been reduced by replacement
with smaller units which also create local domains of relative hydrophilicity
adjacent to the hydrolysable units.
The conclusions drawn from the investigation reported by Davies and
Reynolds [32] have been corroborated and extended by further results more
Table Il.ll Effect of incorporation of minor amount of copolymerized acrylic acid

upon extent of alkaline hydrolysis of a vinyl acetate-vinyl neodecanoate copolymer
in latex form (Davies and Reynolds [32])
Copolymer composition:
Vinyl acetate/vinyl neodecanoate/acrylic acid
/ m/m/m
70/30/0 69/30/1
Average diameter of latex particles

/ nm 170 170
Extent of alkaline hydrolysis
after 24 hours / % 15.1 48.4
Acrylic latices 353
recently reported by Martin, Smith and Bassett [33]. The same technique
for effecting and following hydrolysis was used as that described by Davies
and Reynolds. The conclusion is drawn that the ability of comonomers to
reduce the rate of alkaline hydrolysis of vinyl acetate polymers is determined
by several factors, the more important of which are:
1. the intrinsic resistance of the comonomer units to alkaline hydrolysis;
2. interruption of long sequences of hydrolysable vinyl acetate units by the
comonomer units;
3. steric protection of hydrolysable units from attack by hydroxide ions.
The resistance of the comonomer units to alkaline hydrolysis is enhanced
by structural features such as the presence of a-branched alkyl groups in the
case of vinyl esters.
11.6 LATICES OF POLYACRYLATE AND POLYMETHACRYLATE

ESTERS: ACRYLIC LATICES
The term acrylic latices is used somewhat loosely to delineate as a group
those latices in which are dispersed polymers which contain a substantial
proportion of monomers derived from esters of acrylic acid or methacrylic
acid. The general structures of these monomers are shown as XXII and XXIII
CH 2=CH CH 2 =C.CH 3
I I
CO.OR CO.OR
XXII XXIII
respectively, in which R is usually an alkyl group or a derivative of an alkyl
group. There are various types of latex which are not usually included in the
class of acrylic latices, although perhaps it might be thought that they would.
Thus latices of polymers which contain only minor proportions of units
derived from esters of acrylic and methacrylic acids are not usually included;
neither are latices of hydrophilic polymers which contain mainly units derived
from hydrophilic esters of acrylic and methacrylic acids. Also excluded are
the latices of polymers such as styrene-butadiene copolymers and acrylonitrile-
butadiene copolymers which have been carboxylated by copolymerization
with minor amounts of acrylic or methacrylic acids. Varying degrees of
interest have been shown in latices of copolymers of acrylate or methacrylate
esters and various non-acrylic monomers, such as butadiene, styrene and
acrylonitrile, the units derived from the acrylic monomers forming a
substantial proportion of the polymer. These latices tend to be regarded as
being acrylic latices. Acrylonitrile itself is sometimes regarded as being an
acrylic monomer, being, of course, the nitrile of acrylic acid.
Each polymerized acrylate- or methacrylate-ester unit introduces a chiral
carbon atom into the polymer sequence. Thus the structural variations
associated with the presence of such carbon atoms are possible, as are those
arising from variations in the relative dispositions of pairs of successive chiral
carbon atoms in the polymer sequence. Further structural complications can
arise in principle, because it is possible for the ester groups in the polymer
contained in acrylic latices to become hydrolysed through contact with the
aqueous dispersion medium, analogously to the hydrolysis of vinyl acetate
units in vinyl acetate polymers and copolymers (see section 11.5.4 above). In
practice, the hydrolysis of polymers of acrylic and methacrylic esters in latex
form does not occur so readily as does that of vinyl acetate polymers and
copolymers, for the reason given in section 11.5.4. However, some slight
hydrolysis of the polymers will tend to occur, especially if the ester units are
derived from acrylates rather than methacrylates. Unlike vinyl acetate
polymers, the carboxylic-acid groups generated by hydrolysis remain attached
to the main chain of the polymer. They therefore tend to enhance the colloid
stability of an anionic and alkaline acrylic latex. Some hydrolysis of acrylate
and methacrylate ester monomers may occur during the production oflatices
by emulsion polymerization, especially if the aqueous phase is alkaline.
Acrylic latices are used in large quantities for various applications, which
include as binders for latex-based surface coatings of many types, and as
binders for textile fibres. A large number of esters of acrylic and methacrylic
acids is available for the preparation of acrylic latices, many of them on an
industrial scale. The most important single factor which decides the monomer
mixture from which an acrylic-ester polymer is to be made is the requisite
glass-transition temperature for the polymer. This determines the minimum
film-forming temperature of the latex, provided that film-formation has not
been facilitated by the additioQ of external plasticizers. The glass-transition
temperature also determines the physical character of the polymer, the
mechanical properties of the polymer, and the manner in which the polymer
character and mechanical properties are affected by changes in temperature.
The glass-transition temperatures of homo polymers of acrylate and methacrylate
esters vary widely, depending upon the nature of the moiety which was used
to form the ester. Published values for the glass-transition temperatures of
these homo polymers vary somewhat. Table 11.12 gives reported glass-
transition temperatures for various atactic poly-n-alkyl acrylates and poly-
n-alkyl methacrylates. As the alkyl chain length of a poly-n-alkyl acrylate
increases to ca. octyl, the glass-transition temperature decreases progressively.
This is attributed to the alkyl group facilitating segment motion ofthe polymer
chain by increasing the free volume of the polymer. As the alkyl chain length
increases beyond octyl, the glass-transition temperature increases, this effect
being attributed to increasing intermolecular forces developing between the
alkyl groups. For poly-n-alkyl methacrylates having alkyl groups up to ca.
octyl, the glass-transition temperature is considerably higher than that of the
corresponding poly-n-alkyl acrylate. This effect is attributed to stiffening of
the main chain of the polymer by the (X-methyl substituent of the poly-n-alkyl
Acrylic latices 355
Table 11.12 Glass-transition temperatures of various atactic poly-n-alkyl acrylates

and poly-n-alkyl methacrylates (Peyser [34])
Glass-transition temperature / ·c of
polyacrylate polymethacrylate
n-alkyl group (R) ...-CH2CfCH3)-'"
Name Structure C0 2 R
Methyl - CH 3 10 105
Ethyl -CH 2CH 3 -24 65
n-propyl -(CH 2hCH 3 -37 35
n-butyl -(CH 2hCH 3 -54 20
n-hexyl -(CH 2lsCH 3 -57 -5
n-octyl -(CH 2),CH 3 -65 -20
n-dodecyl -(CH 2)IICH 3 -3 -65
n-hexadecyl -(CH 2)ISCH3 35 15
methacrylate. The effect of the (X-methyl substituent becomes less as the alkyl
chain length increases. As the alkyl chain length increases beyond octyl, the
glass-transition temperature ofthe poly-n-alkyl methacrylates falls below that
of the poly-n-alkyl acrylates. The effect of the isomeric structure of the alkyl
group upon the glass-transition temperature of various atactic poly-n-alkyl
acrylates is illustrated in Table 11.13. Values are given for a selection of
isomeric polyalkyl acrylates up to octyl. They are grouped according to the
number of carbon atoms in the alkyl group. As a generalization, increasing
branching in the alkyl group is accompanied by an increase in the glass-
transition temperature of the poly-n-alkyl acrylate. This effect is as expected,
on the basis that the ability of an alkyl group to reduce the glass-transition
temperature of a polyalkyl acrylate correlates with its effect upon the free
volume of the polymer: the more branched and compact is an alkyl group
of a given size, the less free volume will it generate in the polymer, and so
the less will it reduce the glass-transition temperature.
The values given in Tables 11.12 and 11.13 are for homopolymers of various
acrylate and methacrylate esters. For binary copolymers of these monomers,
the glass-transition temperatures are intermediate between those of the two
homopolymers, the value being approximately that expected from the
composition. The same is broadly true for copolymers which contain units
derived from more than two monomers. Thus it is possible to prepare acrylic
latices containing polymers which are chemically similar, but which have a
very wide range of glass-transition temperatures. In this respect, the acrylic
family of polymers and copolymers is unique.
Acrylic latices are produced by conventional emulsion polymerization.
Mast, Smith and Fisher [35] and Mast and Fisher [36] have reported on
the technical aspects of the emulsion polymerization of acrylic esters. The
colloid stabilizers used are said to have a profound effect upon the properties
Table 11.13 Effect of isomeric structure of alkyl group upon glass-transition

temperature of various atactic polyalkyl acrylates, grouped according to the number
of carbon atoms, n, in the alkyl moiety (Peyser [34])
Glass-transition
temperature / ·c
of polyacrylate
Alkyl group (R) ...-CH 2CH-...
I
Name Structure C0 2 R
n=3
n-propyl -37
I-methylethyl (isopropyl) ca.-5
n=4
n-butyl -(CH 2hCH 3 -54
I-methylpropyl (sec-butyl) -CH(CH 3)CH 2CH 3 -22
2-methylpropyl (isobutyl) -CH 2CH(CH 3h -24
1,1'-dimethylethyl (tert-butyl) -C(CH 3h 43-107 (1)
n=5
2-methylbutyl -CH 2CH(CH 3)CH 2CH 3 -32
3-methylbutyl -CH 2CH 2CH(CH 3h -45
2,2' -dimethylpropyl
(neopentyl) -CH 2C(CH 3h 22
l-ethylpropyl -CH(C 2H sh -6
n=6
n-hexyl -(CH 2)sCH 3 -57
2-methylpentyl -CH2CH(CH3)CH2CH2CH3 -38
1,3-dimethylbutyl -CH(CH 3)CH 2CH(CH3h -15
2-ethylbutyl "":"'CH2CH(C2Hs)CH2CH3 -50
n=8
n-octyl -(CH 2),CH 3 -65
2-ethylhexyl -CH2CH(C2Hs)(CH2hCH3 -50
I-methylheptyl -CH(CH 3)(CH 2hCH 3 -45
of the latex produced. It is desirable that the pH of the aqueous phase of

the reaction system be less than 7, in order to minimize hydrolysis of the
monomers. Higher reaction temperatures are said to be required than for
the emulsion polymerization of diene monomers. Other papers describing
the emulsion polymerization of acrylic monomers include those by Zimmt
[37], by Brodnyan et al. [38], and by McCurdy and Laidler [39,40]. Fitch,
Prenosil and Sprick [41] have investigated the mechanism of particle
nucleation during the emulsion polymerization of methyl methacrylate.
Functionalized acrylic latices can readily be prepared by emulsion
copolymerization with appropriate comonomers.
A procedure for the preparation of latices containing copolymers of
acrylonitrile and acrylate esters has been described by Leonard, Cort and
Latices of vinyl chloride-vinylidene chloride copolymers 357
Blevins [42]. An aqueous solution of potassium chloride and a sodium

alkylbenzenesulphonate as a colloid stabilizer is purged at O·C with nitrogen
for 30 minutes. The monomers are then added. The initiator is a redox system
comprising potassium peroxodisulphate and sodium thiosulphate. This is
introduced ca. 2 hours later. The polymerization reaction commences after
an induction period of 10-30 minutes. The reaction is allowed to proceed
for 24 hours at O·e. Unreacted monomers are then removed by the application
of a partial vacuum, followed by a period of heating at 65·e. After this
treatment, the latex should be essentially odour-free, and should have a total
solids content of 52-55% m/m.
Some interest has been shown in the latices of elastomeric copolymers of
acrylonitrile and ethyl acrylate, of acrylonitrile and n-butyl acrylate, and of
acrylonitrile, ethyl acrylate and n-butyl acrylate, for the fabrication of gloves
which simulate the human skin in colour, texture and detail. Gloves of this
type have been developed to be worn by amputees over mechanical hands;
the intention is that the combination of mechanical hand and glove should
restore as fully as possible the amputee's normal prehensile function and
appearance. One of the major problems in developing gloves for this
application has been that of finding materials which meet the stringent criteria
of being sufficiently inert, resistant to staining and optically translucent, whilst
having the requisite mechanical properties. Elastomers derived from acrylate
esters have many of the desired properties, and have also been found to be
suitable for the low-pressure casting processes by which such gloves can be
made, provided that the total solids content is sufficiently high. However,
acrylate-ester latex elastomers have suffered from inadequate strength for
this application. An objective has therefore been to improve the strength of
films from these latices, either by the inclusion of stiffening comonomers in
the polymer, or by the use of reinforcing fillers, without detracting from
desirable properties such as translucency. In relation to this application, the
technology of films derived from latices of copolymers of acrylonitrile and
acrylate esters has been described by Leonard, Cort and Blevins [42], by
Fram et al. [43], and by Fram, Szlachtun and Leonard [44]. The mechanical
properties of the cast films were found to improve as the acrylonitrile content
of the copolymer increased, but, as expected, the tendency to stiffening at
low temperatures is also increased. The conclusion drawn from this investigation
was that a 1O/90m/m acrylonitrile-n-butyl acrylate copolymer provides a
suitable compromise for this application.
11.7 LATICES OF VINYL CHLORIDE-VINYLIDENE CHLORIDE

COPOLYMERS
The structures of vinyl chloride and vinylidene chloride (1,1-dichloro-
ethylene) are shown at XXI and XXIV respectively. Incorporation of a vinyl
chloride unit into a polymer sequence introduces a chiral carbon atom, but
introduction of a vinylidene chloride unit does not. The glass-transition
temperatures of polyvinyl chloride and polyvinylidene chloride are ca. SO°C
and ca. -ISoC respectively. The glass-transition temperatures of copolymers
of the two monomers are intermediate between these values. Thus either
vinyl chloride may be regarded as a stiffening comonomer for polyvinylidene
chloride, or, alternatively, vinylidene chloride may be regarded as a plasticizing
comonomer for polyvinyl chloride. Latices which contain only vinyl chloride
homopolymers dry down at normal ambient temperatures to powdery
deposits which lack coherence. Film integration at these temperatures
becomes possible only if sufficient of a suitable external plasticizer, e.g.
di-2-ethylhexyl phthalate, is first added to the latex.
CH 2 =CCI 2
XXIV
The emulsion polymerization (and also the bulk polymerization) of both

vinyl chloride and vinylidene chloride is complicated by the limited solubility
of the polymer in the monomer. Thus the polymer tends to precipitate within
the reaction loci as polymerization proceeds. Like polychloroprene latices,
the pH of the latices containing polymers of vinyl chloride and vinylidene
chloride tends to fall slowly on ageing. Again this is a consequence of slow
hydrolysis of the polymer, with concomitant liberation of chloride ions into
the latex aqueous phase. The pH of the latices should be adjusted occasionally
by addition of an alkali such as ammonia. The latices are best stored in glass,
enamelled, aluminium or stainless-steel vessels. They should not be stored
in contact with mild steel, copper or zinc.
I1.S FUNCTIONALIZED SYNTHETIC LATICES
II.S.1 General considerations

The preparation offunctionalized synthetic latices by emulsion copolymerization
has been considered in section 10.3 of Chapter 10. Functionalized variants
of many types of synthetic latices, of both rubbery and glassy polymers, have
been, and are currently, produced industrially. The purpose of this section
is to deal with function ali zed synthetic latices as a class, regardless of the
nature of the polymer type dispersed in the latex. The emphasis will be upon
the subgroup of functionalized synthetic latices which comprises the most
industrially-important class, namely, carboxylated synthetic latices. A very
important type within this class comprises the latices of carboxylated
styrene-butadiene copolymers, especially carboxylated rubbery styrene-
butadiene copolymers. Notable amongst the other types of synthetic latices
produced as functionalized variants are acrylonitrile-butadiene copolymer
latices, vinyl acetate copolymer latices, and various acrylic latices. Many of
Functionalized synthetic latices 359
the comments which follow concerning the effects of functionalization upon

the properties of the latices and of the dispersed polymers are made with
specific reference to styrene-butadiene copolymer latices. Nevertheless, these
comments usually apply generally to the other types of functionalized
synthetic latex. In considering these effects, it should be noted that the
proportion of polymer units derived from the copolymerized functionalized
comonomer is usually small. This proportion (not that of the functionalizing
group itself) is frequently in the range 0.5-5 pphp, but larger amounts can
be incorporated. As noted in section 10.3.2 of Chapter 10, the levels of
functionalization commonly introduced into latices of industrial interest are
not usually such that the glass-transition temperature of the dispersed
polymer is much affected. .
The rationale and principles which underlie the functionalization of
synthetic latices are discussed in the appropriate sub-sections of section 10.3
of Chapter 10, mainly with reference to carboxylated latices. The important
matter of the distribution of functional groups in functionalized latices is
considered in section 10.3.4. The technological advantages which accrue from
functionalization are discussed in section 10.3.2. It is noted there that one of
the important technological advantages which accrue to the dispersed
polymer itself, as opposed to the latex as a colloidal system, is that the
polymer becomes susceptible to additional chemical reactions. This matter
has now to be considered in further detail.
II.S.2 Additional chemical reactions to which functionalized latex polymers

are susceptible
I1.B.2.1 Introduction
The main reason for interest in the additional reactions made possible by
polymer functionalization is that the polymer can become crosslinked by
means of these reactions. It is, of course, necessary that the crosslin king
reagents used should be at least difunctional, and that at least some
intermolecular linkages should form between polymer molecules as a result
of these reactions. However, it is not always easy to obtain experimental
evidence that the expected crosslinks are present in the final polymer, as
distinct from evidence that some change in physical properties has occurred
which might be a consequence of crosslinking. Such evidence as does exist
in respect of crosslinking tends to be rather circumstantial. As a rule,
functionalization in general, and carboxylation in particular, does not
interfere with the ability of polymers to participate in other chemical reactions
to which the non-functionalized variants are susceptible. Thus carboxylated
diene rubbers are susceptible to crosslinking by conventional vulcanization
systems, such as sulphur and organic accelerators, in much the same way as
are their non-carboxylated equivalents.
In exploiting these reactions industrially, the interest has usually been in

effecting thern on the polyrner in bulk forrn after it has been isolated frorn
the latex, e.g., by gelation or by rernoval of the dispersion rnediurn by
evaporation. However, there has also been sorne interest in the possibility
of exploiting sorne of these reactions using the polyrner in the latex forrn. In
this context such latices thernselves are described as being reactive. Attention
in this section is confined to reactions which are norrnally carried out on
functionalized polyrners after isolation frorn the latex. An exarnple of
exploitation of the reactivity of functionalized polyrners in latex forrn is given
in section 18.6 of Chapter 18 (Volurne 3), where the use of reactive latices
and co-reactive water-soluble substances in the rnanufacture of .Iatex-foarn
rubber is described. The principles underlying this exploitation are discussed
in section 5.4.3.3 of Chapter 5 (Volurne 1).
An extensive review of the novel crosslinking chernical reactions to which
functionalized latex polyrners are susceptible is available in six papers by
Bufkin and Grawe [45-50]. These papers contain nurnerous references,
rnainly to the patent literature. The rnain concern in these papers is with
aqueous latices. Bufkin and Grawe propose the following three-fold broad
classification of crosslin king systerns of which polyrners in latex forrn are a
part, according to the rnanner in which the crosslinks are distributed within
the final dried rnaterial:
1. homogeneous systems, in which the crosslinks are distributed uniforrnly

throughout the rnaterial, apart frorn statistical fluctuations;
2. interracial systems, in which rnost of the polyrner crosslinking sites are
located near the surface of the particles, the concentration of crosslinking
sites decreasing towards the centre of the particles;
3. interstitial systems, which cornprise essentially unreacted latex particles
ern bedded in a rnatrix of crosslinked polyrner.
The concern here is with systerns of types (1) and (2). A useful broad
classification of the functionalized latex polyrners thernselves is into those
which are inherently self-crosslink able and those which require the presence
of additional reagents. In what follows, brief consideration is first given to
the rnore irnportant of the reactions which are available for carboxylated
polyrners in latex forrn. Then follows an outline of the reactions available to
polyrners functionalized by the inclusion of N-rnethylolarnide and related
groups. The discussion concludes with reference to reactions rnade possible
by the inclusion of other types of functional groups in latex polyrners.
11.8.2.2 Special reactions to which carboxylated polymers are susceptible

Carboxylation of a polyrner irnparts susceptibility to a nurnber of chernical
reactions not otherwise possible. The crosslinking reagents which are used
are usually polyfunctional with respect to reactions with carboxylic-acid
groups. The reactions can therefore lead to the formation of intermolecular

and intramolecular linkages. The more important of the additional reactions
to which carboxylated polymers are susceptible fall under the following headings:
1. interaction with polyvalent metal ions to form metal-carboxylate linkages;
2. interaction with polyhydric alcohols to form ester linkages;
3. interaction with methylol-containing compounds to give crosslinks of
rather uncertain chemical structure;
4. interaction with epoxides to form ester linkages;
5. interaction with polyamino compounds to form amide linkages;
6. interaction between pairs of carboxylic-acid groups to form acid-anhydride
linkages.
Heating a carboxylated polymer with an oxide or salt of a polyvalent
cation is generally supposed to result in the formation of crosslinks between
the polymer molecules. For an oxide of a divalent cation, and assuming that
the carboxylic-acid groups are un-ionized, the crosslin king reaction is
conventionally represented as follows:
--T-- ~
CO CO
I I
OH 0"-
I
+ MilO ----+ M26+ + H 20
I
OH 0"-
I I
CO CO
____ L__ _
~
However, the evidence that metal-carboxylate crosslinks actually do form
under the conditions usually employed in manufacturing practice is questionable,
partly because metal-carboxylate crosslinks tend to be labile and subject to
rearrangement under the influence of heat and polymer solvents. Even if
crosslinks of this type actually do form, the nature of the bonding between
the metal ions and the carboxylate ions remains obscure; it is probably best
indicated as shown above. The nature of the bond probably depends upon
the particular metal ion, tending to be ionic in some cases and covalent in
others. In most cases, the bond is probably intermediate between the two,
being effectively a polarized covalent bond. The interaction between the metal
ions and the carboxylate ions may also be supplemented by coordinate
covalent bonds formed between the metal ions and the lone pairs of the
second oxygen atom of the carboxylate anion. Zinc is probably the metal
ion which has been most widely used for the modification of carboxylated
latex polymers by formation of metal-carboxylate linkages. Other metal ions
which have been used include those of aluminium, zirconium and chromium.
Ester linkages between the carboxylic-acid-bearing units in carboxylated

polymers can be formed by interaction with polyhydric alcohols. Denoting
the polyhydric alcohol as HO.R.OH, the ester-forming reaction is as follows:
""'" -I -....,..
~
co CO
1 1
OH HO o
1 1
+ R -----+ R
1
o
1
OH HO
1 1
CO CO
~
L ___
Various types of methylol-containing compounds can be used to form

linkages of uncertain chemical structure by reaction with carboxylic-acid
groups. Denoting the methylolic compound as HOCH 2.R.CH 20H, the
reaction can be represented in at least two ways: either as an ester-forming
reaction, or as a ketone-forming reaction in which the methylol groups are
eliminated as formaldehyde. These representations are respectively as follows:
~ ~
CO CO
1 1
OH HOCH 2 O-CH2
1 1
+ R -----+ R +2H 2O
1 1
OH HOCH 2 O-CH2
1 1
CO CO
~ 1
and
~
CO ~
1 1
OH HOCH 2 CO
1 1 + 2H 2O
+ R -----+ R
1 1 +2H.CHO
OH HOCH 2 CO
1
CO
~
~
The principal methylol-containing compounds used to crosslink the polymers

from carboxylated latices by this reaction are those derived by reactions
between phenol and formaldehyde (Structure XXV), between urea and
formaldehyde (Structure XXVI), and between melamine and formaldehyde
(Structure XXVII). It may be desirable to reduce the reactivity of methylolic
OH
,
NH.CH 2 0H
,
co
,
NH
,
CH 2
,
CH 2
o,
o,
xxv XXVI XXVII
compounds towards carboxylated polymers; if so, the methylol groups can

be etherified by partial or complete alkylation. The alkyl group is eliminated
as the corresponding alcohol when interaction with carboxylic-acid groups
occurs. The alkyl group is most commonly methyl. An example of a fully
alkylated product is hexamethoxymethylmelamine.
Epoxides form ester linkages by reaction with carboxylic-acid
groups, behaving effectively as anhydrides of 1,2-diols. The reaction is as
follows:
~ ---1----
CO
, co
,
OH CHR O-CHR
I
+0~1
---+
O-CHR
I
OH CHR co
I
____co
1__ _
----,----
Interaction with a polyamino compound, such as polyethyleneimine, to

form amide linkages can be represented as follows:
~ ~
co co
I I
OH NH2 NH
I I
+R --+ R + 2H 2 0
I I
OH NH2 NH
I I
co CO
~ ~
Interaction between pairs of adjacent carboxylic-acid groups to form
acid-anhydride linkages occurs by elimination of water as follows:
~
CO
I
OH
OH
I
CO
~
This reaction indicates that carboxylated latex polymers should be self-

crosslinkable by heating alone. However, it is doubtful whether this method
of crosslinking has ever been employed industrially to any extent.
11.8.2.3 Special reactions to which latex polymers junctionalized with N-

methylolamide and related groups are susceptible
The best-known N-methylolamide-containing units incorporated in func-
tionalized latex polymers are those derived from N-methylolacrylamide
(Structure XXVIII) and N-methylolmethacrylamide (Structure XXIX).
CH 2=CH.CO.NH.CH 20H CH 2=C{CH 3 ).CO.NH.CH 20H

XXVIII XXIX
Polymers containing units derived from these monomers are in principle

self-crosslinkable through reactions such as that in which methylene-
bisacrylamido linkages are formed between adjacent molecules. This reaction
is catalyzed by acid. The acidity can be provided either by adding a small
amount of a strong acid, such as p-toluenesulphonic acid, or by incorporating
acid groups in the copolymer by copolymerization of a small amount of an
acidic monomer. If a carboxylic-acid monomer is used for this purpose, then
reactions of the type described in section 11.8.2.2 above between methylol

compounds and carboxylic-acid groups can also occur. The self-crosslin king
reaction to form methylene-bisacrylamido linkages can be represented as follows:
~ ---1----
co co
I I
NH NH
I I +H 2 0
CH 20H ----+ CH 2
I + H.CHO
CH 2 0H NH
I I
NH.CH 20H co
I ~
co
----'--------
Latex polymers are also functionalized with reactive units of the
N-methylolamide type by copolymerization with N-alkyl ethers of
N-methylolacrylamides and N-methylolmethacrylamides. These monomers
are usually known as N-alkoxymethylacrylamides and N-alkoxymethylmeth-
acrylamides respectively. Incorporation of units derived from these monomers
gives self-crosslink able polymers which are less susceptible to premature
crosslinking than are those containing units derived from N-methylolacrylamide
and N-methylolmethacrylamide. Relative to the latter monomers, these
monomers are less hydrophilic, and so are more miscible with hydrophobic
monomers. They are also less sensitive to water, and have reduced toxicity.
Specific examples include N-methoxymethylacrylamide (Structure XXX),
N-n-butoxymethylacrylamide (Stucture XXXI) and N-isobutoxymethylac-
rylamide (N-2-methylpropoxymethylacrylamide) (Structure XXXII). Other
comonomers which can be used to introduce reactive units of the
N-methylolamide type into functionalized latex polymers include allyl-
N-methylolcarbamate (Structure XXXIII) and hydroxymethylated N-
formyl-N'-acryloylmethylendiamine (Structure XXXIV).
CH 2=CH.CO.NH.CH 20CH 3 CH 2=CH.CO.NH.CH 20 (CH 2hCH 3

XXX XXXI
CH 2=CH.CO.NH.CH 20CH 2CH(CH 3 h

XXXII
CH 2=CH.CH 20.CO.NH.CH 20H CH 2=CH.CO.NH.CH 2·NH.CO.H

XXXIII XXXIV
11.8.2.4 Special reactions made possible by the inclusion of other functional

groups in latex polymers
C-methylol groups can be introduced into latex polymers by copolymerization
with hydroxyalkyl esters of acrylic and methacrylic acids, such as 2-hydroxyethyl
acrylate (Structure XXXV) and 2-hydroxyethyl methacrylate (Structure
XXXVI). They can also be introduced by copolymerization with hy-
droxymethylated diacetoneacrylamide (Structure XXXVII).
CH 2=CH.CO.OCH 2 CH 2 0H CH 2 =C(CH 3 )·CO.OCH 2 CH 2 0H
XXXV XXXVI
CH 2 =CH.CO.NH.C(CH 3 h·CO.CH 3
XXXVII
C-methylol groups can react in at least three ways:
1. They can react with external methylol-containing crosslin king compounds
of the type described in section 11.8.2.1 above. The expectation is that
formaldehyde and water will be eliminated, and methylene linkages formed
between the polymer and the added compound. They can similarly react
with any internal N-methylol groups which have been introduced into
the polymer by copolymerization with a monomer such as N-methylol-
acrylamide. A further possibility arises if a functionalized latex polymer
containing N-methylol groups, derived from a monomer such as
N-methylolacrylamide, and C-methylol groups, derived from a monomer
such as 2-hydroxyethyl methacrylate, is reacted with an external
methylol-containing crosslinking compound. It is then in principle possible
to exercise some control over the distribution of crosslinks in the final poly-
mer by varying the ratio ofN-methylol to C-methylol groups in the polymer.
2. They can react with copolymerized acid groups to form ester crosslinks. If
the acid groups have been incorporated by copolymerization of a
carboxylic-acid monomer, then the crosslinking reaction can be represented
as follows:
-I -- - -j
CO CO
1
o o
1
1
CH 2
1
CH 20H
-- 1
CH 2
1
CH 2
+ H 20
1
OH o
1 1
CO CO
~ ~
3. They can react with copolymerized ester groups by ester-interchange to

form ester crosslinks. If the ester groups have been incorporated by
copolymerization of a carboxylic-acid ester, such as ethyl acrylate, then the
crosslinking reaction can be represented as follows:
~ ~
C CO
I I
o o
I I
CH 2 CH 2
I I +ROH
CH 20H -------> CH 2
I
R o
I I
o CO
I ~
CO
~
Peroxo groups can be introduced into synthetic latex polymers by

copolymerization with monomers such as tert-butyl peroxoacrylate (Structure
XXVIII) and 2-tert-butylperoxoethylacrylate (Structure XXXIX). It is
apparently possible to emulsion copolymerize such monomers, in small
concentration at least, without significant decomposition. Latex polymers
functionalized in this way have been used to effect subsequent graft copolymer-
ization on to the polymer which contains the peroxo groups.
CH 2=CH.CO.OOqCH 3h CH 2=CH.CO.OCH 2CH 200qCH 3h

XXVIII XXXIX
There has also been some interest in preparing synthetic latex polymers
which are functionalized by the inclusion of minor amounts of various silane
monomers, such as vinyltrimethoxysilane (Structure XL), y-methacryloxy-
propyItrimethoxysilane (Structure XLI) and f3-methacryloxyethoxy-
trimethoxysilane (Structure XLII). There has been considerable doubt in some
cases as to the extent to which the monomer has undergone hydrolysis-
condensation reactions as well as, or instead of, addition copolymerization. The
carbon-oxygen-silicon bond in the second and third of these monomers is
to some extent sterically shielded. It is claimed that latex polymers which have
been functionalized with small amounts of certain silane groups are capable
of producing films which subsequently crosslink at normal ambient temperatures.
CH 2=CH.Si(OCH 3h CH 2=C(CH 3)·CO.O(CH 2hSi(OCH 3h

XL XLI
CH2=C(CH3)·CO.O(CH2hOSi(OCH3h
XLII
11.8.3 Alka6-thickenable synthetic latices

Alka6-thickenable latices, also known as alkali-swellable latices, are latices
which undergo substantial increases in viscosity when they are made alkaline.
They have found use as thickening agents for aqueous systems, in particular,
for aqueous latices. One convenient way of using them to increase the viscosity
of latices is to blend the latex with the alkali-thickenable latex at a pH below
that at which the viscosity of the latter increases sharply, and then to raise
the pH of the blend to an appropriate level by addition of alkali. This technique
is, of course, applicable only to latices which are colloidally stable at pH values
at which the viscosity of the alkali-thickenable latex is relatively low.
Alkali-thickenable synthetic latices are produced by emulsion copolym!=rizing
one or more hydrophobic monomers with an appropriate amount of a
hydrophilic comonomer which contains weak-acid groups, the pH of the
reaction system being such that the weak-acid groups are mainly un-ionized.
The amount of hydrophilic comonomer must be such that, when the resultant
latex is subsequently made alkaline and the weak-acid groups ionize, the latex
particles absorb substantial amounts of the dispersion medium. Some of the
functionalized polymer molecules may dissolve in the aqueous phase. Partly
because the effective volume fraction of disperse phase in the latex has increased,
and partly because hydrophilic macromolecules may have dissolved into the
aqueous phase, the system undergoes a very substantial increase in viscosity.
Ionization of the weak-acid groups combined in the polymer molecules has
at least two effects, both of which tend to cause the latex particles to swell:
1. The thermodynamic tendency of the polymer molecules to mix with water

molecules increases markedly because the hydrophilicity of the acidic sites
increases. In effect, the XI-value for the copolymer and water is decreased.
2. Ionization also has the effect of causing the polymer molecules to adopt
more expanded conformations because of electrostatic repulsions between
the ionized groups.
The weak-acid groups are usually of the carboxylic-acid type. Indeed, all
carboxylated synthetic latices are to some extent alkali-thickenable, in that the
viscosity always tends to increase as the pH is raised and the carboxylic-acid
groups ionize. However, the adjective alkali-thickenable is usually reserved for
synthetic latices containing polymers which have been carboxylated to a
considerably greater extent than that which is appropriate for other applications.
In consequence, the viscosity can increase by several orders of magnitude as

the pH is raised. The effect of absorption of dispersion medium by particles
upon latex viscosity is further considered in section 6.3.3 of Chapter 6 (Volume
1) (see especially Figure 6.7). The effect of soluble lyophilic macromolecules
upon latex viscosity is discussed in section 6.3.4 of Chapter 6 (Volume 1).
Papers in which information is available concerning the preparation,
characterization and properties of alkali-thickenable carboxylated synthetic
latices include those by Muroi [51], by Muroi, Hosoi and Ishikawa [52], by
Verbrugge [53,54], and by Nishida et al. [55]. The paper by Nishida et al. is
particularly informative, in that it describes alkali-thickenable latices produced
by both batch emulsion polymerization, in which all the ingredients were
charged initially, and by semi-continuous emulsion polymerization under
monomer-starved conditions. The main monomer was methyl methacrylate, the
carboxylating monomer methacrylic acid. The methyl methacrylatelmethacrylic
acid ratio was varied from 9515 to 75/25 m/m. Emulsion copolymerization was
effected at 60·C using potassium peroxodisulphate as initiator, a small amount
of n-dodecyl mercaptan as modifier, and a combination of sodium n-dodecyl
sulphate and an ethoxylate as colloid stabilizer. The water content of the
reaction systems was such that, after complete polymerization, the total solids
contents of the latices were ca. 10% m/m. The final pH values were ca. 3.5;
thus the copolymerized methacrylic-acid units were substantially un-ionized.
The monomer reactivity ratios for methyl methacrylate and methacrylic acid,
as determined by Nishida et at., are such that methacrylic acid is expected to
enter the copolymer preferentially. Hence the copolymers formed initially in
the batch copolymerizations were relatively rich in methacrylic acid, and those
formed at the end of the reaction relatively rich in methyl methacrylate.
Nevertheless, the acid-rich copolymers were found to be preferentially situated
near the surface of the particles, presumably as a consequence of the greater
hydrophilicity of methacrylic acid relative to methyl methacrylate. As expected,
the increase in viscosity of the resulting latices on being made alkaline increased
as the level of carboxylation increased. However, for any given level of
carboxylation, the carboxylated latices produced by semi-continuous emulsion
polymerization under monomer-starved conditions were found to undergo
much greater increases in viscosity than did those produced by batch emulsion
polymerization. This was interpreted as being a consequence of a more even
distribution of carboxylic-acid groups in the copolymer produced by the
semi-continuous emulsion polymerization. The lesser increases in the viscosity
of alkali-thickenable latices prepared by batch emulsion copolymerization was
attributed to the alkali-swellable polymer being confined largely to the surface
regions of the particles. Whatever the true explanation, these results demonstrate
that the extent to which the viscosity of an alkali-thickenable carboxylated
latex increases when the pH is raised depends not only upon the level of
carboxylation but also upon the way in which the latex was prepared. Figures
11.l4(a) and (b) show the effect of increasing pH upon the flow behaviour of
~
• ••
105
: 104
a..
E
"'-
~1oa
0
0
.!Il
~c: 102
I!!
111
a.
.¥" 101
10°
10-1
3 5 7 9 11 13
(b) pH
Figure 11.14 (a) Effect of pH upon the relationship between apparent viscosity and rate of
shear strain for an 85/15mjm methyl methacrylatejmethacrylic acid copolymer latex (Nishida
et at. [55]). The latex was prepared by semi-continuous emulsion copolymerization at 60°C
under monomer-starved conditions. Significance of points: • pH 4.2; • pH 8.2; • pH 8.4; ... pH 11.5.
(b) Effect of pH upon apparent viscosity at a single rate of shear strain (215s- l ) for methyl
methacrylatejmethacrylic acid copolymer latices containing copolymers of various compositions
(Nishida et at. [55]). The latices were prepared as for (a). Different styles of point refer to latices
of copolymers having different methyl methacrylatejmethacrylic acid mjm ratios as follows:
.95/5; • 90/10; • 85/15; ... 80(20; • 75(25.
Miscellaneous synthetic latices 371
the alkali-thickenable latices produced by semi-continuous emulsion copolymer-

ization under monomer-starved conditions. Figure 11.l4(a) shows the effect
of increasing pH upon the relationship between the apparent viscosity and
the rate of shear strain for the latex which contained an 85/15 m/m methyl
methacrylate/methacrylic acid copolymer. As the pH increased, so the viscosity
at any given rate of shear strain increased at first, but subsequently decreased
somewhat. Furthermore, the more viscous the latex at any given rate of shear
strain, the more non-Newtonian was the system, in that the more strongly did
the apparent viscosity depend upon the rate of shear strain. Figure 11.l4(b)
shows the effect of pH upon the apparent viscosity at a single rate of shear
strain (215 s - 1) for latices containing copolymers of various methacrylic acid
contents. It is evident that pronounced increases in viscosity occurred as the
pH was raised if the methyl methacrylate/methacrylic acid ratio was less than
ca. 85/15m/m. The viscosity increased to a maximum at pH ca. 9, and then
decreased somewhat with further increase in pH.
The large increase in viscosity which occurred as the pH of the latices of
the copolymers rich in methacrylic acid was raised to ca. 9 is attributed to
extensive swelling of the latex particles, with possible dissolution of some of
the polymer molecules into the aqueous phase. Furthermore, the appearance
of the latices changed from a milky-white colour typical of many synthetic
latices to a slightly hazy appearance at the pH corresponding to maximum
viscosity. This indicates that the system had in effect changed from a
hydrophobic sol to a hydrophilic colloid in which the water molecules of the
dispersion medium were intimately mixed with the segments of the polymer
molecules. In consequence, the difference between the refractive index of the
disperse phase and that of the dispersion medium had largely disappeared.
The shape of the curves for the variation of apparent viscosity with rate of
shear strain (Fig. 11.14(a)) is attributed to the development of shear-destructible
structure in the system through extensive hydrogen bonding. The considerable
decrease in viscosity as the pH was raised above ca. 9 is a consequence of
shielding of the ionized carboxylic-acid groups by the cations of the alkali.
11.9 MISCELLANEOUS SYNTHETIC LATICES
11.9.1 Cationic and amphoteric synthetic latices

It has long been recognized that, relative to the usual aqueous anionic latices,
aqueous cationic latices, containing as they do positively-charged particles,
can in principle have distinctive properties which make them especially suitable
for certain applications. Examples of these applications include cathodic
electrodeposition and the impregnation of textile fabrics, webs and mats, the
fibres of which are negatively charged. Concerning the second ofthese examples,
it was believed that, because the particles in a cationic latex are of opposite
polarity to the textile fibres, such latices would have advantages relative to
anionic equivalents, in that the particles would deposit spontaneously and
uniformly upon the fibres, thereby providing more effective bonding of the
fibres. However, these advantages have not been realized in practice, and
interest in cationic latices for this type of application has waned. Although
there has long been industrial interest in the potentialities of cationic latices,
the production of these latices has never been more than a marginal activity
relative to the production of anionic latices. The preparation of cationic
synthetic latices has been extensively reviewed by Wessling [56].
Like anionic latices, cationic latices can be classified broadly into those
derived from natural sources (of which class natural rubber latex is the only
representative in which any interest has been expressed in this respect), those
produced by emulsion polymerization, and those produced by dispersion of
bulk polymers in aqueous media. The exclusive concern in this section is with
those produced by emulsion polymerization. However, it may be noted that:
1. Cationic natural rubber latex and cationic synthetic latices have been
produced by modifying the surface of the particles in an anionic latice by
addition of excess of a cationic surface-active substance. Surface-active
cations thereby become adsorbed at the particle surface. In consequence,
the polarity of the effective charge carried by the particles is reversed. The
preparation of cationic latices by this procedure is described in section
14.1 1.1 of Chapter 14.
2. Using cationic surface-active dispersing agents and colloid stabilizers instead
of the usual anionic substances, any of the methods described in section
12.3 of Chapter 12 for the production of artificial latices can in principle
be used to obtain cationic artificial latices.
The positive electric charge at the surface of cationic latex particles is usually
provided by the presence of surface-bound ammonium-ion derivatives. As with
anionic latices, the surface binding may be by either physical adsorption or
covalent bonding. The ammonium ions may be either quaternary or
non-quaternary derivatives. The charge conferred by the former is essentially
independent of the pH of the aqueous phase, because substituted quaternary
ammonium hydroxides are strong bases. The charge conferred by surface-
bound non-quaternary ammonium ions is dependent upon the pH of the
aqueous phase, because the corresponding substituted ammonium hydroxides
are weak bases; the charge therefore tends to be lost as the pH is raised. The
pH range over which the charge is neutralized depends upon the base strength
of corresponding substituted ammonium hydroxide. Other types of surface-
bound cation which have been of interest for the production of cationic latices
include phosphonium ions (R 4 P+) and sulphonic ions (R 3S +).
Reaction systems for the direct production of cationic synthetic latices by
emulsion polymerization fall into two broad classes, depending upon the way
in which the positive electric charge is conferred to the latex particles. The
first category is reaction systems for which the charge is conferred primarily
by the use of cationic colloid stabilizers which become physically adsorbed at
the surface of the particles. The second category comprises those for which
the charge is conferred primarily by the use of cationic initiators and/or
(especially) functionalizing cationic comonomers which provide covalently-
bonded cationic groups at the surface of the particles. Cationic latices can also
be produced by emulsion polymerization reaction systems such that the particle
charge is derived from both adsorbed surface-active cations and covalently-bonded
cationic groups.
Formulations given by Howland et al. [57] for emulsion polymerization
reaction systems for the preparation of styrene-butadiene copolymer latices
of the first of the above types are given in Table 11.14. Formulation A uses
n-dodecylammonium chloride as the colloid stabilizer and tert-butyl hydroperoxide
as the initiator; Formulation Buses n-dodecylammonium acetate and a
combination of diisopropylbenzene hydroperoxide and iron(II) ions. The
surface-active cations which confer colloid stability on both the resultant latices
are of the non-quaternary type. Formulation A is intended for polymerization
at ca. 50'C, the conversion being ca. 85% after 22 hours. It is recommended
that an additional 1 pphm of n-dodecylammonium chloride can be added after
ca. 10% conversion. The only added initiator is tert-butyl hydroperoxide.
However, it is likely that a redox couple is formed between the hydroperoxide
and the small amounts of n-dodecylamine present from hydrolysis of
n-dodecylammonium ions. It is the possibility of such interactions with amines
Table 11.14 Typical formulations for emulsion polymerization reaction systems for
preparation of cationic synthetic styrene-butadiene copolymer latices for which the
positive electric charge is conferred to the latex particles by adsorbed cationic colloid
stabilizers (Howland et al. [57])
Parts by mass
Ingredient A B
Styrene 50 30
Butadiene 50 70
Water 120 180
n-dodecylammonium chloride 5
n-dodecylammonium acetate 3
tert-hexadecyl mercaptan 0.35
Mixed tert-mercaptans 0.15
tert-butyl hydroperoxide 0.30
Diisopropylbenzene hydroperoxide 0.05
Iron(II) chloride 0.05
Potassium chloride 0.3 0.10
Aluminium chloride hexahydrate 0.36 0.30
Polymerization temperature / 'C ca. 50 ca. 5

derived from non-quaternary ammonium surface-active cations which make

hydroperoxides suitable as the sole added initiator for reaction systems of this
type intended to be polymerized at relatively high temperatures. Formulation
B is intended for polymerization at ca. 5°C, the conversion being ca. 60"10 after
12 hours. Initiation is by the redox couple formed between diisopropylbenzene
hydroperoxide and iron(II) ions. The rate of polymerization depends upon the
pH of the aqueous phase. It is low at both low and high pH, and passes
through a sharp maximum at pH ca. 6. It is probable that polymerization is
retarded at high pH because of the hydrolysis of both the iron(II) ions and
the n-dodecylammonium ions, and at low pH because of suppression of the
solubility in the aqueous phase of the diisopropylbenzene hydroperoxide. It
has been found that either acetic acid, or mixtures of acetic and hydrochloric
acids, are suitable for the adjustment and control of the pH of this type of
reaction system.
Initiation systems based upon hydroperoxides are well-suited to the production
of cationic latices by emulsion polymerization, compared with anionic initiators
such as peroxodisulphate ions. The latter tend to confer negative electric charges
on the latex particles, and so reduce the effective density of surface-bound
positive charge. Notwithstanding this consideration, many reaction systems
using peroxodisulphates as initiator have been proposed for the preparation
of cationic synthetic latices. Thus Wessling [56] considers that many of the
problems which have been encountered with cationic synthetic latices can be
traced to the use of peroxodisulphates as initiator.
Suitable functionalizing comonomers which can be used for the preparation
of cationic synthetic latices for which the positive electric charge is conferred
to the latex particles primarily by the comonomer include 2-aminoethyl
methacrylate (Structure XLIII) and N, N-dimethyl 2-aminoethyl methacrylate
(Structure XLIV). A typical cationic dissociative initiator is the water-soluble
azo compound oc,oc'-azobisisobutyramidinium chloride (Structure XL V).
CH2=C(CH3)·Co.OCH2CH2NH2 CH 2=C(CH 3)·CO.OCH 2CH 2N(CH 3h

XLIII XLIV
2+
Some interest has also been shown in amphoteric synthetic latices. In these
latices, the nature of the surface ionicity is such that the polarity of the
surface-bound charge reverses from positive to negative as the pH is raised;
these latices are therefore cationic at low pH and anionic at high pH. At some
intermediate pH, the latex is in the isoelectric state in which the particles do
not carry an overall surface-bound electric charge. This situation can in
principle arise in two distinct ways. The isoelectric condition may be such that
the particles are entirely uncharged electrically, or it may be such that the
particles carry equal charge densities of positive and negative electric charge.
In either case, it is expected that the colloid stability will be low because the
contribution from electrostatic stabilization is negligible. It is also expected
that the contribution from hydration stabilization will be greater in the latter
case than in the former, because there should be a much stronger tendency
for water molecules to bind to the particle surface by dipole interactions.
Amphoteric synthetic latices can be prepared by the use of mixtures of cationic
and anionic functionalizing comonomers in reaction systems of appropriate
pH. A typical combination is 2-aminoethyl methacrylate and methacrylic acid.
11.9.2 Synthetic latices which contain microvoid particles

There has been interest in recent years in producing synthetic latices which
contain microvoid particles. The latter are hard particles which contain a
number of small voids. Because of the marked difference in refractive index
between the body of the particle and the micro voids when water has been
removed from the latter, these particles can be used as opacifiers for films.
Being themselves in the form of aqueous latices in the first instance, they can
if necessary be incorporated directly into surface-coatings compositions based
upon aqueous latices.
Latices which contain microvoid particles can in principle be produced by
core-shell emulsion polymerization in which a core of an alkali-swellable
carboxylated latex polymer is encapsulated within a hard crosslinked polymer,
and the core polymer then subsequently neutralized by reaction with an alkali
such as ammonia. When the water evaporates from the latex, microvoids form
within the particles. Two essential requirements for the success of this process are:
1. that the alkali-swollen core of the particles must be of sufficient volume
relative to the volume of the particle to shrink upon drying to form
micro voids inside the particles;
2. that the rigid shell of the particles must be sufficiently strong to maintain
the integrity of the particle during the swelling and drying of the core.
Although straightforward in principle, processes of the type outlined in the
previous paragraph are in practice beset with the problem that thermodynamic
factors favour the formation of core-shell particles having the opposite
morphology to that which is desired (see section 10.2.4.5.3 of Chapter 10). The
tendency is for particles to form which have rigid hydrophobic cores encapsulated
with hydrophilic alkali-swellable shells, rather than particles which have
alkali-swellable cores surrounded by rigid hydrophobic shells. Vanderhoff,
Park and EI-Aasser [58] have described a sequential emulsion polymerization
procedure for overcoming this problem. Methyl methacrylate and methacrylic

acid are first emulsion copolymerized in the absence of added colloid stabilizer
to give a seed latex which provides the cores of the particles to be formed.
This seed latex is next sequentially emulsion polymerized with a mixture of
methyl methacrylate, methacrylic acid and ethyleneglycol dimethacrylate to
provide a crosslinked seed latex. The carboxylic-acid groups are then neutralized
with ammonia and the particles encapsulated with a first shell layer by seed
copolymerization using a mixture of methyl methacrylate, butyl acrylate and
methacrylic acid. A crosslinked rigid second shell is then applied by
copolymerization with a mixture of styrene and divinylbenzene. The final step
is drying of the latex to remove the water from the internal particle micro voids
to give hollow particles. The success of this process is in part at least a
consequence of fixing the core-shell particles in the desired morphology by
incorporating covalent crosslinks during polymerization.
Okubo, Ichikawa and Fujimura [59] have described a so-called stepwise
alkali-acid process for preparing microvoid latex particles. They used:
74.3/17.0/8.7 mol/mol/mol styrene/n-butyl acrylate/methacrylic acid copolymer
latex prepared by emulsion copolymerization in the absence of added colloid
stabilizer. A surface-active substance was then added to the latex to prevent
colloidal stabilization during the subsequent two-step treatment. In the first
step, the pH of the latex was raised to various values in the range 7.0-13.0
by addition of 0.2 mol dm - 3 aqueous potassium hydroxide solution, and the
latex heated at various temperatures for various times, typically 3 hours at
75T. After this treatment, the latex was rapidly cooled to normal ambient
temperature. In the second step, the pH was reduced to 2.2 by addition of
0.2 mol dm - 3 hydrochloric acid, heated for 3 hours at 75·C, and rapidly cooled
to normal ambient temperature. The nature of the particles obtained after
drying was ascertained by electron microscopy. Figure 11.l5(a) shows the
relationship between the expansion of the particles at the end of the second
step and the level to which the pH was raised in the first step. Figure 11.15(b)
shows the relationship between the microvoid size and the level to which the
pH was raised in the first step, and also that between the number of micro voids
per particle and the level to which the pH was raised in the first step. These
results indicate that considerable control can be exercised over the nature of
the microvoid particles formed in this process by varying the parameters of
the two-step treatment.
A patent to Kowalski and Vogel [60] describes in detail the preparation
and properties of latices containing microvoid particles intended for use as
opacifiers. A typical latex contains particles having average diameter ca. 300 nm,
with average void diameter ca. 130 nm. The latices are to be non-film-forming at
20·e. A further patent to Kowalski, Vogel and Blankenship [61] describes
latices of this type claimed to be suitable for use as thickeners of aqueous
media, and also for various other applications such as adhesives for paper and
cardboard, and for the sealing, priming and coating of leather.
50 0.5
~ 40 0.4
--.§
-;{!. E
--Ie
0 I
:::l.
I
I
30 0.3 I
UI
r:::: ·iii I
as
Q.
"0 I
I
~ 20 .~ 0.2 I
..9! 0
0
'E ~
8!. 10 0.1
0 0 ~-+--~~--~--~~~O
7 8 9 10 11 12 13 14 7 8 9 10 11 12 13 14
(8) Initial pH (b) Initial pH
Figure 11.15 (a) Relationship between expansion of the particles at end of second step and
level to which the pH was raised in the first step, for the Okubo-Ichikawa-Fujimura [59]
stepwise alkali-acid process for preparing microvoid latex particles. (b) Relationship between
microvoid size and level to which pH was raised in the first step (points .), and that between
the number of microvoids per particle and level to which pH was raised in the first step (points ... ),
for the Okubo-Ichikawa-Fujimura [59] stepwise alkali-acid process for preparing microvoid
latex particles.
11.9.3 Synthetic latices which contain polymers functionalized such that they
strongly absorb ultraviolet light
Polymers which strongly absorb ultraviolet light are of industrial interest for
several reasons. The tendency of the polymer to be degraded by ultraviolet
light is reduced. The polymers can be used as protectives for other materials,
articles, etc., which are susceptible to undesirable deterioration and degradation
by ultraviolet light. They can also be used to encourage the occurrence of
desirable photochemically-initiated chemical reactions. Polymers which strongly
absorb ultraviolet light can be prepared in various ways. One method is to
copolymerize the principal monomers of the polymer with a minor proportion
of one or more monomers which contain functional groups which are capable
of strongly absorbing ultraviolet light. Incorporation of the absorbing moieties
by copolymerization obviates problems such as dispersion, subsequent
volatilization, migration and extraction of the absorbing moiety; such problems
may be encountered if other methods of incorporation are used. If the
copolymerization is carried out in the emulsion mode, then the copolymer is
initially obtained as a synthetic latex.
The inherent tendency of a molecule to absorb incident electromagnetic
radiation of a particular frequency is conveniently quantified by its molar
absorptivity (or molar extinction coefficient), denoted by B. It is defined as the
quantity -(l/cl)loglO(lllo), where 10 is the incident intensity, 1 is the intensity
after the radiation has travelled a distance I through the substance, and c is
the concentration of absorbers. It is usual for I to be expressed in cm and c

in mol dm - 3; the units of e are then dm 3 mol- 1 cm - 1. In general, the value
of e depends upon the frequency of the radiation. Some idea of the significance
of e-values in the context of minor concentrations of absorbing moieties
dispersed in a non-absorbing matrix can be gained by noting that, over a path
length of t mm, ca. 0.5% mlm of an absorbing moiety of molecular mass 300
daltons and molar absorptivity t 000 dm 3 mol- 1 cm -1 attenuates the intensity
of radiation by ca. 98%.
Amongst the various ethylenic monomers which have been used to incorporate
ultraviolet-absorbing moieties into polymers by free-radical copolymerization
are derivatives of2-hydroxybenzophenone, derivatives of2-hydroxyacetophenone,
and derivatives of phenyl salicylate. Examples of the first of these types of
compound are 2-hydroxy-4-acryloxybenzophenone (Structure XL VI) and
2-hydroxy-4-methoxy-5-methylolbenzophenone acrylate (Structure XL VII); an
example of the second is 2-hydroxy-4-methacryloxyacetophenone (Structure
XLVIII). Compounds of these two types have similar structures. In particular,
OH
OH
CH 2
O.CO.CH=CH 2 I
O.CO.CH=CH 2
XLVI XLVII
OH
O.CO.C(CH 3 )=CH 2
XLVIII
they both contain a benzene ring which has substituent carbonyl and hydroxyl
groups 0- to each other. Hydrogen bonding occurs between these two
substituents, thereby giving a further six-membered planar ring. The presence
of this hydrogen-bonded structure is believed to have at least two important
consequences. Firstly, the reactivity of the hydroxyl group is reduced. In
particular, it is reduced relative to that of other hydroxyl substituents which
may be present in the molecule, thus permitting derivatization of the molecule
by reaction with those other hydroxyl groups without the complication of
simultaneous derivatization of the o-hydroxyl group. Secondly, strong absorption
of ultraviolet light quanta can occur with concomitant photoenolization of
the hydrogen-bonded structure. The electronically-excited state which results

from the photoenolization process can then revert to the initial state, with
degradation ofthe energy of the original light quantum into the comparatively
harmless form of heat. It is significant that the processes of photoenolization
and subsequent reversion to the initial state do not entail the cleavage of
covalent bonds. The preparation of a styrene homopolymer latex and a
styrene-butadiene copolymer latex functionalized with 2-hydroxy-4-methoxy-
5-methylolbenzophenone acrylate has been described by Barabas, Mallya and
Gromelski [62].
11.9.4 Synthetic latices which contain unusual polymers
J1.9.4.1 Synthetic latices which contain hydrophilic polymers

Yen et al. [63] have described the preparation by conventional emulsion
polymerization of synthetic latices which contain crosslinked hydrophilic
polymer particles of various diameters, functionalized with carboxylic-acid
groups. The main monomers are methyl methacrylate and 2-hydroxyethyl
methacrylate, the functionalizing monomer is methacrylic acid, and the
crosslinks are introduced by copolymerization with ethyleneglycol dimethacrylate.
The formulation (in parts by mass) for producing a latex containing particles
of diameter 340nm comprises methyl methacrylate 58.7, 2-hydroxyethyl
methacrylate 31.0, methacrylic acid 10.3, ethyleneglycol dimethacrylate 3.2,
sodium n-dodecyl sulphate 0.3, ammonium peroxodisulphate 0.03, and water
191.4. Polymerization is effected by heating for 1 hour at 98°C in a sealed
tumbled container. The conversion is 99-100%, and the polymer content of
the resultant latex ca. 35% m/m. The level of added colloid stabilizer is low
(0.1 % m/m on the whole latex); colloid stability is presumably considerably
supplemented by the combined carboxylic-acid groups and (to a lesser extent)
the combined hydroxyl groups present at the surface of the particles. It is said
that, by varying the water content of the reaction system, latices having a wide
range of particle sizes can be prepared by this procedure. The particle size of
the final latex decreases linearly with the overall monomer content of the
reaction system from 340 nm when the monomer content is 35% m/m to 35 nm
when it is 3% m/m. According to Yen et al., latices of this type which contain
larger particles (from 300 to 3000 nm) can be prepared by emulsion/dispersion
polymerization in water initiated by cobalt-60 y-rays. The preparation of this
type of latex is also described in a paper by Rembaum ei al. [64].
11.9.4.2 Synthetic polypyrrole latices

Some interest is currently being shown in latices which contain polypyrrole
(Structure XLIX). This polymer is intensely black in colour, and also conducts
electricity moderately well. Both these properties are presumably consequences
XLIX
offacile electronic transitions within its conjugated chain structure. It is unusual

for latex particles to be intrinsically non-white. The intense black colour of
polypyrrole latex particles makes them well-suited for immunoassay applications.
The polymer can be produced by oxidation of pyrrole with an iron(III)
compound. Procedures for the preparation of these latices have been described
by Armes, Miller and Vincent [65] and by Tarcha et al. [66]. In the procedure
described by the latter workers, a dilute aqueous solution of polyvinyl alcohol
is prepared by dissolving O.60g of the polymer in l00cm 3 of water. To this
solution is added a filtered solution of 17.6 g of iron(III) chloride hexahydrate
dissolved in lOOcm 3 of water. 2.0cm 3 of distilled pyrrole is then added all at
once with stirring. The solution turns black within a few seconds, indicating
that polymerization has commenced. The reaction continues for ca. 4 hours.
The particles can be washed and, if desired, the dispersion medium replaced
with some non-aqueous liquid by temporarily compacting the particles by
centrifugation and then separating the particles from the supernatant liquid.
11.9.4.3 Synthetic latices which contain composite particles comprising

polymers covalently bonded to inorganic particles
There has for many years been some interest in aqueous latices which contain
composite particles comprising a polymeric component and an inorganic
particulate component. The preferred morphology for the composite particle
is usually such that each particle contains one, but only one, inorganic particle,
and that the inorganic particles are entirely encapsulated by polymer. Various
inorganic particles have been considered as the inorganic component, including
silica, insoluble metal silicates, calcium carbonate and barium sulphate. As
well as being intrinsically interesting, latices of such particles have several
potential industrial applications as stiffening and reinforcing fillers for rubbers
and pigments for latex-based surface coatings. The usual way in which it is
attempted to form such particles is by emulsion polymerization in the presence
of a fine stable colloidal dispersion of the inorganic particles. If it is desired
that the eventual latex particles should contain only one inorganic particle,
then it is necessary that steps be taken to disintegrate any agglomerates which
may have formed from the primary inorganic particles. The intention is that
most of the reaction loci created during the nucleation period should form at
the surface of the inorganic particles. The polymer produced during the
subsequent stages of the reaction should then form mainly at the surface of
the inorganic particles, thereby achieving the objective of encapsulation.
Opportunities for particle nucleation by conventional mechanisms should be

minimized.
Espiard et al. [67] have reported the results of an investigation into the
products obtained when various monomers (styrene, ethyl acrylate, n-butyl
acrylate and methyl methacrylate) were emulsion-polymerized in the presence
of fine silica particles. The reactions were carried out at 60°C using potassium
peroxodisulphate as initiator. The monomer/water ratio was kept constant at
1O/90m/m, but the silica/monomer ratio was varied from 1/1 to 1/64m/m.
The silica particles were dispersed in water with the aid of various surface-active
substances (sodium n-dodecyl sulphate, n-dodecyltrimethylammonium bromide
and various ethoxylates). These latter presumably contributed to the colloid
stability of the final latex, and may also have assisted in the formation of
non-particulate reaction loci. Various monomer-addition regimens were used.
The nature of the surface of the silica particles was also varied. Some
experiments were carried out using an unmodified silica. In this case, the
particle surface was hydrophilic, comprising many silanol groups. Other
experiments were carried out using silicas which had been surface-modified in
one of the following two ways:
1. by reaction in dry toluene with y-methacryloxypropyltrimethoxysilane
(Structure XLI);
2. by reaction in dry toluene with y-aminopropyltriethoxysilane (Structure L),
followed by reaction of the y-aminopropyltriethoxysilane-modified silica in
dry tetrahydrofuran with 4,4' -azobis-4-cyanopentanoic acid (Structure LI)
in the presence of triethylamine and chloroform.
The following is the type of reaction which probably occurred at the surface
of the silica particles with treatment (I):
(CH 3 0b Si(CH 2 b O.CO.C(CH 3 ) = CH 2

XLI
+ CH 3 0H
The surface produced by this reaction is expected to be hydrophobic, with

surface-bound groups which are in principle capable of copolymerizing with
any monomers polymerizing at the particle surface. The types of reaction which
occurred at the surface of the silica particles with treatment (2) are probably
as follows:
~ ~C2HS
(C2HsOhSi(CH2)3NH2 ~i-O-fi(CH2)3NH2 +
L ~ OC 2 HS
CN CN
I I
H02C(CH2)2C-N=N-C(CH2~C02H
I I
CH 3 CH 3
LI
The surface produced by these reactions is again expected to be hydrophobic,

but is now in principle capable of initiating polymerization of monomers at
the particle surface by thermal decomposition of the surface-bound 4,4'-azobis-
4-cyanopentanoic acid groups to form one surface-bound free radical and one
non-surface-bound free radical. Thus both types of treatment were expected
to produce surface-modified silicas capable offorming covalent bonds between
the particle surface and the polymer which formed there subsequently. Some,
at least, of the polymer which forms subsequently is expected to be, in effect,
grafted on to the particle surface. The nature and extent of the encapsulation
of the silica particles by polymer were investigated by various techniques; these
were determination of particle size, elemental analysis and Fourier-transform
infrared spectroscopy, and comparison of the number of initial silica particles
with the number of final latex particles.
The conclusions reached from this investigation were that the following
reaction conditions are conducive to satisfactory encapsulation of fine silica
particles during emulsion polymerisation:
1. The surface ofthe silica particles must be sufficiently hydrophilic to disperse

thoroughly in the water. This implies that some unreacted hydroxyl groups
must be present at the silica surface.
2. The monomer must have sufficient solubility in water to encapsulate the
hydrophilic silica particles. Methyl methacrylate and ethyl acrylate have
sufficient solubility.
3. The amount of unreacted monomer present at any instant during the
emulsion polymerization should be limited. Thus semi-continuous emulsion
polymerization under monomer-starved conditions is appropriate.
4. If ethoxylates are used as dispersing agents for the silica, then ethoxylates
of higher rather than lower hydrophilicity are preferred.
Using unmodified hydrophilic silica, no encapsulation was observed, the silica

being recoverable unaltered from the final latex by centrifugation. The modified
silicas were found to be difficult to disperse satisfactorily in water, because
they were too hydrophobic. Better dispersion could be achieved if the
hydrophobicity of the particle surface was reduced by partial removal of surface
groups by hydrolysis with dilute aqueous ammonia solution, preferably in the
presence of a small amount of isobutanol. Conversion-time curves reported
for the emulsion copolymerization of ethyl acrylate in the presence of silica
surface-modified by reaction with y-methacryloxypropyltrimethoxysilane show
that the duration of the locus-initiation stage increases as the amount of silica
in the reaction system increases. This provides good evidence that a substantial
proportion of the reaction loci are formed around the silica particles.
Espiard et al. [67] have also described an alternative technique in which
functionalized silica particles are formed by the co-hydrolysis and co-
condensation of tetraethoxysilane and y-methacryloxypropyltrimethoxysilane
with a water-in-toluene emulsion in the presence of di-2-ethylhexyl sodium
sulphosuccinate as colloid stabilizer. The resultant dispersion is then modified
by addition and polymerization of acrylic acid, followed by phase inversion
to give an oil-in-water emulsion. The toluene is removed by evaporation, and
the dispersion used as part of a subsequent emulsion polymerization reaction
system. Further information concerning this latter type of process is available
in a paper by Espiard, Mark and Guyot [68].
11.9.5 Redispersible synthetic latices

As with natural rubber latex (section 9.4.7 of Chapter 9), some interest has
been shown in the possibility of producing synthetic latices which dry down
to deposits capable of being readily redispersed in aqueous media to give
reconstituted synthetic latices. The intention has been to provide a synthetic
latex analogue of powdered milk. Although the advantages in respect of latex
transportation and storage are obvious, this concept has not so far been
exploited industrially to any extent, principally because of the technological
difficulties in obtaining reconstituted latices which are satisfactory for use in

existing latex processes. A major problem is the water-sensitivity of films which
form from the reconstituted latex. This problem has its origin in the high level
of hydrophilic substances which must be present in the latex to ensure ready
redispersibility after drying.
Earlier investigations into the possibility of preparing synthetic latices which
can be dried down to redispersible powders include those of Floria and Mock
[69], of Pateman [70] and of Guziak and Maclay [71]. Floria and Mock
reported the preparation of spontaneously redispersible synthetic latex polymers
by the emulsion copolymerization of styrene and sodium p-styrenesulphonate
(Structure LII). The redispersible polymers described by Pateman were
emulsion copolymers of vinyl acetate and diethyI2,2' -dichloromaleate (Structure
LIII). One significant observation from the work of Pateman is that, if the
polymer is to be redispersible after drying, then no particle coalescence should
have occurred during the drying of the initial latex. Thus if the initial vinyl
acetate-diethyl 2,2'-dichloromaleate copolymer latex, or the reconstituted
latex, was compounded with sufficient of an external plasticizer, then coherent,
non-redispersible films were obtained after drying. Presumably the latices
prepared by Floria and Mock and by Pateman dried down to fine powders
which comprised particles which had not coalesced together, and were such
that, on addition of water, forces of repulsion between the particles sufficient
to cause redispersion were established.
LlI LlII
Guziak and Maclay [7 t] have reported an investigation of synthetic latices

of styrene- 2-ethylhexyl acrylate- methacrylic acid copolymers, some of which
were found to be redispersible. A series of latices of various copolymer
compositions was prepared by emulsion polymerization. The styrene content
of the polymers was maintained constant at 55% m/m. The remaining
45% m/m comprised varying amounts of 2-ethylhexyl acrylate and methacrylic
acid. The characteristics of these latices (unneutralized) are summarized in
Table 11.15. It appears that the copolymer must contain rather more than
10% m/m of methacrylic acid ifthe latex is to be redispersible in alkaline water.
Redispersion was not observed if attempts were made to reconstitute the latex
in pure water, although, as would be expected, some absorption of water did
occur. Thus it appears that ionization of at least some of the carboxylic-acid
Table 11.15 Characteristics of unneutralized 55/(45 - XliX m/m/m styrene/

