G. Pritchard - Developments in Reinforced Plastics - Resin Matrix Aspects-Springer Netherlands (1980)

DEVELOPMENTS IN
REINFORCED PLASTIC8-1
Resin Matrix Aspects
THE DEVELOPMENTS SERIES
Developments in many fields of science and technology occur at such a pace

that frequently there is a long delay before information about them
becomes available and usually it is inconveniently scattered among several
journals.
Developments Series books overcome these disadvantages by bringing
together within one cover papers dealing with the latest trends and
developments in a specific field of study and publishing them within six
months of their being written.
Many subjects are covered by the series including food science and
technology, polymer science, civil and public health engineering, pressure
vessels, composite materials, concrete, building science, petroleum
technology, geology, etc.
Information on other titles in the series will gladly be sent on application to
the publisher.
DEVELOPMENTS IN
REINFORCED PLASTIC8-1
Resin Matrix Aspects
Edited by
G. PRITCHARD
School of Chemical and Physical Sciences,
Kingston Polytechnic, Kingston upon Thames, Surrey, UK
APPLIED SCIENCE PUBLISHERS LTD

LONDON
APPLIED SCIENCE PUBLISHERS LTD
RIPPLE ROAD, BARKING, ESSEX, ENGLAND
British Library Cataloguing in Publication Data

Developments in reinforced plastics.-(Developments
series).
I :ResIn matrix aspects
I. Reinforced plastics
I: Pritchard, Geoffrey II. Series
668.4'94 TPII77
ISBN-13: 978-94-009-8726-5 e-ISBN-13: 978-94-009-8724-1
001: 10.1007/978-94-009-8724-1
WITH 69 TABLES AND 87 ILLUSTRATIONS
© APPLIED SCIENCE PUBLISHERS LTD 1980

Softcover reprint of the hardcover 1st edition 1980
All rights reserved. No part of this pubhcatlOn may be reproduced, stored In

a retrIeval system, or transmItted In any form or by any means, electromc,
mechanical, photocopying, recording, or otherwise, without the prIor
WrItten permISSIOn of the pubhshers, Apphed Science Pubhshers Ltd,
RIpple Road, BarkIng, Essex, England
FOREWORD
It needs only modest foresight to see that the future of plastics lies
in their use as composites. The article made of plastics alone will, I
believe, become a rarity. At the less critical end of the applications
spectrum, users will 'dilute' plastics with cheap, solid fillers. Unstressed
neutral axes will be foams, for air is a splendidly cheap filler. From here, it is
a short stepto get the diluent to reinforce, or alter, the material in someway.
Reinforcement is needed because plastics tend to be weak, to lack stiffness
and, above all, to creep. Glass fibre, especially, offers inherent cheapness,
strength and stiffness and, although almost unusable by itself, it shows real
advantages when married with a matrix. Greater use of filling and
reinforcement will bring into prominence a relatively neglected group of
polymers-the thermosets and chemisets. The great range of potential
properties, obtained from these materials with their varied chemistry,
makes them well suited for use in composites and they must surely move
nearer to parity in tonnage with thermoplastics.
The key to all filling and reinforcement is to retain the one really
outstanding property of plastics-their ease of fabrication. This is a
property which is no less important for not being quantifiable. Consider
the production of complex articles; these may be made in only one
step using injection moulding, how many steps would it take with metal?
Look at a GRP moulding made by hand lay-up; consider the complexity
of making the same article in metal, in wood, or even in concrete. It
is easy to go over the top, chipboard is an example. It has less than
10 % polymer, and uses, otherwise, waste wood, but the real ease of
fabrication is lost as it has to be cut, drilled and fastened like wood
itself, not like plastic.
v
vi FOREWORD
Plastics are complex materials. Attempts to use them without employing

even our present inadequate knowledge of their science leads rapidly
to disaster. Changes in morphology due to such factors as cooling, crystal-
lisation, molecular orientation, or weld lines, can make one quite afraid
to push melt into mould! The situation with composites is much more
complex. To the existing factors is added the nature of the bond with
the second material, i.e. the nature of the interface, the influence of
another material on the polymer morphology, and the different mechanism
of absorbing stress and of eventual failure.
The materials scientist, like other scientists, tends to attempt only
that which he thinks has a good chance of success. He has tended to
keep away from composites. Look at the relative amounts of literature
on crack formation and propagation in unfilled and filled plastics! Con-
sider our inadequate understanding of basic data on the match needed
between matrix stiffness and fibre stiffness. If we have a thin interfacial
layer with different properties to either fibre or matrix, then there is
virtually no information on what those properties would best be.
This situation must change. The scientist, or the technologist, driven
by what he has to do, must move to an intensive attack on the science
of composites. Indeed, in the case of high performance materials for
military applications, this is already happening.
There is a need to set out concisely, but precisely, where we start
from. This series of volumes attempts to do just that, and is timely and
welcome.
A. A. L. CHALLIS
Polymer Engineering Directorate,
Science Research Council,
London, UK
PREFACE
The subject of reinforced plastics is a fast-developing one. It is hoped

that this book, and subsequent volumes, will help readers to keep up
to date with some of the more important changes taking place. One
of the benefits of a book such as this one may be that it encourages
the consideration of reinforced plastics as a multidisciplinary field; some
of the chapters are distinctly chemical, others discuss engineering proper-
ties, while some are a mixture. It should be beneficial if people engaged
in research and development attempt to understand the relationships
between synthesis, structure, properties and applications.
Perceptive readers may note one or two places where different authors
give different opinions about the same matter. This is a fair reflection
of the degree of uncertainty still persisting in our understanding of fibre-
resin composites.
In his Foreword, Dr Challis stresses the importance of fabrication
in determining the future of plastics. This subject is a large one and
deserves a book to itself. Indeed it is hoped that a subsequent volume
in this series will be devoted to that subject.
The editor would like to thank all the authors who have helped by
contributing chapters, and also to acknowledge the part played by their
companies and universities, etc., in facilitating publication.
G. PRITCHARD
VIi
CONTENTS
Foreword v
Preface vii
List of Contributors Xl
1. Thermosetting Resins for Reinforced Plastics

G. PRITCHARD
2. Vinyl Ester Resins 29

THOMAS F. ANDERSON and VIRGINIA B. MESSICK
3. Polyester Resin Chemistry 59

T. HUNT
4. Phenol-Aralkyl and Related Polymers 87

GL YN I. HARRIS
5. Initiator Systems for Unsaturated Polyester Resins 121

V. R. KAMATH and R. B. GALLAGHER
6. High-Temperature Properties of Thermally Stable Resins 145

G. J. KNIGHT
7. Structure-Property Relationships and the Environmental

Sensitivity of Epoxies . 211
ROGER J. MORGAN
ix
x CONTENTS
8. Some Mechanical Properties of Crosslinked Polyester Resins 231

W. E. DOUGLAS and G. PRITCHARD
9. Crack Propagation in Thermosetting Polymers 257

ROBERT J. YOUNG
Index 285
LIST OF CONTRIBUTORS
THOMAS F. ANDERSON
Senior Research Specialist Resins TS & D, Dow Chemical USA,
Texas Division, Freeport, Texas 77541, USA
w. E. DOUGLAS
School of Chemical and Physical Sciences, Kingston Polytechnic,

Penrhyn Road Centre, Penrhyn Road, Kingston upon Thames,
KT12EE, UK
R. B. GALLAGHER
Project Leader-Applications, Lucidol Division, Pennwalt Cor-
poration, PO Box 1048, Buffalo, New York 14240, USA
GL YN I. HARRIS
Advanced Resins Ltd, Unit 8, Llandough Trading Estate,
Penarth Road, Cardiff, South Glamorgan, UK
T. HUNT
BP Chemicals Limited, Research and Development Department,

South Wales Division, Sully, Penarth, South Glamorgan CF6 2YU,
UK
v. R. KAMATH
Lucidol Division, Pennwalt Corporation, PO Box 1048, Buffalo,
New York 14240, USA
Xl
Xli LIST OF CONTRIBUTORS
G. J. KNIGHT
Ministry of Defence PE, Materials Department, Royal Aircraft
Establishment, Farnborough, Hants, UK
VIRGINIA B. MESSICK
Dow Chemical USA, Texas Division, Freeport, Texas 77541,

USA
ROGER J. MORGAN
Lawrence Livermore Laboratory L-338, University of California,
PO Box 808, Livermore, California 94550, USA
G. PRITCHARD
School of Chemical and Physical Sciences, Kingston Polytechnic,
Penrhyn Road Centre, Penrhyn Road, Kingston upon Thames,
KT12EE, UK
ROBERT J. YOUNG
Department of Materials, Queen Mary College, University of
London, Mile End Road, London E1 4NS, UK
Chapter 1
THERMOSETTING RESINS FOR REINFORCED

PLASTICS
G. PRITCHARD
Kingston Polytechnic, Surrey, UK
SUMMARY
The resins used in the reinjorced plastics industry are mainly of the thermo-
setting variety. This means that processing and fabrication operations
are accompanied by chemical crosslinking reactions, which make the pro-
duction of moulded objects less straightforward than is the casefor thermo-
plastics. However, many oj the desirable physical and thermomechanical
properties oj thermosetting resins derive jrom their crosslinked structure.
Vanous classes oj resin are discussed in bnef outline. The common
jeature oj these resins, their three-dimensional network st ructure, renders
analysis and characterisation difficult, but neW physical and chemical tech-
niques are helping polymer scientists to achieve a better understanding
oj resin structure and properties. These techniques, together with other
technical advances,for example in processing technology, Will be important
in determining the juture oj thermosetting resins.
1.1. INTRODUCTION
This book is mainly about resins-their synthesis, structure and properties.

It is difficult to discuss resins without any reference at all to their practical
uses, and without considering the fabrication procedures by which they
are converted to marketable products. Nevertheless, it is going to be
assumed that these important topics will receive much closer attention
I
2 G. PRITCHARD
in later volumes, so the emphasis here is on chemical and physical or

mechanical aspects.
It seems best to concentrate on thermosetting resins because, whatever
the future may hold, thermoplastics constitute only about 10% of the
present reinforced plastics matrix usage. Even this share is heavily
dependent on a single market, i.e. automotive components.
Basic concepts and terminology will be introduced here as a preparation
for the later, more specialised chapters.
1.2. THE FUNCTION OF THE MATRIX
Every component of a composite material has its function. The reinforce-

ment carries mechanical stresses, imparting resistance to creep, together
with toughness, strength and stitTness. For example, the Young's modulus
Ec of the composite, may be considered (as a first approximation) to
be
Ec = KVrEr + VrnErn
where Vr and Vrn refer to the volume fraction of fibre and matrix respec-
tively, and Er and Em refer to the moduli of fibre and matrix. K is
an orientation factor, which is much greater for unidirectional composites,
with alignment parallel to the principal stress axis, than for randomly
oriented fibre arrays.
The reinforcement can operate properly only if there is good adhesion
between it and the matrix, to facilitate load transfer. On the other hand
very good adhesion reduces the ability of fibres to stop cracks. A balance
must be found.
Fibrous reinforcements are vulnerable to mechanical damage, and it
is a function of the resin to protect the fibres as well as to bind them
together. Sometimes fibres also require protection against short-term
chemical attack (although most resins available at present are as much
atTected by common chemical environments, especially organic solvents,
as the fibres; but there are exceptions).
1.3. SUPRAMOLECULAR STRUCTURE
Thermosetting resins are chemically reactive substances which undergo

hardening to produce insoluble, infusible products. The three-dimensional
THERMOSETTING RESINS FOR REINFORCED PLASTICS 3
network structures resulting are much more dense than those 0btained
during the vulcanisation of rubber. But it is still uncertain whether the
crosslink densities of common crosslinked resins are (i) approximately
uniform (as in Fig. l.l(a», or (ii) non-uniform but with interconnecting
chain segments (Fig. l.l(b». The widely held view is that the latter is
more likely. A third possibility is that the network consists of densely
------------------- ---'
I
I
I
I
I
b I
I
I
I
'\..N~'!- I
FIG. 1.1. Three models representing the structure of network polymers. (a)
Continuous network of uniform density, (b) continuous network of non-uniform
density, (c) discontinuous network.
crosslinked regions dispersed in non-bound matter. Figure l.l(c) shows

the dense micelles without the non-bound component. This third model
has been suggested for epoxy resins/ the evidence against the first model
comes from electron microscopy, 2 and from direct visual 0 bservation
of swelling processes. 3 (See also Chapter 7.)
The structure of thermosetting resins is amorphous, and the behaviour
of most resin matrices is that of rather brittle organic glasses. (Brittleness
increases with crosslink density.)
1.4. PROPERTIES OF THERMOSETTING RESINS
Prior to crosslin king, thermosetting resins may be viscous liquids (in

practice, they may be handled as dilute solutions); or they may be glassy
solids which have been powdered and, perhaps, mixed with additives.
4 G. PRITCHARD
The change from the linear, uncrosslinked, soluble, fusible state to the
true thermohardened network polymer occurs at a point called gelation.
Liquid resins or resin solutions need to undergo flow before gelation
occurs, so as to adopt the mould shape. Viscosity is, therefore, of
prime importance and many studies have been made of viscosity-temper a-
ture relationships. There is typically a departure from the classical
behaviour of Newtonian fluids, especially at high shear rates. 4 The
crosslinking or curing process is not necessarily carried out at high
temperatures, and if a low-temperature cure is envisaged, the viscosity
of the resin must initially be low.
After gelation, the resin becomes a soft, weak gel and later, slightly
rubber-like, before finally becoming a relatively hard solid. Heating the
final product sometimes shows up a relaxation, after which rubber-like
behaviour again becomes apparent, but chemical decomposition can mask
the onset of this development, especially with the more dense networks.
Whatever model is taken to represent network structures, a small 'sol'
fraction is invariably found, which can be estimated by solvent extraction.
It would be unwise to generalise about the properties of thermosetting
resins, but most are insufficiently tough for rough handling, and their
practical uses are restricted unless they are reinforced in some way. One
important respect in which thermosetting resins perform relatively well
is in hardness; another is their maximum working temperature. In both
FIG. 1.2. Scanning electron micrograph of the surface of a pigmented, decorative

laminate with a melamine-formaldehyde resin veneer, scratched with a diamond.
these cases they are rather better than thermoplastics, and these two
advantages are utilised in, for example, the domestic kitchen worktop.
Nevertheless, there is room for improvement in both respects. Figure 1.2
shows a scanning electron micrograph of a scratched piece of white
melamine-formaldehyde paper-based decorative laminate.
The properties of some unreinforced thermosetting resins are given
in Table 1.1. It will be seen that there is a very wide range of properties
TABLE 1.1
PROPERTIES OF UNFILLED, CROSSLINKED THERMOSETTING RESINS
Property Units Unsaturated Glycidyl ether Phenol

polyester epoxide formaldehyde
Tensile strength MNm- 2 25-80 30-100 20-65

Compressive strength MNm- 2 60-160 60-190 45-115
Tensile modulus GNm- 2 2'5-3·5 2,5-6,0 2·0-6·5
Elongation at break % 1'3-10'0 1,1-7,5 1·0-3·5
Flexural strength MNm- 2 70-140 60-180 45-95
Flexural modulus GNm- 2 2'5-3·5 1,8-3,3 2·5-6·5
Poisson's ratio c.0·35 0·16--0·25"
Specific gravity 1,11-1,15 1·15-1·25 1·31
Water absorption (24 h) mg 10-30 7-20 15-30
d.c. volume resistivity ohm-cm 10 12_10 14 10 10-10 18 109 _10 14
" Calculated.
for a given resin, because of differences in precise formulation and

structure. Even greater variations are found in the data sheets of commer-
cial manufacturers, because these refer mostly to filled or reinforced
compositions, and depend on the nature, orientation and volume fraction
of the second phase, as well as on interfacial bound strength.
Table 1.1 is inadequate in that it gives no indication of how easily
one specified property can be reconciled with another. In fact it is com-
monly found that tensile modulus decreases as the %elongation at break
increases, so the highest values quoted for these two properties cannot
actually be combined in the same material. It is usually found that high
heat distortion temperature, hardness and modulus are combined with
increased brittleness, i.e. lower impact strength.
Modification of resin properties is achieved partly by alteration of
chemical formulation and partly by the use of particulate, fibrous or
laminar reinforcement (see Table 1.2). There is such a variety of reinforce-
ment forms that quoted values can differ greatly even when the materials
0--
TABLE 1.2
PROPERTIES OF FILLED THERMOSETTING RESINS
Property Units Material

Minerai filled Glass fibre Woodflour Cellulose Paper filled
alkyd filled alkyd phenolic filled phenolic melamine Cl
'1:1
Tensile strength MNm- 2 20-35 25-75 24-35 30-40 45-75 :;:tl
Compressive strength MNm- 2 65-90 100-130 55-70 130-180 140-300 =l

(")
Flexural strength MNm- 2 45-70 100-150 24-60 55-75 65-95 :t
>
:;:tl
Specific gravity 1·8-2·5 1·8-2·2 1·3 1-4 1·4-1·6
0
Volume resistivity ohm-<:m 10 13 _10 15 10 13_10 15 10 12_10 13 10 11 _10 13 10 8 _10 11
Power factor (I MHz) 0·012-0·016 0·019-0·021 0·02-0·05 0·035-0·05 0·02-0·04
Dielectric constant (I MHz) 4-5 4 6·0 4·8-5·4 6-8
Dielectric strength volts/mil 180-320 200-300 50-250 70-120 60-200
Water absorption (24 h) mg 7-75 10-25 30-65 30-60 10-50
are ostensibly similar. However, certain limitations are apparent in all

the common thermosetting resins. There is a need for improved heat
resistance (measured not only by the onset of chemical decomposition
but also by the temperature at which mechanical properties are adversely
affected). The present generation of thermosetting resins is also inadequate
in resistance to combustion, to plasticisation or softening by solvents,
liquids or water vapour, and to oxidation. But major advantages have
been demonstrated by reinforced thermosets in their resistance to dilute
aqueous inorganic fluids, and some other chemicals, at moderate tempera-
tures-resulting in their extensive use for chemical process equipment. s .6
1.5. PHENOL-FORMALDEHYDE RESINS
Phenol condenses with aldehydes to give products which can be either

linear and soluble (provided that phenol is in excess, and the pH is
low) or branched, sometimes crosslinked, and insoluble (if formaldehyde
is in excess). The linear, soluble products are called novolaks, while the
multifunctional products from formaldehyde-rich formulations result in
resoles. The novolaks can be safely heated while remaining soluble, and
they can only be crosslinked by the addition of further aldehyde. So
the novolak moulding powder will contain a hardening agent such as
hexamethylene tetramine:
This compound decomposes on heating to generate ammonia and

formaldehyde.
The first stage in novolak formation is the slow reaction between phenol
and (usually) formaldehyde, to give ortho and para methylol phenols:
8 G. PRITCHARD
OH
© and
Further addition to the ring does not occur, because condensation with
phenol is more rapid:
OH OH OH
~CH'OH
+
©¢ ---+
CH 2
etc.
~OH
Several two-ring products can form, but the major products are:
OH OH
~ ~
CH 2 and CH z
¢ ~OH
OH
These intermediates undergo further reaction with formaldehyde and

phenol alternately, eventually reaching six or seven rings in length, but
without crosslinking. A typical novolak has a structure such as:
OH
& OH
~, )Q>-CH'~CH'-@-OH
& CH, HO
~CH,--@-OH
OH
Crosslinking with hexamethylene tetra mine brings about the formation

of additional -CH 2 0H groups and their mutual condensation, elimina-
ting water, to give complex network structures with -CH 2 - and
-CH 2 0CH 2- bridges between aromatic rings.
Resole formation can lead to crosslinked products, because addition
of aldehyde to phenol occurs rapidly, and results in reactive polyfunctional
intermediates such as:
These molecules undergo self-condensation to form insoluble, infusible

networks. A strong acid catalyst will bring about the curing of a thin
film of resole at ambient temperatures.
Cold curing is used to produce castings, which can be made water-white.
It is found that hot-cured phenolic resins are generally dark, and this
affects their application for domestic use. The dark colour is believed
to be the result of a reaction between the oxygen of -CH 2-O-CH 2
bridges and two neighbouring hydrogen atoms of phenolic OH groups:
10 G. PRITCHARD
This reaction takes place when the temperature reaches 160°C and becomes
rapid at 180 0c.
Phenol-formaldehyde resins are used, with fillers, as moulding com-
pounds (sand is used for foundry moulds, alumina for grinding wheels,
asbestos for friction linings, glass and mineral wool for insulation, wood-
flour for general purpose applications, cotton flock for impact resistance).
A typical formulation might be:
Resin 35 to 45 parts by weight

Hardener 6 parts by weight
Filler 45 to 55 parts by weight
Magnesium oxide 1 to 2 parts by weight
A stearate 1 to 2 parts by weight
A plasticiser 1 to 2 parts by weight
A dark pigment 1 to 2 parts by weight
Alternatively resins can be dissolved in solvent so as to impregnate

cloth or paper for the production of decorative laminates (using a
melamine-formaldehyde veneer) or of industrial laminates (such as printed
circuit boards).
The last decade has seen a great advance in the injection moulding
of phenolic resins. This is more economical than compression moulding
because of sharply reduced cycle times, and lower labour costs. About
60 %of phenolic moulding powder consumed in some European countries

is already injection moulded, although in the United Kingdom the figure
is nearer 35 % (in 1979). There are several problems; the resins are not
easily obtained in convenient granular form, and tend to suffer from
inconsistent properties. Long barrel life is desirable, but almost incom-
patible with fast cure cycles unless there is a cure 'trigger' temperature
of around 120°C. Injection moulding imparts fibre orientation, but the
use of 'injection-compression'-i.e. injection into a slightly opened com-
pression mould-allows for the recovery of anisotropy. There is still
a need for compounding techniques which do not shorten fibre length
too much.
Space has been devoted to phenolic resins because (a) they have under-
gone surprising adaptation in the past decade, and (b) they have certain
promising advantages, which could maintain these oldest of the true
synthetic plastics in a competitive position while other materials are
adversely affected (see Section 1.17). They are now marketed for hand
lay-up, spray-up, and filament-winding as well as for foams. Their excellent
fire-retardant properties and low smoke emission offer great advantages
over many competitive plastics. On the debit side, the high concentration
of polar groups results in moisture uptake, and rather poor electrical
property retention.
1.6. AMINO RESINS
The reaction of aldehydes (again, usually formaldehyde) with amines

or amides gives rise to crosslinked products. Several amines and amides
have been tried, but only two starting materials are of commercial impor-
tance. These are urea (Structure (I)) and melamine, i.e. 1,3,5-tri-amino-
2,4,6-triazine (Structure (II)). Both can be obtained from non-petroleum
sources.
NH2
I
c
N/""N
I II
c c
/'\-/""
H2N N NH2
(II)
12 G. PRITCHARD
Amino resin synthesis is similar to that for phenolic resins. Excess

formaldehyde can lead to reactive intermediates with up to four methylol
groups in urea (Structure (III)) or six in melamine (Structure (IV)). Acidic
conditions favour the formation of polyfunctional intermediates. Subse-
quent crosslinking leads to network structures with bridges such as:
(a)
CH2-~
/
(b) ~N
"" CH2-~
CH2-~
/
~N
(c)
""/ CH 2
~N
"" CH2-~
Both urea and formaldehyde resins are used as moulding powders, with
formulations which include a filler and a hardener. The number of fillers
used with urea resins is limited, because of problems in achieving a
wide range of colours, whereas melamine resins can be produced with
various colours and also have the advantages of superior hardness, heat
resistance, water absorption characteristics and stain resistance.
Urea resins are widely used for adhesives, in chipboard manufacture,

and as foams. These applications are mainly outside the field of reinforced
plastics as it is generally understood. On the other hand, melamine resins
are important in the production of electrical grade laminates. Together,
the amino resins constitute one of the world's major categories of thermo-
setting resin.
1.7. SILICONE RESINS
These are the only resins discussed in this chapter which are not carbon-
based. Despite the availability of silica, they are expensive and therefore
very limited in their present use. The preparation of silicone resins is
by condensation reactions, as illustrated below:
(I) Monochlorosilanes give hexamethyldisiloxane:
condensatIOn
---+. (CH 3hSi-O-Si(CH 3h + H 2 0
(2) Dichlorosilanes give linear polymers:
CH 3 CH 3
I hydrolysIS I
CI-Si-Cl + 2H2 0 --~. HO-Si-OH + 2HCI
I I
CH 3 CH 3
condensation
---....
< I.
CHj
SI
I
CH 3 n
(3) Trichlorosilanes give network polymers:
CH 3
I
CI-Si-CI
I
Cl
14 G. PRITCHARD
Because of their Si-O and Si-CH 3 bonds, the silicone resins have
the advantage of good heat resistance, but their mechanical properties
are poor. This is believed to be partly because crosslinking has to compete
with cyclisation reactions, so the final network is not comparable with
those of organic carbon-based resins. They have very good water resistance,
and electrical properties, and can be blended with organic resins, provided
that they (the silicones) contain sufficient phenyl substituents. As a result
of their electrical properties, silicone-glass laminates are used for printed
circuit board manufacture, electric motor components and transformer
formers.
It is interesting to note that their poor mechanical properties are also
reflected in the laminates. This demonstrates the importance of the matrix
in composite materials.
1.8. UNSATURATED POLYESTER RESINS
Unsaturated polyester (U P) resins constitute an important and growing

sector of the thermosetting resin industry. The resins are produced by
condensation of dibasic acids (including one unsaturated acid or
anhydride) with diols, followed by crosslinking with a reactive diluent,
usually styrene. Most UP resins use phthalic anhydride, maleic anhydride
CH 3
Ho-@-?-@-OH
CH 3
(4,4' dihydroxy, 2,2' dipheny\propane) (bispheno\ A)
and propane-l,2-diol, although 'bisphenol A' and its hydrogenated,

alicyclic analogue are also used, and aliphatic dibasic acids or long-chain
diols may be added. The crosslinking reaction produces no volatile pro-
ducts, provided that the reaction mixture does not become hot enough
to evaporate the styrene. It is, therefore, possible to produce mouldings
at ambient temperatures, without pressure, i.e. by 'contact laminating'.
This is one of the reasons for the adoption of UP resins in very large
structures such as swimming pools and large hulls. However, low-
temperature cure is slow, and high-temperature, rapid cures require
pressure. The unfilled resins are brittle and most applications are for glass
laminates. Two of the most successful fields are the marine market and the
chemical process equipment sector.
Premix compounds have been developed for the compression, transfer

and, recently, injection moulding of intricate parts. These are called dough
or bulk moulding compounds. Sheet moulding compound is similar,
but contains a thickening agent, and is used for shallow-profile mouldings
such as vehicle body parts, furniture and trays.
UP resins are not fire resistant unless specially formulated. These
formulations are effective but expensive, and they also give uncertain
weathering resistance. Some of the attempts to optimise fire retardancy
have been summarised by Wilson.7 Other advances in UP chemistry
are described more fully in Chapter 3, and are also discussed by Bruins. 8
1.9. EPOXY RESINS
Epoxy resins 9 are like polyester resins in having no volatile reaction

(cure) products, although there may be a volatile solvent present. They
have superior mechanical and electrical properties, but are more expensive
than polyester resins and present greater fabrication problems, at least
in the field of large glass laminate products.
Epoxy resins are compounds containing two or more epoxy groups,
the commonest being the reaction product of epichlorhydrin with 4,4' -di-
hydroxy-2,2' -diphenyl propane:
ClCH 2 · CH--CH 2
""/
o CH
+ 3
HO---IQ\-J-0-0H + CH 2-CH-CH 2 · Cl
~~ ~/
CH 3 0
Larger molecules may be produced, but the average molecular weight

is low. Hardening occurs by additive reactions, opening the epoxide ring,
using primary or secondary amines, amides or anhydrides; or catalytically,
e.g. by tertiary amines, or boron trifluoride.
16 G. PRITCHARD
FIG. 1.3. Heavy duty filled epoxy resin screed, able to withstand tracked military
vehicles. (Courtesy of Structoplast Ltd, Leatherhead, England.)
FIG. 1.4. Low viscosity epoxy adhesive, specially formulated for the repair of
cracks in concrete structures. (Courtesy of Structoplast Ltd, Leatherhead,
England.)
Sometimes substantial quantities of hardener are required, and the

nature of the hardener has an important effect on resin properties. Some
hardeners are effective only at elevated temperatures.
Other epoxide resins include epoxy-novolaks; cycloaliphatic epoxy
resins (formed by epoxidising cycloaliphatic intermediates such as dicyclo-
pentadiene); acyclic (chain type) aliphatic epoxides; epoxidised polybuta-
diene, and epoxidised drying oils.
The range of possible properties of epoxy resins is wide. Apart from
the base resin itself, additives such as diluents, flexibilisers and reactive
rubbers are added.
Epoxide resins are widely used for high performance laminates for
first class mechanical and electrical applications. They also have a range
of other uses: adhesives, coatings, abrasion-resistant floorings and road
surfaces. Figure 1.3 shows a specially formulated, highly filled epoxy
resin screed, able to withstand not only ordinary traffic but tracked
military vehicles. There has been a growth in the use of epoxy resins
for the repair of concrete structures; this too requires special formulations
(see Fig. 1.4).
1.10. VINYL ESTER RESINS
These resins have chemical structures intermediate between those of

epoxides and polyesters. They are similar to epoxides at the prepolymer
stage, except for the presence of terminal unsaturation. This un saturation
facilitates polyester-type crosslinking with styrene, thus making hand
lay-up fabrication practicable. These resins are discussed fully in Chapter 2.
1.11. FURAN RESINS
Furan resins have excellent chemical resistance. This has led to their
useful employment as anti-corrosion linings in chemical plant. Unfortu-
nately, they have fabrication difficulties, and have met with little success
as laminates or moulding compounds. Recently, spraying equipment has
been successfully adapted for use with furan resins, and this may lead
to increased growth in the future.
The starting material is furfural, obtained from oat husks and corn
cobs by digestion with sulphuric acid and steam. It is then hydrogenated
to furfuryl alcohol (Structure (V».
18 G. PRITCHARD
HC-CH
II ~
""/""
HC C
o CH 20H
(V)
which self-condenses on heating with acid, giving Structure (VI):
The resins tend to crosslink in bulk unless exothermic heat is removed,

and normal crosslinking is achieved by addition of 4 %p-toluene sulphuric
acid, causing a loss of some of the unsaturation; the final product can
be represented by Structure (VII):
~CH'
~VCH2
(VII)
The furan resins not only have chemical resistance, but also withstand
heat well, and burn only with difficulty at high temperatures, producing
no smoke problem.
1.12. POLYIMIDES
Polyimides have been developed for high temperature applications where

cost is of secondary importance. In supersonic aircraft, for example,
the temperature of the outer walls is proportional to the square of the
velocity, and at Mach 3, the maximum temperature would exceed 300°C.
Cyclic-chain structures such as Structure (VIII) (where R may be aliphatic
or aromatic) possess satisfactory physical and thermomechanical proper-
ties. But not all polyimides are genuinely thermosetting.
-tN/C~R/C-ZN-Rlt
""'/""'/
CO CO
(VIII)
n
First produced fairly early in the twentieth century, polyimides were

developed commercially by Du Pont Co. in the USA in the late 1950s.
Reaction of a diamine with the dianhydride of a tetracarboxylic acid,
in a polar solvent, gives a polyamic acid, e.g. Structure (IX),
COOH
NHC~CONH
COOH n
(IX)
The formation of the polyimide is then achieved by a further reaction

to remove water from the poJyamic acid, either by heating or by a
dehydrating reaction, to give:
Polyimides may be cast as films while still in the intermediate form,

and subsequently baked to produce the final product. But the polyamic
acid (Structure (IX)) is hydrolytically unstable and it melts at a temperature
very close to that needed for conversion to the polyimide; so fabrication
is not easy. If the melt impregnation is obviated by the use of solvents,
another difficulty arises; the most suitable solvents are polar, difficult
to remove, and responsible for void formation. Voids cause a serious
deterioration in the mechanical properties.
Crosslinking by radical reactions has been developed. This avoids the
20 G. PRITCHARD
problem of volatiles and requires a crosslinking agent such as Structure (X)

together with the intermediate product from the reaction of Structure (XI)
with Structure (XII):
QX
I
CO",-
°
CO/
(X)
H2N-@-CH2~NH2
(XII)
Polymerisation in situ (with the reinforcement already immersed in the

original reagents) has been tried. This gives water elimination problems.
Several other routes are still being developed.
The addition polyimides suffer from, as yet, inadequate toughness
and thermo-oxidative stability, and from a susceptibility to micro-cracking
during cure. The thermoplastic varieties require a high temperature and
pressure for fabrication, and suffer from creep at high temperatures.
1.13. OTHER RESINS
Many exotic high-temperature polymers have been developed.!O The more

practical ones are discussed in Chapter 6, and one particular class of
resin having moderately good thermal stability is discussed in Chapter 4.
All the examples of laminating resins mentioned so far include several
polar groups. Apart from the silicones, none have really low water absorp-
tion capacity. This is a limitation not only for their electrical properties,
but also for their retention of mechanical properties under hot, humid
conditions.
The bis-diene resins, developed in the 1960s, represented an attempt
to produce crosslinked, non-polar resins.!! They were made by heating
oligomeric bis-cyclo-pentadienyl compounds to produce monocyclopenta-

dienyl radicals, which then reacted with the remaining oligomer to give
a crosslinked product.
1.14. IMPROVEMENTS IN RESIN PROPERTIES
As implied in the previous paragraph, there is a need for resins with

improved resistance to hygrothermal ageing. Attempts are also being
made to achieve a higher extensibility without sacrificing modulus, and
to obtain improved hardness and heat distortion temperature without
loss of toughness.
Attempts have been made to toughen epoxide and polyester resins
by addition of elastomers, notably, carboxyl-tipped butadiene-acrylo-
nitrile rubbers,12 and urethane rubber. 13 Polyethersulphones have also
been tried. The problem is to achieve the correct particle size distribution
and the right degree of compatibility between phases. Success has been
very limited indeed with polyesters, and although the toughness of epoxy
resins has certainly been increased by carboxyl tipped nitrile addition,
the viscosity-pressure characteristics leave room for improvement. More-
over, improvements in matrix toughness are not always reflected in
improved composite properties. The addition of any second phase affects
weathering, UV absorption, and other aspects of durability.
Perhaps the most important single improvement sought in the past
decade for thermosetting resins has been the non-burning polyester resin.
It is unfortunate that the material most suited for the convenient fabrica-
tion of fire risk products such as boats, caravans, car bodies, building
panels, electrical components and chemical process equipment is
combustible. Polyesters are also prone to smoke generation.
Halogen-containing derivatives are increasingly used as starting
materials for the synthesis of polyesters, and additives such as antimony
oxide are employed to react with the halogen during combustion. The
choice of halogenated derivatives is restricted to halogenated acids or
diols capable of being esterified without loss of halogen during the
synthesis. Some possible candidates have to be eliminated on the grounds
of cost, and others do not provide sufficient halogenation (chloromaleic
acid, for instance). Two acids, tetrachlorophthalic anhydride, (Struc-
ture (XIII)) and 'HET' acid (chlorendic anhydride, Structure (XIV)) have
become established choices.
Diols prepared from decachlorodiphenyl and similar phenyl compounds
22 G. PRITCHARD
CI
CI
CI~CO"o
CI¥CO/
CI CI
(XIII) (XIV)
result in polyesters of exceptionally high chlorine content. 14 (See also

Chapter 3.)
Disadvantages of some fire-retardant systems can be seen in the
increased rate of photodegradation, causing yellowing and eventual
darkening in sunlight. Addition of halogenated additives, e.g. chlorinated
paraffin wax, also affects long term mechanical properties. The incorpora-
tion of aluminium trihydrate has a very beneficial effect on fire resistance.
Protection of glass-polyester laminates can also be achieved by
intumescent coatings.
Smoke generation is very undesirable, increasing the number offatalities
in fires. Polyester resins produce smoke from the styrene and phthalic
components,. both of which are virtually essential constituents. Efforts
have been made to reduce smoke emission by:
(I) substituting non-aromatic monomers for styrene, at considerable
cost,
(2) inducing char formation or intumescence.
The fire retardant properties of urea-formaldehyde (U-F) and some
other thermosetting resins offer great advantages in the market for
insulating foam, despite the sometimes inferior thermal insulation capa-
bility in comparison with expanded polystyrene and polyurethane foam.
1.15. HANDLING AND PROCESSING OF RESINS
It has already been noted that an adequate account of this subject is

beyond the scope of this book. Nevertheless, the chemistry of thermosetting
resins is inseparable from their processability.
Considerable effort has been directed to controlling the hardening
reaction, so that long shelf lives are achieved along with rapid gelation
and cure at the right time. One-component epoxy resins are now available
which rely on heat-curing systems and give rapid cure at 150°C, but
still allow long storage life without refrigeration.
The search for more convenient methods of curing unsaturated polyester

resins has led to further investigation of ultra-violet initiation 15 and visible
light curing. 16 Savings in materials, cleaner operation, and reduced styrene
evaporation have been claimed for visible light cured polyesters. On
the other hand special light sources are required, and the method is
restricted to open mould fabrication methods with thin sections.
Bulk handling of polyester resins is economic for high throughputs.
One method is to pump resin directly from a thermostatted, walled-in
storage tank, while another requires the use of smaller, mobile containers
filled from road tankers.
Injection moulding of dough moulding compounds and of phenolic
resins has become widespread. Polyester sheet moulding compounds have
been developed to achieve better product consistency and improved surface
quality. There is a better understanding of how the resin structure affects
the thickening process and the final SMC properties. I 7
Recycling of scrap thermoset material has been advocated and prac-
tised. 18 This is because the increasing cost of raw materials has made
the policy seem attractive economically. Up to 15 %of reground phenolic,
amino or alkyd is said to leave the mechanical properties of compression-,
transfer-, and injection-moulded parts unchanged. Above this point,
impact and flexural strength are severely affected. Addition of scrap
acts like inert filler in reducing mould shrinkage.
However, the use of reground thermosetting material has been severely
criticised because it could lead to wide variations in properties and occa-
sional product failure. Contamination with metal dust from the grinding
process could seriously affect electrical properties. Regrinding also
increases the production of undesirable dust.
1.16. NEW METHODS OF RESIN QUALITY CONTROL
May et a/. 19 have pointed out a growing realisation of the need for
quality control in the manufacture of high performance resins. The new
instrumental techniques for chemical analysis and for physical examination
of materials, many of which were developed in the 1950s and 1960s,
have now begun to demonstrate their usefulness in the study of thermo-
setting resins.
Variability between batches of epoxy resin affects prepreg processing
by causing changes in tack, flow and gel times. The effect of batch
variability on composite durability has yet to be determined.
24 G. PRITCHARD
Infra-red spectroscopy can determine whether a reactive diluent contain-

ing a major functional group (such as carbonyl) has been accidentally
omitted from a batch. It can also determine quantitatively the percentage
of some curing agents such as dicyandiamide or diaminodiphenylsulphone.
Gel permeation chromatography20 determines the molecular weight
distribution of a resin. Figure 1.5 shows a G PC trace obtained from
an unsaturated polyester resin in styrene solution. The GPC indicates
styrene
100
molecular size nm
FIG. 1.5. Gel permeation chromatograph of an unsaturated polyester resin III

styrene solution.
whether too much or too little epoxy diluent has been used, detects
the presence of two or more resin types in a mixture and measures
free phenol in phenolic resins.
Thin layer chromatography facilitates separation of the resin com-
ponents on a silica gel plate. Differential scanning calorimetry or measure-
ment of electrical loss tangent, can indicate the degree of ageing of a
resin after prolonged storage.
More recent innovations which may prove useful to the resin technologist
include laser-Raman spectroscopy,21 13C NMR neutron scattering, and
ESCA 13C NMR has already been used to study the curing mechanisms
of acetylene-terminated imide monomers when reacted to form polyimide
resins. Tomita and Haton0 22 used the same technique to characterise
random urea-formaldehyde structures. Proton NMR has been used for
the analysis of phenol-formaldehyde 23 and polyester 24 resins. (See
Fig. 1.6.)
Pyrolysis gas chromatography has been a standard method for studying
chemIcal shIll ppm

-H
FIG. 1.6. NMR spectrum of an unsaturated polyester resin.
crosslinked rubbers for many years. It has some potential for studying
crosslinked polymers in general. 25
There are still many problems. It is not easy to determine quantitatively
the hydroxyl content of thermosetting resins. Hase and Hase 26 proposed
a silylation method based on NMR spectroscopy designed to give a
nine-fold enhancement of the proton signal. This technique was applied
to polyester resins. Fritz et al. 27 substituted the hydroxyl group of a
polyethylene glycol by a chromophoric siloxy group, purified the silylated
polymer, and determined the chromophoric concentration photometri-
cally. This method was claimed to be a thousand times more sensitive
than the acetylation procedure.
The quality control of finished products can be still more difficult.
Attention has been directed to developing non-destructive methods for
the determination of void content and distribution 28 .29 of resin, fibre
rati0 30 and of the magnitude and location of built-in stresses. Barrett
and Predecki 31 have described how stresses in epoxy resins can be measured
by high angle X-ray diffraction. Voids and other defects can be detected
and assessed by ultrasonic scanning 29 and by radiotracer methods,32
but there is a need for further development of practical non-destructive
techniques.
1.17. THE FUTURE PROSPECTS FOR THERMOSETIING

RESINS
It has been estimated that about 24 % of the world's plastics moulding

materials are of the thermosetting variety. The share has declined in
26 G. PRITCHARD
the last thirty years, because thermoplastics based on cheap oil and
easy processing had many economic advantages. In the future, the ability
of thermosetting formulators to increase the inorganic proportion of
their moulding compositions may be very important. Also the development
of injection moulded thermosets has reduced their processing
disadvantage.
About 90 % of the major organic chemicals for the chemical industry
are derived from just six feedstocks:
(I) synthesis gas from coal or natural gas
(2) ethylene
(3) propylene
(4) butadiene
(5) benzene
(6) para-xylene
The cost of plastics, and of metals too, depends on oil prices, but
in different ways. Plastics are mostly made from oil with alternative
sources being coal, fermentation for ethanol, and biomass for glycerol,
ethanol and furfural.
The total energy requirement for phenolic resin production, including
both feedstock and conversion, is about 60 kJ/dm 3 . This is less than
for engineering thermoplastics (about 100 to 300kJ/dm 3 ) and also less
than for metals, (up to 400 kJ/dm 3 ). This is partly responsible for the
relatively low cost of phenolic and melamine resin materials. But the
cost of fabricating finished articles can sometimes be higher for thermosets
than for thermoplastics.
The rise in material prices since 1973 has been dramatic for both
thermosetting and thermoplastic materials. (See Table 1.3.)
Attempts will be made to increase the volume fraction offibres, inorganic
fillers, hollow micro spheres and foam material to cushion the effect of
oil price rises. Styrene monomer (used in polyester resins as well as
polystyrene) increased in price from 18 US cents/lb to 31 US cents/lb
in the twelve months up to mid 1979. This compares with 9 cents/lb
in 1972.
Reinforced thermosets offer certain energy advantages. They save energy
in various ways: in their transportation and in their marine, chemical
process, building and aerospace applications. Low weight, low main-
tenance costs and good thermal insulation characteristics will be useful
characteristics in the future. Taking these factors together, there are
good prospects for structural foam composites.
TABLE 1.3
UK PRICE RISE~ fOR MOULDING MhHRlhL5
(Units: pence per kg, tonne lots)
September 1973 December 1979 %

1ncrease
Phenol-formaldehyde 20·2 62·5 209

Urea-formaldehyde 24·0 56-4 135
Melamine-formaldehyde 38·5 85·0 121
Polyester DMC 26·7 75·0 181
Polypropylene 23·6 55·0 133
Nylon 6 61·5 171·0 178
Acetal 69·4 144·0 107
General purpose polystyrene 28·3 63-0 123
Source: Plastics and Rubber Weekly.
The materials producers and the moulders will need to maintain and
improve their standards of pollution control and health and safety. Dust,
fumes and skin-sensitising substances are common in the reinforced plastics
industry, and Sweden has already legislated to improve practices. 33 ,34
Increasing research into the toxicological properties of chemicals in
general, and the inevitable dissemination of over-simplified accounts of
the results, are facts of life.
Fundamental research into the science of crosslinked polymers has
not been so extensive as that directed towards more tractable linear
polymers. It is nevertheless increasing and may be expected to result
in improved commercial products. Some examples of the research and
development recently carried out in the thermosetting field are given
in the remaining chapters of this book.
REFERENCES
I. MORGAN, R. J. and O'NEAL, J. E. J. Macromol. SCI., Phys., 1978, B15, 139.

2. NELSON, B. E. and TURNER, D. T. J. Polym. Sci., Polym. Phys. Ed., 1972, 10,
2461.
3. FUNKE, W. J. Polym. Sci., 1967, Part C3, 1497.
4. ALEMAN, J. V. Polym. Eng. Sci., 1978, 18, 1160.
5. FOWLE, D. J. Chem. 1nd., 1978, 361.
6. DESAI, R. R. Composites, 1974,5, 16.
7. WILSON, E. L. in Flame Retardancy oj Polymeric Materials, Vol. 3, ed. Kuryla,
W. C. and Papa, A. J., 1975, Marcel Dekker, New York, p. 254.
28 G. PRITCHARD
8. BRUINS, P. F., ed., Unsaturated Polyester Technology, 1976, Gordon and

Breach, New York.
9. POTTER, W. G. Epoxide Resins, 1970, Iliffe, London.
10. IDRIS JONES, J. Chem. Brit., 1970,6,251.
II. JUDD, N. C. W. and WRIGHT, W. W. Reinforced Plastics, 1978,22,39.
12. RINDE, J. A., MONES, E. T., MOORE, R. L. and NEWEY, H. A. 34th SP]
Remforced Plasflcs/Composites Conj., New Orleans, La., USA, 1979, Paper
17-A.
13. HANCOX, N. L. and WELLS, H. 32ndSP] Reinforced Plastics/Composites Conj.,
Washington D.C., USA, 1977, Paper 9-C.
14. ANON. Plastics and Rubber Weekly, 1979, Sept. 21, pp.24-6.
15. SCHICK, 1. P. Plastica, 1978, 31, 4.
16. LONGENECKER, D. M. and GRETH, G. G. Plastics Eng., 1977,33, 52.
17. BURNS, R., LYNSKEY, B. M., GANDHI, K. S. and HANKIN, A. G. Plastics and
Polymers, 1975,43,228.
18. BAUER, S. H. Plastics Eng., 1977,33,44.
19. MAY, C. A., HADDAD, D. K. and BROWNING, C. E. 33rd SP] Reinforced
Plastics/Composites Conj., Washington, D.C., USA, 1978, Paper 15-D.
20. LEE, W. Y. J. Appl. Polym. Sci., 1978,22,3343.
21. KOENIG, J. L. and SHIH, P. T. K. J. Polym. Sci., 1972, 10, Part A2, 721.
22. TOMITA, B. and HATONO, S. J. Polym. Sci., Polym. Chem. Ed., 1978, 16,2509.
23. POSPISIL, L. and NAVRATIL, M. Chem. Prum., 1979,29,34.
24. BIRLEY, A. W., DAWKINS, J. V. and KYRIACOS, D. Polymer, 1978, 19, 1433.
25. HAEUSLER, K. G., SCHROEDER, E., GROSSKREUZ, G. and HUBE, H. PlastIc u.
Kaut, 1978, 5, 691.
26. HASE, A. and HASE, T. Analyst, 1972,97,998.
27. FRITZ, D. F., SAHIL,A., KELLER, H. P. and KOVATS,E. S.Analyst, Chem., 1979,
51,7.
28. STONE,D. E. W.andCLARKE, B. Technical Report No. 74162(Dec.1974)Royal
Aircraft Establishment, Farnborough, England.
29. DEAN, G. Characterization of fibre composites using ultrasonics. Proceedings
of Conference 'Composites-Standards, Testing and DeSIgn" 1974, I PC Science
and Technology Press, Guildford, England, p. 126.
30. TORP, S., FORLI, O. and MALMO, J. 32nd SP] Reinforced Plastics/Composites
Conj., Washington, D.C., USA, 1977, Paper 9-A.
31. BARRETT, C. S. and PREDECKI, P. Polym. Eng. Sci., 1976, 16,602.
32. JOINER, J. C. The determination of voids in carbon fibre composites, Report
AQD/NM 00296, (July 1973), Ministry of Defence, Aircraft Quality
Directorate, Woolwich, England.
33. ANON. Reinforced Plastics, 1975, 19, 148.
34. BRIGHTON, C. A., PRITCHARD, G. and SKINNER, G. A. Styrene Polymers:
Technology and Environmental Aspects, Chapters 5 and 7, 1979, Applied
Science Publishers Ltd, London.
Chapter 2
VINYL ESTER RESINS
THOMAS F. ANDERSON and VIRGINIA B. MESSICK

Dow Chemical USA,t Texas, USA
SUMMARY
Vinyl ester resins are relatively recent additions to the thermosettingfamity.