2-ethylhexyl acrylate/methacrylic acid copolymer synthetic latices (Guziak and
Maclay [71])
Level of methacrylic acid / pphm (i.e., X)
Characteristic 10 11 15 20 25 35 45
Continuous film when dried

at normal ambient
temperature? No No No No No No No
Redispersible at pH 9-1O? No No Yes Yes Yes Yes Yes
Thickening at pH 9-1O? No No No Yes Yes Yes Yes
Latex stable to calcium
chloride? No No No Yes Yes Yes Yes
Latex stable to alum? Yes Yes Yes Yes Yes Yes Yes
Latex freeze-thaw stable? Yes Yes Yes Yes Yes Yes Yes
Latex mechanically stable? Yes Yes Yes Yes Yes Yes Yes
groups of the polymer is necessary if repulsive interparticle forces sufficient to

cause redispersion are to develop in an aqueous medium. However, it was also
found that polymers which contained 20% m/m or more of methacrylic acid
tended to dissolve in alkaline media, giving very viscous solutions. The latices
of those copolymers which were redispersible, yet insoluble in aqueous alkali,
could be dried down to granular, free-flowing solids which examination by
microscope showed consisted of colourless transparent flakes. It appears that
the latex particles became compressed into one another as water was removed,
and that deformation occurred without full interparticle coalescence. It is likely
that interparticle coalescence was discouraged by the presence of surface
carboxylic-acid groups and adsorbed hydrophilic substances. Although the
appearance of the dry deposits under the microscope indicates that sufficient
coalescence occurred to destroy the interparticle boundaries, comparison
between the particle-size distribution and various particle-size averages for the
initial and the reconstituted latices provided convincing evidence that it was
essentially the original latex particles which were redispersed. Furthermore,
the redispersed particles, like the initial particles, were spherical in shape. Other
significant observations reported by Guziak and Maclay include the following:
I. Deposits dried from a reconstituted latex were water-sensitive, but not

completely redispersible a second time. The reason for this is unclear.
2. Baking the deposit from the initial or reconstituted latices for 30 minutes
at 150°C made the deposit water-insensitive and non-redispersible. The
reason is presumably that complete interparticle coalescence was induced
by this treatment. Likewise, drying the initial or reconstituted latex at an

elevated temperature resulted in the formation of hard, continuous deposits
which were water-insensitive.
3. Addition of small amounts (5-8% m/m on the dried polymer) of film-
coalescence aids, such as glycols, gave latices which dried to clear continuous
films at normal ambient temperatures. These films were water-sensitive but
not redispersible.
4. Incorporation of a chain transfer agent (tert-dodecyl mercaptan) or a
tetrafunctional monomer (divinylbenzene) in the emulsion polymerization
reaction system used to prepare the latices had almost no effect upon the
water-dispersibility characteristics of deposits from the latices.
Greene, Nelson and Keskey [72] have reported the preparation of redispersible
synthetic styrene-butadiene copolymer latices by copolymerization with
various mono-n-alkyl maleates (i.e. n-alkyl half-esters of maleic acid) (Structure
LIV) during emulsion polymerization. The esters used were methyl, ethyl,
H0 2 C.CH =CH.Co.OC n H 2n + 1
cis
LIV
n-propyl and n-butyl. Emulsion polymerization was carried out at 90·C using
potassium peroxodisulphate as initiator and a sodium alkylarylsulphonate as
colloid stabilizer. The total solids contents ofthe final latices were ca. 47% m/m.
The redispersibility of these latices, the tendency to which again increased as
the pH was raised, appears to be a consequence of several factors, of which
carboxylation of the dispersed polymer is clearly one. The extent of carboxylation
is greater than that indicated by the formulation for the emulsion polymerization
reaction, because hydrolysis of the half-esters tends to occur in aqueous media
either before or after copolymerization, thereby generating additional
carboxylic-acid groups. As noted in Section 10.3.4 of Chapter 10, the order of
decreasing tendency to hydrolysis is as follows:
methyl> ethyl> n-propyl > n-butyl
Table to.3 of Chapter 10 summarizes the properties oflatices of 45/50/5 m/m/m
styrenejbutadiene/monoalkyl maleate copolymers. Results are given for the
pH of the latex formed by emulsion polymerization, the average particle
diameter, and the distribution of carboxylic-acid groups. The variations in
latex properties are attributed mainly to the joint effect of increasing hydrophobicity
of the carboxylic-acid monomer and reduced tendency to hydrolysis, as the
size of the alkyl groups is increased. As the alkyl groups of the half-ester
increase in size, so the tendency for the carboxylic-acid monomer to become
buried in the particle increases, and the tendency for additional carboxylic-acid
groups to form by hydrolysis decreases. Both factors reduce the concentration
of carboxylic-acid groups at the particle surface. The surface concentration of
these groups is especially low for latices produced using the mono-n-butyl
References 387
maleate as the carboxylic-acid monomer. A further complication arises because,

as noted in section to.3.4 of Chapter to, the maleic acid formed by hydrolysis
tends to undergo a Diels-Alder reaction with unreacted butadiene, thereby
producing 1,2,3,6-tetrahydrophthalic acid (Structure XXV of Chapter to). This
substance is somewhat surface-active, and is believed to contribute significantly
to the redispersibility of the deposits from this type of latex. It is possible that
other water-soluble substances oflow molecular mass also enhance redispersibility.
The particle-size distributions for the reconstituted latices are similar to those
of the corresponding initial latex. This is illustrated in Fig. 11.16, which shows
hydrodynamic chromatograms for the initial and reconstituted latex obtained
using a 45/50/5 m/m/m styrenejbutadiene/monomethyl maleate monomer mixture.
Marker
Original latex
x= 258 nm
Marker
Reconstituted latex
x= 263 nm
Figure 11.16 Particle-size distributions. as determined by hydrodynamic chromatography. ror

initial and reconstituted latices prepared using a 45/50/5 mjrn/m styrenejbutadiene/monomethyl
maleate monomer mixture (Greene. Nelson and Keskey [72]).
REFERENCES
1. Gordon, M. and Taylor, 1. S. (1952) Journal of Applied Chemistry, 2, 493.
2. D'Ianni, 1. D., Hess, L. D. and Mast, W. C. (1951) Industrial and Engineering
Chemistry, 43, 319.
3. Hampton, R. R. (1949) Analytical Chemistry, 21, 923.
4. Smith, H. S., Werner, H. G., Madigan, J. C. and Howland, L. H. (1949) Industrial

and Engineering Chemistry, 41, 1584.
5. Tsailingold, V. L., Farberov, M. I., Epshtein, V. G., Uzina, R. V., Peizner, A. B.,
Boguslavski, D. B., Bugrova, G. A., Basin, V. E. and Shmurak, I. L. (1959) Soviet
Rubber Technology, 18(3), 6.
6. Frank, R. L., Adams, C. E., Blegen, J. R., Smith, P. V., Juve, A. E., Schroeder, C. H.
and Goff, M. M. (1948) Industrial and Engineering Chemistry, 40, 879.
7. Hofmann, W. (1964) Rubber Chemistry and Technology, 37(2[2]), 1.
8. Breon Nitrile Rubbers (1967) BP Chemicals (UK) Ltd, London.
9. Maynard, J. T. and Mochel, W. E. (1954) Journal of Polymer Science, 13, 251.
10. Neal, A. M. and Mayo, L. R. (1954) Neoprene, in Synthetic Rubber (eds
G. S. Whitby, C. C. Davies and R. F. Dunbrook), John Wiley, New York.
11. Morton, M., Cala,J. A. and Altier, M. A. (1956) Journal of Polymer Science, 19,549.
12. Morton, M. and Piirma, I. (1956) Journal of Polymer Science, 19,563.
13. Andersen, D. E. and Kovacic, P. (1955) Industrial and Engineering Chemistry, 70,171.
14. Cook, G. S. and Fitch, J. C. (1954) Rubber Chemistry and Technology, 27, 277.
15. Shepelev, M. I., Sandomirskii, D. M., Chernaya, V. V. and Tromfimovich, D. P.
(1962) Soviet Rubber Technology, 21(1), 17.
16. Carl, J. C. (1962) Neoprene Latex: Principles of Compounding and Processing, E.
I. du Pont de Nemours and Company (Inc.), Wilmington, Delaware, USA.
17. Shell Resins, VeoVa 10, VeoVa Technical Manual VV 1.1, 5th edn, undated.
18. El-Aasser, M. S. and Vanderhoff, J. W. (eds) (1981) Emulsion Polymerization of
Vinyl Acetate, Applied Science, London.
19. Elgood, B. G. and Gulbekian, E. V. (1973) British Polymer Journal, 17, 249.
20. Port, W. S., Hansen, J. E., Jordan, E. F., Dietz, T. J. and Swern, D. (1951) Journal
21. Warson, H. (1960) Paint Technology, 24(June), 19.
22. Lohr, G. (1979) Plastics and Rubber: Materials and Applications, 4, 141.
23. Shell Resins, Veo Va Copolymer Latex Preparation, VeoVa Technical Manual VV 2.1,
5th edn, undated.
24. Hartley, F. D. (1959) Journal of Polymer Science, 34, 397.
25. Bondy, C. and Coleman, M. M. (1970) Journal of the Oil and Colour Chemists
Association, 53, 555.
26. Reynolds, G. E. J. (1969) British Polymer Journal, 1, 233.
27. Reynolds, G. E. J. (1970) Journal ofthe Oil and Colour Chemists Association, 53, 399.
28. Edser, M. H. and Bulezuik, B. W. (1974) Polymers, Paint and Colour Journal,
4 December, p. 1051.
29. Farmer, D. B. and Edser, M. H. (1983) Chemistry and Industry, 21 March, p. 228.
30. Vinyl Products Limited (1975) Presentation to Paint Seminar, Spring 1975,
Pigment and Resin Technology, September, p. 11.
31. Lee, G. C. and Edser, M. H. (1973) Polymers, Paint and Colour Journal, 9 May,
p.478.
32. Davies, R. F. B. and Reynolds, G. E. J. (1968) Journal of Applied Polymer Science, 12,47.
33. Martin, P. S., Smith, 0. W. and Bassett, D. R. (1992) Alkali Stability of Vinyl Acetate
Copolymer and Terpolymer Latexes, in Polymer Latexes: Preparation, Characterization,
and Applications (eds E. S. Daniels, E. D. Sudol and M. S. El-Aasser), ACS Symposium
Series No. 492, American Chemical Society, Washington, DC, p. 203.
34. Peyser, P. (1989) in Polymer Handbook, 3rd edn (eds J. Brandrup and E. H.
Immergut), John Wiley, New York, pp. Vlj215-219.
35. Mast, W. c., Smith, L. T. and Fisher, C. H. (1945) Industrial and Engineering
Chemistry, 37, 365.
36. Mast, W. C. and Fisher, C. H. (1949) Industrial and Engineering Chemistry, 41, 790.
References 389
37. Zimmt, W. S. (1959) Journal of Applied Polymer Science, 1, 323.

38. Brodnyan, J. G., Cala, J. A., Konen, T. and Kelly, E. L. (1963) Journal of Colloid
Science, 18, 73.
39. McCurdy, K. G. and Laidler, K. J. (1964) Canadian Journal of Chemistry,
42,818.
40. McCurdy, K. G. and Laidler, K. J. (1964) Canadian Journal of Chemistry, 42, 825.
41. Fitch, R. M., Prenosil, M. S. and Sprick, K. J. (1969) Journal of Applied Polymer
Science: Part C, No. 27, 95.
42. Leonard, F., Cort, I. and Blevins, T. B. (1951) Industrial and Engineering Chemistry,
43,2500.
43. Fram, P., Szlachtun, A. J., DeFries, M. G. and Leonard, F. (1954) Industrial and
Engineering Chemistry, 46, 1992.
44. Fram, P., Szlachtun, A. 1. and Leonard, F. (1955) Industrial and Engineering
45. Bufkin, B. G. and Grawe, J. R. (1978) Journal of Coatings Technology, 50(641), 41.
46. Grawe,1. R. and Bufkin, B. G. (1978) Journal of Coatings Technology, 50(643), 67.
47. Bufkin, B. G. and Grawe, 1. R. (1978) Journal of Coatings Technology, 50(644), 83.
48. Grawe, J. R. and Bufkin, B. G. (1978) Journal of Coatings Technology, 50(645),71.
49. Bufkin, B. G. and Grawe, 1. R. (1978) Journal of Coatings Technology, 50(647), 65.
50. Grawe,1. R. and Bufkin, B. G. (1979) Journal of Coatings Technology, 51(649), 35.
51. Muroi, S. (1966) Journal of Applied Polymer Science, 10, 713.
52. Muroi, S., Hosoi, K. and Ishikawa, T. (1967) Journal of Applied Polymer Science,
11, 1963.
53. Verbrugge, C. J. (1970) Journal of Applied Polymer Science, 14,897.
54. Verbrugge, C. J. (1970) Journal of Applied Polymer Science, 14, 911.
55. Nishida, S., EI-Aasser, M. S., Klein, A. and Vanderhoff, J. W. (1981) Preparation
and Characterization of Alkali-swellable Carboxylated Latexes, in Emulsion
Polymers and Emulsion Polymerization (eds D. R. Bassett and A. E. Hamie1ec),
ACS Symposium Series No. 165, American Chemical Society, Washington, DC, p. 291.
56. Wessling, R. A. (1983) Cationic Polymer Colloids, in Science and Technology of
Polymer Colloids: Volume II: Characterization, Stabilization and Application
Properties (eds G. W. Poehlein, R. H. Ottewill and 1. W. Goodwin), NATO ASI
Series E: Applied Sciences No. 68, Martinus Nijhoff, The Hague, p. 393.
57. Howland, L. H., Neklutin, V. c., Brown, R. W. and Werner, H. G. (1952) Industrial
and Engineering Chemistry, 44, 762.
58. Vanderhoff, J. W., Park, J. M. and EI-Aasser, M. S. (1992) Preparation of Particles
for Microvoid Coatings by Seeded Emulsion Polymerization: Soft Hydrophilic
Polymer Core- Hard Hydrophobic Polymer Shell, in Polymer Latexes: Preparation,
Characterization, and Applications (eds E. S. Daniels, E. D. Sudol and M. S.
EI-Aasser), ACS Symposium Series No. 492, American Chemical Society, Washington,
DC, p. 272.
59. Okubo, M., Ichikawa, K. and Fujimura, M. (1992) Production of Multihollow
Polymer Particles by Stepwise Alkali-Acid Method: Alkali-treatment Process,
in Polymer Latexes: Preparation, Characterization, and Applications (eds E. S.
Daniels, E. D. Sudol and M. S. EI-Aasser), ACS Symposium Series No. 492,
American Chemical Society, Washington, DC, p. 282.
60. Kowalski, A. and Vogel, M./Rohm and Haas Company, European Patent No.
22,633,20 February, 1985.
61. Kowalski, A., Vogel, M. and Blankenship, R. M./Rohm and Haas Company,
European Patent No. 73,529, 23 September, 1987.
62. Barabas, E. S., Mallya, P. and Gromelski, S. J./GAF Corporation, United States
Patent No. 4,310,687, 12 January, 1982.
63. Yen, S. P. S., Rembaum, A., Molday, R. W. and Dreyer, W. (1976) Functional
Colloidal Particles for Immunosearch, in Emulsion Polymerization (eds I. Piirma and
J. L. Gardon), ACS Symposium Series No. 24, American Chemical Society,
64. Rembaum, A., Yen, S. P. S., Cheong, E., Wallace, S., Molday, R. S., Gordon, I.
L. and Dreyer, W. S. (1976) Macromolecules, 9, 328.
65. Armes, S. P., Miller, 1. F. and Vincent, B. (1987) Journal of Colloid and Interface
Science, 118,410.
66. Tarcha, P. 1., Misun, D., Finley, D., Wong, M. and Donovan, J. J. (1992) Synthesis,
Analysis and Immunodiagnostic Applications of Polypyrrole Latex and its
Derivatives, in Polymer Latexes: Preparation, Characterization, and Applications
(eds E. S. Daniels, E. D. Sudol and M. S. EI-Aasser), ACS Symposium Series No.
492, American Chemical Society, Washington, DC, p. 347.
67. Espiard, P., Revillon, A., Guyot, A. and Mark, J. E. (1992) Nucleation of Emulsion
Polymerization in the Presence of Small Particles, in Polymer Latexes: Preparation,
Characterization, and Applications (eds E. S. Daniels, E. D. Sudol and M. S.
E1-Aasser), ACS Symposium Series No. 492, American Chemical Society, Washington,
DC, p. 387.
68. Espiard, P., Mark, J. E. and Guyot, A. (1990) Polymer Bulletin, 24, 173.
69. Floria, V. D. and Mock, R. A. (1956) Paper presented at the 129th Meeting of
the American Chemical Society, Dallas, Texas, USA, 1956.
70. Pateman, R. A. W./The British Oxygen Company Ltd., British Patent No. 851,535,
19 October, 1960.
71. Guziak, L. F. and Maclay, W. N. (1963) Journal of Applied Polymer Science, 7, 2249.
72. Greene, B. W., Nelson, A. R. and Keskey, W. H. (1980) Journal of PhYSical
12
Artificiallatices
12.1 INTRODUCTION
Artificial latices are produced by the dispersion or redispersion of solid
polymers in an appropriate medium. In the present state of latex technology,
the dispersion medium is exclusively aqueous. Artificial latices have been of
some industrial interest for many years. Notwithstanding this, they have
never attained the level of industrial application which has been enjoyed by
natural and synthetic latices. As will appear subsequently, artificial latices
are not easy to prepare. Attempts are made to produce them only if the
polymer cannot be conveniently produced as a synthetic latex, or if it is not
available as a natural latex. If latices of such polymers are required, then
they must be produced by dispersion of the bulk polymer. As a generalization,
the problems associated with the production and application of artificial
latices are tolerable only if the dispersed polymer offers some special
properties, and only if it is particularly advantageous to manufacture the
desired product from a latex rather than from the polymer in bulk. Quite
apart from the difficulties associated with their production and utilization,
artificial latices suffer from two important fundamental disadvantages relative
to synthetic latices: the methods of preparation available at present are such
that it is not possible to produce artificial latices which contain structured
particles; and it is not possible conveniently to produce functionalized
artificial latices, unless the initial bulk polymer is already functionalized.
The oldest artificial latices of any industrial interest were dispersions of
reclaimed natural rubber in aqueous media. The usual processes for the
reclaiming of waste vulcanized rubber give the product in bulk form. The
latter must therefore be dispersed in a suitable medium if a latex is required.
Artificial latices of reclaimed rubber have been widely used in adhesive
formulations (see Chapter 22 (Volume 3», especially in latex bonding agents
for promoting rubber-to-textile adhesion (see section 22.5 of Chapter 22
(Volume 3)).
Of more recent industrial interest have been artificial latices of various
synthetic rubbery polymers which are produced by processes other than

392 Artificial latices
emulsion polymerization. These processes include ionic solution polymerization,

heterogeneous polymerization in non-aqueous media, and condensation/chain-
extension polymerization. Serious attempts have been made to obtain
industrially-useful artificial latices from several types of polymer produced
by such processes. These polymer types include copolymers (commonly
known as butyl rubbers) of isobutene (isobutylene) with minor amounts of
isoprene, synthetic stereo regular cis-l,4-polybutadienes and cis-l,4-polyisop-
renes, and polymers in the classes generally known as polyurethanes and
silicones. There has also been some industrial interest in artificial latices of
rubbery polymers which are obtained by chemically modifying other synthetic
polymers not produced by emulsion polymerization. One example of such
a polymer is chlorosulphonated polyethylene. Apart from latices of polymers
of the polyurethane and silicone types, artificial latices of synthetic polymers
have failed to attain significant industrial application. Artificial latices of
isobutene-isoprene copolymers were originally developed primarily for
application in the manufacture of tyres from isobutene-isoprene rubbers.
They were required as the basis for tyre-cord treatments to ensure adequate
adhesion between the cords and the surrounding rubber matrix. Other
applications which have been claimed for these latices include use as a binder
for non-woven fabrics, use as a basis for certain specialized types of surface
coating, and use in connection with paper treatments. There has also been
interest in producing gloves from this type of rubber by latex dipping; such
gloves have the desirable attributes of good chemical resistance and low
permeability to gases. However, the production problems have been such
that, as far as is known, no isobutene-isoprene rubber gloves are made by
this method at the time of writing. Artificial latices of stereo regular polydiene
rubbers were developed primarily as competitors for natural rubber latices
and styrene-butadiene rubber latices in the manufacture oflatex foam rubber.
Artificial latices of chlorosulphonated polyethylene have been considered for
the manufacture by latex dipping of gloves having the good chemical
resistance which is an attribute of this type of rubber. However, problems
associated with the production and use of these latices appear to have largely
eliminated them from serious consideration. Artificial latices containing
various polyurethane polymers have been of some industrial interest for
several years now for applications such as textile coating, leather treatments,
paper coating, sizing of glass fibres, and the manufacture of products by dipping.
Closely related to the matter of artificial latices is that of water-redispersible
latex polymers. These are polymers which are obtained as latices in the first
instance, and are capable of being dried down to give powders which can
subsequently be reconstituted to latices by admixture with water. Interest
has been shown in producing latices of this type from both natural and
synthetic rubber latices. They are considered in the chapters which deal
respectively with these type of latex (see sections 9.4.7 of Chapter 9 and 11.9.5
of Chapter 11 respectively).
Effect of latex particle size upon creaming or sedimentation 393
12.2 EFFECT OF LATEX PARTICLE SIZE UPON RATE OF

CREAMING OR SEDIMENTATION
It is difficult to achieve small particle size when attempts are made to form
stable colloidal dispersions of bulk polymers in aqueous media. It therefore
follows that problems of creaming or sedimentation are likely to be
encountered if there is an appreciable difference between the density of the
polymer and that of the aqueous phase in which it is dispersed. Although
this is a general problem which may be encountered with all types of latex,
it can be especially acute with artificial latices, because their particles tend
to be relatively large. The elementary theory of latex creaming is expounded
in section 9.4.3.1 of Chapter 9 in connection with the concentration of natural
rubber latex by creaming. As noted there, the tendency to creaming or
sedimentation is opposed by the random Brownian motion of the latex
particles. It is also opposed by any convection currents which arise from
temperature gradients within the latex. A rough criterion which has sometimes
been used is that the tendency for density difference to effect creaming or
sedimentation will be nullified by Brownian motion and convection currents
if the size of the latex particles is such that the rate of creaming or
sedimentation under the influence of density difference is less than ca.
1 mm (24 hours) - I, i.e., less than ca. 1 x 10- 8 m s - I. Equation (9.3) of Chapter
9 can be rearranged to calculate the maximum particle size which meets this
criterion, if the density difference between the particles and the dispersion
medium is known, if the viscosity of the dispersion medium is known, and
if the particles are assumed to be spherical. Equation (9.3) gives
-2" dh
x=3 (12.1 )
(p - O')g dt
where x is the particle diameter, dh/dt is the rate of movement of particles
upwards through the dispersion medium, p and 0' are the densities of the
latex particles and dispersion medium respectively, " is the viscosity of the
dispersion medium, and 9 is the gravitational acceleration. Figure 12.1 shows
predictions given by this equation for the variation of x with (p - 0') for
dh/dt = 1 mm (24 hours)-I, and for" = 1,10,50 and l00mPas." = 1 mPas
corresponds to water at ca. 20·e. Also shown in this diagram are the values
of (p - 0') for a few common polymers dispersed in a medium of density
1 Mg m - 3. The curve for" = 1 mPa s shows that, for dispersions of polymers
in media having approximately the viscosity of water:
1. the maximum particle diameter which meets the criterion dh/dt:::;;
1 mm (24 hours) - 1 is sensitive to the value of (p - 0') when the latter is
small relative to p and 0' separately;
2. the maximum particle diameter which meets the criterion dh/dt:::;;
1 mm (24 hours)-I is insensitive to the value of (p - 0') when the latter
exceeds ca. 0.04 Mg m - 3;
A ....... A
>c
o
-0.12 -0.08 -0.04 o 0.08 0.12
(p-a) / Mg m-3
= ,,= ,,= =
Figure 12.1 Predictions given by equation (12.1) for variation of x with (p - u) for dh/dt =
I mm (24 hours)-t, and for" I (curve A), 10 (curve B~ 50 (curve q and" 100
(curve 0) mPa s. Also shown are the values of (p - u) for the following common polymers for
u = 1.000 Mgm- 3 : I 44/42/14 cis-I,4- /trans-I,2- /1,2 polybutadiene; " cis-I,4-polyisoprene; III
isobutene-isoprene rubber; IV cis-I,4-polybutadiene; V atactic polystyrene.
3. the densities of some common polymers which it might be desired to

obtain in the form of artificial aqueous latices are such that x is relatively
insensitive to (p - u), being ca. 0.5 ~m.
Figure 12.1 also shows that, for any particular value of (p - u), the value of
x given by equation (12.1) is considerably increased if" is increased. However,
it is doubtful to what extent increasing the viscosity of the dispersion medium
is effective in nullifying the effect of large particle size in respect of rate of
creaming or sedimentation, because increased " will also reduce the speed
of the Brownian motion and of any convection currents in the latex.
12.3 METHODS OF PRODUCING ARTIFICIAL LATICES

Many of the processes which have been developed for the production of
artificial latices based upon aqueous media exploit one of the following four
fundamentally distinct techniques:
Methods of producing artificial latices 395
1. In the first of these, usually referred to as the solution-emulsification

technique, the polymer is dissolved in, or is swollen by, a volatile solvent
to form a solution or cement. The polymer solution or cement is then
emulsified in water with the aid of one or more surface-active substances
which function as a dispersing agent, as a colloid stabilizer for the emulsion
which is formed, and also as a colloid stabilizer for the eventual latex. To
the extent that the surface-active substances are effective in assisting the
formation and stabilization of the emulsion of polymer solution or cement
in water, they are appropriately described as an emulsifier. The final step
is removal of the solvent from the emulsion, usually by vacuum or steam
distillation. A dispersion of the polymer in an aqueous medium should
be left behind.
2. The second type of process is appropriately referred to as the phase-
inversion technique. Increasing amounts of an aqueous phase are dispersed
into the polymer. Initially, a dispersion of aqueous medium in polymer
forms. As more aqueous phase is incorporated, so it becomes increasingly
likely that an inversion of the phases will occur, thereby creating the
desired dispersion of the polymer in an aqueous medium. It is necessary
to incorporate in one or both of the two phases surface-active substances
which are able to function as a dispersing agent and as a stabilizer for
the dispersions which are formed, i.e., for both the water-in-polymer
dispersion which forms initially, and the polymer-in-water dispersion
which forms subsequently. The polymer is usually rubbery in character,
and so both the mixing with a surface-active agent and the progressive
incorporation of the aqueous phase can be effected using conventional
rubber-mixing equipment.
3. The third type of process, appropriately called the mutual solvent technique,
is applicable only to more polar types of polymers. The polymer is dissolved
in a water-miscible solvent. This solution is then mixed with water to give
a three-component system comprising a dispersion of the polymer in an
aqueous phase with the polymer solvent distributed between the two
phases. The final stage of the process is the removal of the polymer solvent
by evaporation. Unless the polymer is effectively self-dispersible in water,
it is necessary to include a colloid stabilizer in the components used to
form the dispersion.
4. In the fourth type of process, appropriately called the self-dispersion
technique, the polymer is functionalized with ionizable groups, which may
be either acidic or basic. The polymer is then dispersed in an aqueous
phase which contains an alkali or an acid, according as the polymer is
functionalized with acidic or basic groups. The latex which forms is
colloidally stabilized by the ionic groups which form as the functional
groups ofthe polymer interact with the alkali or acid in the aqueous phase.
Many variants of the above processes are possible. In particular, in the

case of processes of types (1), (2) and (3), it is possible to form the dispersing
agent and colloid stabilizer concurrently with dispersion by the so-called
soap in situ technique for the preparation of oil-water emulsions (see section
15.4.2 of Chapter 15 (Volume 3)), rather than by adding a pre-formed
surface-active substance to the system. The procedure is to mix either the
acid or the base component of an ionic surface-active substance with the
non-aqueous phase, and the other component with the aqueous phase. The
intention is that the ionic surface-active substance should be formed at the
expanding interface between the two phases, this being where the two
components of the substance come into contact and are able to interact. In
a typical implementation of the phase-inversion technique, the ionic surface-active
substance is a fatty-acid soap such as potassium oleate. The fatty-acid
component is incorporated in a rubbery polymer using conventional
rubber-mixing equipment, and the alkali component is dissolved in the
aqueous phase. When the acid component is thoroughly dispersed in the
rubber, the dilute aqueous solution of the alkali component of the ionic
surface-active substance is slowly worked into the mixture, again using
conventional rubber-mixing equipment. In principle, this procedure is applicable
to many types of ionic surface-active substances.
It will be evident from the foregoing outlines that these methods are
inherently inconvenient for large-scale implementation. This in part accounts
for the failure of artificial latices to attain widespread industrial application.
There are at least two important factors relating to the nature of the product
which have further inhibited industrial application of artificial latices. These
are as follows:
I. For reasons given below, the ratio of colloid stabilizer to polymer is almost
always considerably higher in artificial latices than in natural rubber and
synthetic latices, notwithstanding that attempts have been made to reduce
the level of colloid stabilizer after the latex has been formed, e.g., by
dilution and centrifugation.
2. Notwithstanding the high ratio of surface-active substances to polymer
which is characteristic of artificial latices, the average particle size obtained
is usually considerably greater than that of natural rubber and synthetic
latices. This leads to problems of creaming or sedimentation, as well as
to possible processing problems. The matter of the maximum particle size
which can be tolerated if serious problems of creaming or sedimentation
are to be avoided is considered in section 12.2 above.
There are two principal reasons why the ratio of colloid stabilizer to
polymer is high for artificial latices produced by the solution-emulsification
technique. The first is that the polymer is initially diluted by admixture with
a solvent to form a solution or cement. In some cases, it may be desirable
that the polymer concentration should be no more than, say 10% m/v, or
even less. Initially it is necessary to produce a stable colloidal dispersion of
this solution or cement, and the ratio of surface-active substances to polymer
solution must be such as to make this possible. When the solvent is eventually
removed, the ratio of surface-active substances to polymer is higher than
that of stabilizer to solution by a factor equal to the ratio of solvent to
polymer in the polymer solution which was emulsified. The second reason
is that, if the solvent is removed by steam-distillation, then it is necessary
that the initial emulsion should be extremely stable colloidally in order to
withstand the intense destabilizative conditions to which it is subjected during
steam-distillation. For artificial latices produced by the phase-inversion
technique, the intense mechanical shearing to which the developing latex is
subjected necessitates high colloid stability. This implies that the ratio of
surface-active substances to polymer must be high. Furthermore, if dispersion
is facilitated by an ionic colloid stabilizer formen by the soap in situ technique,
then the concentration of both in the system as a whole may be high because
the two components of the surface-active substance are initially dissolved in
the two phases. Regardless of the method used to produce the latex, a high
ratio of surface-active substances to polymer gives rise to several problems
in the manufacture of products from the latex. The colloid stability of the
latex is often unduly high. Films from the latex may be difficult to dry
satisfactorily under industrial production conditions. Even if satisfactory dry
films are obtained, they will have a marked tendency to absorb water because
of the high level of hydrophilic substances which remain. The problem of
drying films from such latices may be exacerbated if, as for latices ofisobutene-
isoprene rubbers, the dispersed polymer has low permeability to water-
vapour.
Having outlined the various methods which are available for the production
of artificial latices, and the problems associated with these methods and the
products obtained, we turn now to describe a selection of specific processes
which have been proposed.
12.3.2 A selection of processes which have been proposed for the production
of artificial latices
12.3.2.1 Early processes for the production ofartificial latices ofnatural rubber
The earliest published process for the production of an artificial latex of
natural rubber appears to have been that of Sievier [1]. Solid rubber was
immersed in dilute ammonia solution for a long time, and the solution then
evaporated to give a dispersion of rubber particles in the aqueous medium.
A later attempt was made by Alexander [2,3]. His process consisted in
dissolving the rubber in a solvent, heating the solution with sodium hydroxide,
and then removing the solvent by distillation. It was claimed that a viscous
mass ofrubber was produced which could be diluted with water. Later again,
the Traun process [4] was developed as a means of utilizing waste vulcanized
rubber. The waste was ground to a fine crumb which was then slurried with
an aqueous medium and subjected to intensive mechanical action in a

high-speed colloid mill.
12.3.2.2 The Pratt processes and modifications thereof

The first practicable process for the redispersion of natural rubber, gutta
percha and balata in aqueous media appears to have been that disclosed by
Pratt in 1924 [5]. Although the process is claimed to be applicable to gutta
percha and balata, as well as to natural rubber, it was primarily of interest
in connection with the latter polymer. In what follows, reference is made
specifically to natural rubber only. The claim is made by Pratt that, by means
of his process, it is possible, without significant depolymerization, to 'disperse
in water crude rubber so that the particles thereof are substantially the same
shape and the same order of magnitude as the original rubber particles in
the latex; in which case the disperse particles may be recoagulated in the
form of crude rubber, vulcanized and otherwise treated in precisely the same
way and with the same results that the original coagulated or crude rubber
could be treated'. The Pratt method uses a combination of the solution-
emulsification and phase-inversion techniques. The rubber is dissolved in a
volatile solvent. Water is then gradually added to the solution with stirring
until an inversion of phase occurs. The process is facilitated by the inclusion
of a saponifiable substance, such as oleic acid, in the rubber solution, and
an alkali, such as ammonia, in the water, thereby producing an emulsifier in
situ. Initially, a dispersion of aqueous phase in rubber solution is formed;
but after the addition of sufficient aqueous phase, the rubber solution becomes
the disperse phase. An important feature of the process is that dispersion of
the rubber solution is accomplished without the application of heat, although
continuous agitation is necessary. The final stage of the process is removal
of the volatile solvent by evaporation, and its recovery for re-use. In principle,
a large number of saponifiable animal and vegetable oils can be used in this
process, but Pratt states that excellent results can be achieved using oleic
acid as this component. In particular, he found the combination of oleic acid
and ammonia to be especially effective. He further claims that it is possible
to form the soap in the aqueous phase first, and then to add the soap solution
to the rubber solution, but states that the degree of dispersion thereby attained
is not as great as that obtained by the soap in situ method.
In an example of the Pratt [5] process, 250 g of natural rubber and 20 g
of oleic acid are dissolved in 750 g of benzene at normal ambient temperature.
20 g of 26° Beaume ammonia solution (the concentration being specified by
way of the solution density as determined by the Beaume hydrometric
method) is dissolved in 750 g of water, and the mixture added slowly to the
rubber solution. As the ammoniated water is gradually added to, and stirred
into, the rubber solution, the water first forms the disperse phase of the
emulsion. But as the total volume of added water increases, a change of
phase occurs, and the water then forms the continuous phase. The final
emulsion is a milky white liquid which can be diluted to any reasonable
extent with water. It may be used as thus produced, or the solvent may be
removed by evaporation. This is preferably accomplished in a vacuum still
at a low temperature, not exceeding 50·C. The resultant latex can be
coacervated by addition of acetic acid or other chemical coacervant, or by
heating, or by vigorous agitation, or by excessive dilution.
In an improved version of this process developed by Pratt [6], disclosed
in 1925, the use of volatile solvents was avoided. Pratt was aware that such
solvents are costly and often flammable, that they must be recovered, and
that care must be exercised during the recovery to avoid colloid destabilization
of the latex. In this second process, water is incorporated into the mass of
the rubber, and the resulting mixture then dispersed in water. Pratt claimed
that pure water alone can be used successfully, because of the presence of
proteinaceous and other non-rubber substances in the bulk rubber. However,
in order to accelerate the dispersion process, it is helpful to dissolve various
hydrocolloids in the water. Suitable hydrocolloids include glue, albumin,
casein, starches, and various water-soluble gums. Pratt considered that such
substances function by assisting the water to penetrate the interparticle spaces.
The current interpretation of their function is, of course, rather different. The
initial absorption of water can be accomplished by mechanical working on
a rubber-mixing mill, or by using a dough mixer. The final dispersion of the
rubber is best carried out using a dough mixer. Again, it seems that faulty
reasoning led to the development of a workable process. Pratt envisaged
that the shearing and stretching of the rubber would facilitate the initial
absorption of water, because such mechanical deformation would tend to
separate the particles, thereby facilitating penetration by the water.
A typical example of this second Pratt [6] process is as follows: 100 parts
by mass of natural rubber are placed in a mixing mill and milled until warm
and plastic. 5 pphr of glue is then added, either in dry granular form or in
the form of a thick paste made by heating the glue with a small quantity of
water. The rubber-glue blend is then transferred to a two-blade dough mixer,
and sufficient water added to bring the total water content up to ca. 10% m/m
on the rubber. The mixer is then covered and allowed to run until all the
water has become absorbed by the rubber. An aqueous solution comprising
5 pphr of saponin and 10 pphr of water is then added slowly to the rubber
mass until the latter has the consistency of putty. The remainder of the
saponin solution is gradually added, and the entire mass then diluted with
a further 45.5 pphr of hot water. Mixing is continued until the water is
thoroughly incorporated throughout the mass, at which stage it should be
found that a smooth paste is produced in which the rubber is dispersed in
the form of globules. This dispersion may be stored as such, or else diluted
with water to any desired consistency.
It was the view of Pratt [6] that the original particles of a latex such as
that of natural rubber preserve their identity after coacervation to form solid
rubber, and possibly after vulcanization as well. He envisaged that, during
the coacervation process, the particles adhere to each other mechanically,
thus forming a coherent mass, but nevertheless retain their identity sufficiently
to be separated and redispersed. To effect redispersion in his second process,
Pratt thought it merely necessary to introduce water into the interstices which
he imagined to exist between the rubber particles, and to restore the functions
of the protective coatings which formerly surrounded the particles. The
particles would then be restored to their original shapes, and, because of
reduction of the 'interfacial tension' between them, physical separation would
follow. Whilst few today would subscribe to this view of the nature of dried
coacervated natural rubber, or of the mechanism by which redispersion
occurs, the efficacy of the process to which these considerations led Pratt is
not in doubt.
According to Pratt [6], microscopic examination of the dispersions
obtained from solid natural rubber using the second of the techniques he
describes showed the particles to resemble those in natural rubber latex. It
should be noted that, although in his second type of process the requirement
to use a volatile solvent is eliminated, nevertheless the rubber still has to be
subjected to mechanical shearing (mastication). Whilst neither he nor his
contemporaries are likely to have recognized that this caused depolymerization
of the rubber, this will probably have occurred to some extent. Account must
be taken of this observation in evaluating the claim made in the patent
covering the first of the processes that no depolymerization of the rubber
occurs when it is subjected to this redispersion procedure.
A significant difference between the two Pratt [5,6] processes is that,
whereas the colloid stabilizer is formed in situ in the first process, it is added
as such to the rubber or to the water in the second. Klein and Szegvari [7]
subsequently discovered that a greatly improved product could be obtained
by the second of the Pratt processes if the colloid stabilizer was again formed
in situ. On the basis of this distinction, Klein and Szegvari obtained a patent
for a further non-solvent process. In a typical example of their method, the
rubber is compounded with oleic acid, and a dilute aqueous ammonia solution
then gradually introduced into the mass. Being a liquid miscible with natural
rubber, oleic acid has the additional effect of plasticizing the rubber when
mixed with it. For this reason alone, incorporation of the ammonia solution
into the mass of rubber is facilitated.
In a further modification, described by Madge [8], water, not dilute
aqueous alkali solution, is added to the rubber until the rubber content of
the mixture is reduced to ca. 75% m/m. The soap is then formed by addition
of a small amount of a concentrated aqueous solution of an alkali such as
potassium hydroxide. The initial dispersion ofthe water in the rubber should
be as fine as possible. The batch should be cold-worked as far as possible.
Phase inversion, which occurs quite sharply, is achieved by heating the
mixture. It is accompanied by a transition from a soft, elastic, coherent mass

to a stiff aqueous paste. The mixture should be diluted somewhat before
being cooled down, otherwise a second phase inversion to a dispersion of
aqueous medium in rubber may occur on cooling.
12.3.2.3 The Trumbul/- Dickson process

In the preferred practice of this process, disclosed by Trumbull and Dickson
[9] in 1924, a hydrocolloid is first mixed with water to provide a high-viscosity
solution or paste which is placed upon a tight rubber-mixing mill. Softened
rubber is then added in small pieces to the bank of hydrocolloid paste, and
is gradually worked into the viscous paste. In order to avoid phase inversion
in the mixture which forms on the rolls, the rubber is added slowly, small
quantities of water being added frequently to allow for that which is lost by
evaporation. Suitable hydrocolloids include glue, casein, rosin soaps and
various gums. The rubber can be softened either by mechanical working
(mastication) or by the addition of plasticizers. By this method, it is claimed
to be possible to produce dispersions having total solids contents of up to
80% mlm or higher. It is further claimed that, by the addition of water to
the putty-like mass which eventually forms upon the mill, and in which all
the rubber has been dispersed into very small particles, it is possible to
prepare colloidal dispersions which closely resemble natural rubber latex in
their physical properties. The addition of small amounts of alkaline substances,
such as sodium carbonate, sodium borate or ammonium hydroxide, to the
aqueous phase was found to be advisable in certain cases. Such substances
increased the colloid stability of the resultant latex. It is also stated that, if
excess rubber is added, either at the end of the mixing operation or before
the ultimate high concentration of rubber has been attained, then phase
inversion may occur. A dispersion in which this has taken place differs from
the normal type of dispersion in being difficult to wet with water, in being
more viscous, and in being confined to one roll of the mill instead of two.
A phase-inverted batch can be inverted to the desired dispersion of rubber
in an aqueous phase by gradually working into it an aqueous paste of
hydrocolloid which is contained as a thin bank on a rubber-mixing mill.
12.3.2.4 The Murphy-Madge process

In the patent which covers this process [10], it is stated that all the hitherto
published processes for the production of satisfactory artificial latices required
the incorporation of quantities of soaps and protective colloids far in excess
of the amounts present in natural rubber latex. Attempts to reduce the levels
of such substances during the formation of the artificial latex may lead to
phase inversion, or to the formation of dispersions of doubtful colloid stability.
However, a number of important disadvantages accrue from the presence of
high levels of hydrophilic substances of these types. Coagula obtained from
the latices are weak and friable, and also readily absorb water. If the latices
are to be made heat-sensitive, then high levels of heat-sensitizing agents are
necessary. The excessive levels of colloid stabilizers also interfere with certain
other latex-manufacturing processes, notably those for latex foam rubber
which were being developed in the 1930s. The distinguishing feature of the
Murphy-Madge process for the production of artificial latices is that, in
order to overcome these disadvantages, a substantial proportion of the added
dispersing agents and colloid stabilizers is removed by various means, such
as by centrifuging or by creaming. In a typical example, an artificial latex
containing 6 pphr of oleic acid, 2 pphr of potassium hydroxide and 4 pphr
of glue, made from masticated pale crepe natural rubber, is diluted to
20% m/m total solids content using 1% m/v potassium oleate solution. It is
necessary to use dilute soap solution rather than water for this dilution,
because otherwise the latex tends to be colloidally destabilized. After dilution,
the latex is passed through a centrifuge to give a cream having 55% m/m
total solids content. In a second example, the artificial latex is first diluted
to 20% m/m total solids content as before. 1% m/m of sodium alginate on
the mass of the diluted latex is then added, and the mixture allowed to stand.
Separation into three layers occurs as follows: (I) a top layer which contains
cream having total solids content ca. 50% m/m; (2) an intermediate layer
which is partially creamed; and (3) a bottom layer which is clear. In a further
procedure covered by this patent, a proportion of the colloid stabilizer in
the latex is rendered ineffective by treating the artificial latex with a substance,
such as formaldehyde or trioxymethylene, which hardens and insolubilizes
proteinaceous substances, indigenous or added, present in the latex. It is
claimed that, by treatments such as these, artificial latices can be modified
so that their properties conform closely to those of natural rubber latex.
Coagula from the modified latices are said to be strong and coherent.
Heat-sensitivity can be induced by the addition of normal levels of
heat-sensitizers. Processes such as the manufacture of latex foam rubber can
be carried out successfully.
12.3.2.5 The Hunter-Segura-Small-Miller process for artificial latices of

isobutene-isoprene rubbers
Consideration is now given to three more recent processes which have been
proposed for the production of artificial latices. These are all intended for
the production of artificial latices of synthetic polymers which so far cannot
be conveniently produced in latex form by emulsion polymerization. There
has been almost no industrial interest in artificial latices of natural rubber
for many years now, partly for economic reasons, and partly because the
product is not well-suited to conventional latex manufacturing processes, for
reasons given above.
The first of the processes to be considered was disclosed in a patent to

Hunter et al. [11] published in 1960. It is a solution-emulsification process
intended to be used for the production of artificial latices ofisobutene-isoprene
rubbers. The emulsifiers comprise an alkali-metal soap of a fatty acid
containing between 12 and 24 carbon atoms per molecule, polyvinyl alcohol,
and a 7-12-mole ethoxylate of an alkylphenol for which the alkyl group is
octyl or nony!. A wide range of organic solvents can be used to dissolve the
rubber. After emulsification, the solvent is removed by stripping under
vacuum at a temperature above 38°C. If it is desired to remove some water
as well, so as to obtain a more concentrated latex, then the temperature
should be raised to ca. 88°C. Some agitation of the emulsion is necessary
during stripping, but this should be as mild as possible to minimize the
mechanical destabilizative influence to which the latex is subjected during
the stage when its colloid stability is at its lowest. In a typical example,
3.4 pphr of oleic acid is added to a 23% mjv solution of the rubber in hexane.
This solution is then emulsified in an equal volume of an aqueous phase
which contains 3.1 pphr of an 8-IO-mole octylphenol ethoxylate, 2.7 pphr of
partially-hydrolysed polyvinyl acetate and 0.7 pphr of potassium hydroxide.
The hexane and some water are then removed by stripping. It is claimed
that no foaming or other difficulties are experienced, and that the product
is an excellent latex. The principal emulsifier in this system is the soap formed
by interaction between the oleic acid and the potassium hydroxide. The
partially-hydrolysed polyvinyl acetate is said to function as a protective
colloid during the emulsification process, and also as a stabilizer for the
system during and after stripping. The alkylphenol ethoxylate is said to be
effective in suppressing foam formation, and also in stabilizing the final latex.
It is advantageous to use slightly more alkali than the equivalent of the fatty
acid, to enhance the colloid stability of the system; too great an excess should
be avoided, because it leads to foaming. It is advised that the pH of the
aqueous phase be maintained in the range 10-11.
Hunter, Small and Miller [12] have described a continuous process for
stripping the solvent from raw isobutene-isoprene rubber artificial latices
made by the solution-emulsification process. To prevent colloidal destabilization
of the latex, the stripping column is first charged with a dilute aqueous
solution of a surface-active substance such as potassium oleate or ammonium
n-dodecyl sulphate, and the temperature raised to a level sufficient subsequently
to strip the solvent from the raw isobutene-isoprene rubber latex. The raw
latex is then added to the column and stripped until it is essentially free of
solvent, whereupon it is removed from the column and fresh raw latex added
at the same rate. In an example, the stripper was first charged with 1.5% mjm
aqueous potassium oleate solution, the temperature raised to ca. 90°C, and
a raw isobutene-isoprene rubber latex containing potassium oleate as
emulsifier and hexane as solvent added over a period of 2 hours. Raw latex
was then pumped continuously into the stripper, the temperature being
maintained at ca. 88·C and the pressure atmospheric. Hexane-free finished

latex was removed from the stripper at such a rate that the mean residence
time was 3 hours.
12.3.2.6 The Beerbower-Burton-Malloy process for artificial latices of

isobutene-isoprene rubbers and ethylene-propylene-diene rubbers
This process, which is disclosed in a patent to Beerbower, Burton and Malloy
[13] published in 1978, is intended for the production, by a solution-
emulsification procedure, of artificial latices of isobutene-isoprene rubbers
and of ethylene-propylene-diene rubbers. The finished latices are said to
have total solids contents in the range 50-68% mlm, and average particle
diameters of less than 1 ~m. They are claimed to be mechanically stable, to
have a viscosity in the range 0.5-2.0 Pa s, and to be coagulatable. The process
is distinguished by the type of emulsifier system used. This comprises a
mixture of two components: an ammonium or alkali-metal soap of one or
more fatty acids which contain between 12 and 24 carbon atoms per molecule;
and an ionic hydrocolloid which has number-average molecular mass in the
range 500- 3500 daltons, and which is an ammonium, alkali-metal or amine
salt of a polycarboxylic acid or acid anhydride obtained by adding an a-olefin
polymer across the olefinic double bond of an a, p-olefinically-unsaturated
aliphatic mono- or dicarboxylic acid or anhydride. It is claimed that the
second component of the emulsifier system confers improved mechanical
stability on both the raw and finished latex, without undesirable increase in
viscosity and without affecting the ease of coagulation. The nature and method
of preparation of the hydrocolloids used as the second component are
described in two patents to Le Suer and Norman [14, 15]. The a-olefin
polymer is first halogenated by reaction with a reagent such as chlorine,
bromine, iodine, N-bromosuccinamide or N-chlorophthalimide. The method
by which halogenation is effected is said not to be critical. The halogenated
a-olefin polymer or copolymer is then reacted with the a, p-olefinically-
unsaturated aliphatic mono- or dicarboxylic acid or anhydride under
conditions such that addition of the polymer across the double bond of the
latter occurs. The mechanism of this reaction appears to be uncertain, as
does the structure of the product, but it seems clear that the feasibility of the
reaction depends upon the presence of the halogen atoms in the modified
a-olefin polymer. If the reaction is indeed a straightforward addition across
the double bond of the acidic compound, then the product must contain
halogen. It is said that this reaction can be effected by merely mixing the
reactants and heating the mixture to 80·C or above, the preferred temperature
being within the range 120-250·C. The ratio of reactants should be such
that at least one equivalent of a,p-olefinically-unsaturated aliphatic acidic
compound is used for each equivalent of halogenated polymer. It is
advantageous to effect the reaction in an inert solvent such as mineral oil,
cyclohexane, chlorobenzene or toluene. The final hydrocolloid-forming step

is neutralization of the polycarboxylic acid or acid anhydride with an
appropriate alkali. Thus the formation of the hydrocolloid from the IX-olefin
polymer can be represented by the following series of reactions, in which the
IX-olefin polymer is taken as polyisobutene, the lX,p-olefinically-unsaturated
aliphatic acidic compound as maleic acid, and the alkali as potassium
hydroxide. The reaction between the IX-olefin polymer and the IX,P-
olefinically-unsaturated aliphatic acidic compound is assumed to involve
straightforward addition across the double bond of the latter:
el 2
... -CH 2-C(CH 3)2- ... -------. ... -CHCI-C(CH 3h-···
... -CHCI-C(CH3h- ... + CH.CO-------. ... -CH-C(CH3h-···
II >0 I
CH.CO CH.CO
I >0
CICH.CO
... -CH-C(CH3h-··· ... -CH-C(CH3h-···

I KOH
I
CH.CO -------. CH.CO;K+
I >0 I
CICH.CO CICH.CO;K+
The products obtained by reactions of this type are named as salts of
polyolefin derivatives of the saturated acid corresponding to the IX,P-
olefinically-unsaturated aliphatic acidic compound used in the preparation.
Thus the product of the above example is described as potassium
polyisohutenylsuccinate. The molecular mass of the product is determined
primarily by that of the initial IX-olefin polymer.
Preferably, the emulsifier system comprises approximately equal proportions
by mass of the two components. The preferred hydrocolloid component of
the emulsifier system is the potassium salt of either polyisobutenylsuccinic
acid or a polyisobutenylpropionic acid having number-average molecular
mass in the range 850-1500 daltons. It is understandable that a polyiso-
butenylcarboxylate salt is preferred for the preparation of artificial latices of
isobutene-isoprene rubbers; the polyisobutenyl moiety of the hydrocolloid
component should bind effectively to the rubber particle, because the repeat
units in this moiety are chemically similar to those of the rubber molecule.
For the same reason, it is expected that the preferred hydrophobic moiety
for the hydrocolloid component of the emulsifier system for preparation of
artificial latices of ethylene-propylene-diene rubbers would be polypropenyl.
The mechanism of colloid stabilization in the artificial latices obtained by
this process is presumably predominantly electrostatic, arising from the
electric charge conferred to the particle surface by adsorbed fatty-acid soap
anions and hydrocolloid carboxylate anions. It is to be expected that the
latter will be particularly tightly bound, because of the large number of

anchorage points which are available. Although limits are set for the molecular
mass of the hydrocolloid component ofthe emulsifier system, no limits appear
to be specified for the extent of carboxylation. Presumably the latter must
be at least sufficient to ensure that the hydrocolloid is water-soluble. The
maximum possible extent of carboxylation is determined by the extent to
which the initial IX-olefin polymer is halogenated.
In a typical example of this process, the polymer to be dispersed is a
98.5/1.5 mol/mol isobutene-isoprene copolymer of number-average molecular
mass ca. 200000 daltons. It is dissolved in toluene to form a 16% m/m cement.
5 pphr of an industrial mixture of fatty acids is added to this cement, together
with ca. 5 pphr of a polyisobutenylsuccinic anhydride of number-average
molecular mass ca. 1000 daltons, the latter being added as a 50% m/m solution
in toluene. In this example, the mixture of fatty acids comprises mainly oleic
acid, with minor amounts of linoleic acid, palmotoleic acid, palmitic acid
and myristic acid. The mixture of rubber cement and emulsifier acids is then
blended with 625 pphr of water and ca. 8.33 pphr of a 10% m/v aqueous
solution of potassium hydroxide. The order of addition of the two phases is
unclear, but the blending is to take place over a period of 3-5 minutes, with
stirring. The pH of the blend after mixing is ca. 10.1; this is raised to ca. 10.5
by addition of 10% m/v aqueous potassium hydroxide. The resultant partial
emulsion is then run through a homogenizer for ca. 15 minutes. and stripped
of toluene and some of the water by evaporation, to give a latex having a
total solids content of 61.5% m/m. The advantages conferred by using the
two-component emulsifier, as compared with using the fatty-acid emulsifier
alone, either in the amount stated in this example, or in an amount equal
to the total mass of emulsifiers in this example, is evident from the results
shown in Table 12.1. By the use of the second component, it is possible to
obtain a latex which is of low viscosity and has good mechanical stability.
However, it should be noted that the ratio of hydrophilic colloid stabilizers
to rubber is high, being ca. II pphr.
J2.3.2.7 The Vanderhoff- EJ-Aasser- Ugelstad process

This process [16] depends upon a colloid phenomenon which has been known
for many years. This is that very fine sub-micron (0.05-0.5 ~m), stable
oil-in-water emulsion can be prepared if a combination of a water-soluble
surface-active substance, which is usually ionic, and a more hydrophobic,
less water-soluble organic compound is used as the emulsifier, provided that
an appropriate emulsification procedure is employed. The more hydrophobic
organic compound is sometimes known as a co-surfactant. A typical example
of such a combination is sodium n-dodecyl sulphate as the ionic surface-active
agent and n-hexadecylalcohol (cetyl alcohol) as the less water-soluble organic
compound. It is believed that the primary function of the latter is to retard
Table 12.1 Comparison between an isobutene-isoprene rubber artificial latex prepared

using only potassium oleate as emulsifier, and a similar latex prepared using a
combination of potassium oleate and potassium polyisobutenylsuccinate (Beerbower,
Burton and Malloy [13])
High-temperature Viscosity of
mechanical stability of finished
Amount raw latex· latex
Emulsifier / pphr / % m/m coagulum on solids / Pas
Potassium oleate 5.5 15 0.825

Potassium oleate 5.5
+ + 0.9 1.14
Potassium polyisobutenyl- 5.1
succinate
Potassium oleate 11.0 2.5 6.60
* Determined by stirring 50 g of the raw (i.e., unstripped) latex in a 400-cm J beaker for 30
minutes at rotational frequency 19 ()()() minute - I and S2°C, diluting with water, filtering through
a l00-mesh screen and weighing the dried solids retained; expressed as retained solids as a
percentage of the solids content of the initial latex.
diffusion of the oil molecules out of the droplets. In some cases, such as the
combination of sodium n-dodecyl sulphate and n-hexadecylalcohol, it is
thought that the colloid stability of the emulsion is further enhanced by the
formation of intermolecular complexes at the oil-water interface, these
complexes conferring electric charge at the surface and giving rise to low
interfacial free energy and high resistance to droplet coalescence.
The Vanderhoff-El-Aasser-Ugelstad [16] process is potentially very wide
in scope. It is in principle applicable to any polymer which is insoluble in
water, and which can be presented to the aqueous phase as a liquid of
sufficiently low viscosity to permit emulsification. This requirement is
generally met, provided that the viscosity is less than ca. 100 Pa s. In many
cases, this requirement will be achieved by dissolving the polymer in an
essentially water-immiscible organic solvent of low molecular mass; the
process is then of the solution-emulsification type, and the solvent must be
stripped from the emulsion to give the finished latex. However, the patent
covering this process also contemplates the possibility that a solvent-free
polymer may be convertible to a liquid of sufficiently low viscosity by heating,
an obvious constraint being that the temperature to which the polymer is
heated must not exceed the boiling point of water. The range of polymer
types to which the process is applicable is very wide; it includes unusual
polymers (as regards latex science and technology) such as epoxide resins
and water-insoluble cellulose derivatives, e.g. ethylcellulose. A wide range of
surface-active substances can be used as the primary emulsifier; sodium
n-dodecyl sulphate, n-hexadecyltrimethylammonium bromide and a 4O-mole
ethoxylate of n-octylphenol are respective examples of anionic, cationic and non-
ionogenic surface-active substances which are cited as suitable. The water-

insoluble co-surfactant is an aliphatic hydrocarbon or hydrocarbon derivative,
the hydrocarbon moiety of which contains at least 8 carbon atoms. Suitable
hydrocarbon derivatives include alcohols, amines, ethers and esters. The
solubility in water should be low, being less than ca. to- 3 g dm - 3, and
preferably less than ca. to - 4 g dm - 3; the molecular mass should not be too
high, being preferably in the range ca. 110-500 daltons. n-decane, n-
tetradecane and n-hexadecane are cited as suitable aliphatic hydrocarbons
for this application; suitable hydrocarbon derivatives include n-tetradecyl
alcohol, n-hexadecyl alcohol, n-tetradecylamine and di-n-octyl ether. The two
components of the emulsifier system can be incorporated in either the polymer
phase or the aqueous phase, as is convenient. Water-soluble surface-active
substances will usually be dissolved in the aqueous phase. If the co-surfactant
is a hydrocarbon, it will probably be convenient to dissolve it in the polymer
phase. Less hydrophobic co-surfactants, such as alcohols, can be incorporated
in the aqueous phase together with water-soluble surface-active substances.
In a typical example of this process, the polymer to be dispersed is a
polystyrene of molecular mass 20000 daltons. It is dissolved in 300 pphp of
benzene, and the solution is emulsified in 1184 pphp of water using 5.3 pphp
of sodium n-dodecyl sulphate and to.7 pphp of n-hexadecyl alcohol. The
procedure is to dissolve the sodium n-dodecyl sulphate and n-hexadecyl
alcohol in the water by heating for 30 minutes at 63°C. A crude emulsion is
then prepared by adding the polymer solution to the aqueous phase and
stirring for 30 minutes at 63°C. The crude emulsion, which contains droplets
of a wide range of sizes, e.g. 1-100 11m, is then passed through a homogenizer
to reduce the droplet size. After this treatment, most of the droplets have
diameters ca. 0.1 11m, but some have diameters up to ca. 0.3 11m. It should be
noted that the ratio of emulsifiers to polymer is still high, being 16 pphp,
although 67% of the emulsifier is not hydrophilic in the usual sense.
12.3.2.8 Methods for polyurethane artificial latices
12.3.2.8.1 Introduction
In recent years, there has been increasing interest in the production of
polyurethane polymers in latex form. This subject has been reviewed in detail
by Dieterich [17]. Much ofthe summary which follows is based upon that review.
The term polyurethane is used to denote those macromolecular substances
which contain what are known generically as urethane linkages (i.e.
-NH.CO.OR) at regular intervals in the polymer chain. These linkages are
formed by reaction between an isocyanate R.N:C:O) and an active-hydrogen
compound (X-H) as follows:
R.N=C=O + X-H ---+ [R.N=r-OH ] ---+ R.~----<r=o

X H X
If, for example, the active-hydrogen compound is an alcohol, R'OH, then the
product of the reaction is a urethane ester R.NH.C0 2 R' in the first instance,
although this can undergo reaction with further isocyanate, because it also
contains an active-hydrogen atom. If the active-hydrogen compound is an
amine, R'NH 2 , then the product is a di-substituted urea, R.NH.CO.NHR' in
the first instance. If the active-hydrogen compound is urea or a urea derivative,
R'NH.CO.NHR', then the product is a biuret, R.NH.CO.NR'.CO.NHR' in
the first instance. Reactions of this type are used to chain-extend and crosslink
polymers of lower molecular mass, such as polyethers and polyesters.
Although the repeat units within these latter polymers are not joined together
by urethane-type linkages, the final products are usually known as polyurethanes.
Polyurethane polymers are prepared by bulk or solution polymerization
and chain-extension reactions. Thus it is inevitable that, if required in latex
form, they must be dispersed in a suitable medium to form artificial latices.
When it is liquid polyurethane precursors which have been dispersed, it is
more correct to describe the dispersions as emulsions rather than as latices;
however, little purpose is served by maintaining this distinction. In what
follows, a selection of methods for preparing polyurethane artificial latices
is considered under the following headings:
1. dispersions of isocyanate-terminated prepolymers;
2. chain-extension of dispersions of isocyanate-terminated prepolymers;
3. the acetone process;
4. the melt-dispersion process;
5. dispersions in which chain-extension and crosslinking occur by non-
isocyanate reactions.
12.3.2.8.2 Dispersions of isocyanate-terminated prepolymers

It appears that it is possible to prepare stable dispersions of isocyanate-
terminated prepolymers by shearing the prepolymer in the presence of water
and a suitable dispersant. Small amounts of water-insoluble organic liquids
may be added to facilitate dispersion. Typically, the prepolymers in these
dispersions are subsequently reacted with diamines to form polyurethane-
polyureas of high molecular mass. The matter of the stability of such
dispersions has two distinct aspects. There is, of course, as always the matter
of the colloid stability of the dispersion. However, for isocyanate-
terminated prepolymers, there is also the special problem of the chemical
stability of the polymer in an aqueous environment. Water is an active-hydrogen
compound, and shows the same reactivity towards isocyanates as do other active-
hydrogen compounds. The reaction between water and isocyanates fits into
the general pattern indicated above. The initial product is a carbamic acid
(Structure I), but these acids are unstable, dissociating to give primary amines
and carbon dioxide. The overall reaction between an isocyanate and water
can be represented as foIIows:
R.N=C=O + HO-H ---+ [R·f-Y=O] -+ R·f- H + o=c=o

H OH H
Thus it is at first sight somewhat surprising that it is possible to prepare

aqueous dispersions of isocyanate-terminated prepolymers which remain
stable for relatively long periods of time. To the extent that these aqueous
dispersions are chemically-stable for significant periods of time, that stability
is a consequence oflimited access of water molecules to the isocyanate groups.
Limited access is presumably primarily a consequence of the low solubility
of water in the pre polymer particles. Notwithstanding short-term chemical
stability, aqueous dispersions of isocyanate-terminated prepolymers are not
chemically stable in the long term, in the sense in which most latices are
stable. In the absence of more reactive reagents to bring about chain-extension
and crosslinking, the isocyanate groups react with water, the rate of reaction
being influenced by factors such as temperature, pH, the intrinsic reactivity
of the particular isocyanate group, and the hydrophilicity of the prepolymer.
The latter factor influences the equilibrium concentration of water molecules
in the prepolymer particles, and therefore the equilibrium concentration of
water molecules available to the isocyanate groups for reaction.
Interesting variants of the straightforward dispersion of isocyanate-
terminated prepolymers in water include:
I. the use of isocyanate-terminated prepolymers which are self-dispersing

because they contain ionic groups or hydrophilic polyether segments;
2. the use of isocyanate-masking reagents which render the isocyanate group
unreactive at normal ambient temperatures towards water (and also
towards other active-hydrogen compounds with which they usuaIIy react),
but which form compounds capable of reacting as isocyanates when the
temperature is subsequently raised, either by thermal dissociation to
regenerate the isocyanate group or by some other mechanism.
Self-dispersing isocyanate-terminated prepolymers are prepared by incorporating

hydrophilic moieties such as sulphonate or carboxylate groups in the
prepolymer. It is said that, if the molecular mass of the prepolymer is below
ca. 8000 daltons, and if sufficient hydrophilic groups are incorporated, then
the pre polymer disperses in water spontaneously even without stirring. These
prepolymers can also be used as aids to the dispersion of isocyanate-
terminated prepolymers which do not contain hydrophilic groups. Isocyanate

groups which have been reversibly reacted so as to make them temporarily
unreactive are said to have been masked, capped or blocked. The best-known
isocyanate-masking reagent is phenol; the capping reaction causes the
formation of a phenyl urethane ester. The capped isocyanate groups are
capable of reacting normally with active-hydrogen compounds, such as
polyols and polyamines, when the prepolymer and the active-hydrogen
compound are subsequently heated to an elevated temperature after being
separated from the aqueous dispersion medium. Reaction at elevated
temperatures is presumed to occur by way of thermal regeneration of the
isocyanate groups. The masking-unmasking reaction with phenol is usually
represented as follows:
The subject of the masking of isocyanate groups is of considerable relevance

to the production of aqueous dispersions of isocyanate-terminated prepolymers.
It has been extensively reviewed in two papers by Wicks [18,19].
12.3.2.8.3 Chain-extension of dispersed isocyanate-terminated pre polymers to

give artificial latices of polyurethanes of high molecular mass
In an early process, a diisocyanate was emulsified in water and subsequently
reacted with an equivalent amount of a diamine under vigorous stirring
conditions. The amine diffused into the droplets of diisocyanate, where it
reacted to give a polyurea in the form of a fine dispersion. Surface-active
substances and hydrocolloids were used to stabilize the dispersion colloidally.
This process may be regarded as the precursor of processes in which
isocyanate-terminated prepolymers are dispersed in an aqueous medium and
then chain-extended with a diamine to give an artificial latex of a polyurethane
polyurea of high molecular mass. Dispersion of the initial isocyanate-
terminated prepolymers can be facilitated by reducing the viscosity by
addition of toluene. A further development is the use of ionic polyurethane
polymers as dispersants. These are said to be superior to conventional surface-
active substances as dispersing aids for non-ionic polyurethanes. An example
of an anionic polyurethane is the polymer which is formed when an
isocyanate-terminated polyether prepolymer is reacted with a hydroxycarboxylic
acid, and the carboxylic-acid groups then ionized by reaction with an alkali.
Precautions must be taken to avoid significant reaction occurring between
the carboxylic-acid groups and the isocyanate groups of the prepolymer,
because such reaction leads to undesirable branching and crosslinking. It is
advantageous if the carboxylic-acid group of the hydroxycarboxylic acid is
sterically hindered, as in, say, 2,2-dimethylolpropionic acid (Structure II).
Sulphonate groups are conveniently introduced by reacting an isocyanate-
terminated prepolymer with a diaminoalkanesulphonate.
CH 3 C(CH 2 0HhC0 2 H
II
In another variant of this type of process, the reactivity of a chain-extending

diamine towards isocyanate groups is reduced by reaction with a ketone to
form a bis-ketimine prior to the prepolymer and diamine being dispersed in
the aqueous phase:
R'
R"/
'" C=N-R-N=C /
R'
"'R"
+ 2H
2
°
Diamines masked with ketones to form ketimines can be mixed with
isocyanate-terminated prepolymers without reaction occurring. The mixture
is then dispersed in an aqueous phase, whereupon the ketimine hydrolyses
to regenerate the diamine. Chain-extension then occurs. This process is known
as the ketimine process.
In a further variant of this type of process, known as the ketazine process,
the isocyanate-terminated prepolymer is mixed with a ketazine (Structure
III), an aldazine (Structure IV) or a hydrazone (Structure V) before being
dispersed in the aqueous phase. After dispersion, these compounds are
hydrolysed by the water to produce hydrazine (H 2 N.NH 2 ), which then effects
chain-extension by reacting with the isocyanate groups of the prepolymer.
The hydrolysis of ketazines occurs more slowly than does that of ketimines;
the process is therefore said to be particularly suitable for the chain-extension
of prepolymers which are terminated with especially reactive isocyanate
groups, notably, with aromatically-bound isocyanate groups.
R R R
'"
"'C=N-N=C/ R-CH=N-N=CH-R "'C=N-NH
/ 2
R'/ R' R'
III IV V
12.3.2.8.4 The acetone process