They have some features in common with unsaturated polyesters, and
most have structural features similar to those of epoxides. They are notable
for their high-temperature properties, their chemical resistance, their high
elongation and their convenient processing characteristics.
This chapter describes the history, the synthesis and typical structures
oj vinyl esters, and relates these structures to the properties of cast resins
and laminates.
Applications in corrosIOn-resistant equipment, land transport, electrical
insulation, marine, and several other fields are described. The toxicological
properties oj vinyl ester resins are summarised.
2.1. INTRODUCTION
Vinyl ester resins are resins produced by the addition of an ethylenically

unsaturated monocarboxylic acid to a backbone (usually epoxy) producing
terminal un saturation and which can be cured with vinyl monomers similar
to those used for crosslinking polyesters. Vinyl ester resins combine the
excellent thermal and mechanical properties of epoxy resins with the
ease of processing and rapid curing of polyester resins. They have opened
up broad new applications for thermoset resins since they (1) can be
t Dow CRI No. B-600-024-80.

29
30 THOMAS F. ANDERSON AND VIRGINIA B. MESSICK
cured rapidly with relatively nontoxic catalysts, as can polyester resins,

(2) have excellent wetting and bond to glass fibres, as epoxy resins do,
(3) retain high elongation at moderate and high heat distortion tempera-
tures due to controlled crosslinking structure and (4) have excellent high-
temperature and heat ageing properties when novolac epoxy resins are
a part of the basic structure.
2.2. EARLY HISTORY
The chemistry of the vinyl ester resins was developed in the late 1950s
and early 1960s by several people, each having a different base. Bowen,
working for the US Commerce Department, was looking for a way to
get greater toughness and improved bonding to teeth in the acrylic
polymers used in dental applications. He reacted glycidyl acrylate and
glycidyl methacrylate with bisphenol and used the resulting difunctional
resin he obtained to give crosslinked polymers. 1 •2 These resins were so
reactive that they did not have a useful working life and it was several
years before stable resins for the dental market were commercialised.
Fekete et al. developed resins primarily for use in electrical insulation
and corrosion resistant equipment. 3.4 These resins were homopolymerised
through their acrylic end groups and copolymerised with monomers.
One development with these resins was the appearance of very fast curing
moulding compounds made with a blend of the vinyl ester resin with
a polyester resin. 5 These compounds were unique in that they cured
faster than compounds made with either of the other resins by itself.
Bearden worked on polyalkyl acrylate polymers and reacted several
different vinyl unsaturated compounds with diepoxides to give di- and
tetra-functional unsaturated resins. 6 These resins were developed to be
cut in styrene monomer and designed for use in matched metal die mould-
ing. Bearden, in early work, discovered methods of stabilising these tremen-
dously reactive resins so that they could be shipped to moulders without
polymerising in the drum, and yet would cure rapidly and completely
when properly catalysed. These stabilised systems were the key that allowed
commercialisation of these resins in manufacturing structural and electrical
matched metal die moulded laminate structures in 1964. Electrical
laminates made with these resins were found to have exceptional retention
of physical and electrical properties when heat aged continuously at tem-
peratures of 180 ° to 200°C, and in special cases at 220°C. Yet they
could be processed easily and rapidly like polyester resins. As a result,
VINYL ESTER RESINS 31
insulation based on these resins made possible the development of more

compact (higher operating temperature) motors, generators, and dry type
transformers.
Bearden and coworkers, Jernigan, Najvar and Hargis, recognised that
the heat resistance, high reactivity activity and toughness of these resins
would offer real advantages in the corrosion resistance field. They modified
one of the resins by substituting methacrylic acid for acrylic acid on
the ends of the molecule. The resultant shielding of the ester linkage
gave a resin with excellent resistance to hydrolysis and broad general
corrosion resistance. This group recognised the need for fire retardant
formulations and thickenable formulations and developed such resins. 7 - 9
The fast glassfibre wet-out and rapid, complete curing of these resins
was used to develop highly automated filament winding technology.!O
2.3. SYNTHESIS
Vinyl ester resins are based on the reaction product of an epoxy resin
and an ethylenically unsaturated carboxylic acid which results in terminal
unsaturation. Various epoxy resins are used, including the diglycidyl
ether of bisphenol A, or higher homologues thereof, the diglycidyl ether
of tetra bromo bisphenol A, epoxylated phenol-formaldehyde novolac and
polypropylene oxide diepoxide. The most commonly used acids are acrylic
and methacrylic acids, although use of other unsaturated acids such
as cinnamic and crotonic acids has been reported in the literature.!!
The acid-epoxide reaction is straightforward and is catalysed by tertiary
amines, phosphines, alkalis or onium salts.12 The acid-epoxide reaction
results in pendant hydroxyl groups which provide adhesion and/or reactive
sites for further modification with compounds such as anhydrides or
isocyanates.13 Vinyl ester resins are diluted with a reactive monomer
such as styrene, vinyl toluene or dicyclopentadiene acrylate.
2.4. STRUCTURE-IMPARTED CHARACTERISTI CS
RO
I
C==c-C-O
I f
The generalised structure of a vinyl ester resin is shown below.
~-c-o-@-~-@-o
OH
C}~-c-o-c-c==c
C
I OR
n OH
I I
(I)
R = H or CH 3
2.4.1. Toughness
Toughness is imparted to vinyl ester resins by the epoxy resin backbone 14
(the portion enclosed in brackets in Structure (I)). The molecular weight
of the epoxy resin portion can be controlled by reacting the diglycidyl
ether of bisphenol A with specific amounts of bisphenol A. Physical
properties such as tensile strength, tensile elongation and heat distortion
temperature are relatable to the molecular weight of the epoxy resin
portion of the vinyl ester resin. Thus, vinyl ester resins can be 'tailor
made' to meet the requirements of specific applications.
2.4.2. Corrosion Resistance

Corrosion resistance is obtained first from the phenyl ether linkage of
the epoxy resin backbone. This linkage is quite stable in many chemical
environments. The ester linkage appears to be shielded somewhat by
the pendant methyl group. Furthermore, the ester linkages are formed
only at the ends of the molecule, so the number of ester groups present
are minimal. This is in contrast to polyesters where ester linkages form
part of the repeating unit.
2.5. STRUCTURES
The following structures illustrate the diversification of vinyl ester resins

that has taken place over the last 10 years to meet specific end-use
needs.
2.5.1. Basic Vinyl Ester Resin (Bisphenol A-Epoxy)

Example: Derakane® 411 vinyl ester resin.15
CH2~J-O{CH2-CH-CH2-0-1Q\-_1H~tH2-CH-CH2-0J-C~1
I I -~~ I I
CH 3 OH CH 3 n OH CH 3
123 4 5 (I
I. Terminal vinyl un saturation-located at the ends of the molecule

where they are very reactive, causing the vinyl ester resin to cure
rapidly to give fast green strength, and enable the vinyl ester resin
to homopolymerise or copolymerise to give polymers with high
corrosion resistance.
® Trademark of The Dow Chemical Company.
2. Methyl grou~shields the ester linkage increasing resistance to

hydrolysis.
3. Ester groups-vinyl ester resins have 35 to 50 %fewer ester groups
per unit of molecular weight than corrosion resistant polyesters.
This contributes to their resistance to hydrolysis by alkaline
solutions.
4. Secondary hydroxyl-the interaction of the chain secondary
hydroxyl groups with the hydroxyl groups on the surface of glass
fibre gives improved wetting and bonding. This is one of the factors
responsible for the higher strengths obtained with vinyl ester resin
laminates.
5. Epoxy resin backbone-imparts toughness and allows controlled
molecular weight for low viscosity. The epoxy ether bond provides
superior acid resistance.
2.5.2. SMC Resins

Example: Derakane 786 vinyl ester resin.
A B-stageable vinyl ester resin was developed to overcome problem areas

in preimpregnated moulding mats or sheet moulding compound (SMC). 16
For a resin to be appropriate for use in SMC technology, it must be
chemically thickenable (using divalent metallic oxides and/or hydroxides)
while retaining the molecular un saturation which is polymerised during
the moulding operation. This is accomplished by introducing acid function-
ality on the vinyl ester resin molecule. 9 This resin has a more rapid
B-staglllg rate, good glass wettability, and better rheology (allowing lower
moulding pressures) than polyesters.
The SMC resins have found application in structural parts, particularly
in the automotive industry, where weight reduction is an important factor
for improved fuel consumption.17.18.19 Addition of a thermoplastic
materiaFO has greatly improved the surface characteristics and shrinkage.
2.5.3. Epoxy Novolac Vinyl Ester Resins for High-Temperature Applications

Example: Derakane 470 vinyl ester resin.
CH 2 CH 2 CH 2
II II II
C-CH 3 C-CH 3 C-CH 3
I I I
C=O C=O C=O
I I I
0 0 0
I I I
CH 2 CH 2 CH 2 (IV)
I I I
CH-OH CH-OH CH-OH
I I I
CH 2 CH 2 CH 2
~~H'+~~H,t~
In 1972 a new high-temperature organic resistant vinyl ester resin was
made commercially available. 21 The heat resistance and high thermal
stability were achieved by incorporating an epoxy resin based on phenol-
formaldehyde novo lac into the vinyl ester resin backbone, increasing
the crosslink density when the resin is cured. This resin, which has a
heat distortion temperature of 270--300 of (132-149°C) significantly
extended the useful operating temperature of vinyl ester resins while
retaining excellent corrosion resistance, especially in environments con-
taining chlorine or organic solvents.
2.5.4. Flame Retardant Resins

Example: Derakane 510-A-40 vinyl ester resin.
Brominated vinyl ester resins were developed to meet the industry's need
for reduced flammability characteristics while retaining corrosion resist-
ance. This flame retardant vinyl ester resin is used extensively in RP
TABLE 2.1
THE PERFORMANCE OF A FLAME RETARDANT RESIN IN FIRE RESISTANCE TESTS"
Test type Test designation Derakane 510A resin
60 s Burning test b ASTM D757 0·4in/min (l0'2mm/min)

Intermittent exposure teste HLT 15 100
Tunnel test flame spread
Rating teste ASTM E84
unfilled 30
with 5 % Sb 2 0 3 10
Limiting oxygen index teste ASTM D2863
unfilled 29·7 % oxygen
with 5 % Sb 2 0 3 40·8 % oxygen
" The results shown in this table were obtained from controlled and/or small scale
bench tests. They are not necessarily predictive of behaviour in a real fire situation.
Vinyl ester resins are organic materials, and the resins and products made therefrom
will burn under the right conditions of heat and oxygen supply.
b 60% glass, press moulded at 60 psi (0-41 MNm- 2 ); benzoyl peroxide cure.
e 25 % glass, 0·125 inch (3'2 mm), hand lay-up with Methylethylketone peroxide
(MEKP) and Cobalt Naphthenate cure. Glass sequence: 10 mil. C-veil (V), three
1·05 oz ft - 2 (457 gm - 2) mats (M), 10 mil. veil.
hoods, ductwork and stack applications where decreased flammability

is desirable. The performance of this flame retardant resin in fire resistance
tests is shown in Table 2.1.
2.5.5. Radiation Curable Resins

Example: Dow XD-9002 experimental vinyl ester resin.
(VI)
Radiation curable resins were developed to meet needs in several areas

of application including coatings and printing inks. The ability to cure
when exposed to ultraviolet or electron beam radiation is obtained by
replacing the terminal methacrylate groups of vinyl ester resins with
acrylate end groups. Photo initiators such as benzophenone or benzoin
ethers are used in ultraviolet curing to absorb the UV energy and transfer
it to the resin system so as to effect vinyl polymerisation. The resins
have a low viscosity in the undiluted state, or reactive diluents such

as 2-ethyl hexyl acrylate or 2-hydroxypropyl acrylate can be used to
further reduce viscosity. Cure times are quite rapid, and are measured
in seconds rather than minutes, as for conventional coatings.
2.5.6. Bisphenol A Fumaric Acid Condensation Polyester

Example: Atlac® 382.
oII 0II t O O II II
HO-C---CH-CH---C-O H-CH 2 -R---CH 2 ---CH-O---C---CH-CH---C-
I I
CH 3 CH 3
CH-CH 2-R---CH 2---CH-OH

I I
R = Blsphenol A CH 3 CH 3
(VII)
The structure of a bisphenol A fumarate polyester has been included

for comparison. The physical properties are shown in Table 2.2.
t
2.5.7. Urethane-Based Vinyl Ester Resin
-t
Example: Atlac 580.
I
R2
H 2 C==C-U
°
I
O-R 1-O-C-CH==CH-C-O-R 1-O
I ° I
R2
U-C==CH 2
n
(VIII)
Rl = Bisphenol A as shown in ICI literature

R2 = Alkyl group or hydrogen
U = Urethane Connecting Group
A urethane-based vinyl ester resin was introduced in 1975. 13 It is reported

to combine the best properties of polymers containing both internal
and terminal unsaturation. There is some controversy as to whether or
not this resin can be properly termed a 'vinyl ester resin' due to a large
amount of internal un saturation. The urethane modification is said to
improve glass wettability and adhesion.
® leI United States Inc.

TABLE 2.2
TYPICAL PHYSICAL PROPERTIES OF VINYL ESTER RESINS
/I III IV V VI VII VIII >IX

B,s A epoxy vrnyl S..\1C vmy! ester Epoxy novo!ac hased Brol1wwred vinyl Radlallon cured BlSphenol Urethane-based Ruibber-
ester resin resin + Vinyl ester resin ester resin vmyl ester resin A-fumarate Vinyl ester resm! modljl'ed VIn;!
thermoplastIc polyester est~,.- reSin
addlllVe
-<
LiqUId Properlles Z
VIscosity kmematlc. 7rF (25 0C). cs 450 2250 200 300 4700' 10001 -<:
SpecIfic gravity 104 102 107 I 22 I 15 I 07 t'"'
77 'F (25 'C) Gel time (mm)' 28 12 24 I'T'I
on
180'F (82'C) Gel time (mm)' 12 300 13 10 -l
% Styrene 45 44 36 40 neat 50 450 I'T'I
Clear Castmg ProperllesJ ::0
TensIle strength, pSI (MN rn - 2) 12000 (82 7) 6000' (414) 11000 (758) 10600(761) 10000 (69 0) 13100 (90 3) 100001 (69 0) ::0
I'T'I
% Elongation 5 6 3 5 I 5 42 IIlI on
Flexural strength, PSI (MN m- ) 2 18000 (1240) 11000 (75 8) 20000 (137 9) 18000 (1240) 13500 (93 I) 22600 (ISS 8) 17000 (1172)
Flexural modulus PSI x 10 5 (GN m- 2 ) 45(3 I) 34 (2 3) 55 (3 8) 52 (3 6) 49 (34) 41 4 (3 0)
Z
on
HOT. 'F ('C) 215 (102) 220 (104) 290-300 (143 9) 230 (110) 270 (132) 221 (lOS) 17D1 (77)
Barcol hardness 35 40 40 37 37 301
<lI 5% MEKP solutIOn (60% MEKP In dImethyl phthalate) and 0 5/~cobalt naphthenate solution (6% Co III mmeral spmts)
b 1 % Benzoyl peroxide
, 60°C
d I 0% MEKP solutIOn and 0 3% cobalt naphthenate solution, cured 16h at 77'F (25°C) and 2h at 311°F (I55'C)
, I % Benzoyl perOXide and 0 I % N.N-<hmethyl anilme, cured 16 h at 77 'F (25 'C) and 2 h at 311 'F (ISS 'C)
f Data taken from 'Introducing Atlac<!) 580 Vmyl Ester Resm', lei United States Inc, 1975
<.;J
--l
2.5.8. Rubber-Modified Vinyl Ester Resin

Example: Dow XD-SOS4 experimental vinyl ester resin. To meet the
need for improved toughness for applications where severe mechanical
abuses are encountered, a rubber-modified vinyl ester resin was developed.
Rubber is incorporated into the vinyl ester molecule. In addition to
TABLE 2.3
TYPICAL REVERSE IMPACT DATA ON 156 INCH (7·9 mm) HAND
LAY-UP LAMINATES
Resin First crack
Bisphenol A-fumarate polyester 16 in Ib (I. 77 J)

Epoxy novo lac based vinyl ester resin 28inlb (3'04J)
Basic vinyl ester resin S7inlb (6'38J)
Rubber-modified vinyl ester resin 207 in Ib (22'96 J)
improved toughness, the rubber-modified resin has improved adhesion,

lower exotherm temperatures and less shrinkage than the basic vinyl
ester resin.
Tables 2.3 and 2.422 illustrate the improved impact toughness and
adhesion of the rubber-modified vinyl ester resin compared to the basic
vinyl ester resin.
TABLE 2.4
TYPICAL ADHESIVE STRENGTH OF RUBBER-MODIFIED VINYL ESTER RESIN COMPARED TO
THE BASIC VINYL ESTER RESIN AND BISPHENOL A FUMARATE POLYESTER
Test Result
t-Peel (ASTM D-1876-6IT) on 2024 T3 Aluminium

Rubber-modified vinyl ester resin 4· SIb/linear in (779 N m -I)
BaSIC vinyl ester resin 2·9Ib/linear in (S02Nm- l )
Bisphenol A-fumarate polyester 0'9Ib/linear in (lS6Nm- l )
Lap Shear (ASTM D-I002) on 2024 T3 Aluminium

Rubber-modified vinyl ester resin > I 000 psi (> 6·9 MNm -2)
Basic vinyl ester resin 500 psi (3·5MNm- 2 )
Bisphenol A-fumarate polyester SOOpsi (3·S MNm- 2 )
2.6. CURING
The choice of peroxide catalysts is determined by the particular resin

in question and the temperature at which it is to be cured. Generally,
methyl ethyl ketone peroxide (MEKP) is used for room temperature
curing, and benzoyl peroxide (BPO) or I-butyl per benzoate is used for
elevated temperature curing. BPO with N,N-dimethyl aniline as an
accele'rator may also be used for curing vinyl ester resins at room tempera-
ture. Cobalt causes ketone peroxides to dissociate into free radicals.
Thus the room temperature cure can be effected without the application
of external heat. 23 N ,N-dimethyl aniline or other aromatic tertiary amines
may also be used to further accelerate the MEKP curing system. MEKP
with a high dimer content is more reactive than low dimer MEKP in
curing vinyl ester resins, while the opposite is true with polyester resins. 24,25
Criteria for determining the catalyst system and judging the degree
of cure include exotherm temperature, residual monomer, physical proper-
ties, working time and development of hardness or tack-free state. The
exotherm temperature should be high enough to cure the resin but not
so high as to cause cracking of the resin, This is particularly important
in cast parts of large mass and thickness where the heat dissipation
will be slower. Residual monomer will deleteriously affect physical pro-
perties and corrosion resistance. 26
Oxygen in the air will inhibit the complete cure of an exposed vinyl
ester resin surface. This inhibited surface may vary in depth, and may
result in reduced weatherability, poor chemical resistance and/or prema-
ture failure. This problem is overcome by preventing or reducing contact
of the curing surface with air.
2.7. INHIBITORS
Phenolic inhibitors such as hydroquinone or the monomethyl ether of

hydroquinone are used during the synthesis of vinyl ester resins to prevent
polymerisation during processing. An inhibitor such as one of the
phenolics may also be added at the completion of synthesis to extend
the shelf-life. Sulphur-containing compounds have also been claimed to
be effective in extending shelf-life.27 The effectiveness of an inhibitor
can be determined by following its consumption using gas or liquid
chromatography or by uncatalysed shelf-life studies under controlled
conditions.
The ideal inhibitor would give infinite shelf-life while not interfering
with, or retarding, the peroxide-catalysed cure of the vinyl ester resin.
Frequently the catalyst system and the curing temperature for a particular
application will determine the choice of inhibitors as with polyester resins.
Periodic aeration of vinyl ester resins has also been found to be helpful
in extending the shelf-life.
2.8. EFFECTS OF CAST RESIN HIGH TENSILE ELONGATION

ON PERFORMANCE OF LAMINATED STRUCTURES
The effect of the very high tensile elongation of basic vinyl ester resin
castings shows up in the performance of laminated structures in areas
such as, (I) the strain level at first resin cracks, (2) impact load at first
reverse side cracking and (3) hydraulic pressure in pipe at first weeping.
(I) In Table 2.5 it is shown that there is direct linear relationship
between cast resin tensile elongation and the %tensile strain at first
visible cracks in SPI laminates made with the same resins. The
laminates were made from: veil (V)-Owens-Corning Fibreglass
M5l4 Treatment 236; mat (M)-Owens-Corning Fibreglass
M- 711 I-I' 5 oz chopped strand mat (457 g m - 2). Construction was
V-M-M-V. Mould surface was Mylar® film on both sides.
The first fine cracks were visible in the chopped strand layers
only at the values shown in the table. These fine cracks were
TABLE 2.5
PERCENT STRAIN AT FIRST VISIBLE RESIN CRACKS IN CHOPPED
STRAND MAT LAMINATE V. TENSILE ELONGATION OF RESIN
CASTINGS
Resin type % Tenstle % Tensile stram

elongation oj in laminate to first
cast resin visible resin cracks
Basic vinyl ester 5-7 1-1'2

Epoxy novo lac vinyl ester 3-4 0·6--0·7
Polyester (rigid) 1'5-2 0,3-0,5
Polyester (very ngid) <I 0'2-0'3
® Trademark of the E. I. Du Pont Company.

TABLE 2.6
REVERSE IMPACT ON t6 (7'9mm)
INCH THICK HAND LAY-UP
LAMINATES
Resin t}pe % Tenszle Impact at

elongatIOn oj first crack
cast resin
Polyester (rigid) 1·5-2 16 in Ib (1'77 J)

Vinyl ester (epoxy novo lac) 3--4 28 in Ib (3'04J)
Vinyl ester (basic) 5-7 57inlb (6'38J)
visible when the samples were illuminated with a parallel light beam
at a 45 0 angle to the laminate surface. They disappeared when the
stress was removed.
(2) Table 2.6 shows the impact load to first reverse surface cracking
on 156 in (7·9 mm) thick laminate v. resin tensile elongation. In
this case we see that there is a direct relationship between resin
tensile elongation and impact resistance. In the resins tested here
the relationship is also linear. However, the linear relationship dis-
appears with high elongation rubber modified resins, which show
stress marks, but do not reverse crack at impact loads of almost
twice that which would be obtained by linear extrapolation of the
points shown in Table 2.6. This information is previously given in
Table 2.3 in the discussion on rubber modified resin.
(3) Filament wound pipe having a 2 in (50 mm) inside pipe diameter
was made using a 0·0 lOin (0·254 mm) C-veil liner and nominal
0·080 in (2'0 mm)wall from two of the polyester resins and the basic
vinyl resin used in the previous tests. The results are shown in
Table 2.7. Note that the pipes made from the 'very rigid' and 'rigid'
polyester resins started to weep at less than half and slightly more
than half of the weep pressure of the high elongation vinyl ester
resin although their 'Split ring Tensile' strength is equal or almost
equal to that of the vinyl ester resin. The burst pressure of the poly-
ester pipes could not be determined because the weep rate became
so high that the pump capacity was reached not very far above the
first weep pressure. The epoxy novo lac type of vinyl ester resin was
not run in this test. In tests on 4 in (102 mm) diameter pipe made
from this 3-4 % tensile elongation resin it started to weep at
approximately 75 % of the weep pressure of pipe made from the
basic vinyl ester resin.
TABLE 2.7
WEEP AND BURST PRESSURE AND HOOP STRENGTH OF 2 INCH (SOmm) FILAMENT
WOUND PIPE V. TENSILE ELONGATION OF CAST RESIN 0·080 INCH (2'Omm) WALL
THICKNESS, 0·010 INCH (0·2S4mm) C-VEIL LINER
Resin type % Tensile Pressure Burst Hoop

elongatIOn of to weep pressure tensile"
cast resin psi (MNm- 2) x 10 3 psi X 10 3 psi
(GNm- 2) (MNm-2)
_b
Polyester (very rigid) <I 960 (6'6) 30 (207)
b
Polyester (rigid) 1'5-2 1280 (8'8) - 36 (248)
Vinyl ester (basic) 5-7 2180 (15'0) 2310 (15'9) 36 (248)
" Split ring tensile.

b Rate of weeping exceeded pump capacity so that pipes could not burst.
The examples given above show the effects of cast resin elongation on
performance oflaminate in three different types of tests. From the practical
side, this shows up dramatically in performance in applications that range
from chemical equipment to ballistic armour. This is covered in Section 2.ll
(applications).
2.9. MONOMERS
While styrene is the most common monomer used to dilute vinyl ester
resins, other monomers which have been used include divinyl benzene,28
vinyl toluene, chlorostyrene,29 a-methyl styrene, t-butyl styrene and
dicyclopentadiene acrylate.
300
HDT OF
280
260
~
240 0
20 30 40 50
% Styrene
FIG. 2.1. The effect of styrene concentration on the heat distortion temperature
(HDT) of Derakane 470.
TABLE 2.8
MECHANICAL PROPERTIES OF DERAKANE 411 RESIN AT 45 % MONO~:lERa
Monomer Tensile % Flexural Flexural HTD Barcol

strength Elongation strength modulus CF (a e» hardness -<
(psi (MNm- 2 )) (psi (MNm- 2 )) (psi x ]05 (GNm- 2 ))
Z
-<
r
tr1
Chlorostyrene 12600 (87) 3·80 24100 (166) 5·66 (3·9) 231 (111) 45 en
-l
tr1
Dicyclopentadiene acrylate ::tI
(70 % monomer) 7800 (54) 1·36 20400 (141) 5·39 (3·7) 191 (88) 46 ::tI
tr1
Divinyl benzene 3600 (25) 0·94 12200 (84) 5·76 (4·0) 226 (107) 53 en
Styrene 12000 (83) 6·29 22400 (154) 5·68 (3·9) 222 (105) 41 Z
en
Vinyl toluene 11500 (79) 4·15 21 300 (147) 5-44 (3·7) 205 (96) 44
aCured with O· 3 % Cobalt naphthenate solution (6 % Co in mineral spirits) and 1·0 % MEKP solution (60 % MEKP in dimethyl
phthalate) 16h at room temperature and 2h at 311 OF (155 a C).
.j:>.
~
Styrene monomer offers advantages of low cost, desirable reactivity

and many years of formulating experience in the field. The generalisation
may be made that for a given vinyl ester resin, there is an optimum
styrene concentration with respect to physical properties. Thus with
Derakane 470, epoxy novolac vinyl ester resin, the maximum heat distor-
tion temperature (HDT) is achieved with a 36 %styrene dilution as shown
in Fig. 2.1.
The effect of the styrene content on the corrosion resistance of vinyl
ester resins has been reported. 30 In contrast to polyester resins, vinyl
ester resins have good corrosion resistance at low monomer levels.
Vinyl toluene diluted vinyl ester resins were developed for use in electrical
applications. These resins were shown to have less water absorption,
higher flexural strength and electrical strength retention v. heat ageing
than a conventional polyester resin diluted in styrene. Vinyl toluene diluted
vinyl ester resins offer the advantages of easier processability (especially
in matched metal die moulding) and lower cost than epoxy resins in
electrical applications. Vinyl toluene is less volatile and has a higher
flash point than styrene. The former property is important in view of
the push for lower monomer emissions during fabrication and processing.
Dicyc10pentadiene acrylate-diluted vinyl ester resins have the advantage
of longer shelf-life than conventional styrene-diluted vinyl ester resins.
Dicyclopentadiene acrylate is more compatible with vinyl ester resins
than styrene, which is advantageous since higher levels of dicyc1openta-
diene acrylate are required to produce comparable viscosities. A disadvan-
tage of dicyc10pentadiene acrylate monomer is that it yields resins which
are more brittle.
Table 2.8 shows the mechanical properties of a vinyl ester resin in
several monomers.
The low tensile strengths and elongations of the resin diluted in dicyc1o-
pentadiene acrylate and divinyl benzene are indicative of brittleness.
2.10. VINYL ESTER RESIN LAMINATE PROPERTIES
Although laminates have been made and tested from vinyl ester resins
using graphite or carbon fibres, Nexus® polyamide fibres, and a wide
variety of polyester fibres, glass fibres are commercially by far the most
important reinforcements. Since this book is primarily about resins, only
a few basic glass fibre laminates will be discussed. Hand lay-up, filament
® Trademark of Burlington Industries.
TABLE 2.9
TYPICAL PROPERTIES OF HAND LAY-UP LAMINATE (in (6-4mm) THICK)
Property PS 15-69 Structure II Structure IV Structure V

requirement basic resin epoxy novolac resin flame retardant resin
~o Monomeric styrene 45 36 40
Cast resin HDT of (0C) 210-215° 290-305 225-235
Flexural strength, psi (MN m - 2)
Room temperature 19000 (131) 29600 (204) 24000 (165) 23800 (164) <:
150 of (66°C) 28500 (196) 23800 (164) Z
-<
t""
200 of (93 0C) 27400 (189) 24500 (169) 24000 (165)
225°F (107°C) 14700 (101) 21000 (145) ..,1Jl
250°F (121 0C) 5000 (34) tTl
24100 (166) 12000 (83) :0::1
300°F (149°C) 3200 (22) 21000 (145) :0::1
tTl
325 of (163°C) 12000 (83) CI'J
350 of (177 0C) 8000 (55) Z
CI'J
Flexural modulus, x 10 5 psi (GNm- 2 )
Room temperature 8 (5,5) 10·3 (7·1) 12·5 (8'6) 11 (7'6)
150°F (66°C) 10·1 (7'0) 11(7'6)
200°F (93°C) 8·5 (5'9) 11·8 (8,1) 9·7 (6'7)
225°F (lOrC) 4·9 (3-4) 8·2 (5,7)
250°F (121 0C) 2·3 (1'6) 10·6 (7,3) 5·8 (4,0)
300°F (149°C) 2·3 (1'6) 8·3 (5'7)
325 of (163 0C) 6·1 (4,2)
350°F (177°C) 5·2 (3'5)
Vl
"""
TABLE 2.9-contd. -"'"
0"-
Property PS 15-69 Structure II Structure IV Structure V

reqUirement baSIC reSIn epoxy novolac resin flame retardant resin
Tensile strength, psi (MN m - 2) -l

Room temperature 12000(83) 20700 (143) 18000 (124) 16400 (113) :t
0
150°F (66°C) 25100 (173) 18300 (126) s:::
>
CJ>
200°F (93°C) 21800 (150) 18600 (128) 19500(134)
225°F (107°C) 18200 (126) 18500 (128) :-'1
250°F (121 0c) II 700 (81) 18800 (130) 17000 (117) z>
300°F (149°C) 7700 (53) 17000 (117) 0
tTl
325 OF (163 0c) 14400 (100) :;o:l
CJ>
350 OF (177 0c) 11000 (76) 0
z
Tensile modulus, x 10 5 psi (GN m - 2)
Room temperature 17-4 (12·0) 16·5 (11·4) 15 (10·3) z>
0
150°F (66°C) 18·1 (12·5) 17 (11·7) <:
200°F (93°C) 14·9 (10·3) 17·1 (11·8) 13 (9·0) ;.;
225 OF (107 0c) Cl
11·1 (7·7) 12·6 (8·7)
250°F (121 0c)
Z
7·6 (5·2) 17·1 (11·8) 12(8·3) ;;
300°F (149°C) 10-4 (7·2) ~
325 OF (163 0c) 9·1 (6·3) s:::
tTl
350°F (177°C) 7·3 (5·0) CJ>
CJ>
Wr = Woven roving glass.

n
~
Glass Content: 40%.
Laminate Construction: V-M-M-Wr-M-Wr-M.
V = Std. 10 mil. corrosion-grade C-glass veil.
M = Chopped strand mat of 1·50zft- 2 (457gm- 2 ).
Courtesy of The Dow Chemical Company.
wound and matched metal die moulded laminates make up the bulk of
the commercial laminates being sold.
The biggest use of vinyl ester resin hand lay-up laminate is in making
chemical resistant equipment. Table 2.9 shows the typical properties of
-i- in (6-4 mm) thick, 40 % glass fibre laminate, made with three different
vinyl ester resins. Their properties are compared with the US Bureau
of Standards' requirement PS 15-69 which was developed for polyester
corrosion resistant equipment. Note the high physical properties and
excellent retention of physical properties at elevated temperatures. This
typical data has been used extensively in the design of reinforced plastics
equipment. Four pressure vessels designed using this data were pressure
tested up to 1·5 times operating pressure. Only one of more than 100
strain gauges used showed a significantly higher strain level than was
predicted.
The properties of filament wound laminate vary with the wind angle
of the glass fibre. In pipe wound at an angle of 54·5 a to the axis, having
a nominal 65 %glass content, the tensile stress at break is 37000-43000 psi
(255-296 MN m - 2).
TABLE 2.10
PHYSICAL PROPERTIES OF MATCHED DIE MOULD VINYL ESTER SMC AND HSMC
Property SMC HSMC HSMC
CompOSition
% glass 40 50 65
% resin 30 25 35
% filler 30 25 0
Tensile Strength psi (MN m - 2) 19000 (131) 26000 (179) 32000 (221)
Flexural Strength psi (MN m - 2) 43000 (296) 50000 (345) 62000 (428)
Matched die laminates of vinyl ester resins have slightly better static
physical properties and significantly better dynamic properties than similar
laminates made with polyester resins. One of the sizeable use areas is
automotive parts moulded from SMC and HSMC (high strength sheet
moulded compound) vinyl ester resins. Table 2. \0 gives some of the
physical properties of automotive type moulded laminate.
2.11. VINYL ESTER RESIN APPLICATIONS
Because vinyl ester resins are made from higher cost raw materials and
often use extended process times, they have higher costs than do standard
polyester resins. For this reason they are only used commercially III
applications where their superior properties are needed.
The major uses of vinyl ester resins will be discussed by markets.
The markets are:
corrosion resistant,
land transportation,
electrical insulation,
marine and radiation curing,
others.
2.11.1. Corrosion Resistant Applications

By far the largest use of vinyl ester resins is in making corrosion resistant
equipment. When the basic vinyl ester resin is made with methacrylic
acid end caps (Structure (II» the resultant shielding of the ester linkages
gives a resin with exceptional resistance to a broad range of acids and
alkalis. The 'E' type glass fibres used as reinforcements in making corrosion
resistant equipment have high strength but poor resistance to acids and
alkalis. The pendant hydroxyl groups on the vinyl ester chain assist
in wetting and bonding to glass fibres. Superior protection of the glass
fibre results. These characteristics, coupled with the high tensile elongation
FIG. 2.2. Pulp mill recovery boiler scrubber demister.

(toughness) of the resins, have been responsible for their wide use in
this market. Unstressed corrosion resistance testing in the laboratory
shows the vinyl ester resins to be very similar to the bisphenol A fumarate
and chlorendic anhydride polyesters in many chemicals. In plant service,
however, the higher toughness of the resin decreases the dangers of
mechanical damage to the corrosion liner during fabrication, shipping,
installation, operation and maintenance. The resulting superior perform-
ance has led to wide use of the vinyl ester resins in tanks, piping, absorption
.... " '"ppOrl
FIG. 2.3. 400 ft (122 m) chimney liner installation (schematic).

towers, process vessels, hoods, ducts, scrubbers and stacks. The trio
of scrubbers, demisters and stacks shown in Fig. 2.2 are an excellent
example of a cost effective use of the basic VInyl ester resin. These units,
13ft8in (4.17m) in diameter at the base and 100ft (30·5m) tall, scrub
sulphur dioxide gases at 180 OF (82°C) from a pulp mill's two recovery
boilers. These large structures were transported hundreds of miles to
the mill site and erected. They had been in service over six years at
the time that an inspection showed them to be essentially in original
condition.
Some idea of the importance of toughness in resins for corrosion
applications can be gained by looking at the possibilities for rough
mechanical overstressing shown in Figs. 2.3 and 2.4. These show an
11 ft 3 in (3-43 m) diameter stack liner that was made from the fire retardant
vinyl ester resin (Structure (V)) in 45 ft (13·7 m) long sections and installed
in a 400 ft (122 m) tall concrete chimney. Figure 2.3 shows schematically
FIG. 2.4. Inserting chimney section through breaching.

the method of installation, while Fig. 2.4 shows one of the 45 ft (13, 7 m)
long sections being inserted through the breaching in the concrete chimney.
No cracks were found in this 400ft (122m) tall stack liner when it was
originally inspected after installation, nor when inspected after seven
years of service handling smelter gases.
The vinyl ester resins made with a novo lac backbone (Structure (IV))
retain good general corrosion resistance, and much of the resistance
to solvents and high temperatures that the phenolic resins, used in their
synthesis, display. Equipment made with these resins has been in service
over 10 years at 340 OF (172 0c) handling chlorine and hydrogen chloride
wet vapour. These resins are widely used to make tanks and piping
that handle organic contaminated waste water and by-product hydro-
chloric acid that is contaminated with organic and chlorinated organic
compounds. Figure 2.5 shows a tank trailer, made with the epoxy novo lac
vinyl ester on the inside with the fire retardant resin on the exterior, that
hauls by-product hydrochloric acid one direction and 50 % caustic on
the return.
FIG. 2.5. Reinforced plastic tank trailer.
2.11.2. Land Transportation Applications

The second largest use of vinyl ester resins is in making body and structural
parts for land transportation equipment. Much of this equipment,
e.g. automobiles, trucks, buses, snowmobiles and tractors, is very competi-
tive and is assembled in large numbers on assembly lines from painted,
press moulded parts. These parts must have a high heat distortion tempera-
ture to withstand the paint ovens, while being tough enough to withstand
rough assembly line handling and the vibration, impact and other mechani-
cal stresses of the particular service.
In production of the body for the Chevrolet Corvette, there were prob-
lems with cracked polyester SMC (sheet moulding compound) exterior
parts at the end of the assembly line. When vinyl ester SMC parts replaced
the polyester parts, cracks due to assembly line handling disappeared.
With the emphasis on improved fuel economy and weight reduction
in automobiles, many development programmes for structural parts are
moving rapidly. The excellent fatigue life of parts made from vinyl ester
resins makes them the logical choice for structural parts. One of the
most promising is the development of automobile wheels from vinyl
ester HSMC (Fig. 2.6). The glass fibre, high strength sheet mouldings
FIG. 2.6. Vinyl ester HSMC wheel on test car.
weigh only 50--60 % as much as the steel wheel. 31.32 They have, to date,
passed all the specifications for steel wheels, while wheels made from
other resins (polyester and epoxy) have not passed all the required tests.
Railroad hopper car covers made from vinyl ester HSMC pass all
the operation requirements, while being light enough for one man to
open and close easily.
2.11.3. Electrical Insulation Applications

Large motors, generators, trans[grmers and other heavy electrical equip-
ment require insulation that is very tough and does not shrink or crack
under continuous operation at 160 to 200°C. Special vinyl ester resins

are used to make rigid insulation that retains adequate electrical strength
and dimensional stability for more than 10 years at 200 0C. 33
2.11.4. Marine Applications

Use of vinyl ester resins in marine applications has been limited to high
performance craft such as kyaks, canoes, speed boats and special fishing
rigs. In many cases Kevlar® fibre is used as all or part of the reinforcement
to give lighter weight boats with more puncture resistant hulls.
2.11.5. Radiation Curing Applications

Resins of the Structure (VI) type are used in can coatings, printing inks,
panel coatings and many other areas where rapid cures and short finishing
lines are required. 34 ,3"
2.11.6. Other Applications

I. The aircraft industry has made relatively little use of vinyl ester
resins. Developed in the late 1960s, the Cessna AG truck crop
spraying aeroplane shown in Fig. 2.7 has been a continuing success.
The spray tank shown in Fig. 2.8, made with the basic vinyl ester
resin, is built in between the pilot and the engine as part of the
aircraft structure. The resin was selected because of good resistance
to severe mechanical stresses coupled with a broad resistance to
agricultural spray chemicals.
2. Dental filling material consistently uses a small amount of vinyl
ester resins. The resin's toughness, good resistance to food staining,
FIG. 2.7. Cessna AG truck crop duster.
® Trademark of the E. I. Du Pont Company.

FIG. 2.8. Spray tank for Cessna AG truck.
good bonding chracteristics and ability to gel and cure rapidly and
completely in the presence of moisture all contribute to its per-
formance.
3. Ballistic armour of very light weight has been developed by com-
bining the very high tenacity of Kevlar® fibres with the toughness
characteristics of the vinyl ester resins. This armour is light
enough to be used for personnel and vital equipment protection on
helicopters. It is used in armouring cars and, when made with fire
retardant vinyl ester resin, is used aboard ships.
2.12. NEW DEVELOPMENTS
Any time a new resin is developed by combining the best properties of two
other resins, there is a strong effort to make the new resin better in
all respects. The vinyl ester resins did combine the ease of processing
and curing of polyester resins with the toughness and bonding character-
isti~s Qf eVQ;,\Y resins,
By chemical structure the vinyl ester resins should, and in many cases
did, cure better than polyester resins. Total processibility as good as,
or better, than that obtained with polyester resins evolved over a period
of many years. The better bonding chracteristics obtained from the epoxy
backbone caused many parts to stick in the mould. The high reactivity
of the end of chain double bonds made stabilising the resins a very
delicate procedure; the problem was to prevent polymerisation during
shipping and storage without interfering with catalysed cure. The first
difficulty has been solved, but new resin producers still struggle with
resin stabilisation.
While the vinyl ester resins gave better bonding than polyester resins,
they did not bond as well as the epoxy resins to materials like stainless
steel and aluminium. The rubber modified vinyl ester resins were developed
to give improved bonding to a wide range of substrates (Table 2.4) and
to provide resins with vastly superior impact resistance. 22 Urethane
modified resins 13 (Structure (VIII» were developed that gave improved
bonding.
New vinyl ester SMC resins have recently been developed that are
thickened through a network produced by reacting a polyisocyanate with
a polyol instead of the widely used reaction of a group II oxide or
hydroxide with acid end or pendant groups. Very fast SMC thickening
times and high laminate strengths are reported for these new resins. 36
A brominated, novo lac-based vinyl ester resin was recently developed
to provide a fire retardant resin with superior organic resistance and
superior heat ageing resistance. The improved organic resistance was
developed to handle the requirements of the process of solvent extraction of
uranium from crude phosphoric acid and from copper mine tailings. 37 The
improved heat ageing resistance was developed to provide a matrix resin
for high temperature ductwork and stacks for coal fired power plants,
smelters and related processing plants.
2.13. TOXICOLOGICAL ASPECTS OF VINYL ESTER RESINS
Vinyl ester resins in the uncured state are not highly toxic by ingestion,
but are capable of causing significant eye, skin and respiratory disorders.
Eye contact may cause slight irritation or transient corneal injury. Skin
contact may cause slight to moderate irritation on prolonged contact.
Inhalation of concentrated va pours may cause irritation of respiratory

passages and slight anaesthesia. It is recommended that eye, skin and
vapour contact with uncured vinyl ester resins be avoided. Normallabora-
tory and plant standards for good housekeeping and personal hygiene
should be followed.
Vinyl ester resins that are completely polymerised (cured) are considered
to be toxicologically inert and should present no health problems from
handling. The cured resins may, however, present a health problem from
inhalation of dust generated during machining or grinding, especially
if they contain glass, silica powders, asbestos or metal powders.
Users should consult with the manufacturers of initiators and promoters
or other materials used in formulating with vinyl ester resins for safe
handling procedures.
REFERENCES
I. BOWEN, R. L., US Patent No. 3,066,112 issued to the US government, Dental

filling material comprising vinyl silane jused silica and a binder consisting oj the
reaction product oj blsphenol and glycldyl acrylate, Nov. 27, 1962.
2. BOWEN, R. L., US Patent No. 3,179,623 issued to the US government, Method
oj preparing a monomer having phenoxy and methacrylate groups linked by
hydroxy glyceryl groups, April 20, 1965.
3. FEKETE, F., KEENAN, P. J. and PLANT, W. J., US Patent No. 3,221,043 issued to
H. H. Robertson Co., Ethylenically unsaturated dihydroxy dlesters, Nov. 30,
1965.
4. FEKETE, F., KEENAN, P. J. and PLANT, W. J., US Patent No. 3,256,226 issued to
H. H. Robertson Co., Hydroxy polyether polyesters having terminal
ethylenically unsaturated groups, June 14, 1966.
5. SZOBODA, G. R., SINGLETON, F. and ESHLEMAN, L., US Patent No. 3,621,093
issued to H. H. Robertson Co., Process jor making reinforced thermoset
articles, Nov. 16, 1971.
6. BEARDEN, c., US Patent No. 3,367,992 issued to The Dow Chemical Company,
2-Hydroxy alkyl acrylate and methacrylate dlcarboxylic aCid part tal esters and
the oxyalkylated derivatives thereof, Feb. 6, 1968.
7. NAJVAR, D. J., US Patent No. 3,524,901 issued to The Dow Chemical
Company, Flame retardant vinyl esters containing acrylic or methacrylic
phosphate esters, Aug. 18, 1970.
8. HARGIS, S. R., JR, US Patent No. 3,524,903 issued to The Dow CheIl)ical
Company, Flame retardant vinyl ester containing alkyl hydrogen phosphate
resin and a halogenated epoxide reSin, Aug. 18, 1970.
9. SWISHER, D. H. and GARMS, D. c., US Patent No. 3,564,074 issued to The Dow
Chemical Company, Thermosetting vinyl resins reacted with dicarboxylic acid
anhydrides, Feb. 16, 1971.
10. PENNINGTON, D. W. and NORTON, F. E., US Patent No. 3,594,247 issued to The
Dow Chemical Company, Filament winding process, July 20, 1971.
11. YOUNG, R. E. in Unsaturated Polyester Technology, ed. Bruins, P. F., 1976,
Cordon and Breach Publ;shers, New Vork, p. 31B.
12. YOUNG, R. E. in Unsaturated Polyester Technology, ed. Brums, P. F., 1976,
Gordon and Breach Publishers, New York, p. 316.
13. LEWANDOWSKI, R. J., FORD, E. c., JR, LONGENEEKER, D. M., RESTAINO, A. J.
and BURNS, J. P. 30th SP! Reinforced Plasflcs/Composites Conj., 1975, New
high performance corrosion resistant resin, Paper 6-B.
14. NIESSE, J. and FENNER, O. Effect of glass content and laminate structure on the
mechallical properties of FRP structures, Nat. AssocwtlOn of Corrosion
Engineers, CorrosIOn 78, March, 1978, Paper 105.
15. LINOW, W. H., BEARDEN, C. R. and NEUENDORF, W. R. 21st SP! Reinforced
Plastics/Composites Conj., 1966, The DERAKANE resins-a valuable
addition to thermosetting technology, Paper I-D.
16. JERNIGAN,J. W., BEARDEN,C.R. and PENNINGTON, D. W. 22ndSP! Reinforced
Plastics/Composites Conj., 1967, High speed 'B-staging' with vinyl ester resins
-a route to the automation of remforced plastics moulding, Paper 8-D.
17. STAVINOHA, R. F. and MACRAE, J. D. 27th SP! Reinforced Plastics/Composites
Coni, 1972, DERAKANE vinyl ester resins-unique chemistry for unique
SMC opportunities, Paper 2-E.
18. CUTSHALL, J. E. and PENNINGTON, D. W. 27th SPI Reinforced Plastics/
Composites Coni, 1972, Vinyl ester resin for automotive SMC, Paper 15-A.
19. THOMAS, R. E. and ENOS, J. H. 33rd SP! Reinforced Plasflcs/Composlfes Conj.,
1978, Very high strength SMC in automotive structures, Paper 4-E.
20. NOWAK, R. M. and GINTER, T. 0., US Patent No. 3,674,893 issued to The Dow
Chemical Company, Thermosettable resins containing vinyl ester reSin,
polydiene rubbers and Vinyl monomers, July 4, 1972.
21. CRAVENS, T. E. 27th SP! Reinforced Plasl/cs/Composlles Conj., 1972,
DERAKANE 470-45, a new high temperature corrosion resistant resin, Paper
3-B.
22. HAWTHORNE, K., STAVINOHA, R. and CRAIGIE, L. 32nd SP! Reinforced
Plastics/Composites Conj., 1977, High bond-super tough-CR resin, Paper
5-E.
23. BRINKMAN, W. H., DAMEN, L. W. and SALVATORE, M. 23rd SP! Reinforced
Plast ics/Composites Conj., 1968, Accelerators for the organic peroxide curing
of polyesters and factors influencing their behaviour, Paper 19-D.
24. CASSONI, J. P., HARPELL, G. A., WANG, P. C. and ZUPA, A. H. 32nd SP!
Reinforced Plastics/Composites Conj., 1977, Use of ketone peroxides for room
temperature cure of thermoset resins, Paper 3-E.
25. THOMAS, A., JACYSZN, 0., SCHMITT, W. and KOLCZYNSKI, 1. 32nd SP!
Reinforced Plastics/Composites Conj., 1977, Methyl ethyl ketone peroxides,
relationship of reactivity to chemical structure, Paper 3-B.
26. VARCO, P. 30th SP! Reinforced Plastics/Composites Conf., 1975, Important
cure criteria for chemical resistance and food handling applicat ons of
reinforced plastics, Paper 6-C.
27. AWAJI, T. andATOBE, D., US Patent No. 4,129,609 issued to Nippon Shokubai
Kagaku Kogyo Co., Ltd, Methodjor Improving storage stability of epoxy ester
thermosetting resins with thiuram compounds, Dec. 12, 1978.
28. YOUNG, R. E. in Unsaturated Polyester Technology, ed. Bruins, P. F., 1976,