The acetone process makes use of the observation that polyurethanes which
have some ionic groups incorporated in them are able spontaneously to form
stable dispersions when solutions of the polymer in water-miscible organic
solvents, such as acetone, methylethyl ketone and tetrahydrofuran, are mixed
with water. If the organic solvent is subsequently removed by distillation, a
solvent-free latex of the polyurethane is obtained. In the dispersion which
forms initially when the polymer solution is mixed with water, the organic
solvent is partitioned between the polymer and aqueous phases. The
mechanism by which the initial dispersion forms appears to depend upon
factors such as the concentration of the solution, the concentration of ionic
groups in the polymer, and (presumably) the order of mixing of the two
phases. At least two distinct types of mechanism can be envisaged. In the
first, the hydrophobic segments of the polyurethane macromolecule separate
from the aqueous phase because the solvent medium becomes progressively
less favourable as the ratio of water to organic solvent increases. In the
second, a dispersion of water in polymer solution forms initially, and then phase-
inversion occurs as the ratio of water to organic solvent increases.
Dieterich [17] has given the following description of the preparation of a
cationic polyurethane artificial latex by the acetone process: An isocyanate-
terminated prepolymer is prepared by reacting a hydroxy-terminated adipic-
acid polyester of molecular mass ca. 2000 daltons with excess hexamethylene-
diisocyanate. The prepolymer is so viscous that further reactions must be
carried out in solution. An equimolar amount of N-methyldiethanolamine
dissolved in acetone is added. The viscosity increases as chain-extension
occurs. As the viscosity increases, so further amounts of acetone are added
to keep the mixture stirrable. The next step is to quaternize the tertiary
nitrogen atoms of the segmented polyurethane by reaction with dimethyl
sulphate. Quaternization is accompanied by a further increase in viscosity,
attributed to interchain association as electric charges are conferred along
the polymer chains. Water is now slowly and progressively added to the
mixture, whereupon a succession of changes is observed. Initially the viscosity
falls to approximately the level before quaternization. This is attributed to
the water molecules reducing ionic interchain associations, presumably by
hydration of the ionic sites and because electrostatic forces are reduced. A
clear solution, in which the ionic polyurethane is molecularly dispersed, is
obtained. As more water is added, the viscosity increases again, notwithstanding
that the polymer concentration is decreasing. This is attributed to reduced
solvation of the hydrophobic segments of the polymer chain as the concentration
of acetone in the dispersion medium decreases. This is the first stage of the
formation of a heterogeneous system. Further addition of water causes
turbidity to develop. Eventually the viscosity decreases again. As the acetone
concentration in the dispersion medium continues to decrease, so the
hydrophobic segments of the polymer chains agglomerate progressively to
form the body of the eventual latex particles. At this stage, the system
comprises a continuous aqueous phase in which are dispersed particles of
polyurethane swollen with acetone. The formation of this dispersion is a
gradual process; it is not possible to say with accuracy at what point the
system becomes heterogeneous. The final step in the preparation of the latex
is the removal of acetone by distillation. As the acetone is removed, the
turbidity of the latex increases, and the viscosity decreases, because of
decreasing solvation of the hydrophobic segments of the polyurethane chain.
The physical properties of latices prepared by this method are said to depend
upon several factors, including the chemical composition of the polymer, the
type and concentration of ionic groups, the molecular mass of the
isocyanate-terminated prepolymer, and the precise procedure used to prepare

the latex. It appears that the diameter of the latex particles can vary widely
over the range 10-5000 nm. In consequence, the appearance can vary between
an opaque translucent sol and a milky white dispersion. Ionic polyurethanes
of the type described in this example can also be used as dispersants for
preparing artificial latices of non-ionic polyurethanes by this method. The
procedure is said to be applicable to all linear polyurethanes which can be
prepared in organic solvents and modified by incorporation of ionic groups.
An important feature of the process is that all the polymerization and
chain-extension reactions take place in homogeneous phases. One disadvantage
of the process is said to be its cost.
12.3.2.8.5 The melt-dispersion process

As the name implies, the melt-dispersion process is one in which the
polyurethane is dispersed in the melt condition at an elevated temperature.
The polymer is sufficiently hydrophilic as to be self-dispersing. Chemical
modification subsequent to dispersion increases the molecular mass of the
polymer, and also causes it to become less hydrophilic. No organic solvents
are required, and at no stage do isocyanate groups come into contact with
water. Dieterich [17] has given the following description of a process of this
type, in which the polyurethane is made self-dispersing by developing cationic
sites along the polymer chain: An isocyanate-terminated prepolymer is
prepared in the melt by reacting a polyether or polyester diol having molecular
mass within the range 1500- 5000 daltons, a diisocyanate and a glycol which
contains a tertiary amino group. The product is then reacted successively
with two types of reagent as follows:
l. a compound which reacts with the isocyanate groups by addition, and
which leaves terminal groups capable of methylolation by reaction with
formaldehyde; and
2. a compound which quaternizes the tertiary amino groups which have
become incorporated in the isocyanate-terminated prepolymer.
In the example given by Dieterich, the compound which reacts with the
isocyanate groups is urea, and the quaternizing agent is 2-chloroacetamide
(Structure VI). The urea reacts with the terminal isocyanate groups to form
biuret groups which are subsequently methylolated by reaction with

formaldehyde. The fourth group which becomes attached to the quaternized
nitrogen atom is the acetamido moiety. Thus not only is the resultant polymer
terminated exclusively by amide groups; there are also amide groups attached
to the ionic sites. The reactions which lead to the formation of the polymer
at this stage can be represented as follows:
R'
1
OCN.R.NCO + H~H + OCN.R.NCO + HD-N-OH + OCN.R.NCO
1 R'
1
OCN.R.NH.CO.()....-...O.OC.NH.R.NH.CO.D-N-O.OC.NH.R.NCO
1 + H,N.CO.NH,
+ CICH,.CO.NH,
R' a-
1+
H2N.CO.NHCO.NHR.NH.co.~.oc.NH.R.NHCO.D-~.oc.NH.R.NHCO.NHCO.NH2
1
CH 2·CO.NH2
The polymer is now methylolated by reaction with aqueous formaldehyde
in homogeneous phase. The resulting methylolated polyurethane is sufficiently
hydrophilic to be capable of being diluted with almost unlimited amounts
of water at temperatures in the range 50-130·C, thereby giving a stable
aqueous dispersion. If the pH is then lowered, polycondensation occurs by
reaction between the methylol groups and the amino groups. The final
product is a dispersion of a modified polyurethane polymer of high molecular
mass and reduced hydrophilicity. Thus the polymer has its maximum
hydrophilicity at the stage where this is most desirable, namely, when it is
to be dispersed in the water. The initial absorption of water into the melt
produces an apparently homogeneous mixture. On the molecular level,
however, it is likely that the water molecules are preferentially associated
with the various hydrophilic sites in the polymer, both ionic and non-ionic.
As more water is absorbed, these water-rich regions grow in size, turbidity
develops, and the system becomes effectively an emulsion of water droplets
in a continuous phase of polymer melt, the hydrophilic groups of the polymer
tending to accumulate at the interface between the aqueous droplets and the
polymer melt. Further addition of water causes the droplets to grow in size,
until, just before phase-inversion occurs, the system is a viscous white paste
in which the droplets of aqueous phase have compressed the hydrophobic
domains into the spaces between the droplets. Still further addition of water
causes the system to phase-invert, with appropriate restructuring of the
particles. The hydrophilic groups of the polymer tend to remain at the
interface, which is now between droplets of the polymer melt and the
continuous aqueous phase. It is these surface-bound hydrophilic groups which
confer colloid stability upon the final artificial latex. This method is said to
be of wide applicability. It is necessary that the conditions should not be
such that phase-inversion is prevented because the system has become 'fixed'
as an emulsion of aqueous phase in polymer melt. At least two circumstances
can lead to this latter situation: insufficient hydrophilic groups being present
in the polyurethane polymer, in which case the absorption of water into the
system ceases before phase-inversion has occurred; and the occurrence of too
much polycondensation between the methylol groups and amino groups
prior to phase-inversion.
12.3.2.8.6 Dispersions in which chain-extension and crosslinking occur by

non-isocyanate reactions
Reference has been made above (section 12.3.2.8.2) to the matter of the
inherent reactivity of isocyanate groups towards water, and the problems to
which this gives rise in relation to aqueous dispersions. One way of
circumventing this problem is to react the isocyanate groups under anhydrous
conditions with reagents which leave the prepolymer terminated with other
types of group which can be used to effect chain-extension and crosslinking
by non-isocyanate reactions. The modified pre polymer can then be dispersed
in an aqueous medium, and the chain-extension and crosslinking reaction
carried out, provided, of course, that the modified prepolymer is not itself
susceptible to reaction with water molecules. A process of this type has been
described by Vanderhoff, EI-Aasser and Hoffman [20]. The essential features
are as follows: An isocyanate-terminated prepolymer is reacted under
anhydrous conditions with an olefinically-unsaturated hydroxy compound
to produce a prepolymer which has polymerizable olefinic double bonds in
its terminal groups. The olefinically-unsaturated hydroxy compound is
preferably a hydroxyalkyl acrylate or methacrylate, but may also be allyl
alcohol. The modified prepolymer is dispersed in an aqueous phase using
the system described by Vanderhoff, EI-Aasser and Ugelstad [16] (see section
12.3.2.7 above). Because the isocyanate groups of the initial prepolymer have
been reacted and thereby rendered inert, the modified prepolymer can be
dispersed in an aqueous medium without the possibility of the terminal
groups being involved in premature reactions. Free-radical polymerization
conditions are then established, so that the modified pre polymer molecules
become linked together by polymerization through their terminal olefinic
unsaturation.
Several variations of this process are possible, in addition to the more
obvious variations associated with the choice of olefinically-unsaturated
hydroxy compound with which the isocyanate-terminated prepolymer is
reacted, and with the emulsifier system. Thus the olefinically-unsaturated
hydroxy compound can advantageously be mixed with a non-olefinically-
unsaturated hydroxy compound. The effect is to reduce the concentration of
olefinic unsaturation introduced into the modified pre polymer, thereby
moderating the chain-extension and crosslinking which occur subsequently.
Such moderation is said to be especially desirable if the olefinically-
unsaturated hydroxy compound is a hydroxyalkyl acrylate or methacrylate
rather than allyl alcohol. A further variation is that, prior to dispersion in
the aqueous phase, the modified prepolymer is dissolved in an olefinically-
unsaturated monomer. The dispersion process then becomes one of solution-
emulsification. When subsequently conditions are established for free-radical

polymerization through olefinic double bonds, the olefinically-unsaturated
monomer polymerizes, and the modified prepolymer is expected to become
combined with the growing polymer chains by way of copolymerization
through the olefinic double bonds in the terminal groups of the prepolymer.
The particles of the resultant latex are expected to comprise an interpenetrating
polymer network of chains derived from the initial prepolymer and chains
derived from the polymerized olefinically-unsaturated monomer. Admixture
of the modified prepolymer with an olefinically-unsaturated monomer confers
at least three advantages:
1. Dispersion of the modified prepolymer in the aqueous phase is facilitated
because the viscosity of the prepolymer is reduced by dilution with the
monomer.
2. Whereas, when a polymer is dispersed in an aqueous phase by the
solution-emulsification technique, it is usually necessary subsequently to
remove the solvent, this requirement is obviated in this process because
the monomer polymerizes when polymerization conditions are established.
3. There is considerable scope for modifying the properties of the final
polymer by appropriate variation of the nature and amount of the
olefinically-unsaturated monomer in which the modified prepolymer is
dissolved.
The olefinically-unsaturated monomer should be sufficiently hydrophobic to
be absorbed into the prepolymer particles. Alkyl acrylates and methacrylates
are suitable. Provided that (as will usually be the case) the olefinically-
unsaturated monomer is unreactive towards isocyanate groups, it can
advantageously be used as the vehicle for the reaction in which the
isocyanate-terminated pre polymer is modified by reaction with the hydroxy
compound(s). The polymerization conditions to be established after the
solution of modified prepolymer has been emulsified in the aqueous phase
are typically those for free-radical aqueous emulsion polymerization. Free
radicals then enter the prepolymer/monomer particles from the aqueous
phase, as they do in a conventional emulsion polymerization, and there
initiate polymerization of the monomer.
Typically the isocyanate-terminated prepolymer is dissolved in a mixture
of 2-hydroxypropyl methacrylate, n-butyl alcohol, n-butyl acrylate and
isobutyl methacrylate. Dibutyltin dilaurate is added to catalyse the reaction
between the hydroxy compounds and the isocyanate groups. 48 hours at
normal ambient temperature is said to be sufficient for complete reaction.
The resulting solution is then coarsely emulsified in water at ca. 63°C, using
a mixture of sodium n-dodecyl sulphate and n-hexadecyl alcohol as emulsifier.
The crude emulsion is passed through a homogenizer to reduce the droplet
size to 0.03-0.40 ~m. An aqueous solution of sodium n-dodecyl sulphate,
n-hexadecyl alcohol and sodium bicarbonate is then added to the fine
emulsion, the temperature raised to ca. 63°C, and the mixture purged with
oxygen-free nitrogen. Polymerization is initiated by addition of potassium per-
oxodisulphate. It is complete after 3-4 hours reaction. The product is said to
be a stable latex, the disperse phase of which is a fully-polymerized polyurethane.
12.3.2.9 Effect of solution-emulsification process upon polymer properties

An investigation of some of the effects of the solution-emulsification process
upon the properties of the polymer contained in an artificial natural rubber
latex has been reported by Rogers, Ryan and Hecker [21]. They compared
some of the properties of red is persed natural rubber with those of the polymer
in natural rubber latex. Apart from the artificial latex being rather more
viscous than the natural latex, it was found that the process of solution
emulsification tends to reduce the modulus and tensile strength of the
vulcanized rubbers obtained from the dispersed polymers. However, the
effects are small, and it was noted that they can be offset by altering the
vulcanization system. These effects are illustrated in Figs 12.2(a) and (b); these
give the results for tensile strength and modulus at 100% extension
respectively. They are not unexpected, and are presumably consequences of
1.1
45
~ 1.0
--
~
<II c B
~ 40 .~ 0.9
c
CD
xCD
~ 0.8
8
1ii
'" 0.7
'"
"5
"8
~ 0.6
300L----........110----2.1...0---.....
30 0.5 0 10 20 30
(a) Time of vulcanization / minute (b) Time of vulcanization / minute
Figure 12.2 Effects of the solution-emulsification process upon tensile stress-strain properties
of sulphur-postvulcanized films obtained from natural rubber latex (Rogers, Ryan and Hecker
[21]): (a) tensile strength; (b) modulus at 100"10 extension. Results are shown for films from
natural rubber latex (curve A), redispersed natural rubber (curve 8), and redispersed purified
natural rubber (curve q. Latex formulation (parts by mass dry): rubber 100, sulphur 2, zinc
diethyldithiocarbamate 0.5, zinc mercaptobenzthiazolate 1.5, zinc oxide 3, antioxidant I, diphenyl-
guanidine 0.6, sodium silicofluoride 1. Vulcanization conditions: various times at lOO"C.
Particular types of artificial latex 419
reduction in the molecular mass of the rubber to facilitate dissolution in the

solvent prior to emulsification. Rogers, Ryan and Hecker also give results
for the effects of purifying the natural rubber prior to redispersion. These
are also shown in Figs 12.2(a) and (b). The tensile srength of the vulcanizates
is significantly lower than that ofvulcanizates from the redispersed unpurified
rubber. The vulcanizate modulus is also lower, but the effect is not so marked
as that upon tensile strength. These effects are presumably a consequence of
removal from the rubber of substances, such as nitrogenous bases, which are
able to function as vulcanization accelerators.
12.4 PARTICULAR TYPES OF ARTIFICIAL LATEX
12.4.1 Artificial latices of isobutene-isoprene rubbers

Isobutene-isoprene rubber (butyl rubber) is a copolymer principally of
isobutene with a small amount of isoprene to provide a low concentration
of olefinic unsaturation which renders the rubber vulcanizable by heating
with sulphur and conventional organic accelerators. Its structure can
therefore be represented as shown at VII, in which the number of isobutene
... -CH 2-C(CH 3)2-'" ... -CH 2-C(CH 3)=CH-CH 2-' ..
VII
units far exceeds the number of isoprene units. The mole fraction of isoprene
units in the rubber is typically within the range ca. 0.01-0.05. This type of
rubber is noted for its low permeability to gases, its general chemical
unreactivity, and its resistance to deteriorative influences such as heat,
sunlight and ozone. These desirable characteristics have contributed, directly
or indirectly, to industrial interest in this type of rubber in latex form. They
diminish progressively in magnitude as the proportion of isoprene in the
polymer increases. Especially is this so for resistance to deterioration by
exposure to ozone. Insofar as this rubber is available as a latex, it has to be
as an artificial type, because the rubber is produced by a cationic solution-
polymerization reaction. Details of two grades of isobutene-isoprene rubber
latex which have been available from the Burke- Palmason Chemical
Company [22,23] in recent years are given in Table 12.2. The MD-602 grade
contains a co-emulsified hydrocarbon resin. It is evident that these latices
are colloidally stable over a wide pH range. They are also resistant to colloidal
destabilization by many electrolytes. It is therefore not possible to coagulate
and gel these latices by conventional methods for rubber latices. In particular,
it is not possible to practise conventional coagulant dipping (see section
17.2.3 of Chapter 17 (Volume 3)) using coagulants such as calcium nitrate.
These latices are, however, susceptible to colloidal destabilization by
water-soluble alcohols and ketones, and by concentrated aqueous solutions
of strong alkalis, e.g., 50% m/v sodium hydroxide. Use can be made of the
Table 12.2 Information concerning two grades of isobutene-isoprene rubber latex

which have been available from the Burke-Palmason Chemical Company in recent
years (Burke- Palmason Chemical Company [22, 23])
Grade of isobutene-isoprene rubber latex
Property BL-100 MD-602
Total solids content / % m/m 61-63 55-57

pH 4.5-6.5 4.5-6.0
Average particle diameter / ~m <1 <1
Viscosity at 25T· / Pa s 1.8-5 0.8-1.6
Surface free energy / mJ m - 2 38
pH range of colloid stability 2-12 2-12
* Brookfield LVT, # 3 spindle, rotational frequency 12 minute - I.
colloid-destabilizative effect of water-soluble alcohols and ketones to sensitize

compounded isobutene-isoprene rubber latices to metal-salt coacervants.
The amount required depends upon factors such as total solids content, pH
and temperature; 5-10% v/v is recommended as a starting-point. As expected,
isobutene-isoprene rubber latices have excellent mechanical stability.
Although the chemical stability of artificial latices of isobutene-isoprene
rubbers is high, because of the high concentration of colloid stabilizers they
contain, nevertheless their viscosity can be greatly reduced by the addition
of small amounts of inorganic electrolytes. This effect is illustrated by results
for the effects of two added electrolytes, ammonium chloride and sodium
chloride, reported by Miller and Powers [24] and reproduced here in Fig.
12.3. Initially the viscosity of the latex increases, but a substantial decrease
in viscosity occurs once a certain level of electrolyte addition has been
attained. The cause of these viscosity changes is obscure. One possibility is
that they are associated with partial flocculation of the latex particles. Another
possibility is that the latex contained one or more macromolecular colloid
stabilizers the conformation of which in aqueous solution is sensitive to
electrolyte concentration.
Artificial isobutene-isoprene rubber latices were originally developed to
provide bases for tyre-cord adhesives to be used in the construction of
isobutene-isoprene rubber tyres. Other proposed applications for these
latices have included binders for paper coatings, treatments for barrier papers,
binders for non-woven fabrics, vehicles for latex-based surface coatings, tyre
sealants, and various types of adhesive.
12.4.2 Artificial latices of synthetic cis-I, 4-polyisoprene rubbers

There has been much industrial and academic interest in the past in
polymerizing both butadiene and isoprene to produce polymers which could
Particular types of artijiciallatex 421
tJ)
8!.
.......
.a-2 B
.~
:>
1 2 3 4 5 6
Level of added electrolyte I pphr
Figure 12.3 Effect of added ammonium chloride (curve A) and added sodium chloride (curve
0) upon viscosity of an artificial latex of an isobutene-isoprene rubber (Miller and Powers
[24]). Total solids content of latex: ca. 54% mjrn. Viscosities determined using a Brookfield
Model LVF viscometer with various spindles and various rotational frequencies.
be used as synthetic equivalents of natural rubber. There has been especial

interest in polyisoprenes which can be so used, because of the chemical
similarity between the repeat unit of the hydrocarbon polymer which is
synthesized in the rubber tree and the polymer which is obtained by
polymerizing isoprene. The important features of the natural rubber
macromolecule are summarized in section 9.5.2.2.2 of Chapter 9. It was soon
realized that, in order to obtain a synthetic polyisoprene which would be
broadly equivalent to natural rubber in respect of processing behaviour and
vulcanizate properties, it would be necessary to reproduce the stereoregularity
of the polyisoprene sequence in natural rubber. Although it is possible to
emulsion polymerize by free-radical mechanism both butadiene and isoprene
to give stable latices, the polymers dispersed in these latices are not
stereo regular (see section 11.2.l of Chapter 11). In the case of polyisoprenes
produced by free-radical emulsion polymerization, substantial proportions
of the repeat units become polymerized in the 1,2-, 3,4- and trans-l,4-
modes, as well in the desired cis-l,4-mode. It was at one time thought that
it might be possible to produce rubbery stereoregular polymers of butadiene
and isoprene by non-free-radical emulsion polymerization using various metal
salts as initiator. However, these hopes have not been realized so far. Although
stereoregular polybutadienes and polyisoprenes have been produced for many
years now, and are widely used as equivalents for natural rubber, they are
obtained by solution polymerization under strictly anhydrous conditions.
Insofar as latices of these rubbers are required, they have to be produced by
dispersing the bulk polymers in aqueous media. Because the term artificial
cis-l,4-polyisoprene rubber latex could be understood to mean an artificial
latex which had been produced by redispersing bulk natural rubber in an
aqueous medium, the type oflatex under consideration here is best described,
perhaps rather awkwardly, as artificial synthetic cis-l,4-polyisoprene rubber latex.
It is believed that most, if not all, of the attempts which have been made
to produce artificial latices of synthetic cis-l,4-polyisoprene rubbers have
used processes of the solution-emulsification type. These latices have been
developed primarily as competitors for natural rubber latex. However, they
have failed to establish themselves in any of the conventional applications
for natural rubber latex, or in any other industrial applications. As far as is
known, there are few, if any, artificial cis-I,4-polyisoprene latices produced
and used industrially at the time of writing. A typical artificial synthetic cis-I,4-
polyisoprene latex which was available from the Shell Chemical Company
some years ago had a high total solids content (ca. 65% m/m), a pH of ca.
10, an average particle diameter of ca. 0.7 ~m, and a surface free energy of
ca. 38 mJ m - 2. The amounts of non-rubber substances, other than water,
were small. Minimal quantities of dispersing agents, colloid stabilizers and
phenolic antioxidant were used in the production of the latex. The function
of the phenolic antioxidant was to protect the polymer during transportation
and storage. Being somewhat volatile, it tended to be lost during processing,
so that there might be little residual antioxidant in the dried and vulcanized
product unless other antioxidants were included in the compounding
formulation. However, this had the advantage that, for the first time, the
formulator of a cis-I,4-polyisoprene rubber latex had complete control over
the antioxidants and accelerators which were to be present in the latex
compound. One interesting feature of this artificial synthetic cis-l ,4-polyisoprene
latex is that it combined high mechanical stability with low chemical stability.
It was hoped that this combination of colloidal characteristics would make
it suitable for the manufacture oflatex foam rubber and ofthin-walled rubber
articles by dipping. This expectation has not been realized. The latex was
also sensitive to colloidal destabilization by pressure; thus there were hopes,
again unrealized, that it would find use in pressure-sensitive adhesives.
The particle-size distribution in the artificial synthetic cis-I ,4-polyisoprene
latex produced by the Shell process has been reported by Wales [25] to be
approximately logarithmic-Gaussian with respect to particle diameter. This
means that the proportion of the total number of particles within any specified
infinitesimal range of diameters is Gaussian if the measure of particle size is
taken as the logarithm of the diameter. The average particle size of latices
having particle-size distributions ofthis type can be conveniently characterized

by the mass median particle diameter, i.e., by the diameter which is such that
half the mass of the dispersed polymer is present as particles smaller than
this diameter, and the other half is present as particles larger than this
diameter. For the particular artificial synthetic cis-l,4-polyisoprene latex
investigated by Preiss, Sawyer and Simpson [26] (see further below), the
mass median particle diameter was ca. 0.9 11m; this is similar to that of many
natural rubber latices.
The polymer contained in the Shell latex was a polyisoprene of number-
average molecular mass ca. 2 x 106 daltons, the distribution of molecular
sizes about that average being relatively narrow. The predominant (92-97%)
mode of enchainment of the isoprene units in. the polymer was cis-l,4, the
remainder being trans-l,4. This latex differed from natural rubber latex in
at least two important respects. Firstly, the dispersed polymer was essentially
linear and free from the branched structures and polymer microgel which
are present in the polyisoprene in natural rubber latex. Secondly, as noted
above, the non-rubber substances present in the latex were restricted to
minimal levels of surface-active substances, and to antioxidant. The absence
of proteinaceous substances, carbohydrates, phospholipids, etc. has the
consequence that preservation of the type required for natural rubber latex
is unnecessary.
The effect of the gel content of the dispersed polymer upon the properties
of postvulcanized films from artificial synthetic cis-l ,4-polyisoprene latex has
been reported by Rogers, Ryan and Hecker [21]. Their results for the effect
of vulcanization time upon modulus at 100% extension for postvulcanized
latex films are shown in Fig. 12.4. Results for films from natural rubber latex
postvulcanized using the same vulcanization system are also shown. It is
evident that the behaviour of the films from the artificial synthetic cis-
1,4-polyisoprene latex approximated closely to that of films from natural
rubber latex if the gel content of the dispersed polymer was ca. 28%. As
expected, the moduli of the vulcanized films from the artificial synthetic
cis-l,4-polyisoprene latex which contained gel-free polymer were somewhat
lower. However, the tensile strengths of the vulcanized films from the latex
containing the gel-free synthetic cis-l,4-polyisoprene were found to be higher
than those derived from the latex of synthetic polymer which contained 28%
gel, and also of those derived from natural rubber latex.
Results for the tensile properties of postvulcanized films from artificial
synthetic cis-l,4-polyisoprene latex, natural rubber latex, and blends of the
two, have been reported by Schmit [27]. The formulations and vulcanization
conditions are shown in Table 12.3, and the results for tensile properties in
Table 12.4. Five formulations, designated A, B, C, 0 and E, were used; the
first three of these enable comparisons to be made between the behaviour
of the synthetic polymer and that of the natural polymer. From the results
for films from the first three formulations, it appears that the behaviour of
1.1-
B
1.0 C
8!.
--
::E
I::
.~ 0.9
I:: A
CD
j
::e
0
0
0
11i
III
:::l
"3
"8
::E
0.6
0.5 0 10 20 30
Time of vulcanization I minute
Figure 12.4 Modulus at 100"1. extension for sulphur-postvulcanized films from artificial
synthetic cis-I ,4-polyisoprene latices containing 0"1. gel (curve A) and 28% gel (curve B) (Rogers,
Ryan and Hecker [21]). Results are also shown (curve q for films from natural rubber latex
vulcanized using the same vulcanizing system. Latex formulation (parts by mass dry): rubber
100, sulphur 2, zinc diethyldithiocarbamate I, zinc mercaptobenzthiazolate 3, zinc oxide 3,
antioxidant I, diphenylguanidine 0.6, sodium silicofluoride I. Vulcani7.ation conditions: various
times at 100·C.
Table 12.3 Latex formulations for postvulcanized films from artificial synthetic
cis-I,4-polyisoprene rubber latices, from natural rubber latex, and from blends of the
two (Schmit [27])
Parts by mass dry
Ingredient A B C D E
Synthetic cis-l,4-polyisoprene 100 50 100 100

Natural rubber 50 100
Potassium oleate 0.5 0.5 0.5 0.5
Di-n-butylammonium oleate 0.75
Sulphated alkylphenol ethoxylate 0.5 0.5 0.5 0.5
Potassium hydroxide 0.5
Sulphur 1.5 1.5 1.5 2 2
Zinc diethyldithiocarbamate 1 1 1 1.25 0.5
Zinc oxide 2 2 2 2
Antioxidant 1 1 1 1
Vulcanization conditions ? ? ? 15 min 15 min

at at
l00·C l00·C
Table 12.4 Tensile stress-strain properties of postvulcanized films from artificial

synthetic cis-l,4-polyisoprene rubber latices, from natural rubber latex, and from
blends of the two (Schmit [27])
Formulations and vulcanization conditions as shown in Table 12.3
Formulation for post vulcanized film
Property A B C D E
Modulus at 500%
extension / MPa 2.41 2.62 2.41
Modulus at 700%
extension / MPa 6.21 10.14 15.86 5.52 8.96
Tensile strength / MPa 41.4 41.4 42.4 41.0 41.7
Extension at break / % 950 930 880 970 920
the polymer contained in the artificial synthetic cis-l,4-polyisoprene latex

was broadly similar to that of the polymer contained in natural rubber latex.
However, an important difference is that the 700% modulus ofthe vulcanizate
from the artificial latex was significantly lower. The tear strength was also
lower. These effects are attributed to the absence from the artificial poly isoprene
latex of organic substances, in particular, nitrogenous bases, which are present
in natural rubber latex, and which function as accelerators of vulcanization.
The results for formulations D and E demonstrate that a significant increase
in 700% modulus can be obtained if zinc oxide is omitted, and if part of the
accelerator is replaced by an alkylammonium carboxylate soap. Presumably
the amine moiety of the latter acts as a vulcanization accelerator.
The results of a more detailed investigation into the properties of films
from artificial synthetic cis-l,4-polyisoprene rubber latex and natural rubber
latex have been reported by Preiss, Sawyer and Simpson [26]. These results
corroborate in a general way the observations outlined above. A gel-free
artificial polyisoprene latex produced by the Shell process was used. It was
found that the tensile properties of the postvulcanized unfilled films from the
synthetic and natural synthetic cis-l,4-polyisoprene latices were similar.
Using a vulcanizing system comprising sulphur, zinc diethyldithiocarbamate
and zinc mercaptobenzthiazolate (but no zinc oxide), the films from natural
rubber latex vulcanized slightly faster at lOO·C, and gave slightly higher
optimum tensile strength and modulus at 300% extension. The 500% moduli
of postvulcanized films from the two types of latex were, however, almost
identical. The extensions at break of the postvulcanized films from the natural
rubber latex were somewhat higher than those of films from the artificial
polyisoprene rubber latex. For both types of film, the extension at break was
found to be essentially independent of the state of vulcanization over the
range investigated.
Preiss, Sawyer and Simpson [26] have also reported results for the effects
of crosslink concentration, determined by measurements of equilibrium
swelling, upon the tensile strength, extension at break and modulus at 500%
extension for sulphur-vulcanized films from artificial synthetic cis- 1,4-
polyisoprene rubber latex. The films contained various levels of sulphur, and
were postvulcanized at l00'C for various lengths of time over the range
10-60 minutes. The results are shown in Figs 12.5(a), (b) and (c) respectively.
They may be compared with the results for unvulcanized latex films shown
in Table 12.5; because the synthetic polyisoprene was gel-free, the crosslink
concentration in films derived from this latex was presumably zero. It is
evident from Fig. 12.5(a) that the tensile strength of films from the artificial
polyisoprene rubber latex was highest when the crosslink concentration was
in the range ca. 1.0- 1.2 x 10 - 4 mol cm - 3, gradually declined as the crosslink
concentration increased over the range 1.2- 1.6 x 10- 4 mol cm - 3, and then
declined sharply with further increase in crosslink concentration. Although
the tensile strength and extension at break were relatively insensitive to
increases of crosslink concentration over the range 1.0- 1.4 x 10 - 4 mol cm - 3,
the modulus did increase significantly.
As regards the tensile properties of unvulcanized films from natural and
artificial cis-l,4-polyisoprene rubber latices shown in Table 12.5, it is seen
that the films from the latter latex had much lower tensile strength, lower
modulus and higher extension at break than did those from natural rubber
latex. These differences are attributed to structural differences between the
dispersed unvulcanized polymers in the two latices. In particular, they are a
consequence of the absence of crosslinks and microgel in the polymer in the
artificial latex. Preiss, Sawyer and Simpson have also shown that small, but
significant, concentrations of crosslinks can be introduced into the polymer
dispersed in artificial synthetic cis-l,4-polyisoprene rubber latex by heating
the latex for various periods of time at 50'C with 0.5 pphr of sulphur and
Table 12.5 Tensile stress-strain properties of unvulcanized films from artificial

synthetic cis-1,4-polyisoprene rubber latex and from natural rubber latex (Preiss,
Sawyer and Simpson [26])
Film from natural Film from artificial

cis- J.4-polyisoprene synthetic cis- J.4-polyisoprene
Property rubber latex rubber latex
Modulus at 300%
extension / MPa 0.55 0.10
Modulus at 500010
extension / MPa 1.03 0.17
Tensile strength / MPa 13.1 0.7
Extension at break / % 1450 2000
50
~ 40...._--.J1L..l!....~
--.s~
CI
30
!
u; 20
.!!1
•
·iii
c:
~ 10
•
o~ __~~__~____~__~~~~
1.0 1.2 1.4 1.6 1.S 2.0
(a) Concentration of crosslinks x 10 4 / mol cm-3
1000
900 •
--~
~
.:.t! SOO •
.0 700
•
iii
.§ 600
II)
Iii 500
Ji 400
300~--~-----L----~ ____~__~
1.0 1.2 1.4 1.6 1.S 2.0
(b) Concentration of crossllnks x 10 4 / mol cm-3
12
ttl •
~
--
10
c:
.Q
c: S
II)
CD
i 6
::!!
0
8
It)
iii 4
II)
:::l
"S
"8
~
2
O~--~~--~----~--~~--~
1.0 1.2 1.4 1.6 1.S 2.0
Concentration of crosslinks x 10 4 / mol cm-3
(c)
Figure 12.5 Effects of crosslink concentration, detennined by measurements of equilibrium

swelling, upon tensile stress-strain properties of sulphur-postvulcanized films from artificial
synthetic cis-I ,4-polyisoprene rubber latex (Preiss, Sawyer and Simpson [26]): (a) tensile strength;
(b) extension at break; (c) modulus at 500"/. extension. Latex formulation (parts by mass dry):
rubber 100, potassium oleate 1.5, sulphur variable, zinc diethyldithiocarbamate 1.5, zinc
mercaptobenzthiazolate 1.5, antioxidant I. Vulcanization conditions: 10-60 minutes at lOOT.
0.75 pphr of zinc diethyldithiocarbamate. In effect, the latex becomes slightly

prevulcanized, as is evident from the results shown in Table 12.6. As the
crosslink concentration increased, so the tensile properties of the dried film
followed the expected trends. However, although films from an artificial
synthetic polyisoprene latex prevulcanized to a crosslink concentration of
ca. 0.05 x 10 - 4 mol cm - 3 approximated closely to those from natural rubber
latex in respect of extension at break and modulus at 500% extension, the
tensile strength was still much below that of unvulcanized films from natural
rubber latex. In this connection, it may be noted that the concentration of
physically-effective crosslinks in the rubber deposited from un vulcanized
natural rubber latex appears to be ca. 0.01 x 10- 4 mol cm - 3.
An extensive comparison between artificial synthetic cis-1, 4-polyisoprene
latices and high-ammonia-preserved natural rubber latex centrifuged concentrate
has been provided by Gorton [28]. This comparison covers not only the raw
latices, but also their behaviour in important rubber latex applications. The
following is a summary of some of the conclusions to be drawn from this
comparison:
1. The viscosity-concentration curve for the artificial latex differs significantly
from that of ammonia-preserved natural rubber latex concentrate, as is
shown in Fig. 12.6. The viscosity of the artificial latex is lower than that
of natural rubber latex at any given total solids content. This is attributed
to differences in particle-size distribution. Electron microscopy has
revealed that, unlike natural rubber latex centrifuged concentrate, the
artificial latex contained predominantly large spherical particles with
very few, if any, small particles.
Table 12.6 Effect of low concentrations of crosslinks upon tensile stress-strain

properties of sulphur-prevulcanized films from artificial synthetic cis-I .4-polyisoprene
rubber latex (Preiss. Sawyer and Simpson [26])
Latex formulation (parts by mass dry):
Rubber 100
Sulphur 0.5
Zinc diethyldithiocarbamate 0.75
Prevulcanization conditions: various times at 50°C.
Modulus at
Crosslink 300"1. 500"1. Tensile Extension

concentrat ion extension extension strength at break
/ molcm- 3 / MPa / MPa / MPa /%
0 0.10 0.17 0.69 2000
2 x 10- 6 0.28 0.48 1.38 1000
4 x 10- 6 0.34 0.55 2.21 1220
6 x 10- 6 0.41 0.62 3.17 1150
10~_~--------~--------~------~
55 60 65 70
Total solids content 1% mlm
Figure 12.6 Viscosity-concentration curves for artificial synthetic cis-l,4-polyisoprene rubber

latex (curve A) and natural rubber latex concentrate (curve B) (Gorton [28]). Viscosities
determined using a Brookfield viscometer at rotational frequency 60 minute - I.
2. The mechanical stability of the artificial latex, as determined by the

conventional high-speed stirring test used for natural rubber latex (see
section 8.7.1.2 of Chapter 8 (Volume 1)), was much higher than that of
the natural rubber latex: 5000 and 1500 seconds respectively are cited.
This is presumably a consequence of the high levels of colloid stabilizers
contained in the artificial latex.
3. Notwithstanding its high mechanical stability, the artificial latex displayed
much lower chemical stability towards ionic substances than did the
natural rubber latex, necessitating that precautions be taken to avoid
colloidal destabilization during compounding.
4. The viscosity of compounded artificial latex increased much less during
storage at normal ambient temperatures than did similarly-compounded
natural rubber latex. In both cases, the increases in viscosity could be
effectively eliminated by addition of small amounts of an ethoxylate
surface-active substance. The compounded artificial latex was less stable
at elevated temperatures than was the compounded natural rubber latex.
5. The equilibrium swelling in benzene Of films from the compounded
artificial latex was significantly greater than that of films from the natural
rubber latex. No results are given for sol contents. The indication from
these results is that, insofar as polymer gel is present in the artificial latex,
the concentration of crosslinks in the gel is less than that in the polymer
gel contained in natural rubber latex.
6. The polymer in the artificial latex had lower resistance to oxidative ageing
than did that in the natural rubber latex. This difference is attributable
to the absence from the artificial latex of the non-rubber substances in
natural rubber latex which are able to function as antioxidants. This is

confirmed by the observation that film from multiply-creamed natural
rubber latex behaved similarly to that from the artificial latex.
7. Unvulcanized films from the artificial latex had lower tensile strength
and modulus than did similar films from natural rubber latex. The results
given for tensile strength are 0.03 and 3.9 MPa respectively, and for
relaxed modulus at 100% extension 0.03 and 0.28 MPa. Extensions at
break were 105% and 815% respectively.
8. The tensile properties of unfilled postvulcanized films from the two types
of latex were broadly similar, but those from the artificial latex tended
to have the lower values. Very roughly, optimum tensile strengths were
ca. 30 MPa and extension at break ca. 800%.
9. The ageing resistance of unfilled postvulcanized films from the artificial
latex was greatly inferior to that of films from the natural rubber latex,
notwithstanding that all the formulations contained 2 pphr of a good
phenolic antioxidant. These results demonstrate the importance of the
non-rubber substances in determining the ageing behaviour of films from
natural rubber latex.
10. As regards the tensile properties of unfilled postvulcanized films from
blends of the two latices, the following changes were observed as natural
rubber latex was progressively replaced by the artificial latex: tensile
strength was reduced somewhat, extension at break increased, modulus,
relaxed modulus and tensile set decreased, and tear strength decreased.
Resistance to oxidative ageing was also progressively reduced.
12.4.3 Artificial latices of cis-l,4-polybutadiene rubbers

Artificial latices of cis-I,4-polybutadiene rubbers, like those of synthetic
cis-I,4-polyisoprene rubbers, have been developed primarily as competitors
for natural rubber latex. Again, they have failed to establish themselves in
any of the conventional applications for natural rubber latex, or in any other
industrial applications. As far as is known, there are no artificial cis-
I,4-polybutadiene latices industrially available at the time of writing. Some
information concerning latices of this type intended for industrial application
is available in the paper by Rogers, Ryan and Hecker [21], which has already
been cited. It appears that the polymer dispersed in this latex vulcanizes very
rapidly if conventional sulphur-vulcanizing systems are used. The consequence
is gross over-vulcanization under normal conditions. A less active vulcanizing
system, comprising sulphur 1, zinc mercaptobenzthiazolate 2.5 and zinc oxide
3 parts by mass dry per 100 parts by mass of rubber, has been found to be
more satisfactory. It is claimed that, by the use of less active vulcanizing
systems, it is possible to obtain films having tensile strengths of up to ca.
14 MPa. Such values are, of course, considerably less than those (up to ca.
35 MPa) which are obtainable from natural rubber, processed either in bulk
1.2
60
r -
~
::?! 1.0
-...
B §
'iii
c:
i 0.8
~
g
00 10 20 30 0.40 10 20 30
(a) Time of vulcanization / minute (b) Time of vulcanization / minute
Figure 12.7 Tensile stress-strain properties of sulphur-postvulcanized films from cis-I,4-

polybutadiene rubber latex (curve A) and natural rubber latex (curve B) (Rogers, Ryan and
Hecker [21]): (a) tensile strength; (b) modulus at 100"10 extension. Latex fonnulation (parts by
mass dry): rubber 100, sulphur I, zinc mercaptobenzthiazolate 2.5, zinc oxide 3, antioxidant I,
diphenyl guanidine 0.6, sodium silicofluoride 1. Vulcanization conditions: various times at lOOT.
form or in latex form. Figures 12.7(a) and (b) show results reported by Rogers,
Ryan and Hecker for the sulphur-post vulcanization characteristics
of films from artificial cis-l ,4-polybutadiene rubber latex using zinc mercapto-
benzthiazolate as the vulcanization accelerator. At (a) is shown the variation
of tensile strength with time of vulcanization at 100°C; at (b) is shown the
variation of modulus at 100% extension. Results for films from natural rubber
latex vulcanized using the same system are included for comparison. The
tendency for the tensile strength of sulphur-vulcanized cis-l ,4-polybutadiene
rubber latex deposits to revert on over-vulcanization is evident from in Fig.
12.7(a). Notwithstanding this tendency, vu1canizates from cis-l ,4-polybutadiene
rubber latices appear to display excellent resistance to deterioration on ageing.
12.4.4 Artificial latices of reclaimed rubbers

Reclaimed rubber has been described by Smith [29] as 'the product resulting
from the treatment of ground vulcanized scrap rubber tires, tubes and
miscellaneous waste rubber articles by the application of heat and chemical
agents, followed by intense mechanical working, whereby a substantial
"devulcanization" or regeneration of the rubber component to its original
plastic state is effected, thus permitting the product to be compounded,
processed and re-vulcanized'. The combined sulphur is not removed from

the product. The principal molecular change brought about by the various
rubber reclaiming processes is scission of the network chains of the vulcanized
rubber. This is what is meant by depolymerization in this context. The term
devulcanization in this context is usually taken to refer to the change in
physical properties, the principal of which is increased plasticity, rather than
to the chemical changes which bring about those changes. The product of a
rubber-reclaiming process is inevitably complex and heterogeneous. The
polymer phase comprises fragments of crosslinked gel dispersed in a plastic
continuum which has suffered so much network scission that it is of finite
molecular size. Also dispersed in the material are particles of the fillers which
were present in the vulcanized rubber from which the reclaim was produced.
These particles are usually present in essentially unchanged form. There are
several processes which have been used for the production of reclaimed
rubber. They all involve comminuting the rubber, removing any unwanted
substances such as metal particles, and then heating with reclaiming agents
such as strong aqueous alkali or acid solutions, rubber solvents and specific
depolymerizing agents such as aromatic disulphides.
There has been considerable industrial interest in artificial latices of
reclaimed rubbers in the past, but there appears to be almost no interest
now. The past interest was primarily in artificial latices of reclaimed rubbers
based predominantly upon natural rubber. Applications for such latices
included tyre-cord adhesives, adhesives for paper, cloth, wood, masonry, etc.,
binders for paper, felt, etc., coatings for paper and cloth, and as extenders
for other rubber latices, notably for natural rubber latex.
A typical process for the manufacture of artificial latices of reclaimed
rubber has been described by Patterson [30]. This is a phase-inversion process
which uses ca. 5 pph of reclaimed rubber of soap as the principal dispersing
agent. Potassium oleate, formed in situ from oleic acid and potassium
hydroxide, is favoured as the soap. A water-soluble hydrocolloid is used as
a 'protective colloid'; casein is the preferred substance, the amount being ca.
3% pph of reclaimed rubber. In a process for the dispersion of alkali-reclaimed
whole-cover reclaim, a conventional dough mixer is used. Some care in the
selection of the most suitable type of mixer blade is necessary. The mixer is
first warmed by passing steam through the outer casing. The reclaimed rubber
is cut into small pieces, placed in the mixer, and allowed to become plastic.
This operation takes ca. 3 minutes. 5 pph of reclaimed rubber of oleic acid
is added and allowed to mix thoroughly, ca. 5 minutes being required. The
potassium hydroxide (1.25 pph of reclaimed rubber) is added next. If this is
added as an aqueous solution, then as soon as the first drops touch the
rubber, the latter breaks up into small pieces which slip over each other,
thereby preventing effective mixing. Gradually the potassium hydroxide
solution is absorbed by the rubber, and reintegration of the batch occurs.
However, disintegration of the batch occurs every time a few drops of the
alkali solution are added, and so the addition of the potassium hydroxide is
a lengthy operation. One satisfactory way of overcoming this difficulty is
said to be to comminute the solid alkali with an equal mass of an inert
non-hygroscopic fine powder, such as zinc oxide or kaolinite clay. It appears
that the fine particles of powder coat the potassium hydroxide particles and
prevent the latter from becoming a sticky mass as they absorb moisture from
the atmosphere. It is claimed that such alkali/powder blends can be stored
for several weeks without becoming sticky, and that, if the alkali is added
to the mixture of reclaimed rubber and oleic acid in this form, the time
required for the addition of the alkali is greatly reduced. Thus after the
addition ofthe alkali/powder blend, only ca. 2 minutes is required for intimate
mixing. After the alkali has been added, the addition of the casein as a
12.5% m/v solution in 0.5% m/v aqueous ammonia is commenced. The casein
solution is added in small amounts. Each addition causes the batch to
disintegrate, with the evolution of steam. No more solution is added until
the batch is once again uniform. The casein solution is absorbed by the batch
increasingly readily as the batch becomes softer. Up until the point of phase
inversion, the batch has the appearance of normal masticated rubber. At the
point of phase inversion, it suddenly becomes very sticky, loses its toughness,
and is no longer elastic. The casein solution is now absorbed much more
readily, and the batch soon becomes paste-like in consistency. At this stage,
the remainder of the casein solution, and also any water which is necessary,
is added as quickly as possible, consistent with maintaining uniformity. The
time taken from the commencement of the addition of the casein solution
to the point of phase inversion is ca. 10 minutes. From the point of phase
inversion to the completion of the batch is another 10 minutes. The total
time for the production of the batch is ca. 30- 35 minutes. Phase inversion
occurs when ca. 15% of the mass of the batch is water; it is therefore possible
in principle by this method to produce artificial latices of reclaimed rubber
which have total solids contents up to ca. 85% m/m. A similar process has
been described in outline by Cubberley [31], who also gives reference to the
United States patent literature. As expected, reclaimed rubbers produced by
heating the rubber with acids are more difficult to disperse in alkaline aqueous
media than are reclaimed rubbers produced by alkali-reclaiming processes.
Patterson [30] has reported the interesting observation that small additions
of certain artificial reclaimed rubber latices to compounded natural rubber
latex can increase the tensile strength of postvu1canized films obtained from
the latex. This effect was observed with artificial latices ofwhole-tyre reclaimed
rubber, but not with those of reclaimed rubber from inner tubes. This
difference in behaviour does not appear to reside in differences in the
respective reclaiming processes, because both types of rubber were reclaimed
by similar alkali processes. The probable explanation is that the whole-tyre
reclaimed rubber contained a higher level of reinforcing carbon black than
did the reclaimed rubber from inner tubes, and that the level of carbon black
in the blend of natural rubber latex and artificial reclaimed rubber latex was
sufficient to effect some reinforcement of films from the latex blend. This
observation is interesting, because inorganic fillers such as carbon blacks do
not usually reinforce natural rubber latex films in the same way that they
do vulcanizates produced by dry-rubber technology. The reason is the absence
in latex technology of any step analogous to mastication and mixing, in
which the intensity of mechanical shearing is sufficient to promote molecular
interaction between the rubber chains and the filler particles (see further
section 16.6.1.1 of Chapter 16 (Volume 3». In the case of an artificial latex
of whole-tyre reclaimed rubber, there will have been ample opportunity for
interaction between the carbon black particles and the rubber phase in which
they were dispersed, not only during the initial compounding of the rubber
but also during the reclaiming process and the subsequent manufacture of
the artificial latex. It may be that the overall effect of interaction between
the carbon-black particles and the rubber of the reclaimed rubber, and the
subsequent interaction between the particles of the reclaimed rubber and the
matrix of latex rubber in which they become embedded, provides sufficient
indirect interaction between the carbon-black particles and the latex rubber
to effect some degree of reinforcement.
12.4.5 Artificial latices of polyurethane polymers

A brief description of a range of artificial polyurethane latices intended for
application as coatings has been given by Taub [32]. The latices described
are apparently produced by a process in which a polyurethane polymer
having incorporated anionic groups is dispersed in an aqueous medium
without the aid of other surface-active substances. An isocyanate-terminated
prepolymer oflow molecular mass is formed from a diol, a dihydroxycarboxylic
acid and a diisocyanate. The carboxylic-acid groups of this prepolymer are
then ionized by reaction with a tertiary amine, dispersed in water, and
chain-extended by reaction with a diamine. Typical diols include poly tetra-
methylene ether diols, polyester diols and polycaprolactone diols. The
molecular mass of the diol ranges from ca. 500 to ca. 3000 daltons, depending
upon the application for which the final polymer is required. The higher is
the diol molecular mass, the more flexible is the final polymer. A suitable
dihydroxycarboxylic acid for this application is dimethylolpropionic acid
(Structure II shows the 2,2-isomer). The amount used determines the colloid
stability of the dispersion produced, and also has an influence upon the
hardness of the final polymer. The diisocyanate can be either aliphatic (e.g.,
hexamethylenediisocyanate) or aromatic (e.g., toluenediisocyanate). The
isocyanate-terminated prepolymer can be formed in the absence of a solvent.
However, it is preferred to prepare the prepolymer in an inert water-miscible
organic liquid; otherwise the high viscosity of the prepolymer hinders
subsequent dispersion and the chain-extension reaction. Being water-miscible,
these liquids assist greatly the formation of the dispersion; in addition to

reducing the viscosity of the phase to be dispersed, they also facilitate the
mixing of that phase with the water. They are known as co-solvents,
presumably because they partition between the polymer phase and the
dispersion medium in the final latex. Suitable organic liquids include
N-methylpyrrolidone, N, N-dimethylformamide, tetrahydrofuran and methyl-
ethyl ketone. These co-solvents can also function as coalescing agents during
subsequent formation of films from the latex. As regards alkalis to ionize the
carboxylic-acid groups of the isocyanate-terminated prepolymer, triethylamine
is suitable if the subsequent surface-coating is to be air-dried, but a less
volatile tertiary amine, such as N, N-diethylethanolamine, is preferred if the
surface-coating is to be baked. The amine-neutralized isocyanate-terminated
prepolymer is preferably dispersed in water by incremental addition to the
water under high agitation. The water can also be added incrementally to
the prepolymer, but this is more difficult because of the high viscosity of the
initial dispersion before phase inversion occurs. The final step is chain-extension
of the dispersed isocyanate-terminated prepolymer by reaction with a diamine
such as hydrazine, ethylenediamine or methylene-bis(4-cyclohexylamine). The
properties of typical artificial polyurethane latices made by this technique
are given in Table 12.7. Information concerning artificial polyurethane latices,
and the films derived from them, is also available in a paper by Ney, Regos
and Labb [33].
12.4.6 Artificial latices of other polymers
J2.4.6. J Artificial latices of ethylene-propylene rubbers

Attempts have been made to develop and market artificial latices of ethylene-
propylene rubbers. These latices are of interest in the same applicational
areas as are latices of isobutene-isoprene rubbers. In particular, they are of
interest as tyre-cord adhesives. The adhesion to synthetic fibres, such as
polyamide and polyester, is said to be very good. The Burke- Palmason
Chemical Company [22,23] have developed an artificial latex of an
ethylene-propylene diene rubber which seems to resemble closely their
isobutene-isoprene rubber latices (see section 12.4.1 above). The latex is
anionic, having total solids content 49-51% mlm, pH 4.5-6.0, average
particle diameter ca. 1 J,lm, and viscosity (Brookfield LVT # 3 spindle,
rotational frequency 12 minute - I, 24°C) 0.2-0.4 Pa s. It is colloidally stable
over the pH range 2-12. Ethylene-propylene diene rubbers resemble
isobutene-isoprene rubbers in that they are essentially saturated hydrocarbon
polymers which contain low concentrations of olefinic unsaturation introduced
by copolymerization. However, there can be a fundamental difference between
the two types of rubber in respect of the relationship of the residual olefinic
unsaturation to the polymer chain; this can have important consequences
Table 12.7 Properties of artificial polyurethane latices made by the anionic self-
emulsification process, and of films derived from the latices (Taub [32])
Co-solvent: N-methylpyrrolidone
Latices prepared using an Latices prepared using an

aliphatic diisocyanate aromatic diisocyanate
Property A B C D E F
Latex properties
/%m/m 40 30 30 33 33 33
Water/co-solvent
ratio / m/m 76/24 80/20 75/25 70/30 85/15 73/27
Kinematic viscosity
/cm 2 s- 1 10-35 1 20 10 2 10
pH 8 8 8 9 8 9
Appearance Hazy Hazy Clear Hazy Hazy Hazy
Film properties
Modulus at 100%
extension / MPa 2.07 7.58 20.68 3.45 16.55 20.68
Modulus at 300%
extension / MPa 3.45 6.89 31.03
Tensile strength / MPa 24.1 38.6 48.3 27.6 40.0 41.4
Extension at break / % 830 410 300 550 360 300
Shore hardness· 60A 40D 50D 68A 45D 48D
Tear strength,
die C / Ncm- I 683 1051 350 876 1051
• The Shore A scale is used for soft elastomeric polymers, whereas the Shore D scale is used
for harder polymers.
for the behaviour of the vulcanized rubber in service. As their generic name
implies. the olefinic unsaturation is introduced into isobutene-isoprene
rubbers by copolymerization with a small amount of isoprene. As such, the
residual olefinic unsaturation is mostly in the main chain of the polymer
(insofar as the isoprene units polymerize in the 1,4 mode). Ozonolytic
degradation of the polymer therefore brings about scission of the main chain.
For the ethylene-propylene diene rubbers. a wide range of diene comonomers
has been investigated for introducing olefinic unsaturation. If the residual
unsaturation is present in side groups. and not in the main chain of the
polymer. then ozonolytic degradation of the polymer is not accompanied by
scission of the main chain. One diene comonomer which introduces residual
unsaturation in this way, and which is said to be widely used for the
production of these rubbers. is 5-ethylidene-2-norbornene (2-ethylidene-
bicyclo(2.2.1 )-5-heptene) (Structure VIlla). The units which become incorporated
in the polyethylene-propylene chain are said to be mainly of the type shown
as Structure VIIlb. i.e.• they are incorporated mainly through the carbon-carbon
"""q"""
$CH-CH,
CH-CH 3
Villa Vilib
double bond of the ring. Whether or not this is the type of residual olefinic
unsaturation in the polymer dispersed in the currently-available ethylene-
propylene diene rubber artificial latex does not appear to have been
disclosed.
12.4.6.2 Artificial latices of chlorosulphonated polyethylene rubbers

Chi oro sulphonated polyethylene rubber is produced from polyethylene by
reaction in solution with chlorine and sulphur dioxide in an appropriate
ratio. The product is a polymer of the type indicated in Structure IX. Some
of the hydrogen atoms have been replaced by chlorine atoms, and a smaller
number by chlorosulphonyl groups. Chlorosulphonation destroys the regularity
of the polyethylene sequence, thereby inhibiting the tendency of the polymer
to crystallize. The intrinsic flexibility of the polyethylene chain is then
manifest; thus the polymer behaves as a rubber at normal temperatures,
rather than as a plastic. Chlorosulphonation also introduces into the polymer
chain reactive sites by means of which the rubber can be crosslinked.
Vulcanization is usually effected by heating with metal oxides, such as
magnesium oxide. The range of chlorine contents of the grades industrially
available is ca. 25-45% m/m, and the range of sulphur contents ca.
1.0-1.5% m/m. For a chlorosulphonated polyethylene which contains 35% m/m
of chlorine and 1% m/m of sulphur, ca. 20% of the carbon atoms in the
polyethylene chain have been either chlorinated or chlorosulphonated. Of
these 20% of carbon atoms, ca. I in 30 has been chlorosulphonated .
. . . -CH 2-CH 2-' ..... -CH-CH 2-' ..... -CH-CH2-' ..
I I
CI S02CI
IX
Chlorosulphonated polyethylene is a special-purpose synthetic rubber
which gives vulcanizates having excellent chemical resistance, resistance to
weathering, and, in particular, resistance to ozonolytic degradation. If
required for processing as latices, it is necessary to disperse it in an aqueous
medium to form an artificial latex. du Pont [34] have published guidelines
for the manufacture of a chlorosulphonated polyethylene artificial latex by
a solution-emulsification process. The rubber is dissolved in 665 pphr of
I, I, I-trichloroethane together with 50 pphr of a fluorinated solvent to assist
in reducing the particle size of the final latex. A small amount (0.8 pphr) of
n-octyl alcohol is added to reduce foaming. This solution is then emulsified

under high shear in an aqueous solution comprising 500 pphr of water, 8 pphr
of the ammonium salt of a sulphate ester of an ethoxylated nonylphenol,
1 pphr of a dispersing agent and 0.8 pphr of sodium nitrite as a polymer
stabilizer during the subsequent stripping stage. The solvent is then stripped
by steam-distillation at atmospheric pressure. Considerable dilution of the
latex occurs. The total solids content after removal of the solvent is said to
be ca. 8% m/m, and the pH 5.7. The latex is concentrated by creaming using
hydroxyethylcellulose as creaming agent.
12.4.6.3 Other artificial latices

According to Athey [35], there has been some interest in aqueous dispersions
of alkyd-modified drying oils. The disperse phase is produced by reacting
the drying oil with a multifunctional alcohol and a multifunctional carboxylic
acid in a suitable water-insoluble solvent. Typically, a triol and phthalic acid
are used. An excess of carboxylic-acid groups over hydroxyl groups should
be used, so that the product contains unreacted carboxylic-acid groups. If
not already water-soluble, the final product can then be dispersed in water
ifit is first neutralized to a pH of8.5 or higher by addition of alkali, preferably
an amine. For toxicological reasons, the preferred amines are ethanolamine
derivatives. After neutralization of the carboxylic-acid groups, the polymer
solution spontaneously forms an anionic emulsion when mixed with water.
The polymer solution also contains some free fatty acid, formed by ester
interchange between the drying oil and the phthalic acid. This fatty acid
subsequently becomes converted to fatty-acid soap when the alkali is added.
This soap assists with the emulsification of the polymer solution; it also
enhances the colloid stability of the emulsion and of the latex which is formed
when the solvent has been removed by stripping. Being polyesters, the
polymers contained in these latices are in principle subject to alkaline
hydrolysis by the aqueous phase of the latex. In practice, such hydrolysis is
likely to occur only slowly, unless excess alkali is present. These latices are
of interest as vehicles for water-based surface coatings. After deposition on
a substrate, the polymers harden by oxidation-induced crosslinking, the
process being accelerated by the inclusion of various transition-metal
compounds as oxidation catalysts.
Procedures for the preparation of artificial latices of epoxide prepolymers
and of ethylcellulose have been described by Vanderhoff, EI-Aasser and
Ugelstad [16]. Typically, the epoxide prepolymer is emulsified as a 25% m/m
solution in a 50/50 v/v mixture of toluene and methylisobutyl ketone using
sodium n-dodecyl sulphate and n-hexadecyl alcohol as emulsifying agents.
The solvent is removed by distillation after emulsification is completed. A
similar technique is said to be applicable to ethylcellulose. A suitable solvent
References 439
for this polymer is a 70/15/15 v/v /v mixture of toluene, methyl alcohol and
dichloromethane; a suitable polymer concentration in this solvent is
15% m/m.
REFERENCES
1. Sievier, R. W., British Patent No. 7015, 27 August, 1836.

2. Alexander, P., British Patent No. 14,681, application date 17 July, 1905, acceptance
date 26 October, 1905.
3. A[exander, P., British Patent No. 25,735, application date 14 November, 1906,
acceptance date 10 October, 1907.
4. H. O. Traun's Forschungs[aboratorium GMBH, British Patent No. 156,150,31
March, 1922.
5. Pratt, W. B., British Patent No. 2[7,612, 19 June, 1924.
6. Pratt, W. B., British Patent No. 233,370, 7 May, [925.
7. K[ein, P., and Szegvari, A. / The Anode Rubber Company Ltd., British Patent
No. 296,685,4 September, 1928.
8. Madge, E. W., (1943) Transactions of the Institution of the Rubber Industry, 19, 103.
9. Trumbull, H. L. and Dickson, J. B./The B. F. Goodrich Company, United States
Patent No. 1,513,139, 28 October, 1924.
10. Murphy, E. A. and Madge, E. W./Dun[op Rubber Company Ltd., British Patent
No. 445,541, 14 April, 1936.
11. Hunter, E. A., Segura, M. A., Small, A. B. and Miller, A. L./Esso Research and
Engineering Company, United States Patent No. 2,944,038, 5 Ju[y, [960.
12. Hunter, E. A., Small, A. B. and Miller, A. L./Esso Research and Engineering
Company, United States Patent No. 3,062,767, 6 November, 1962.
13. Beerbower, A., Burton, G. W. and Malloy, P. L./Exxon Research and Engineering
Company, British Patent No. 1,497,757, 12 January, 1978.
14. Le Suer, W. M. and Norman, G. R./The Lubrizo[ Corporation, United States
Patent No.3, 172,892,9 March, 1965.
15. Le Suer, W. M. and Norman, G. R./The Lubrizo[ Corporation, United States
Patent No. 3,448,048, 3 June, 1969.
16. Vanderhoff, J. W., E[-Aasser, M. S. and Uge[stad, J., United States Patent No.
4,177,177, 4 December, 1979.
17. Dieterich, D. (1981) Progress in Organic Coatings, 9, 281.
18. Wicks, Z. W. (1975) Progress in Organic Coatings, 3, 73.
19. Wicks, Z. W. (1981) Progress in Organic Coatings, 9, 3.
20. Vanderhoff, J. W., E[-Aasser, M. S. and Hoffman, J. D., United States Patent No.
4,070,323, 24 January, 1978.
21. Rogers, T. H., Ryan, A. O. and Hecker, K. C. (1962) Rubber World, 147(1), 86.
22. Technica[ literature published by Burke- Pa[mason Chemical Company, F[orida,
undated.
23. Burke-Pa[mason Chemical Company (1987) Latexes Produced by Solvent
Polymerization, in The Vanderbilt Latex Handbook, 3rd edn (ed. R. F. Mausser),
R. T. Vanderbilt Company Inc., Norwalk, Connecticut, Chapter 4.
24. Miller, A. L. and Powers, K. W. (1959) Rubber Age, New York, 86, 89.
25. Wales, M. (1962) Journal of Physical Chemistry, 66, 1768.
26. Preiss, D. M., Sawyer, W. M. and Simpson, W. C. (1963) Journal of Applied
27. Schmit, K. H. (1962) Rubber Age, New York, 91, 437.