Gordon and Breach Pubhshers, New York, p. 336.
29. RUBENS, L. C, THOMPSON, C F. and NOWAK, R. M. 20th SP] Reinforced
Plastics/Composites Con!, 1965, The advantage of chlorostyrene diluted
polyesters III production of reinforced plastics, Paper 2-C
30. SAMS, L. L., PIEGSA, 1. G. and ANDERSON, T. F. 28th SP] Remforced
Plastics/Composites Conj., 1973, Monomer content vs. corrosIon resistance of
vinyl ester resins, Paper 13-A.
31. WOELFEL, 1. A., O'NEIL, K. B., BREDE, A. and SMITH, R. W. 34th SP]
Reinforced Plasllcs/Composites Conf., 1979, Reinforced composite wheels:
light-wheel of the future, Paper 21-A.
32. STANLEY, A. 34th SP] Reinforced Plastics/Composites Conj., 1979, Wheel wells
of structural SMC-for transportation industry, Paper 21-C
33. WHARTON, W. H., ESCH, F. P. and ANDERSON, T. F. Proceedings of the lOth
Electncal]nsulatlOn Conj., Sept. 1971, Improving price-performance ratio of
hIgh temperature laminates, IEEE Pub. No. 71-C38-El, p. 40.
34. CRARY, E., US Patent No. 3,661,576 issued to the W. H. Brady Company,
Photopolymerizable compositions and articles, May 1, 1972.
35. SHUR, E. and DABAL, R., US Patent No. 3,772,062 issued to the Inmoat
Corporation, UltravIOlet curable coating composition:" Nov. 13, 1973.
36. FERRANINI, 1., MAGRANS, 1. 1. and REITZ, 1. A., III. 34th SP] Remforced
Plastics/Composites Con!, 1979, New resins for high strength SMC, Paper 2-G.
37. CRAIGIE, L. 1., RUSSELL, D. L. and HARTLESS, A. L. 34th SP] Reinforced
Plastics/Composites Conj., 1979, A new brominated novolac-base vinyl ester
resin, Paper 8-B.
Chapter 3
POLYESTER RESIN CHEMISTRY
T. HUNT
BP Chemicals Ltd, Penarth, South Glamorgan, UK
SUMMARY
This chapter reviews recent developments in the processing andformulation

of unsaturated polyester resins. Considering the condensation polymer
first, refinement oj existing production techniques is discussed and the
alkylene oxide route to the preparation of polyester resins is compared
with the conventional diol based process. New, or relatively new, building
blocks which can be used to prepare resins with improved properties are
also considered.
Turning to the many additives which can be incorporated to modify
the properties ofpolyester resins,jiame retardant formulations are reviewed
and efforts to improve the maturation, shrinkage and toughness of moulding
compounds are summarised.
A novel polyester cement, curable by the addition oj water, an ingenious
newfoaming system and variousjormulations that emit less styrene vapour
duringjabrication than standard grades arefurther examples oj the valuable
properties which can be introduced via additives.
Future trends in polyester synthesis are considered.
3.1. INTRODUCTION
A high proportion of the total sales of unsaturated polyester resins is

made up of general purpose grades. These are mainly styrene solutions
of polyesters derived from maleic anhydride, phthalic anhydride and
59
60 T. HUNT
I ,2-propane diol (propylene glycol). A wide variety of performance speci-

fications can be satisfied by varying the proportions of these three basic
ingredients. Viscosity is varied by adjusting the styrene content; reactivity
is largely dependent on the molar ratio of maleic anhydride to phthalic
anhydride. Increased flexibility, if required, is introduced by partial
replacement of propane diol with a longer chain molecule such as
diethylene glycol. Alternatively phthalic anhydride can be partially
replaced with a linear dicarboxylic acid such as adipic acid.
Speciality resins which, when cured, have improved resistance to
hydrolysis and better mechanical properties are obtained by using iso-
phthalic acid instead of phthalic anhydride.
Polyesters made with a halogenated building block such as tetrachlor-
phthalic anhydride instead of phthalic anhydride are the basis of formula-
tions which give cured products with a reduced tendency to burn.
Further variations can be produced by the use of additives. Dispersion
of a small quantity (0·5-3 %) of fumed silica introduces thixotropy. A
small paraffin wax addition (c. 0·05-0·1 %) reduces residual surface tack
after cure. Tertiary aromatic amines boost the cure rate of resins polymer-
ised with a cobalt salt/ketone hydro peroxide system. Addition of an
organohalogen compound plus antimony trioxide is an efficient means
of reducing the tendency to burn. Refractive index can be reduced, in
order to produce clearer fibreglass laminates, by substituting methyl meth-
acrylate for some of the styrene. This also improves the weathering proper-
ties of the cured resin. These and many other variations have been described
elsewhere. 1.2
In this chapter we shall be considering recent developments in polyester
chemistry and improvements/modifications in performance which can
be achieved by the use of various additives. It is in this latter area
that the main development effort has been concentrated in recent years.
Resin development has been largely confined to refinement of existing
processing techniques and modifications of standard formulations to suit
new application areas or improved methods of fabrication. Some building
blocks new to the unsaturated polyester resin field, which yield products
with improved properties, are also being used.
3.2. POLYESTER MANUFACTURING PROCESSES
3.2.1. The AIkylene Oxide Route to Polyesters

Increased interest is being shown in this process. Unsaturated polyester
resins can be prepared by reacting an alkylene oxide, especially propylene
POLYESTER RESIN CHEMISTRY 61
oxide, with maleic anhydride and optionally phthalic anhydride in the

presence of an hydroxyl-containing initiator (water, alcohols, carboxylic
acids). The reactions taking place are shown below.
(a) Alcohol initiated
o
/
CHC
ROH+
"'-0 CHCOOR
--+ I (1)
/ CHCOOH
CHC
~
o
CHCOOR 0 CHCOOR
II + /"'- --+11
CHCOOH R,CH---------CH 2 CHCOOCH 2 CH(OH)R, (2)
reacts with more anhydride,
and so on
(b) Acid initiated
o
/
CHC
II>
CHC R,
RCOOCH 2 CH(OH)R, ---+) RCOOCH 2 CH "" o I
(4)
I
OCOCH
II
HOOC-CH
reacts with more oxide
(c) Water initiated
o
/
CHC
"'- CHCOOH
HOH+ II 0--+ I (5)
/ CHCOOH
CHC
~
o
62 T. HUNT
CHCOOH o CHCOOH
II + / '" ~ II (6)
CHCOOH R 1CH----CH 2 CHCOOCH 2 CH(OH)R 1
o
HOH /'" 2 ~ R 1CH(OH)CH 2 0H
+ R 1CH----CH (7)
In all these systems, incoming oxide can react with both hydroxyl and
carboxyl terminated fragments. It is important, therefore, to use a selective
catalyst which promotes esterification rather than etherification if resins
of similar structure to diol based polyesters are to be produced. Alterna-
tively, reaction conditions can be adjusted so that esterification is the
dominant activity, without using a catalyst. Both options are described
in patents. Oxides and salts of magnesium, zinc and lead were preferred
catalysts in an early patent. 3 Preferred reaction temperature was
120-130°C in order to get light coloured products and the propylene
oxide addition was carried out over about 5 hours. It was later realised
that maleate to fumarate isomerisation did not take place efficiently
at these low reaction temperatures which explained the somewhat inferior
cure properties of products made in this way, relative to normal diol
based resins. Later processes, therefore, incorporated a higher temperature
isomerisation stage, after the propylene oxide had been reacted. Preferred
catalysts then included alkali metal hydroxides and salts,4lithium chloride
or a quaternary ammonium halide,5 and organic acid salts of zirconium,
titanium or cerium. 6
Addition of a small quantity of phosphoric acid before commencing
the isomerisation stage at 200°C was found to prevent discoloration.
In an uncatalysed process 7 the alkylene oxide is added to a vigorously
stirred mixture of anhydrides and initiator, in the most turbulent zone,
the mixture being maintained within the temperature range 130-220°C.
Pale coloured products are claimed to result.
The alkylene oxide route to unsaturated polyester resins has a number
of advantages over the traditional glycol based process. No aqueous
distillate is produced and the preferred propylene oxide is slightly cheaper
than propylene glycol. The oxide ring opening is exothermic, so heat
input is reduced. Fast cycle times are possible, provided that the addition
rate and dispersion of the oxide are optimised. Also, a continuous process
could be based on this route.
Disadvantages include the relatively limited range of formulations which
can be made in this way, i.e. without a normal esterification stage in
POL YESTER RESIN CHEMISTRY 63
which water is evolved and the higher hazard ratings of the oxides relative
to the corresponding glycols.
At the present time, the majority of resin manufacturers use the glycol
route. When new plants are being planned, however, the oxide process
will be a serious contender, particularly when manufacture of general
purpose grades is the objective. Half the polyester output of one manu-
facturer is now made using propylene oxide. 8
3.2.2. Refinement of Existing Processes

A recent innovation is the use of molten maleic and phthalic anhydrides
to speed up charging and reduce heat-up times. The molten anhydrides
are stored in heated tanks at 60°C for maleic anhydride and 155°C
for phthalic anhydride.
Isophthalic acid based unsaturated polyester resins continue to be made
by a two stage process, because of the low solubility and reactivity of
this acid relative to maleic anhydride. In the first stage, isophthalic acid
and the full glycol charge are reacted together until the hot mix is clear.
Maleic anhydride is then added, and polyesterification is completed. If
the glycol is propylene glycol, the stage-one temperature is limited to
about 180°C at atmospheric pressure, and overall process times are con-
siderably longer than those for the corresponding orthophthalates. Use
of catalysts, such as butyl stannoic acid, or higher temperature processing
under pressure are now being recommended to reduce stage-one process
times.
The kinetics and mechanism of the polyesterification process are also
being examined in efforts to reduce production scale process times and
improve product consistency. 9
3.3. TEREPHTHALIC ACID
Polyester resins in which the main saturated dicarboxylic acid component

is terephthalic acid are now available commercially. From published data 2
the main advantages gained by the use of terephthalic acid relative to
equivalent isophthalic acid formulations appear to be increased heat distor-
tion temperature and some reduction in shrinkage during cure. There
is said to be little, if any, improvement in chemical resistance. Nevertheless,
the terephthalates have gained a foothold, mainly for applications
requiring chemical resistance.
64 T. HUNT
The disadvantages associated with the use of this acid are considerable.
It is the slowest reacting of the three phthalic acids and catalysts or
described for isophthalic acid in Section 3.2.2)

pressure are required in the first stage of the two stage process (as already
to achieve acceptable
process times. There are also more constraints on the formulation of
resins than is the case with isophthalic acid or phthalic anhydride. These
derive from its symmetrical structure and its tendency to give products
with reduced solubility in styrene when reacted with symmetrical glycols
such as neopentyl glycol.
3.4. DICYCLOPENTADIENE (DCPD)
DCPD used to be available in relatively small quantities from coal tar

operations. Large quantities are now recoverable as a by-product of
the steam cracking of naphtha for ethylene production at a price below
those of the basic polyester raw materials.
The following reactions make DCPD an interesting raw material for
the preparation of high performance unsaturated polyester resins.
(CO heat
16(}170°C
20 (8)
DCPD
cyc10pentadiene
0
CH-C'O
+ II ""/ 0 )I ([JCO :=0 (9)

CO
CH-C'O
endomethylene
maleic tetrahydrophthalic
anhydride anhydride (EMTHPA)
(CO+ROH _. catalyst
heat
)
RO
fCC) ~
(10)
ROH may be water, an alcohol or a carboxyhc acid

POL YESTER RESIN CHEMISTRY 65
Until recently, only the Diels~Alder addition of cycIopentadiene

(eqn. (9)) has been utilised commercially in order to obtain unsaturated
polyester resins containing endomethylene tetrahydrophthalate groupings.
There are two main methods of achieving this Diels~Alder addition.
In the first, DCPD (1 mole) is heated with maleic anhydride (2 moles)
under reflux in an atmosphere of nitrogen at 170~ 190°C. The resulting
EMTHPA is then reacted into an unsaturated polyester in the normal
manner.
In the second method, propylene glycol and maleic anhydride are
polyesterified and the polyester is then heated under reflux with DCPD
at about 170°C.
Latterly, attention has been concentrated on the acid catalysed addition
of alcohols or acids to the double bond in the strained ring of DCPD
(eqn. (10)). Three procedures have been described/Q,ll all rely on the
fact that the acidity of a maleic anhydride/glycol 'half ester', or of maleic
acid itself, is sufficient to catalyse the addition to DCPD, as shown
in equations (11) to (13).
Half ester preparation
o
/
CHC
~O + HOCH,CH,OH
c 150'C
II ~ CHCOOCH 2 CH 2 0H (11)
II
CHC CHCOOH
~
o
Reaction of half ester with DCPD
CHCOOCH 2 CH 2 0H
II + ~~ ~
CHCOOH ~ CHCOO~
I
CHCOOCH 2 CH 2 0H
+ (12)
CHCOOCH,CH'O~
II
CHCOOH
66 T. HUNT
Hydrolysis method
o
/
CHC
CHCOOH
CHC
/ "" O+HZO II
CHCOOH
'\-
""'00/([0
o
(13)
CHCOO~
II
CHCOOH
In the so called 'beginning' method, DCPD, maleic anhydride and
propylene glycol, in mole ratio of, for example, 0,25:1:1'1, are heated
to c. 150°C in an atmosphere of nitrogen and are held at this temperature
under reflux for 30-60 min. The condenser is then set for distillation,
and polyesterification is completed in the usual way at 200°C.
In the 'half ester' method, maleic anhydride and propylene glycol are
heated at 150°C under nitrogen to prepare the half ester. DCPD is
then dripped in, under reflux whilst slowly increasing the temperature
to about 175°C. When no DCPD refluxes, the temperature is raised
to 200°C and polyesterification is completed.
In the hydrolysis method 11 maleic anhydride is converted to maleic
acid by heating with an equimolar quantity of water at 80-90°C. DCPD
is then added and, when reaction is complete, the glycol and, optionally,
a modifying acid are charged prior to polyesterification at 200°C.
Polyester resins prepared by the last three methods have lower molecular
weights than general purpose orthophthalate resins and Diels-Alder
addition types. This is due to the chain terminating nature of the DCPD
addition reaction. At a given concentration in styrene they are, therefore,
significantly less viscous than the other resins.
DCPD modified polyesters, whether Diels-Alder addition or DCPD
addition types, give cured products having outstanding resistance to heat
and ultraviolet light. They also have high heat distortion temperatures
and enhanced chemical resistance relative to orthophthalate resins (see
Ref. 10 for figures).
At the present time, with the high relative costs of maleic anh~dride
and phthalic anhydride, modification with DCPD looks an attractive
proposition. The DCPD addition route is more economical than the
Diels-Alder addition, since the latter requires a higher proportion of
expensive maleic anhydride in the initial charge to produce a resin with
reasonable reactivity. In the DCPD addition, the maleate unsaturation
remains intact and there is scope for incorporation of modifying acids
or phthalic anhydride.
3.5. GLYCOLS
3.5.1. Glycols in Standard Resins

The vast majority of unsaturated polyester resins are based on 1,2-propyl-
ene glycol, which currentiyt costs c. £370/tonne. Ethylene glycol
(c. £500/tonne) based polyesters have reduced solubility in styrene, but
this glycol is preferred in fire resistant formulations, particularly with
a halogenated acid component. The latter usually improves solubility
in styrene, and the polyesters have less tendency to burn than those
containing propylene glycol.
Diethylene glycol (c. £450/tonne) is used where a cured resin with
enhanced flexibility is required.
The glycols described in this section are all significantly more expensive
than propylene glycol. However, the improved properties obtainable have
enabled them to gain a foothold where high performance resins are
necessary.
3.5.2. Neopentyl Glycol (NPG)

CH 3
I
HOCHz-C-CHzOH
I
CH 3
NPG costs c. £500/tonne. It has frequently been advocated for use when
cured products with improved hydrolytic stability are required. In the
USA where gel coated polyester is widely used to make baths, etc, the
preferred gel coat resins for this application are now based mainly on
isophthalic acid, maleic anhydride and neopentyl glycol. In the UK,
t All prices in this chapter relate to the last quarter of 1979.

68 T. HUNT
baths and wash basins are more frequently made with a vacuum formed
acrylic outer surface, reinforced behind with fibre glass, plus a general
purpose polyester resin. There has, therefore, been less demand for these
high performance gel coat resins here.
N PG gel coats are also making a significant impact on the marine
market. They are claimed to reduce the incidence of gel coat blistering
which can occur when boats have been in service for some time. 1z However,
there is no single, simple solution to this problem as workshop practice,
choice of back up resin and the type of glass reinforcement used are
all contributory factors. 13 •14
The symmetrical structure of N PG can be used to produce resins
with novel properties. Polyesters made from a glycol component (compris-
ing at least 80 mole % NPG, the remainder preferably consisting of
other symmetrical glycols such as ethylene glycol) and an acid component
(made up of mainly fumaric acid and 12-25 mole % isophthalic acid)
yield blends with styrene which are solid at normal ambient temperatures.
This is believed to be due to the increased crystallinity of these polyesters. ls
The novel polyester/styrene mixtures have melting points ranging from
about 15°C to 100 0c. Some have a buttery consistency at ambienttempera-
ture; others can be ground to yield free flowing powders. One application
for these 'solid' polyesters is the preparation of a sheet moulding compound
which does not require the addition of a maturation agent (see Section 3.8.1)
to produce the necessary thickening. The only disadvantage would seem to
be that the resin needs to be heated in order to liquefy it during the
impregnation stage.
A further application for the above, highly crystalline systems involves
blending them with normal liquid unsaturated polyester resins in order
to introduce thixotropy. Compositions containing between 10 %and 50 %
of the 'solid' polyester can be used.
3.5.3. 2,2,4-Trimethyl-l,3-Pentanediol (TMPD)
CH 3 CH 3
I I
HOCH z-C-CH-CH-CH 3
I I
CH 3 OH
TMPD gives hydrolysis-resistant polyesters as a result of its sterically

hindered hydroxyl groups. It currently costs c. £640/tonne in the UK,
but enjoys a slight price advantage over NPG in the USA.
To prepare polyesters from TMPD, a stabiliser such as triethanolamine

%
(c. 0·2 of the acid + TMPD charge) is usually added at the start of
the reaction. This minimises cyclisation reactions.
Extensive data on TMPD based polyester formulations suitable for
a wide range of end uses have been published by Eastman Chemical
workers. 16 •17
Significant applications for TMPD polyesters in the USA include
corrosion resistant storage tanks, gel coat resins for sanitary ware and the
preparation of cultured marble.
3.5.4. 1:4 Cyclohexane Dimethanol (CHDM)
CHDM costs c. £1050/tonne. It reacts readily with dicarboxylic acids

and anhydrides to give polyesters having excellent hydrolytic stability.
The symmetrical structure of this glycol can be used to produce highly
crystalline polyesters in a similar manner to that already described for
NPG. For this application at least 55 mole % of the glycol component
should be CHDM and at least 50 mole %of the acid component should
be fumaric acid. When hot solutions of the products in styrene are cooled
to ambient temperature, solid blends having similar applications to those
based on NPG are obtained. 18
In a further interesting application, polyesters made from, for example,
CHDM, maleic anhydride and tetrahydrophthalic anhydride are claimed
to give cured products with excellent chemical resistance, thermal stability
and electrical properties. 19 The absence of aromatic structures in the
polyester molecule is believed to be responsible for the high performance
of these resins relative to general purpose grades based on phthalic
anhydride. In this respect they are comparable with the dicyclopentadiene
modified resins already discussed.
To date, there is little evidence that CHDM is being used to any
significant extent in commercially available unsaturated polyester resins.
70 1. HUNT
3.5.5. Dibromo Neopentyl Glycol (DBNPG)

CH 2 Br
I
H OCH 2 -C-CH 2-OH
I
CH 2 Br
DBNPG costs c. £1500/tonne. It is increasingly being used as a source
of 'built in' bromine in unsaturated polyester formulations designed to
have a reduced tendency to burn. 20 A major disadvantage is that some
HBr is released during polyesterification. Glass lined or highly acid
resistant stainless steel vessels are therefore needed to manufacture satis-
factory products, although the possibility of vessel corrosion can be
reduced by the use of, for example, endomethylene tetra hydro phthalic
anhydride as part of the modifying acid charge. This acts as an HBr
scavenger. 21
With appropriate production vessels, light coloured products with high
levels of combined bromine can be prepared rapidly at a maximum tem-
perature of about 180°C. The recommended procedure is to use DBNPG
as the sole glycol. The products can then be blended with a general
purpose grade if a lower bromine content is required.
In combination with suitable fire retardant additives, (see Section 3.7)
cured products with much reduced tendency to burn can be obtained.
3.5.6. Bis (2-Hydroxyethyl Ether) of Tetrabromo Bisphenol A

Br CH 3 Br
HOCH,CH,O*?~OCH,CH,OH
Br CH 3 Br
This ethylene oxide/brominated Bisphenol A reaction product is also
being offered as a source of 'built in' bromine. It esterifies readily and,
in contrast to DBNPG, does not release HBr at polyesterification tempera-
tures. To date, nothing has been heard of its commercial use in the
unsaturated polyester resin field. It costs c. £2000/tonne.
3.6. NEW MONOMERS
There is no sign of a serious competitor to styrene (currently c. £420/tonne)

for the vast majority of applications. Methyl methacrylate is used to
adjust refractive index downwards where this is required, e.g. in resins

for the manufacture of transparent roofsheet.
Bromostyrene (c. £1300jtonne) is proposed as a source of 'built in'
bromine in flame retardant formulations. Various glycol acrylates and
methacrylates are also available.
3.7. FLAME RETARDANT FORMULATIONS
3.7.1. General Principles and Mechanisms of Fire Retardance

The most common way of making unsaturated polyesters burn less readily
is to introduce organo halogen groupings either in the form of additives
(e.g. chlorinated paraffin wax) or by chemical modification of the polyester
(e.g. by the use ofhexachloroendomethylene tetrahydrophthalic acid (HET
acid) as part of the dicarboxylic acid charge).
The fire resistance of these halogen containing formulations is greatly
improved by the addition of antimony trioxide and/or phosphorus
compounds.
Various mechanisms have been proposed to explain the fire retarding
properties of halogenated formulations. It is believed that HBr or HCl
released on combustion interferes with the free radical reactions proceeding
in the flame. The important reactions in the flame are:
·OH + CO ~ CO 2 + H· (14)
H· + O2 ~ ·OH + o· (15)
The HBr and/or HCl reacts with the hydroxyl radicals, replacing them
with the less reactive Br· or Cl· and R· where RH is an organic compound
containing hydrogen in the flame, according to the following equations:
·OH + HBr ~ H 2 0 + Br· (16)
Br· +RH~HBr+R· (17)
The synergistic action of antimony trioxide in halogen containing formu-

lations may be due to the endothermic formation of volatile antimony
halides, at least one of which, SbCI 3 , is believed to be an effective free
radical trap. These can pass through a number of decompositions and
recombinations, eventually giving the oxide again as a fine dust and
releasing the halogen.
Zinc borate is being offered as a relatively cheap partial replacement
for antimony trioxide. The latter currently costs c. £1750/tonne. It is
72 T. HUNT
thought that the borate functions by forming a protective glass matrix

at the burning surface.
Phosphorus compounds are thought to act by accelerating the evolution
of hydrogen halide from halogen containing formulations. A heat shield
of polymeric meta-phosphoric acid is also said to be formed, which stops
'afterglow' caused by the highly exothermic solid phase oxidation of
carbon.
2C + O 2 ~ 2CO
The heat shield probably accounts for the higher levels of residual
carbon char produced when phosphorus containing formulations are
burned.
The use of alumina trihydrate and molybdenum compounds as fire
retarding additives is discussed in 3.7.6 and 3.7.7.
Flame retardant polyester formulations have been reviewed 22 ,23 and
extensive data on fire retardants collected. 24
3.7.2. Formulation Guidelines