28. Gorton, A. D. T. (1970) Rubber Chemistry and Technology, 43, 1255.
29. Smith, F. G. (1978) Reclaimed Rubber, in The Vanderbilt Rubber Handbook (ed.
R. O. Babbit), R. T. Vanderbilt Company Inc., Norwalk, Connecticut, p. 319.
30. Patterson, P. D. (1932) Transactions ofthe Institution of the Rubber Industry, 8, 80.
31. Cubberley, R. H. (1954) Reclaim Rubber Dispersions, in The Vanderbilt Latex
Handbook (ed. G. G. Winspear), R. T. Vanderbilt Company Inc., New York, p. 75.
32. Taub, B. (1983) Water Dispersible Urethanes, paper presented at Water-Borne
and Higher-Solids Coatings Symposium, New Orleans, Louisiana, 1983.
33. Ney, E. A., Regos, N. and Labb, P. (1976) Journal of Elastomers and Plastics, 8, 210.
34. Guidelines for the Manufacture of Hypalon Latex, E. I. du Pont de Nemours,
Wilmington, Delaware, undated.
35. Athey, R. D. (1991) Emulsion Polymer Technology, Marcel Dekker Inc., New York,
p.103.
13
Chemically-modified latices:
1. Prevulcanized latices
13.1 INTRODUCTION
The subject of this chapter and the next is chemically-modified latices. These
are defined as latices which contain polymer particles which have been
sUbjected to some form of chemical modification after the latex has been
formed as such. Thus, for example, functionalized synthetic latices which
have been produced in the usual manner (i.e., as described in section 10.3 of
Chapter 10) are not usually regarded as being chemically-modified,
notwithstanding that the dispersed polymers have been chemically modified
relative to the polymers dispersed in equivalent non-functionalized latices.
However, a functionalized latex would be regarded as having been chemically
modified if the functional groups had been introduced into the dispersed
polymer after the latex had been prepared as such. A hypothetical example
of a preparative procedure of this latter type is one in which the first step is
the production by emulsion polymerization of a functionalized latex which
contains a polymer having pendant functional groups capable of being
hydrolysed to carboxylic-acid groups; then, in a second step carried out after
the emulsion polymerization is complete, some or all of those functional
groups are hydrolysed to carboxylic-acid groups.
There are two further general comments to be made concerning chemically-
modified latices:
1. It is usually understood that almost all the particles have been modified
in the same manner, but not necessarily to the same extent; nor is the
extent of chemical modification necessarily uniform throughout the
individual particles. Little is known concerning the uniformity of modification
either between particles or within particles. If, as is usual, the reagents

442 Chemically-modified latices: 1
which bring about the modification are initially present in the aqueous
phase of the latex, then there is a prima facie reason for doubting that the
extent of modification is uniform throughout individual particles; it seems
more likely that the extent of modification will be greatest at the particle
surface, and then decrease progressively towards the centre of the particle.
2. It is also usually understood that the chemical modification is deliberate,
and has been effected for some technological reason. Latices containing
polymers which happen to have undergone some undesirable modification
in the course of storage or processing are not usually regarded as being
chemically modified in the sense in which the term is commonly used.
Examples of latices containing polymers which may undergo unwanted
chemical changes include polychloroprene rubber latices and latices of
vinyl acetate polymers and copolymers; the dispersed polymers in both
types are susceptible to slow hydrolysis (see sections 11.4.3 and 11.5.4
respectively of Chapter 11).
By far the most industrially-important type of chemically-modified latex

is that commonly known as prevulcanized natural rubber latex, produced
almost exclusively by sulphur-prevulcanization of natural rubber latex. For
this reason, a separate chapter is devoted to such latices; other types of
chemically-modified latex are considered in the following chapter. Although
it is possible to prevulcanize several types of diene-rubber latex, it is only
prevulcanized natural rubber latex which has achieved any industrial importance.
Vulcanization is a process to which most rubbers are subjected in order
to convert them from plastic temperature-sensitive materials to technologically-
useful elastic materials. As outlined in section 1.9 of Chapter 1 (Volume I),
in conventional rubber technology the rubber is first softened and mixed
with compounding ingredients appropriate for the required application, then
shaped, and then vulcanized. Chemically, the vulcanization process entails
crosslinking the individual polymer chains of the rubber matrix sufficiently
to combine the hitherto individual polymer molecules into a lightly-crosslinked
network of unlimited extent. The physical consequence of the chemical change
is that the material is converted from a plastic material which can flow under
shear stress to a viscoelastic material in which the elastic character predominates
over the viscous character. The usual manner of effecting crosslinking of general-
purpose diene-derived rubbers, such as natural rubber and the styrene-
butadiene rubbers, is by heating with sulphur and organic vulcanization
accelerators.
It is a curious, and still largely unexplained, property of natural rubber
latex, and also of certain other types of diene-rubber latex, that it is possible
to vulcanize the individual particles in the latex without destroying the
colloidal character of the latex as a stable hydrophobic sol. The product is
a colloidally-stable latex, the dispersed polymer of which is crosslinked into
networks of indefinite extent. It is this product that is commonly known as
Introduction 443
prevulcanized latex, although some consider that a more appropriate name

is vulcanized latex. In any event, it has become common practice now to
abbreviate the name of the product obtained from natural rubber latex as
PVNR latex, or PVNRL.
The following two unexpected features of prevulcanized natural rubber
latex call for comment at this point:
I. At least as far as sulphur-prevulcanization is concerned, the reaction in
latex occurs with much greater ease than would be expected from
experience with the sulphur-prevulcanization of dry natural rubber.
2. Because the segmental motion of the polymer molecules within the
individual particles is inevitably restricted by the presence ofthe crosslinks,
it would not be expected that prevulcanized natural rubber latex would
dry down to coherent strong films. In fact, it does dry down to such films,
the general character ofthese films being similar to that oflightly-vulcanized
rubber films produced from dry rubber or from natural rubber latex by
postvulcanization.
Both these matters are discussed further in sections 13.2.6 and 13.2.8
respectively below. At this stage, it is sufficient to note the expectation that
latex prevulcanization would not occur under the conditions usually employed
for the reaction. Furthermore, if it did occur, then the product would not be
expected to have practical utility, because it should not form strong, coherent,
integrated films on drying.
In appearance, prevulcanized natural rubber latices are very similar to
un vulcanized natural rubber latex. The original fluidity is retained. It appears
that the crosslinking of the rubber takes place in each individual particle
separately, without interaction with other particles, and without altering the
state of dispersion of the particles appreciably. Thus the particles in a
prevulcanized natural rubber latex have essentially the same shape, size and
size distribution as did those in the initial unvulcanized latex. Furthermore,
the particles in the vulcanized latex exhibit a similar Brownian motion.
There are three principal ways in which natural rubber latex can be
prevulcanized. They are:
t. by allowing the latex particles to react with sulphur and one or more
organic vulcanization accelerators, and possibly also with an inorganic
vulcanization activator;
2. by heating the latex with substances such as organic peroxides and
hydroperoxides, and possibly also with activators for those compounds;
3. by exposing the latex to high-energy radiation in the form of electrons or
electromagnetic waves.
These three ways of effecting latex prevulcanization provide the main divisions
for this chapter. Of these three ways, sulphur-prevulcanization is by far the
most important industrially. Almost all the prevulcanized natural rubber
latex produced at present is made by this method. There has for many years
been interest in alternative processes. The search for such processes has been
motivated in recent years by the problem of the formation of N-nitrosamines
in sulphur-vulcanized rubber. This problem is discussed in section 16.3.6 of
Chapter 16 (Volume 3). It is sufficient to note here that N-nitrosamines are
formed from the organic compounds commonly used to accelerate the sulphur-
prevulcanization and postvulcanization of natural rubber in latex form, and
that their presence constitutes a health hazard. The health hazard is especially
serious in rubber products which are to come into contact directly or indirectly
with the human body. Although this problem can be mitigated to some extent
by appropriate choices of vulcanization accelerators, there has nevertheless
been considerable interest in finding ways of prevulcanizing natural rubber
latex without the use of sulphur and organic vulcanization accelerators.
As has already been noted above, industrial interest in latex prevulcanization
is confined almost exclusively to natural rubber latex. This situation has
arisen from a conjunction of three circumstances:
1. Natural rubber latex is eminently suited for the production of general-
purpose thin-walled rubber products by the process of latex dipping. As
far as is known, apart from speciality products which require polymer
properties not offered by natural rubber, no other type of rubber latex is
used for dipping.
2. Partially-prevulcanized natural rubber latex has been found to be capable
of yielding, by dipping, thin films which have very satisfactory mechanical
properties, especially if those films have been further vulcanized. The
principal mechanical properties of interest in applications of thin rubber
films are tensile strength, tear strength, puncture strength, and extension
at break. It is questionable whether the mechanical properties of
post vulcanized films produced from unvulcanized natural rubber latex are
generally superior or inferior to those of films from prevulcanized natural
rubber latex. One might be inclined to the former view. What is not open
to dispute is that the mechanical properties of films obtained from
prevulcanized natural rubber latex are generally perceived to be adequate
for the applications for which they required.
3. In the context of latex dipping, the use of partially-prevulcanized natural
rubber latex offers a considerable economic advantage relative to post-
vulcanizable natural rubber latex, in that its use enables a large quantity
of rubber to be partially vulcanized as bulk latex, instead of having to be
entirely vulcanized when spread out as a thin film over the surface of
innumerable formers.
It seems probable that most latices containing olefinically-unsaturated
rubbers are capable of being prevulcanized if heated with appropriate
reagents. However, there appears to have been little industrial or academic
interest in such reactions or their products, presumably because of the lack
Sulphur-prevulcanization of natural rubber latex 445
of industrial applications for the products. As will appear subsequently

(section 13.2.6 below), it is known that both a synthetic polyisoprene latex
(i.e., a polyisoprene rubber latex produced by emulsion polymerization) and
an artificial polyisoprene rubber latex (i.e., a polyisoprene rubber latex
produced by dispersing polyisoprene in an aqueous medium) are capable of
being pre vulcanized by heating with sulphur and organic vulcanization
accelerators in much the same way as natural rubber latex. Amongst the
other rubbery polymers which have been claimed to be successfully vulcanized
in latex form are styrene- butadiene rubbers, acrylonitrile- butadiene rubbers,
and isobutylene-isoprene (butyl) rubber. The prevulcanization of other
rubber latices is discussed briefly in section 13.5 below.
One final introductory comment is necessary concerning sulphur-
prevulcanized latices. This is that, as noted in section 15.1 of Chapter 15, it
is common practice to allow a compounded rubber latex to stand for a period
at normal ambient temperature before being used in the process for which
it is intended. Certain subtle changes occur in the latex during this period.
These changes are known collectively as maturation. It has been found
generally advantageous to allow opportunity for such changes to occur before
the compounded latex is processed further. In so far as sulphur and organic
vulcanization accelerators are present during maturation, one of the changes
which occur during this period is that the rubber in the particles become
prevulcanized to a small extent. No further reference is made to this matter
in this chapter, but it should be understood that this aspect of the maturation
process is presumed to ha ve all the characteristics of sulphur-prevulcanization
described below, even though the degree of prevulcanization which occurs
during maturation may be very slight.
13.2 SULPHUR-PREVULCANIZATION OF NATURAL RUBBER

LATEX
13.2.1 History of sulphur-prevulcanization of natural rubber latex

The possibility of vulcanizing the disperse phase of natural rubber latex
without any concomitant colloidal destabilization was first investigated by
Schidrowitz [1-3] in the early years of this century. His interest was in
making a cellular rubber directly from natural rubber latex. He had the idea
that the production time would be significantly reduced if the initial latex
were a dispersion of particles of vulcanized rubber. In this way, the
requirement for subsequent vulcanization would be obviated. It has subsequently
been recognized that prevulcanized natural rubber latex is not a suitable
material for the production of latex foam rubber; it has never been widely
used for this purpose. Thus, although Schidrowitz did succeed in prevulcanizing
natural rubber latex, he did not achieve his principal objective. The method
of prevulcanizing natural rubber latex disclosed in the first of his patents [1]
was to heat the latex with sodium polysulphide, sulphur and zinc oxide at
I45"C for ca. 30-45 minutes. At the conclusion of the process, it was necessary
to reduce the steam pressure very slowly to prevent the latex from boiling
over. Relatively high vulcanization temperatures were necessary because very
active organic accelerators of sulphur-vulcanization were not available at
first. Subsequently, water-soluble accelerators of high activity became available.
By their use, it was possible to reduce the intensity of the vulcanization
conditions to, say, 60 minutes at 70-80·C, preceded by a I-hour rise to that
temperature. In the years since the experiments of Schidrowitz, prevulcanized
natural rubber latex has become widely used industrially. It has become
available commercially as such for users who do not wish to prepare their
own prevulcanized latex.
13.2.2 Preparation of sulphur-prevulcanized natural rubber latex

Sulphur-prevulcanized natural rubber latex is invariably prepared by allowing
the rubber molecules in the particles to react with sulphur under the influence
of one or more organic accelerators, and possibly of an inorganic vulcanization
activator as well. These reagents are present in the aqueous phase of the
latex. It is not usual to allow sulphur-prevulcanization to proceed to the
maximum extent which is possible. What is usually sought is an optimum
balance between, on the one hand, the convenience of effecting as much as
possible of the total vulcanization of the final product at the prevulcanization
stage, and, on the other, maximum desirable mechanical properties of the
final product. It is common practice to achieve this optimum balance by
partially prevulcanizing the latex, and then completing the vulcanization of
the rubber after the final product has been formed from the latex. It may be
necessary to add further vulcanizing ingredients to the latex after prevulcanization
and before the final processing, in order to achieve the desired degree of
vulcanization in the final product.
Accelerators for the sulphur-vulcanization of diene-based rubbers in latex
form are discussed in section 16.3 of Chapter 16 (Volume 3). Those most
commonly used for the sulphur-prevulcanization of natural rubber latex are
the metal dialkyldithiocarbamates. Thiazoles have been used as secondary
accelerators, but have fallen into some disfavour in recent years because of
associated health hazards. Both dialkyldithiocarbamates and thiazoles are
available as either water-insoluble or water-soluble variants; both variants
can be used. Examples of water-insoluble accelerators are zinc diethyl-
dithiocarbamate and zinc mercaptobenzthiazolate; the corresponding sodium
salts provide examples of water-soluble accelerators. Water-insoluble accelerators
are added as aqueous dispersions, water-soluble accelerators as aqueous
solutions. Although they are less convenient to add to the latex, the water-
insoluble accelerators appear to be generally preferred, partly because they
are not removed from the subsequent rubber film during the water-leaching
which usually precedes the final postvulcanization. Not only is the presence
of residual accelerator from the prevulcanization stage desirable to assist
with the final postvulcanization reaction; it is also found that residual
dithiocarbamate accelerators function as efficient antioxidants in films derived
from natural rubber latex. The presence of residual accelerators therefore
advantageously retards the oxidative degradation of such films. This is
important for very thin films, because the ratio of surface area to mass is
high, and thus oxygen from the air has ready access to virtually all the rubber
in the film. However, removal of residual dithiocarbamate during leaching
may be advantageous if it is desired to produce thin rubber films having
maximum optical clarity. A further advantage of the water-insoluble zinc
dialkyldithiocarbamates over the water-soluble equivalents is that they tend
to be more active as accelerators of sulphur-prevulcanization, at least if no
inorganic activator is present in the reaction system.
The inorganic vulcanization activator is usually a sparingly-soluble metal
compound, and thus has to be added to the latex as an aqueous dispersion.
The substance most commonly used for this purpose is zinc oxide; zinc
carbonate is an alternative sometimes used. An important function of the
zinc compound is to regenerate the accelerator as the reaction proceeds (see
section 13.2.5.3 below). Occasionally, water-soluble zinc salts, such as zinc
acetate, have been used as vulcanization activators. However, it is to be
expected that the presence of such compounds would drastically reduce the
colloid stability of the latex. The use of water-soluble activators might be
advantageous if maximum optical clarity is required in the eventual rubber
film, because much of the zinc compound should then be removed during
water-leaching. The presence of zinc oxide or zinc carbonate in the eventual
rubber film is, of course, detrimental to optical clarity.
Although all processes for the sulphur-prevulcanization of natural rubber
latex conform to the outline description given in the preceding paragraphs,
there are several features of the process which are subject to variation. These
include:
1. the level of sulphur;
2. whether the vulcanization accelerators are to be water-soluble or water-
insoluble, or a combination;
3. the natures and levels of the accelerators;
4. the particle size of the sulphur dispersion;
5. the particle size of dispersions of any water-insoluble accelerators which
are used;
6. whether or not to use an inorganic activator, and, if so, the type and amount;
7. the temperature-time profile which is imposed for the reaction;
8. the extent to which the reaction is allowed to continue.
It is clear from this list that many variants of the sulphur-prevulcanization
process are possible. The effects of some of these variables upon the
prevulcanized latex itself, and upon films obtained from the latex, will become
apparent in what follows.
In a typical prevulcanization process, the latex is first colloidally stabilized
by the addition of small amounts of a caustic alkali and a hydrocolloid
stabilizer. Suitable hydrocolloid stabilizers include sodium caseinate and
sodium carboxymethylcellulose. Ethoxylate stabilizers are not generally
suitable for this purpose, because their ability to stabilize hydrophobic sols
is lost at elevated temperatures. After the latex has been colloidally stabilized,
it is then partially de-ammoniated by aeration, and compounded with
appropriate amounts of sulphur, accelerator and zinc oxide. The purpose of
reducing the ammonia content and replacing it with a fixed alkali is to
minimize the risk of zinc-oxide thickening when the latex is subsequently
heated to effect vulcanization. The amounts of the vulcanizing ingredients
are somewhat in excess of those required to achieve the desired degree of
vulcanization. These ingredients are added as relatively coarse slurries, rather
than as the fine dispersions which are commonly used for compounding
latices with water-insoluble powders (described in section 15.3 of Chapter 15
(Volume 3», in order that the excess vulcanizing ingredients can be readily
separated from the latex at the end of the reaction, either by sedimentation
or by centrifugation. It may be desirable to warm the compounded latex for
ca. 30 minutes at ca. 30·C prior to the addition of the vulcanizing ingredients.
This can have the effect of reducing the variability of the latex to the
vulcanization reaction by accelerating certain chemical changes which tend
to occur slowly at normal ambient temperatures. A typical formulation for
the preparation of prevulcanized natural rubber latex is shown in Table 13.1.
Vulcanization is effected by heating the latex slowly to a temperature in the
range 55-S0·C, and then maintaining the temperature at this level until the
desired degree of vulcanization has been achieved. Constant stirring must
be maintained, with especial attention to the need to keep the coarse particles
Table 13.1 Typical formulation for prevulcanization of natural rubber latex

Parts by mass
Ingredient Dry Wet
Natural rubber (as 60% mlm ammonia-preserved

centrifuged latex) 100 167
Potassium hydroxide (as 10% mlm aqueous solution) 0.3 3
Hydrocolloid stabilizer (as 10% mlm aqueous solution) 0.2 2
n
Sulphur ) as 50% mlm
Zinc diethyldithiocarbamate aqueous 8
Zinc oxide slurry
of sulphur, accelerator and zinc oxide in suspension. Care must also be taken
to minimize skin formation. A typical vulcanizer suitable for carrying out
this process on an industrial scale is one which is jacketed for steam heating
and water cooling. The temperature should preferably be controlled
thermostatically, and the vulcanizer equipped with a mechanical stirrer.
Various tests are available for ascertaining the degree of vulcanization which
has been attained. The more important of these are described in section
13.2.3 below. When the desired degree of vulcanization has been attained,
the latex is cooled, run off into containers, strained and then bulked. Residual
vulcanizing ingredients sediment during this time, and so can be eliminated
at this stage. If removal of residual vulcanizing ingredients is very important,
e.g. if optimum optical clarity in the eventual rubber film is required, then
the latex may also be centrifuged. It is preferable to allow the prevulcanized
latex to mature for a period of ca. 7 days at normal ambient temperature
before releasing it for further factory processing. During this period, certain
changes occur in the latex, the consequences of which can be an initial drop
in the tensile strength of dried films, followed by a gradual increase. Little
further change occurs after ca. 7 days, even during prolonged storage. Overall,
the advantages which accrue from this further period of maturation comprise
improvements in uniformity, both as regards subsequent processing behaviour
and the mechanical properties of the films which are formed from the latex.
In a variant of this process, the latex is heated with the minimum amounts
of vulcanizing ingredients required to attain the desired degree of
prevulcanization. If they are water-insoluble, the vulcanizing ingredients are
added to the latex as fine dispersions of the type commonly used for
compounding latices with water-insoluble powders (see section 15.3 of
Chapter 15 (Volume 3)). This method depends upon the assumption that a
high proportion of the vulcanizing ingredients will be consumed during the
vulcanization reaction. Separation of unreacted vulcanizing ingredients is
therefore unnecessary. Having these ingredients present as fine dispersions
ensures that, in comparison with coarse dispersions, availability for reaction
is increased and tendency to sediment is reduced.
It has become increasingly realized in recent years that, to effect satisfactory
prevulcanization of natural rubber latex, there is no necessity to heat the
compounded latex to the high temperatures which have previously been
regarded as essential. It has been found possible to produce satisfactorily
prevulcanized natural rubber latex by allowing the compounded latex to
stand for a long period at warm ambient temperature, such as is normal in
tropical regions, provided that one of the more active zinc dialkyldithiocar-
bamates, such as zinc di-n-butyldithiocarbamate, is used as the principal
accelerator.
It is said to be possible to sulphur-prevulcanize natural rubber latex
satisfactorily using a so-called suIphurless vulcanizing system, in which a
thiuram polysulphide, e.g., tetraethylthiuram disulphide, is used as a sulphur-
donor. Presumably an activator, such as thiourea, is also required. Tensile

strengths of films from latices prevulcanized in this way appear to be
somewhat lower than those from conventional sulphur-prevulcanized natural
rubber latex. As far as is known, such vulcanizing systems are not used
industrially.
In principle, it would be possible to prepare a sulphur-prevulcanized latex
by a continuous vulcanization process, rather than by the batch process
described above. However, as far as is known, such processes are not used,
possibly because the scale of utilization of the product does not warrant
continuous production. There appears to be no reason why the procedures
outlined in section 10.2.4.4 of Chapter 10 for the production of synthetic
latices by continuous emulsion polymerization should not be used for this
purpose. The principal advantage would be the elimination of batch-to-batch
variations, with potential concomitant improvement in product uniformity.
However, if a single continuous-stirred-tank reactor were used, there would
be a wide distribution of residence times in the reactor. This would be expected
to result in wide differences in degree of vulcanization between the various
particles. Although this might at first sight be perceived as disadvantageous,
it might not necessarily be so. As discussed in section 10.2.4.4 of Chapter 10,
the distribution of residence times could be reduced by using a train of
continuous-stirred-tank reactors in series, and reduced to effectively zero by
using some form of tubular reactor.
13.2.3 Methods for assessing the degree of vulcanization of prevulcanized

There are four principal types of test which are commonly used to assess the
degree of vulcanization of prevulcanized natural rubber latex. They are:
l. the chloroform-coagulation test, in which the latex is stirred with a fixed
proportion of chloroform, and the degree of vulcanization judged from
the appearance of the coagulum which forms;
2. the equilibrium-swelling test, in which the degree of vulcanization is judged
from the extent of equilibrium swelling of films dried down from the latex,
the swelling being carried out under controlled conditions using a suitable
rubber solvent;
3. the relaxed-modulus test, in which the degree of vulcanization is judged
from the relaxed modulus at 100% extension (M R 1(0) of films dried down
from the latex; and
4. the prevulcanizate relaxed-modulus (PRM) test, in which the degree of
vulcanization is again judged from the relaxed modulus at 100% extension
(MR 1(0) of films dried down from the latex.
It might be thought that the determination of unreacted (or free) sulphur in
the vulcanizate would also provide an indication of the degree of vulcanization
of prevulcanized natural rubber latex. This property is certainly of considerable

interest in its own right. It is also broadly true that, for a given type of
sulphur-vulcanization system, the degree of vulcanization tends to increase
as the concentration of unreacted sulphur in the rubber decreases. However,
free sulphur content is not a good index of degree of vulcanization, because
it provides information relating only to the disappearance of sulphur; it is
silent as to the fate of the sulphur. Combined sulphur can in principle be
present in various forms in a vulcanizate. Even that which is combined with
the rubber can be present in forms other than crosslinks between rubber
molecules. Even the proportion which is combined in the form of crosslinks
provides no satisfactory index of the concentration of crosslinks, because the
important matter of the average number of sulphur atoms per crosslink
remains unclear. Of the test methods listed above, only the chloroform-
coagulation test is peculiar to prevulcanized natural rubber latex, in that it
is a test which is applied to the latex itself, and not to films dried down from
the latex. The other three tests are applied to films from the latex, and are
in principle applicable to any elastomeric films, including postvulcanized
films derived from rubber latices.
The chloroform-coagulation test provides a very simple, rapid semi-
quantitative procedure for assessing degree of vulcanization in prevulcanized
natural rubber latex. A small volume (typically lOcm 3 ) of the latex is mixed
with an equal volume of chloroform, and the mixture stirred until coagulation
occurs as a consequence of the latex particles having absorbed much of the
chloroform. The physical character of the coagulum is then examined; it
depends upon the degree of vulcanization of the latex particles. The more
highly crosslinked are the particles, the less readily do they coalesce together
to form a coherent coagulum. As the degree of vulcanization increases, so
the coagulum changes from a soft, plastic well-integrated mass to a crumbly
material. The state of vulcanization can therefore be judged from the
appearance of the coagulum. An arbitrary number, known as the chloroform
number, is assigned to the latex on the basis of the appearance. Four stages
of prevulcanization are usually distinguished by this test, and are assigned
chloroform numbers as follows, the perceived degree of vulcanization
increasing with increasing chloroform number:
1. chloroform number 1: the coagulum is a tacky mass, breaking in a stringy
manner when stretched;
2. chloroform number 2: the coagulum is a weak lump which breaks short
when stretched;
3. chloroform number 3: the coagulum has the form of non-tacky agglomerates;
4. chloroform number 4: the coagulum has the form of small dry crumbs.
Attempts are sometimes made to refine the scale of chloroform numbers by
introducing intermediate gradings designated by numbers such as 2 1/ 2 ,3 1/ 2 ,
etc. The great advantage of the chloroform-coagulation test over other
methods of assessing the degree of vulcanization of natural rubber latex is

its rapidity and consequent suitability for control purposes under factory
conditions. Its disadvantages are that it is a subjective assessment, and that
the chloroform number does not correlate closely with degree of vulcanization
as judged by less subjective methods, in particular, by those methods which
endeavour to determine that concentration of crosslinks in the rubber network.
In the equilibrium-swelling test, a thin film is first cast from the latex. A
test piece is then cut from it, immersed in a rubber solvent, and allowed to
absorb solvent until equilibrium is attained. Toluene is the solvent usually
used for the test. The extent to which the test piece has swollen is then
determined, usually either by weighing the test piece before and after
immersion in the solvent or by measuring one or more dimensions of the
test piece before and after immersion in the solvent. The extent of swelling
can be expressed in various ways, e.g. as the mass or volume ratio of solvent
to rubber in the swollen test piece, as the mass or volume fraction of rubber
in the swollen test piece, as the linear swelling ratio of the test piece, or as
the areal swelling ratio of the test piece. The extent to which the test piece
swells is, of course, an inverse measure of the degree of vulcanization. It is
always possible to make estimates of the concentration of physically-effective
crosslinks in the rubber network from values of equilibrium swelling, if the
appropriate polymer-solvent interaction parameter (XI) is known. The results
obtained in this way for concentrations of physically-effective crosslinks in
films from prevulcanized latices are of doubtful significance, because of
uncertainties concerning the structure of the dried film and the factors which
restrict swelling. Although it provides a quantitative and, if the test is carried
out with care, accurate measure of the degree of vulcanization of a latex, the
method suffers from the disadvantage of being time-consuming. Many hours
can be required for the drying down of the film and the attainment of swelling
equilibrium. The time required for the attainment of equilibrium swelling
depends greatly upon the thickness of the film. Some reduction of the requisite
time is possible if very thin films are cast, but the handling of such films, e.g.
those less than 0.1 mm in thickness, is difficult. In consequence, practical
equilibrium swelling times are usually at least 45 minutes. As a consequence,
this method is not very convenient for factory-control purposes.
A specific procedure for the equilibrium-swelling test has been described
by Deal and Waterman [4]. It was developed primarily for application to
films from prevulcanized styrene-butadiene rubber latices, for which the
characteristics of the elastomer exclude the use of the chloroform-coagulation
test. However, it is in principle applicable to films from any type of
prevulcanized rubber latex, provided that coherent films can be formed from
the latex. The procedure is as follows: A thin film of latex, ca. 0.08-0.25 mm
when dry, is spread upon a glass plate and dried for one hour using a fan.
The surface of the dried film is dusted with talc, the film removed from the
glass plate, and both sides dusted with talc. Removal of the film from the
glass plate can be facilitated by prior immersion of the glass plate in warm
water. After removal from the glass plate, the film is dried for a further period
of 1 hour at 23°C and 50% relative humidity. Circular discs of diameter ca.
38 mm are cut from the film. Excess talc is wiped off, and the discs swollen
by immersion in a suitable liquid contained in a Petri dish. The swelling
fluid for films from prevulcanized styrene-butadiene rubber latex is a
70/30 v/v hexane/10° xylol mixture. The diameter of the disc is measured
after immersion for 25-27 minutes in the swelling fluid. The method of
measurement is to place the Petri dish over graph paper. The diameter is
measured in two directions, and the average value calculated. The diameter
of the swollen disc (or the ratio of the swollen to initial disc diameters) is
taken as an inverse measure of the degree of vulcanization of the film, and
hence of the latex from which it was derived. The composition of the solvent
mixture is said to have been carefully adjusted so as to swell the film sufficiently
for ready discrimination between different states of vulcanization without
undue weakening of the film. A similar method has been described by Panich,
Fodiman and Voyutski [5]. Their procedure uses a square test piece of
thickness 0.4-0.6mm and area 3cm 2 • The test piece is immersed in 25cm 3
of benzene at a temperature in the range 20-22°C. The absorption of benzene
is determined gravimetrically by periodically removing, drying and weighing
the test piece. They found that a constant weight was attained after 6 hours
immersion. A further procedure for assessing degree of latex vulcanization
by swelling has been described by Murray [6].
In the relaxed modulus test, a thin film is cast from the latex, and the
relaxed tensile modulus of a test piece cut from it is determined. The extension
for the modulus determination is usually 100%. The important feature of
this test method is that a considerable time is allowed for the stretched test
piece to relax, so that the elastic modulus which is determined approximates
closely to the equilibrium modulus of the material. As such, it provides an
accurate measure of the concentration of physically-effective crosslinks in
the rubber network, and thus an accurate measure of the degree of
vulcanization. This method is again time-consuming, and therefore unsuitable
if rapid results are required. Further information concerning the use of this
type of test to assess state of vulcanization is available in a paper by Fletcher,
Gee and Morell [7].
The prevulcanizate relaxed modulus (PRM) test, which has been described
in detail by Gorton and Pendle [8], again assesses the degree of latex
prevulcanization by way of measurement of the elastic tensile modulus of a
film dried down from the latex. A thin film is prepared by dipping a glass
tube into the latex, and drying the latex deposit so formed at elevated
temperature (70°C). The dried film is then rolled into a ring, and the modulus
at 100% extension determined after relaxation for 1 minute. It is claimed
that the whole procedure can be completed in less than 10 minutes, because
the layer of wet latex is sufficiently thin to dry completely in less than 4
minutes at 70°C. This period of heating is insufficient to significantly affect

the degree of vulcanization of the rubber.
Notwithstanding the improvements which have been made in recent years
in respect of test methods for routinely assessing the degree of vulcanization
in prevulcanized natural rubber latex, there is still a need for a rapid,
quantitative procedure which is suitable for use as a factory-control test.
Although considerable thought has been given to this problem, it appears
that a satisfactory solution has yet to be found.
13.2.4 ElI'ect of sulpbur-prevulcanization upon properties of natural rubber

latex, and of films derived from natural rubber latex
Sulphur-prevulcanized natural rubber latex is slightly yellow in colour relative
to the initial unvulcanized latex, but is otherwise very similar in appearance.
In particular, the fluidity of the initial latex is retained, and the state of
dispersion of the particles is not significantly altered. Presumably the density
of the latex particles increases somewhat as a consequence of prevulcanization,
because rubbers contract slightly when they are crosslinked. However, it is
doubtful whether this matter has ever been investigated experimentally. It is
therefore to be expected that the rate and tendency to creaming and
centrifugation are considerably reduced as a consequence of sulphur-
prevulcanization (see sections 9.4.3.1 and 9.4.4.2 of Chapter 9). It is difficult
to make quantitative predictions, because of uncertainties concerning appropriate
values to take for the densities of the particles in unvulcanized and
prevulcanized natural rubber latex. However, some indication of the likely
effect can be gained by noting that the densities of dry unvulcanized
polyisoprene and of gum polyisoprene vulcanizates have been given as ca.
0.913 and ca. 0.970 Mg m - 3 respectively.
Typical results, reported by Gorton [9], for the effect of sulphur-
prevulcanization upon the chloroform number of ammonia-preserved natural
rubber latex, and upon the properties of films derived from the latex, are
shown in Table t 3.2. The formulation to which these results refer, given at
the head of the table, is simple. The vulcanizing system comprised sulphur
and zinc diethyldithiocarbamate only; no other zinc compound was present.
The following general features of the sulphur-prevulcanization of natural
rubber latex are illustrated by these results:
1. None of the four methods for assessing the degree of vulcanization, listed
in section 13.2.3 above, provides any evidence of vulcanization reversion,
i.e., that, with increasing time of prevulcanization, the degree of vulcanization
rises to a maximum and then decreases. Taken together, the four methods
provide strong evidence that the degree of vulcanization increases
progressively with increasing time of pre vulcanization.
2. Whilst there is a broad correlation between the indications of state of

vulcanization given by the four methods of assessment, there is no close
correlation. The chloroform number is very sensitive to state of
prevulcanization in the initial stages of the reaction, but does not provide
evidence of continuation of vulcanization after ca. 3 1/ 2 hours of vulcanization
at SO·e. The other three methods indicate clearly a progressive increase
in degree of vulcanization with time of heating. The reduction in the extent
of equilibrium swelling in a good rubber solvent (toluene in this case)
appears to be a more sensitive indicator of degree of vulcanization than
do measurements of modulus.
3. The results for free sulphur indicate that approximately half the sulphur
reacts rapidly, say, during the first hour at SO·C, and thereafter the
concentration of free sulphur decreases only slowly with further time of
prevulcanization.
4. The tensile strength of films derived from the prevulcanized latex decreases
progressively with increasing time of prevulcanization. Unfortunately
Gorton does not give a value for the tensile strength of a film from the
corresponding unvulcanized natural rubber latex. Almost certainly it
would have been considerably lower than any of the values shown in
Table 13.2. Results for the effect of time of prevulcanization at various
temperatures upon the tensile strength of films from natural rubber latex,
reported by Merrill [10] and shown here as Fig. 13.1, indicate that the
tensile strength of films from the unvulcanized latex is ca. 5 MPa. The
values for tensile strength shown in Table 13.2 are typical ofthose for lightly-
crosslinked unfilled sulphur-vulcanizates derived from natural rubber,
regardless of the method of preparation. The decrease of tensile strength
with increasing time of prevulcanization is probably a consequence of
decreasing ability of the latex particles to coalesce and integrate fully as
the concentration of crosslinks in the particle increases. There is no
evidence that it is a consequence of vulcanization reversion.
5. The extension at break decreases somewhat as the time of pre vulcanization
increases. All the values shown in Table 13.2 are typical of those for
lightly-crosslinked unfilled sulphur-vulcanizates derived from natural
rubber, again regardless of the method of preparation. These values are
surprisingly high, bearing in mind that the films were formed by coalescence
and integration of particles of vulcanized rubber.
It is now well-established that the tensile strength of films derived from

prevulcanized natural rubber latex increases considerably if the dried film is
leached in water. The tensile strength tends to increase still further if the
films are leached in acetone, although the statistical significance of the latter
increase is perhaps doubtful. Typical results, obtained by Merrill [10], are
shown in Fig. 13.2. The latex which was used to obtain these results contained
sulphur and zinc diethyldithiocarbamate, but again no other zinc compound
Table 13.2 Effect of time of heating at 50'C upon sulphur-prevulcanization of ammonia-preserved natural rubber latex (Gorton [9])
Natural rubber (as 60% m/m high-ammonia centrifuged latex) 100
Potassium caprylate (n-octanoate) (as 20% m/m aqueous solution) 0.26
Potassium hydroxide (as \0% m/m aqueous solution) 0.25
Sulphur (as 50% m/m aqueous dispersion) 1.5
Zince di-n-butyldithiocarbamate (as 50% m/m aqueous dispersion) 1
Properties of unleached films from latex
Swelling Modulus Free

Time of Chloroform ratio Tensile Extension sulphur
heating number in toluene MRJOO* PRMt M300t M500* strength at break % mlm
I hour of latex I m/m I MPa I MPa I MPa I MPa I MPa 1% on rubber
0.5 1 14.7 0.40 0.31 0.74 1.03 23.7 \030

1.0 2 11.8 0.43 0.33 0.77 1.12 23.4 990 0.68
1.5 2 8.6 0.43 0.34 0.82 1.17 23.4 980
2.0 3 6.9 0.45 0.35 0.85 1.25 22.4 940 0.62
2.5 3 5.7 0.48 0.35 0.86 1.27 21.6 920
3.0 3 5.7 0.48 0.37 0.90 1.37 20.6 890 0.52
3.5 4 5.3 0.50 0.38 0.94 1.47 20.4 880
4.0 4 5.3 0.51 0.40 0.99 1.59 21.4 870 0.55
4.5 4 5.7 0.52 0.41 0.99 1.61 19.8 845
5.0 4 5.0 0.53 0.41 1.02 1.66 20.4 840
5.5 4 4.7 0.54 0.42 1.14 1.84 19.8 835
6.0 4 4.4 0.55 0.43 1.17 1.78 19.8 850 0.49
7.0 4 4.2 0.54 0.42 1.18 1.88 18.8 820
• Relaxed modulus at I(XY'1o extension.

t Prevulcanizate relaxed modulus.
t Conventional modulus at 300"10 extension.
§ Conventional modulus at 500% extension.
40
5 10 15
Time of prevulcanization / hour
Figure 13.1 Effect of time of sulphur-prevulcani711tion at various temperatures upon tensile

strength of films from natural rubber latex (Merrill [10]). Formulation (parts by mass dry):
natural rubber 100, potassium oleate 0.75, sulphur 1, zinc diethyldithiocarbamate 1. Prevulcanization
temperatures: 6Q"C (points e), 70T (points A), 80T (points .).
was present. As well as showing the effect ofleaching upon the tensile strength
of films from prevulcanized natural rubber latex, these results, and those
shown in Fig. 13.1, also confirm that the tensile strength of films from
prevulcanized natural rubber latex increases sharply during the early stages
of the reaction, passes through a shallow maximum, and then decreases
somewhat as the time of prevulcanization further increases. The increase in
tensile strength which follows leaching in water is evidently substantial. The
results of Merrill indicate enhancements of up to ca. 10 MPa. Gorton [9]
has reported even more striking enhancements of tensile strength as a
consequence of leaching in water for 16 hours at 20°C: in one case, the initial
tensile strength was 17.8 MPa, and this increased to 35.0 MPa after leaching.
Regardless of the magnitude ofthe effect of leaching upon the tensile strength
of films from sulphur-prevulcanized natural rubber latex, it is now recognized
that tensile strengths in the range 30-35 MPa are routinely attainable for
films from prevulcanized natural rubber latex, if an appropriate vulcanizing
system has been used, and if the films have been leached in water. The increase
in tensile strength which results from leaching is attributed to extraction of
non-rubber substances from between the latex particles which comprise the
40
Figure 13.2 Effect of leaching in water and then leaching in acetone upon tensile strength of
films from sulphur-prevulcanized natural rubber latex (Merrill [to]). Formulation (parts by
mass dry): natural rubber 100, potassium oleate 0.75, sulphur I, 7jnc diethyldithiocarbamate 1.
Prevulcani7.ation temperature: 70°C. Significance of points: • unleached, '" leached in water,
• leached in water, then acetone.
film, with consequent improvement in the degree of interparticle integration.

Gorton has also shown that the modulus of films from prevulcanized natural
rubber latex increases on leaching; this is attributed to the same cause as the
increase in tensile strength. Leaching has little effect upon extension at break.
Gorton has also investigated the effects of heating (30 minutes at 100°C), and
of leaching in water followed by heating, upon the properties of films from
prevulcanized natural rubber latex. Heating was found to bring about
increases in tensile strength and modulus, but these increases were not as
great as those which followed leaching. The combined effect of leaching and
heating was similar to that of leaching alone. These observations indicate
that the development of maximum tensile strength and modulus in these
films requires extraction of non-rubber substances from the interparticle
boundaries; merely increasing the segmental mobility of the rubber
macromolecules is not sufficient.
Loh [11] has reported some interesting qualitative observations concerning
the clarity of the films obtained from mixtures of natural rubber latex, sulphur
and zinc dialkyldithiocarbamate accelerator after heating for various periods

and then drying at normal ambient temperature. These observations are
relevant to the mechanism of the sulphur-prevulcanization reaction in natural
rubber latex, as well as being of interest in their own right. Loh observed
that, if latex and sulphur alone are heated at 50·C, the films become clear
after only 30 minutes heating, indicating that rapid transfer of the sulphur
to the rubber particles occurs. If the temperature of heating is raised to 70·C,
then the films from latex-sulphur mixtures become clear after only 5 minutes
heating. Films from latex containing only zinc di-n-butyldithiocarbamate
remain opaque even though the latex-accelerator mixture is heated for 10
hours at 70·C. Films from latex containing both sulphur and zinc
di-n-butyldithiocarbamate remain opaque after heating for 4 hours, but
become clear after heating for 5 hours. This indicates that, although in a
typical prevulcanization reaction the transfer of accelerator to the rubber
particles does not occur as rapidly as does the transfer of sulphur, transfer
of accelerator is nevertheless complete long before the prevulcanization has
reached an advanced stage. Furthermore, it appears that the presence of
sulphur is necessary for transfer of accelerator to occur at all.
Another matter of some interest is the average number of sulphur atoms
in the crosslinks introduced into the particles of natural rubber latex by sulphur-
prevulcanization. In principle, the crosslinks could be monosulphidic, disulphidic
or polysulphidic, although it would be expected from the low temperature
at which latex prevulcanization occurs that the crosslinks would be
predominantly polysulphidic. This has been confirmed by Loh [11] using
films, dried from sulphur-prevulcanized latex, which had not been subjected
to any further process likely to change the nature or concentration of the
crosslinks in the rubber. The films were successively treated with two types
of chemical 'probe' reagent, namely, 2-mercaptopropane and piperidine in
heptane, which degrades polysulphidic crosslinks, and then n-hexyl mercaptan
in neat piperidine, which degrades disulphidic crosslinks. The concentrations
of crosslinks in the initial film, and after the successive treatments, were
determined by swelling. The proportion of monosulphidic crosslinks was
found to be so low as to be effectively zero. Typically, the amount of sulphur
combined with the rubber in a well-prevulcanized natural rubber latex was
found to correspond to ca. 10 sulphur atoms for each crosslink present. For
films from sulphur-prevulcanized natural rubber latex which has been
produced using a 'sulphurless' vulcanizing system in which a thiuram
polysulphide is used as a sulphur-donor, it is expected that the crosslinks
will be predominantly disulphidic and monosulphidic. However, as far as is
known, this matter has not so far been investigated experimentally. The
ageing resistance of films from these latices at high temperatures has been
found to be superior to that of films from conventional sulphur-prevulcanized
latices; this observation is consistent with the crosslinks being predominantly
disulphidic and monosulphidic.
13.2.S Effect of some important procedural variables upon sulphur-

prevulcanization of natural rubber latex
13.2.5.1 Temperature
Figure 13.1 shows results obtained by Merrill [10] for the effect of time of
prevulcanization at various temperatures upon the tensile strength of films
cast from sulphur-prevulcanized ammonia-preserved natural rubber latex.
The latex used contained sulphur and zinc diethyldithiocarbamate, but again
no other zinc compound was present. As expected, the initial sharp increase
in tensile strength is accelerated by increasing the prevulcanization temperature.
These results also indicate that the peak tensile strength depends significantly
upon the prevulcanization temperature, although the nature ofthe dependency
is not clear. Further results for the effect of temperature upon the rate of
sulphur-prevulcanization of ammonia-preserved natural rubber latex, reported
by Gorton [9], are shown in Table 13.3. The accelerator was again zinc
diethyldithiocarbamate. A small amount of zinc oxide was also present. Over
the temperature range 40-80·C, the rate of crosslinking increased with
increasing prevulcanization temperature, as is evidenced by the modulus of
the vulcanizate. It also appears that the maximum attainable modulus
increased with increasing prevulcanization temperature, at least up to a
prevulcanization temperature of 70·e. Gorton has also reported results for
the prevulcanization of the same latex compound at 20·C and at 30·C, but
these results are not shown in Table 13.3. Little vulcanizing activity was
apparent at 20·C as judged by modulus increase, notwithstanding that
prevulcanization was allowed to continue for 7 days. The tensile strength of
films from this latex was significantly greater than that expected for an
unvulcanized film from natural rubber latex, but varied little with the time
of prevulcanization. This observation indicates that some prevulcanization
had probably occurred during the early stages of the reaction. Somewhat
increased prevulcanization activity is evident from the results for films from
the latex heated at 30·e.
13.2.5.2 Level of vulcanizing ingredients

Table 13.4 shows results reported by Gorton [9] for the effect of the levels
of sulphur and accelerator upon the sulphur-prevulcanization of ammonia-
preserved natural rubber latex, keeping the accelerator/sulphur ratio constant.
The latex compounds also contained a constant small amount of zinc oxide.
Although the levels of sulphur and accelerator were increased by a factor of
10, the effects upon film properties after any given time of prevulcanization
were found to be relatively slight. As the levels of sulphur and accelerator
were increased, so the modulus after any given time of pre vulcanization tended
to increase, and the extension at break to decrease. Insofar as conclusions
can be drawn concerning the effect upon tensile strength, it appears that
Table 13.3 Effect of heating temperature upon sulphur-prevulcanization of ammonia-preserved natural rubber latex (Gorton [9])
Potassium hydroxide (as 10% m/m aqueous solution) 0.25
Sulphur (as 50% m/m aqueous dispersion) 1
Zinc diethyldithiocarbamate (as 50% m/m dispersion) 0.5
Zinc oxide (as 50% m/m dispersion) 0.2
Properties of unleached films from latex prevulcanized at temperature indicated
40·C 50·C 70·C 80·C

Time of
heating MRlOO· TSt EBt MRlOO· TSt EBt MRJOO· TSt EBt MRlOO· TSt EBt
I hour I MPa I MPa 1% I MPa I MPa 1% I MPa I MPa 1% I MPa I MPa 1%
0.5 0.38 22.2 1060 0.43 23.6 985
1.0 0.31 13.3 1080 0.43 23.1 965 0.50 21.5 870
1.5 0.51 21.2 855
2.0 0.25 9.1 1125 0.34 16.9 1030 0.50 22.4 890 0.50 19.8 850
3.0 0.39 22.4 1010 0.51 21.6 885 0.51 23.9 855
4.0 0.24 9.3 1155 0.44 26.9 980 0.52 17.0 815
5.0 0.47 27.5 965 0.53 17.2 820
6.0 0.26 12.6 1145 0.50 27.1 930
16.0 0.40 23.8 1030 0.48 23.0 915
24.0 0.39 24.3 1050
* Relaxed modulus at 100"1. extension.

t Tensile strength.
t Extension at break.
Table 13.4 Effect of levels of sulphur and accelerator upon sulphur-prevulcanization of ammonia-preserved natural rubber latex at 60°C,
keeping the accelerator/sulphur ratio constant (Gorton [9J)
Sulphur (as 50% m/m aqueous dispersion) X
Zinc diethyldithiocarbamate (as 50% m/m dispersion) X
Properties of unleached films from prevulcanized latex compound having value of X indicated
X =0.2 X= J X=2
Time of heating MRlOO· TSt EBt MRIOO· TSt EDt MRlOO· TSt EDt
/ hour / MPa / MPa /% / MPa / MPa /% / MPa / MPa /%
I 0.34 15.8 1075 0.38 18.2 990 0.39 18.4 995

2 0.36 21.0 1030 0.45 28.6 975 0.46 24.0 965
3 0.40 23.2 1010 0.48 24.7 930 0.49 23.5 910
4 0.42 24.2 1000 0.51 20.7 870 0.50 21.9 865
5 0.44 23.5 960 0.54 23.0 870 0.55 21.7 840
6 0.45 23.6 955 0.57 20.9 830 0.57 21.0 825
• Relaxed modulus at 100'10 extension.

t Tensile strength.
there may be a slight tendency for this property to decrease with increasing
level of sulphur and accelerator at the longer times of prevulcanization.
Notwithstanding these trends, the effect of sulphur and accelerator level upon
the properties of films derived from sulphur-prevulcanized natural rubber
latex is much less than might be supposed. This is presumably a consequence
of the vulcanizing ingredients being initially present as solid phases dispersed
in the aqueous phase of the latex. The amounts which become available for
reaction with the rubber under these conditions are probably determined by
factors other than merely the amounts ofthe ingredients present in the system.
13.2.5.3 Presence or absence of a sparingly-soluble inorganic zinc compound

In designing a compound for the sulphur-prevulcanization of ammonia-
preserved natural rubber latex, one matter which has to be decided is whether
or not to include a sparingly-soluble inorganic zinc compound with the
vulcanizing ingredients. On the one hand, it is desirable to minimize the level
of sparingly-soluble zinc compounds in the compound, in order to minimize
colloidal destabilization arising from the formation of zinc-ammine ions in
the aqueous phase of the latex. Such colloidal destabilization is primarily
manifested as zinc-oxide thickening. The higher is the temperature of
prevulcanization, the more serious is this problem likely to be. Furthermore,
as noted in section 13.2.2 above, it is also desirable to minimize the level of
sparingly-soluble zinc compounds ifit is required to achieve maximum optical
transparency in the eventual rubber film. For both these reasons, it is logical
to exclude sparingly-soluble inorganic zinc compounds from the latex. On
the other hand, certain advantages accrue if such a compound is included.
Thus Gorton [9] has reported results which indicate that, although not
significantly affecting the rate of crosslinking of the rubber, the presence of
a small amount of zinc oxide can increase the final level of vulcanization.
Tensile strength, however, is not significantly affected.
The effect of zinc oxide upon the sulphur-prevulcanization reaction is more
clearly demonstrated by results obtained by Merrill [10], reproduced here
as Figs 13.3(a) and (b). Figure 13.3(a) shows results for the disappearance of
free zinc diethyldithiocarbamate during prevulcanization. In the absence of
zinc oxide, the level of free accelerator decreases sharply as prevulcanization
proceeds, eventually falling to almost zero. The rate of decrease of the level
of free accelerator is greatly reduced if a small amount of zinc oxide is present.
These observations are consistent with modern understanding ofthe chemistry
of the sulphur-vulcanization of natural rubber. The reason the decrease in
the level of free zinc dialkyldithiocarbamate accelerator is retarded by zinc
oxide is that the accelerator is regenerated in the presence of the zinc oxide.
Figure 13.3(b) shows results for the appearance of chemical crosslinks in the
dry films derived from the latices to which the results of Fig. 13.3(a) refer.
Although some retardation in the rate of crosslink insertion as prevulcanization
... 1.0 5
.£:
a.
a.
.......
CI)
~ 0.8
~
~ 0.6
'C
~
'C 0.4
o
c
'N
3l
~ 0.2
'0
~
...J
°0~----~2----~4----~6~--~8 °0~--~2~--~4----~6----~8
(8) Time of prevulcanization I hour (b) Time of prevulcanization Ihour
Figure 13.3 (al Effect of presence of zinc oxide upon disappearance of free zinc diethyldithiocar-
bamate during sulphur-prevulcanization of natural rubber latex at 70·C (Merrill [10]).
Formulation (parts by mass dry): natural rubber 100, sulphur 1, zinc diethyldithiocarbamate 1,
zinc oxide 0.2 (points .) and 0 (points A). (b) Appearance of chemical crosslinks in dry films
derived from the latices to which the results of (a) refer (Merrill [10]).
proceeds is evident for the reaction system which contained zinc oxide, there
was no cessation of vulcanization. It was otherwise for the reaction system
which did not contain zinc oxide: insertion of chemical crosslinks ceased at
about the time the level of free accelerator had fallen to almost zero. From
the results for the zinc-oxide-free compound shown in Fig. 13.3(b), it can be
concluded that the prevulcanization reaction ceases when the free accelerator
level has fallen to zero. This conclusion is supported by the further observation
by Merrill that the level of free sulphur in the latex ceases to decrease when
the level of free accelerator has fallen to zero, and that the disappearance of
free sulphur resumes if further accelerator is added.
13.2.5.4 Nature of dithiocarbamate accelerator

Gorton [9] has reported results for the effect of the nature of the
dialkyldithiocarbamate accelerator upon the sulphur-prevulcanization of
ammonia-preserved natural rubber latex. All the latex compounds investigated
contained zinc oxide. The various dialkyldithiocarbamates were compared
on a basis of equal parts by mass. Thus, although these results are of immediate
interest to the latex technologist, their interpretation by the polymer chemist
is somewhat obscured by there being different numbers of molecules present
in the various latices. Table 13.5 shows the results of Gorton for latices
containing, respectively, zinc diethyldithiocarbamate, zinc di-n-butyldithio-
carbamate, and the corresponding sodium salts. Zinc di-n-butyldithiocarbamate

is clearly rather more active in promoting crosslinking than is zinc
diethyldithiocarbamate. This is confirmed by other results reported by
Gorton, which show that zinc di-n-butyldithiocarbamate is active at lower
prevulcanization temperatures than is zinc diethyldithiocarbamate. Zinc
di-n-butyldithiocarbamate also appears to give significantly higher tensile
strength than does zinc diethyldithiocarbamate. The corresponding sodium
salts were found to behave similarly to the zinc derivatives. However, in
comparing the sodium and zinc compounds, it should be recalled that zinc
oxide was present in the latices, and this would be expected to cause the
sodium compounds to behave similarly to the zinc compounds.
Gorton [9] has also reported results for latices which were sulphur-
prevulcanized using other dithiocarbamate accelerators, namely, zinc
pentamethylene-dithiocarbamate, zinc ethylphenyldithiocarbamate, a zinc
di-n-butyldithiocarbamate-dibutylamine complex, and a sodium dithiocar-
bamate-thiazole blend. The conclusion reached was that there is little
advantage in using these accelerators in preference to those for which results
are given in Table 13.5.
13.2.5.5 Degree of dispersion of insoluble vulcanizing ingredients

Gorton and Pendle [12] have reported results for the effect of the particle
size ofthe sulphur and accelerator dispersion upon the sulphur-prevulcanization
of ammonia-preserved natural rubber latex. Some of their results for the
effect of the particle size of sulphur dispersion are shown in Table 13.6.
Further results are given in Table 15.1 of Chapter 15 (Volume 3). Perhaps
rather surprisingly, the overall conclusion from this work is that, over the
range of particle size investigated, the particle size of the sulphur and
accelerator appeared to have no significant effect upon the tensile properties
of films obtained from the latex. This observation is again presumably a
consequence of the vulcanizing ingredients being initially present as solid
phases dispersed in the aqueous phase of the latex. It indicates that the
amounts of vulcanizing ingredients which become available for reaction with
the rubber under these conditions is not significantly affected by the area of
interface between the sulphur or accelerator particles and the aqueous phase.
13.2.6 Comparison between sulphur-prevulcanization of natural rubber latex

and sulphur-vulcanization of dry natural rubber
It has already been noted above (section 13.1) that the sulphur-
prevulcanization reaction in natural rubber latex occurs at much lower
temperatures, and is generally more facile, than would be expected from
experience with the vulcanization of dry natural rubber with sulphur and
organic accelerators. Figures 13.4(a) and (b), which are based upon results
obtained by Loh [11], illustrate clearly that sulphur-prevulcanization occurs
Table 13.5 Effect of nature of dialkyldithiocarbamate vulcanization accelerator upon sulphur-prevulcanization of ammonia-preserved
0
natural rubber latex at 60 e (Gorton [9])
Natural rubber (as 60"10 m/m high-ammonia centrifuged latex) 100
Sulphur (as 50% m/m aqueous dispersion) 1
Metal dialkyldithiocarbamate (as 50% m/m dispersion or as a solution) 0.5
Properties of unleached films from prevulcanized latex compound containing metal dialkyldithiocarbamate accelerator indicated
Zinc diethyldithio- Zinc di-n-butyldithio- Sodium diethyldithio- Sodium di-n-butyldithio-

carbamate carbamate carbamate carbamate
Time of
heating MR1OO· TSt EBt MRlOO· TSt EBt MRlOO· TSt EBt MRlOO· TSt EDt
1 hour 1 MPa 1 MPa 1% 1 MPa 1 MPa 1% 1 MPa 1 MPa 1% IMPa IMPa 1%
1 0.38 20.1 1015 0.40 29.4 1030 0.32 18.8 1015 0.49 29.3 965
2 0.42 21.8 960 0.45 30.3 1000 0.36 24.1 995 0.51 24.5 910
3 0.46 21.5 905 0.48 28.4 965 0.40 24.6 980 0.54 22.6 840
4 0.49 21.5 880 0.49 26.4 925 0.45 24.4 925 0.56 20.8 815
5 0.51 21.0 845 0.50 25.0 900 0.50 23.2 890 0.57 21.6 800
6 0.54 19.8 820 0.54 26.1 895 0.53 23.3 840 0.59 20.9 790
* Relaxed modulus at 100"1. extension.

t Tensile strength.
Table 13.6 Effect of particle size of sulphur dispersion upon sulphur-prevulcanization of ammonia-preserved natural rubber latex at 60°C
(Gorton and Pendle [12])
Natural rubber (as 60% mlm high-ammonia centrifuged latex) 100
Potassium caprylate (n-octanoate) or laurate (n-dodecanoate) (as 20% mlm aqueous solution) 0.26
Potassium hydroxide (as 10% mlm aqueous solution) 0.25
Sulphur (as 50% mlm aqueous dispersion) 1.5
Zinc diethyldithiocarbamate (as 50% mlm dispersion) 1
Zinc oxide (as 50% mlm dispersion) 0.2
Properties of unleached films from latex pre vulcanized using sulphur dispersion of number-average particle size (in) indicated
Xn = 13.0 JI111 Xn = 1.8 JI111 lt n = 1.7 JI111 lt n = 1.5 JI111

Time of
heating MRlOO· TSt MRlOO· TSt MRlOO· TSt MRIOO· TSt
I hour / MPa I MPa / MPa / MPa I MPa IMPa I MPa I MPa
1 0.37 17.8 0.40 24.0 0.35 24.7 0.34 19.2
2 0.44 23.3 0.44 26.4 0.39 25.6 0.40 24.8
3 0.49 24.1 0.47 27.2 0.42 25.8 0.41 26.2
4 0.53 24.0 0.50 25.5 0.44 26.1 0.44 26.1
5 0.52 22.8 0.52 26.2 0.46 25.3 0.45 25.5
6 0.55 22.1 0.58 24.9 0.49 25.8 0.48 24.3
• Relaxed modulus at 100"1. extension.

t Tensile strength.
1.0 I" 4
i"
..ca. :8
a. 2
~ 0.8 S
:::l
.c "0 3
a. E
"S
rJ)
.......
oil
0
~ 0.6
c: )(
:0 rJ)
E ""u; 2
.5:
~ 0.4 e
!
()
Q)
'0
c: c:
0
""~
cQ)
()
c:
<3
2 4 6 8 00 2 4 6 8
(8) Time of prevulcanization / hour (b) Time of prevulcanization / hour
Figure 13.4 Vulcanization of natural rubber latex and dry natural rubber at sooC (Loh [II]):
(a) variation of network-combined sulphur with time of prevulcanization or vulcani7.ation; (b)
variation of concentration of chemical crosslinks in vulcani7.ates with time of prevulcani7.ation
or vulcani7.ation. Formulation (parts by mass dry): natural rubber 100, sulphur 1, zinc di-
n-butyldithiocarbamate 1.3. Significance of points: • rubber as latex, A dry rubber.
much more rapidly than does the vulcanization of solid natural rubber at
the same temperature using the same vulcanizing ingredients. Figure 13.4(a)
shows the variation with time of prevulcanization or vulcanization at 50·C
of network-combined sulphur for the rubber in a prevulcanizing natural
rubber latex and the rubber in a vulcanizing solid natural rubber of the same
composition. The latter was obtained by evaporating to dryness at low
temperature the compounded latex which was used for the prevulcanization
experiment. Figure 13.4(b) gives results for the concentration of chemical
crosslinks in these vulcanizates. These results confirm the indication from
experience of the sulphur-vulcanization of solid natural rubber that,
notwithstanding the use of extremely fast vulcanization accelerators for the
sulphur-prevulcanization reaction, prevulcanization readily proceeds at far
lower temperatures than would be expected. Indeed, as has already been
noted (see sections 13.2.2 and 13.2.5.1 above), the reaction can proceed at a
measurable rate even at warm ambient temperature.
The unexpected facility of the sulphur-prevulcanization reaction, relative
to the equivalent dry-rubber sulphur-vulcanization reaction, is not peculiar
to natural rubber latex. Loh [II] has obtained analogous results to those
shown in Figs 13.4(a) and (b) for synthetic cis-I,4-polyisoprene rubber in the
forms of, on the one hand, an artificial latex prepared by dispersing the solid
polymer in an aqueous medium, and, on the other, the dry solid polymer.
Again, the sulphur-prevulcanization reaction readily proceeds at far lower
temperatures than would be expected on the basis of the rate of vulcanization
of the dry rubber using an identical vulcanizing system. Furthermore, it is
evident from other results given in section 13.2.7.6 below that it is also possible
to sulphur-prevulcanize at similar temperatures synthetic polyisoprene latices
produced by the emulsion polymerization of isoprene. Thus there is strong
evidence for supposing that the unexpected facility of the sulphur-
pre vulcanization reaction is a consequence of the polyisoprene being in the
form of a stable colloidal dispersion in an aqueous medium, rather than to
the presence of, say, some ofthe non-rubber substances in ammonia-preserved
13.2.7 Kinetics and mechanism of sulphur-prevulcanization of natural rubber

latex
13.2.7.1 Introductory comments

We consider now the matter of the mechanism by which the sulphur-
prevulcanization of natural rubber occurs, together with the related matter
ofthe kinetics of certain processes which occur during sulphur-prevulcanization.
The structure of films derived from sulphur-prevulcanized natural rubber
latex is considered in a subsequent section (section 13.2.9 below). The kinetics
of the various processes which occur during sulphur-prevulcanization are of
interest not only in their own right but also because any satisfactory theory
of the reaction mechanism must be consistent with such facts as have been
established in respect of kinetics.
Before proceeding to discuss the kinetics and mechanism of the sulphur-
prevulcanization reaction, it is desirable to form a clear picture of the reaction
system. The initial state of the reaction system is illustrated schematically in
Fig. 13.5. The vulcanizing ingredients are dispersed as sparingly-soluble
particles in the aqueous phase of the latex, together with the rubber particles.
It seems likely that the vulcanizing ingredients become absorbed into the
rubber particles via the aqueous phase, and not by direct contact between
the rubber particles and those of the vulcanizing ingredients. As the
prevulcanization reaction proceeds, it appears that the sulphur and organic
accelerator become converted to forms which do not cause opacity in films
deposited from the latex. The obvious interpretation of the observations in
this respect made by Loh [11], noted in section 13.2.4 above, is that these
ingredients have become almost entirely absorbed within the rubber particles
in forms which are rubber-soluble.
In addition to consistency with the observed kinetics of the reaction, the
more obvious requirements which any satisfactory theory of the mechanism
Figure 13.5 Schematic illustration of sulphur-prevulcanization reaction system.
of the sulphur-prevulcanization reaction must fulfil are as follows:

1. It must be consistent with the reaction mechanisms which are generally
accepted for the vulcanization of solid natural rubber using sulphur and
organic vulcanization accelerators. This requirement implies that the
sequence of reactions which is postulated to occur within the rubber
particles must be of a similar nature to that which is believed to bring
about sulphur-vulcanization in dry natural rubber.
2. It must explain the processes by which vulcanizing reagents which are
initially confined entirely to the aqueous phase of the latex rapidly become
present in the rubber particles in a form in which they are able to effect
crosslinking of the rubber macromolecules by sulphur bridges.
3. It must explain the facility with which the reaction can occur at relatively
low temperatures.
It has to be said at the outset of this discussion that, although considerable
progress has been made in recent years in elucidating and understanding the
principal features of the sulphur-prevulcanization reaction system, understanding
of the distinctive chemistry of the reaction is still almost non-existent. The
relative facility of the reaction is still essentially a mystery.
13.2.7.2 Kinetics of disappearance of free vulcanization accelerator

Figure 13.6(a) shows the results of Loh [11] for the disappearance of zinc
di-n-butyldithiocarbamate in a sulphur-prevulcanization system which contained
., 0.20
t2.0
Q.
....... '0
~ '5.
-=
*a *
C I Q.
f!! c;: ....... 0.16
~ 1.5
m.l a
8: ~
g m·5€
.-N«I012
€
:2
~ $ g .
0 ..... _
~
Q)-€
1;j 'i5
.5I 1.0
II: ~ 0.08
c::: .c
.l
'C
o
c
'N 0.5 0.04
~
CP
0.8 1.6 2.4
3 00 2 4 6 8 Initial level of
zinc di-n-butyl dithiocarbanate / pphr
(a) (b)
Figure 13.6 (a) Disappearance of zinc di-n-butyldithiocarbamate in a sulphur-prevulcanization

system at SO'C (Loh [11]). Formulations (parts by mass dry): natural rubber 100, sulphur 1,
zinc di-n-butyldithiocarbamate 0.3 (points e), I (points ... ) and 2 (points .). (b) Variation of
zero-order rate of disappearance of zinc di-n-butyldithiocarbamate with initial level of accelerator,
derived from results shown in (a) (Loh [11]).
only sulphur and vulcanization accelerator. It is evident that the concentration

offree dithiocarbamate accelerator in any given reaction system falls linearly
with time of prevulcanization (see also Fig. 13.3(a)). This implies that the
order of reaction with respect to free accelerator is zero. An immediate and
obvious explanation for this observation is that, whatever may be the true
order of reaction with respect to accelerator, the order appears to be zero
because the accelerator is present in the reaction system as a separate
sparingly-soluble phase which maintains the concentration of accelerator in
the aqueous phase, and hence in the rubber particles, constant for most of
the reaction. However, this explanation is difficult to reconcile with the
observation that, in the typicallatex-prevulcanization reaction, the accelerator
seems to have transferred to the rubber particles long before the prevulcanization
has reached an advanced stage. This explanation is also inconsistent with
the observation, evident from Fig. 13.6(a), that the constant rate of disappearance
of accelerator in any given reaction system increases as the initial level of
accelerator increases. The indications are that, in the absence of a zinc
compound such as zinc oxide, the level of free zinc dialkyldithiocarbamate
accelerator always eventually falls to almost zero, regardless of the initial
level. These results show that, although the rate of disappearance of
accelerator is apparently of zero order with respect to accelerator concentration

in anyone reaction system, the zero-order rate is itself dependent upon the
initial level of accelerator. Furthermore, when the zero-order rate of
disappearance of accelerator in anyone reaction system is plotted against
the initial level of accelerator, as in Fig. 13.6(b), a relationship of direct
proportionality is revealed. This implies that the zero-order rate of disappearance
of accelerator in anyone reaction system is itself first order with respect to
the initial level of accelerator. Possible hypotheses to account for this unusual
and apparently contradictory behaviour include the following:
1. that the rate of disappearance is first order in the number of accelerator

particles initially present in the compounded latex;
2. that the reaction is subject to first-order catalysis by the dispersing agents
present in the dispersions of vulcanizing ingredients added to the latex;
3. that the reaction is subject to first-order catalysis by impurities present
in the industrial grades of accelerators used for this investigation.
None of these hypotheses has been confirmed by appropriate experimental

tests carried out by Loh.
The variation of the zero-order rate of disappearance of zinc diethyl-
dithiocarbamates with temperature of prevulcanization follows the Arrhenius
law approximately. The relationship between the logarithm of the zero-order
rate and reciprocal absolute temperature indicates a value of ca. 120 kJ mol- t
for the energy of activation for the disappearance of accelerator.
13.2.7.3 Effect of varying alkyl chain length of zinc dialkyldithiocarhamate

accelerator
Figure 13.7 summarizes results obtained by Loh [11] for the effect of the
alkyl chain length of a zinc di-n-alkyldithiocarbamate accelerator upon the
rate at which free sulphur disappears in a latex sulphur-prevulcanization
reaction. The various reaction systems used to obtain these results contained
the same level of sulphur and equimolar amounts of the various accelerators.
No other zinc compound was present. It is apparent that, as the alkyl chain
length of the zinc di-n-alkyldithiocarbamate is increased, so its ability to
accelerate the rate of combination of sulphur with the rubber passes through
a maximum. On the basis of these results, it is not possible to say exactly
which zinc di-n-alkyldithiocarbamate shows the optimum ability to accelerate
the combination of sulphur with rubber in a latex sulphur-prevulcanization
reaction. Of the various zinc di-n-alkyldithiocarbamates used to obtain the
results shown in Fig. 13.7, namely, the C t , C 2 , C 4 , C 6 , C s and C t2 compounds,
it is the di-n-butyl compound which showed the greatest ability to accelerate
the reaction.
The most obvious explanation for the existence of an optimum alkyl chain
length as regards ability to accelerate sulphur-prevulcanization is that the
1.2
1.0
.§. 0.8
.....Q.
~
~ 0.4
0.2
00 2
Time of prevulcanization I hour
Figure 13.7 Effect of alkyl chain length of a zinc di-n-alkyldithiocarbamate upon rate of
disappearance of free sulphur during sulphur-prevulcani7.ation of natural rubber latex at 70°C
(Loh [II]). Formulations (parts by mass dry): natural rubber 100, sulphur I, 7jnc di-
n-alkyldithiocarbamate equimolar to zinc dicthyldithiocarbamatc 1. Significance of points:
• zinc dicthyldithiocarbamatc; • zinc di-n-butyldithiocarbamatc; ... zinc di-n-hcxyldithiocarbamate;
• zinc di-n-octyldithiocarbamate; 0 zinc di-n-dodecyldithiocarbamate.
accelerative ability depends upon the balance of two factors which change
in opposite directions as the alkyl chain length of the accelerator varies.
These two factors are the solubility of the accelerator in the aqueous phase
of the latex, and the ability of the accelerator, or some derivative of it, to
adsorb at the surface of the rubber particle. Separate experiments carried
out by Loh [11] have shown that, as expected, the solubilty of the zinc
di-n-alkyldithiocarbamates in water at pH ca. to decreases as the alkyl chain
length of the compound increases, although the differences in solubility
between the various homologues are perhaps not as great as might be
expected. Thus the first of these two factors decreases as the alkyl chain
length of the zinc di-n-alkyldithiocarbamate increases. As regards ability to
adsorb at the particle surface, this is expected to increase progressively as
the alkyl chain length increases.
13.2.7.4 Effect of varying counterion of dialkyldithiocarbamate accelerator

Loh [11] has carried out sulphur-prevulcanization experiments using a fixed
level of sulphur and equimolar levels of the diethyldithiocarbamates of zinc,
sodium, lead(II), selenium and tellurium. No other zinc compound was present
in the reaction system. The lead derivative showed little ability to accelerate
the reaction. The other four derivatives were found to be effective as
accelerators, the most effective being the zinc compound.
13.2.7.5 Kinetics of disappearance of free sulphur

Taken as a whole, results obtained by Merrill [10] and by Loh [11] for the
disappearance offree sulphur in a sulphur-prevulcanization reaction accelerated
by a zinc dialkyldithiocarbamate, in the absence of any other zinc compound,
seem to be consistent with the disappearance offree sulphur being approximately
first order in the amount of free sulphur remaining in the reaction system.
However, as is evident from Fig. 13.8(a), the first-order rate coefficient for
the disappearance of free sulphur depends upon the initial level of sulphur
which was present in the system if the initial level is low, but not if it is high.
The reaction systems to which Fig. 13.8(a) refers contained a fixed initial
level of zinc di-n-butyldithiocarbamate accelerator. This suggests that the
first-order rate coefficient for the disappearance of free sulphur may increase
with the initial ratio of accelerator to sulphur, at least over a certain range
of initial levels. The complexity of the kinetics of sulphur-prevulcanization is
further illustrated by the observation that, as shown in Fig. 13.8(b), the
zero-order rate of disappearance of free zinc di-n-butyldithiocarbamate
increases as the initial level of sulphur is increased, even though the initial
level of zinc di-n-butyldithiocarbamate remains fixed.
The variation with prevulcanization temperature of the first-order rate
coefficient for the disappearance of sulphur follows the Arrhenius law
approximately. The relationship between the logarithm of the first-order rate
coefficient and reciprocal absolute temperature indicates ca. 100 kJ mol- 1 as
the energy of activation for the disappearance of sulphur.
13.2.7.6 Effect of varying particle size of latex

A further aspect of the sulphur-prevulcanization of natural rubber latex which
has been discussed in the past is whether or not the rate at which
prevulcanization (however defined) occurs depends upon the size of the
particle in which the rubber macromolecules are present in the latex. This
matter is of some practical interest for at least two reasons:
1. Ammonia-preserved natural rubber latex concentrate contains rubber

particles which differ widely in size (see section 9.5.2.1 of Chapter 9). If
the rate of crosslink insertion under given reaction conditions depends
1.2
.2
Co
10° ~1.0
-
.2
Co
~
Co ~
-e
5
-a 80.8
"3
Ul
;§
-0
~ ~
j 0.6
*
'0
110-' 'is
u
...J c::
·N 0.4
CD
~
'0
§l 0.2
CD
...J
10~~----~----L-----~--~ O~--~~--~~--~~--~
0246802468
(8) Time of prevulcanization / hour (b) Time of prevulcanization / hour
Figure 13.8 (a) Kinetics of diappearance of free sulphur during sulphur-prevulcanization of

natural rubber latex at SO"C accelerated by zinc di-n-butyldithiocarbamate in the absence of
any other zinc compound (Loh [II]). Formulations (parts by mass): natural rubber 100, zinc
di-n-butyldithiocarbamate I, sulphur 0.5 (points e), 1 (points .) and 1.5 (points .). (b) Effect of
level of sulphur upon disappearance of free zinc di-n-butyldithiocarbamate during sulphur-
prevulcanization of natural rubber latex at SO"C (Loh [II]). Formulations and significance of
points as for (a).
significantly upon latex particle size, then the product of the reaction may
be inhomogeneous in that it contains particles of significantly different
degrees of vulcanization. It may be that such inhomogeneity is advantageous,
disadvantageous or of no significance as regards the properties of the final
product. However, regardless of significance for the final product, it is an
aspect of the reaction about which it is desirable to have information.
2. If the rate of prevulcanization does depend upon latex particle size, then
variation of particle-size distribution provides a possible reason for
variation in the rate of prevulcanization of natural rubber latices from
different sources and different batches. Whether or not this variation is
significant in practice would depend upon the magnitude of the effect
relative to variations from other sources, and also upon the conditions
under which the prevulcanization reaction is carried out.
In principle, it would be possible to investigate the effect of latex particle

size upon rate of pre vulcanization by fractionating natural rubber latex with
respect to particle size, and then prevulcanizing the range of latices thus
obtained under the same conditions. However, such experiments would be
tedious. By using synthetic polyisoprene latices obtained by the emulsion
polymerization of isoprene, Siswantoro [13] has investigated the effect upon
rate of prevulcanization of variation of latex particle diameter. The results
shown in Fig. 13.9(a) demonstrate that the variation of concentration of
inserted crosslinks with time of reaction is qualitatively similar to that which
is observed with natural rubber latex, indicating that the reaction in synthetic
polyisoprene latex is essentially similar to that in natural rubber latex. Figure
13.9(a) also shows that the rate of sulphur-prevulcanization increases as the
latex particle size decreases. It might be thought that the initial rate of
insertion of crosslinks would be directly proportional to the aggregate surface
area of the latex particles in the reaction system. However, the limited results
so far available indicate that this is not so (see Fig. 13.9(b». On the other
hand, rather surprisingly, when these limited results are plotted as variation
of initial rate of crosslinks insertion with concentration of latex particles in
the reaction system, a relationship of direct proportionality is observed (see
Fig. 13.9(c». The mechanistic implications of this observation are considered
in section 13.2.7.8 below.
/3.2.7.7 Role of the non-rubber constituents in natural rubber latex

It has sometimes been supposed in the past that the presence of various
non-rubber constituents in natural rubber latex is essential for the occurrence
of the sulphur-prevulcanization reaction. In particular, the presence of various
proteinaceous substances has been regarded as essential. That this is not so
is demonstrated by the following evidence:
1. As has been noted above (see section 13.2.1), it is possible to sulphur-

prevulcanize artificial polyisoprene rubber latices made by dispersing
synthetic cis-l,4-polyisoprene in an aqueous medium. Similar differences
are observed between latex prevulcanization and an equivalent dry-rubber
vulcanization as are observed between natural rubber latex prevulcanization
and an equivalent vulcanization of dry natural rubber.
2. As has also been noted above (see sections 13.2.1 and 13.2.7.6), it is also
possible to sulphur-prevulcanize the latices of various diene rubbers,
including synthetic polyisoprene latex obtained by the emulsion polymer-
ization of isoprene.
3. It has been shown by Ghazaly [14] that it is possible to sulphur-
prevulcanize natural rubber latex from which much of the non-rubber
°0~--~2~--~4----~6~--~
...'"
00 -- 1 2
(a) Time of prevulcanization / hour Aggregate surface area of polymer
particles x 10-19 I nm2 (g latex compoundr 1
(b)
2 4 6
Particle concentration x 10-14 /
(e) number (g latex compoundr 1
Figure 13.9 (a) Kinetics ofinsertion of crosslinks into poIyisoprene during sulphur-prevulcanization
of synthetic polyisoprene latices at SO'C obtained by the emulsion polymerization of isoprene,
demonstrating effect of latex particle size (Siswantoro [13]). Formulation (parts by mass)
polyisoprene 100, sulphur I, zinc diethyldithiocarbarnate 1. Latices stabili7.ed with sufficient
potassium laurate to cover ca. 50"10 of particle surface. Significance of points: • ltO-nm latex;
• 183-om latex; _ 297-nm latex. (b) Variation of initial rate of crosslink insertion with aggregate
surface area of latex particles in reaction system (Siswantoro [13]). Formulation and reaction
conditions as for (a). (c) Variation of initial rate of crosslink insertion with number of latex
particles in reaction system (Siswantoro [13]). Formulation and reaction conditions as for (a).
material has been removed, although the overall rate of insertion of

crosslinks during the prevulcanization reaction is somewhat reduced. This
reduction in overall rate of crosslink insertion is partly a consequence of
the presence of an extended induction period for the commencement of
crosslink insertion, and partly a consequence of a reduced rate of crosslink
insertion once the induction period is over.
Ghazaly [14] has also investigated the effect upon the rate of sulphur-
prevulcanization of adding to highly-purified natural rubber latex substances
which are representative of the non-rubber substances present in ammonia-
preserved natural rubber latex concentrate. In this way, it was hoped to
discover which of the various non-rubber substances are responsible for the
rate of prevulcanization being higher in the normal concentrate than in
natural rubber latex from which much of the non-rubber material has been
removed. Rather surprisingly, the addition of ammonia was found to be
without effect. Potassium hydroxide and sodium hydroxide were found to
accelerate the reaction somewhat, sodium hydroxide having the greater effect
of the two. Ammonium linoleate, ammonium carbonate, diammonium
hydrogen phosphate, fJ-alanine and the serum obtained from high-ammonia
latex concentrate were all found to accelerate sulphur-prevulcanization to a
lesser or greater extent. However, in none of the cases investigated did the
reaction proceed as rapidly as it did with the normal ammonia-preserved
latex concentrate. Linoleic acid was found to retard prevulcanization
somewhat. Thus it is still not possible to say with certainty which of the
non-rubber substances present in normal ammonia-preserved natural rubber
latex concentrate are important in determining the rate at which the reaction
proceeds. It may be that the observed rate of prevulcanization is a
consequence of synergism between one or more of the types of substance
which have been investigated so far. Alternatively, it may be primarily a
consequence of the presence of one or more other types of substance. This
is a matter which merits further detailed investigation, not only for its own
intrinsic interest but also because it may be possible to discover substances
which markedly accelerate the reaction, and which in consequence may be
used to eliminate the variations in rate of prevulcanization which are observed
between different batches of latex concentrate.
Although the presence of the non-rubber constituents of natural rubber
latex are not essential for the occurrence of the sulphur-prevulcanization
reaction, the aqueous phase of natural rubber latex is able to facilitate the
reaction. Thus van Gils [15] demonstrated that sulphur is appreciably soluble
in the aqueous phase of ammonia-preserved natural rubber, whereas it is
believed not to be significantly soluble in dilute aqueous ammonia solution.
Presumably enhanced solubility of sulphur in the latex aqueous phase
facilitates sulphur-prevulcanization. van Gils also showed that sulphur is
appreciably soluble in aqueous solutions of cysteine hydrochloride. Similarly,
Loh [11] found that zinc di-n-butyldithiocarbamate is appreciably more

soluble in the diluted aqueous phase of ammonia-preserved natural rubber
latex than in water or dilute aqueous ammonia solution of pH 10.2. Again,
enhanced solubility of accelerator in the latex aqueous phase presumably
facilitates sulphur-prevulcanization. Porter, Rawi and Rahim [16] have
recently demonstrated that sulphur and zinc dialkyldithiocarbamate accelerators
are able independently to dissolve in ammonia-preserved natural rubber latex
in forms which are able subsequently to effect prevulcanization of the latex
when the other ingredient is added as a solid phase. Thus the simultaneous
presence of both solid sulphur and zinc dialkyldithiocarbamate accelerator
in the latex is not necessary for prevulcanization to occur. These workers
heated uncompounded ammonia-preserved natural rubber latex at 70·C in
a sulphur-coated vessel, removed the latex, compounded the latter with
accelerator (zinc diethyldithiocarbamate), and then heated it in a separate
vessel, again at 70·C. The final latex was found to be prevulcanized, but the
degree of vulcanization was limited by the amount of sulphur which had
dissolved in the latex initially. Similarly, they heated uncompounded
ammonia-preserved natural rubber latex in a vessel coated with solid zinc
diethyldithiocarbamate, removed the latex, compounded the latter with
sulphur, and then heated it in a separate vessel. Again, the final latex was
found to be prevulcanized, but the degree of vulcanization was not limited
by the amount of accelerator which had dissolved in the latex initially,
presumably because the accelerator could be continuously regenerated. In
both types of experiment, prevulcanization was found to be facilitated by
the presence of an aliphatic amine (diethylamine) in the latex when the latter
was heated in the sulphur- or accelerator-coated vessel. It is reasonable to
suppose that such amines facilitate dissolution of sulphur in the latex by
catalysing the ring-opening of cydo-octasulphur, and also the dissolution of
zinc dialkyldithiocarbamate by complexation with the zinc. A related
experiment, which appears so far not to have been performed, is one in which
separate portions of ammonia-preserved natural rubber latex are heated in
separate sulphur-coated and accelerator-coated vessels, and the latices
subsequently blended and heated. There seems little doubt that some
prevulcanization will occur, the extent of which will depend upon the amount
of sulphur which had dissolved in the portion of the latex heated in the
sulphur-coated vessel. A further interesting observation reported by Porter,
Rawi and Rahim concerns the tensile stress-strain properties, notably tensile
strength and modulus, of films cast from latices which had been prevulcanized
in either of the ways described above. Whereas, as has been noted above
(section 13.2.4), leaching in water usually increases the tensile strength and
modulus of such films, this was not the case for the films obtained in this
investigation. The effect of leaching was found to be variable: in some cases,
tensile strength and modulus were decreased. The significance of this
observation is obscure.
13.2.7.8 Conclusions concerning mechanism of sulphur-prevulcanization of

The overall conclusion to be drawn from the observations summarized above
is that it is the presence of the water in the latex which is essential for the
occurrence of the sulphur-prevulcanization of natural rubber latex at
unexpectedly low temperatures. The primary requirement for sulphur-
prevulcanization to occur at these low temperatures seems to be no more
than that the reaction should be carried out on polyisoprene in the form of
a stable aqueous colloidal dispersion. The reaction is capable of being
accelerated by small amounts of various additional substance, in particular,
by some of the non-rubber substances which are present in normal
ammonia-preserved natural rubber latex concentrate. Whether the presence
of very small amounts of such substances is an essential prerequisite for the
occurrence of the reaction is not known at present; indeed, it appears that
it would be very difficult conclusively to establish this matter experimentally,
because of the inevitable presence of colloid stabilizer in any stable aqueous
polyisoprene colloidal dispersion.
It is commonly believed that the accelerated sulphur-vulcanization of dry
natural rubber proceeds by a mechanism in which the sulphur first reacts
with the vulcanization accelerator to form an active sulphurating agent. This
agent then reacts with the rubber molecules to form a sulphurated species
which subsequently rearranges to form a network crosslinked with sulphur.
Other substances present in the rubber, such as fatty acids and amines,
facilitate the reaction, probably by complexing with the metal atom of the
accelerator. The average number of sulphur atoms per crosslink in the final
vulcanizate depends upon factors such as the ratio of accelerator to sulphur,
the temperature at which the vulcanization reaction was carried out, and the
length of time for which the vulcanizate was heated.
In the following discussion of a possible outline mechanism for the
sulphur-prevulcanization of natural rubber latex, two assumptions are made:
1. that the reaction proceeds by a mechanism which, in broad outline at

least, is qualitatively similar to that by which dry natural rubber is
vulcanized by sulphur and accelerator;
2. that prevulcanization can proceed only if the active sulphurating species
is present in the rubber particles.
Qualitative observations (see section 13.2.4 above) indicate that the sulphur
can transfer from the original particles to the rubber particles remarkably
rapidly, that the zinc dialkyldithiocarbamate accelerator by itself does not
transfer to any significant extent, and that, in the presence of sulphur, the
accelerator can transfer, although more slowly than does the sulphur on its
own. It therefore appears that the first important step of the reaction is the
formation of a sulphur-accelerator species in the aqueous phase of the latex.
This species, or some derivative of it, then transfers to the rubber phase. If,
as seems likely, this species is surface-active, then the most obvious mode of
transfer would be adsorption from the aqueous phase on to the surface of
the rubber particle. Such a mechanism is consistent with the observations
for the effect upon rate of combination of sulphur of varying the alkyl chain
length of zinc di-n-alkyldithiocarbamate accelerator. On the other hand, the
results of the experiments using synthetic polyisoprene latices of different
particle sizes, and therefore having different particle numbers, are most easily
interpreted in terms of a transfer mechanism which involves collision between
the rubber particles and, say, a colloidal species which contains sulphur and
accelerator and which perhaps has the form of a micelle. However, the
experimental evidence demonstrating direct proportionality between initial
rate of crosslink insertion and particle concentration is at present slender.
Regardless of whether or not the sulphur-accelerator species is of colloidal
dimensions, what might be its chemical nature? An obvious suggestion is
that it is the same type of species as that which is postulated to be formed
in the accelerated sulphur-vulcanization of dry rubber. However, to be
effective in the sulphur-prevulcanization reaction, this species has to be soluble
(although perhaps micellized) in an aqueous medium. It could be rendered
soluble in water to a sufficient extent by complexation with some of the
water-soluble non-rubber substances present in normal ammonia-preserved
natural rubber latex concentrate (or which are present in synthetic and
artificial polyisoprene rubber latices). In the absence of zinc dialkyldithiocar-
bamate accelerator, the sulphur on its own is apparently able to form with
such substances species which are sufficiently water-soluble to permit rapid
transfer of sulphur to the rubber particles at typical prevulcanization
temperatures. If suitable other substances are absent from the aqueous phase
of the latex, then it may be that complexation of the zinc atoms of the
accelerator with hydroxide ions from the water provides the sulphur-accelerator
species with sufficient water-solubility for transfer to occur, albeit at a reduced
rate. Perhaps this is why the rate of prevulcanization is considerably reduced
if most of the non-rubber substances are removed from natural rubber latex.
It is significant that one consequence of removal of the non-rubber substances
is extension of the induction period for crosslink insertion. Perhaps this is
a consequence of a much-reduced concentration of sulphur-accelerator
species in the aqueous phase of the latex. If the concentration of this species
can be increased by increasing the hydroxide-ion concentration, or by
increasing the concentration of other substances which can complex with
the sulphur-accelerator species, then there is the basis for a qualitative
explanation of the results of experiments in which the effects of adding various
substances to highly-purified natural rubber latex have been investigated.
Assuming that the transfer of sulphur and accelerator to the rubber particles
occurs by way of adsorption at the particle surface of a water-soluble
surface-active sulphur-accelerator species, it is now necessary to consider
what might be the mechanism by which the sulphur-accelerator species

enters the interior of the rubber particle. The most obvious possibility is that
the sulphur-accelerator species loses by dissociation some or all of the
molecules which have rendered it sufficiently water-soluble for transfer to
the surface of the rubber particles to have occurred. The lost molecules would
then return to the aqueous phase, and be available for complexation with
other molecules ofthe sulphur-accelerator species in due course. The residual
sulphur-accelerator species at the particle surface would now be sufficiently
hydrophobic to migrate into the interior of the rubber particle.
Once the species responsible for the eventual vulcanization of the rubber
has become present in the rubber particles, a sequence of reactions must
occur which results in the rubber molecules becoming crosslinked by
polydisulphidic bridges. The details of that sequence of reactions is obscure
at present. It is not even known whether the rate-controlling process for the
crosslin king of the rubber at any point within a rubber particle is the rate
at which the reactants diffuse to that point from the particle surface, or the
kinetics of the crosslinking reaction itself. If diffusion is the rate-controlling
process, then the concentration of crosslinks in the outer regions of the
vulcanized particles is expected to be greater than that in the interior, and
the non-uniformity of crosslink concentration within the particle is expected
to increase with increasing particle size. If the kinetics of the crosslinking
reaction itself is the rate-controlling process, then the concentration of
crosslinks within the particles is expected to be uniform.
There remains to be considered the matter of the reason why the
sulphur-crosslinking reaction within the rubber particle occurs at much lower
temperatures than is expected ,on the basis of experience with the sulphur-
vulcanization of dry rubber. That both synthetic and artificial polyisoprene
rubber latices are capable of being prevulcanized by heating with sulphur
and accelerator at similar temperatures indicates that the reason is not to
be found in some special molecular species which is present in the interior
of the particles in natural rubber latex. A more plausible explanation is that
the presence of residual hydrophilic moieties which remain complexed to the
sulphur-accelerator species as it migrates into the interior of the rubber
particle facilitates the subsequent sulphuration and rearrangement reactions
by pathways which are not available when natural rubber is vulcanized in
the dry state.
It should be recognized that the ideas discussed here amount to little more
than unsubstantiated speculations which in general terms are not inconsistent
with reported experimental observations. Together, these speculations constitute
only the barest outline of a possible mechanism for the reaction. There is
clearly wide scope here for investigation of a problem which is both
academically challenging and of considerable practical importance.
Other papers which deal with various aspects ofthe sulphur-prevulcanization
reaction in natural rubber latex include those of Hauser, Ie Beau and Kao
[17], of Humphreys and Wake [18], of Sutton [19], and of Gorton [20].
The paper by Sutton gives a useful description of sulphur-prevulcanization
as practised industrially; that by Gorton is specifically concerned with the
production of sulphur-prevulcanized natural rubber latex for dipping processes.
A summary of some of the recent investigations into the kinetics and
mechanism of the sulphur-prevulcanization reaction has been given by
Blackley [21]. The paper by Porter, Rawi and Rahim [16] cited above gives
a useful introduction to some of the issues which have to be resolved in order
to advance our understanding of the mechanism of the reaction.
13.2.8 Properties of films derived from sulphur-prevulcanized natural rubber latex
13.2.8.1 Mechanical properties of.films

Films which have been dried down from sulphur-prevulcanized natural
rubber latex closely resemble films which have been cast from an equivalent
compounded unvulcanized natural rubber latex and then postvulcanized.
However, in making comparisons between the two types of film, it must be
recognized that differences between them tend to be minimized if, as is usual
industrial practice, films from sulphur-prevulcanized latex are subjected to
subsequent postvulcanization. If precautions are taken to avoid postvulcan-
ization, implying the use of relatively low drying temperatures, then tensile
strengths tend to be rather low, being typically ca. 10-15 MPa. Extensions
at break are ca. 700-800%. If, as is usual, the drying conditions are such
that significant postvulcanization occurs after the particles of the film have
integrated, then, as indicated in Tables 13.2-13.6, higher tensile strengths
(ca. 20- 30 M Pa) and extensions at break (800-1000%) are obtained. A further
increase in tensile strength usually occurs if the film is then leached in water
(see section 13.2.4 above).
The tensile stress-strain curve for films from sulphur-prevulcanized natural
rubber latex formed under conditions which preclude significant post-
vulcanization are intermediate in character between those for unvulcanized
latex film and those for postvulcanized latex film. At low extensions (up to
ca. 500%), the curve for the prevu\canized film resembles that for the
unvulcanized film, having a modulus somewhat higher than that for
unvulcanized film but lower than that for postvulcanized film. Furthermore,
at low extensions, the tensile stress-strain curve for prevulcanized film is
rather fiat, as is that for unvulcanized film. Then, at extensions exceeding ca.
500%, the modulus increases sharply in the same way as for a postvu\canized
film. The tensile strength of vulcanized films depends upon the conditions
under which the film was dried, but is always considerably greater than that
of un vulcanized films.
13.2.8.2 Behaviour towards rubber solvents
When immersed in rubber solvents such as benzene or carbon tetrachloride,

it is found that films from sulphur-prevulcanized natural rubber latex swell
very considerably to give gels which are mechanically rather weak. This
weakness is especially apparent if care has been taken to minimize
postvulcanization during the drying of the film. Humphreys and Wake [18]
observed that, if significant postvulcanization is avoided, then films from
sulphur-prevulcanized natural rubber latex tend to disintegrate into a mass
of very small separate particles of vulcanized rubber swollen with solvent. It
should be noted that, although extensive disintegration and dispersion may
occur, true dissolution does not occur. This behaviour of deposits from sulphur-
prevulcanized natural rubber latex contrasts with that of deposits from
unvulcanized latex on the one hand, and that of postvulcanized latex deposits
on the other. Unvulcanized deposits swell extensively with considerable
dissolution of the rubber, whereas postvulcanized latex deposits behave like
normal rubber vulcanizates in that they swell to a limited extent which
depends upon the nature of the solvent and the concentration of crosslinks
in the rubber network. Observations on the behaviour of films from prevulcanized
natural rubber latex towards rubber solvents accord well with the inferior
resistance of these films to oils, greases and solvents, as compared with
postvulcanized latex films. Thus a prevulcanized film can readily rupture if a
drop of solvent is placed upon it whilst it is held in an extended condition.
This does not happen with postvulcanized latex films. It is for this reason that
latex-dipped gloves which have been produced from prevulcanized natural
rubber latex are more readily damaged by oil and grease than are those which
have been vulcanized after the glove film has been formed. However, this
disadvantage of prevulcanized films relative to post vulcanized films can be
mitigated to some extent by subjecting the prevulcanized film to postvulcanization.
A further interesting observation reported by Humphreys and Wake [18]
is that the equilibrium swelling of deposits from sulphur-prevulcanized
natural rubber latex is considerably less sensitive to combined-sulphur
content than are conventional sulphur-vulcanizates from natural rubber. The
results of Humphreys and Wake are shown in Fig. 13.10; this shows the
variation of equilibrium swelling in benzene with combined-sulphur content
for both prevulcanized and postvulcanized latex deposits. The significance
of this observation is far from clear. For instance, it is not clear to what
extent the difference in behaviour between the two types of vulcanizate
towards the solvent is reduced if the prevulcanized deposit is subsequently
postvulcanized. Nor is it clear whether the origin of the difference in behaviour
lies in a difference in the relationship between equilibrium swelling and
crosslink concentration for the two types of vulcanizate, or whether it lies
in a difference in the relationship between combined sulphur content and
crosslink concentration. If the latter is the case, then the implication is that,
1500
~
'#
.......
CD
5i 1000
N
C
CD
.c
.!:
0>
.!:
1E
::s 500
g
'5
0-
w
°0~----------~1-----------2L---------~3
Combined sulphur content 1% mlm
Figure 13.10 Variation of equilibrium swelling in benzene with combined-sulphur content for
prevulcanized (points e) and postvulcanized (points &) natural rubber latex deposits (Humphreys
and Wake [18]). Formulations not specified, but said to be 'similar' for both latices.
at low combined-sulphur contents, but not at higher combined-sulphur

contents, the average number of sulphur atoms per crosslink can be
considerably lower for prevulcanized latex film than for postvulcanized latex
film. Humphreys and Wake themselves incline to the view that the reason
for the phenomenon illustrated in Fig. l3.10 is that the concentration of
crosslinks in the surface regions of the particles is higher than in the interior
when the combined-sulphur content is low.
Humphreys and Wake [18] also found that the chloroform extract of
acetone-extracted deposits from prevulcanized natural rubber latex is negligible.
From this observation, they infer that the rubber within the individual latex
particles is completely crosslinked. However, this leaves unresolved the matter
of whether or not the distribution of crosslinks throughout the particles is uniform.
13.2.9 Structure of films derived from sulphur-prevulcanized natural rubber latex
13.2.9.1 Resume of theories

We turn now to consider a matter which has been the subject of considerable
discussion and controversy, namely, the structure of films derived from
sulphur-prevulcanized natural rubber latex. It seems unlikely that the final

word has yet been said concerning this matter. It has been noted above (see
section 13.1) that it would not be expected from general principles that the
particles in a prevulcanized latex would dry down to coherent, well-integrated
films, because the crosslinks between the rubber macromolecules would
restrict the ability of the molecular segments in one particle to interdiffuse
with those of contiguous particles. Three types of theory have been proposed
to explain the strength and coherence of films derived from prevulcanized
natural rubber latex. They are:
1. a theory in which the strength is attributed to primary valence bonds
formed between the rubber macromolecules in contiguous particles;
2. a theory in which the strength is attributed to the interparticle bonding
effect of various non-rubber substances derived from the latex, such
substances acting as an adhesive which bonds otherwise discrete particles
of unvulcanized rubber into a coherent film;
3. a theory in which the strength is attributed to secondary valence bonds
which form between segments of the rubber macromolecules in the surfaces
of contiguous particles.
Each of these theories is now considered in turn, together with their
implications for the structure of pre vulcanized latex films. Much of what
follows is based upon ideas discussed in the paper by Humphreys and Wake
[18], to which reference has already been made.
13.2.9.2 The primary-valence-bond theory

According to the first of the above theories, the strength and coherence of
films from sulphur-prevulcanized natural rubber latex depends primarily
upon the formation of interparticle covalent chemical bonds between the
rubber macromolecules in contiguous particles, such bonds being formed as
the latex film dries. The interparticle covalent chemical bonds between the
rubber macromolecules formed at the prevulcanization stage serve principally
to immobilize the rubber macromolecules within the individual particles.
According to this view, deposits from sulphur-prevulcanized latex have a
similar structure to postvulcanized latex films. In both, the identity of the
original latex particles may be partially retained, but the strength and
coherence derive from the existence in the film of three-dimensional covalent
networks of indefinite extent which are considerably larger than were the
original latex particles. However, the structures of the two types of film differ
in that, in principle, two types of network can be distinguished in deposits
from sulphur-prevulcanized latices, whereas only one type is likely to be
present in post vulcanized deposits. One type of network in prevulcanized
films is that which was formed during the prevulcanization reaction; the
second is that which was formed as the film dried. Deposits from prevulcanized
latex could therefore be visualized as comprising a dispersion of prevulcanized

latex particles embedded in, and covalently bonded to, a reticulum of
postvulcanized rubber.
It is probable that interparticle covalent chemical bonds between the rubber
macromolecules in contiguous particles do form when films from sulphur-
prevulcanized natural rubber latex are dried under the normal conditions of
elevated temperature. Indeed, as has been stated above, it is often the intention
in industrial practice that such covalent bonds should be formed, and that
concomitant improvements in mechanical properties should accrue. There
is a ready mechanism by which such interparticle bonds can form. It is
well-known that polysulphidic crosslinks are thermally labile, and are able
to rearrange amongst themselves at elevated temperatures. It is presumably
such rearrangements, together with the formation offurther sulphur crosslinks
if residual vulcanizing ingredients are present, which lead to the formation
of interparticle crosslinks.
However, there are several cogent reasons for believing that the formation
of interparticle crosslinks during drying is not essential to the formation of
strong coherent deposits from prevulcanized latices. It is for these reasons
that the primary-valence-bond theory has been rejected. The more important
of them are as follows:
1. Strong coherent deposits can be formed from sulphur-prevulcanized

natural rubber latex when drying takes place under conditions such that
no rearrangement of crosslinks is likely to have occurred, nor is further
vulcanization likely to have occurred.
2. Strong coherent deposits can also be formed from non-sulphur-prevulcanized
natural rubber latices which are such that crosslink-rearrangement is
unlikely to occur. Vulcanization by heating with organic peroxides, or by
exposure to high-energy irradiation, is believed to be a consequence of
the formation of thermally-stable carbon-carbon crosslinks which are
unlikely to rearrange during the subsequent drying of latex films. Some
post vulcanization may perhaps occur if films from these latices are dried
at elevated temperatures, and if residual peroxides are present in
peroxide-prevulcanized latices, or ifradiation-prevulcanized latices contain
free radicals trapped in the particles or peroxides which have formed
during irradiation in the presence of oxygen. These possibilities do not
appear to have been investigated so far.
3. If the strength and coherence of sulphur-prevulcanized films is primarily
a consequence of interparticle covalent chemical bonds, then it is difficult
to explain the differences between prevulcanized and postvulcanized films.
In particular, it is difficult to understand why
(a) the tensile stress-strain curves for prevulcanized films are intermediate
in character between that for unvulcanized films and that for
postvulcanized films (see section 13.2.8.1 above);
(b) the behaviour of pre vulcanized films towards rubber solvents can differ
significantly from that of postvulcanized films (see section 13.2.8.2 above).
Thus, to the extent that the primary-valence-bond theory explains anything
at all, it explains rather too much.
13.2.9.3 The 'bonding-adhesive' theory

According to the second of the theories for the strength and coherence of
deposits from sulphur-prevulcanized latex, the deposits comprise a close-
packed assemblage of small particles of vulcanized rubber bonded together
by non-rubber substances situated in the interfaces between the particles.
This theory is at first sight inherently implausible, because the phase inversion
which occurs when a latex dries out to a coherent deposit would be expected
to tend to leave the non-rubber substances dispersed in small pockets
throughout a continuous matrix of rubber, rather than spread out over the
surface of the particles in such a way that they could exert an appreciable
adhesive effect. However, in considering the problem of film-formation from
aqueous latices in section 4.3 of Chapter 4 (Volume 1), it is noted that just
such a theory has been proposed to account for the effect of hydrocolloids
in enhancing the strength of films from certain synthetic latices (see section
4.3.5). However, notwithstanding this latter hypothesis, it is obvious that a
crucial test of the theory as applied to films from sulphur-prevulcanized
latices is to remove as much as possible of the non-rubber substances from
a latex film, and ascertain whether or not the strength and coherence of the
film are reduced. If all the non-rubber substances are removed from the film,
and the theory is correct, then it would be expected that the film would
disintegrate. The results of such a test have been reported by Humphreys
and Wake [18]. Ammonia-preserved natural rubber latex was creamed four
times in the presence of a carboxylate soap (sodium laurate) to displace the
proteinaceous substances adsorbed at the surface of the particles. The latex
was then sulphur-prevulcanized using sodium diethyldithiocarbamate as the
accelerator and sodium methylene-bis(naphthalenesulphonate) as the dispersing
agent for the other ingredients. Some coagulum formed during prevulcanization,
because the colloid stability of the latex was low. This was removed by
straining. Films were dried from the prevulcanized latex on glass plates at
normal ambient temperature over a period of2 weeks. When drying appeared
to be complete, the film was examined and found to have a strength similar
to that of a normal sulphur-prevulcanized latex film. A second film was
refluxed with ethyl alcohol for 2 hours, and then dried for 1 week at normal
ambient temperature. The strength of this film was also found to be similar
to that of a normal sulphur-prevulcanized latex film. Because most of the
non-rubber substances had been replaced by an ethanol-soluble soap through
the repeated creamings, and the dispersing agent used to prepare dispersions
of the insoluble compounding ingredients was also ethanol-soluble, it was
inferred that little interparticle adhesive substance could have remained in

the film after extraction with ethyl alcohol. The conclusion drawn is that,
even if non-rubber substances capable of functioning as interparticle bonding
adhesives are present in sulphur-prevulcanized natural rubber latex, they
contribute little to the strength and coherence of films formed from the latex.
Further evidence to support rejection of this theory has accumulated since
the investigation reported by Humphreys and Wake [18]. Thus Ghazaly
[14] has found it possible to obtain coherent films of reasonable strength
from sulphur-prevulcanized natural rubber latices prepared from ammonia-
preserved natural rubber latex from which much of the adventitious non-rubber
substances had been removed by enzymatic deproteinization followed by
multiple centrifugation. Coherent films of reasonable strength can also be
formed from sulphur-prevulcanized artificial and synthetic poly isoprene
rubber latices and, indeed, from other synthetic latices of diene rubbers which
have been sulphur-prevulcanized. These further observations support the
view that the strength and coherence of the films is not primarily dependent
upon the presence of appropriate non-rubber substances in the latex.
13.2.9.4 The secondary-valence-bond theory

According to the third of these theories for the strength and coherence of
deposits from sulphur-prevulcanized latex, the deposits have a structure
similar to that envisaged by the second of the theories, except that bonding
is now by way of secondary-valence attractive interactions between segments
of rubber macromolecules in the surfaces of contiguous particles, rather than
by way of non-rubber substances acting as a bonding adhesive. These
interactions are envisaged as developing as the latex dries and the surfaces
of contiguous particles come into increasingly close contact, and as the region
of overlap between the surfaces of the two particles increases in volume. The
situation for a single pair of contiguous particles is illustrated schematically
in Fig. 13.11. Individual secondary-valence bonds are, of course, much weaker
than are individual primary covalent bonds. The credibility of this theory
depends upon two considerations. The first is that the number of secondary-
valence bonds which can form between the surface regions of a pair of
contiguous particles is very large. The second is that all the rubber
macromolecules within any given particle will have been covalently linked
into a single network of indefinite extent, or, at most, into a small number
of such networks, so that each particle is effectively either a single molecular
network or comprises a few interpenetrating molecular networks. Thus, in
order to separate two contiguous particles, it is necessary to break all the
secondary-valence bonds which have been formed in the region of overlap
between the surfaces ofthe two particles. Some idea ofthe order of magnitude
of the consequent scaling-up of the secondary valence forces which exist
between pairs of primary rubber macromolecules can be gained from the
Region of overlap between two particies
Figure 13.11 Schematic illustration of single pair of contiguous particles in films derived from
sulphur-prevulcanized natural rubber latex.
consideration that a spherical particle of diameter I J.lm contains ca. 106

primary rubber macromolecules ifthe molecular mass of those macromolecules
is 3 x 10 5 daltons (see section 2.2.3.3 of Chapter 2 (Volume 1». Thus, although
individual bonds are very weak, the total force required to separate the pair
ofparticIes is sufficient to account for the strength and coherence of the deposit.
This is the theory now generally accepted for the strength and coherence
of deposits from all types of pre vulcanized rubber latex. It accords well with
the observed properties of films from these latices. As regards the tensile
stress-strain behaviour, the initial extension of the film leads to rearrangement
of the rubber macromolecules in the region of overlap between two particles.
Because at this stage these macromolecules are able to move essentially
independently of each other, the tensile stress-strain curve will approximate
to that of an unvulcanized latex film, it being generally agreed that the
strength and coherence of such films are almost entirely a consequence of
secondary-valence forces. As extension proceeds, there comes a point at which
the rubber macromolecules in the region of overlap are not any longer able
to move independently of each other because they are attached to large
numbers of other primary rubber macromolecules by crosslinks. In order to
continue extension of the film, it is now necessary simultaneously to overcome
the secondary-valence forces which operate between increasingly large
numbers of primary rubber macromolecules. In this way, the general shape
of the tensile stress-strain curve can be explained. The initial flat region of
the curve, where the modulus is low, is a consequence of the primary rubber
macromolecules in the region of overlap moving independently of each other,
so that the behaviour is similar to that of un vulcanized rubber. The increasing
modulus observed after ca. 500% extension is a consequence of increasing
interdependence of the motion of these macromolecules. The behaviour of
prevulcanized latex films towards rubber solvents can also be satisfactorily

explained on the secondary-valence-bond theory. In the presence of a rubber
solvent, the particles swell as they absorb solvent, and the secondary-valence
forces weaken as the average separation between the segments of the rubber
macromolecules increases. If no covalent bonds have formed between the
macromolecules in contiguous particles, then the swollen gel is expected
eventually to disintegrate to a dispersion of particles of swollen, vulcanized
rubber of a size corresponding to that of the individual particles of the initial
unvulcanized latex.
13.2.10 Industrial grades of sulphur-prevulcanized natural rubber latex

Some users of sulphur-prevulcanized natural rubber latex produce their own
prevulcanized latex; others purchase the latex already prevulcanized. Table
13.7 gives details of the principal grades of sulphur-prevulcanized natural
rubber latex concentrate which are currently available from Revertex under
the trade name 'Revultex'. No further vulcanization of films from these latices
is said to be necessary. There is thus no requirement to compound the latices
with additional vulcanizing ingredients. These latices are normally supplied
containing a rubber antioxidant approved by the United States of America
Table 13.7 Information concerning some of the grades of sulphur-prevulcanized

natural rubber latex concentrate currently available from Revertex [22]
Grade of Revultex
Property LR MR HR LA
Properties of latices
Total solids content 1% m/m 60.5 60.5 60.5 60.5
Ammonia content I % m/m on whole latex 0.6 0.6 0.6 0.3
pH 10.5 10.5 10.5 10.2
Viscosity at 30T (Ford Cup # 3) Is 30 35 40 32
Properties of unaged leached and desiccated films·
Modulus at 700% extension I MPa 9.5 12.0 18.0 11.0
Tensile strength I MPa 31.0 31.0 27.0 31.0
Extension at break I % 1000 900 800 950
Properties of unaged leached and desiccated films after conditioningt
Modulus at 700% extension I MPa 8.5 11.0 \6.0 10.0
Tensile strength I MPa 30.0 30.0 26.0 30.0
Extension at break I % 1000 900 800 950
• Films prepared by pouring S<flo m/m latex on to glass plates, drying at 23"C and S<flo relative
humidity until clear (2- 3 days) to give films of thickness ca. 0.5 mm, leaching for at least
16 hours in running water, redrying, then storing in a desiccator over anhydrous calcium chloride
for at least 48 hours. The films were tested at 23·C immediately after removal from the desiccator.
t Films prepared as above, but, after removal from the desiccator, conditioned for at least
16 hours at 23·C and S<flo relative humidity, and then tested under the same conditions.
Food and Drug Administration (FDA) and the German Bundesgesundheitsamt

(BGA) (Federal Health Department). Three grades of high-ammonia and one
grade of low-ammonia pre vulcanized latex are available. The high-
ammonia grades are distinguished as low-modulus (LR), medium-modulus
(MR) and high-modulus (UR), according to the modulus of films which dry
from the latex. The low-ammonia grade gives a film of medium modulus. It
has been developed specifically for the manufacture, by heat-sensitized
dipping, of films having high optical clarity. Other grades of sulphur-
prevulcanized natural rubber latex are also available, e.g. grades which have
been double-centrifuged to reduce the concentration of non-rubber substances,
and grades which have been formulated to minimize the levels of extractable
N-nitrosamines and N-nitrosatables in films derived from the latex.
13.3 PREVULCANIZATION OF NATURAL RUBBER LATEX BY

REACTION WITH ORGANIC PEROXIDES (INCLUDING
HYDRO PEROXIDES), AND WITH HYDROGEN PEROXIDE
13.3.1 Introduction
The possibility of pre vulcanizing natural rubber latex by heating with organic
peroxides has been known for many years. It was inevitable that it should
be investigated, because it has been known since ca. 1915 that it is possible
to vulcanize dry natural rubber latex by compounding with organic peroxides
and then subsequently heating. The vulcanization of natural rubber in both
bulk and latex forms occurs through the formation of carbon -carbon
crosslinks between the primary rubber macromolecules. The mechanism by
which crosslinking is believed to occur in dry natural rubber is in principle
simple. Under the influence of heat, the oxygen-oxygen bond of the organic
peroxide dissociates to give two free radicals. Each of these radicals can then
abstract one hydrogen atom from a rubber macromolecule. The hydrogen
atoms which are abstracted are believed to be principally those of the
IX-methylene groups adjacent to the olefinic double bonds. Subsequent
combination of the resultant hydrocarbon free radicals leads to the formation
of covalent carbon-carbon bonds between the rubber molecules. Ideally, one
carbon -carbon crosslink is formed for each molecule of peroxide which
decomposes. The number of crosslinks formed will be less than this if some
of the decomposed peroxide is wasted through side reactions, and more than
this if the decomposing peroxide induces crosslin king by a chain-reaction
mechanism.
There is little doubt that the peroxide-pre vulcanization of natural rubber
latex proceeds by a broadly similar mechanism to that for dry natural rubber,
although the details are obscure. Thus, for example, it is not clear to what
extent the initial decomposition of the peroxide occurs in the aqueous phase,
Prevulcanization by reaction with organic peroxides 493
rather than in the rubber particles after the peroxide has been absorbed into
them. Presumably this balance depends upon the nature of the organic
peroxide. Another uncertainty concerns the extent to which, for prevulcanization
systems in which the peroxide is the only added vulcanization ingredient (i.e.,
no activator for the decomposition of the peroxide is deliberately added),
the initial radical-forming process occurs by redox-induced decomposition
of the peroxide, rather than by simple thermal dissociation. Unless very
carefully purified, natural rubber latex contains organic reducing substances
which are able to form redox systems with peroxides. It may be that activated
peroxide-decomposition then occurs. A third uncertainty concerns the extent
to which other peroxide-induced reactions occur in addition to, or instead
of, crosslinking. Peroxide-induced polymer oxidation, to which further
reference is made below, is an obvious possibility. Another possibility is
epoxidation of the rubber molecules. This latter reaction is described in
section 14.2 of Chapter 14.
With aqueous latices, one additional peroxide can be used which is not
available for dry-rubber vulcanization, namely, hydrogen peroxide. The
peroxide then partitions almost exclusively in the aqueous phase of the latex,
and the initial radical-forming reaction is expected to occur almost entirely
within that phase.
Notwithstanding the inherent simplicity of the peroxide-vulcanization
reaction, there was for many years little industrial interest in either the
peroxide-vulcanization of dry natural rubber, or the peroxide-prevulcanization
of natural rubber latex. Lack of interest in the peroxide-vulcanization of
natural rubber was, in part at least, a consequence of the relatively poor
ageing behaviour of peroxide-vulcanizates as compared with that of those
obtained by sulphur-vulcanization. This defect is particularly serious for
thin-walled products made by latex dipping; the ratio of surface area to bulk
is relatively high for these products, and thus access of oxygen to the rubber
is facilitated. That the ageing behaviour of deposits from peroxide-prevulcanized
natural rubber latex should be bad is not surprising, bearing in mind that
the oxidative degradation of hydrocarbon rubbers has for long been known
to occur by way of a peroxide-induced chain oxidative reaction which leads
to scission of the main chains of the rubber network. However, the crosslinks
which are introduced into hydrocarbon rubber networks by peroxide-
vulcanization are more thermally stable and generally less reactive chemically
than are sulphur crosslinks. Lack of interest in peroxide-prevulcanization of
natural rubber latex has been a consequence partly of reservations about the
ageing behaviour of the resultant vulcanizates. A second factor has been the
absence of any perceived need for non-sulphur-prevulcanization systems.
Sulphur-prevulcanization has been found to give a product which has been
generally satisfactory for those applications in which it is chiefly used.
Renewed interest in the possibility of using peroxides to prevulcanize natural
rubber latex has been stimulated mainly by recognition of the need to
minimize the formation of N-nitrosamines and N-nitrosatable substances in

vulcanized rubber products which are intended to come into contact with
the human body. In principle, peroxide-vulcanization provides one route to
the production of prevulcanized natural rubber latex which is low in levels
of such substances. The reason for this is that no conventional organic
vulcanization accelerators are used in conjunction with peroxides, these
accelerators being a major source ofN-nitrosamines and N-nitrosatables (see
section 16.3.6 of Chapter 16 (Volume 3». Furthermore, rubber articles made
from natural rubber latex which has been prevulcanized without the use of
sulphur, conventional vulcanization accelerators and zinc oxide, have the
following additional advantages:
1. If used products are disposed of by incineration, the evolution of
sulphur-containing gases such as sulphur dioxide is minimized, as also is
the amount of residual ash if the latex has not been compounded with
inorganic substances.
2. The absence of conventional sulphur-vulcanization accelerators implies
the absence of physiological disadvantages, such as toxicity, irritancy and
bitter taste, which can be associated with these accelerators (see section
16.3.7 of Chapter 16 (Volume 3». However, there is the possibility of
physiological problems arising from the presence in the rubber product
of residual peroxide and peroxide-decomposition products.
3. Insofar as rubber-insoluble additives are absent from the prevulcanized
latex, it is to be expected that films from peroxide-prevulcanized natural
rubber latex will be more transparent than those from sulphur-prevulcanized
natural rubber latex, especially if the latter contains a rubber-insoluble
compound such as zinc oxide.
13.3.2 Older investigations

Early processes for the peroxide-prevulcanization of natural rubber latex
include those described in patents to Stevens [23], to Stott [24,25], to
Hercules Powder Company [26], and to Ropp [27]. More recently, Rodaway
and Calvert [28] have described a process for the peroxide-prevulcanization
of various latices of olefinically-unsaturated rubbers, including natural rubber
latex, by reaction with hydrogen peroxide and a hydrogen-peroxide activator
which does not encourage decomposition of the hydrogen peroxide to
molecular oxygen under the conditions of the prevuIcanization reaction.
Specifically, inorganic and organic compounds which yield peroxo acids or
peroxo salts by reaction with hydrogen peroxide in aqueous medium are
said to be suitable as activators. The inorganic compounds include sodium
and potassium molybdates, tungstates, stannates, borates and metasilicates.
The organic compounds include formic acid, formaldehyde and fluoroacetic
acid. The question arises as to the extent to which epoxidation, rather than
crosslin king, occurs when reagents such as hydrogen peroxide and formic
acid are used in this process (see further section 14.2 of Chapter 14). When
this process is used with natural rubber latex, it is said to be advantageous
to sequester traces of transition-metal ions, such as those of copper and
manganese, by the addition of a metal-ion complexing agent, in order to
discourage the rapid decomposition of the hydrogen peroxide to water and
oxygen. It is also said to be advantageous to carry out the prevulcanization
reaction under acidic conditions, for the same reason. In a typical process,
100 parts by volume of 60% mlm ammonia-preserved natural rubber latex
(0.7% mlm ammonia) is treated at normal ambient temperature with 2 parts
by volume of 5% mlv aqueous disodium ethylenediaminetetraacetate solution,
t part by volume of 5% mlv aqueous sodium tungstate solution, and 5 parts
by volume of tOO-volume hydrogen peroxide. The reaction time is not
disclosed. The tensile strength of films obtained from the treated latex is
relatively low (ca. 12 MPa), but is higher than that of films from the untreated
latex (ca. 7.6 MPa). No results for modulus are given. Examples using latices
of other types of olefinically-unsaturated rubbers indicate that the treatment
brings about some increase in modulus, and hence some changes in the
rubber. These changes are consistent with crosslinking, although they do not
establish it unequivocally.
13.3.3 More recent investigations

Said, Pendle and Blackley [29] have reported a summary of a recent extensive
investigation into the peroxide-prevulcanization of ammonia-preserved natural
rubber latex. This investigation was motivated by the desirability of developing
prevulcanized natural rubber latex containing low levels of N-nitrosamines
and N-nitrosatable substances. The primary purpose of the investigation was
to evaluate the effectiveness of a range of organic peroxidic substances which
are currently available industrially. The substances investigated were
representative of the classes of alkyl hydroperoxides (Structure I), diacyl
peroxides (Structure II), alkyl esters of peroxoalkanoic acids (Structure III),
alkyl esters of peroxobenzoic acid (Structure IV), alkyl peroxocarbonates
(Structure V), alkyl peroxodicarbonates (Structure VI), and dialkyl peroxoketals
(Structure VII). Some peroxides were used without the addition of activators,
although it was recognized that compounds capable of functioning as
peroxide-activators would inevitably be present in the latex. In other cases,
an activator was added with the peroxide. Attempts were made to correlate
effectiveness as a prevulcanization agent with the rate of decomposition of
the peroxide, as quantified by the half-life for the non-activated decomposition
of the latter at the prevulcanization temperature. The half-life of a peroxide
is expected to give an indication of the minimum rate of decomposition in
the prevulcanization reaction system. Amongst a range of activators investigated,
fructose was found to be particularly suitable. As well as effectiveness as an
activator, it has other advantages, such as lack of toxicity and other
RO-OH R.CO.O-O.CO.R R'.CO.O-OR C 6 HS·CO.O-OR

I II III IV
RO-O.CO.O-OR RO.CO.O-O.CO.OR R~C(OORh
V VI VII
health hazards, and lack of significant effect upon the colloid stability of the
latex. Further significant activation could sometimes be achieved by addition
of trace amounts of iron(III) ions to a peroxide-fructose combination. The
iron was added in the form of an iron(III)-pyrophosphate complex. The
peroxide was emulsified in a small amount of water prior to addition to the
latex. A non-ionogenic ethoxylate was used as emulsifier for prevulcanization
reactions at temperatures below 90·C; potassium oleate was used when the
temperature was lOO·C. The emulsifier level was generally in the range
15-20% m/m on the peroxide. A typical reaction system for the prevulcanization
of natural rubber latex as 60% m/m ammonia-preserved concentrate comprised
(parts by mass dry) rubber 100, 75% tert-butyl peroxoisobutyrate l.lS5, and
fructose (as a 20% m/v aqueous solution) I. The levels of peroxide and fructose in
this system are both equivalent to 5.55 x 10- s mol(g rubber)-l. Non-ionogenic
ethoxylate stabilizer 0.2 was used to emulsify the peroxide in water.
The conclusions from this investigation can be summarized as follows:
1. Di-n-octanoyl peroxide, di-n-decanoyl peroxide and bis-2-ethylhexyl

peroxodicarbonate are ineffective as prevulcanizing agents under a
reaction condition of 6 hours at 70·C. The half-lives of these peroxides
at 70·C indicate that at least 75% of these peroxides should have
decomposed during this period. Ineffectiveness as prevulcanization agents
is therefore attributed to low reactivity of the free radicals formed in the
system.
2. Using di-tert-butyl peroxide without added activator gives slight
prevulcanization after 24 hours heating at lOO·C.
3. The peroxoketal I, l-bis-(tert-butylperoxo)-3,3,5-trimethylcyclohexane is
ineffective at SO·C, even if an added activator is present.
4. As shown in Table 13.S, the use of tert-butyl hydroperoxide without
added activator gives a relatively low concentration of crosslinks after
24 hours heating at lOO·C. The half-life of tert-butyl hydroperoxide at
lOO·C in the absence of activators is long (> 1500 h). It is therefore
probable that the decomposition of the peroxide is strongly activated by
certain non-rubber substances in the latex.
5. Also as shown in Table 13.S, various peroxoesters without added activator
are more effective after 24 hours heating at lOO·C than is tert-butyl
hydroperoxide.
6. Of the range of activating substances investigated in conjunction with
tert-butyl hydroperoxide, fructose is the most effective.
Table 13.8 Concentration of crosslinks found in films obtained from ammonia-

preserved natural rubber latex prevulcanized with a hydroperoxide and with various
peroxoesters in the absence of added activators (Said, Pendle and Blackley [29])
Prevulcanization conditions: 24 hours at 100°C.
Concentration of
Concentration of chemical cross links
peroxide added found in film from
to latex prevulcanized latex
Peroxide ! mol (g rubber) - 1 ! mol (g rubber) - 1
tert-butyl hydroperoxide 11.1 10- 5
X 0.04 X 10- 5
tert-butyl peroxo-2-ethylhexanoate 11.3 10- 5
X 3.55 x 10- 5
tert-butyl peroxoisobutyrate 11.2 x 10- 5 2.88 x 10- 5
t ert-butyl peroxoisopropylcarbonate ILl X 10- 5 1.39 x 10- 5
tert-butyl peroxoacetate 11.0 x 10- 5 0.96 x 10- 5
tert-butyl peroxobenzoate 10.6 X 10- 5 0.05 x 10- 5
7. PrevuIcanization at 40"C using fructose-activated tert-butyl peroxo-2-

ethylhexanoate occurs only very slowly. However, the rate can be
significantly increased by the addition of a trace amount of an iron(III)
pyrophosphate complex, notwithstanding that trace amounts of transition-
metal ions which would be expected to activate peroxide decomposition
are known to be present in ammonia-preserved natural rubber latex.
8. The variation of concentration of crosslinks in the rubber network with
time of reaction is kinetically almost exactly first order, consistent with
the expectation that the decomposition of a peroxide in natural rubber
latex is kinetically of first order in peroxide concentration.
9. Crosslinking efficiencies, defined as average number of crosslinks formed
per molecule of peroxide decomposed, are in the range 0.05 -0.15 for
prevuIcanizations using fructose-activated peroxoesters at 60 e. Crosslinking
0
efficiency increases approximately linearly with prevuIcanization tem-

perature over the range 50-80·e. There appears to be no obvious
correlation between crosslinking efficiency and chemical structure. Of
the combinations investigated, fructose-tert-butyl peroxo-2-ethylhexanoate
and fructose-tert-butyl peroxoisobutyrate are the most efficient.
10. Films having tensile strengths in the range 25-30 MPa can be obtained
from peroxide-prevuIcanized natural rubber latex, but the tensile strength
can depend rather sensitively upon the concentration of crosslinks in the
rubber network.
11. The ageing behaviour of films from peroxide-prevuIcanized natural rubber
latex is generally poor, but can be significantly improved by addition of
an antioxidant to the latex. By far the most effective of the antioxidants
investigated is zinc diethyldithiocarbamate. This is unfortunate, because
dithiocarbamates are potential sources of secondary amines, and the latter
are nitrosatable to N-nitrosamines (see section 16.3.6 of Chapter 16 (Volume 3)).
13.4 PREVULCANIZATION OF NATURAL RUBBER LATEX BY

EXPOSURE TO HIGH-ENERGY RADIATION
13.4.1 Introduction
It has been known since the late 1920s that natural rubber can be crosslinked
by exposure to electrons from a cathode-ray tube. Serious interest in the
possibility of crosslin king rubber by exposure to various forms of high-energy
radiation dates from the period after the end of World War II when convenient
sources of these radiations became available for industrial applications. The
mechanism by which crosslinking is believed to occur in dry natural rubber
is probably rather more complex than that by which peroxide-crosslinking
occurs. Absorption of the radiation by the polyisoprene molecules leads to
the formation of transitory excited molecular states which can in principle
decay in at least three ways:
l. by elimination of a hydrogen atom, most probably from an IX-methylene

group adjacent to an olefinic double bond, thereby leaving a free-radical
site on the polymer chain;
2. by elimination of an electron, leaving a cationic site on the polymer chain; and
3. by main-chain scission, giving two polymeric free radicals.
Polymer crosslinking is believed to occur mainly by the subsequent combination

of polymeric free radicals with the formation of covalent bonds between the
polymer molecules. The covalent bonds which form are between carbon
atoms, and so are in principle very stable thermally.
Having discovered that bulk natural rubber could be crosslinked by
exposure to high-energy radiations, it was inevitable that the possibility of
prevulcanizing natural rubber latex by similar means would be investigated.
Insofar as prevulcanization does occur, and incident radiation is absorbed
by the rubber particles and not by the aqueous phase, it is to be expected
that the mechanism by which crosslinking occurs is similar to that for the
radiation-crosslinking of dry natural rubber. However, with natural rubber
latex, the further possibility arises that free radicals are generated in the
aqueous phase of the latex by interaction between the high-energy radiation
and water and other molecules. These radicals may then be absorbed into
the rubber particles, and may bring about crosslinking by hydrogen abstraction
and subsequent combination ofthe resultant polymeric free radicals, analogously
to crosslin king by free radicals derived from the decomposition of peroxides.
Any crosslin king which occurs by this mechanism will be additional to that
which results from direct interaction between the high-energy radiation and
the polymer molecules within the rubber particles. It is known that high-
energy electromagnetic radiations can interact with water molecules in
various ways. One possibility is decomposition into hydrogen and hydroxide
free radicals. Several fates for these radicals are possible. They may, for
Prevulcanization by exposure to high-energy radiation 499
instance, be adsorbed at the surface of the rubber particles, and then

subsequently abstract hydrogen atoms from rubber molecules; or they may
recombine to form water or hydrogen plus hydrogen peroxide; or they may
interact with other water molecules to form various free-radical species.
Industrial interest in the prevulcanization of natural rubber latex by
exposure to high-energy radiation has been inhibited by the inconvenience
and expense of equipment for carrying out the necessary exposure, and by
potential health hazards associated with the process. Nevertheless, the process
is of considerable potential industrial interest now, mainly for the same reason
that interest in peroxide prevulcanization has been renewed. The three
additional potential advantages accruing to peroxide-prevulcanized natural
rubber latex, noted in section 13.3.1 above, are expected to be realized to an
even greater extent with latex which has been prevulcanized by exposure to
high-energy radiation. Products from the latex which are eventually disposed
of by incineration lead to minimum evolution of sulphur-containing gases
and leave minimum amounts of residual ash. Physiological disadvantages
from added organic chemicals (including organic peroxides and their activators)
are absent. Thin films are expected to have maximum transparency, unless
rubber-insoluble substances have been added to the latex, or unless the
radiation has interacted with non-rubber constituents to produce substances
which impair transparency. Prevulcanization by exposure to high-energy
radiation has the considerable advantage compared with peroxide-prevul-
canization in that neither the prevulcanized latex nor films obtained from it
contain residual peroxidic vulcanizing agent. The problem of bad ageing
behaviour (see section 13.3.1 above) is not therefore encountered.
Two types of high-energy radiation have been used to induce prevulcanization
in rubber latices: high-energy electromagnetic waves, and high-energy electrons.
Brief consideration is now given to prevulcanization phenomena using each
of these types of radiation.
13.4.2 Prevulcanization by irradiation with high-energy electromagnetic waves

Natural rubber latex which has been prevulcanized by exposure to high-
energy electromagnetic waves is known as radiation-prevulcanized natural
rubber latex (RVNRL). The principal types of electromagnetic waves which
have been used to prevulcanize natural rubber latex are those which fall
within the class known as y-rays. These are waves having wavelengths in
vacuo less than ca. 3 x 10- 11 m, and frequencies greater than ca. 1 x 10 19 Hz.
The magnitude of the quantum associated with such waves is greater than
ca. 6 x 10- 15 J, i.e., 40keV. The type ofy-ray most commonly used has been
that emitted during decay of atoms of the cobalt-60 isotope. The principal
components of this radiation are two electromagnetic waves of frequencies
such that the associated quanta are 1.173 and 1.332 MeV respectively.
Electrons having energy 0.32 MeV are also emitted when cobalt-60 atoms
decay. Electromagnetic waves of lower energy falling within the class known
as X-rays have also been used. These waves have wavelengths in vacuo in
the range ca. I x 10 - 8 - 3 x 10 - 11 m, frequencies in the range ca. 3 x 10 16 -
1 X 10 19 Hz, and associated quanta in the range ca. 2 x 10- 17 _6 X 10- 15 J,
i.e., ca. 0.1 -40 keY.
The extent to which a material has been irradiated by such waves is
generally known as the dose. It can be quantified in two different ways: as
the amount of radiation to which the material has been exposed, or as the
amount of radiation which has been absorbed by the material. Radiation
doses expressed in these two ways are conveniently referred to as
radiation-exposure dose and radiation-absorption dose respectively. The extent
of exposure is quantified by the amount of electric charge which the given
amount of radiation would produce by the ionization of molecules in air.
The modern unit is the C (kg air) - 1. The older unit, which is encountered
in some of the older publications dealing with the irradiation of polymers
and latices, is the rontgen (R or r); it is equal to 0.258 mC (kg air) - 1. The
extent of adsorption is quantified as the energy which has been imparted to
unit mass of the material. The modern unit is the J (i.e., 6.242 x 10 18 eV)
(kg material) - 1, known as the gray (Gy). The older unit, which is again
encountered in some of the older publications, is the rad; it is equal to 10mGy.
These interconversions should be borne in mind when endeavouring to
compare experimental results in the older literature with more recent results.
In what follows, results from the older literature have been converted to the
modern units.
Early patents claiming processes for the prevulcanization of natural rubber
latex by exposure to high-energy electromagnetic waves include those to
Kemp [30] and to Pinner [31]. In an example given in the Kemp patent,
60% m/m (ammonia-preserved?) centrifuged natural rubber latex contained
in sealed glass tubes was exposed to y-rays from a cobalt-60 source contained
within a heavy concrete shield. Tubes were withdrawn after various times of
exposure such that the radiation absorption doses were respectively 8, 80
and 400 kGy. The tensile strengths of films from these latices, and also from
the initial latex, before and after hot-air ageing for 14 days at 70'C, are shown
in Table 13.9. Not only does the film tensile strength increase sharply with
increasing radiation dose; so too does the fraction of the tensile strength
which is retained after ageing in hot air. The patent to Pinner discloses that
the extent of prevulcanization by exposure to high-energy radiation can be
increased if a small amount of a polyfunctional olefinic monomer, in
particular, divinylbenzene, is added to the latex prior to irradiation.
The results of an extensive investigation of the effects of irradiation with
cobalt-60 y-rays upon ammonia-preserved natural rubber latex have been
reported in two papers by Minoura and Asao [32,33]. The first of these
papers gives results for the irradiation of the latex in the absence of additives.
The principal conclusions to be drawn from these results are as follows:
Table 13.9 Tensile strengths of films from natural rubber latex prevulcanized by
exposure to cobalt·60 y-rays, before and after hot-air ageing for 14 days at 70·C
(Kemp [30])
Radiation- Tensile strength / MPa % retention of
absorption dose tensile strength
/kGy Before ageing After ageing after ageing
o 16.6 3.4 21
8 16.2 9.3 57·
80 25.8 18.3 71
400 25.1 21.l 84
* Reference gives 52%.
1. As the radiation-exposure dose is increased, the viscosity of the latex

decreases initially, passes through a minimum, and then increases. The
initial decrease may be a consequence of agglomeration of the latex
particles, and the subsequent increase to changes in the colloid stabilizers.
2. The concentration of crosslinks found in films from the latex after a given
radiation-exposure dose increases somewhat as the latex is diluted with
water, and is in all cases significantly greater than the crosslink concentration
found in dried film from non-irradiated latex which is subsequently
subjected to the same radiation-exposure dose. This indicates that the
crosslin king reaction is facilitated by the water in the latex. The relationship
between crosslink concentration and rubber content in the latex is
approximately linear.
3. The latex pH and the pH at the point of coagulation (determined by
titration of diluted latex with 0.1 mol dm - 3 aqueous hydrochloric acid)
are essentially unaffected by irradiation.
4. As the radiation-exposure dose is increased, the KOH number of the latex
increases slightly initially, passes through a maximum, and then decreases.
These changes are probably a consequence of the effect of the radiation
upon the fatty-acid anions in the latex.
5. The mechanical stability of the latex increases progressively and significantly
with increasing radiation-exposure dose.
6. The concentration of crosslinks found in films from the latex increases
progressively and significantly with increasing radiation-exposure dose,
as does the modulus of the film. The relationship between crosslink
concentration and radiation-exposure dose is approximately linear. There
is no evidence of overall vulcanization reversion.
7. The tensile strength of films from the latex passes through a maximum
as the radiation-exposure dose increases. The maximum film tensile
strength attainable appears to be ca. 18 M Pa. The tensile strength of films
from latex which has received a large radiation-exposure dose is effectively
that of film from the non-irradiated latex.
8. The extension at break of films from the latex passes through a maximum
as the radiation-exposure dose increases, the maximum being ca. 1000%.
The extension at break of films from latex which has received a large
radiation-exposure dose is considerably less than that of film from the
non-irradiated latex.
9. No oxidation of the rubber, or other reactions, appears to occur,
notwithstanding that irradiation takes place in the presence of air. In this
respect, irradiation of natural rubber in latex form appears to contrast
with that of dry natural rubber. The chemical evidence in respect of natural
rubber irradiated in latex form does not exclude the possibility that some
chain scission occurs. The low tensile strength and high modulus of film
from latex which has received large radiation-exposure doses indicates
that crosslinking and chain scission have occurred concurrently.
Typical results reported by Minoura and Asao for the properties of films
obtained from y-irradiated natural rubber latex are shown in Table 13.10.
In a second paper [33], Minoura and Asao have reported results for the
effect of various organic chlorine compounds at levels in the range 1-5 pphr
upon the prevulcanization of natural rubber latex by exposure to cobalt-60
y-radiation. The compounds used were carbon tetrachloride, chloroform,
1,2-dichloroethane. After addition ofthe compounds to the latex, the mixtures
were stirred rapidly in a sealed vessel for 1 hour at normal ambient
temperature; it was assumed that most of the compound was absorbed into
the rubber particles after this treatment. Any enhancement of radiation-
crosslinking is therefore to be interpreted as being a consequence of additional
interactions between the radiation and the contents of the rubber particles.
Table 13.10 Properties of films obtained from ammonia-preserved natural rubber

latex prevulcanized by exposure to cobalt-60 )I-rays (Minoura and Asao [32])
Radiation-exposure dose rate: ca. 26C (kg of air) - 1 h - I.
Radiation- Concentration
exposure Tensile Extension Volume of
dose M300· M500t strength at break swelling crosslinks
/ KC (kgair)-I / MPa / MPa / MPa /% ratiot / moicm- J
0 0.36 0.52 1.8 770 56.6

1.3 0.63 1.03 7.6 880
2.6 0.77 1.17 930
3.1 14.3 1000 11.5 1.35 x 10- 5
12.9 0.83 1.34 11.3 850 6.7 3.29 x 10- 5
19.4 1.53 2.66 6.0 670 5.5 4.58 x 10- 5
25.8 1.53 1.7 330 4.8 5.75 x 10- 5
* Modulus at 300"10 extension.

t Modulus at 500"10 extension.
t Immersed in benzene at 25°C for 40 hours.
2.0
o~--~~~~~~~--~~~~~~~--~~~~~~
10- 2 10-' 100 10'
Radiation-exposure dose / kC (kg airr'
Figure 13.12 Results of Minoura and Asao [33] for effect of level of carbon tetrachloride upon
relationship between radiation-exposure dose (cobalt-60 )I-rays) and modulus at 300"1. extension
of films derived from radiation-prevulcanized natural rubber latex. Significance of points:
• t pphr carbon tetrachloride; '" 3 pphr carbon tetrachloride; _ 5 pphr carbon tetrachloride.
The most likely possibility is the formation of free radicals from the organic
chlorine compound by elimination of chlorine atoms from molecules which
become excited by the absorption of radiation. It is expected that free radicals
will then effect additional crosslinking of the rubber macromolecules by the
hydrogen-abstraction mechanism. All three organic chlorine compounds were
found to enhance the concentration of crosslinks formed after a given
radiation-exposure dose. Of the compounds used, carbon tetrachloride was
found to be the most effective. Figure 13.12 shows the results of Minoura
and Asao for the effect of level of carbon tetrachloride upon the relationship
between radiation-exposure dose and the modulus at 300% extension of films
derived from the irradiated latex. The effects of radiation-exposure dose upon
the tensile strength of films from the irradiated latices which contained the
organic chlorine compounds were found to be broadly similar to those for
films from the control latex which did not contain a compound. Again, tensile
strength passed through a maximum as the radiation-exposure dose increased,
but the radiation-exposure dose at which the maximum tensile strength was
achieved decreased as the level of the organic chlorine compound increased.
The tensile strengths of films from latices which had received large radiation-
exposure doses were in all cases similar to that of film from the non-irradiated
latex. Chlorine atoms were not introduced into the rubber in sufficient
amounts to be detectable by infra-red analysis, but the Beilstein test indicated
the presence of traces of combined chlorine which could not be eliminated
by vacuum drying. Minoura and Asao also report the effect of added benzene
upon the pre vulcanization of natural rubber latex by cobalt-60 y-radiation

to be negligible.
Kartowardoyo and Sundardi [34] have confirmed that the addition of
organic chlorine compounds, such as carbon tetrachloride and chloroform,
to ammonia-preserved natural rubber latex in amounts up to 5 pphr reduces
the radiation-absorption dose required to achieve optimal crosslinking during
prevulcanization from ca. 150kGy to ca. 10kGy. A low-ammonia latex
concentrate was used for their experiments. The organic chlorine compounds
were added as aqueous emulsions.
Contemporaneously with Minoura and Asao [32,33], Gregson et al. [35]
also reported results for the effect of irradiation with cobalt-60 y-rays upon
ammonia-preserved natural rubber latex and the polymer which it contains.
Their results for the effect of radiation-absorption dose upon various latex
properties, and upon various properties of films dried down from the latex,
are shown in Table 13.11. Figure 13.13 shows results reported by them for
the effect of radiation-absorption dose upon the tensile stress-strain curve
for films from the latex. In contrast to Minoura and Asao, Gregson et al.
observed the KOH number to increase progressively with radiation-absorption
dose. This is attributed to radiation-induced decomposition of various
non-rubber substances in the latex. Gregson el al. also observed that, as
Table 13.11 Effect of radiation-absorption dose upon various latex properties, and
upon various properties of films dried down from the latex, for ammonia-preserved
natural rubber latex (ca. 0.6% mjm ammonia on latex aqueous phase) prevulcanized
by exposure to cobalt-60 y-rays (Gregson et al. [35])
Radiation dose j kGy
Property 0 10 20 50 200
Properties of latices
pH 9.90 9.88 9.92 9.94 9.82
KOH number 0.567 0.624 0.667 0.670 0.739
VFA number 0.0550 0.0567 0.0560 0.0535 0.0545
Surface free energy j mJ m - 2 35.1 35.6 35.7 36.4 37.9
Brookfield viscosity· j mPa s:
at 60% mjm total solids content 60 60 52 56 67
at 50% mjm total solids content 14 14 14 14 15
Colour White White Slight Light Light
off-white tan brown
Properties of films dried down from latices

Gel content j % mjm 6.8 66.9 69.5 91.7 100.0
Water extract j % mjm 0.5 0.7 0.6 1.1 1.7
Acetone extract j % mjm 3.3 3.2 3.2 3.2 3.0
Equilibrium swelling in benzene
j mjm on mass of film 388 45 38 15 10
• Rotational frequency 30 minute - I; spindle not specified.