Flame retardant formulations are usually designed to pass a specified
fire test. A common requirement in the UK is a Class I or Class 2
rating to the BS.476 part 7, 1971 Surface Spread of Flame test. A useful
screening test on experimental formulations is to determine their oxygen
index (ASTM 02863-76). The table below gives a very rough guide
to the performance that can be expected in the BS.476 part 7 test for
a given oxygen index.
Oxygen index % Rating by BS.476 part 7, 1971
25 minimum Class 2
c. 38 minimum Class I
N.B. A general purpose polyester specimen would have

an oxygen index of about 19-20%.
An oxygen index of about 25 %is readily obtainable from a combination

of a general purpose resin and one or more fire retardant additives.
A typical composition of this type would comprise 100 parts by weight
of general purpose polyester, 10 parts of chlorinated paraffin wax, contain-
ing 70 % combined chlorine, and 5-10 parts of antimony trioxide. To
achieve the higher indices it is normal practice to use a halogenated
resin plus additives, as a general purpose resin would require excessive

levels of the latter which could reduce the strength of the cured resin
and might also adversely affect weathering properties. A suitable halo-
genated resin might only require the addition of antimony trioxide to
give the desired performance in the fire test.
The additives used depend on whether the particular application needs
a clear resin or win tolerate an opaque one. If a clear resin is required
antimony trioxide cannot be used as it functions as a white pigment.
When an opaque resin is acceptable antimony trioxide win almost certainly
be one of the additives used.
The cost of the formulation, as always, is a major factor. Toxicity
of the additives is another. Tris (2,3-dibromo propyl) phosphate was
frequently used in fire resistant polyester formulations until American
tests branded it as a carcinogen. It is now prudent to insist that a negative
Ames test for mutagenicity is obtained on a new halogen containing
additive or raw material before expensive development work is started.
3.7.3. Polyesters with 'Built-In' Halogen

Where these are required, the preferred halogen containing building blocks
are RET acid, tetrachlorophthalic anhydride and, less frequently, tetra-
bromo phthalic anhydride.
CI
Cl COOH
COOR
CI
hexachloro endomethylene tetra hydro phthalic aCid
(HET acid)
Usage of dibromo neopentyl glycol is increasing (see Section 3.5.5). The

diol shown below prepared from decachlordiphenyl and ethanolamine
is used in France. 25
CI CI CI CI
CI CI CI CI
74 T. HUNT
The corrosion difficulties associated with processing NPG polyesters

have already been discussed. HET acid also tends to liberate Hel if
the polyesterification temperature goes above about 180°C. The remaining
materials, with aromatic bound halogen, do not have this problem. Tetra-
bromo phthalic anhydride, however, as supplied, normally contains traces
of sulphuric acid which must be neutralised. This is achieved by adding
a neutraliser such as sodium acetate or diethanolamine to the polyesteri-
fication charge. 26
3.7.4. Halogenated Additives

A wide range of organohalogen compounds are available. 24 As a general
rule, bromine compounds are more effective than their chlorine equiva-
lents, since the chlorine-carbon bond is stronger than the corresponding
bromine linkage. Similarly, aliphatic halogen compounds are more effec-
tive than aromatic ones, since the aliphatic halogen is more labile. The
thermal stability of the additive needs to be considered when formulating
heat curable systems.
It is advantageous if the halogen compound is soluble in the polyester
styrene solution. Insoluble materials have restricted application.
A selection of the halogenated compounds currently available is given
TABLE 3.1
HALOGENATED ADDITIVES
Additive % Physical form Solubility ApprOXimate

Halogen in styrene price
£/tonne
(late 1979)
Tetrabromo vinyl cyc10hexane 74 m.p. 70-77 °C soluble 2500

Hexabromocyc1o dodecane 73 m.p. 180°C insoluble 3000
Penta bromo ethyl benzene 80 m.p.136-138°C soluble 2400
Hexabromobenzene 86 m.p.330°C insoluble
Tetrabromoxylene 75 m.p.250°C slightly
soluble 1500
Decabromodiphenyl oxide 83 m.p. 304°C insoluble 2700
Octabromodiphenyloxide 80 m.p. 75~125°C soluble 2400
Pentabromodiphenyl oxide 70 viscous liquid soluble 2300
Pentabromotoluene 81 m.p.280°C insoluble 2900
Tetrabromo bisphenol A 59 m.p. 180°C insoluble 1300
Tetrabromo phthalic anhydride" 67 m.p. 270°C soluble 2000
Chlorinated paraffin wax up to 70 viscous liquids soluble 500
" Used here as an additive.

in Table 3.1. As can be seen, chlorinated paraffin wax is, by far, the
cheapest source of halogen.
3.7.5. Additives Containing Phosphorus or Phosphorus and Halogen

Clear fire resistant polyester formulations normally contain both halogen
and phosphorus compounds. Some or all of the halogen may be built
into the polyester molecule as already described in Section 3.7.3. The
phosphorus is almost always in an additive, which may also contain
further halogen. There are a large number of phosphorus containing
additives available. 24 A selection of these is given in Table 3.2.
TABLE 3.2
PHOSPHORUS BASED FIRE RETARDANT ADDITIVES (ALL LIQUIDS)
AddItive % Phosphorus % Halogen

Tris (2-chloroethyl) phosphate 10·85 37·26
Tris (1,3 dichloropropyl) phosphate 7·2 49
Tris (monochlorpropyl) phosphate 9-4 32
Triethyl phosphate 17
Triphenyl phosphate 10
It should be noted that phosphorus compounds having a significant

P-OH content are unsuitable for use in polyester systems promoted
with cobalt salts as interaction occurs and the cobalt is deactivated.
3.7.6. Alumina Trihydrate

This filler has two main advantages over other fire retarding additives.
It is cheap (currently from about £180 to £220/tonne, depending on particle
size and quality) and it also reduces smoke emission.
In a fire situation, the trihydrate releases 34·5 % of its weight as water
vapour via a strongly endothermic reaction. The water dilutes any com-
bustible gases present, while the endothermic reaction removes heat from
the burning polymer .
Alumina trihydrate is used with both general purpose and halogenated
polyesters and with other additives (e.g. antimony trioxide or trichloroethyl
phosphate). Relatively high loadings of the filler (40-60 % of the compo-
sition) are required to produce adequate fire resistance when it is the
sole fire retarding component. Published data 2 7 indicates that a general
purpose polyester formulation containing 60 % by weight of alumina
trihydrate would have an oxygen index of about 38 %. This compares
76 T. HUNT
with an index of about 20 % for a general purpose polyester without

any additive.
Not surprisingly, polyesters with these high loadings of filler are more
difficult to handle than unfilled resins, and laminators are less than
enthusiastic about using them. It has been shown 28 that silane treated
alumina trihydrate gives much lower viscosities when dispersed in polyester
resins than those obtained with the uncoated filler at the same level
of addition. Possibly use of the coated alumina will increase the accept-
ability of this type of formulation.
3.7.7. Molybdenum Compounds as Fire Retardants and Smoke Suppressants

Molybdenum compounds, in particular molybdenum trioxide and
ammonium dimolybdate, have been found to be effective flame retard-
ants for halogen containing polyester formulations. While less effective
on its own, relative to antimony trioxide, molybdenum trioxide can be
used to replace up to 50 %by weight of the latter without loss of perform-
ance. Whereas antimony trioxide increases smoke generation, molyb-
denum compounds reduce it substantially (up to 50 %). They are believed
to function as char promoters in the solid phase and possibly catalyse
the evolution of Hel and HBr from the halogenated components. Ash
from burnt molybdenum containing formulations contains over 90 %
of the molybdenum added. 29 . 30
Molybdenum trioxide costs approximately twice as much as antimony
trioxide.
3.8. MOULDING COMPOUNDS
3.8.1. General Description

The main development activity in this growing field of application of
unsaturated polyesters has concentrated on improving two properties;
shrinkage and maturation rate. The second property applies mainly to
sheet moulding compounds (SMC), and before going any further it would
be useful to explain what moulding compounds actually are.
There are two types: dough moulding compounds (DMC) and SMC.
Both are designed for hot press moulding.
In its simplest form, DMC consists of resin, chopped glass, mineral
filler, mould release agent, and catalyst. These are thoroughly mixed
to a dough-like consistency in a Z-blade mixer, the glass being added
last, and care being taken to avoid breaking down the fibres. The catalyst
is a heat activated organic peroxide, with decomposition temperature

chosen to suit the moulding temperature to be used. At ambient tempera-
ture the moulding compound has a shelf life of several weeks. A simple
DMC formulation is shown in Table 3.3.
SMC is made by dispensing premixed resin, fillers, maturation agent,
mould release agent and catalyst onto two moving sheets of polythene
TABLE 3.3
A TYPICAL DMC FORMULATION
Component Parts by weight
Polyester resin 25
Filler (calcium carbonate) 55
Zinc stearate (lubricant) 1·5
t-Butyl perbenzoate 0·25
Chopped glass (!in) (6·3mm) 20
film, the lower of which also receives chopped glass rovings or glass
mat. The two coated sheets are brought together, to blend the coatings
and glass, by passing between two rollers. The compacted sheet is then
wound into a roll and set aside to thicken. A simple SMC formulation
is shown in Table 3.4.
The function of the maturation agent in SMC is to thicken up the
mix so that the sheet compound becomes tack free and stiff enough
to handle. The polythene can then be stripped off when the moulding
operation is ready to start.
The preferred maturation agent is magnesium oxide, although a variety
of suitable inorganic thickening agents have been described. 31 Other
TABLE 3.4
A TYPICAL SMC FORMULATION
Component Parts by »'eighl
Polyester reSIn 28
Precipitated calcium carbonate 42
Chopped glass roving <!-2in) (l2'5-50mm) 30
Zinc stearate (lubricant) 1-4
Magnesium hydroxide (thickening agent) 0·7
I-Butyl perbenzoate 0·28
78 T. HUNT
favoured agents are magnesium hydroxide and calcium hydroxide. The

mechanism of thickening has received considerable attention,32,33 it is
believed to be salt formation with the acid end groups of the polyester.
2RCOOH + MgO -+ RCOOMgOOCR + H 2 0
Where R = polyester chain.
If the maturation rate is too fast complete impregnation of the glass
during SMC production is made more difficult. Slow maturation causes
storage problems. Normally SMC should become stiff enough to handle
24-48 h after mixing, when stored at normal ambient temperatures.
Maturation rate is affected by many factors. Resin molecular weight,
acid value, moisture content, and free glycol content can all have a
significant influence. Fillers, r-elease agents, and other minor additives
must also be carefully selected. Polyesters intended for use in SMC are,
therefore, made to extremely tight acid value and molecular weight speci-
fications. Care is taken to control levels of water and free glycol in
the products, in order to produce a compound with a consistent maturation
rate.
3.8.2. Shrinkage Control

Moulding compounds described so far in this section suffer from one
major defect, i.e. shrinkage. All polyesters shrink during use, typically
by about 8 % by volume, but this is not a serious problem in hand
lay-up operations, etc. However, when making complex articles to fine
tolerances with a good surface finish, a moulding material that shrinks
is unacceptable and shrinkage control additives must be incorporated.
These additives are thermoplastics which are soluble in the styrene
of the polyester resin and may be soluble or partially soluble in the
unpolymerised polyester/styrene solution. There effectiveness as shrinkage
controllers depends on their insolubility in the cured resin.
A wide range of such thermoplastics have been patented. These include
polymers or copolymers from one or more C 1-C 4 alkyl esters of acrylic
or methacrylic acid,34 cellulose acetate butyrate and/or cellose acetate
propionate,34 polyvinyl acetate 35 polycaprolactones,36 and saturated
liquid polyesters, e.g. polypropylene adipate, plus sufficient of a thermo-
plastic polymer capable of being plasticised by the liquid polyester to
prevent exudation of it during the moulding operation. 37
In general, a highly reactive polyester (i.e. one made using a high
proportion of maleic anhydride) is required in formulations including
a shrinkage controlling thermoplastic. In some of the patents just
described, the preferred polyester is the reaction product of maleic

anhydride and propylene glycol.
The mechanism of shrinkage control has been investigated. 38 •39 The
following sequence of events is believed to occur. Once polymerisation
at the moulding temperature has got under way, the thermoplastic, if
it was in solution, begins to separate. In some systems, the thermoplastic
may already be present as a well dispersed second phase in the initial
moulding compound. In either case, the droplets of thermoplastic, which
also contain some monomer and possibly polyester as well, expand therm-
ally and thus comp(:llsate for the polymerisation shrinkage occurring
in the continuous phase.
Microscopic voids are generally but not always found in the dispersed
thermoplastic phase when cured, unfilled samples are examined. These
are said to result from either migration of monomer from the dispersed
droplets or from polymerisation in the droplets with subsequent shrinkage.
The differences in the performance of various thermoplastics as
shrinkage control agents are believed to be related to differences in their
TABLE 3.5
NON-SHRINK DMC FORMULATION
Component Parts by weight
Polyester resin" 35·0

Polypropylene adipate (Hexaplas PPA)b 7·0
Polyvinyl chloride (Breon 121Y 1·5
Benzoyl peroxide paste 1·0
Internal lubricant 2·0
Mineral filler 38·5
(±
Chopped glass fibre in) 15·0
" From maleic anhydride/isophthalic acid/phthalic an-

hydride/diethylene glycol/propylene glycol in mole ratio
5: 1: I :O· 5: 7·3. Styrene content 40 % by weight.
b Omission of the polypropylene adipate from the above
formulation gave mouldings with a shrinkage of
0·0035 in/in. These showed considerable surface ripple
and inferior gloss, thicker sections revealed extensive
internal voids.
C The polyvmyl chloride is plasticised by the poly-
propylene adipate.
A moulding tested to British Standard 2782, 1958 showed
a shrinkage of 0·001 in/in. The surface of the moulding
was free from surface ripple and had a high gloss.
80 T. HUNT
thermal coefficients of expansion, glass transition temperatures and polari-

ties. In the case of thermoplastics, which are soluble in the unpolymerised
polyester styrene solution, precipitation during the cure can also contribute
an increase in volume.
A typical 'nil shrinkage' DMC formulation is shown in Table 3.5.
This example is taken from Ref. 37.
Shrinkage control can be adversely affected when a maturation agent
such as magnesium oxide is incorporated in some systems. This is said
to occur with the thermoplastic resins described in Ref. 34. Low shrinkage
systems of this type which can be chemically thickened are described
in a later patent. 40 The preferred thermoplastic is a copolymer with
pendant acid groups, e.g. from methyl methacrylate, ethyl acrylate and
acrylic acid.
3.8.3. Toughened DMC and SMC

Moulding compounds which give moulded products with improved impact
strength and resistance to internal cracking have been obtained by incor-
porating a butadiene acrylonitrile copolymer with terminal and pendant
vinyl groupS.41 DMC containing 5-12·5 parts copolymer per 100 parts
polyester and, SMC containing 5-10 parts per 100 parts polyester plus
shrinkage control agent, were investigated. The copolymer forms discrete
particles in the cured compound producing a structure very similar to
that of high impact polystyrene.
3.9. ESTERCRETE®
This polymer cement comprises a polyester resin, Portland cement and

a free radical forming catalyst which is insoluble in polyester resin but
readily soluble in water. 42 It is cured by the addition of water (c. 8 %
by weight).
The preferred catalyst is ammonium persulphate, and the cement is a
water repellent stearic acid-coated grade which disperses readily in the
resin. On addition of water to the polymer cement, the catalyst dissolves,
and calcium hydroxide is released by reaction of water with the cement.
In turn, this causes decomposition of the catalyst to produce free radicals
which initiate addition polymerisation of styrene with the polyester resin.
The overall effect is a gradual transition from the liquid cement to a
solid with a low rate of heat evolution. Consequently, most of the shrinkage
® Estercrete is a registered trade name of the Cement Marketmg Co. Ltd.

strains created are relieved by plastic flow. Adhesion to normal concrete

is extremely good.
Mlllor modifying additives include finely ground dicarboxylic acids.
These improve the shelf life by neutralising any free lime which develops
in the cement as a result of water in the resin. Other additives can
reduce the settlement of the cement on storage and help redisper-
sion. 43 .44.45 The basic formulation and typical properties are shown in
Table 3.6.
TABLE 3.6
ESTERCRETE® FORMULATION AND PROPERTIES
BasIc formulatIOn Parts by l'mght
Polyester resin 60
Portland cement 40
Catalyst 2
Stabilising acid 0·25
Anti-sedlmenting agent 0·2
Water to be added for cure 9
Proper lies T}plcal results
Specific gravity 1-45

Gel time at 20°C with the specified water addition 6Q-90min
Set time at 20°C 3-5h
Storage life at 20°C 3 months
Appearance dark grey lIquid
The polymer cement is used in combination with aggregates and sand.

These formulations have compressive strengths after 24 h cure of the
same order as those of Portland cement mixes after 28 days. Chemical
resistance is superior to that of normal concrete and the polymer cement
has been used successfully for factory floors, e.g. in breweries, where
the traditional product had proved to be unsatisfactory. Other applications
include airstrips, road surfacing-particularly on bridges and other areas
where a fast hardening topping is required to reduce traffic dislocation,
and mortars for pointing and surfacing.
3.10. POLYESTER FOAM
A novel foaming system patented by Farbenfabriken Bayer AG 46 is

being used to make wall panels, prefabricated bathroom walls, etc. 47
92 T. HUNT
The foaming agent is a carbonic acid ester anhydride which is decomposed

at normal ambient temperatures to the corresponding carboxylic acid
ester and CO 2 in the presence of certain metal salts, e.g. cobalt
naphthenate.
The carbonic acid ester anhydrides are prepared by reacting the appro-
priate carboxylic acid sodium salt with a chlorocarbonic acid ester. A
preferred foaming agent is isophthalic acid bis (carbonic acid methyl
ester anhydride). This is prepared by the reaction shown below:
COONa COOCOOMe
©L + 2CICOOMe--> (18)
COONa ©LCOOCOOMe
When the dian hydride is treated with one of the metal salts listed in
the patent,46 decomposition takes place according to the following
equation.
©l ©l
COOCOOMe COOMe
+2CO, (19)
-->
COOCOOMe COOMe
If the carbonic acid ester anhydride is derived from an unsaturated

acid such as acrylic acid, the acrylic ester produced by the foaming
reaction will be copolymerised with the polyester. This could be an advan-
tage, as no residual material will be left to contaminate the foam.
To prepare a foam, the polyester resin which already contains the
metal salt is mixed with the foaming agent and a suitable peroxide catalyst.
In the building applications already mentioned, the above mix is poured
into a mould which has been partially filled with special lightweight
aggregate such as expanded forms of glass or slate. When the resin
has penetrated the filler, further aggregate is added to fill the mould,
which is then closed before foaming starts. The same type of operation
can be used to make sandwich panels using GRP, plasterboard, etc.,
as the skin material.
3.11. FORMULATIONS GIVING REDUCED STYRENE

EMISSION
When polyester resins are applied to open moulds by hand lay-up or

spray-up techniques some styrene evaporates. Where ventilation is poor,
high atmospheric levels of styrene can build up. To date, there is no
evidence to suggest that long term exposure to styrene vapour is a health
hazard. Nevertheless, a threshold limit value (TLV) of 100 ppm for styrene
vapour in the atmosphere is now in force. 48
To ensure that workshop levels remain below the TLV good ventilation
is essential. Most resin suppliers now market so called 'environmental'
or 'low emission' grades which give rise to significantly less styrene vapour
than unmodified equivalents. However, it should be emphasised that
where ventilation is poor, an environmental resin is unlikely to solve
a problem of high styrene levels. Efficient ventilation must be the first
priority.49 The environmental grades are a means of reducing levels of
styrene still further.
Early environmental resins contained a small amount of paraffin wax,
as an evaporation inhibiting additive. This tends to form a film at the
resin surface during application and fully separates at gelation. The waxy
layer can cause delamination problems, and additives preferred in later
formulations include a combination of a high and a low melting point
wax,50 various polymers, e.g. polydibutyl fumarate, polybutyl acrylate, 51
and a combination of an evaporation inhibitor such as paraffin wax
and an 'attachment promoter'. 52 This latter patent is particularly interest-
ing. The attachment promoter can be an acyclic, hydrophobic ether or
ester with at least two hydrocarbon groups, having at least one double
bond in each, or an unsaturated isoprenoid compound, or an ether or
ester of such a compound. Examples are linseed oil, dipentene and tri-
methylol propane diallyl ether. The preferred wax content is 0·05-0·5 %;
the preferred attachment promoter addition is 0'1-2%. Examples give
interlaminar strengths of compositions with and without the attachment
promoter.
Other means of reducing styrene emission include lowering the styrene
content of the resin and replacing some or all of the styrene with a
less volatile monomer. Both these approaches have disadvantages.
Reducing the styrene content would also involve lowering resin molecular
weight to achieve a similar viscosity product which, in turn, could lead
to inferior cured resin properties.
Styrene is considerably cheaper than any of the higher boiling point
84 T. HUNT
monomers which might be considered as replacements, e.g. vinyl toluene,

methacrylates, etc. It is, therefore, unlikely that a polyester dissolved
in one of these expensive monomers would be economically acceptable,
except for very specialised applications. Moreover, the safety in use of
alternative monomers, even at lower concentrations than styrene, would
have to be established.
If permitted atmospheric styrene levels are drastically reduced in the
future, the most likely outcome is a changeover to closed mould techniques,
such as resin injection.
3.12. FUTURE TREN DS
Future development will have to keep pace with changes in the cost
and availability of basic raw materials. At present polyesters are main-
taining their competitive position in the market place. Nevertheless, close
attention will have to be paid to resin costs to ensure that this position
is not eroded.
Rising petrol prices and the efforts of the motor industry to produce
cars with reduced fuel consumption should lead to increased use of poly-
esters and particularly SMC in car body construction. This will increase
the demand for resins with more consistent and easily controlled matura-
tion rates, and thus, may necessitate the development of new methods
of maturation. Possibilities in this area include the highly crystalline
systems already discussed,15 or perhaps the use of an external gelling
agent such as sodium stearate. 53
High impact strength, rubber modified SMC and DMC have been
briefly mentioned in this chapter (3.8.3). This should be a fruitful area
for further development, with automotive applications, particularly, in
mind.
The toughening of polyester formulations in general may become neces-
sary as they could be faced with competition from glass reinforced thermo-
plastics. Possible ways of building improved impact properties into the
polyester molecule include the use of butadiene polymers and copolymers
with hydroxy or carboxyl functionality.
Legislation specifying resins with reduced tendency to burn for a variety
of end products is likely and long overdue. This will almost certainly
include a requirement for resin formulations which produce low levels
of smoke in a fire.
As mentioned earlier, a drastic reduction in the TL V for styrene vapour
in the atmosphere would probably lead to an increase in the use of

closed moulds. Fortunately the knowledge and techniques for meeting
this eventuality and many others are already available. These should
provide a firm base from which further advances can be made.
REFERENCES
I. BOENIG, H. V. Unsaturated Polyesters: Structure and Properties, 1964, Elsevier
Publishing Company, Amsterdam.
2. Amoco Chemicals Corporation. How ingredients influence unsaturated
polyester propertIes, Bulletin IP-70, 1979.
3. British Patent No. 1000534 issut:d to Chemische Werke Hiils AG, 1965.
4. US Patent No. 3254060 issued to Allied Chemical Corporation, 1966.
5. US Patent No. 3355434 issued to Jefferson Chemical Company, 1967.
6. British Patent No. 1101247 issued to Scott Bader & Company Ltd, 1968.
7. British Patent No. 1375038 issued to Produits Chimiques Pechiney Saint
Gobain, 1974.
8. ANON. Hydrocarbon Process., December 1977, pp.1I5-16.
9. ALEXANDER, J. British Plastics Federation Reinforced Plastics Congress, 39,
1976, Paper A.
10. SMITH, P. I., MCGARY, e. W. and COMSTOCK, R. I. 22nd SPI Reinforced
Plastics/ComposItes Conf., 1967, Paper I-e.
II. NELSON, D. L. 34th SPI Reinforced Plastics/Composites ConI, 1979,
Paper I-G.
12. DAVIS, J. H. and HILLMAN, S. L. 26th SPI Reinforced Plastics/Composites
Conj., 1971, Paper 12-e.
13. EDWARDS, H. R. 34th SPI Reinforced Plastics/Composites Conf., 1979, Paper
4-D.
14. British Plastics Federation, Publication No: 220/1, 1978.
15. British Patent No. 1319243 issued to Scott Bader Company Ltd, 1973.
16. GOTT, S. L., SUGGS, J. L. and BLOUNT, W. W. 28th SPI Reinforced
Plastics/Composites Conj., 1973, Paper 19-e.
17. Technical data available from Eastman Chemical.
18. British Patent No. 1318517 issued to Scott Bader Company Ltd, 1973.
19. British Patent No. 1190116 issued to Koppers Company, Inc., 1970.
20. LARSEN, E. R. and WEAVER, W. e. 28th SPI Reinforced Plastics/Composites
Conj., 1973, Paper 2-A.
21. German Patent No. 2819446 issued to Freeman Chemical Corporation, 1979.
22. NAMETZ, R. e. Ind. Eng. Chem., 1967,59,99.
23. BELL, K. M. and CAESAR, H. J. British Plastics Federation Reinforced Plastics
Congress, 7, 1970, Paper 7.
24. Flame Retardancy of Polymeric Materials eds. Kuryla, W. e. and Papa, A. J.,
1973, Marcel Dekker, New York.
25. French Patent No. 1336751 issued to Pechiney Ugine Kuhlmann, 1963.
26. GOINS, O. K., ATWELL, R. W., NAMETZ, R. e. and CHANDIK, B. 29th SPI
Reinforced Plastics/Composites ConI, 1974, Paper 23-B.
86 T. HUNT
27. BONSIGNORE, P. V. and MANHART, J. H. 29th SPI Remforced

Plastics/Composites Con/., 1974, Paper 23-C.
28. ATKINS, K. E., GENTRY, R. R., GANDY, R. c., BERGER, S. E. and SCHWARTZ,
E. G. 32nd SPI Reinforced Plastics/Composites Con/., 1977, Paper 4-D.
29. SKINNER, G. A. Fire and Materials, 1976, 1, pp.154-9.
30. MOORE, F. W. and CHURCH, D. A. Paper presented at the International
Symposium on flammability and flame retardants, Toronto, Ontario, 1976.
31. FEKETE, F. 27th SPI Reinforced Plastics/Composites Conf., 1972, Paper 12-D.
32. WARNER, N. K. 28th SPI Reinforced Plastics/Composites Conj., 1973, Paper
19-E.
33. BURNS, R., GHANDI, K. S., HANKIN, A. G. and LYNSKEY, B. M. Plastics and
Polymers, December 1975,43, pp. 228-35.
34. British Patent No. 1201087 issued to Rohm & Haas Company, 1970.
35. German Patent No. 2104575 issued to Union Carbide Corporation, 1971.
36. US Patent No. 3549586 issued to Union Carbide Corporation, 1970.
37. British Patent No. 1098132 issued to The Distillers Company Ltd, 1968.
38. KROEKEL, C. H. and BARTKUS, E. J. 23rd SPI Reinforced Plastics/Composites
Conf., 1968, Paper 18-E.
39. ATKINS, K. E., KOLESKE, J. V., SMITH, P. L., WALTER, E. R. and MATTHEWS,
V. E. 31st SPI Reinforced Plastics/Composites Conj., 1976, Paper 2-E.
40. British Patent No. 1276198 issued to Rohm & Haas Company, 1972.
41. MCGARRY, F. J., ROWE, E. H. and RIEW, C. K. 32nd SPI Reinforced
Plastics/Composites Conj., 1977, Paper 16-C.
42. British Patent No. 1065053 issued to Cement Marketing Company Ltd, 1967.
43. British Patent No. 1091325 issued to The Distillers Company Ltd, 1967.
44. British Patent No. 1157292 issued to British Resin, 1969.
45. British Patent No. 1292333 issued to BP Chemicals Ltd, 1972.
46. British Patent No. 1160476 issued to Farbenfabriken Bayer AG, 1969.
47. ANON. Mod. Plastlcs InternatIOnal, December 1973, pp. 20-2.
48. Guidance Note EH 15/17-TLV, Health & Safety Executive, HM Factory
Inspectorate.
49: EVANS, P. D. Remforced Plastics, 12, 1976, 364.
50. Europatent No. 941 issued to BASF AG, 1979.
51. Belgian Patent No. 859773 issued to BIP Ltd, 1978.
52. Norwegian Patent No. 783398 issued to AB Syntes, 1979.
53. British Patent No. 1440345 issued to BP Chemicals Ltd, 1976.
Chapter 4
PHENOL-ARALKYL AND RELATED POLYMERS
GL YN I. HARRIS
Advanced Resins Ltd, Cardiff, UK
SUMMARY
The synthesis of polymers by means of the Friedel-Crajts condensation

reaction dates from 1881. More recently, a wide variety oj aromatic
compounds have been polymerised by means of catalysts such as aluminium
chloride and stannic chloride. The condensation reactions involve the
elimination oj small molecules, usually hydrogen halide or methanol.
The polymerisation reactions can be made to produce crosslinked net-
works. The curing process, however, is liable to produce voids which impair
the mechanical properties oj the resultant products. The more promising
materials are derived from phenol and (xlx.'-dimethoxy-para-xylene; in this
case, the phenolic component facilitates crosslinking by either hexa-
methylene tetramine, or a diepoxide.
These resins have now been developed commercially. They are used
in mouldings and laminates because of their good chemical resistance,
electrical properties and high-temperature performance. They also have
some potential as coatmgs. Blending with conventional phenol-formalde-
hyde resins improves the properties oj the latter.
4.1. INTRODUCTION
The possibility of making polymers by a Friedel-Crafts condensation

dates back to 1881 1 when polymeric materials were reported to be formed
by the action of aluminium chloride on benzyl chloride. These polybenzyl
polymers were of periodic academic interest 2 - 5 up until the 1950s when
87
88 GL YN I. HARRIS
they became the subject of more intense examination as part of the

search for polymers to meet increasingly demanding industrial and military
applications. In this chapter the progressive polymer development from
polybenzyl through the so called Friedel-Crafts resins to the now com-
mercially established phenol-aralkyl resins is reviewed.
4.2. POLYBENZYL POLYMERS
The first serious examination of the polymerisation of benzyl chloride

with small quantities of aluminium, ferric and stannic chlorides was
reported by Jacobson 2 in 1932. Aluminium trichloride yielded mainly
an insoluble hydrocarbon and a small amount of soluble resin having
the same empirical formula, namely (C 7 H 6 )n.
@- CH 2Cl
Fnedel-Crafts catalyst
heat
•
-f@-c H2
t
n
+ nHCl
FIG. 4.1. Preparation of polybenzyl.
The proportions of these were reversed when ferric chloride was employed,
while stannic chloride yielded entirely the soluble polymers. The latter
had molecular weights in the range 1260 to 2250 and were found to
be relatively resistant to attack by oxidising agents. The properties of
the soluble polymers were accounted for by assuming that they are hydro-
carbon chains containing the group -C 6 H 4 -CH 2- as the structural
unit. The investigations of Shriner and Berger 5 led them to the view
that polybenzyl prepared from the polymerisation of both benzyl halides
and benzyl alcohol were predominantly linear para-substituted polymers.
Detailed studies by Haas and co-workers 6 using infrared, X-ray and
chemical methods were interpreted in favour of a more complex structure.
They proposed that these polymers contain a nucleus of almost completely
substituted phenyl rings and have a periphery of pendant benzyl groups.
On the basis of the results of oxidative and thermal degradation, a highly
branched, globular structure was proposed by Parker. 7 Similar findings
were also reported by Ellis and White,8 while Ellis and co-workers 9
attributed the fluorescence of the polybenzyl formed by treating benzyl-
chloride in the presence of air at room temperature to benzyl substituted
PHENOL-ARALKYL AND RELATED POLYMERS 89
9-phenyl or 9:10-diphenylanthracene. The only recently reported lO prep-

aration of a linear polybenzyl has been that of treating polybenzyl chloride
with aluminium chloride at low temperatures (-100°C).
The oxidutive degrudution mechuni9m for l'olybenzyl was established
by Conley, II but more important, Parker l2 at the Royal Aircraft Establish-
ment (RAE) established that a polybenzyl, polymerised with the aid of
aluminium chloride, was more thermally sta~le than the best phenolic
resin. By 1964 investigators 13 were, however, resigned to the view that
these polymers in an unmodified state were of little practical use, the
thermoplastic form being extremely weak and brittle, while the insoluble
infusible forms are impossible to process.
4.3. POLYDIPHENYL ETHER RESINS
One of the alternative classes of monomers to benzyl halides, investigated

particularly in the USA, was the bis-chloromethylated diphenyl ethers.
Doebens 14 and later Geyer and co-workers 15 prepared polyaromatic ethers
by heating monomers such as 4,4'-di-(chloromethyl) diphenylether with
selected metallic halides. The condensation reaction proceeds with forma-
tion of methylene bridges and the concurrent elimination of hydrogen
chloride as illustrated in Fig. 4.2. The crosslinking as well as the further
polymerisation can occur through continued reaction of the residual halo-
methyl groups with other diphenyl ether moieties. Since the reaction
to form crosslinking methylene bridges is the same as that involved in
polymer chain extension, the final condensation polymer is, in practice,
nearly always a highly crosslinked resinous polymer insoluble in most
zncl,l heat
C!CH'~~H'C! + HC!
CH2-@--O~H2CJ
FIG. 4.2. Condensation of 4,4'-dl-(chloromethyl)diphenylether.
90 GL YN I. HARRIS
polar and non-polar solvents. Only when there is a deficiency of halo methyl
groups or reactive sites, are soluble polymeric materials obtained. The
practical value of these polymers is, however, limited by the unacceptability
of the hydrogen chloride released during the final cure of these poly-
mers. This problem was resolved by Spreng ling and co-workers l6 at
Westinghouse Electric Corporation by using methoxymethyl- rather than
chloromethyl-substituted diphenylether. They prepared monomeric com-
pounds of the formula:
@-O-@
H
where Y is a lower alkyl substituent, generally methyl, and x has an
average value of about 0·8 to about 3·0.
Polymers derived from these monomers were reported to have the
general formula shown in Fig. 4.3, where n is an integer in the range
one to nine. The reaction is catalysed by Friedel-Crafts catalysts such
as aluminium trichloride, zinc chloride, boron trifluoride, etc.; solids
FIG. 4.3. Condensation of methoxymethyldiphenylether.
such as silica, diatomaceous earths, bentonites, some metals in the form

of their organic soluble chelates, notably ferric acetyl acetonate and soluble
acids such as p-toluene sulphonic acid. The change from using chloro-
methyl- to methoxymethyl-substituted diphenyl ethers is also reported
to decrease chain branching and consequently reduce the tendency for
undesirable gel formation during an early stage in the polymer
conversion. These observations were exploited by the Westinghouse
Corporation for the manufacture of the first commercially promoted
Friedel-Crafts resin. Chloromethyldiphenyl ether and a small amount
of diphenyl ether are condensed to give a low molecular weight soluble
pre-polymer which is chlorine-free. This is, subsequently, partially con-
densed with a methoxymethyldiphenyl ether to yield a soluble and fusible
resinous composition which readily cures in the presence of suspended
silica to give a hard, tough resin and methanol as the sole by-product.
This new resin 17 -19 was launched by the Westinghouse Electric
Corporation under the tradename 'DoryI' in about 1961 in the form

of varnishes and prepregs. The new materials were claimed to be capable
of operating under Class 'H' (ISO°C) electrical insulating conditions and
to have several other outstanding characteristics, namely:
(1) bond strength maintained over 3000 h at 250°C;
(2) electric strength retention for 10 years at 250°C in form of insula-
tion tube;
(3) imperviousness to chemical and solvent attack;
(4) after initial cure, the material maintains a hornlike structure
even after extreme and abrupt temperature changes, and
(5) considerably lower cost than previous Class 'H' materials, long
shelf life and ease of application.
Despite their many attractive properties, the 'Doryl' resins have, in the
years since their launching, enjoyed only limited commercial success.
This has been, at least partially, due to the curing catalyst being hetero-
geneous and, so, requiring it to be stirred into the system immediately
before use.
4.4. FRIEDEL-CRAFTS RESINS BASED ON

(1.(1.'- DICHLORO-p-XYLENE
In the early 1960s efforts were made in the UK to improve the processa-
bility and physical properties of polybenzyl polymers by crosslinking
them with arl-dichloro-p-xylene in the presence of a mild Friedel-Crafts
catalyst such as stannic chloride. This approach proved relatively successful
and Phillips 20 reported the preparation of asbestos felt reinforced compo-
sites with these resins. The laminates were found to be very similar in
appearance to conventional Durestos® boards with phenolic resins. They
differed in two respects.
(1) Resistance to concentrated alkali Even after twelve weeks immer-
sion in 40 %caustic soda the specimens were unchanged, whereas
the phenolic laminates were swollen and weak.
(2) Strength retention after heat ageing There was no substantial loss in
room temperature flexural strength after exposure to 240°C for
1000 h, whereas the Durestos grade had virtually no strength after
250 h exposure.
® Durestos is a registered tradename of Turner Bros. Asbestos Co.
92 GL YN I. HARRIS
The implications of the above investigation proved far reaching both

from the synthetic viewpoint and from the performance appraisal and
expectations for these resins. Since phenyl, phenylene and phenoxy groups
are present in a variety of compounds it encouraged the synthesis of a wide
range of Friedel-Crafts resins. This is exemplified in Fig. 4.4 by the
condensation of diphenyl with txlx'-dichloro-p-xylene in the presence of
trace amounts of stannic chloride to give a pre-polymer and hydrogen
chloride. Crosslinking of the pre-polymer to a hard resinous product was
readily achieved by mixing with further quantities of r:t.r:t.' -dichloro-p-xylene
and heating.
@-@ + ClCH2~H2Cl heat
~H'~H'i. +2nHCI
FIG. 4.4. Condensation of diphenyl and aa' -dichloro-p-xylene.
The reactivity of a wide range of organic compounds was examined

by Phillips 20 who found they fell into three categories summarised in
Table 4.1. The first class included dihydric phenols which give a very
strongly exothermic reaction with r:t.r:t.' -dichloro-p-xylene on heating
together in the absence of a solvent. For controlled reactions the condensa-
tion of the co-reactants in the first two categories was found to be best
undertaken in a 1,2-dichloroethane solution.
The reaction kinetics and structure of the condensation products formed
from the reaction of r:t.r:t.' -dichloro-p-xylene with benzene 21 .22 and diphenyl-
methane 23 .24 in 1,2-dichloroethane were examined by Grassie and
Meldrum. They found in the first reaction system that the rate constant
for the reaction of the second chloromethyl group in r:t.r:t.' -dichloro-p-xylene
was twenty times as great as the first rate constant. This was offered
as an explanation to account for the observation that the high molecular
weight polymers, isolated by gel Permeation chromatography, have no
detectable chloromethyl groups even though they must undoubtedly exist
in low concentrations as intermediates between successive hydrocarbons.
It was established in the condensation of (X(X' -dichloro-p-xylene (Pi) with
benzene ( Po) that when the concentration of first condensation products,
namely p-benzylchloromethylbenzene (P 2) and dibenzylbenzene (P 3), be-
come large then more complex products begin to appear. They are formed
TABLE 4.1
REACTIVITIES OF VARIOUS AROMATIC COMPOUNDS TO IXIX'mCHLORO-p-XYLENE
Qualitative guide to chemical reactIVity Co-reactant
I. Reacts without catalyst Resorcinol

Catechol
Diphenyl ether
Diphenylene oxide
2. Reacts with aid of stannic chloride catalyst Naphthalene
Fluorene
Anthracene
Mesltylene
Phenanthrene
0, m and p Terphenyl
Triphenylene
Diphenylmethane
Triphenylmethane
Diphenylsulphone
3. Unreactive with stannic chloride catalyst present Monochlorobenzene
0- Dichloro benzene
Nitrobenzene
Benzophenone
by two routes. The first occurs when the concentrations of PI and Po

are still quite high, and products will be formed predominantly by the
reaction of PI with an earlier product, followed by reaction of benzene
(Po) with the free chloromethyl group as in the sequence shown in Fig.
4.5.
The second route becomes important when an appreciable concentration
of products accumulate in the reaction mixture. The pendant chloromethyl
Po Po P,
PI~ P2~ P3~ P6
P4
I '\
1
PI Ps
Po ipo
P6 P6
FIG. 4.5. Condensation polymers from IXIX'-dlchloro-p-xylene (PI) and benzene
(Po)·
94 GL YN I. HARRIS
groups are then considered to react increasingly with aromatic nuclei

to result, for example, in the formation of seven nuclei products (P7)
by reaction of P4 and P 5 with P 3 or of P 2 with P 6. Thus the reaction
rapidly becomes highly complex with a proliferation of isomers at each
level of molecular weight.
The parallel studies 23 •24 on the condensation of aa' -dichloro-p-xylene
with diphenylmethane established, as might be expected, that there were
many common features. In each case there are two kinds of products.
Hydrocarbons are again obtained in much larger concentrations than
chloromethylated products which can likewise be attributed to the second
chloromethyl group in the aa'-dichloro-p-xylene molecule, being very
much more reactive than the first. The frequency of occurrence of chloro-
methyl groups in product molecules was, however, observed to increase
from about one in every 24 aromatic nuclei to about one in six nuclei
when the initial ratio of aa'-dichloro-p-xylene to diphenylmethane is
changed from 1:2 to 4:l. From this it was concluded that the extent
of reaction required for gelation would be quite sensitive to the relative
initial proportion of reactants and this has been confirmed in practice.
4.5. ORGANOMETALLIC RESINS
The only serious attempt to promote the use of the Friedel-Crafts resins
based on aa' -dichloro-p-xylene was undertaken by the RAE for military
applications. Phillips 13.2o successfully extended his condensation studies
with aa' -dichloro-p-xylene to include some phenyl-substituted organo-
metallic compounds. OctaphenyJcyclotetrasiloxane and aa' -dichloro-p-
xylene, dissolved in a mixture of o-chlorobenzene and 1,2-dichloroethane,
were found to react immediately the system was heated to reflux and
catalysed with stannic chloride. Likewise triphenylphosphate and aa'-di-
chloro-p-xylene were found to react slowly in an o-dichlorobenzene
solution.
The impregnation of asbestos felts with siloxane-based resin and their
subsequent processing into laminates showed two disadvantages. The
impregnated felts were stiff and rather awkward to handle, moreover,
during the post cure stage there was persistent blistering so that many
laminates were spoiled. By contrast, asbestos felts impregnated with the
phosphate resin were soft and rather sticky to the touch, but it was
observed that the boards made from them never blistered. Accordingly
a number of siloxanejphosphate blends were examined and, with a ratio
{}[ 10~ J gilmUlne!~h{}g~hllte, II firm llnd yet flexible im~regnllted felt W1l9
reported to be obtained. 20 .25 This cured reasonably quickly to give
laminates which did not blister during postcure. The resultant composites
when subject to extensive heat-ageing are characterised by:
(1) higher strength, at room temperature and at temperatures up to
400°C, than commonly found with conventional asbestos rein-
forced resin composites,
(2) retention of an inorganic residue which prevents gross delamina-
tion on prolonged exposure at 300°C or more, and
(3) no catastrophic failure even on exposure to 350 ° and 400 dc.
Infrared studies 25 show that the methylene bridges in the resins are oxidised
to carbonyl groups during ageing at high temperatures and the thermal
stability of the siloxane!phosphate laminate is below that of a conventional
asbestos felt reinforced polysiloxane resin composite. Moore 26 extended
evaluation studies to chemical resistance testing. He found that silica
cloth reinforced boards maintain, at room temperature and 90°C, an
overall strength retention of 77·5 % after 5000 h exposure.
The siloxane!phosphate composites were proposed for use in the field
of guided weapons because of their ability to withstand 5 h exposure
up to 400°C. The commercial use of these laminates has, however, been
inhibited due to handling difficulties and, in particular, to 'sweating'
which is caused by the slow escape of some of the trapped hydrogen
chloride produced during the condensation. This, coupled with the general
processing problems due to the release of corrosive hydrogen chloride
gas during the curing reaction, has prevented any serious commercial
exploitation of these and related resins based on lJ.(l-dichloro-p-xylene.
4.6. FRIEDEL-CRAFTS RESINS BASED ON cxcx'-DIMETHOXY-

p-XYLENE
The processing problems associated with the manufacture and cure of

Friedel-Crafts resins based on aralkyl chlorides encouraged a search
for other classes of monomers. The aralkyl ethers were found, inde-
pendently in 1964 by Harris 27 at Midland Silicones Ltd and Phillips28
at the RAE, to be attractive alternative resin intermediates. The condensa-
tion of IJ.IJ.' -dialkoxy-p-xylenes with aromatic compounds has the twin
advantage over aralkyl chlorides of being a less exothermic reaction and
yielding a relatively innocuous alcohol as a by-product. This type of
96 GL YN I. HARRIS
©© + MeOCH'~H,OMe
heat 1
SnC!.
-fOO--cH,-\Q)-cH+ + 2nMeOH
FIG. 4.6. Condensation of naphthalene and r:x.r:x.'-dimethoxy-p-xylene.
condensation is illustrated in Fig. 4.6 by the condensation of r:x.rl-dimeth-

oxy-p-xylene and naphthalene to yield a pre-polymer and methanol as
the by-product.
Harris 29 investigated the reactivity of a wide range of aromatic, hetero-
cyclic and organometallic compounds with r:x.r:x.'-dimethoxy-p-xylene. The
results of this study are summarised in Table 4.2.
The technique employed was to heat, and subject to slow agitation,
a mixture of r:x.r:x.'-dimethoxy-p-xylene and the co-reactant together with
0·001 moles of stannic chloride/mole of the aralkyl ether. The reactivity
and overall course of the reaction was followed in each case by the
elimination of methanol. The r:x.r:x.' -dimethoxy-p-xylene reacted with all
the aromatic hydrocarbons examined except anthacine and with all the
TABLE 4.2
REACTIVITY TO r:x.r:x.'DIMETHOXy-p-XYLENE
ReactIVe Unreactive
Anisole Anthracene
Benzene Benziminazole
N-benzylaniline Chlorobenzene
Dibenzylether Diphenylurea
Diphenyl N-methylaniline
Diphenylamine Nitrobenzene
Diphenylether Tetra phenylsilane
p-Diphenylbenzene Tetramethyltetraphenyltrisiloxane
Naphthalene Poly(methylphenylsiloxane)
Octaphenylcyclotetrasiloxane
Terphenyl
1,3,5 Triphenylbenzene
Triphenyl phosphate
other aromatic compounds except those containing electron withdrawing

substituents. By contrast, organosilicon compounds containing phenyl
groups were found to be generally unreactive. This is not surprising
in view of the cleavage of silicon-aryl bonds previously reported 3o . 31
during attempts to carry out Friedel-Crafts reactions on methyl phenyl-
silanes. The exception is octaphenyltetrasiloxane in which the silicon-aryl
bonds are probably protected by steric hindrance.
The most promising resins in this group have been found to be those
based on the condensation of cax'-dimethoxy-p-xylene and diphenyl ether.
This is due to the inherent heat stability of the diphenyl ether moiety
and the ease of the condensation reaction between these two intermediates
in the presence of a relatively weak Friedel-Crafts catalyst such as stannic
chloride. The reaction proceeds smoothly, with the loss of methanol
on heating the system to about 165°C. It is advanced to a suitable
application viscosity and the final cure is best achieved by heating after
the addition of a stronger Friedel-Crafts catalyst such as ferric chloride. 32
Harris and Edwards 33 studied the effect of functionality on the physical
properties of these resins by preparing a series with the r:trl-dimethoxy-p-
xylene content ranging from 50 to 70 mole %. The changes in flexural
strength of glasscloth (Marglass® 7T /methacrylato-silane finish) rein-
forced laminates based on these resins on heat ageing for up to 1000h
at 250°C are plotted in Fig. 4.7. All the laminates are characterised
by excellent initial flexural strength at room temperature but low strength
retention at 250°C. The increase in functionality is reflected in a faster
build-up in high-temperature strength and a more rapid decline in the
room temperature strength on heat ageing at 250°C.
The overall assessment of these resins is that they offer good
thermal stability and would be suitable for applications requiring pro-
longed exposure to temperatures up to at least 200°C. The slow build-up
in high-temperature mechanical strength and the need to use a strong
Friedel-Crafts catalyst to achieve a rapid and controlled cure has, un-
fortunately, prevented their widespread commercial acceptance 111
reinforced composites.
Other investigations on the above system have been reported by
Paxton 34 of Associated Electrical Industries Ltd. He developed a solvent-
less varnish which was promoted commercially under the tradename
Caldura®. It was recommended for the impregnation of the windings of
® Marglass is the registered trademark of Marglass Ltd.