16
12
8!.
--
~
t J)
tJ)
l!? 8
'lii
~
'iii
c
{E
4
A
00 200 400 600 1200

Extension 1%
Figure 13.13 Results of Gregson et al. [35] for effect of radiation-prevulcanization by exposure
to cobalt-60 y-rays upon tensile stress-strain curve for films from natural rubber latex. Curves
A, B, C, D and E show the tensile stress-strain curves for latices which had received radiation-
absorption of respectively 0, 10, 20, 50 and 200 kGy.
expected, the tendency of the latex to putrefy during storage was reduced by
exposure to y-radiation, indicating partial sterilization.
The effect of various olefinic monomers as promoters of radiation-induced
latex prevulcanization is the subject of various papers by Makuuchi and his
co-workers [36-38]. One reason for investigating this possibility is that,
although organic chlorine compounds such as carbon tetrachloride are
effective as radiation-crosslinking promoters, most of the compound added
remains unchanged in the prevulcanized latex, and is subsequently released
into the environment during the drying of films from the latex. It is therefore
desirable to use low-volatility promoters which preferably have low inherent
toxicity also. It is to be expected that the effectiveness of olefinic monomers
will depend partly upon their ability to become absorbed in the rubber
particles when added to the latex, and partly upon their inherent efficiency
as promoters of y-radiation-induced crosslinking. It is also important that
the presence of the monomer during irradiation should not adversely affect
the colloid stability of the latex. Makuuchi and Hagiwara [36] have reported
on the performance of various polyfunctional olefinic monomers as promoters
for the y-radiation-prevulcanization of ammonia-preserved natural rubber
latex. Wide variations in effectiveness were found. Of those investigated,
neopentylglycol dimethacrylate was found to be most effective. Advantages
as compared with carbon tetrachloride are said to be the high colloid stability
of the irradiated latex and the high thermal stability of films derived from
the irradiated latex. The tensile strength of films from the latex could be
increased significantly by the addition of 10 pphr of a C 6 - 8 aliphatic solvent

(ligroin) after irradiation. A further paper by Makuuchi, Hagiwara and
Serizawa [37] compares neopentylglycol dimethacrylate and neopentylglycol
diacrylate as promoters of latex prevulcanization by y-radiation. Although
the diacrylate is more hydrophilic than the dimethacrylate, it is also the more
reactive of the two. It also reduces the colloid stability of the irradiated latex.
This is attributed to a tendency for the monomer to concentrate at the particle
surface. The colloid stability of the prevulcanized latex can be increased by
addition of a hydrocarbon solvent before irradiation, the effect being to
encourage the monomer to be adsorbed into the rubber particles. Addition
of 10 pphr of a C 6 8 aliphatic solvent (ligroin) was found to be particularly
effective. It appears that films having tensile strengths up to ca. 35 MPa could
be obtained from natural rubber latex which contained 3 pphr of neopentylglycol
diacrylate and 10 pphr of ligroin, and which had received a radiation dose
of ca. 30 kGy from a cobalt-60 source. The generally-accepted mechanism for
the promotion of radiation-crosslinking by polyfunctional olefinic monomers
is in outline as follows: Initially the monomer becomes graft-copolymerized
on to the rubber macrmolecules, thereby giving pendant groups which contain
unreacted olefinic double bonds. Some polymerization subsequently occurs
through these double bonds or through activated adjacent sites, resulting in
the formation of crosslinks between the primary rubber macromolecules.
However, in a further paper, Makuuchi and Tsushima [38] have shown that
the y-radiation-prevulcanization of natural rubber latex can also be promoted
by monofunctional acrylates and methacrylates. Ofthe various monofunctional
monomers investigated, n-butyl acrylate and 2-ethylhexyl acrylate were found
to be particularly effective. Films from latices containing 5 pphr of either of
these monomers, together with 20 pphr of ligroin, and subjected to a
radiation-absorption dose of 30 kGy from a cobalt-60 source, were reported
as having tensile strengths of ca. 30 M Pa. The latex containing 2-ethylhexyl
acrylate was more colloidally stable than that containing n-butyl acrylate.
The above outline mechanism for the promotion of radiation-crosslinking
by graft copolymerization is clearly inapplicable to monofunctional olefinic
monomers. Makuuchi and Tsushima suggest that the observed changes in
the mechanical properties of films from the irradiation of latices containing
the monofunctional monomers is primarily a consequence of physical
entanglement of polyacrylate macromolecules with the rubber molecules.
Evidence in support of their view that no significant graft copolymerization
of the acrylate occurs is provided by a gel-permeation chromatogram of an
irradiated mixture of a liquid polyisoprene (a model for natural rubber) and
2-ethylhexyl acrylate.
Makuuchi et al. [39] have recently described a process for the production
of y-radiation-prevulcanized natural rubber latex which is claimed to be
suitable for the manufacture of gloves by a conventional dipping process
using a calcium salt as coagulant. High-ammonia centrifuged natural rubber
latex concentrate was used without further purification. It was diluted to ca.
53% mlm total solids content with 1% mlv aqueous ammonia solution.
Radiation-crosslinking was facilitated by adding 5 pphr of 2-ethylhexyl
acrylate and 1 pphr of carbon tetrachloride to the latex prior to irradiation.
After these additions, the latex was allowed to stand for 16 hours before
being irradiated. The radiation-absorption dose was 12 kGy from a cobalt-60
source. Dose rates ranged from 0.57 to 1.22 kGy h - 1. During irradiation, the
latex was contained within a stainless-steel drum which was continuously
rotated on rollers. After irradiation, 0.5 pphr of 2,2-methylene bis(4-
ethyl-6-tert-butylphenol) was added to the latex as an antioxidant. The
mechanical properties of gloves produced from this latex were considered to
be satisfactory, although in some respects perhaps not as good as some which
are produced from sulphur-prevulcanized natural rubber latex. Tensile
strength and extension at break before ageing are given as 37.8 MPa and
897% respectively, and the tensile strength after hot-air ageing for 48 hours
at 70·C as 37.0 MPa. The acid-resistance of gloves produced from the radiation-
prevulcanized latex is said to be inferior to that of gloves produced from
sulphur-prevulcanized latex, but it is suggested that the acid-resistance could
be improved by using a non-calcium-based coagulant in the dipping process
for the production of the gloves.
Other papers concerned with the prevulcanization of natural rubber latex
by irradiation with high-energy electromagnetic waves include those by
Lamm and Lamm [40], by Puig [41], and by Ridwan et at. [42]. The
mechanism of the radiation-prevulcanization reaction has been discussed by
Todorov [43].
13.4.3 Prevulcanization by irradiation with high-energy electrons

Although it is possible to prevulcanize natural rubber latex by irradiation
with high-energy electrons, there appears to have been far less interest in
such processes than in prevulcanization by irradiation with high-energy
electromagnetic waves. The source of the electrons is usually an electron-
accelerating apparatus. The extent of exposure to the high-energy electrons
is quantified either as exposure dose or as absorption dose, in the same way,
and using the same units, as for high-energy electromagnetic radiation. The
effect of the incident electrons upon the latex is to generate within either the
aqueous phase, or the rubber particles, reactive species which eventually
cause covalent crosslinks to form between the rubber molecules.
A typical process has been described by Kemp [30]. 60% mlm ammonia-
preserved centrifuged natural rubber latex concentrate is placed in a shallow
aluminium trough, covered with a thin sheet of aluminium, and exposed for
various periods to a beam of high-energy electrons from a van de Graaf
apparatus. The effects of increasing absorption dose upon the mechanical
properties of films dried down from the resultant latex are shown in
Table 13.12 Tensile stress-strain properties of films from natural rubber latex
prevulcanized by exposure to high-energy electrons (Kemp [30])
Modulus at extension indicated

Absorption / MPa Tensile Extension at
dose strength break
/kGy 300% 500% 700% /MPa /%
0 0.58 1.35 4.42 10.0 966
100 0.76 1.76 5.85 17.4 1008
200 1.15 2.56 7.80 19.0 961
300 1.27 2.73 9.27 18.5 892
370 1.39 3.04 11.44 14.9 812
450 1.41 3.50 14.0 757
Table 13.12. These results provide clear evidence that crosslin king of the
rubber has occurred. As in the case of prevulcanization by irradiation with
high-energy electromagnetic waves, film tensile strength passes through a
maximum with increasing electron dose, and the tensile strength of the film
obtained from latices which have been exposed to very high doses is similar
to that offilm from the control latex which has not been exposed to electrons.
13.5 PREVULCANIZATION OF OTHER RUBBER LATICES
13.5.1 Styrene-butadiene rubber and acrylonitrile-butadiene rubber synthetic

latices
An investigation into the possibility of prevulcanizing styrene-butadiene
rubber and acrylonitrile-butadiene rubber synthetic latices has been reported
by Lebedev et al. [44]. They found that prevulcanization occurs if these
latices are heated with sulphur, dithiocarbamate accelerators and zinc oxide.
As expected, the tensile strengths ofthe dried films obtained from the resultant
latices were markedly inferior to those of films obtained from sulphur-
prevulcanized natural rubber latices. Of the accelerators which were investigated,
the most satisfactory were found to be dimethylammonium dimethyl-
dithiocarbamate and zinc ethylphenyldithiocarbamate. The published
formulations indicate that the levels of sulphur used were high relative to
those conventionally used with natural rubber latex; ca. 6 pphr was used in
one instance. These workers have also reported that these synthetic rubber
latices can be prevulcanized by exposure to cobalt-60 y-rays.
Lebedev et al. [44J have also provided evidence for the essential correctness
of the secondary-valence-bond theory for the strength and coherence of the
dry deposits which are formed from prevulcanized rubber latices (see section
13.2.9.4 above). They compared the mechanical properties of films obtained
from styrene-butadiene rubber latices and from acrylonitrile-butadiene
rubber latices which had been prevulcanized by the conventional sulphur
Prevulcanization of other rubber latices 509
method on the one hand, and by exposure to y-rays on the other. They found
that the tensile strengths of films from the irradiated styrene-butadiene
rubber latex were considerably less than those of films from the sulphur-
prevulcanized latex, whereas there was much less difference between the
strengths of the films from the radiation- and sulphur-prevulcanized
acrylonitrile-butadiene rubber latices. From these observations, it was
inferred that the polarity of the polymer is an important factor in determining
the tensile strength of a film from a prevulcanized rubber latex, and that
therefore the strength-determining bonds are predominantly of the secondary-
valence type. Furthermore, it appears that the process of sulphur-vulcanization
significantly enhances the polarity of the polymer dispersed within a
prevulcanized latex. It may be that this is an important factor in determining
the tensile strength of films from sulphur-prevulcanized rubber latices.
13.5.2 Isobutene-isoprene (butyl) rubber artificial latices

According to Esso Research [45], it is possible to prevulcanize isobutene-
isoprene (butyl) rubber artificial latices by heating with sulphur and vulcanization
accelerators. This observation is perhaps surprising, in view of the slow rate
at which bulk isobutene-isoprene rubber vulcanizes when heated with
sulphur and vulcanization accelerators, even at elevated temperatures, and
also in view of the further retardation of the reaction by small concentrations
of other substances, notably of olefinically-unsaturated compounds. Both
these latter phenomena are a consequence of the low concentration of olefinic
unsaturation in this rubber. In order satisfactorily to sulphur-prevulcanize
isobutene-isoprene rubber latex, it is necessary to employ fairly high levels
of organic vulcanization accelerator. The formulation shown in Table 13.13
has been recommended.
Table 13.13 Formulation for sulphur-prevulcanization ofisobutene-isoprene (butyl)

rubber artificial latices (Esso Research [45])
Parts by mass
Ingredient Dry Actual
Isobutene-isoprene (butyl) rubber

(as 55% mlm aqueous latex) 100 182
Sodium caseinate (as 10% m/m aqueous solution) 1 10
Sulphur (as 50% m/m aqueous dispersion) 2 4
Zinc di-n-butyldithiocarbamate
(as 50% m/m aqueous dispersion) 3 6
Zinc mercaptobenzthiazolate
(as 50% m/m aqueous dispersion) 0.5 1
Zinc oxide (as 50% m/m aqueous dispersion) 3 6
Silicone antifoam 0.5 0.5
REFERENCES
1. Schidrowitz, P. and Goldsbrough, H. A., British Patent No.1, 111, application
date 15 January, 1914, acceptance date 15 April, 1915.
2. Schidrowitz, P., British Patent No. 193,451, 26 February, 1923.
3. Schidrowitz, P., British Patent No. 208,235, 14 December, 1923.
4. Deal, K. M. and Waterman, R. R. (1962) Rubber Age, New York, 92, 93.
5. Panich, R. M., Fodiman, N. M. and Voyutski, S. S. (1959) Soviet Rubber
Technology, 18(2), 14.
6. Murray, P. J. (1982) NR Technology, 13, 31.
7. Fletcher, W. P., Gee, G. and Morell, S. H. (1952) Transactions of the Institution
of the Rubber Industry, 28, 85.
8. Gorton, A. D. T. and Pendle, T. D. (1976) N R Technology, 7, 77.
9. Gorton, A. D. T. (1979) NR Technology, 10, 9.
10. Merrill, R. W. T. (1980) Prevulcanization of Natural Rubber Latex, M. Phil. thesis,
United Kingdom Council for National Academic Awards.
11. Loh, A. c.-P. (1982) Further Investigations of the Prevulcanisation of Natural
Rubber Latex, Ph. D. thesis, United Kingdom Council for National Academic Awards.
12. Gorton, A. D. T. and Pendle, T. D. (1981) N R Technology, 12, 21.
13. Siswantoro, O. (1985) Preparation and Destabilisation of Model Polyisoprene
Latices, Ph. D. thesis, United Kingdom Council for National Academic Awards.
14. Ghazaly, Hafsah bte. Mohd. (1988) Prevulcanisation of NR Latex using Purified
Reactants, M. Phil. thesis, United Kingdom Council for National Academic Awards.
15. van Gils, G. E. (1977) Rubber Chemistry and Technology, SO, 141.
16. Porter, M., Rawi, R.and Rahim,S. A. (1992) Journal of Natural Rubber Research, 7,85.
17. Hauser, E. A., Ie Beau, D. S. and Kao, 1. Y. L. (1942) Journal of Physical Chemistry,
46,1099.
18. Humphreys, N. C. H. and Wake, W. C. (1950) Transactions of the Institution of
the Rubber Industry, 25, 334.
19. Sutton, S. D. (1951) Transactions of the Institution of the Rubber Industry, 27,193.
20. Gorton, A. D. T. (1988) Proceedings International Rubber Technology Conference,
Penang, Malaysia, 1988, p. 85.
21. Blackley, D. C. (1988) Proceedings International Rubber Technology Conference,
Penang, Malaysia, 1988, p. 3.
22. Revertex General Bulletin GBl, undated.
23. Stevens, D. M./Dewey and Almy Chemical Company, British Patent No. 324,287,
23 January, 1930.
24. Stott, G./Revertex Ltd., British Patent No. 738,279, 12 October, 1955.
25. Stott, G./United States Patent No. 2,868,859, 13 January, 1959.
26. Hercules Powder Company, British Patent No. 819,669, 19 September, 1959.
27. Ropp, W. S.jHercules Powder Company, United States Patent No. 2,975,151,
14 March, 1961.
28. Rodaway, B. K. and Calvert, K. o./Dunlop Holdings Ltd., United States Patent
No. 3,755,232, 28 August, 1973.
29. Said, Ma'zam Md., Pend Ie, T. D. and Blackley, D. C. (1990) Journal of Natural
Rubber Research, 5, 27.
30. Kemp, I./Revertex Ltd., British Patent No. 816,230,8 July, 1959.
31. Pinner, S. H./TI(Group Services) Limited, British Patent No. 831, 197, 23 March, 1960.
32. Minoura, Y. and Asao, M. (1961) Journal of Applied Polymer Science, 5, 233.
33. Minoura, Y. and Asao, M. (1961) Journal of Applied Polymer Science, 5,401.
34. Kartowardoyo, S. and Sundardi, F. (1977) Journal of Applied Polymer Science,
21, 3017.
References 511
35. Gregson, T. c., Rogers, T. H., Bangs, L. B. and Peabody, D. W. (1961) Rubber
Age, New York, 89, 81.
36. Makuuchi, K. and Hagiwara, M. (1984) Journal of Applied Polymer Science, 29, 965.
37. Makuuchi, K., Hagiwara, M. and Serizawa, T. (1984) Radiation Physics and
Chemistry, 24, 203.
38. Makuuchi, K. and Tsushima, T. (1985) Proceedings International Rubber Coriference,
Kuala Lumpur, Malaysia, 1985, Volume II, p.502.
39. Makuuchi, K., Yoshi, F., Ishigaki, I., Thushima, K., Mogi, M. and Saito, T. (1988)
Proceedings International Rubber Technology Coriference, Penang, Malaysia, 1988,
p. 104.
40. Lamm, A. and Lamm, G. (1962) Proceedings ofthe Tihany Symposium on Radiation
Chemistry, Akademiai Kiad6, Budapest, 1962, p. 245.
41. Puig, J. R. (1971) Atomic Energy Review, 9, 373.
42. Ridwan, M., Sundardi, F., Utama, M. and Kartowardojo, S. (1979) Radiation
Physics and Chemistry, 14, 747.
43. Todorov, J. M. (1967) Proceedings of the Second Tihany Symposium on Radiation
Chemistry, Akademiai Kiad6, Budapest, 1967, p. 749.
44. Lebedev, A. V., Fermor, N. A., Mints, S. M. and Zakharchenko, P.1. (1958) Kauch
i. Rezina, 17(5), 3.
45. Esso Research (1963) Private communication.
14
Chemically-modified latices:
2. Types of chemically-modified latex
other than prevulcanized latices
14.1 INTRODUCTION
This chapter is concerned with chemically-modified latices other than
prevulcanized latices. Because of their industrial importance, the latter have
been considered separately in Chapter 13. The reader is referred to section
13.1 of Chapter 13 for a general introduction to all types of chemically-modified
latices.
As with prevuIcanized latices, most of the published information concerning
other types of chemically-modified latices relates to natural rubber latex.
This is not surprising. If it is desired to prepare a chemical variant of a
synthetic latex, then the most obvious route is by way of an appropriately-
modified emulsion-polymerization reaction system. Likewise, if it is desired
to prepare a chemical variant of an artificial latex, then the most obvious
method will commence with a bulk polymer which has the desired chemical
composition and structure. These options are not, of course, available for
natural latices. The chemical nature of the polymer dispersed in the latter is
determined entirely by the biochemical pathway by which the polymer is
synthesized in the plant. The only remaining option is chemical modification
after the polymer has been obtained from the plant. If the modified polymer
is itself required in latex form, then the most promising route is chemical
modification of the polymer in the latex form.
The types of chemical modification considered in this chapter are:
1. epoxidation of natural rubber in latex form;

2. graft-copolymerization reactions in latices;
3. cis-trans isomerization of natural rubber in latex form;
4. cycIization of natural rubber in latex form;

Epoxidation of natural rubber in latex form 513
5. halogenation of olefinically-unsaturated polymers in latex form;

6. hydro halogenation of olefinically-unsaturated polymers in latex form;
7. halogenoalkylation of natural rubber in latex form;
8. depolymerization and oxidation of natural rubber in latex form;
9. modifications involving only the surface of the latex particles.
For the reactions to be considered under the headings (1)-(8), the intention
is that the modification should take place more or less uniformly throughout
the latex particles. For the processes to be considered under heading (9), the
intention is that the modification is confined essentially to the surface regions
of the latex particle.
14.2 EPOXIDATION OF NATURAL RUBBER IN LATEX FORM

Ideally, the epoxidation of an olefinically-unsaturated polymer involves the
addition of one oxygen atom to a proportion of the carbon-carbon double
bonds, thereby converting them to oxirane (epoxide) rings:
In practice, various other reactions may also occur, depending upon the
reagent used to effect epoxidation. This matter is discussed further below.
The reaction of natural rubber with reagents which might be expected to
convert olefinic double bonds to oxirane rings has been known since at least
the early 1920s. Interest in modifying natural rubber in this way was dormant
for several years, but has been revived in recent years. The effect of epoxidation
is to increase the polarity of the polymer, and hence its glass-transition
temperature and its resistance to swelling in non-polar organic liquids. The
magnitude of these effects depends upon the extent to which the rubber has
been epoxidized, i.e., upon the proportion of olefinic double bonds which
have become converted to oxirane rings. Currently this is in the range
25-50% mol/mol. It has been found convenient to carry out the epoxidation
reaction upon natural rubber in latex form, and then to separate the modified
polymer for application as bulk rubber.
There is also considerable interest in developing an epoxidized natural
rubber latex which is suitable for application as such. The immediate product
of the epoxidation reaction is a rather dilute latex. Difficulty has been
encountered in concentrating this dilute latex to give a product which is
economical to transport and is suitable for use in latex-manufacturing
processes. Creaming, centrifugation and electrodecantation are not feasible,
because the density of the epoxidized rubber particles is very similar to that
of the aqueous phase in which they are dispersed. Evaporation is not feasible
because of the nature of the dispersion medium. It is understood that a

filtration method has recently been developed, and that it is promising.
A wide range of reagents has been used to effect partial epoxidation of
the olefinic double bonds in cis-l,4-polyisoprenes. For epoxidation of natural
rubber in latex form, mixtures of hydrogen peroxide and a water-soluble
carboxylic acid are the preferred type of reagent, on grounds of safety,
convenience and economy. It is believed that the actual epoxidizing reagent
is the corresponding peroxycarboxylic acid formed in situ by interaction
between the hydrogen peroxide and the carboxylic acid. Within this group
of reagents, the most cost-effective reagent has been found to be a mixture
of formic acid and hydrogen peroxide; consequently, most interest has been
shown in this combination. The effective epoxidizing reagent is thought to
be the peroxoformic acid, generated by reversible reaction between the
hydrogen peroxide and the formic acid as follows:
The oxygen which effects epoxidation of the polyisoprene is derived almost

exclusively from decomposition of the peroxoformic acid back to formic acid.
Thus formic acid effectively functions as a catalyst for epoxidation of the
polymer by hydrogen peroxide.
Important side-reactions which can occur when peroxycarboxylic acids
are used as epoxidizing reagents for olefinically-unsaturated polymers include
acid-catalysed opening of the oxirane ring to give eventually vic-diols, and
the formation of ether crosslinks between polyisoprene molecules by various
reactions. Overall, the reaction by which vic-diols are formed can be
represented as follows:
"'-C--C/ + H 20 --
"'-C--C/
/"'- /"'- /1 1"'-
o OH OH
If the acidity arises from the presence of a carboxylic acid, then the reaction
may occur via the formation, and subsequent hydrolysis of, a monoester of
the carboxylic acid and the eventual vic-diol:
'c-c/
. /1 I'
OH O.OC.R
Processes which can lead to crosslinking include acid-catalysed interaction

between oxirane rings and vic-diol groups by reactions such as:
'c-c/
/1 I'
.. o OH
'!-c/
+
/1 I'
OH OH
One consequence of such crosslinking reactions is, of course, that the gel
content of the rubber increases. There is evidence that many other side-
reactions can occur during the epoxidation of cis-I,4-polyisoprene in latex
form by peroxocarboxylic acids. A recent comprehensive review of these has
been given by Farley [I].
Typical results for the effects of progressive epoxidation at 60·C by
hydrogen peroxide and formic acid upon the properties of the polymer
contained in ammonia-preserved natural rubber latex are shown in Table
14.1. As well as introducing oxirane rings into the polymer, the initial effect
of epoxidation is to reduce the gel content ofthe dispersed polymer drastically,
Table 14.1 Typical results for the effects of progressive epoxidation by hydrogen
peroxide and formic acid at 60·C upon properties of polymer dispersed in ammonia-
preserved natural rubber latex (Farley [I])
Reaction system (parts by mass):
Natural rubber (as 30% m/m latex) 100
Non-ionogenic ethoxylate stabilizer 2
Formic acid (98%) 13.5-
Hydrogen peroxide 37t
- Corresponds to 20% mol/mol on the isoprene units.
t Corresponds to 74% mol/mol on the isoprene units.
Number-average
Gel molecular mass
Reaction content- Extent of epoxidationt of sol content
time of rubber of rubber of rubber
/ hour /% m/m / % mol/mol oxirane groups / dalton
0 81 0 2.10 x 10 5
1 13 5 1.74 X 10 5
2 19 11 2.06 x 10 5
3 17 16 1.92 x 10 5
20 17 51 1.28 x 10 5
22 33 53 1.57 x 10 5
24 45 56 1.19 x 10 5
* Detennined by extraction with tetrahydrofuran.

tDetennined by IH-NMR spectroscopy at 200 MHz.
and to reduce the number-average molecular mass of the sol fraction

somewhat. The reduction in gel content is presumably a consequence of
rupture of the covalent and/or other types of bonds which are responsible
for the presence of gel in the cis-l,4-polyisoprene dispersed in natural rubber
latex (see section 9.5.2.2.2 of Chapter 9). After prolonged reaction, the gel
content of the rubber increases again as a consequence of crosslinking
reactions of the type indicated above, but not necessarily to the level initially
present in the rubber. The gel content does not appear to correlate with the
extent of epoxidation, being a consequence of prolonged reaction time rather
than of high extent of epoxidation. The extent to which ring-opening and
crosslinking occur are dependent upon the concentration of formic acid in
the latex aqueous phase, because, as noted above, these reactions are
acid-catalysed. Thus Ng and Gan [2,3] have shown that epoxidation of
natural rubber in latex form by mixtures of hydrogen peroxide and formic
acid is accompanied by extensive ring-opening (and also by extensive particle
coagulation) ifthe concentration offormic acid is ca. 8 mol dm -3. Farley [1]
has reported results which demonstrate the effect of increasing the concentration
of formic acid (from 0.81 moldm- 3 to 3.22moldm- 3 ) and concomitantly
reducing the hydrogen peroxide concentration (from 3.16 mol dm - 3 to
1.58 mol dm - 3) upon epoxidation for 24 hours at 6O.C. The pH of the reaction
system was reduced from 3.8 to 2.4. The gel content of the resultant polymer
increased from 45% to 95%, and the extent of epoxidation decreased from
56% mol/mol to 38% mol/mol. The proportions of hydrogen peroxide
consumed in the two reactions were 79% and 100% respectively. The ratio
of oxirane groups incorporated to molecules of hydrogen peroxide consumed
increased from 0.92 to 1.00.
As stated above. epoxidation of natural rubber is accompanied by increase
in glass-transition temperature. The oxygen content of the polymer increases.
and consequently the latter becomes more polar. The oxirane rings, and
other oxygen-containing moieties, introduced in the course of the reaction
may also restrict the motion of the polymer segments by steric effects. The
relationship between glass-transition temperature and extent of epoxidation
is approximately linear. This is illustrated schematically in Fig. 14.1, which
is taken from a review of epoxidized natural rubber by Gelling [4]. The rate
of increase of glass-transition temperature is ca. 0.93·C per 1% mol/mol
increase in extent of epoxidation. Epoxidation to an extent of 50% mol/mol
increases the glass-transition temperature from ca. - 70·C to ca. - 20·C.
A convenient summary of the chemistry of the epoxidation of natural
rubber by peroxocarboxylic acids has been given by Gelling and Porter [5].
The reaction is stereospecific, in that the configuration ofthe four substituents
in the planar oxirane ring is the same as that of the four substituents about
the original olefinic double bond, i.e., cis-olefins epoxidize to cis-oxiranes,
trans-olefins to trans-oxiranes. The epoxidation reaction is kinetically of first
order with respect to the concentrations of both peroxoacid and olefinic
V 40
.......
l!!
e~
::::I
0
~
§
;:;
.~ -40
I!!
ia-80~----~----~~----~----~~--~
o 20 40 60 80 100
Extent of epoxidation / %mol/mol
Figure 14.1 Illustrating relationship between glass-transition temperature of epoxidized natural

rubber and extent of epoxidation (Gelling [4]).
double bonds; thus overall the reaction is kinetically of second order. The
entropy of activation is large and negative, indicating that the transition state
is highly ordered. The generally-accepted mechanism is that proposed by
Bartlett [6]. This involves electrophilic attack of the peroxocarboxylic acid
on the olefinic double bond, with the formation of a bicyclic transition state
,-
which then dissociates to an oxirane derivative and the carboxylic acid:
V
-
0 R \ /
\C/
# .0.~
o/ \C/ Ii ·-··0'''·
c. R
IIC + : ....",-..."".",.". 1 ' 0 + R.C0 2 H
c/
I 0II C"" "
/\ H / \ ~------~ /\
Campbell [7] has provided a summary of an investigation into the
consequences of phase heterogeneity for the epoxidation of natural rubber
in latex form by formic acid and hydrogen peroxide. It is generally agreed
that the rate-controlling step for the epoxidation reaction is the formation
of peroxoformic acid from these two reagents. In the outline mechanism for
the epoxidation by this mixture of an olefin dispersed in an aqueous phase
proposed by Campbell, it is assumed that the hydrogen peroxide and formic
acid partition almost exclusively in the aqueous phase of the latex, so that
it is in that phase that the peroxoformic acid is generated. The latter then
partitions reversibly between the aqueous phase and the rubber particles.
Diffusion of the peroxoformic acid within the rubber particle, and reaction
with the olefinic double bonds of the rubber macromolecules, occur rapidly
relative to the formation of peroxoformic acid in the aqueous phase. Thus
neither diffusion nor reaction with the olefinic double bonds significantly
affects the overall rate of epoxidation of the rubber. Being very hydrophilic
in nature, the formic acid which is formed during the epoxidation reaction
diffuses from the rubber particles back into the aqueous phase. Campbell
also concludes that a critically important feature of the reaction system which
has contributed to the success of the formic acid-hydrogen peroxide
combination as epoxidizing reagent is that the epoxidized product is retained
within a hydrophobic organic phase in which access to water is severely
restricted. The importance of this feature is demonstrated by the observation
that the epoxidation of I-methylcyclo-I-hexene in aqueous emulsion using
hydrogen peroxide and formic acid as the reagent leads principally to the
formation of the hydrolysis product l-methylcyclo-I,2-hexanediol even from
very early stages of the reaction. The most likely reason for this is that,
although the epoxidation reaction occurs within the emulsified olefin droplets,
the molecular mass of the epoxidized product is sufficiently low for the
molecules of this product to move to the surface of the droplets and there
undergo hydrolysis. Another possibility is that the molecules ofthe epoxidation
product diffuse out of the emulsion droplets into the aqueous phase and
undergo hydrolysis in that phase.
Other publications relevant to the epoxidation of natural rubber in latex
form include a patent application by Gelling [8], and papers by Burfield,
Lim and Law [9], and by Perera, Elix and Bradbury [10]. The review by
Gelling [4], cited above, provides a useful summary of the properties of the
polymer contained in epoxidized natural rubber latex.
Bulk epoxidized natural rubber can be vulcanized by heating with sulphur
and conventional organic vulcanization accelerators, by heating with peroxides,
and by heating with dibasic acids. The latter reagents effect crosslinking
through di-ester formation by reaction between the acid and neighbouring
pairs of oxirane rings. There seems little doubt that it is possible to
prevulcanize epoxidized natural rubber latex by heating with sulphur and
organic vulcanization accelerators. However, as far as the writer is aware,
this possibility has not so far been seriously investigated.
14.3 GRAFT-COPOLYMERIZATION REACTIONS IN LATICES
14.3.1 Outline of graft-copolymerization reactions

The terms block copolymer and graft copolymer have been introduced in
section 2.2.3.1 of Chapter 2 (Volume I). A graft copolymer is a specialized
type of block copolymer in which blocks of one monomer unit are covalently
bonded to a main-chain (backbone) polymer comprising exclusively units
derived from the other monomer. The monomer units constituting the
attached blocks are said to be grafted on to the main-chain polymer. Many
methods are available for the preparation of graft copolymers. In one of
Graft-copolymerization reactions in latices 519
these, an olefinic monomer is polymerized by free-radical mechanism within

the latex of a polymer derived from another monomer. If the reaction
conditions are favourable, much of the added monomer graft-copolymerizes
on to the polymer molecules of the latex particles. The particles then contain
a graft copolymer which comprises the original polymer of the latex as the
main-chain component and blocks derived from the added monomer as grafts.
One obvious prerequisite for graft-copolymerization to occur in a reaction
system of this type is that the polymerization of the added monomer takes
place in the immediate presence of the polymer molecules. This implies that
most of the added monomer must be absorbed into the latex particles, and
most of the polymerization must take place within those particles. The
mechanism by which graft-copolymerization occurs when a monomer is
polymerized by free-radical mechanism in the presence of another polymer
is essentially independent of whether the reaction is carried out with the
polymer in homogenous solution or in the form of a latex. However, some
ofthe details of the reaction may differ between the two types of reaction system.
If the main-chain polymer does not contain residual olefinic unsaturation,
then the most likely reaction by which the polymer grafts become covalently
bonded to the main-chain polymer is that of radical-transfer to polymer, to
which brief mention is made in sections to.2.1.2. t and to.2.2.4 of Chapter
to, and section t 1.2.2 of Chapter t 1. Let the backbone polymer be denoted
by PX, where X is a relatively labile moiety which is abstractable from the
polymer by interaction with a free radical, thereby forming a polymeric
radical p .. Let the added monomer be denoted by M, and a radical species
in the system by R·. Then graft-copolymerization by the reaction of
transfer-to-polymer can be represented generally as comprising the following
steps:
R· + PX ---+ RX + p.
+M +M
p. + M ---+ PM l · - - - + ... PM n· ---+
The result is the formation of the propagating block Mn· attached to the
inactive polymer P. The species R· may be a free radical of low molecular
mass such as is derived directly or indirectly from an initiator, or it may be
a macromolecular radical of the type Mn· propagating as the normal
free-radical polymerization of the monomer M proceeds. In the latter case,
the non-radical species RX is an inactive homopolymer MnX. To the extent
that it is desired to graft all the added monomer to the main-chain polymer,
added monomer which becomes polymerized as homopolymer is regarded
as having been wasted. For hydrocarbon polymers, the labile moiety X is
invariably a hydrogen atom. If the polymer contains halogen atoms covalently
bonded to carbon atoms, then these too are susceptible to abstraction by
free radicals. A labile atom can also be abstracted from the polymer if the
latter is irradiated by high-energy radiation, such as y-rays; this in principle

provides an alternative method of initiating graft-copolymerization.
If the main-chain polymer contains olefinic double bonds, then graft-
copolymerization can also in principle occur by the copolymerization of the
monomer M with the main-chain polymer. This reaction can be represented
generally as follows:
~=~+R·--+ ~-~
I •
R
1 +M
-----c-~
I I
R Mn·
In this reaction scheme, M again denotes the added monomer, and R· a

radical species in the system. Again, R· may be a free radical oflow molecular
mass, or it may be a macromolecular radical of the type Mn·. If the former,
then only one graft is added at each olefinic site copolymerized; if the latter,
then two grafts are added. Again, the added monomer is wasted insofar as
concurrent homo polymerization occurs. The presence of olefinic unsaturation
in the main-chain polymer may also facilitate transfer-to-polymer reactions
at an adjacent site. In particular, radical abstraction of allylic (IX-methylenic)
hydrogen atoms may be facilitated, as has been noted in section 11.2.2 of
Chapter 11.
The various radical species present in graft-copolymerization reaction
systems of the types described in the preceding paragraphs can interact with
all the components of the system in the ways which are usual for free-radical
addition polymerization. In particular, a radical can interact with another
radical to produce an inactive non-radical species. Termination of polymerization
usually occurs by way of bimolecular reactions of this type, the outcome
being either combination or disproportionation of the radicals. Bearing in
mind the number of different types of radical which may be present in
graft-copolymerization reaction systems of the types envisaged above, it is
clear that the end-product of the reaction can be a complex mixture of
polymeric species. This is particularly so if termination of propagating chains
occurs by radical combination.
In principle, free-radical graft-copolymerization to a polymer in latex form
can be initiated by many types of substance, and also by high-energy
irradiation. As has been noted, the propagating polymer chains must be
located in the monomer-swollen polymer particles if graft-copolymerization
is to occur, but this does not imply that the initiating free radicals should
necessarily be generated within the particles. As with the typical free-radical
emulsion polymerization (see section 10.2.2 of Chapter 10), the free radicals
may be generated within the aqueous phase, and then subsequently enter
the swollen latex particles by diffusion or collision, either in the form as
generated, or after reaction with some other species such as water or monomer
dissolved in the aqueous phase. As with free-radical emulsion polymerization
reactions, chemical initiators fall into two broad classes (see section 10.2.3.1
of Chapter 10), namely, dissociative initiators and redox initiators.
Representatives of both classes of initiator are used for effecting graft-
copolymerizations in latices.
Several factors influence the nature of the polymer particles present in a
latex in which graft-copolymerization has been effected by the addition of
an olefinic monomer and the establishment of free-radical polymerization
conditions. These include the relative rates at which the various reactions
occur, and the miscibility of the various types of polymer block which
comprise the final polymer. Full characterization ofthe product of a latex graft-
copolymerization reaction of the types envisaged above requires a knowledge
of several properties of the product, such as:
1. the proportion of the polymerized added monomer which has become

grafted to the main-chain polymer, rather than having become homo-
polymerized - this proportion is usually known as the grafting efficiency;
2. the proportion of the main-chain polymer initially present in the reaction
system which has acquired grafted blocks derived from the added monomer;
3. the average composition of the graft copolymer;
4. the average molecular mass of the blocks of the graft copolymer;
5. the distribution of molecular masses for the graft copolymer blocks;
6. the spatial distribution of the graft copolymer blocks in relation to the
main-chain polymer;
7. any changes, e.g. crosslinking, which have occurred within the main-chain
polymer, other than graft-copolymerization;
8. the manner in which the graft copolymer and any homopolymers are
distributed within the particles of the latex obtained at the end of the reaction.
Usually, the amount of such information available for the products of

particular reaction systems is very limited. Separation of graft copolymer
from homopolymer and the initial main-chain polymer has often been effected
by using the differing solubility characteristics ofthe three potential components
of the mixture. The polymer mixture may be dissolved in a common solvent
for the components, and then subjected to fractional precipitation by addition
of increasing amounts of a miscible non-solvent for the polymers. Alternatively,
the polymer mixture may be subjected to fractional extraction using one or
more solvents having different characteristics. By such means, it is often
possible to demonstrate that a copolymer of the added monomer and the
initial polymer has formed, as well as to gain some indication of the relative
proportions of the three possible polymeric components in the mixture.
However, methods for separating the components in the product of a

graft-copolymerization reaction which depend upon differences in solubility
behaviour can give misleading results, principally because the graft copolymer
is sometimes able to solubilize one or other of the homopolymers. Some
information concerning the average molecular mass of the grafted blocks of
the graft copolymer can be obtained from the difference between the average
molecular mass of the graft copolymer and that of the initial polymer, if the
assumption is made that the molecular mass of the main-chain polymer is
essentially unaffected by the graft-copolymerization reaction. The average
molecular mass of these blocks, and also the distribution of block lengths,
can in principle be determined if the main-chain polymer can be degraded
without influencing the grafted blocks. For graft copolymers derived from
cis-l,4-polyisoprenes, this has been achieved by carefully degrading the
polyisoprene component of the copolymer by reaction with ozone. The
morphology of the particles formed by graft-copolymerizations in latices is
important, because this morphology can influence the nature of the solid
phase obtained from the latex. Mechanical reinforcement of the continuous
phase, whether it be rubbery or glassy, requires that the reinforcing phase
be dispersed within the principal phase as very fine particles. If the reaction
conditions within the latex of a rubbery polymer are such that grafting of
blocks of a glassy polymer occurs preferentially near the surface of the
particles, then integration of the particles to form a coherent film may be
restricted or prevented altogether.
In the last 40 years or so, there has been great academic and industrial
interest in block and graft copolymers produced by all methods, including
latex graft-copolymerization reactions. This is because block and graft
copolymers represent extreme ways in which two or more types of monomer
unit can be assembled to form a copolymer. Because the various types of
unit are present as large blocks, the copolymer molecule often displays a
combination of the characteristics of the separate homopolymers, rather than
characteristics intermediate between those of the separate homopolymers.
Two important uses which exploit this tendency are as adhesives for pairs
of adherends of disparate character, and applications which depend upon
the strong tendency for the respective blocks of the copolymer to separate
into distinct microscopic phases. In the first of these applications, the
copolymer is designed to be such that one block has an affinity for one type
of adherend, and the other has an affinity for the other type. Examples of
the second of these applications include uses of phase-separated microdomains
of a glassy polymer within a matrix of a rubbery polymer
1. to provide self-reinforcement and stiffening of the rubber phase;

2. to produce so-called thermoplastic rubbers, in which the glassy domains
provide thermally-reversible crosslinking sites within the rubber phase; and
3. to improve the impact-resistance of the glassy phase.
14.3.2 Graft-copolymerization reactions in natural rubber latex
14.3.2.1 General considerations

There has for many years been interest in modifying the properties of natural
rubber by graft-copolymerization, both in bulk and by reaction in latex form.
One important motivation has been the desire to produce rubber which,
when vulcanized, swells less when in contact with hydrocarbon oils than do
vulcanizates from unmodified natural rubber. Graft-copolymerization with
polar monomers such as acrylonitrile has been considered as a possible route
to the production of oil-resistant natural rubber. Another important motivation
has been the production of self-reinforced and thermoplastic natural rubbers,
in which the desired modification is achieved by way of phase-separation of
blocks of a glassy polymer which have become covalently bonded to the
rubber molecules by graft-copolymerization. As regards modified natural
rubber produced by this reaction in natural rubber latex, by far the most
important product has been a self-reinforced rubber obtained by graft-
copolymerization with methyl methacrylate. This product is available industrially
under the name 'Heveaplus MG'. Various grades of this material have been
produced. An important variable is the mass ratio of poly methyl methacrylate
to natural rubber. This material finds use in applications which require a
hard rubber with optimum properties such as tensile strength and impact
strength, and where the use of reinforcing carbon blacks is undesirable.
Free-radical graft-copolymerization reactions involving an olefinically-
unsaturated polymer such as natural rubber can occur by both the mechanisms
outlined in section 14.3.1 above. The polyisoprene chain of natural rubber
is an example of a polymer which contains allylic (oc-methylenic) hydrogen
atoms susceptible to abstraction by interaction with free radicals, thereby
generating reactive sites from which copolymer grafts can propagate. The
reaction by which such sites are generated in polyisoprene by this means is
as follows:
... --CH 2 --C(CH 3 )=CH--CH 2- ... + R· - - ... -CH 2--C(CH 3)=CH-CH-... + RH
Grafting can also occur by polymerization through the olefinic double bond
of the polyisoprene chain as follows:
... -CH 2-C(CH 3 )=CH-CH 2 - ... + R· - - ... -CH 2-CR(CH 3 )-CH-CH 2 - ...
The indications are that, for natural rubber, grafting tends to occur more
readily by hydrogen-abstraction than by copolymerization.
A convenient summary of work on graft-copolymerization reactions
involving natural rubber in general, and on graft-copolymerization reactions
in natural rubber latex in particular, up to the early 1960s, has been given
by Allen [11]. Somewhat more recent summaries have been given by Ceresa
[12] and by Pend Ie [13]. The latter deals with the properties and applications
of products obtained by reactions in both latex and solid rubber. In early

attempts to polymerize various vinyl monomers in ammonia-preserved
natural rubber latex, reported by Bacon, Farmer and Schidrowitz [14],
problems of initiation of polymerization were encountered. Bloomfield [15]
has described some of the difficulties experienced in the course of early
attempts to polymerize vinyl monomers in ammonia-preserved natural rubber
latex. The principal difficulty was that of finding an initiating system which
would function in the presence of air and the ammonia preservative, both
of which have a retarding effect upon many conventional free-radical
initiators. One approach was to use large amounts (up to 8% m/m) of
dibenzoyl peroxide; even so, polymerization occurred only when the latex
was completely de-ammoniated. Experience in Bloomfield's laboratory
indicated that methyl methacrylate could be persuaded to polymerize in
natural rubber latex by the use of dibenzoyl peroxide only if the ammonia
was removed, if the temperature was raised to nearly 80·C, and if a substantial
quantity of a non-ionogenic stabilizer was added to prevent colloidal
destabilization. Peroxodisulphate initiators were found to be more effective
than dibenzoyl peroxide in the presence of ammonia. Bloomfield reported
that potassium peroxodisulphate, with or without a mercaptan promoter, is
effective for initiation of the polymerization of methyl methacrylate in natural
rubber latex at 50·C. De-ammoniation of the latex was found to be
unnecessary, but it was necessary to exclude air.
An important advance was made when it was discovered that the initiating
efficiency of the polyamine-hydroperoxide redox combination is unaffected
by the presence of ammonia, and is only slightly reduced by the presence of
air. A combination of tert-butyl hydroperoxide and tetraethylenepentamine
initiates rapid and smooth polymerization of both methyl methacrylate and
styrene in ammonia-preserved natural rubber latex, the former monomer at
normal ambient temperatures. These observations have provided the basis
for industrial processes for effecting graft-copolymerization in ammonia-
preserved natural rubber latex. One component of these initiating systems
(the hydroperoxide) partitions mainly in the rubber phase; the other (the
polyamine) partitions mainly in the aqueous phase. It is of interest that
Sekhar [16] has found that it is not necessary to add a hydroperoxide; if the
latex is merely aerated by rotation in a half-filled container for several days,
various monomers can be slowly polymerized in the latex if small amounts
of a polyamine and an iron(II) compound are added. The ability to initiate
polymerization is associated with the rubber phase of the latex, and not with
the aqueous phase. It is probably a consequence of the formation of
hydroperoxidic sites on the polyisoprene during aeration. Furthermore,
ammonia-preserved natural rubber latex which has been pre-treated with a
hydroperoxide and a polyamine retains its ability to initiate the polymerization
of added monomers even though the period which has elapsed since the
pre-treatment is such that almost all the added hydroperoxide is expected
to have disappeared. It has also been found that polymerization in

ammonia-preserved natural rubber latex can be initiated by various other
combinations, such as dibenzoyl peroxide activated by dimethylaniline, and
by hydrazine and copper(II) ions. Some interest has also been shown in the
initiation of graft-copolymerizations in ammonia-preserved natural rubber
latex by irradiation with y-rays.
14.3.2.2 Graft-copolymerizations initiated by added substances

Typical formulations for the graft-copolymerizations of methyl methacrylate
and of styrene in ammonia-preserved natural rubber latex using a
hydroperoxide-polyamine initiation system are shown in Table 14.2.
Approximately half the mass of rubber in the reaction system becomes
modified by graft-copolymerization, and approximately half the added
monomer participates in graft-copolymerization reactions, i.e., approximately
half of the added monomer is present in the final product as homopolymer.
The extent to which grafting occurs depends upon the polymerization initiator
which is used. Significant extents of grafting occur ifhydroperoxide-polyamine
combinations or dibenzoyl peroxide are used as initiator. By contrast, if
azobisisobutyronitrile is used as initiator, almost no grafting occurs. The
Table 14.2 Typical formulations for the graft-copolymerizations of methyl methacrylate

and of styrene in ammonia-preserved natural rubber latex (based upon recommendations
given by Bloomfield [15])
Parts by mass
Methyl methacrylate Styrene
Ingredient Dry Actual Dry Actual
Natural rubber (as 30% m/m

latex, 0.4% m/m ammonia) 100 333 100 333
Non-ionogenic stabilizer
(as 20% aqueous solution) 3 15
Methyl methacrylate 33 33
Styrene 55 55
Tert-butyl hydroperoxide 0.18 0.18 0.25 0.25
Tetraethylenepentamine
(as 10% m/m aqueous solution) 0.21 2.1 0.10 1.0
Time of polymerization / hour 3 6.5

Polymerization temperature / ·C 12 55
Conversion / % 90 95
extent of grafting is low if peroxodisulphates are used as initiators. The

dependence of the extent of grafting upon the nature of the initiator is
inconsistent with reaction mechanisms in which grafting occurs principally by
copolymerization or by hydrogen-abstraction by interaction between a
propagating polymer chain and a rubber macromolecule; neither of these
reactions would be expected to depend upon the nature of the free radical which
initiated the polymerization. For this reason, it is believed that grafting occurs
primarily by interaction between rubber macromolecules and the primary
radicals which form from the initiator. The absence of significant grafting
when azobisisobutyronitrile is used as initiator is attributed to inability of
the 2-cyanopropyl free radical to interact with rubber macromolecules.
A detailed investigation of the kinetics of the polymerization of methyl
methacrylate and styrene in ammonia-preserved natural rubber latex has
been reported by Allen, Bell and Cockbain [17]. The special type of
dilatometer illustrated in Fig. 14.2 was used. Approximately 50cm 3 of latex
was confined over mercury such that volume changes in the reaction system
of less than 1 x 10- 4 % could be recorded. Correspondingly accurate
thermos tatting (±O.OOloq was necessary. 60% mlm ammonia-preserved
centrifuged natural rubber latex was used. It was diluted with distilled water
0.7 mm precision-bore
capillary glass tubing
ca. SO cm 3 of
latex confined
over mercury
ca.6-mm 10
glass tubing
wire
contact
Figure 14.2 Dilatometer used by Allen, Bell and Cockbain [17] for investigation of kinetics
of polymerization of methyl methacrylate and styrene in ammonia-preserved natural rubber latex.
as required (to ca. 5% mlm dry rubber content). Potassium laurate was added
as a colloid stabilizer to the extent of 0.5% mlm on the diluted latex. The
intention was that, at this level of addition, there would be so few micelles
present derived from the laurate anions that negligible polymerization would
occur outside the rubber particles. The pH of the reaction system was adjusted
to ca. 11 by addition of potassium hydroxide. Ammonia and dissolved oxygen
were removed by blowing nitrogen through the system for 72 hours at 50·e.
Two types of initiator were used: azobisisobutyronitrile, added as a solution
in the monomer, and potassium peroxodisulphate, added as a dilute aqueous
solution. An induction period was observed with both monomers. Typical
conversion-time curves are shown in Fig. 14.3{a). As expected, the duration
ofthe induction period decreased as the initiator level increased. For reaction
systems which contained methyl methacrylate, it was found that the rate of
conversion goes through a sharp maximum at low conversion, and then falls
very rapidly away from this maximum as conversion progresses, as shown
in Fig. 14.3{b). This effect is thought to be a consequence of the presence of
micro-aggregates of phase-separated polymethyl methacrylate within the
latex particles, and to the partitioning of the monomer between the rubber
phase and the micro-aggregates within the particles. This view is confirmed
by the observation that the rate-conversion relationship for the styrene-natural
rubber latex reaction system corresponds more nearly to that which would
be expected for a simple monomer-depletion effect. Polystyrene, having a
greater tendency to mix with natural rubber than does polymethyl methacrylate,
will have a lesser tendency to phase-separate to form micro-aggregates; thus
the distribution of monomer within the composite latex particles is expected
to be more uniform.
An important contribution to our understanding of the reactions which
occur when a vinyl monomer undergoes free-radical polymerization in the
presence of polyisoprene is provided by the results of an investigation reported
by Allen, Ayrey and Moore [18] . Gutta percha, a trans-l,4-polyisoprene,
was used instead of natural rubber, and a benzene solution of the polymer
instead of a latex system. Nevertheless, the results obtained are of relevance
for graft-copolymerization reactions in natural rubber latex. The monomer
was methyl methacrylate. The initiator was 14C-labelled dibenzoyl peroxide.
It was found that, when di-{aryl-14C)-benzoyl peroxide was used to initiate
graft-copolymerization, a graft copolymer was obtained which contained
ca. 0.9 14C-initiator fragments per grafted polymer molecule. Initiation by
di-{carbonyl-14C)-benzoyl peroxide under similar conditions produced a graft
copolymer which contained ca. 0.8 14C-initiator fragments per grafted
polymer molecule. Assuming, as is usual, that the radical species generated
by the thermal decomposition of dibenzoyl peroxide are phenyl and benzoyloxy
radicals, it follows that, of the total radicals incorporated in the copolymer,
ca. 90% are of the latter type and only ca. 10% are of the former type.
Furthermore, nearly all of the 14C-radioactivity was found to be retained
60
rfl.
......
c
0 40
~
N
'5i
E
>-
"0
a.
'0
EQ) 20
><
w
50 100 150
(a) Time / minute
25
en
.'::J2 20
~
I!!
:e
......aI 15
c
0
~
N
'53
E
>-
8.
'0
OJ
til
a:
00 10 20 30 40 50 60 70
(b) Extent of polymerization / %
Figure 14.3 (a) Conversion-time curves for polymerization of styrene at 70·C (curve A) and
methyl methacrylate at 70·C (curve B) and 50T (curve q in ammonia-preserved natural rubber
latex (dry rubber content 5% m/m) using ammonium peroxodisulphate (0.08% m/m on latex)
as initiator (Allen, Bell and Cockbain [17]). Initial monomer/rubber ratio was 1/1 m/m.
(b) Variation of rate of polymerization with conversion during polymerization of methyl
methacrylate in ammonia-preserved natural rubber latex (dry rubber content 5% m/m) at 5O·C
using ammonium peroxodisulphate as initiator (Allen, Bell and Cockbain [17]). Initiator levels:
0.08% (curve A), 0.23% (curve B) and 0.77% (curve q m/m on the latex. Initial monomer/rubber
ratio was 1/1 m/m.
after ozonolytic degradation of the graft copolymer. The inference is that the
initiator fragments which are incorporated in the graft copolymer are
associated almost exclusively with the grafted methyl methacrylate, i.e. they
are either directly attached to the polymethyl methacrylate graft, or else they
are located on the polyisoprene chain at positions adjacent to the site of
grafting. These results are consistent with the view that, for methyl methacrylate
and polyisoprene, grafting results primarily from direct interaction between
the radicals derived from the initiator and the polymer molecule. The presence
in the graft copolymer of ca. 0.9 initiator fragments per polymethyl methacrylate
graft is sufficiently large effectively to exclude transfer-to-polymer as an
important pathway to graft-copolymerization. Allen, Ayrey and Moore
consider that interaction between polyisoprene and initiator radicals may be
by way both of addition to the olefinic double bonds of the polyisoprene
and of abstraction of allylic hydrogen atoms.
Allen, Ayrey and Moore [18] consider that, at least under the reaction
conditions employed by them, termination reactions occur exclusively by
bimolecular interaction between propagating polymethyl methacrylate radicals,
resulting in either combination or disproportionation. They regard as
negligible the probability that a polyisoprene radical reacts with another
polymeric species under the particular conditions employed in their reactions.
As a corollary, interaction between polymeric radicals is not thought to be
an important pathway to the formation of graft copolymer.
Andrews and Turner [19] have reported the results of an electron-
microscope examination of graft copolymers of natural rubber and methyl
methacrylate prepared in latex using, on the one hand, a redox system
expected to partition almost exclusively in the rubber phase of the latex, and,
on the other, a redox system such that one component is expected to partition
mainly in the rubber phase and the other component mainly in the aqueous
phase. The findings from this investigation are in agreement with the
expectation that the distribution of polymethyl methacrylate within the
rubber particle is likely to be more uniform if the former type of initiator is
used than if the latter type is used. The use of a redox initiator, one component
of which is in the rubber phase and the other component in the aqueous
phase, is expected to favour polymerization at or near the particle surface,
because this is the region where the initiating free radicals are most readily
generated by interaction between the two components of the system. The
use of a redox initiator, both components of which are in the rubber phase,
is expected to generate radicals uniformly throughout the particle. The
findings reported by Andrews and Turner also corroborate qualitative
observations of the behaviour of natural rubber-polymethyl methacrylate
graft-copolymer latices obtained using the two types of redox initiator.
Graft-copolymer latices prepared using an initiation system comprising a
rubber-soluble hydroperoxide and a water-soluble polyamine tend to flocculate
rather than coagulate, and to dry down to discontinuous films of limited
coherence. By contrast, comparable graft-copolymer latices prepared using

a completely rubber-soluble redox initiation system, such as dibenzoyl
peroxide and dimethylamine, give coherent coagula when colloidally destabilized,
and dry down to continuous coherent films.
Most graft-copolymerizations in natural rubber latex, of the type described
above, have been carried out using hydrophobic monomers. There has also
been limited interest in graft-copolymerizations in this latex using added
initiators and hydrophilic monomers which themselves yield water-soluble
or water-swell able homopolymers. Thus Burfield and Ng [20] have investigated
the graft-copolymerization of methacrylamide in ammonia-preserved natural
rubber latex at 60·C using potassium peroxodisulphate as initiator. The
overall course of the graft-copolymerization was found to be similar to that
of the corresponding homopolymerization. After an initial induction period,
the rate of polymerization increased to a maximum, and then gradually
declined as a consequence of monomer depletion. The rate of polymerization
during the steady-rate region was found to be kinetically of first order in
monomer concentration, at least up to 70% conversion. One surprising
observation was that, although the induction period was slightly longer than
the corresponding homopolymerization (which was expected because of the
presence of oxygen and other non-rubber substances), the rate of polymerization
was significantly faster, in contrast to what has been observed in other
graft-copolymerizations in natural rubber latex. The rate of polymerization
increased markedly with increasing concentration of rubber in the latex. This
effect is attributed to retardation of the termination step of the polymerization.
The polymerization reaction is envisaged as taking place predominantly in
the aqueous phase of the latex. Thus it is feasible to suppose that bimolecular
termination of macromolecular radicals may be retarded by factors such as
increased viscosity of the latex arising from increased concentration of rubber
particles, and adsorption of the macromolecular radicals at the surface of
the rubber particles. Evidence is presented that some graft-copolymerization
of methacrylamide on to natural rubber in latex form does occur. Efficiency
and degree of grafting depends upon the polymerization conditions. Grafting
efficiency is typically ca. 30%. It increases with increasing polymerization
temperature. This may be a consequence of increased concentration of
hydroxide radicals formed by interaction between sulphate radical-ions and
water molecules. It is expected that hydroxide radicals will be better able to
penetrate negatively-charged latex particles than are negatively-charged
sulphate radical-ions, and so are the better able to initiate graft-copolymerization
rather than homopolymerization in the aqueous phase.
14.3.2.3 Graft-copolymerizations initiated by high-energy irradiation

Cockbain, Pendle and Turner [21] have shown that the graft-copolymerization
of methyl methacrylate on to natural rubber in latex form can be initiated
by irradiation with cobalt-60 y-rays. The resultant latices give coherent

coagula when colloidally destabilized, and dry down to continuous coherent
films. It is therefore surmised that the grafted copolymer is uniformly
distributed throughout the latex particles. This accords with the expectation
that free radicals are generated uniformly throughout both the polymer
particles and the dispersion medium when a latex is irradiated with
high-energy electromagnetic waves. Cockbain, Pendle and Turner also report
the following interesting and significant observations:
1. Addition of a rubber-miscible polymerization-retarder, such as tert-
dodecyl mercaptan or zinc diethyldithiocarbamate, to the latex prior to
irradiation gives a product which flocculates on addition of acids, and
which dries down to incoherent films. The obvious interpretation of this
observation is that graft-copolymerization within the interior of the latex
particles is discouraged to a greater extent than that at the particle surface.
2. Addition of a water-soluble polymerization retarder, such as sodium
diethyldithiocarbamate, to the latex prior to irradiation gives a product
which coagulates on addition of acids, and which has excellent film-forming
properties. The implication is that, in the presences of such additives,
graft-copolymerization at the particle surface is discouraged to a greater
extent than that in the particle interior.
Cooper et at. [22] have also reported upon the formation of graft
copolymers from methyl methacrylate and natural rubber in latex form, using
electromagnetic waves of various frequencies to initiate polymerization.
60% mlm ammonia-preserved natural rubber latex was used. Sodium caseinate,
to the extent of 10% mlm on the monomer, was added to supplement the
colloid stability of the mixture. The overall concentration of disperse material
in the reaction system was adjusted to 40% m/m. After polymerization, the
latices were colloidally destabilized by addition of ethyl alcohol. The coagula
obtained were extracted with acetone to remove any methyl methacrylate
homopolymer which had formed. The residual polymer was then degraded
ozonolytically to remove the polyisoprene component. The latex-monomer
mixtures were irradiated with either cobalt-60 y-rays, or with the radiation
(not monochromated) from a mercury-arc or a tungsten-filament source.
Induction periods were reported to be relatively short. The rate of polymerization
was found to be constant, and kinetically of first order with respect to the
monomer concentration. It was inferred from this that the rate of initiation
was independent of the monomer concentration, and that the predominant
mode of initiation was the formation of free-radical sites on the polyisoprene
chain. This finding may be compared with that of Allen, Ayrey and Moore
[18] for the mechanism of the initiation of the graft-copolymerization of
methyl methacrylate in natural rubber latex by radicals formed by the thermal
decomposition of dibenzoyl peroxide. Cooper et at. also reported that the
effect of varying the intensity of the y-rays was such as to indicate an intensity
exponent of ca. 0.5. This suggests that in this reaction the termination of
propagating polymer chains occurs by bimolecular interaction between
radicals, as is so for many bulk free-radical addition polymerizations. For
initiation by irradiation with ultraviolet light, however, an intensity exponent
of 0.25 was observed. This result is more difficult to interpret. Several
substances were found to function as sensitizers for graft-copolymerization
initiated by irradiation with ultraviolet light. Of the compounds investigated,
l-chloro-9, 1O-anthraquinone (Structure I) was found to be the most satisfactory.
o CI
It is unlikely that the intensity exponent of ca. 0.5 reported by Cooper et

al. [22] for initiation by irradiation with y-rays is related to the exponent of
0.4 for the dependence of the rate of a free-radical emulsion polymerization
reaction upon the aqueous-phase initiator concentration, predicted by the
Smith-Ewart theory of micellar nucleation (see section 10.2.2.2.1 of Chapter
10). The Smith-Ewart prediction for the effect of initiator concentration
upon the rate of polymerization in Interval II arises as a consequence of the
prediction for the variation with initiator concentration of the concentration
of reaction loci which form during Interval I. The same dependence is then
predicted for the rate of polymerization during Interval II, because that rate
is postulated to be directly proportional to the concentration of reaction loci
formed during Interval I. In the experiments reported by Cooper et al., most,
if not all, of the polymerization will have occurred within the latex particles
which were present at the beginning of the reaction. Insofar as analogies
with emulsion polymerization reactions are made, latex graft-copolymerizations
resemble seeded reactions more closely than they do ab initio reactions.
Indeed, latex graft-copolymerization reactions are probably best regarded as
micro-bulk polymerizations in which the individual reaction loci are provided
by the pre-existing particles. Similar comments apply to results for the effect
of initiator concentration upon rates of latex graft-copolymerization reaction
systems where initiation is by added substances.
Cooper et al. [22] also noted that, although the initial fraction of polymer
grafted was high in all cases, being ca. 90-100%, this fraction decreased as
polymerization proceeded. Sometimes the decline was sharp. In general, the
higher the initial monomer concentration, the lower was the conversion at
which non-grafted polymethyl methacrylate homopolymer began to form; in

some cases, nearly all the polymer produced was ungrafted. The rapid decrease
in the fraction of monomer grafted as polymerization proceeded was
attributed to phase-separation of polymethyl methacrylate blocks within the
rubber particles. The formation of mainly homopolymer in the latter stages
of the reaction is then explained as being a consequence of partitioning of
the monomer between the rubber phase and the polymethyl methacrylate
phase in such a way as greatly to favour the latter phase. This explanation
for the formation of homopolymer is similar to that proposed by Allen, Bell
and Cockbain [17] that the reduction in rate of polymerization of methyl
methacrylate in natural rubber latex initiated by added substances is a
consequence ofthe formation of micro-aggregates of polymethyl methacrylate
with the latex particles.
Papers by Cooper and Vaughan [23] and by Cooper, Sewell and Vaughan
[24] provide further information relating to y-ray-initiated graft-copolymer-
izations of various monomers in ammonia-preserved natural rubber latex.
That by Cooper and Vaughan discusses in detail the kinetics of these reactions.
Experimental results for the graft-copolymerization of methyl methacrylate
are said to be consistent with slow propagation from polyisoprene radicals
and rapid cross-termination between propagating polymer and polyisoprene
radicals. The high proportion of graft copolymer which forms in these
reactions is explained by assuming the radicals produced initially from the
monomer to be of such high activity that they rapidly attack the polyisoprene
molecules to form initiation sites before significant propagation to form
homopolymer occurs.
14.3.2.4 Properties of graft copolymers of natural rubber

Conventional sulphur-vulcanizing systems can be used with the dry rubbers
obtained from graft-copolymer natural rubber latices. The most obvious
effect of increasing levels of polymerized methyl methacrylate and styrene
upon the mechanical properties of vulcanizates from such rubbers is increase
in the stiffness and hardness of the material. The tensile strength tends to
fall at the higher levels of polymerized monomer. The polymerization
initiation system itself causes some hardening of the subsequent vulcanizate,
even if added monomer is absent. However, hardening does not occur when
the polyamine component of the initiation system is omitted, from which it
is concluded that hardening is not caused by the hydroperoxide alone.
Vulcanized graft copolymers of natural rubber and methyl methacrylate are
said to display low hysteresis together with outstanding resistance to
flex-cracking and to fatigue. The tear resistance, however, appears to be
somewhat reduced. Styrene is less efficient as a reinforcing comonomer than
is methyl methacrylate. Comparison between graft copolymers of natural
rubber and polymer blends of the same overall composition has shown that
the blends are stiffer at low extension than are the graft copolymers, but
have similar stiffnesses at extensions in excess of 300%. As would be expected,
the reinforcing efficiency of the esters of methacrylic acid decreases as the
size of the alkyl group increases. Results illustrating this effect reported by
Bloomfield et al. [25] are shown in Table 14.3. The stiffening effect correlates
broadly with the glass-transition temperature of the homopolymer of the
grafted comonomer: the higher is that glass-transition temperature, the
greater is the stiffening effect at normal ambient temperatures. This is
presumably a consequence of two effects. The principal effect is that, as the
glass-transition temperature of the homopolymer of the grafted comonomer
decreases, so the modulus of the phase-separated grafted polymer at normal
ambient temperatures decreases. A secondary effect is that the tendency of
the grafted copolymer to phase-separate decreases as the chain length of the
alkyl group of the ester increases, and increasing this chain length also has
the effect of decreasing the glass-transition temperature of the homopolymer
of the grafted comonomer.
Graft copolymers of natural rubber of the type obtained by polymerizing
a monomer such as methyl methacrylate in natural rubber latex do not
behave as thermoplastic rubbers in the sense in which the term is conventionally
used. The reasons for this have been discussed by Campbell [26]. They are
believed to be associated with the primary molecular mass of the rubber
Table 14.3 Effect of size of alkyl group of methacrylic-acid ester upon mechanical
properties of vu1canizates of natural rubber-alkyl methacrylate graft copolymers
containing 30 pphr of alkyl methacrylate (Bloomfield et 01. [25])
Vulcanizate formulation (parts by mass):
Rubber 100
Sulphur 2.5
N-cyclohexyl-2-benzthiazylsulphenamide 0.6
Zinc oxide 5
Stearic acid 1
Phenyl-2-naphthylamine 1
Vulcanization conditions: 20 minutes at 140°C.
Alkyl group of Tensile Extension

methacrylate MlOO· M300t strength at break Shore
ester / MPa / MPa / MPa /% hardnesst
Methyl 3.14 11.77 27.5 560 75

Ethyl 2.35 8.43 23.5 553 63
Isobutyl 1.67 5.98 22.1 578 55
n-butyl 1.37 3.73 18.6 613 56
• Modulus at 100"1. extension.

t Modulus at 300"1. extension.
t A measure of indentation hardness: the greater the number, the greater the resistance to indentation.
macromolecules in natural rubber latex, and with the distribution of grafted

sites along the individual rubber molecules.
Burfield and Ng [20] have demonstrated that peroxide-vulcanizates
derived from the graft copolymers of natural rubber and methacrylamide
prepared by graft-copolymerization in natural rubber latex (see section
14.3.2.2 above) have significantly better resistance to swelling in hydrocarbon
liquids than do similar vulcanizates from unmodified natural rubber. This is
to be expected, bearing in mind the hydrophilic nature of polymethacrylamide.
Vulcanizate modulus and hardness were also observed to increase, possibly
indicating that some crosslinking may have been induced by the polymerization
initiation system. Vulcanizate tensile strength and extension at break tended
to be reduced by graft-copolymerization with this monomer.
14.3.3 Graft-copolymerization reactions in other latices

Hayes [27] has reported the results of attempts to graft-polymerize several
monomers on to various polymers in latex form, the latter having been
prepared by the emulsion polymerization of the appropriate monomers. Four
synthetic latices were investigated: those of polyvinyl chloride, polyacrylonitrile,
polystyrene and polyvinyl acetate. The formulations used for the preparation
of the initial latices are shown in Table 14.4. The grafting experiments were
carried out at 60·C, using as initiator potassium peroxodisulphate either by
itself or in combination with other substances. Sufficient water was added
to increase the ratio of dispersion medium to total disperse phase to 2/1. No
additional colloid stabilizer was added. All the graft-copolymerization
reactions except those involving vinyl chloride and vinyl acetate with
polystyrene were complete within 16 hours. Reaction systems which contained
these two monomers required between 24 and 40 hours for complete
conversion. The resultant polymers were isolated from the latices by colloidal
destabilization using 10% m/v aqueous calcium chloride solution, followed
Table 14.4 Formulations used by Hayes [27] for preparation of synthetic latices to
be used for graft-copolymerization experiments
Parts by mass for monomer indicated
Vinyl Vinyl
Ingredient chloride Acrylonitrile Styrene acetate
Monomer 100 100 100 100

Potassium peroxodisulphate 0.3 0.25 0.5 0.5
Sodium n-dodecyl sulphate 1.0 2.5 1.5 1.5
Water 233 625 200 200
Polymerization temperature / °C 45 60 60 60
Conversion / % 100 98 99 100
by filtration, washing with water, and drying at 50·C. In all cases, the yields
were in the range 98-100%, and were assumed to be exactly 100% for the
purpose of calculating copolymer compositions and grafting efficiencies. Some
of the results reported by Hayes for the system vinyl acetate-polyvinyl
chloride are shown in Figs 14.4(a), (b) and (c). It is evident that the grafting
efficiency increased as the ratio of monomer to polymer was increased, but
was drastically reduced by the addition of a chain-transfer agent such as
n-dodecyl mercaptan. The initiator concentration also had an effect upon
grafting efficiency. Hayes regards the effect of n-dodecyl mercaptan upon
grafting efficiency as indicating that the principal mechanism for graft-
copolymerization in this type of system is that which commences with transfer
1.0 1.0
>- 0.8 >- 0.8

o o
c: c:
II) II)
~ 0.6 ~ 0.6
II)
Ol Ol
c: 0.4 ~ 0.4
~
(!) 0.2 (!) 0.2
00 20 40 60 80 100 00 1 2 3 4
Vinyl acetate content of monomer Level of n-<lodecyl mercaptan! pphm
(a) mixture! % mlm (b)
1.0
>- 0.8
o
c:
II)
:§ 0.6
:m
Ol
~ 0.4
~
(!) 0.2
00 0.4 0.8 1.2 1.6

(c) Square root of initiator level! pphm112
Figure 14.4 (a) Variation of grafting efficiency with concentration of monomer in reaction
system during graft-copolymerization of vinyl acetate in polyvinyl chloride latex at 6O"C (Hayes
[27]). Initiator was 0.5 pphm of potassium peroxodisulphate. (b) Effect of added n-dodecyl
mercaptan upon grafting efficiency during graft-copolymerization of vinyl acetate in polyvinyl
chloride latex at 6O"C (Hayes [27]). Initiator was 0.5 pphm of potassium peroxodisulphate;
monomer/polymer ratio was 5O/50m/m. (c) Variation of grafting efficiency with square root of
concentration of initiator in reaction system during graft-copolymerization of vinyl acetate in
polyvinyl chloride latex at 6O"C (Hayes [27]). Monomer/polymer ratio was 5O/50m/m.
Cis-trans isomerization of natural rubber in latex form 537
of radical activity to polymer. Additional support for this mechanism was

adduced from the observation that the grafting efficiency decreased as the
temperature was reduced. This correlates with the transfer coefficient (i.e.,
the ratio of rate coefficient for transfer-to-polymer to that for propagation
of the monomer) also decreasing with decreasing temperature.
Hayes [27] has also reported information concerning the relative rates of
grafting of various vinyl monomers on to various polymers in latex form.
For the monomers, he gives the order of decreasing activity as
vinyl chloride > vinyl acetate> styrene
He claims that this is the order which would be predicted from knowledge
of their respective transfer coefficients. For the polymers, the order of
decreasing activity is given as
polyvinyl chloride> polyacrylonitrile > polyvinyl acetate> polystyrene
The high activity of polyvinyl chloride correlates with the well-known
tendency of chlorinated hydrocarbons to participate in radical-transfer reactions.
Gardner and Harper [28] have reported the results of an investigation
into the graft copolymerization of ammonium polyacrylate, ammonium
polymethacrylate and various metal-chelating vinyl monomers on to styrene-
butadiene copolymers in latex form. The simplest of the metal-chelating vinyl
monomers was 2-(vinylphenyl)aminoacetic acid, the p-isomer of which is
shown here as Structure II. Polymerization was initiated by exposure to
I-Mev electrons. A typical dose was lOkGy.
14.4 CIS-TRANS ISOMERIZATION OF NATURAL RUBBER IN LATEX

FORM

Interest in the possibility of effecting the partial cis-trans isomerization of
the natural rubber hydrocarbon has its origin in the observation that both
raw and vulcanized natural rubber tend to suffer irreversible hardening and
loss of elasticity if stored for long periods at low temperatures. For vulcanized
rubber, the effect is especially marked if the rubber is maintained under an
appreciable strain. The hardening is believed to be a consequence of slow
crystallization of segments of the rubber macromolecules. If rubbery properties
are to be retained during long periods of storage at low temperatures, it is
necessary to suppress, or at least greatly to retard, the crystallization of the
polymer. The crystallization tendency of unmodified natural rubber is
attributed to the high degree of stereoregularity of its molecular structure;
thus in principle the tendency to low-temperature hardening can be mitigated
by reducing somewhat that degree of stereoregularity. It is, of course, essential
that any reduction of the degree of stereoregularity should not adversely
affect the processing behaviour of the rubber, or the mechanical properties
and ageing behaviour of vulcanizates obtained from the rubber.
The natural rubber macromolecule is stereo regular in two important respects:
1. If, as is usual, it is regarded as having been produced by the addition
polymerization of isoprene (which it is not - see section 9.7 of Chapter 9),
then most of the isoprene units are linked together through their 1,4
positions; and
2. the configuration of the substituents about the majority (possibly all) of
the olefinic double bonds is cis, in the sense that the two polymeric
substituents at each double bond (and hence also the two non-polymeric
substituents) are on the same side of the double bond as each other, not
on opposite sides.
It is impracticable and undesirable to attempt to interfere with the first of
these two aspects of stereoregularity. The most attractive way of reducing
the crystallization tendency of the macromolecule is therefore to isomerize
a proportion of the olefinic double bonds to the trans configuration.
Cunneen and Shipley [29] have reported that one way of satisfactorily
achieving partial cis-trans isomerization of natural rubber is to heat the
latex with a small amount of an alkane- or arylenecarbothiolic acid
(R.CO.SH), together with a free-radical generator. The product of such
reaction is a rubber which is much less prone to irreversible low-temperature
hardening than is unmodified natural rubber. The change is believed to be
a consequence of partial cis-trans isomerization which has occurred by the
following reaction mechanism:
X· + R.CO.SH R.CO.S· + XH
"
---+
C=C / + R.CO.S· == " C-C /

/ " /1S.OC.R
"
In these reactions, X· denotes a free radical which has been formed from the
radical generator. In the first of these reactions, a carbothiolyl free radical
is formed. This then adds reversibly to an olefinic double bond of the rubber
macromolecule. The olefinic double bond which is formed when the second
of the above reactions reverses can have either the cis or the trans
configuration. In this way, the isomerization and the loss of stereoregularity
are explained.
The original explanation for the effect of treatment with carbothiolic acids
was that the acid merely added across the olefinic double bonds of the rubber
macromolecule, thereby incorporating bulky side-groups which were effective
in suppressing the crystallization of natural rubber. A plausible mechanism
for such a free-radical addition is the following:
••• -CH 2 -C(CH 3 )=CH-CH 2 - ••• + R.CO.S·
••• -CH 2 -C(CH 3 )-CH-CH 2 - •••
~SH
S.OC.R
••• -CH 2 -CH(CH 3 )-CH-CH 2 - ••• + R.CO.S·

I
S.OC.R
However, the alternative isomerization explanation is preferred because it

satisfactorily explains why a large effect is brought about by small additions
of the carbothiolic acid. The isomerization theory is also consistent with the
observation that thiophenols are able to effect the isomerization of both cis-
and trans-2-butene.
14.4.2 Experimental details
14.4.2.1 Preparation of isomerized natural rubber from latex

The method of Cunneen and Shipley [29] was to stabilize 60% mlm
ammonia-preserved centrifuged or creamed natural rubber latex with ca.
2 pphr of a fatty-amine ethoxylate added as an aqueous solution. The 50-mole
ethoxylate of n-octadecylamine was found to be especially suitable for
stabilizing the latex against the high levels of acidity required for the
satisfactory completion of the reaction. Sufficient dilute hydrochloric acid
was added to reduce the pH to ca. 0.6. The low pH was presumably necessary
to ensure that the solubility of the carbothiolic acid in the aqueous phase
was a minimum. Up to ca. 1 pphr of the carbothiolic acid (see further section
14.4.2.3 below) was then added to the latex, either as a solution in benzene
or as an aqueous emulsion of a solution in a benzene-carbon tetrachloride

mixture. Several radical-generating substances were found to be effective as
a catalyst, including tert-butyl hydroperoxide, isopropyl benzene (cumene)
hydroperoxide, ascaridole (a naturally-occurring terpene peroxide shown as
Structure III) and azobisisobutyronitrile. The amount of catalyst used was
in the range 10-20% mlm on the carbothiolic acid. It was added in the same
manner as the latter. Isomerization was effected by heating the mixture for
ca. 5 hours at 60°C. The resultant latices were coagulated by pouring into a
polar water-miscible liquid such as ethyl alcohol or acetone, or by the addition
of substances which form large polyvalent anions, e.g., tannic acid,
polymethacrylic acid, etc. The coagula were then either acetone-extracted
and vacuum-dried, or washed on a washing mill and then oven-dried.
III
14.4.2.2 Assessment of resistance to crystallization

The rates of crystallization of both raw and vulcanized natural rubber are
maximal at ca. - 26°C. This temperature has therefore often been selected
for making comparative measurements of rates of crystallization. Cunneen
and Shipley [29] followed the crystallization of the raw isomerized and
unmodified rubber dilatometrically; that of corresponding vulcanizates was
assessed by way of the stress-relaxation of test pieces maintained at a constant
extension of 150%. It is known that the time required for the stress in a
vulcanizate from a crystallizable rubber to relax to half its initial value is an
accurate measure of the time required for half of the primary crystallization
to occur. The stress half-life is therefore an inverse measure of the rate of
crystallization. Typical stress-relaxation curves for vulcanizates based upon
isomerized and unmodified rubbers are shown in Fig. 14.5. The effect of
cis-trans isomerization is evident. The details of the vulcanizate formulation
are given in the caption to this diagram; a so-called 'sulphurless' thiuram
polysulphide vulcanizing system was used, because the type of vulcanizate
obtained using this system is known to be more prone to crystallization than
are conventional sulphur vulcanizates.
1.0 r--~"""'''''--'-~'--''''''-_
o~----------~~--------~~--------~~----------~
101 103 104 105
Time I minutes
Figure 14.5 Stress-relaxation curves for vulcanizates based upon isomerized (points e) and
unmodified (points .) natural rubbers (Cunneen and Shipley [29]). Vulcanizate formulation
(parts by mass): rubber 100, tetramethylthiuram disulphide 4, zinc oxide 5, stearic acid 2.
Vulcanization conditions: 40 minutes at 14O'C.
14.4.2.3 Effectiveness of various carbothiolic acids in facilitating cis-trans

isomerization, as judged by vulcanizate crystallization half-life
A few of the many results reported by Cunneen and Shipley [29] are
reproduced in Table 14.5. This shows the effectiveness of various carbothiolic
acids in facilitating the cis-trans isomerization of natural rubber in latex
form, as judged by the crystallization half-lives of vulcanizates. The effect of
treatment with I-naphthalenecarbothiolic acid as an isomerizing agent is
evident from these results. For both this acid and benzenecarbothiolic acid,
the lowest concentration at which the isomerizing effect becomes discernible
is of the order of one molecule per 400000 daltons of polymer molecular
mass. This is the principal reason for rejecting the alternative explanation
for suppression of crystallization, outlined in section 14.4.1 above. The
hydroxycarbothiolic acids are less effective as isomerizing agents than are
the corresponding non-hydroxy acids. One explanation is that the former
are more soluble in the aqueous phases of the latex than are the latter; so
for any given overall concentration in the latex, the equilibrium concentration
in the rubber phase is lower. However, it is not sufficient that the carbothiolic
acid should preferentially partition in the rubber phase; n-heptadecanecar-
bothiolic acid is not very effective as a cis-trans isomerizing agent, although
it probably partitions almost exclusively in the rubber phase.
A further phenomenon reported by Cunneen and Shipley [29] concerns
the effect ofthe presence of an inert solvent such as benzene upon the efficiency
of cis-trans isomerization by carbothiolic acids. It appears that the more solvent
Table 14.5 Effectiveness of various thiolic acids as cis-trans isomerizing agents for natural rubber in latex form using 20% m/m ascaridole
on carbothiolic acid as catalyst, as judged by effect upon crystallization half-life of vulcanizate (Cunneen and Shipley [29])
Formulation for vulcanizates (parts by mass):
Rubber 100
Tetramethylthiuram disulphide 4
Zinc oxide 5
Stearic acid 2
Vulcanization conditions: 40 minutes at 140°C.
Molecular Amount used Crystallization

mass of for cis-trans half-life
acid isomerization at -26°C
Thiolic acid Structure of acid I dalton I pphr I days
None 0.09
n-heptadecanecarbothiolic acid CH3(CH2)16·CO.SH 301 0.16 0.42
Benzenecarbothiolic acid 138 0.08 10
I "', CO.SH
I I
V-·
" '--' "
o-hydroxybenzenecarbothiolic acid 154 0.32 8.3
I
':" I COSH
'--"
I-naphthalenecarbothiolic acid
0-
" " CO.SH
188 0.109 ca. 69
OH
I-hydroxy-2-naphthalenecarbo-
COSH
thiolic acid 204 0.32 14
OH
2-hydroxy-3-naphthalenecarbo-
COSH
thiolic acid 204 0.32 9.7
Cyciization of natural rubber in latex form 543
is present, the more efficient is the isomerization. The reason for this may be
that the carbothiolic acid and catalyst become more uniformly distributed in
the latex particles as the amount of solvent carrier is increased. Although the
overall extent of cis-trans isomerization may be the same for a given level of
thiolic acid and catalyst, the distribution of isomerized sites in the presence of
solvent may be such that the effect upon subsequent crystallization is greater.
Further information concerning the technological properties of cis-trans
isomerized natural rubber is available in a paper by Cunneen et al. [30].
14.5 CYCLIZATION OF NATURAL RUBBER IN LATEX FORM
14.5.1 Introduction
Cyclized rubber is the name given to the hard, resinous substance which is
obtained from natural rubber by treatment with strong acids of the
conventional type, and with acids of the Lewis type, such as tin(IV) chloride.
Cyclization is a reaction which can be carried out on natural rubber in
solution and in latex form. In the latter case, the use of Lewis acids is not
feasible. Thus the cyclization of natural rubber in latex form is effected by
treatment with conventional strong acids, in particular, with sulphuric acid.
In outline, the process consists in stabilizing the latex against acidification,
adding sulphuric acid, heating the latex, and then separating the cyclized
rubber from the latex. Apart from eliminating any requirement for expensive
and hazardous solvents, the advantages which have been claimed for the
cyclization of natural rubber in latex form compared with cyclization in
solution include the following:
l. elimination of the de polymerization step which is necessary (usually
achieved by means of mechanically-induced degradation) if a tractable
rubber solution is to be obtained;
2. the possibility of co-coagulating cyclized rubber latex with uncyclized latex
to give an intimate mixture of cyclized and uncyclized rubber, which can
then be used as a masterbatch for incorporating cyclized rubber into
natural rubber compounds.
Cyclized natural rubber has been of industrial interest for several applications,
the most important of which have been as a hardening and reinforcing resin
for rubber compounds, and as a shoe-soling material. In applications such
as these, cyclized natural rubber has now been largely superseded by
high-styrene styrene-butadiene copolymer resins.
14.5.2 Preparation of cyclized natural rubber from natural rubber latex

In the process described by Edwards [31], centrifuged natural rubber latex
of at least 60% m/m dry rubber content is colloidally stabilized against
acidification, and then transferred to a jacketed vessel which is lined with

vitreous enamel and equipped with a stirrer. Sufficient of an industrial grade
of concentrated (ca. 98%) sulphuric acid is added with continual stirring to
raise the temperature to ca. l00·C. This temperature is maintained for 21/2
hours, after which time cyclization should be complete. If desired, the extent
of cyclization can be ascertained by determining the specific gravity of the
reaction mixture. The latex is then cooled to 50·C, and an equal mass of
unmodified natural rubber added in the form of uncyclized, but suitably-stabilized
latex. The blended latex is then coagulated by pouring into boiling water. A
minimum of four volumes of water is required for each volume of blended
latex. The product is a co-coagulum of equal amounts of cyclized and
uncyclized natural rubber. It can be washed and processed on the conventional
type of creping or washing mill used in the production of solid natural rubber
from natural rubber latex. After thorough washing, the product is thinly
sheeted and allowed to soak for 24 hours in 2% m/v aqueous ammonia
solution to neutralize any residual acid. It is then washed again in clean
water, and is air-dried at a maximum temperature of 40·C. Part at least of
the uncyclized component of the blend can be derived from centrifuge
washings, and from the skim latices obtained as by-products of the various
processes used for the concentration of natural rubber latex.
Suitable latex stabilizers for this process include fatty-acid ethoxylates and
fatty-amine ethoxylates. Fatty-acid ethoxylates are usually considered to be
non-ionogenic (see section 16.7.6 of Chapter 16 (Volume 3». However, at the
very low pH values which prevail in reaction systems for the cyclization of
natural rubber latex, oxonium cations may form because some of the oxygen
atoms of the ethoxylate chain become protonated. Fatty-amine ethoxylates
are undoubtedly cationic at low pH, because substituted ammonium ions
form by protonation of the nitrogen atom of the amine group. A typical
cyclization formulation for natural rubber in latex form comprises 7.5 pphr
of ethoxylate colloid stabilizer, and 100 pphr of 98% sulphuric acid.
14.5.3 Mechanism and kinetics of cyclization of natural rubber in latex, and

structure of the product
Conflicting views have been held concerning the structure of cyclized natural
rubber and the mechanism whereby it is formed. During the cyclization
reaction, resinification occurs, and there is partial loss of olefinic unsaturation,
but no change in the empirical formula of CsHs. One of the principal matters
of controversy has been the extent to which the olefinic unsaturation can be
reduced if the reaction is carried on indefinitely. This aspect has been reviewed
by Lee, Scanlan and Watson [32]; the reader is referred to their paper for
a detailed discussion. Various methods have been used for the determination
of residual unsaturation in cyclized natural rubber, most of them being suspect
to a lesser or greater extent. Of the procedures which they investigated, Lee,
Cyciization of natural rubber in latex form 545
Scanlan and Watson considered the most satisfactory to be reaction with

excess peroxobenzoic acid, followed by estimation, by iodine titration, of the
residual peroxobenzoic acid which had not added to the olefinic double bonds.
As regards the structure of cyclized natural rubber, van Veersen [33] has
postulated that it comprises a series of condensed polycyclohexyl groups
which are connected by shorter or longer polyisoprene chains, depending
upon the extent of cyclization, as shown in Structure IV. Consistent with
this structure, van Veersen claimed to have shown by hydrochlorination
studies that cyclized natural rubber contains two types of olefinic unsaturation;
these he identified as being associated with the polyisoprene chains and the
condensed polycyclohexyl groups respectively. van Veersen considers the
mechanism of cyclization to be analogous to a type of ionic polymerization.
IV
The polymeric products obtained by Gordon [34] from the latices resulting
from experiments on the cyclization of natural rubber in latex form were
found to be readily soluble in rubber solvents. It was therefore inferred that
they were not crosslinked, although solubility may have been facilitated to
some extent by concurrent oxidative degradation. Gordon [34,35] has
reported results for the kinetics of the cyclization reaction using sulphuric
acid as the cyclizing reagent. The progress of the reaction was followed
dilatometrically. Typical curves reported by Gordon [35] for the variation
of extent of reaction with time of reaction are shown in Fig. 14.6. The
occurrence of an initial period of reduced rate was attributed to the need for
the sulphuric acid to diffuse into the rubber particles before the maximum
rate of reaction could be achieved. According to Gordon [34], the rate of
cyclization correlates well with the Hammett acidity function, H o' for the
aqueous phase of the reaction system. This function is defined as
(14.1 )
1.0
~ 0.8
2
~
g,o 0.6
.s
tJ)
CI
c
.;::
'0
c 0.4
o
:u
~
~
~
50 100 150 250

Time I minute
Figure 14.6 Variation of extent of reaction with time of reaction during acid-induced cyclization
of natural rubber in latex fonn (Gordon [35]). Points. refer to a system containing 16.3% mjrn
rubber and 78.1% mjrn sulphuric acid reacted at 65.4'C; points. refer to a diluted system
containing 70.9% mjrn sulphuric acid reacted at 84. 1'c. The natural rubber latex was colloidally
stabilized by addition of a non-ionogenic ethoxylate.
where the as denote activities, the cs concentrations, and subscripts B, BH +

refer respectively to a Brensted base and its conjugate acid. The first term
on the right-hand side of equation (14.1) is equal to pK. for the conjugate
acid. Gordon regards this correlation between Ho and the rate of cyciization
as precluding a chain mechanism for the reaction, suggesting instead a
mechanism which involves proton transfer. Gordon also attaches significance
to the observation of Fisher and McColm [36] that the maximum extent to
which the olefinic unsaturation of natural rubber can be reduced by latex
cyclization corresponds to ca. 43% of that originally present in the rubber,
i.e., the minimum product-unsaturation attainable by latex cyciization is ca.
57% of that originally present in the rubber. Gordon argues that, if cyciization
proceeds by a mechanism which involves the interaction of adjacent but
isolated pairs of isoprene units, then a certain proportion of those units will
be left as 'widows', unable to interact because they have no partners. It can
be shown that, if all the isoprene units have an equal chance of reacting,
then the proportion of unreacted units left at the completion of the reaction
Cyclization of natural rubber in latex form 547
is e- 2 • Taking into account that the formation of each cyclohexene ring

results in the loss of one olefinic double bond of the isoprene units from
which it was formed, the maximum extent to which the unsaturation of the
product can be reduced should correspond to a residual unsaturation equal
to a fraction 1/2 (1 - e - 2) + e - 2 of the original, i.e., 56.8%. This agrees closely
with the observation of Fisher and McColm. On the basis of these
considerations, Gordon proposes the following proton-transfer mechanism
for the cyclization reaction:
fast
protonation
fast
transfer
CH 2-CH 2
-H+ / \ +
CH 2 C -CH3
~w \ /
deprotonation
• • • 2 I
-CH -C--CH-CH _
2 •••
CH 3
Further information concerning the kinetics of acid-induced cyclization of

natural rubber in latex is available in the second paper by Gordon [35] cited
above.
Having carefully investigated the nature of the products obtained by the
cyclization of natural rubber in solution, Lee, Scanlan and Watson [32] have
concluded that the residual unsaturation can fall considerably below 57%
of the original. Their lowest proportion is ca. 18%. They conclude that,
although the Gordon [34] concept of isolated pairs of isoprene units cyclizing
to leave unattached 'widows' must be abandoned, his reaction mechanism
involving protonation and deprotonation is still inherently plausible. However,
it must now be assumed that cyclization and condensation can proceed
beyond the isolated pairs envisaged by Gordon, and that unreacted units do
not become isolated and incapable of reaction. Lee, Scanlan and Watson
attach significance to the absence from the infra-red spectrum of cyclized
rubber of an absorption at wave number 830 cm - 1. This absorption is present
in the spectrum of uncyclized natural rubber. It is attributed to C-H
out-of-plane deformations of the trialkyl-substituted carbon-carbon double
bonds. It was therefore assumed that the deprotonation stage of the
cyclization reaction occurs in such a way as to form olefinic double bonds
of the tetraalkyl-substituted type, analogous to those postulated In the

Gordon mechanism reproduced above.
14.6 HALOGENATION OF OLEFINICALLY-UNSATURATED

POLYMERS IN LATEX FORM
Chlorinated natural rubber has for some years been used industrially,
principally in applications such as paints and adhesives. It is usually prepared
by passing chlorine into a dilute solution of the rubber in an organic solvent.
However, interest has also been shown in the possibility of chlorinating
natural rubber and other olefinically-unsaturated polymers in latex form
using molecular chlorine. Several advantages accrue from carrying out the
reaction in this way, the principal of which are as follows:
1. The initial molecular mass of the polymer is much higher than for that
used in the solution process, because there is no requirement for any
degradation step in order to obtain a solution of tractable viscosity.
Furthermore, the relatively high initial polymer molecular mass is
substantially retained during chlorination.
2. The reaction can be carried out at a higher polymer concentration than
if carried out in solution.
3. Heat dissipation is a lesser problem than if the reaction is carried out in
solution, because the viscosity of the reaction system is lower.
4. Provided that the reaction system has been appropriately formulated, the
immediate end-product is a colloidally-stable latex of potential technological
interest in its own right.
However it is produced, chlorinated natural rubber obtained using
molecular chlorine as the reagent is a substance of indefinite composition
and constitution. It appears that the chlorination reaction in solution
proceeds in part by way of addition across the olefinic double bonds of the
rubber macromolecule, and in part by way of substitution:
"'-C=C/ + Cl 2 -------.
"'-C-C/
/ "'- /1 1"'-
CI CI
-CH 2 - + CI 2 -------. -CHCI + HCI
A complication which can occur when natural rubber is chlorinated in an
aqueous medium using molecular chlorine is that the product contains a
substantial amount of oxygen. It is probable that the oxygen becomes
incorporated through chlorohydrin formation, the latter occurring because
hypochlorous acid forms in the aqueous medium and then adds to the olefinic
double bonds:
Halogenation of olejinically-unsaturated polymers in latex form 549
"-c=c / + HOCI ---+

"-c-c /
/ "- /1
CIOH
1"-
One adverse consequence of chlorohydrin formation is that the thermal

stability of the product is reduced. It has usually been assumed that the
chlorination of natural rubber by molecular chlorine in aqueous media
proceeds mainly by the addition and substitution reactions which are believed
to occur when the reaction is carried out in solution. Chlorohydrin formation
has been regarded as a minor perturbation. However, as is noted in section
17.3.6.2.2 of Chapter 17 (Volume 3) in connection with the surface chlorination
of thin films of natural rubber produced by latex dipping, analogy with
reactions between simple olefins and chlorine in aqueous media indicates
that chlorohydrination may be the principal outcome. Although this seems
likely for surface chlorination from a contiguous aqueous phase, it is not
clear whether it is likely as the outcome of chlorination in the interior of
latex particles by molecular chlorine which is initially present in the latex
aqueous phase.
One disadvantage associated with the chlorination of natural rubber in
latex form is that the hydrochloric acid liberated (whether as a consequence
of substitution reactions or as a concomitant of chlorohydrination) tends
greatly to reduce the colloid stability of the latex. Some cyclization of the
natural rubber also occurs during both solution and latex reactions, probably
as a consequence of the formation of hydrogen chloride.
van Amerongen [37] has described a latex-chlorination procedure which
largely overcomes both these difficulties. The latex is colloidally stabilized
against acidification, and is then strongly acidified with either concentrated
hydrochloric acid or gaseous hydrogen chloride. Gaseous chlorine is passed
through the latex at a given temperature until the desired degree of
chlorination has been attained. Provided that a suitable stabilizer is employed,
a stable latex of chlorinated natural rubber is produced, from which the
chlorinated rubber can be isolated by colloidal destabilization with methyl
alcohol, or by gently warming the latex whilst adding sodium chloride. The
dried product is usually in the form of a finely-divided white powder. Suitable
colloid stabilizers for this process include non-ionogenic ethoxylates, added to
the extent of ca. 2 pphr, and cationic stabilizers of the quaternary ammonium
type, added to the extent of ca. 3-5 pphr. Specific stabilizers cited by van
Amerongen include an ethoxylate of n-octadecenyl alcohol, n-hexadecyl-
dimethylethylammonium bromide, n-hexadecyltrimethylammonium bromide,
n-hexadecylpyridinium bromide and n-tetradecylpyridinium bromide. Some
chlorination of ethoxylate stabilizers is said to occur concurrently with
chlorination of the rubber, but this does not prevent the formation of stable
latices of chlorinated natural rubber when these substances are used as colloid
stabilizer. The function of the hydrochloric acid in this procedure is twofold.
Firstly, the two ions which it generates suppress the formation of hypochlorous
acid by the common-ion effect, the hydrogen ions suppressing the concentration of
hydroxide ions and the chloride ions reducing any residual tendency for the
formation of hypochlorite anions. Secondly, an acidic medium is created upon
which any hydrogen chloride formed in the course of the chlorination can
have little effect. The initial acidification should be as rapid as possible, so as to
attain maximal latex colloid stability under acid conditions. Although a high level
of acidity is desirable to suppress chlorohydrin formation, complete saturation
of the latex with hydrochloric acid should be avoided, otherwise hydrochlorination
of the rubber tends to occur concurrently with chlorination (see section 14.7 below).
The rate of chlorination depends upon the rate at which chlorine is injected
relative to the rubber content of the reaction system. Figure 14.7 illustrates
the course of a typical latex chlorination reaction. The initial rate of
absorption of chlorine is high. The rate decreases sharply as the chlorine
content of the product approaches 50% m/m. It is generally advisable to
chlorinate as rapidly as possible, so as to avoid, or at least to minimize,
hydrochlorination as a side reaction. Temperature has little effect upon either
the rate or extent of chlorination. It appears that the tendency of the rate of
reaction to be increased by raising the temperature is offset by the reduced
solubility of chlorine in the aqueous phase of the reaction system. The
temperature of chlorination does, however, affect the character of the product
to some extent. Products obtained at temperatures below 10°C, and which
contain more than 48% m/m of chlorine, are usually soluble in solvents such
as chloroform, although crosslin king of the macromolecules occurs with great
~ 80
'#
......
CD 60
~
8.
"040
~
8 20
~
~
.c
o 00~--------~10~------~2~0~------~~=-------~40
Time/hour
Figure 14.7 Variation of extent of reaction with time of reaction during chlorination of
deproteinized natural rubber in latex fonn (45% mjm dry rubber content) at O·C (van Amerongen
[37]). The latex was colloidally stabilized by addition of a non-ionogenic ethoxylate (2 pphr),
and acidified with hydrochloric acid.
Halogenation of olejinically-unsaturated polymers in latex form 551
ease, e.g., by heating. Slow chlorination at temperatures above 50·C produces

discoloured materials which are insoluble in chloroform. The chemical
stability of chlorinated natural rubber increases with increasing chlorine content.
van Amerongen [37] has also succeeded in chlorinating certain synthetic
rubbers in latex form, in particular, styrene-butadiene rubber, polychloroprene
rubber, and acrylonitrile-butadiene rubber. He points out, however, that
some of these polymers contain substantial concentrations of functional
groups which are resistant to chlorination. The ultimate chlorine content of
the products is therefore correspondingly lower. It has also been claimed
that polyvinyl chloride in latex form can be chlorinated.
Baker [38] has described a different procedure for the chlorination of
natural rubber in latex form. His method depends upon the generation of
nascent chlorine in situ by interaction between sodium chlorate and hydrochloric
acid. In simple terms, this interaction can be represented as follows:
ClO; + 5Cl- + 6H+ ----+ 3Cl 2 + 3H 2 0
However, in reality the reaction is undoubtedly more complex than is
indicated by this simple scheme, and the reaction system probably contains
other species beside those indicated. In any event, it is most unlikely that
this scheme can represent anything other than the overall stoichiometry of
the reaction. A typical procedure which makes use of this method is as follows:
l00cm 3 of ammonia-preserved natural rubber latex of 40% mlm dry rubber
content is mixed with 10cm 3 of a sodium alkyl sulphate stabilizer. The
requisite amounts of hydrochloric acid and sodium chlorate are then added
in that order, the former as a 32% mlv aqueous solution and the latter as a
ca. 50% mlv aqueous solution. The addition of the sodium chlorate solution
should be made slowly, say over a period of 1112 hours, during which time
the vessel containing the latex should be surrounded by cold water. After
the addition of the sodium chlorate is complete, the stirring is continued for
a further 1112 hours, and the mixture then allowed to stand for 24 hours at
normal ambient temperature. It is then possible to remove the supernatant
chlorine water by decantation. The wet product is made just alkaline with
dilute aqueous sodium hydroxide solution, and is then filtered, washed
thoroughly with water, air-dried for 2 days, and kept over anhydrous calcium
chloride in a vacuum desiccator. The extent of chlorination, and the character
of the product, depend upon the amount of chlorine which is available; the
latter is determined by the amounts of hydrochloric acid and sodium chlorate
added initially, as is evident from the results reported by Baker and
summarized in Table 14.6. It appears that insolubility of the products in
organic solvents is associated not only with a low chlorine content, but also
with a rise in temperature during the reaction. The products of this process
inevitably contain some oxygen, probably arising from chlorohydrin formation.
Baker [38] has also described a method for brominating natural rubber
in the form of ammonia-preserved latex using bromine water. The latex is
stabilized by addition of saponin (a naturally-occurring hydrocolloid), and

is then de-ammoniated with formaldehyde. The bromine water is added, and
the mixture allowed to stand overnight. The brominated rubber can then be
separated from the reaction mixture by filtration through paper using a
Buchner funnel, and compacted by hand prior to sheeting on a mill.
Brominated natural rubber appears to be less easily oxidized than is
chlorinated natural rubber. It is a reasonably stable substance. Baker found
that halogenated natural rubbers of relatively low halogen content retain
their elasticity, and can be vulcanized by conventional means.
It is noted in section 17.3.6.2.2 of Chapter 17 (Volume 3) that various
organic chlorinating agents have been considered as alternatives to chlorine
water and aqueous sodium hypochlorite-hydrochloric acid mixtures for the
surface chlorination of thin films of natural rubber produced by latex dipping.
It is not known whether any of these has ever been investigated as a reagent
for the chlorination of olefinically-unsaturated polymers in latex form. One
particular organic chlorinating agent which is widely used for surface
chlorination is sodium dichloroisocyanurate (Structure I of Chapter 17),
usually, but not necessarily, in the presence of added acid.
The relative extents to which addition, substitution and cyclization occur
during the chlorination of natural rubber in solution has been investigated
by Kraus and Reynolds [39]. They concluded that chlorination at 30·C
proceeds in three stages. In the first stage, substitution occurs almost
exclusively to the extent of approximately 1 chlorine molecule per isoprene
unit. In the second stage, a further ca. 1.2 molecules of chlorine react both
by substitution and by addition. Finally, when ca. 2.2 molecules of chlorine
per isoprene unit have reacted, the addition reaction ceases abruptly, and all
further reaction occurs by substitution. Some cyclization takes place in the
early stages of the chlorination concurrently with the initial substitution.
Substitutive chlorination is catalysed by oxygen and peroxides, whereas
additive chlorination is catalysed by ultraviolet light. Cyclization is partially
prevented by an increase in the addition reaction. Although these conclusions
Table 14.6 Effect of available chlorine upon composition and character of chlorinated
natural rubber produced by latex chlorination according to the method of Baker
(Baker [38])
A vailable chlorine: Chlorine content of product

atoms per C6Ha unit / % m/m Character of product
3.85 37.4 Buff-coloured insoluble powder

5.77 45.6 White readily-soluble powder
Hydrohalogenation of olejinically-unsaturated polymers 553
were reached from an investigation of the chlorination of natural rubber in

solution, it seems likely that they also apply in a general way to the
chlorination of natural rubber in latex form.
Further information concerning chlorinated natural rubber is available in
two papers by Bloomfield [40,41].
14.7 HYDROHALOGENATION OF OLEFINICALLY-

UNSATURATED POLYMERS IN LATEX FORM
For many years, the addition product of natural rubber and hydrogen
chloride, often referred to loosely as rubber hydrochloride, was of industrial
interest as a material for the manufacture of thin transparent films for
packaging purposes. Conventionally, it is made by passing gaseous hydrogen
chloride through a solution of natural rubber in an organic solvent. It is,
however, possible to carry out the reaction upon natural rubber in latex
form. As with the chlorination of natural rubber, there are in principle several
advantages which accrue from hydrochlorinating the polymer in latex form
rather than in solution. These advantages are essentially those outlined above
in section 14.6 for chlorinated natural rubber.
van Veersen [42] has described a procedure for hydrochlorinating natural
rubber in latex form in which l00cm 3 of 40% mlm ammonia-preserved
natural rubber latex is de-ammoniated by drawing air through it, and is then
diluted and colloidally stabilized against acidification by the addition of
40cm 3 of a 5% m/v aqueous solution of n-hexadecylpyridinium bromide.
Gaseous hydrogen chloride is passed through the latex. Hydrochlorination
of the rubber commences as soon as the aqueous phase is saturated with
hydrogen chloride. Figure 14.8 shows the course of the reaction as it appears
from the results reported by van Veersen. The rate of absorption of hydrogen
chloride is increased by increasing the temperature and pressure of the gas.
van Veersen reports that it is also feasible to use non-ionogenic stabilizers
to protect the latex against colloidal destabilization by acid. A specific
formulation for a reaction system stabilized in this way comprises lOOcm 3
of 40% m/m latex stabilized by addition of 5 cm 3 of a 20% m/v aqueous
solution of an ethoxylate of n-octadecenyl alcohol. After passing hydrogen
chloride through the stabilized latex at normal ambient temperature and
pressure, a product containing 32.7% mlm of chlorine was isolated from the
reaction system by filtration, after first adding sodium chloride to the
hydrochlorinated latex and then heating. An interesting feature of the
hydrochlorination of natural rubber in latex form is that it appears to be
possible to carry the reaction almost to completion (98-99%), contrasting
with solution hydrochlorination, where ca. 90% of the theoretical is the
maximum attainable.
According to van Veersen [42], the rate-determining step in the
hydrochlorination of natural rubber in latex form is the diffusion of hydrogen
40
30
Time/hour
Figure 14.8 Variation of extent of reaction with time of reaction during hydrochlorination
reaction of natural rubber in latex form (van Veersen [42]). Reaction system comprised 40"10 m/m
dry rubber latex to which was added 5 pphr n-hexadecylpyridinium bromide dissolved in 100 pphr
water. Reaction conditions were as follows: points. lOoe, normal atmospheric pressure (ca.
0.1 MPa); points • 60°C, ca. 3 atmospheres pressure (ca. 0.3 MPa); points _ 90°C, ca. 3
atmospheres pressure (ca. 0.3 MPa).
chloride molecules into the rubber particles. On the other hand, Crampsey,
Gordon and Taylor [43] have reported evidence which indicates that the rate-
determining step is the formation of hydrogen chloride ion-pairs in the
aqueous phase. This step can be represented as
Crampsey, Gordon and Taylor consider that some of the protons will be
present as 'solvation' complexes with hydrogen chloride, rather than with water.
The experiments of Crampsey, Gordon and Taylor [43] were carried out
upon natural rubber latex and upon a synthetic polyisoprene latex which
had been prepared by emulsion polymerization using a non-ionogenic
ethoxylate as the colloid stabilizer. This synthetic latex was found to be
completely stable to acidification, and could not be flocculated by dilution
with boiling water. Colloidal destabilization by addition of acetone was the
method used to isolate the hydrochlorinated polymer. The progress of the
hydrochlorination reaction was followed by determination of the specific
gravity of the isolated product. The rates of the hydrochlorination of both
natural rubber and the synthetic polyisoprene were found to be kinetically
of almost zero order with respect to the concentration of unreacted polymer.
Thus under given conditions of temperature and pressure (27"C and 0.1 or
0.2 M Pa), the rate of hydrochlorination was essentially constant over a large
Halogenoalkylation of natural rubber in latex form 555
part of the reaction, provided that the hydrogen chloride was present in
considerable excess. Two interesting anomalies relating to the initial rate of
reaction were reported by Crampsey, Gordon and Taylor. The first is that
for both polymers there seemed to be a very fast initial hydrochlorination.
This was interpreted as being a reaction which occurs in the surface regions
of the particles. The extent of this surface region was thought to be limited
to ca. 1.5 nm, because of the low rate of diffusion of unreacted segments of
rubber macromolecules from the interior of the particle to its surface. The
second anomaly relates to the hydrochlorination of natural rubber latex. A
period of retardation was observed, associated with the presence of retarding
substances in this latex. Furthermore, the extent of the initial fast reaction
was less with natural rubber latex than with the synthetic polyisoprene latex.
It has been known for some years that the product obtained by
hydrochlorinating natural rubber is stereoregular, and that this is the reason
why this substance will crystallize to give tough films. The stereo regularity
is generally agreed to arise from a conjunction of two circumstances: the
addition of hydrogen chloride to the olefinic double bonds occurs in
accordance with the Markownikoff rule; and the asymmetric centres thereby
introduced are alternatively (R) and (S) configurations, so that sequences of
repeat units are effectively syndiotactic. The molecular structure of
hydrochlorinated natural rubber can therefore be represented as shown in
Structure V. It is clear that the addition of hydrogen chloride to natural
CH 3 CI
I I
... -C-CH2-CH2-CH2-C-CH2-CH2-CH2-···
I I
CI CH 3
V
rubber must take place in a stepwise manner, proceeding from unit to unit
through the polymer chain, and that the entry of one hydrogen chloride
molecule into the rubber chain must in some way influence the manner in
which the next enters. However, the precise manner in which this occurs is
not clear at present. For a discussion of this matter, the reader is referred to
a paper by Gordon and Taylor [44].
14.8 HALOGENOALKYLATION OF NATURAL RUBBER IN LATEX

FORM
One defect of natural rubber is its ready flammability. This has important
safety implications for certain applications, notably for application in the
form oflatex foam rubber. Although this product is bulky, it presents a large
ratio of surface area to volume of combustible material. There has long been
interest in the possibility of modifying natural rubber so as to reduce its
flammability whilst retaining its strength and elasticity. Increasing the halogen
content does reduce flammability, but halogenation also tends to convert

natural rubber from a rubbery, elastic material to a non-elastic, resinous
substance of limited thermal stability. It appears, however, that the process
of halogenoalkylation can give an elastic, vulcanizable polymer which also
has some degree of flame-resistance, or at least of flame-retardance. In the
process of halogenoalkylation, a polyhalogenoalkane is added across the
olefinic double bond of the rubber macromolecule. For the case of a
tetrahalogenomethane, CX 4 , where X is a halogen atom, the addition can
be represented generally as follows:
,,/ ,,/
'"
C-C + CX4 --+ C-C
/" /,X CX 3
It should be noted that, in this connection, the term 'polyhalogenoalkane' is
used loosely and incorrectly to denote a polyhalogeno derivative of any
aliphatic hydrocarbon, whether saturated or not. The halogenoalkylation
reaction appears to proceed by a free-radical mechanism, being facilitated
by the presence offree-radical generators in the reaction system. A convenient
method for carrying out the reaction is to have the natural rubber in the
form of a latex, to swell the particles with the requisite amount of
poly hal ogenoal kane, and to initiate the reaction by means of a free-radical
generator such as a combination of tert-butyl hydroperoxide and tetra-
ethylenepentamine. The reactivities of polyhalogenoalkanes in this reaction
vary widely. Table 14.7 shows results reported by Cockbain et al. [45] for
the proportions of the added polyhalogenoalkane combined with the rubber
under comparable reaction conditions, using tert-butyl hydroperoxide and
tetraethylenepentamine as initiator. It appears that the most reactive compounds
are those in which all the hydrogen atoms of the parent hydrocarbon have
been replaced by halogen atoms, and in which at least one of the halogen
atoms is bromine. The necessity for at least one bromine atom in the
polyhalogenoalkane is attributed to the activation energy for homolytic
scission of the carbon-bromine bond being significantly less than that for
homolytic scission of the carbon -chlorine bond. The consequence is that the
propagation reaction
,,/ ,,/
/0,''
c=c + CX 4
CX 3
--+
/, ,,,
C-C
X CX 3
+ 'CX 3
is facilitated if at least one bromine atom is present in the CX 4 molecllie. The

polyhalogenoalkane selected by Cockbain et al. for detailed investigation
was trichlorobromomethane, the reasons being relatively low cost, high
reactivity, and ease of absorption into the rubber particles when added to
The character of the modified polymer obtained from natural rubber latex
Halogenoalkylation of natural rubber in latex form 557
Table 14.7 Proportion of added polyhalogenoalkane combined with natural rubber

in latex form under similar halogenoalkylation reaction conditions (Cockbain et al. [45])
Initiation system: tert-butyl hydroperoxide + tetraethylenepentamine.
Reaction system and conditions not specified.
Proportion of added
polyhalogenoalkane
combined with rubber
Polyhalogenoalkane Structure /%
Carbon tetrachloride CCI 4 10
Carbon tetrabromide CBr 4 97
Trichlorobromomethane CCI3Br 95
Bromoform CHBr 3 40
Tribromofluoromethane CFBr 3 80
Dibromodifluoromethane CF 2 Br 2 53
1,2-dibromotetrafluoroethane CF 2Br.CF 2Br 3
1,1,2,2-tetrabromoethane CHBr 2 ·CHBr2 45
Ethyl bromide CH 3·CHBr2 10
1,3-dibromopropane CH 2Br.CH 2·CH 2 Br <10
1,2,3,4-tetrabromobutane CH 2Br.CHBr.CHBr.CH 2Br 45
1,2-dibromoethylene CHBr:CHBr 10
Bromobenzene C 6 H s ·Br 10
which has been halogenoalkylated by reaction with trichlorobromomethane

depends upon the amount of the latter combined with the rubber. The
maximum proportion attainable experimentally is such that the original
rubber constitutes ca. 25% m/m of the final product. The proportion
corresponds almost exactly to the addition of one trichlorobromomethane
molecule to each double bond in the rubber macromolecule. This particular
adduct is resinous in character. To obtain a modified latex which is
film-forming at normal ambient temperatures, and the polymer of which
retains some elasticity, it is necessary that the proportion of trichlorob-
romomethane in the modified polymer should not exceed ca. 30% m/m.
The two formulations of halogenoalkylation reaction systems given in
Table 14.8 were used by Cockbain et at. [45] to obtain modified natural
rubbers which contained respectively 20% and 65% m/m combined
trichlorobromomethane. The reaction system for the preparation of the latex
of modified rubber containing 20% m/m trichlorobromomethane is made up
as follows: The tert-butyl hydro peroxide initiator is dissolved in the
trichlorobromomethane, and the solution coarsely emulsified with the
remaining ingredients of Part B. The emulsion is then added with stirring
to the latex which has first been colloidally stabilized by addition of sodium
n-dodecyl sulphate. Finally, the tetraethylenepentamine solution is added.
The storage stability of the latex obtained at the end of the reaction is said
to be adequate, provided that the total solids content does not exceed
55% m/m. The reaction system for the preparation of the latex of modified
Table 14.8 Formulations for halogenoalkylation reaction systems for production of

latices of modified natural rubbers which contain respectively 20% and 65% m/m
combined trichlorobromomethane (Cockbain et al. [45])
Parts by mass
20% m/m CCIJBr 65% m/m CCIJBr
Ingredient Dry Actual Dry Actual
Part A
Natural rubber (as 60% m/m
ammonia-preserved latex) 100 167 100 167
Sodium n-dodecyl sulphate (as
20% m/m aqueous solution) 5 2 to
Part B
Trichlorobromomethane 26 26 196 196
tert-butyl hydroperoxide 0.87 0.87
Sodium n-dodecyl sulphate
(as 20% m/m aqueous solution) 5 2 to
Ammonia (as 2% aqueous m/m
solution) 0.66 33 2.16 108
Hydrazine hydrate 3.92 3.92
Part C
Tetraethylenepentamine (as 10% m/m
aqueous solution) 0.35 3.5
rubber containing 65% m/m trichlorobromomethane is made up by adding

a coarse emulsion of Part B to Part A. Hydrazine is rather less effective as
an initiator than is the combination of tert-butyl hydroperoxide and
tetraethylenepentamine. Typically, the system is allowed to react for 16 hours
at normal ambient temperature. The temperature should not be allowed to
rise above 30·C; if it does, then the colloid stability of the latex may decrease.
Increasing levels of combined trichlorobromomethane increase the flame-
resistance of the rubber, but also decrease the tensile strength of vulcanizates
from the rubber. To be self-extinguishing, the modified rubber must contain
at least 15% m/m of the modifying polyhalogenoalkane. At these levels, the
rubber macromolecule retains sufficient of its initial olefinic unsaturation to
permit vulcanization to be effected by conventional methods.
14.9 DEPOLYMERIZATION AND OXIDATION OF NATURAL

RUBBER IN LATEX FORM
The matters of the depolymerization and oxidation of natural rubber in latex
form are considered together because, although it is easy to distinguish
Depolymerization and oxidation of natural rubber in latex form 559
between the two conceptually, in practice they are closely related. By

depolymerization is meant reduction in the chain-length of the rubber
macromolecules without affecting the essential nature of the polymer. By
oxidation is meant a chemical reaction iIi which oxygen becomes combined
with the rubber macromolecule. In practice, oxidation is usually accompanied
by reduction in the chain-length of the rubber macromolecule, and controlled
oxidation of natural rubber provides the most convenient method for effecting
partial depolymerization. The reason why oxidation is usually accompanied
by reduction in the chain-length of the rubber macromolecule is that the
oxidation of natural rubber is a free-radical chain reaction, one step of which
is occasionally accompanied by scission of the main chain of the polymer,
although this is not evident from the outline sequence of reactions which is
usually postulated to explain the combination of oxygen with the rubber
macromolecules.
Limited interest has been shown in the depolymerization and oxidation
of natural rubber in latex form. A major interest in oxidized natural rubber
latex has been as an adhesive base for materials such as leather. The
oxidized/depolymerized polymer is softer, more plastic and more tacky, than
the initial unoxidized ·rubber. Surfaces treated with oxidized natural rubber
latex readily adhere to each other, and prolonged pressure is 'Dot necessary
in order to achieve satisfactory union.
The procedure described by Bloomfield, Farmer and Schidrowitz [46] for
the concurrent oxidation and depolymerization of natural rubber in latex
form is typical. In this procedure, oxidation is effected by treating the latex
with hydrogen peroxide solution. 60% mlm ammonia-preserved centrifuged
natural rubber latex is diluted with water to 30% mlm dry rubber content,
and is colloidally stabilized by the addition of a small amount of ammonium
caseinate; ca. 1 g of casein per l00cm 3 of the original latex is suggested. The
latex is then de-ammoniated as rapidly as possible by drawing air through
it. 150 cm 3 of the de-ammoniated latex is placed in a vessel which is sufficiently
large to accommodate the froth generated subsequently. 12.5 cm 3 of 20-
volume hydrogen peroxide solution is added. The mixture is then slowly but
efficiently stirred and warmed on a steam bath until the frothing subsides;
this takes ca. 25-30 minutes. A further 12.5 cm 3 of 20-volume hydrogen
peroxide solution is added, and the mixture heated for 3-4 hours on a steam
bath. Constant stirring must be maintained. If any tendency to colloidal
destabilization is observed, then a little dilute ammonia solution should be
added. Finally, the mixture is cooled, strained and colloidally stabilized and
preserved by the addition of ammonia. A similar process is described in a
patent to Stevens [47].
In recent years, there has been renewed interest in the production of
depolymerized natural rubber obtained by oxidation of the polymer in latex
form. An important motivation for these endeavours has been a desire to
produce polyisoprenes of reduced molecular mass and viscosity which can
be used as plasticizers for natural rubber and other diene rubbers in the dry
state; polyisoprene plasticizers suitable for use with dry natural rubber are
particularly desirable because of the variability of the molecular mass and
molecular-mass distribution of natural rubber, and the absence of methods
for controlling these characteristics of the polymer as it is formed in the
Hevea brasiliensis tree. Gazeley, Menthe and Morris [48] have recently
described a novel process for effecting oxidative depolymerization of natural
rubber in latex form using various combinations of inorganic oxidizing and
reducing agents. Typical components of suitable redox combinations are
sodium chlorite, sodium nitrite and hydrogen peroxide. If the final product
is required as a latex, then a non-ionogenic stabilizer is added to maintain
colloid stability. The degradation is effected at a somewhat elevated temperature
in the range 30-60°C. The extent of reaction, and hence the molecular-mass
distribution of the product, is determined by the amounts of reagents used
and the reaction time and temperature. Coomarasamy et at. [49] have
described a similar process which uses a mixture of hydrogen peroxide and
sodium hypochlorite as the oxidizing reagent. The claim is made that the
effective oxidizing agent is singlet molecular oxygen, rather than the more
usual triplet molecular oxygen. Triplet molecular oxygen has the structure
of a dioxygen diradical, appropriately represented symbolically as ·0-0·.
Singlet molecular oxygen is double-bonded dioxygen; its structure is
analogous to that of ethylene, and is appropriately represented symbolically
as 0=0. Certainly interaction between hydrogen peroxide and hypo-
chlorite anions is recognized as producing singlet molecular oxygen by the
reaction
In the process described by Coomarasamy et at., aqueous hydrogen peroxide

solution is added to stabilized natural rubber latex contained in a reaction
flask which is either tinted or covered with aluminium foil to prevent light
penetration. Over a period of ca. 1 hour, aqueous sodium hypochlorite
solution is added dropwise below the surface of the latex whilst the latter is
stirred. The temperature is maintained within the range 0-30°C. The latex
is stirred for further period of 5 hours after the addition of the sodium
hypochlorite solution is complete. Table 14.9 shows results reported by
Coomarasamy et at. for the effect of level of hydrogen peroxide and sodium
hypochlorite upon the intrinsic viscosity and calculated viscosity-average
molecular mass of the rubber obtained from the treated latex. Although some
of the details of the treatments and of the results are unclear (see footnotes
to Table 14.9), the trends indicated by these results are clear. Drastic
reductions in rubber molecular mass are possible by this type of treatment.
Increasing levels of both reagents leads to reduced polymer molecular mass,
but increasing level of sodium hypochlorite appears to be more effective than
Latices containing interpenetrating polymer networks 561
Table 14.9 Effect of level of hydrogen peroxide and sodium hypochlorite upon
intrinsic viscosity and calculated viscosity-average molecular mass of rubber obtained
from natural rubber latex treated with these reagents (Coomarasamy et al. [49])
Level of reagent
/ % m/m * on latex Properties of rubber obtained from treated latex
Uscosity-average
Hydrogen Sodium Intrinsic viscosity molecular mass
peroxide hypochlorite / l00cm 3 g-It / dalton
o o 5.25 1062000
10 5 2.60 370000
10 10 2.20 288500
40 10 1.47 157500
10 30 0.96 83000
16.7 50 0.15 5100
• These levels probably refer to aqueous solutions of the reagents. The concentrations of the
aqueous reagent solutions are not given.
t The units of intrinsic viscosity are not specified, but are assumed to be the usual units of
lOOcm 3 g-I; neither is the solvent stated, nor the values of K and ex used to obtain values of
Nt. from those of intrinsic viscosity.
increasing level of hydrogen peroxide. Coomarasamy et al. also report that

the reduction in polymer molecular mass increases as the pH of the latex is
reduced. This indicates that it is un-ionized hydrogen peroxide, and possibly
also un-ionized hydrogen hypochlorite, rather than peroxide and hypochlorite
anions, which interact to form singlet molecular oxygen. It may be that
oxidation by singlet molecular oxygen also occurs in the process described
by Gazeley, Mente and Morris.
14.10 LATICES CONTAINING INTERPENETRATING POLYMER

NETWORKS
An interpenetrating polymer network is defined as an intimate mixture of two
or more crosslinked polymers of different types which is such that there are
no covalent bonds between the networks of different types. Such systems
offer a third way - the other two being mechanical blending and copolymerization
- in which polymer sequences of different types can be brought together in
intimate mixture. In consequence, they offer the possibility of forming
composite polymeric materials having mechanical properties not attainable
by other means. Any thermodynamic tendency for the different types of
polymer sequence in an interpenetrating polymer network to separate is
inhibited. by the inevitable molecular constraints resulting from interpenetration.

Closely related are mixtures known as semi-interpenetrating polymer networks.
These are intimate mixtures of two types of polymer sequence, one of which
is crosslinked and the other not. Again, there are no covalent bonds between
the polymer sequences of different types. Interpenetrating polymer networks
can be prepared either sequentially or simultaneously. If prepared sequentially,
one of the networks is synthesized or crosslinked in the presence of the other;
if prepared simultaneously, both are synthesized and crosslinked together
using non-interfering modes of polymerization.
Interpenetrating polymer networks can be produced sequentially with one
polymer in the form of a latex. The products are then known as latex
interpenetrating polymer networks. They can be produced by a two-stage
process. In the first stage, one of the polymer types is formed as a latex
containing particles of crosslinked polymer. In the second stage, another
monomer is added with crosslinker, and then polymerized within or upon
the particles of the first polymer. Care must be taken to avoid the nucleation
of new particles at the second stage. Hourston and Romaine [50] have
described the preparation and properties of certain latex interpenetrating
polymer networks based upon natural rubber latex. The second polymer was
either polystyrene or polymethyl methacrylate, the crosslinker divinylbenzene,
and the initiation system a redox combination comprising 0.3 pphr of
tert-butyl hydroperoxide and 0.117 pphr oftetraethylenepentamine. Ammonium
oleate was used as a supplementary colloid stabilizer; it was formed in situ
by adding 0.5 pphr oleic acid. The procedure was to dilute 60% mlm natural
rubber latex concentrate to ca. 47% mlm rubber content by addition of
2% mlv aqueous ammonia solution, deoxygenate the diluted latex by
bubbling nitrogen through it whilst maintaining the pH at 8.5 by addition
of ammonia as necessary, raise the temperature to 50·C, then add the
monomer, crosslinker, tert-butyl hydroperoxide and stabilizer dropwise over
a period of 15 minutes. After a further 15 minutes, the tetraethylenepent-
amine was added as a 10% mlv aqueous solution. The polymerization
reaction was then allowed to proceed for 24 hours under continuous stirring.
Whether polymer mixtures obtained in this way are to be regarded as
interpenetrating polymer networks, semi-interpenetrating polymer networks,
or a mixture of the two, depends upon the gel content of the initial natural
rubber component. It is unlikely that such polymer mixtures are entirely
devoid of covalent bonds between the two types of polymer sequence which
they contain.
A comprehensive review of interpenetrating polymer networks and related
polymer compositions, including, inter alia, latex interpenetrating polymer
networks, is available in a book by Sperling [51] published in 1981. Further
detailed information concerning latex interpenetrating polymer networks is
available in papers by Hourston and Satgurunathan [52] and by Hourston,
Satgurunathan and Varma [53-55].
Modifications involving only the surface of the latex particles 563
14.11 MODIFICATIONS INVOLVING ONLY THE SURFACE OF THE

LATEX PARTICLES
14.11.1 Reversal of charge of latex particles

All the chemical modifications of polymers in latex form which have been
considered in this chapter so far, and also in Chapter 13, are such that the
principal effect is an alteration of the chemical constitution or structure of
the polymer which makes up the bulk of the latex particles. This chapter
concludes with brief consideration of certain chemical modifications of latex
polymers which are such that the effect is confined wholly or mainly to the
interfacial region between the polymer particle and the dispersion medium.
The concern will be exclusively with aqueous latices. The first type of surface
modification to be considered is changes intended to reverse the polarity of
the effective electric charge bound to the surface of the particles. It is also
necessary to consider the matter of possible interactions between the
constituents of the interfacial region and certain ions which may be present
in the aqueous phase; the nature and extent of such interactions can have a
determining influence upon whether or not a colloidally-stable latex can be
obtained. For this reason, the matter of modification of the interfacial region
of an aqueous latex has to be considered in relation to the composition of
the aqueous phase which is contiguous to the modified interface. Of the
various aspects of the composition of the aqueous phase which may be
relevant in this connection, the most important is that of acidity or alkalinity.
In principle, it is possible to distinguish four types of latex insofar as the
relationship between the pH of the aqueous phase and the polarity of the
electric charge bound at the particle surface is concerned. These are:
1. latices such that the aqueous phase is alkaline and the effective bound
charge carried by the particles is negative;
2. latices such that the aqueous phase is alkaline and the effective bound
charge carried by the particles is positive;
3. latices such that the aqueous phase is acidic and the effective bound charge
carried by the particles is positive;
4. latices such that the aqueous phase is acidic and the effective bound charge
carried by the particles is negative.
The majority of latices of industrial and academic interest are of type (I); it
is latices of this type which are the subject of most of this book. Considerable
industrial interest has been shown in latices of type (2) in the past. The
reasons for this interest are discussed in section 19.7 of Chapter 19 (Volume
3). The generic name positex has been coined for these latices. Latices of type
(3) include various synthetic cationic latices which are prepared using cationic
colloid stabilizers and acidic aqueous phases (see section t t .9. t of Chapter
11). Alternatively, latices of this type can in principle be prepared by rapid
acidification of an anionic latex which has an alkaline aqueous phase and is

stabilized by an amphoteric stabilizer such as an adsorbed protein. The
objective is to reduce the pH of the aqueous phase to a value below the
isoelectric point of the stabilizer so rapidly that there is insufficient time for
the latex to undergo gross colloidal destabilization when it is in a condition
oflow colloid stability. The effect of acidification is then to reverse the effective
polarity of the particles, as well as to lower the pH of the latex aqueous
phase. Latices of type (4) are of industrial interest. They are formed when
synthetic latices of polymers functionalized with acidic comonomers, such as
carboxylic acids, are produced (see section to.3.3 of Chapter to and 11.8 of
Chapter 11); they are sometimes, but not usually, used in the acidic condition.
By analogy with the generic name positex for latices of type (2), the name
negatex has been proposed for latices of type (4).
The principal concern here is with the preparation of latices of type (2) by
addition of a considerable excess of a quaternary ammonium surface-active
substance to a latex of type (1). The cation of the surface-active substance is
strongly adsorbed at the surface of the particles in an anionic latex. The
effect of small additions is to reduce the surface charge density of effective
bound negative charge. The colloid stability of the latex is correspondingly
reduced. If sufficient of the cationic surface-active substance is added, the
adsorption occurs to such an extent that the effective charge of the particles
is changed from negative to positive. The particles are nevertheless colloidally
stable in an alkaline aqueous medium, because, being of the quaternary-
ammonium type, the cations which confer colloid stability do not interact
with the hydroxide ions which are present in the aqueous phase in appreciable
concentration if the pH is high. Were a cationic surface-active substance of
the non-quaternary-ammonium type used, then it is to be expected that the
stabilizing cations would be neutralized by interaction with the hydroxide
ions, and the colloid stability of the latex lost. Typical cationic surface-active
substances suitable for the preparation of latices of this type include
n-hexadecyltrimethylammonium bromide and n-hexadecylpyridinium bromide.
It is necessary to make the addition of excess cationic surface-active substance
rapidly, otherwise the latex may be irreversibly colloidally destabilized. The
objective is to reverse the effective polarity of the particles so rapidly that
there is insufficient time for the latex to undergo gross colloidal destabilization
when it is in a condition of low colloid stability. An alternative procedure
is to pour the latex into an aqueous solution of the cationic surface-active
substance, so that an excess of the latter is always encountered.
Blow [56], who played a major role in the development of cationic latices
of this type, has described a method for investigating the sign and magnitude
ofthe effective charge carried by the latex particles. His apparatus is illustrated
in Fig. 14.9. The principle is to determine the electrophoretic mobility of the
latex particles by the moving-boundary method. Because the latex particles
are usually less dense than the medium in which they are dispersed, it is
necessary that the latex should always be above the aqueous phase with
which it forms a boundary. Furthermore, so that the boundary remains as
sharp as possible, it is desirable that the conductivities of the latex and the
contiguous aqueous phase be as similar as possible. Blow achieved this by
using as the aqueous phase clear serum obtained from a further sample of
latex by filtration. The procedure for using the apparatus is as follows: The
electrodes are removed and the apparatus completely filled with the aqueous
phase, the tap, T, being closed so that the funnel, F, remains empty. Much
of the liquid in the side limbs is then removed by pipette, and the electrodes
replaced in position. Latex is run in slowly from the funnel, F, so as to give
sharp boundaries in the region of the line XX'. A potential difference of ca.
lOOV is applied across the two electrodes, and the movement of the
boundaries observed. The average electrophoretic mobility of the particles
in the latex is taken to be equal to the rate of movement of the boundaries.
The problem of relating electrophoretic mobility to electrokinetic potential
is notoriously difficult; if a geometrically-complex apparatus such as that of
Blow is used, it is preferable merely to quote the results as speed of movement
per standard potential difference in the given apparatus. By means such as
this, it has been confirmed that addition of excess amounts of a quaternary
ammonium surface-active substance does indeed bring about reversal of the
polarity ofthe effective charge carried by the latex particles. If only qualitative
evidence of polarity change is required, then it is possible, for natural rubber
x-- -- ------ - --x'
Figure 14.9 Illustrating apparatus used by Blow [56] for investigating the sign and magnitude
of the effective charge carried by the latex particles.
latex at least, to observe the direction of migration of particles when a

potential gradient is applied to a small quantity of diluted latex contained
in a glass cell suitable for use with a microscope. Further information
concerning reverse-charge natural rubber latex which has an alkaline aqueous
phase is available in a second paper by Blow [57].
Latices of type (4) of the above classification must contain anionic colloid
stabilizers which are effective in acidic media. Many sulphate and sulphonate
types fulfil this requirement. Acidification of a latex of type (1) which contains
a sufficient quantity of such colloid stabilizers results in the formation of a
latex of type (4). Alternatively, latices of this type can be formed by emulsion
polymerization in an acidic medium using as colloid stabilizer a surface-active
sulphate or sulphonate. Reaction systems of this type are used for the
production of synthetic latices of polymers functionalized by acidic co-
monomers.
14.11.2 Surface dyeing of latex particles

Latices are usually whitish liquids of variable fluidity. The white colour of
the latex is primarily a consequence of a difference in refractive index between
the particles and the medium in which they are dispersed. Light of all
wavelengths is scattered indifferently; hence when illuminated by white light,
a latex scatters white light. The particles themselves are presumably almost
colourless usually, although they may be turbid to some extent because they
have absorbed a small amount of the dispersion medium. There are few
latices which are non-white because the particles are non-white. One example
is the latex of polypyrrole (Structure XLIX of Chapter II). This is black,
presumably because free radicals are de localized throughout the polymer
chain. Also, monodisperse latices when left to stand can develop iridescent
colours (see section 3.7 of Chapter 3 (Volume 1)). This is believed to be a
consequence of the existence of a secondary minimum in the potential
energy-separation curve for pairs of particles (see section 3.6 of Chapter 3).
Because the particles all have almost the same size, the interparticle separation
at which the secondary minimum occurs is the same for all pairs of particles.
Consequently a regular structure can develop within the latex as the particles
'crystallize' on a 'lattice', the spacing of which is determined by the nature
of the potential energy-separation curve. The colours seen in such latices are
a consequence of diffraction of light by the ordered particles. They are
temporary, disappearing if the structure within the latex is destroyed by
agitation or heating. Temporary slight iridescence is also sometimes observed
when thin films of synthetic latices are spread on glass. In neither case are
these colours properties of the individual particles themselves; they are a
consequence of temporary interparticle structure.
Some interest has been shown in the possibility of making synthetic latices
which contain coloured particles. There are several potential applications for
such latices. Various methods have been proposed for preparing these latices,
including the incorporation of copolymerizable dyes or transfer-active dyes
in the emulsion polymerization reaction system by which a synthetic latex
is made. Another possibility is to chemically modify uncoloured latex particles
so that chromophoric moieties become covalently bonded to the surface of
the particles. A patent to Shih, EI-Aasser and Vanderhoff [58] describes a
procedure for preparing polystyrene latices which contain particles having
dye molecules bonded to their surface. The modification can be essentially
confined to the surface regions of the particles. The initial step of a typical
process requires the preparation of a cationic synthetic copolymer latex of
styrene and p-vinylbenzyl chloride (Structure VI). Preferably, the particles
are produced by seeded emulsion polymerization using a polystyrene seed
and a mixture of styrene and p-vinylbenzyl chloride for the second-stage
polymerization. By this means, the p-vinylbenzyl chloride units, which provide
the sites for subsequent reaction, are confined mainly to the surface regions
of the particles. The latex particles are then subjected to surface amination
by reaction with N-methylaniline, and the surface-bound N-methylanilino
groups thereby formed are reacted with various diazonium salts to give
covalently-bonded dye molecules. The sequence of preparative reactions
where the diazonium salt is derived from p-nitroaniline can be represented
as shown below. The ami nation reaction using N-methylaniJine takes place
at normal ambient temperature, but requires ca. 10 days for completion.
Prior to reaction with the diazonium compound, unreacted N-methylaniline
is removed by the serum-replacement technique (see section 23.13.4.2 of
Chapter 23 (Volume 3)), using first dilute aqueous hydrochloric acid solution
of pH 2.3 and then distilled deionized water. Other comonomers containing
halogen atoms of suitable reactivity may be copolymerized with the main
monomer in the initial step.
The patent to Shih, EI-Aasser and Vanderhoff [58] also provides a useful
review of earlier attempts to prepare coloured latices by various methods.
14.11.3 Surface modifications of latex particles for immunological applications

There has been interest in using the large aggregate interfacial area of latices
for binding immunological reagents in such a way that they can be accessed
from the aqueous phase (see section 23.11 of Chapter 23 (Volume 3)). A paper
by Yen et al. [59], published in 1976, provides an introduction to this subject;
it also provides references to the earlier literature. Industrial polystyrene
latices have been used as immunological markers for investigations using
scanning electron microscopy. However, applications of such latices in this
6
_~~CH3
., .,
, ,
"--'
N
II
N
way are limited because of the hydrophobic nature of the particle surface.
This causes the particles to adhere non-specifically to many surfaces of living
cells, and frequently to induce a physiological response such as the clotting
of blood cells. Use has also been made of the ability of antibodies to become
physically adsorbed on to the surface of polystyrene latex particles. The latex
particles can then interact with specific antigens in blood sera, thereby causing
particle aggregation to occur. A number of diagnostic tests have been based
upon this phenomenon. However, because the adsorption of the antibody
molecule to the particle surface is weak, and also because polystyrene interacts
non-specifically with living cells, such agglutination tests have not always
been satisfactory. A more satisfactory procedure is to use latex particles which
have been suitably chemically modified so that the antibody molecules are
covalently bonded to the particle surface.
Yen et al. [59] have described the preparation of aqueous latices of
hydrophilic polymers which have been modified such that antibodies are
covalently bonded to the particle surface. The initial latices (see also section
11.9.4.1 of Chapter 11) were prepared by emulsion copolymerization of methyl
methacrylate, hydroxyethyl methacrylate and methacrylic acid, with
ethyleneglycol dimethacrylate as a crosslinking comonomer. Sodium n-dodecyl
sulphate was the colloid stabilizer and ammonium peroxodisulphate the
initiator. The pH of the latex was adjusted to 7, and the latex then purified
by ion-exchange. Use was made of the presence of hydroxyl and carboxylic-acid
groups at the surface of the latex particles to enable a few antibody molecules
to be covalently bonded to each particle. Methods of attachment making
use of cyanogen bromide, of carbodiimide, and of glutaraldehyde (pentandial)
are described in their paper.
In the cyanogen-bromide method, the hydroxyl groups of the polymer are
reacted with cyanogen bromide under alkaline conditions, and then coupled
to amino groups on the antibody protein molecule at a pH in the range
7-10. The following reaction mechanism is proposed, H2NR denoting the
antibody molecule:
and
In the carbodiimide method, a water-soluble carbodiimide (Structure VII)

is used to couple the amino groups of the antibody molecule to the
carboxylic-acid groups at the surface of the latex particles by means of a
peptide bond. The reaction is carried out with the pH in the range 6- 7 in
the absence of added buffer, or in the presence of a buffer which does not
interfere with the reaction. The proposed reaction mechanism is as follows,
in which again the antibody molecule is denoted by H2NR:
r-
~Co,H R'-N=C=N-~ - ~-O-C=I-RII
+
VII
r-o NH-R'
~N-O-,=N-RII ~N-NHR
0
+ H2NR- + R'-I-U-I-RII
o NH-R' 0 H H
Although this method is simple in principle, intermolecular and intramolecular
crosslinking of the antibody molecule can occur, because the latter also
contains carboxylic-acid groups. If such crosslin king occurs, then aggregates
can form, with loss of immunological activity.
The glutaraldehyde (pentandial) (Structure VIII) method is believed to
involve the addition of the amino groups of the antibody molecule to the
lX,fJ-unsaturated aldehyde oligomers which are present in aqueous solutions
of glutaraldehyde. These oligomers are believed to be formed by successive
aldol condensations, with concomitant elimination of water to give olefinic
unsaturation rx.,P to aldehyde groups. The initial steps resulting in the

formation of a glutaraldehyde dimer (Structure IX) can be represented as
follows:
H
2 HC-CH 2-CH 2-CH 2-CH
H
VIII 0 "'-
~ CHO 0
II I II
HC-CH2-CH2-CH2-CH(OH)-CH-CH2-CH2-CH
IX
The corresponding rx.,p-unsaturated aldehyde, formed by elimination of water

from IX, has Structure X. Various further aldol condensations, with concomitant
elimination of water to give olefinic unsaturation, can then occur. In
particular, it is postulated that a molecule having Structure XI forms as the
next step. XI will be formed if aldol condensation occurs between X and one
of the C=O group of a further pentandial molecule, and if the hydrogen
atom lost from X is one in the final right-hand methylene group of the
sequence shown in X. The rationale is presumably that the hydrogen atoms
in this methylene group are expected to be particularly susceptible to
participation in aldol condensation because of the aldehyde groups attached
to the adjacent carbon atom. It is then proposed that molecules such as XI
can cyclize to give structures of type XII, because the number of carbon
o CHO 0
II I II
HC-CH 2-CH 2-CH 2-CH=C-CH 2-CH 2-CH
X
o CHO CHO 0
II I I II
HC-CH2-CHz-CHz-CH=C-CHz-C=CH-CHz-CHz-CHz-CH
XI
CHO
I
C
/ ~
CH 2 CH CHO 0
I I I II
"~
CH 2 C-CH 2-C=CH-CH 2-CH 2-CH 2-CH
CH
XII
atoms which separate the aldehyde groups in glutaraldehyde is conducive
to cyclization in such derivatives. A possible reaction mechanism is internal
aldol condensation with concomitant elimination of water. Continuation of

such reactions is postulated to give polymeric structures which, because of
the presence ofthe aldehyde groups, contain olefinic double bonds of sufficient
reactivity to bond covalently with antibody molecules (H2NR) by reactions
of the following type:
CHO CHO
I I
-CH=C- + H2NR ~ -CH-CH-
I
NHR
Tarcha et al. [60] have recently reported the surface functionalization of
polypyrrole (Structure XLIX of Chapter 11) latices, and the subsequent
covalent binding of antibody molecules at the particle surface, to produce
latices for immunodiagnostic applications. This type of latex is well-suited
for visual immunoassay applications because of the intense black colour of
the particles. The latex was prepared by aqueous emulsion polymerization
(see section 11.9.4.2 of Chapter 11), and the dispersion medium replaced by
N-methylpyrrolidone by centrifugally concentrating the latex and then
diluting with the replacement dispersion medium. Several methods of
attaching antibody molecules to the particles in this latter latex are described,
including the following, in which> NH and ~CH represent respectively NH
and CH moieties in the cyclic repeat unit of polypyrrole:
l. N-acylation of the polymer by reaction with bromoacetyl bromide,
followed by reaction with the antibody molecule (H2NR):
+H2NR
>NH + CH2Br.COBr~ >N.CO.CH2Br - - - + >N.CO.CH 2.NHR
2. C-acylation of the polymer by reaction with bromoacetyl bromide,
followed by reaction with the antibody molecule (H2NR):
+H2NR
~CH + CH 2Br.COBr ~ ~C.CO.CH2Br - - - + ~C.CO.CH2.NHR
3. N-acylation of the polymer by reaction with bromoacetyl bromide,

followed by reaction with thioglycollic (mercaptoacetic) acid to functionalize
the surface of the latex particles with carboxylic-acid groups, followed by
reaction with the antibody molecule (H 2NR):
> NH + CH 2Br.COBr ~ > N.CO.CH2Br + HBr
> N.CO.CH2Br + HS.CH 2.C0 2H ~
> N.CO.CH 2.S.CH 2.C0 2H + HBr
> N.CO.CH 2.S.CH 2.C0 2H + H2NR ~
>N.CO.CH 2.S.CH 2.CO.NHR + H 20
References 573
l-ethyl-3-(3-dimethylaminopropyl) carbodiimide was used as catalyst for

the latter reaction.
4. N-acylation of the polymer by reaction with bromoacetyl bromide,
followed by reaction with triethylenetetramine to functionalize the surface
of the latex particles with amine groups. These amino groups were then
coupled with the amino groups of the antibody molecules by reaction
with glutaraldehyde (pentandial). The coupling reactions are presumably
similar to those described above in connection with the work reported by
Yen et al. [59].
REFERENCES
1. Farley, P. S. (1992) The Preparation, Characterization and Hydrolysis ofEpoxidized
Natural Rubber, Ph.D. thesis, University of London.
2. Ng, S. C. and Gan, L. H. (1981) European Polymer Journal, 17, 1073.
3. Gan, L. H. and Ng, S. C. (1986) European Polymer Journal, 22, 573.
4. Gelling, I. R. (1985) Rubber Chemistry and Technology, 58, 86.
5. Gelling, I. R. and Porter, M. (1988) Chemical Modifications of Natural Rubber,
in Natural Rubber Science and Technology (ed. A. D. Roberts), Oxford University
Press, Oxford, Chapter 10, especially Sections 8, 9 and 10.
6. Bartlett, P. D. (1950) Record of Chemical Progress, 11, 47.
7. Campbell, D. S. (1989) Preprints of Plastics and Rubber Institute Polymer Latex
III International Conference, London, 1989, Paper No. 17.
8. Gelling, I. R./The Malaysian Rubber Producers' Research Association, UK
Patent Application No.2, 113,692, 10 August, 1983.
9. Burfield, D., Lim, K. -L. and Law, K. -So (1984) Journal of Applied Polymer Science,
29, 1661.
10. Perera, M. C. S., Elix, J. E. and Bradbury, J. H. (1988) Journal of Polymer Science:
Polymer Chemistry Edition, 26, 637.
11. Allen, P. W. (1963) Graft Copolymers from Natural Rubber, in The Chemistry
and Physics of Rubber-Like Substances (ed. L. Bateman), MacLaren, London,
Chapter 5.
12. Ceresa, R. J. (1973) Syntheses and Characterization of Natural Rubber Block and
Graft Copolymers, in Block and Graft Copolymerization, Volume 1 (ed. R. 1.
Ceresa), John Wiley, London, Chapter 3.
13. Pendle, T. D. (1973) Properties and Applications of Block and Graft Copolymers
of Natural Rubber, in Block and Graft Copolymerization, Volume 1 (ed. R. J.
Ceresa), John Wiley, London, Chapter 4.
14. Bacon, R. G. R., Farmer, E. H. and Schidrowitz, P. (1938) Proceedings of the
Rubber Technology Conference, London, 1938, p. 525.
15. Bloomfield, G. F. (1952) Rubber Developments, 5,34.
16. Sekhar, B. C. (1958) Rubber Chemistry and Technology, 31, 430.
17. Allen, P. W., Bell, C. L. M. and Cockbain, E. G. (1959) Proceedings of International
Rubber Conference, Washington, 1959, p. 521.
18. Allen, P. W., Ayrey, G. and Moore, C. G. (1959) Journal of Polymer Science, 36, 55.
19. Andrews, E. H. and Turner, D. T. (1960) Journal of Applied Polymer Science, 3, 366.
20. Burfield, D. R. and Ng, S. C. (1975) Proceedings of International Rubber Conference,
Kuala Lumpur, 1975, Volume IV, p. 235.
21. Cockbain, E. G., Pend Ie, T. D. and Turner, D. T. (1958) Chemistry and Industry,
p.759.
22. Cooper, W., Vaughan, G., Miller, S. and Fielden, M. (1959) Journal of Polymer
Science, 34, 651.
23. Cooper, W. and Vaughan, G. (1959) Journal of Polymer Science, 37, 241.
24. Cooper, W., Sewell, P. R. and Vaughan, G. (1959) Journal of Polymer Science,
41, 167.
25. Bloomfield, G. F., Merrett, F. M., Popham, F. J. and Swift, P. McL. (1954)
Proceedings of the Third Rubber Technology Conference, London, 1954, p. 185.
26. Campbell, D. S. (1988) Graft Copolymers from Natural Rubber, in Natural Rubber
Science and Technology (ed. A. D. Roberts), Oxford University Press, Oxford,
Chapter 14.
27. Hayes, R. A. (1953) Journal of Polymer Science, II, 531.
28. Gardner, 1. B. and Harper, B. G. (1965) Journal of Applied Polymer Science, 9, 715.
29. Cunneen, J. I. and Shipley, F. W. (1959) Journal of Polymer Science, 36, 77.
30. Cunneen, J. I., Fletcher, W. P., Shipley, F. W. and Wood, R. I. (1958) Transactions
of the Institution of the Rubber Industry, 34,260.
31. Edwards, E. P. B. (1953) India Rubber Journal, 125, 334.
32. Lee, D. F., Scanlan, J. and Watson, W. F. (1963) Proceedings of the Royal Society,
A173,345.
33. van Veersen, G. J. (1951) Rubber Chemistry and Technology, 24, 957.
34. Gordon, M. (1951) Industrial and Engineering Chemistry, 43, 386.
35. Gordon, M. (1951) Proceedings of the Royal Society, Al04, 569.
36. Fisher, H. L. and McColm, E. M. (1927) Industrial and Engineering Chemistry.
19, 1325, 1328.
37. van Amerongen, G. J. (1951) Industrial and Engineering Chemistry, 43, 2535.
38. Baker, H. C. (1938) Proceedings of the Rubber Technology Conference, London,
1938, p. 209.
39. Kraus, G. and Reynolds, W. B. (1950) Journal of the American Chemical Society,
72,5621.
40. Bloomfield, G. F. (1943) Journal of the Chemical Society, p. 289.
41. Bloomfield, G. F. (1944) Journal of the Chemical Society, p. 114.
42. van Veersen, G. J. (1948) Proceedings ofthe Second Rubber Technology Conference,
London, 1948, p. 87.
43. Crampsey, E., Gordon, M. and Taylor, J. S. (1953) Journal of the Chemical Society,
p.3925.
44. Gordon, M. and Taylor, J. S. (1953) Journal of Applied Chemistry, 3, 537.
45. Cockbain, E. G., Pendle, T. D., Pole, E. G. and Turner, D. T. (1962) Proceedings
of the Fourth Rubber Technology Conference, London, 1962, p. 498.
46. Bloomfield, G. F., Farmer, E. H. and Schidrowitz, P. (1934) Bulletin of the Rubber
Growers' Association, 16, 116.
47. Stevens, H. P./Rubber Producers' Research Association, British Patent No.
442,136, 27 January, 1936.
48. Gazeley, K. F., Menthe, P. G. and Morris, M. D. (1987) Paper No. 4.2 presented
at Malaysian Rubber Producers' Research Association International Conference
on Developments in the Plastics and Rubber Product Industries, Kuala Lumpur,
1987.
49. Coomarasamy, A., Hoffmann, U., Nadarajah, M., Seneviratne, W. M. G. and
Wickremasinghe, N. (1989) Preprints of Plastics and Rubber Institute Polymer
Latex III International Conference, London, 1989, Paper No. 16.
50. Hourston, D. J. and Romaine, J. (1989) Preprints of Plastics and Rubber Institute
Polymer Latex III International Conference, London, 1989, Paper No. 18.
51. Sperling, L. H. (1981) Interpenetrating Polymer Networks and Related Materials,
Plenum Press, New York.
References 575
52. Hourston, D. J. and Satgurunathan, R. (1984) Journal of Applied Polymer Science,

29,2969.
53. Hourston, D. J., Satgurunathan, R. and Varma, H. (1986) Journal of Applied
56. Blow, C. M. (1938) Proceedings of the Rubber Technology Conference, London,
1938, p. 186.
57. Blow, C. M. (1938) Journal of Society of Chemical Industry, 57, 116.
58. Shih, y'-J., El-Aasser, M. S. and Vanderhoff, J. W. United States Patent No.
4,487,855, 11 December, 1984.
59. Yen, S. P. S., Rembaum, A., Molday, R. W. and Dreyer, W. (1976) Functional
Colloidal Particles for Immunoresearch, in Emulsion Polymerization (eds I. Piirma
and J. L. Gardon), ACS Symposium Series No. 24, American Chemical Society,
60. Tarcha, P. J., Misun, D., Finley, D., Wong, M. and Donovan, J. J. (1992) Synthesis,
Analysis, and Immunodiagnostic Applications of Polypyrrole Latex and its
Derivatives, in Polymer Latexes: Preparation, Characterization, and Applications (eds
E. S. Daniels, E. D. Sudol and M. S. El-Aasser), ACS Symposium Series No. 492,
American Chemical Society, Washington, DC, p. 347.
Index
Page numbers appearing in bold refer to figures and page numbers appearing in
italics refer to tables.
Ab initio emulsion polymerization mechanical agitation 279,284-9

199-200 natural rubber latex 53-5
Accelerators partial agglomeration 264-89
dithiocarbamates 464-5, 466, 470-4, physical methods 265--6, 278-89
479-82 pressure agglomeration 285-9
sulphur-prevuIcanization 446-7, 460, solvent addition 266-72
462-6,470-4,479-82 styrene-butadiene copolymers
Acetate, as polyisoprene precursors 264-89
120-4 surfactant addition 278
Acetone process, polyurethane artificial. ultrasonic techniques 288-9
latices 412-14 Aggregation, natural rubber latex 53-5
Acrylic latices 353-7 Alkali-thickenable synthetic latices
chemical structure 353-4 242, 304-5, 368-71
glass-transition temperature 354-6 Alkyd-modified drying oils, artificial
Acrylonitrile-butadiene copolymers latices 438
323-5 American leaf blight 10-11
glass-transition temperature 323-4 American Society for Testing and
Adhesives Materials (ASTM), natural rubber
function ali zed latices 241-2 latex concentrates 73-5
graft-copolymers, use of 522 Amino acids, natural rubber latex
styrene-vinylpyridine-butadiene aqueous phase 96-100, 105
terpolymer latices 320--1 Aminophenols, preservative for natural
Agglomeration rubber latex 37
chemical methods 265-78 Ammonia
co-agglomeration 266 alkalinity 30-1
colloid stabilizer surface-activity as bactericide 29-30
272-3 hazards 28-9, 32-3
effect upon viscosity~oncentration multivalent metal ion deactivation
relationship 260--4, 274-6 31-2
electrolyte addition 269-72 preservative for natural rubber latex
freeze-thaw process 278-84 27-40
hydrocolloid addition 273-8 reduction of allergenic reactions 100
578 Index
Ammonium alginate, creaming agent potassium hydroxide 41
40,48,289-92 for synthetic latices 143
Amphoteric synthetic latices 371-5 volatile fatty acid formation 108
Anaphylactic shock, reaction to Balata latex 115-18
natural rubber latex 98-100 Batch emulsion polymerization 200-1,
Antibiotics, preservative for natural 317
rubber latex 41 Beerbower-Burton-Malloy process,
Archras zapote 115-17 artificial latices 404-6, 407
Artificial latices 391-440 Benzene, solvent agglomeration 267
see also Redispersible latices Binary copolymerization 206-32
Beerbower-Burton-Malloy process particle morphology 224-32
404-6,407 Biogenesis, polyisoprenes 118-31
chlorosulphonated polyethylene Block copolymers 518,522
rubbers 437-8 Boric acid
creaming 393-4 inhibitor for volatile fatty acid
ethylene-propylene rubbers 435-7 formation 111-12
Hunter-Segura-Small-Miller process preservative for natural rubber latex
402-4 37-8
isobutene-isoprene rubbers 402-4, Bromination, natural rubber latex
419-20,509 551-2
Murphy-Madge process 401-2 Bulk polymerization 147-8
mutual solvent technique 39-7 Butadiene
natural rubber latex 391, 397-8 see also Styrene-butadiene
particle size 393-4 copolymers
phase-inversion 395-7 acrylonitrile-butadiene copolymers
cis-I,4-polybutadiene rubbers 429-30 323-5,508-9
cis-I,4-polyisoprene rubbers 420-30 chemical structure 306-9
polyurethane polymers 408-18,434-5 emulsion polymerization 175-6, 181,
Pratt processes 398-401 183
production 394-419 styrene-vinyl pyridine-butadiene
reclaimed rubbers 391,397-8,431-4 terpolymer latices 320--1
re-vulcanization 431-4 Butyl rubbers 392
sedimentation 393-4
self-dispersion 395-7 Carageenin, natural rubber latex
soap in situ technique 396 creaming agent 51
solution-emulsification 395-7, 418, Carbohydrates, natural rubber latex
422 aqueous phase 96
Trumbull-Dickson process 401 Carbothiolic acids
Vanderhoff-EI-Aasser-Ugelstad cis-trans isomerization of natural
process 406-8 rubber latex 539-40,541-3
Azobisisobutyronitrile, emulsion Carboxylated latices
polymerization initiator 177 advantages and disadvantages 241-3
alkali-thickenable 242, 369
Bactericides carboxylic-acid group distribution
aminophenols 37 248-60
ammonia 29-30 colloid stabilizers 244-6
formaldehyde 40,143 crosslinking 360--4
for natural rubber latex 29-30,35-7, electrolytes 247
40-1 hydrophilicity 249-50, 257
pentachlorophenol 35-7 initiators 246
Index Sl9
modifiers 246-7 Chlorination
particle nucleation 255-60 kinetics 550-1
pH 243-8, 249-55 natural rubber latex 548-58
polymerization rate 255-60 Chloroform-coagulation test, degree of
post-polymerization additives 247-8 vulcanization 450-2
production by emulsion Chlorosulphonated polyethylene
polymerization 243-8, 254-5 rubbers, artificial latices 437-8
water-absorption 242-3 Cis-trans isomerization
Carboxylate soaps carbothiolic acid effectiveness
colloid stabilizers 181-2, 189,234, 539-40,541-3
272-3 mechanism 538-9
spontaneous coagulation of natural natural rubber latex 537-43
rubber latex 25-6, 39 Clones, Hevea brasiliensis
Castilla elastica 115-16 (Euphorbiaceae) 12, 114
Cationic synthetic latices 371-5 Co-agglomeration 266
Caucho rubber 115-16 see also Agglomeration
Ceara rubber 115-16 Coagulation
Centrifugation chloroform-coagulation degree of
concentrate properties 69-77 vulcanization test 450-2
and creaming 56--9, 63-4, 289 natural rubber latex 25-6
degree of ammoniation 62 synthetic latices 164
de Laval centrifuge 60-3 Coagulum
efficiency 63-4 emulsion polymerization 234-8
natural rubber latex concentration synthetic latex agglomeration 276,
55-65, 103-5 281
particle size distribution 58-9 Coenzyme A, polyisoprene biosynthesis
skim rubber 64-5 123-4
synthetic latices 289 Colloid stability
theory 56--9 artificial latices 396--7
Centrosema pubescens 8 electrolyte effects 269, 271
Chain scission, cis-l ,4-polyisoprene 89 emulsifier-free emulsion
Charge reversal, latex particle surfaces polymerization 198-9
563-6 functionalized latices 241
Chemically-modified latices 441-575 isobutene-isoprene rubbers 419-20
see also Functionalized latices natural rubber latex 26, 31, 36, 41,
cis-trans isomerization 537-43 106--8, 112-14
cyclization 543-8 storage effects 106--8, 112-14
depolymerization 558-61 styrene-butadiene copolymers 311-12
epoxidation 513-18 Colloid stabilizers
graft-copolymerization 518-37 for carboxylated latices 244-6
halogenation 548-53 carboxylates 181-2, 189,234, 272-3
halogenoalkylation 555-8 in emulsion polymerization 141-6,
hydrohalogenation 553-5 153,157-65,172-3,181-5,
interpenetrating polymer networks 244-6
561-2 ethoxylates 183-4
oxidation 558-61 hydrocolloids 184, 190, 339, 341-2,
prevulcanization 441-511 448
surface modification 563-73 oligomers 184-5
Chemical stability, isocyanates 409-11 particle nucleation 181
Chicle latex 115-17 polymerizable 185
580 Index
Colloid stabilizers (contd) styrene-ethyl acrylate 221-3
protective colloids 149 styrene-methacrylic acid 255-60
sulphates 182-3 tapered emulsion copolymerization
sulphonates 182-3, 244-6 212
surface-activity partial destruction vinyl acetate 334-53
272-3 vinyl chloride-vinylidene chloride
for synthetic latices 140, 141-2, 150, 357-8
244-6 Copper sulphate, natural rubber latex
volatile base 142 flow stimulation 23-4
Concentrates Creaming
ASTM standard 73-5 artificial latices 393-4
effect of multiple concentration 74-7 and centrifugation 56-9, 63-4, 289
ISO specification 71-3 concentrate properties 69-77
LAZN concentrates 38-9 creaming agents 48-55, 289-92
natural rubber latex 42-78 creaming rate 46-7,52
properties 69-77 and electrodecantation 66-9
Condensation polymerization, mechanisms 49-55, 289-90
isopentenyl pyrophosphate 128-31 natural rubber latex concentration
Contact urticaria, reaction to natural 43-55
rubber latex 98-100 particle size distribution 47-8, 52-3,
Continuous emulsion polymerization 55, 289-90
200,202-6 skim separation 49-51,55
reactors 202-6 synthetic latex concentration 289-92
residence time 202-6 Crosslinking
styrene-butadiene copolymers 317-18 see also Prevulcanization;
Conversion-time curves 146, 154-6, Vulcanization
255-60, 527-9 carboxylated latices 360-4
Copolymerization functionalized latices 359-68
acrylonitrile-butadiene 323-5 natural rubber latex 459, 478, 482,
azeotropic monomer feed 484-508,514-5
composition 208 polychloroprene rubber latices 333-4
binary 206-32, 224-32 cis-l,4-polyisoprene 85--7, 89, 426-9,
block-copolymers 518, 522 476,477,482
composition drift 208-10 polyurethane artificial latices 416-18
copolymer composition equation styrene-butadiene copolymers 311
207-8 Crystallization
emulsion polymerization 153, 206-32 natural latices 116
gradient copolymer latex particles natural rubber latex 540
214-221 polychloroprene rubber latices 327
graft-copolymers 176, 192-4, 213-4, Cultivation, Hevea brasiliensis
518-37 (Euphorbiaceae) 3-24
monomer-addition regimens 211-232 Cyanides, natural rubber latex aqueous
monomer reactivity ratios 207-10 phase 100
multi-stage 212-3 Cyclization
power-feed 214-221 mechanism 544-8
products by process 211 natural rubber latex 543-8
sequential212-3
structured latex particles 211-32 Degree of vulcanization
styrene-butadiene 186-90, 194-6, chloroform-coagulation test 450-2
196-8, 231-2, 305-23 equilibrium-swelling test 450, 452-3
Index 581
natural rubber latex 450-6 Elongation at break see Extension at
relaxed-modulus test 450, 453-4 break
De Laval centrifuge 60-3 Emulsifiers see Colloid stabilizers
Density Emulsion polymerization
aqueous phase of natural rubber see also Free-radical addition
latex 96 polymerization
fresh natural rubber latex 78 ab initio 199-200
rubber particles 83 acrylic latices 355-7
Depolymerization, natural rubber latex amphoteric synthetic latices 371-5
558-61 batch processes 200-1,317
Dispersion polymerization bonding comonomers 213-4
polymethyl methacrylate 295-6 butadiene 175-6, 181, 183, 307-9
Dithiocarbamates carboxylated latices 243-8
preservative for natural rubber latex cationic synthetic latices 371-5
37,463-4 colloid stabilizers 141-6, 153,
sulphur-prevulcanization 157-65,172-3,181-5,244-6
accelerators 464-5, 466, 470-4, compartmentalized propagation 150,
479-82 152-4
Dry natural rubber, sulphur- continuous 200, 202-6,317-18
vulcanization 443, 465, 468-9, conversion-time curves 146, 154-6,
476,480 255-60
Dyeing, latex particle surfaces 566-7 copolymerization 153, 206-32
Dyera costulata 115-16 crosslinker comonomers 192-4,
213-4,267-8
Elastic modulus see Modulus electrolytes 189-91,247
Electrical conductivity, natural rubber emulsifier-free 161, 1989
latex 112 freezing-point depressants 194-6
Electrodecantation functionalized latices 238~0
concentrate properties 69-77 graft copolymers 176
cream removal 68 graftlinker comonomers 192-4,
natural rubber latex concentration 213-4
66-9 high-solid latices 260-4
Electrolytes initiation 145-7, 176-81
in emulsion polymerization 189-91, intervals 154-76
247 kinetics 154-76
synthetic latex agglomeration 269-72 limiting conversion phenomenon 175
Electromagnetic waves literature 143-4
graft-copolymerization initiation mechanism 150-76
530-3 micellar mechanism 151, 156-61,
prevulcanization of natural rubber 164, 165, 181, 189
latex 499-507 modifiers 150, 186-9
Electron irradiation, prevulcanization monomer droplets 152-4, 155-6,
of natural rubber latex 507-8 157, 164-8
Electrophoresis monomer-starved 201, 215
and dissolved metal cations 93 non-aqueous reactions 293-6
natural rubber latex particle surfaces non-latex polymer formation 234-8
90-5 particle nucleation 155-65, 255-60
and pH 90-3 polymerization rate 255~O
Electrostatic effects, natural rubber polymerization temperature 313-15
latex 113 post-reaction additives 198, 247-8
582 Index
Emulsion polymerization (contd) modulus 479
practical methods 199-238 natural rubber latex 98-100, 454-9,
reactions 145-238 483-91,497,501-7
seeded 199-200,232-4 polychloroprene rubber latices 332-4
semi-continuous 200,201,237 cis-l ,4-polyisoprene rubbers 423-30
sequestering agents 196 polyurethane artificial latices 435
shorts toppers 196-8 solvent swelling 484-5
synthetic latices 137-302 styrene-butadiene copolymers 310,
termination reactions 147, 169-72 508-9
transfer reactions 147, 186-9, 197 synthetic latices 141-2,266
Enzymes tensile strength 455-8,479,483,497,
deprotenization of natural rubber 501-9
latex 87-9 vinyl acetate polymers and
natural rubber latex aqueous phase copolymers 343-5, 347
100 water-resistance 141-2,242-3
volatile fatty acid fonnation 108-12 Flammability, natural rubber 555-8
Epoxidation, natural rubber latex Flocculation
513-18 natural rubber latex 51, 53, 55
Equilibrium-swelling test, degree of synthetic latices 163
vulcanization 450, 452-3 Fonnaldehyde
Ethoxylates, colloid stabilizers 183-4 bactericide 40, 143
Ethylene-propylene rubbers partial de-ammoniation of natural
artificial latices 435-7 rubber latex 34-5
Beerbower-Burton-Malloy process preservative for natural rubber latex 40
404-6,407 Free-radical addition polymerization
ethylene-propylene-<iiene rubbers see also Emulsion polymerization
404-6,407 bulk polymerization 147-8
Evaporation copolymerization 206-32,520-1,
concentrate properties 69-77 523-4
Luwa thin-layer evaporator 293, 294 initiation 178-80
Metallgesellschaft Aktiengesellschaft kinetics 167-72
process 44 mechanism 146-155,164-7,172-3
natural rubber latex concentration solution polymerization 148-9
43-4 suspension polymerization 149-50
synthetic latex concentration 292-3 synthetic latices 138, 146-54
Extension at break Free surface energy
graft-copolymers 533-4 fresh natural rubber latex 78
natural rubber 455, 460-3, 483, 491, isobutene-isoprene rubbers 420
502,508 synthetic latices 140, 184,265,
cis-I ,4-polyisoprene rubbers 425-30 274-6,280
styrene-butadiene copolymers Freezing
310-11 critical freeze temperature 281-2
freezing-point depressants 194-6
Fatty acids see Volatile fatty acids synthetic latex agglomeration
Ficus elastica 115-16 278-84
Films Frey-Wyssling particles, natural
acrylonitrile-butadiene rubber rubber latex 79, 102-5, 107
synthetic latices 508-9 Functionalized latices
clarity 458-9 see also Chemically-modified latices
microstructure 485-91 advantages and disadvantages 241-3
Index 583
alkali-thickenable 242, 304-5, kinetics 526-9
368-71 methyl methacrylate 525-35
amphoteric 374-5 in natural rubber latex 523-35
carboxylated latices 239-60,358-71 in synthetic latices 535-7
chemical reactions 359-68 reactions 518-22
colloid stability 241 retardation 531
comonomers 239-40 thermoplastic rubbers 522
crosslinking 359-68 Guayule rubber 115-16
production by emulsion Gum tragacanth, natural rubber latex
polymerization 238-60 creaming agent 51
styrene-butadiene copolymers Gutta percha latex 115-17
24(}"-60, 358-71
synthetic 238-60,304-5,358-71 Halogenation, olefinically-unsaturated
ultraviolet light absorption 377-9 polymer latices 548-53
Halogenoalkylation
Gamma-rays effect upon flame-resistance of
graft-copolymerization initiation natural rubber 556-8
53(}"-3 natural rubber latex 555-8
prevulcanization 499-507 Hardness, graft-copolymers 532-4
Gelation, natural rubber latex 33-4 Heels, emulsion polymerization 200
Gels Heptane ratio 274
liquefaction 19(}"-1 Hevea brasiliensis (Euphorbiaceae)
natural latices 116 1-25
natural rubber latex fraction 80, American leaf blight 1(}"-11
85-89 branch cuttings 17
cis-I,4-polyisoprene 85-6, 423 brown bast 20
Glass-transition temperature bud-grafting 14-17
acrylic latices 354-6 cloning 12, 114
acrylonitrile-butadiene copolymers cover-crops 8
323-4 cultivation 3-24
graft-copolymers 532-4 enzymes 100
natural rubber latex 516-7 flow stimulation 23-4
polyacrylate latices 354-6 germination 12
polychloroprene rubber latices 326 history 9-11
polymethacrylate ester latices 354-6 latex vessels 5-7
styrene-butadiene copolymers 309-10 leaves, flowers and fruit 4
synthetic latices 213, 243 mature trunk 5
vinyl acetate polymers and plantations 7-11
copolymers 334, 336 propagation by seed 11-14
vinyl chloride-vinylidene chloride protoplast culture 18
copolymers 358 regional occurrence 7-11
Glenshiel I 32, 114 rubber content 42
ex-Globulin, natural rubber latex 90, seedlings 13-14
96-7 selective breeding 22-3
Government rubber-styrene latices tapping 18-22
305-6,315,317,319 tissue culture 17-18
Graft-copolymerization vegetative propagation 11-12, 14-18
graftlinker comonomers 192-4, yield improvement 21-4
213-4 Hevea ribonucleoprotein, natural
initiation 524-33 rubber latex aqueous phase 97
584 Index
Hevein, natural rubber latex aqueous Initiators
phase 96-7 azobisisobutyronitrile 177
High-energy radiation for carboxylated latices 246
gamma-rays 499-507,530-3 dissociative 176-7
electromagnetic waves 499-507, in emulsion polymerization 145-7,
530-3 176-81
electron irradiation 507-8 in graft-copolymerization 524-33
graft-copolymerization 530-3 high-energy irradiation 530-3
prevulcanization of natural rubber hydroperoxides 178-80, 524-6
latex 498-508 impurity effects 181
radiation-exposure dose 500-5 peroxodisulphates 176-7,246,530
Homogeneous nucleation 161-3 redox 176, 177-80,524,529-30
Roe theory 162, 168-9 International Organization for
Homogenizers, pressure agglomeration Standardization (ISO), natural
285-6 rubber latex concentrates 71-3
Hunter-Segura-Small-Miller process, Interpenetrating polymer networks
artificial latices 402-4 561-2
Hydrocolloids Ionic strength, natural rubber latex
colloid stabilizers 184, 190,339, 112-14
341-2,448 Iron, emulsion polymerization
natural rubber latex creaming agent initiators 178-80
51,53 Isobutene-isoprene rubbers 419-20
synthetic latex agglomeration applications 420
273-8 Beerbower-Burton-Malloy process
water-soluble 273-8 404-6,407
Hydrocyanic acid, natural rubber latex Hunter-Segura-Small-Miller process
aqueous phase 100 402-4
Hydrogen peroxide prevulcanization 509
epoxidation 514-6 Isocyanate-terminated prepolymers
oxidation 559-61 chain-extension 411-12
prevulcanization 492-7 chemical stability 409-11
Hydrohalogenation, olefinically- polyurethane latices 409-18
unsaturated polymer latices Isomerization, isopentenyl
553-5 pyrophosphate 127-8
Hydrolysis Isopentenyl pyrophosphate
phospholipids 106-7 condensation polymerization 128-31
polychloroprene rubber latices isomerization 127-8
330-4 polyisoprene biosynthesis 126-31
proteins 106-7 Isoprene 119-20
vinyl acetate polymers and isobutene-isoprene rubbers 402-6,
copolymers 345-53 419-20
Hydroperoxides
polymerization initiators 178-80, Jelutong 115-16
524-6
prevulcanization 492-7 Ketimine process, polyurethane
Hydroxyperoxides artificial latices 412
emulsion polymerization initiators Kinetics
178-80 chlorination 550-1
8-Hydroxyquinoline, preservative for cyclization 544-8
natural rubber latex 39 emulsion polymerization 154-76
Index 585
graft-copolymerization 526-9 vinyl acetate polymers and
hydrohalogenation 553-5 copolymers 344, 346
sulphur-prevulcanization 469-83 Mechanical stability
KLAZN latex 39 natural rubber latex 32, 36, 39,
KOH number 112-14,501-7
see also Potassium hydroxide cis-l,4-polyisoprene rubbers 429
natural rubber latex 72--4,76-7, 106, synthetic latices 142
501-4 Melt-dispersion process, polyurethane
artificial latices 414-16
Latent latex particles 157 Mercaptans, emulsion polymerization
Latex dipping 444,506-7 modifiers 186-9
LAZN concentrates 38-9 Metallgesellschaft Aktiengesellschaft
Leaching, natural rubber latex evaporation process 44
455-8 Methyl cellulose, natural rubber latex
Lecithins creaming agent 50,51
chemical structure 93 I-Methyl inositol, natural rubber latex
natural rubber latex 93-5 aqueous phase 96
protein association 94-5 Methyl methacrylate, graft-
Lipids copolymerization 525-35
hydrolysis 106-7 Mevalonic acid, polyisoprene
natural rubber latex 93-5 biosynthesis 125-6
Locust bean gum, natural rubber latex Micellar nucleation 151, 156-61, 164,
creaming agent 50,51 165, 181, 189
Loop reactors, continuous emulsion Smith-Ewart theory 158-61, 168-9,
polymerization 202-206 532
Lutoids Microvoid particles
fibrillar content 101-2 stepwise alkali-acid process 376-7
natural rubber latex 78, 79, 101-2, synthetic latices 375-7
107 Mimusops balata 115-18
Luwa thin-layer evaporator 293, 294 Modifiers
Lyophilic macromolecules, natural for carboxylated latices 246-7
rubber latex creaming agent 54 in emulsion polymerization 150,
186-9
Magnesium ammonium phosphate, mercaptans 186-9
preservation of natural rubber Modulus
latex 31-3, 39 degree of vulcanization testing 450,
Manihot glaziovii 115-16 453--4
Mastication 400 films 479
Maturation, natural rubber latex 445 natural rubber latex 418, 456-63,
McGavack system, preservation of 491-2,508
natural rubber latex 40 cis-l,4-polyisoprene rubbers
Mechanical properties 423-30
graft-copolymers 533-5 styrene-butadiene copolymers
styrene-butadiene copolymers 310-11
312-15 styrene-ethyl acrylate copolymers
styrene-ethyl acrylate copolymers 221-3
221-3 Morphology see Particle morphology
sulphur-prevulcanized natural rubber Mud-cracking, synthetic latices 142
latex films 483 Multi-stage emulsion copolymerization
synthetic latices 140-1 212-3
586 Index
Murphy-Madge process, artificial halogenation 548-53
latices 401-2 halogenoalkylation 555-8
Hevea brasiliensis (Euphorbiaceae)
Natural latices 1-136 1-25,42,73-5,100,114,116
cis-l,4-polyisoprene microstructure hydrohalogenation 553-5
115-17 hydrolysis 31, 39, 106-7
trans-I,4-polyisoprene industrial grades 491-2
microstructure 115--17 KOH number 72-4, 76-7, 106,
sources 115-17 501-4
Natural rubber latex 1-114 lipids 93-5
see also cis-l,4-Polyisoprene lutoid phase 78, 79, 101-2
aeration 34-5 maturation 445
agglomeration 53-5 mechanical stability 32, 36, 39,
aggregation 53-5 112-14, 501-7
alkalinity 30-1, 69-77 micro-flocculation 51,53,55
allergenic reactions 98-100 modulus 418, 456-63, 491-2, 508
ammoniation 27-40, 105--112 negatex 564
aqueous phase 79, 96-100 network formation 54
bacterial infection 108 non-rubber constituents 476, 478-9
bottom fraction 97, 102 oxidation 558--61
bromination 551-2 partial de-ammoniation 34-5
centrifugation 55--65, 103-5 partially-prevulcanized 444
chemical composition 79,83 particle shape 79-80
chlorinated 548-58 particle size 52-3, 55, 80, 474-6
cis-trans isomerization 537-43 particle-size distribution 47-8,
colloidal stability 26,31,36,41, 59-60,71,80-83
106-8, 112-14 particle surface 90-5
concentration 42-78,513-4 pH 30-1, 39, 55, 78,491,501-4
constitution 78-114 positex 563-4
creaming 43-55 preservation 24-41
crosslinking 459, 478, 482, 484-508, prevulcanization 442-511
514-5 putrefaction 25
crystallization resistance 540 redispersible products 77-8, 391,
cultivation 3--24 397-8
cyclization 543-8 reticulum 104-5
degree of vulcanization 450-6 Revertex 70
density 78, 96 Revultex 491-2
depolymerization 558-61 rubber content 42-78
electrodecantation 66-9 rubber phase 79-95
electrophoretic behaviour 90-5 sedimentation 45
enzymatic deprotenization 87-9 skim rubber 64-5,69
epoxidation 513-18 spontaneous coagulation 25-6
evaporation 43-4 storage effects 105-6, 112-4
films 98-100, 454-9, 483-91, 497, sulphur-prevulcanization 442-92
501-7 surface free energy 78
Frey-Wyssling particles 79, 102-5 transportation 42
gelation 33-4 tree tapping 18-22
gel fraction 80, 85-89 vacuolar particles 104
IX-globulin 90 viscosity 69,78,491,501-7,560-1
graft-copolymerization in 523-35 volatile fatty acids 96, 106-14
Index
vulcanization 442-3 solvent agglomeration 266-7
Negatex 564 structured latex particles 219-21,
N-nitrosamines, prevuIcanization of 222,226-30
natural rubber latex 493-7 styrene-butadiene copolymers
Nucleation see Particle nucleation 262-3
Nucleoprotein, natural rubber latex sulphur-prevulcanization ingredients
aqueous phase 97 465,467
synthetic latices 139
OIefinically-unsaturated polymer Particle size distribution
latices measurement 80-2
halogenation 548-53 natural rubber latex 47-8, 59-60,71,
hydrohalogenation 553-5 80-3
Oligomers, colloid stabilizers 184-5 cis-I,4-polyisoprene rubbers 422-3
Organic peroxides, prevulcanization of synthetic latices 139,260-4
natural rubber latex 492-7 Particle surface
Oxidation, natural rubber latex 558-61 charge reversal 563-6
dyeing 566-7
Palaquium gutta 115-17 electrophoresis 90-5
Pale crepe 40 modification 563-73
gel fraction 85 natural rubber latex 90-5
Para rubber 3 protein layer 90, 93-5
Parthenium argentatum 115-16 synthetic latices 140
Partial soap neutralization 272-3 Pentachlorophenol, preservative for
Particle morphology natural rubber latex 35-7
core-shell 212, 224-30 Peptization, polychloroprene rubber
emergent 224-5 latices 328
separate-particle 224-5, 228-30 Peroxocarboxylic acids, epoxidation of
styrene-butadiene copolymers 231-2 natural rubber latex 516-7
two-stage emulsion binary Peroxodisulphates, polymerization
copolymerization 224-32 initiators 176-7,246,530
Particle nucleation pH
carboxylated latices 255-60 alkali-thickenable synthetic latices
coagulative mechanism 164 368-71
emulsion polymerization 155-65 carboxylated latex production 243-8,
free radicals 164-5 249-55
homogeneous 161-3 and electrophoretic mobility 90-3
hydrophobic association 164 effect upon freeze-thaw
locus 155 agglomeration 280-1
micellar 150, 156-61, 164, 165, 181, effect upon pressure agglomeration
189 286-8
microlatices 165 iSQbutene-isoprene rubbers 420
partial 155 natural rubber latex 30-1, 39, 55, 78,
Particle shape 491,501-4
natural rubber latex 79-80 partial soap neutralization 272-3
synthetic latices 139 polychloroprene rubber latices
Particle size 330-3,335
artificial latices 393-4 reduction by electrolytes 190-1
natural latices 116 Phase-inversion process
natural rubber latex 52-3, 55, 80, artificial latices 395-7
474-6 reclaimed rubbers 432-4
588 Index
Phenoxyacetic acid derivatives, natural Polymethacrylate ester latices 353-7
rubber latex flow stimulation 24 Polymethyl methacrylate, dispersion
Phospholipids polymerization 295-6
hydrolysis 106-7 Polypeptides, natural rubber latex
natural rubber latex 93-5 aqueous phase 98, 105
Plasticization, vinyl acetate polymers Polypyrrole
and copolymers 336, 337, 342-5 surface modification 572-3
Polyacrylate latices 353-7 synthetic latices 379-80
Polyacrylonitrile, graft- Polystyrene
copolymerization 535-7 graft-copolymerization 535-7
cis-l,4-Polybutadiene, artificial latices Polyurethane artificial latices 408-18,
430-1 434-5
Polychloroprene rubber latices 326-34 acetone process 412-14
films 332-4 crosslinking 416-18
formulation 328-9, 335 isocyanate-terminated prepolymers
hydrolysis 330-4 409-18
non-functionalized 335 ketimine process 412
pH 330-3, 335 melt-dispersion process 414-16
production 328-30 properties 436
storage effects 332-4 Polyvinyl acetate, graft-
viscosity 335 copolymerization 535-7
vulcanization 331 Polyvinyl chloride, graft-
Polyethylene rubbers, copolymerization 535-7
chi oro sulphonated 437-8 Polyvinylmethyl ether, water-soluble
Polyisoprenes hydrocolloid 273-8
acetate precursor 120-4 Positex 563-4
biogenesis 118-31 Post-polymerization additives,
biosynthesis 119-31 emulsion polymerization 198,
cis-I,4-Polyisoprene 247-8
see also Natural rubber latex Potassium hydroxide
ageing resistance 429-30 see also KOH number
artificial latices 420-30 bactericide 41
chain scission 89 preservative 40-1
chemical structure 84 Power-feed emulsion copolymerization
crosslinking 85-7, 89, 426-9, 476, 214-221,222
477,482 Pratt processes, artificial latices
epoxidation 514-6 398-401
extension at break 425-30 CJO-Prenyl pyrophosphates 128-31
films 423-30 Preserva tives
gel content 85-6, 423 aminophenols 37
microgel 86-7 ammonia 27-40
microstructure 84, 115-17 antibiotics 41
modulus 42~30 bactericides 29-30
in natural latices 115-17 boric acid 37-8
particle size distribution 422-3 formaldehyde 40
synthetic latices 420-30 8-hydroxyquinoline 39
tensile strength 423-30 low-ammonia systems 35-40
vulcanization 423-30, 498 McGavack system 40
trans-l,4-Polyisoprene, in natural natural rubber latex 24-41
latices 115-17 pentachlorophenol 35-7
Index 589
potassium hydroxide 41 Residence time, continuous emulsion
sodium sulphite 40-1 polymerization 202-6
tetraalkylthiuram sulphides 38-9 Reticulum, natural rubber latex 104-5
zinc dialkyldithiocarbamates 37, Revertex 70
463-4 Revultex 491-2
Pressure agglomeration 285-9 Rubber hydrochloride 553
Pressure polymers 340
Prevulcanization Saprota achros 115-17
see also Crosslinking; Sulphur- Sedimentation
prevulcanization artificial latices 393-4
acrylonitrile-butadiene rubber natural rubber latex concentration 45
synthetic latices 508-9 Seeded emulsion polymerization
chemically-modified latices 441-511 199-200,232-4
high-energy radiation exposure Semi-continuous emulsion
498-508 polymerization 200,201,237
isobutene-isoprene(butyl)rubber Sequential emulsion copolymerization
artificial latices 509 212-3
natural rubber latex 442-511 Sequestering agents, emulsion
N-nitrosamine levels 493-7 polymerization 196
organic peroxide reactions 492-7 Serum replacement, natural rubber
partial 444 latex concentration 76
styrene-butadiene rubber latex Shorts toppers, emulsion
508-9 polymerization 196-8
sulphurless vulcanizing system Silicoftuorides, natural rubber latex
449-50 gelation 34
Products by process, emulsion Skim rubber 64-5, 69
copolymerization 211 Sodium alginate, natural rubber latex
Protective colloids 149 creaming agent 51
Proteins Sodium n-dodecyl sulphate, colloid
hydrolysis 106-7 stabilizer 172-3
natural rubber latex aqueous phase Sodium sulphite, preservative for
96-100, 105 natural rubber latex 40-1
natural rubber latex particle surface Sols
90,93-5 molecular-size distribution 87-9
Solubilization 156-7
Quebrachitol, natural rubber latex Solution-emulsification
aqueous phase 96 artificial latices 395-7
Radiation see High-energy radiation cis-l,4-polyisoprene rubbers 422
Radiation-prevulcanized natural Solution polymerization 148-9, 161
rubber latex 499-507 Solvents
Reclaimed rubbers, artificial latices particle swelling 266-7,484-5
391,397-8,431-4 sulphur-prevulcanized natural rubber
Redispersible latices latex film swelling 484-5
see also Artificial latices synthetic latex agglomeration 266-72
natural rubber 77-8,391,397-8 Specific surface area, synthetic latices
synthetic 383-7 140
Regulators see Modifiers Spontaneous coagulation
Relaxed-modulus test, degree of carboxylate soaps 25-6, 39
vulcanization 450, 453-4 natural rubber latex 25-6
590 Index
Stability see Colloid stability; Sulphur-prevulcanization
Mechanical stability see also Prevulcanization
Stirred-tank reactors, continuous accelerators 446-7, 460, 462-6,
emulsion polymerization 202-6 470-4,479-82
Stockpunkt concentration process activators 447-8
269-70 colloidal stabilizers 448
Storage crosslinks 459,478,482,484-92
colloid stability effects 106-8, 112-14 degree of vulcanization 450-4
electrical conductivity effects 112 effect of non-rubber constituents
natural rubber latex 105-8, 112-4 476,478-9,480-1,486-91
polychloroprene rubber latices 332-4 films 454-9, 479, 483-91
Stress-strain relations see Tensile formulation 448
strength free sulphur disappearance 474, 475
Styrene history 445-6
copolymerization 186-90, 194-6, insoluble ingredients 465, 467
196-8 kinetics 469-83
emulsion polymerization 145-6, 161, latex particle size 474-6
181,183 mechanism 469-83
graft-copolymerization 525-30, natural rubber latex 442-92
533-5 particle size effects 465, 467
Styrene-butadiene copolymers latex polyisoprene latices 476, 477, 480
305-23 property effects 454-9,479
see also Butadiene reaction system 469-70
agglomeration 264-89 sulphur content 460, 462-3, 484-5
applications 305 sulphurless vulcanizing system
cationic 373-4 449-50
colloid stability 311-12 temperature effects 460-1
concentration 289-93 water content 480
creaming 289-92 zinc oxide 460-4
films 310-11 Sulphur-vulcanization, dry natural
freezing and thawing 279-84 rubber 443, 465, 468-9,476, 480
functionalized 240-60, 358-71 Surface free energy see Free surface
natural rubber latex replacement energy
305-6 Surface modification
particle morphology 231-2 charge reversal 563-6
particle size 262-3 dyeing 566-7
polydispersible products 386-7 for immunological applications
prevulcanization 508-9 567-73
production by emulsion latex particles 563-73
polymerization 186-90, 194-6, Suspension polymerization 149-50
196-8,312-17 Swelling
viscosity 261-3, 320 see also Alkali-thickenable synthetic
wet-gel strength 315 latices
Styrene-ethyl acrylate copolymer, equilibrium-swelling degree of
dynamic modulus 221-3 vulcanization test 450, 452-3
Styrene-vinylpyridine-butadiene cis-l ,4-polyisoprene rubbers 429
terpolymer latices 320-1 solvent agglomeration 266-7
Sulphates, colloid stabilizers 182-3 Synthetic latices 137-390
Sulphonates, colloid stabilizers 182-3, see also Copolymerization
244-6 acrylonitrile-butadiene rubber 508-9
Index 591
agglomeration 260-289 Hevea brasiliensis (Euphorbiaceae)
alkali-thickenable 242, 304-5, 18-22
368-71 incisions 18-22
amphoteric 371-5 puncture tapping 22
aqueous 303-389 Taraxacum koksaghyz 117
carboxylated 239-60 Tensile strength
cationic 371-5 films 455-8,479,483,497,501-9
colloidal stability reduction 265-89 graft-copolymers 533-4
colloid stabilizers 140, 141-2, 150, leaching effects 455-8, 479
153,157-61, 165 natural rubber latex 418, 455-63,
composite particles 380-3 483,491,501-8
concentration 260-4, 289-93 cis-l,4-polybutadiene rubbers 431
core-shell particles 212, 224-30 polychloroprene rubber latices 333
creaming 289-92 cis-l,4-polyisoprene rubbers 423-30
definition 137-8 styrene-butadiene copolymers
dispersion polymerization 923-6 310-11,313-14,508-9
evaporation 292-3 vinyl acetate polymers and
flocculation 163 copolymers 346
free surface energy 140, 184,265, Testing, degree of vulcanization 450-4
274-6,280 Tetraalkylthiuram sulphides,
functionalized 238-60,304-5,358-71 preservative for natural rubber
gradient copolymers 214-221 latex 38-9
high-solids 260-4 Thiolic acids see Carbothiolic acids
hydrophilic polymers 379 Trumbull-Dickson process, artificial
large-particle 232-4 latices 401
latent latex particles 157 Tubular reactors, continuous emulsion
mechanical stability 142 polymerization 202-3, 206
rnicrovoid particles 375-7 Turbulent-film evaporator 293, 294
monodisperse 232-4
mud-cracking 142 Ultrasonic techniques, synthetic latex
non-aqueous 293-6, 303 agglomeration 288-9
particle shape 139 Ultraviolet light absorption
particle size 139, 219-221, 222, functionalized latices 377-9
224-32 molar absorptivity 377-8
particle-size distribution 139, 260-4
cis-l,4-polyisoprene rubbers 420-30, Vacuolar particles, natural rubber latex
476,477 104
polypyrrole 379-80 Vanderhoff-EI-Aasser-Ugelstad
production by emulsion process, artificial latices 406-8
polymerization 137-302 Versatic acids see Vinyl acetate
redispersible 383-7 polymers and copolymers
structured latex particles 211-32 Vinyl acetate polymers and copolymers
viscosity--<:oncentration relationship 334-53
260-4,274-6 chemical structure 336-8
Synthetic natural rubber 1-2 films 343-5, 347
glass-transition temperature 334, 336
Tapered emulsion copolymerization hydrolysis 345-53
212 mechanical properties 344, 346
Tapping plasticization 336, 337, 342-5
bark consumption 21-2 production 338-42
592 Index
Vinyl acetate polymers and copolymers Vulcanization 442
(contd) see also Crosslinking;
production by emulsion Prevulcanization
polymerization 338--42 natural rubber latex 442-3
tensile strength 346 polychloroprene rubber latices 331
Vinyl chloride-vinylidene chloride cis-l ,4-polyisoprene rubbers
copolymers 423-30
glass-transition temperature 358 styrene-butadiene copolymers
production by emulsion 312-15
polymerization 358
Vinyl neodecanoate see Vinyl acetate Water-resistance, synthetic latex films
polymers and copolymers 141-2,242-3
Vinylpyridine, styrene-vinylpyridine- Wet-gel strength
butadiene terpolymer latices 320--1 styrene-butadiene copolymers 315
Viscoids see Lutoids synthetic latices 140--1
Viscosity World War II
alkali-thickenable synthetic latices Government rubber-styrene latices
242, 304-5, 368-71, 375-7 305-6,315,317,319
isobutene-isoprene rubber latices high-energy radiation
420--1 prevulcanization 498
natural latices 116 shortages 11
natural rubber latex 69, 78, 491, synthetic rubber development 152,
501-7,560--1 265,269-70
polychloroprene rubber latices 335
cis-I,4-polyisoprene rubber latices Zinc dialkyldithiocarbamates
428-9 preservative for natural rubber latex
styrene-butadiene copolymer latices 37,463-4
320 sulphur-prevulcanization
synthetic latices 260--4, 274-6 accelerators 464-5, 466, 470--4,
VLAZN latex 39 479-82
Volatile base colloid stabilizers 142 Zinc oxide
Volatile fatty acids natural rubber latex gelation 33
enzyme-catalysed formation 108-12 preservative for natural rubber latex
natural rubber latex aqueous phase 39
96, 106--14 sulphur-prevulcanization 460--4, 471

Polymer Latices - Science and Technology Volume 2 - Types of Latices PDF

Uploaded by

Copyright:

Available Formats

Polymer Latices - Science and Technology Volume 2 - Types of Latices PDF

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

Polymer Latices - Science and Technology Volume 2 - Types of Latices PDF

Uploaded by

Copyright:

Available Formats

JOIN US ON THE INTERNET VIA WWW, GOPHER, FTP OR EMAIL:

lal11 SPRINGER-SCIENCE+BUSINESS MEDIA, B.V

Library of Congress Catalog Card Number: 96-70800

00 Printed on permanent acid-free text paper, manufactured in accordance

10. Synthetic latices: general principles of production 137

10.4.3 Concentration of synthetic latices 289

11. Synthetic latices: individual types 303

12. Artificial latices 391

13. ChemicaUy-modified latices: 1. Prevulcanized latices 441

14. Chemically-modified latices: 2. Types of chemically-modified

This book is effectively the second edition of my book High Polymer

of indicating the precise significance of numerical entries, especially where

D. C. Blackley, Polymer Latices

would herald the demise of the natural-rubber-producing industry; 20 years

and for application in the large-scale manufacturing processes in which it is

9.2 CULTIVATION OF NATURAL RUBBER

9.2.1 The principal rubber tree

9.2.1.1 General description

Fruit Transverse section of fruit Seed.

9.2.1.2 More detailed structure of the mature trunk

Inner cortex or bark

Outer cortex or bark----1

Transverse section Longitudinal section

9.2.2 The HeJ'ea brasiliensis plantations

9.2.2.1 Conditions required for the growth of Hevea brasiliensis

9.2.2.2 Regions of the world where Hevea brasiliensis is found

9.2.2.3 Outline of the history of the Hevea brasiliensis plantations

development of the plantation industry occurred. By 1938, some 97% of the

9.2.3 Propagation of Hevea brasiliensis

are therefore preferred when it is desired to replicate a variety true-to-type,

9.2.3.2 Propagation by seed

Secondary roots /Taproot

9.2.3.3 Vegetative propagation

Cutting of budpatch Budwood after

Development of viable bud-graft 'Elephant's foot'

having selected canopy characteristics must be set the obvious economic

9.2.4.2 Early tapping systems

(a) (b) (c)

9.2.4.3 Modern tapping systems

9.2.4.4 Factors which affect the yield of latex and rubber

9.2.4.5 Other methods of tapping

9.2.5 Improvement of latex yield

9.2.5.1 By selective breeding

development and intensive planting of a new variety. This period may be

9.2.5.2 By yield stimulation

9.3 PRESERVATION OF NATURAL RUBBER LATEX

9.3.1 Preliminary considerations

9.3.1.1 Why it is necessary to preserve natural rubber latex

As such, it is a complex biochemical system. It is therefore not surprising

9.3.1.2 Mechanism of spontaneous coagulation of natural rubber latex

3. Spontaneous coagulation, and its sensitization by carboxylate soaps,

9.3.2 The ideal preservative for natural rubber latex

9.3.2.1 Efficacy as a preservative

1. It should destroy latex microorganisms, or at least suppress their activity

9.3.2.2 Ancillary requirements

9.3.3 Ammonia as a preservative

9.3.3.1 Historical aspects

9.3.3.2 Ammoniation of natural rubber latex