® Caldura is the registered trademark of Associated Electrical Industries Ltd.
700,.--------------------------------------, 700r'-----------------------------------, '-0
00
600 600
500
'"'E '"'E
z z
:::;: :::;:
400
;: .s::.
c:
'"
~ Cl
t; r
-<
e Z
::>
.,>< r
Li: ~/30
lL. ::c
i30+ :>
::0
::0
r;;
100
00
00 260 400 660 860 1060 1000
Heat ageing t,me at 250·C (h) Heat ageln9 t,me at 250·C (h)
(a) (b)
FIG. 4.7. Flexural strengths of glassdoth reinforced IXIX'-dimethoxy-p-xylenejdiphenylether resin laminates on exposure at
250°C for up to IOOOh, (a) tested at room temperature, (b) tested at 250°C.
electrical machines designed to operate at temperatures up to 240 DC,

and is also claimed to be compatible with polyimide wire enamels and
to give good hot bond strengths. The cured resin is also reported to
be resistant to aircraft lubricants and hydraulic fluids.
A detailed evaluation of carbon fibre reinforced composites based on
resins prepared from the condensation of act.' -dimethoxy-p-xylene with
toluene,35 xylene,36 terphenyl35 and diphenyl ether 37 was undertaken
by Parker. As was to be expected with matrix resins, which give off
volatiles during cure, the composites tended to have high void contents,38
averaging 4 % and ranging from 1-20 %. All the mechanical properties
were adversely affected by the high void contents. In longitudinal tension,
50-60 %of the fibre strength, and 70 %of the fibre modulus, was realised
in the composite. The transverse tensile strength was low, while the trans-
verse modulus was almost the same as for the resin itself. The inter laminar
shear strength was also noticeably affected by void content. Extrapolation
of the results to zero void contents showed that potentially the shear
strength could approach the values of typical epoxy resin composites.
Very little difference was observed between the diphenyl ether and the
terphenyl resin based composites except that the former were faster curing.
Composites of resins cured with ferric chloride were found to have excellent
retention of strength at elevated temperatures after postcure. When boron
trifluoride was used as the catalyst, the composites gave better interlaminar
shear strength than with ferric chloride, but were poorer in strength
retention at high temperature. From Arrhenius plots it was calculated
that 75 % strength retention lives, for these composites, were one year
at 200°C and 10 years at 150 dc.
The possibility of making unsaturated condensation polymers by the
Friedel-Crafts reaction has been reported by Huck and Pritchard. 39 They
prepared a low molecular weight polymer containing substantial olefinic
unsaturation by heating aa' -dimethoxy-p-xylene and stilbene in the
presence of a Lewis acid catalyst. When this condensation polymer was
mixed with maleic anhydride in the ratio of one mole of the latter to
each mole of retained un saturation and heated with benzoyl peroxide,
a hard intractable resin was obtained. Unfortunately, the high cost of
stilbene would appear to preclude a commercial future for this polymer
even if the physical properties prove attractive.
4.7. PHENOL-ARALKYL RESINS
The polyaromatic resins synthesised from IY.IY.' -dimethoxy-p-xylene, des-

cribed in the last section, proved far more acceptable for processing
100 GL YN I. HARRIS
but, nevertheless, have at least two serious deficiencies. These are, the
relatively low mechanical strength of the reinforced composites at elevated
temperatures, and the slow curing rates which preclude their widespread
use in the form of moulding compounds. To overcome these shortcomings,
Harris 4041 prepared a further series of resins based, essentially, on the
condensation of aralkyl ethers and phenols, so presenting the opportunity
of using hexamine (hexamethylene tetramine) or a di- or poly-epoxide
to be used as the curing agent. A simple example is illustrated in Fig. 4.8,
©MeOCH,--@-cH,OMe
OH
heat 1
SoCI.
1: I 1:l
~H'-@-cH,-tqcH,-<QrH'-©
OH
pre-polymer + 2n + 2MeOH
FIG. 4.8. Condensation of phenol and cxcx'-dimethoxy-p-xylene.
namely the condensation of phenol and r:t.r:t.' -dimethoxy-p-xylene to give

a pre-polymer, which on heating with hexamine quickly cures with the
release of ammonia as a gaseous by-product. This novel class of thermosets
-the phenol-aralkyl resins-was launched commercially in 1970 by
Albright & Wilson Ltd under the tradename Xylok. A detailed review
of the preparation, structure and properties of this important addition
to the class of high performance thermosetting resins represents the subject
matter of the remainder of this chapter.
4.8. PREPARATION AND HEXAMINE CURE OF

PHENOL-ARALKYL RESINS
The condensation reaction between r:t.r:t.'-dimethoxy-p-xylene and a phenol

generally occurs readily in the presence of very low levels of Friedel--Crafts
catalyst at temperatures in the range 120 °-165°C. Stannic chloride is

a very suitable reaction catalyst at an addition of 0·001 moles/mole of
the aralkyl ether. Other catalysts which may be used include ferric, zinc
and cupric chloride as well as sulphuric acid and even some acidic clays.
In a typical condensation, such as that between phenol and rtrt' -dimethoxy-
p-xylene, exemplified in Fig. 4.8, the ratios of reactants are arranged
such that the condensation can be taken to completion. The resultant
pre-polymers have number average molecular weights, as determined by
vapour pressure osmometry in 2-ethoxyethanol, generally in the range
450 to 750. Molecular weight distribution investigations on the same
pre-polymers by gel-permeation chromatography indicate a significant
spread. Typical values are given in the table below.
% Molecular weight range
42 > 1500
35 400-1500
23 <400
Some of the low molecular weight fraction «400) is due to residual

free phenol which is ideally present at a level of 2·5 to 5·0 %. Lower
levels of free phenol, as with phenolic novolacs, are undesirable since
the rate of gelation with hexamine becomes unacceptably long.
The softening points of the pre-polymers formed from the condensation
of phenol and wl-dimethoxy-p-xylene generally fall in the range 70-100 °C
and can be influenced by the catalyst used. This suggests that the catalyst,
as in phenolic novolacs 42 . 43 can probably affect the structure by influencing
the distribution of ortho-ortllO, ortho-para and para-para linkages
(Fig. 4.9).
Nuclear magnetic resonance studies 44 directed to measuring the chemical
shifts of the methylene protons for the three linkages have only been
partially successful. It has proved impossible to separate the ortho-ortho
and ortho-para peaks using pyridine as the solvent. Even the use of
additives such as europium compounds and deuterated methanol have
failed to bring about splitting. Quantitative measurements of the peaks
assigned to the combined ortho-ortho and ortho-para linkages and to
the para-para linkage have given values of 71 % and 29 %, respectively.
The results of an extension of this NMR investigation to substitution
102 GL YN I. HARRIS
-@rH,-@-cH,-@-
OH OH
ortho-ortho
~H'-@-cH'-@-OH
ortho-para
para-para
FIG. 4.9. Three possible linkages III phenol-aralkyl pre-polymers.
in a phenol-aralkyl pre-polymer as well as a conventional phenolic and

a high 'ortha-ortho' phenolic are summarised in Table 4.3.
These results indicate that about 65 %of the substitution in the phenol-
aralkyl pre-polymer is in the ortho position. This is lower than in a
high ortha-ortho phenolic, but still leaves a significant proportion of
para positions free to react with hexamine during the curing stage.
The hexamine cure of phenol-aralkyl resins proceeds through di-
methylene amino bridges which slowly decompose at elevated temperatures
to give the more heat stable methylene and azomethine linkages. 45 The
ultimate products are hard brown intractable solids.
TABLE 4.3
SUBSTITUTION IN PHENOLIC NUCLEI
Resin system Condensation Position of

catalyst substitution on
phenolic nuclei
%ortho %para
Phenol/cxcx'-dimethoxy-p-xylene Stannic chloride 65 35
Pheno I/formaldeh yde Hydrochloric acid 44 56
Phenol/formaldehyde (high ortha-ortho) Zinc acetate 89 II
4.9. NATURE AND SCOPE OF THE REACTION
The condensation reaction by which the phenol-aralkyl resins are formed

falls into the broad terms of a Friedel-Crafts reaction and the most
likely reaction mechanism is that outlined below.
RCH 2 0X + SnCI 4 ¢RCHi + SnCli (1)

ArH + RCHi --+ ArCH 2 R + H+ (2)
H + + XSnCli --+ HX + SnCI 4 (3)
where X = alkoxy
R = divalent aromatic hydrocarbon or hydrocarbonoxy radical.
The aralkyl ether is ionised under the influence of the catalyst, followed
by electrophilic substitution of the aromatic nucleus of phenol or any
other reactive aromatic compound by the carbonium ion. Evidence in
support of this mechanism was provided by an examination of the influence
of substituents in the aromatic co-reactant on the ease of reaction. Methyl
or phenolic hydroxy substituents, for example, which are electron donating
overall, increase the rate of reaction. A dihydroxy-phenol reacts more
rapidly than phenol itself, which, in turn, reacts more readily than an
aromatic hydrocarbon, such as a diphenyl. By contrast, halogeno, nitro
and carboxy groups, which are strongly electron withdrawing either retard
or completely inhibit the reaction. Even when sulphuric acid is used
as the catalyst there is little reaction between benzoic acid and rt.rl-dimeth-
oxy-p-xylene. Chlorobenzene is so unreactive that it can be used as the
solvent if it is necessary to carry out the Friedel-Crafts condensation
in solution.
Some indication of the wide range of phenol-aralkyl resins which can
be prepared is given by the examples listed in Table 4.4. They can be
considered to fall into three classes on the basis of the intermediates
used. In the first class are the products of condensation of an aralkyl
ether with a single co-reactant, namely mono- or di-hydric phenols. In
addition to alkyl-substituted phenols it has been found, rather surprisingly,
that the condensation reaction even occurs, if only at a slow rate with
carboxy-, chloro- and nitro-substituted phenols as well as with salicyclic
acid.
The second class of phenol-aralkyl resins is based on the condensation
of an aralkyl ether with two phenols. By using phenols of different function-
ali ties it is possible to formulate resins with predictable physical properties
and chemical reactivities. When there is a considerable difference in the
104 GL YN I. HARRIS
TABLE 4.4
EXAMPLES OF PHENOL-ARALKYL RESINS
Intermediates
Class Aralkyl ether Co-reactant
I. aa'Dimethoxy-p-xylene Phenol
aa'Dimethoxy-p-xylene Phenol
aa'Dimethoxy-m-xylene Phenol
aa'Dimethoxy-p-xylene Diphenylolpropane
aa'Dimethoxy-p-xylene p-t-Butylphenol
aa'Dimethoxy-p-xylene Resorcinol
aa'Dimethoxy-p-xylene 2-Naphthol
aa'Dimethoxy-p-xylene Salicylic acid
IXIX'Dimethoxy-p-xylene 2-Chloro-phenol
aa'Dimethoxy-p-xylene 4-Chloro-phenol
aa'Dimethoxy-p-xylene 2,4-Dichlorophenol
aa'Dimethoxy-p-xylene 4-Nitrophenol
2. aa'Dimethoxy-p-xylene Phenol and p-cresol
aa'Dimethoxy-p-xylene Phenol and 2-hydroxydiphenyl
aa'Dimethoxy-p-xylene Phenol and 2-naphthol
aa'Dimethoxy-p-xylene Phenol and diphenylolpropane
aa'Dimethoxy-p-xylene Phenol and 44' dihydroxydiphenyl
aa'Dimethoxy-p-xylene Phenol and salicycIic acid
ClI:lDimethoxy-p-xylene p-Cresol and resorcinol
aa'Dimethoxy-p-xylene Diphenylolpropane and 2-hydroxydlphenyl
3. aa'Dimethoxy-p-xylene Phenol and diphenyl
aa'Dimethoxy-p-xylene Phenol and diphenyl ether
aa'Dimethoxy-p-xylene Phenol and diphenylamine
aa'Dimethoxy-p-xylene Phenol and carbazole
aa'Dimethoxy-p-xylene Phenol and diphenylsulphone
aa'Dimethoxy-p-xylene 2-Hydroxydiphenyl and diphenyl ether
aa'Dimethoxy-p-xylene p-Cresol and triphenyl phosphate
aa'Dimethoxy-p-xylene p-Cresol and octaphenyIcycIotetrasiloxane
ease of electrophilic substitution of the phenols, the more reactive species

rapidly become poly-substituted and so lead to premature gelation of
the system. In such cases, an alternative synthetic procedure can be used.
The aralkyl ether and the less reactive phenol are condensed to a pre-
determined stage before the second phenol is added and the reaction
taken to completion. Resins based on resorcinol and phenol itself have
been prepared by this method.
Similar considerations apply to the third class of phenol-aralkyl resins
listed in Table 4.4. In this series, the co-reactants are phenols with other
classes of aromatic compounds, or alternatively heterocyclic or organo-

metallic compounds. The non-phenolic co-reactants are introduced with
the object of improving specific properties. Hydrocarbons, such as
diphenyl, lower the permittivity and water absorption, whilst compounds
such as diphenyl sulphone, diphenylamine or carbazole may be used
to introduce sulphone or amino groups and so improve the prospects
of bonding to metal substrates. Organometallic intermediates such as
triphenylphosphate or octaphenylcyclotetrasiloxane may be introduced
into these polymers to improve flexibility and heat stability or to give
a polymer which, on complete thermal decomposition, still leaves an
inorganic residue.
4.10. COMPARISON OF STRUCTURE AND PROPERTIES
The structure of the simple phenol-aralkyl pre-polymer shown in Fig.

4.8 has a close relationship to that of the phenolic novo lacs. The essential
difference is that the phenolic nuclei are linked by xylylene rather than
methylene bridges (Fig. 4.10). This structural difference is reflected in
a number of important physical properties, namely:
(I) chemical reactivity,
(2) thermal stability,
(3) permittivity and loss tangent,
(4) water absorplion and
(5) chemical resistance.
The chemical reactivity of phenol-aralkyl resins in terms of their curing
rate with hexamine is slower than for phenolic novolacs. This is to be
OH OH
(a)
OH
@-cH'~ OH
(b)
-@-cH'-@-CH'~
FIG. 4.10. Comparison ofresm structures, (a) phenolic resin, (b) phenol-aralkyl
resin.
106 GL YN I. HARRIS
expected since the number of reactive sites/unit length of chain IS lower.

Fortunately, in laminate manufacture this is not a problem, while the
cure times of phenol-aralkyl moulding compounds can be advanced to
at least those of phenolics by operating the moulding equipment at a
temperature about 20°C higher.
The phenol-aralkyl resins are far more heat stable than phenolic resins,
and this can be explained in the context of the generally accepted oxidative
degradation mechanism for the latter proposed by Conley and Bieron 46
given in Fig. 4.11. This involves two stages, namely a primary oxidation
1 decompoSlllOn
FIG. 4.11. Oxidative degradation of phenol-formaldehyde condensates.
of the substituted dihydroxydiphenylmethane unit to substituted di-

hydroxybenzophenone, followed by a secondary oxidation leading to
chain scission. Infrared studies by Harris 47 indicate that oxidation of
methylene to carbonyl groups proceeds at a similar rate for both phenolic
and phenol-aralkyl resins. The overall difference in heat stability of the
two classes of resins is, undoubtedly, due to the relative thermal oxidative
stability of the two ketonic species shown in Fig. 4.12. Proton resonance
can occur in the first species between two hydroxy groups and one carbonyl,
~-t&
~-<Q!1-t&
FIG. 4.12. Primary oXidation products.
while in the ketonic species, derived from the phenol-aralkyl resin, there
is only the possibility of resonance between one hydroxy and one carbonyl
group. In addition, the presence of a non-phenolic aromatic ring probably
acts as an energy sink. The overall effect is that on exposure to high
temperatures the phenolic resin develops bonds with a high double bond
character and is consequently more susceptible to secondary oxidation
than the simple phenol-aralkyl resin.
A third difference between the two classes of resins is in the respect
of permittivity and loss tangent. The lower number of phenolic hydroxy
groups in the phenol--aralkyl resin is reflected in its less polar character
and lower dielectric properties. Typical values for glasscloth reinforced
laminates are shown in Table 4.5. The permittivity of the phenol-aralkyl
laminate is nearly one unit less than that of the phenolic, and the lower
dielectric loss of the phenol-aralkyl resin makes it a better insulating
material than the phenolic resin.
The lower polar character of the phenol-aralkyl resin is also reflected
in a water absorption which is about one-third that of a phenolic. This,
TABLE 4.5
COMPARISON OF PERMITTIVITY AND LOSS TANGENT
Property Phenol-aralkyl Phenolic

resin resin
Permittivity at I MHz
dry 4·77 5·73
wet 4·82 5·81
Loss tangent at I MHz
dry 0·011 0·012
wet 0·013 0·038
108 GL YN I. HARRIS
in turn, leads to better electrical properties and dimensional stability

in a damp environment.
The reduced number of hydroxy groups also influences the chemical
resistance. Thus, for example, a glasscloth reinforced phenolic laminate
is completely destroyed after 100 h immersion in 10 %caustic soda solution
at 90°C. By contrast, a phenol-aralkyl laminate shows a 75 % strength
retention.
4.11. UPGRADING PHENOLIC NOVOLAC AND RESOL

RESINS
While the difference in chemical structure results in some significant

changes in physical properties, the overall close structural relationship
between phenolic novolac and phenol-aralkyl resins results in them being
completely compatible. The addition of phenol-aralkyl resins to upgrade
the performance characteristics of phenolic novo lacs is currently being
utilised in a number of product areas including moulding compounds,
lamp capping cements and friction linings. In each instance, an upgrading
in the thermal stability is the most sought after improvement. Some
80r-----------------------------------~
N
'E 60
z
:::0;
..c: 50
0.
c:
~
1;; 40
'"!:
.><
0
~
30
.0
I
'"'"0
U
20
~
co
:;J
Q.
(;
10
0
0
Heal ageing II me al 250°C (h )
FIG. 4.13. Changes in the cross-breaking strengths with heat ageing of asbestos-
filled mouldings, based on phenolic Xylok 225 and blended resins.
indication of the extent to which this can be achieved is provided by

the plots in Fig. 4.13.
The mouldings based on 1:3, Xylok 225:phenolic have a useful life
at 250°C, about twice that of phenolic mouldings, while mouldings based
on 3: 1, Xylok 225: phenolic show a strength retention on exposure at
250°C which is slightly inferior to the Xylok 225 mouldings. In addition
to thermal stability, the solvent resistance, water absorption, dimensional
stability and dielectric properties of the mouldings are improved by the
introduction of the Xylok 225.
The upgrading of phenolic resins with phenol-aralkyl resins is not
limited to phenolic novo lacs. Phenolic resols can in many instances also
be upgraded, but the results are far less predictable. It has been found
by Harris and Golledge 48 that the addition of a phenol-aralkyl pre-
polymer can, depending on the particular resol, upgrade properties such
as:
(1) thermal stability,
(2) mechanical strength and
(3) wet insulation resistance.
The indications to date are that the upgrading is only achieved with
600~-----------------------------------,
500
10 90 Xylok 214 resol

.<:
g. 300
!
VI
e
~ 200
G:
100
o 400 800 1200 1600 2000

Heat ageing time at 180°C (h)
FIG. 4.14. Flexural strengths of glasscloth reinforced phenol-aralkyl modified
phenolic resol laminates.
110 GL YN I. HARRIS
resols having high methylol contents. An indication of the improvement

in thermal stability of glasscloth reinforced resol laminates which can
be achieved by the addition of 5, 10 and 15 % of phenol-aralkyl pre-
polymer to such a resol is given in Fig. 4.14.
4.12. HIGH-TEMPERATURE STABILITY AND STRENGTH
In most applications it is the combination of properties offered by a

reinforced composite rather than individual ones which influences selec-
tion. For high-temperature structural applications it is the combination
of thermal stability and high-temperature mechanical strength and strength
retention which is generally of paramount importance. The glasscloth
reinforced phenol-aralkyl resins fall into this category on prolonged
exposure at temperatures up to 250°C. The results of a comparative
study with other classes of post cured glasscloth laminates are shown
in Fig. 4.15. The laminates based on a phenolic and a special high-
temperature phenolic have a high initial flexural strength and show poor
thermal stability. The methyl nadic anhydride-cured epoxy and epoxy-
novolac, the silicone and the thermosetting acrylic have relatively poor
mechanical strength at 250°C but quite good strength retention. The
laminate based on an American produced polyimide alone has given
an overall performance approaching that of the phenol-aralkyl board,
namely Xylok 210. That is, high initial flexural strength and a moderately
good strength retention, i.e. about 60 %. By comparison, the Xylok 210
laminate gives an initial flexural strength at 250°C of the same order,
and over 80 % retention after 1000 h, dropping to just over 50 % after
2000 h. Even at higher temperatures, the glasscloth reinforced phenol-
aralkyl laminates give quite high strength retentions. They fall to 50 %
of their initial value after 750-1000h at 275°C and after about 300h
at 300°C.
An Arrhenius plot reported by Buchi and Kultzow 49 on glasscloth
reinforced phenol-aralkyl (Xylok 210) and a bis-maleimide system
(Kerimid® 601) is given in Fig. 4.16. The results show that the phenol-
aralkyl composite is the marginally more thermally stable, with a 20000 h
thermal classification temperature for 50 % strength retention at about
183°C. It, consequently, satisfies the Class 'H' insulation requirements
of the International Electrotechnical Commission.
® Kerimid is a registered trademark of Societe Des Usines Chimiques
Rhone~ Poulenc.
900rl----------------------------------------, 900
800
700
'"0
:I:
m
N N 600 Z
I I o
E E tj
z z >
::< ::<
.L;
.;: ~
t"'
c.c c> ~
c
~ ~ -<
t"'
u; u;
~ ~
>
Z
x
Q)
"x ~
" ~Xylo"210
u: "- \.\ :;o:l
"m
300 t"'
>
;;l
200
"25
t"'
-<
m
==
Epoxy ~
0' I I I ! ,
o 200 400 600 800 1000 00 200 400 600 BOO 1000
Heat ageing time at 250"C (h) Heat ageing time at 250°C (h)
(a) ( b)
FIG. 4.15. Flexural strengths of post cured glassc10th laminates on exposure at 250°C for up to IOOOb, (a) tested at room
temperature, (b) tested at 250°C.
112 GLYN I. HARRIS
v/ / .
V
II.:
/
.' / 20,000
//
1/
.~/
-z 10,000
/ h'
/ i'
/
7,1
II/?
I- ["j.
I~/
/!, /
/
1,000 ~
Q)
E
'"
7
77 .'
/
'"o
7
'/
/
--Xylok 210 (50% retention)

_·_·-Xylok 210 (70% retention)
Kenmld 601 (70%retenllon)
300 280 260 240 220 200 180 160 150

Temperature °C
FIG. 4.16. Retention of flexural strength of glasscloth reinforced Xylok 210 and
Kerimid 601 laminates.
4.13. ELECTRICAL PROPERTIES AND CHEMICAL

RESISTANCE
The more important electrical properties of glassc10th reinforced phenol-

aralkyl composites and a series of other post cured boards together
with some data on chemical resistance are given in Table 4.6. The best
combination of electrical properties are given by the phenol-aralkyl,
silicone and epoxy composites, but only the phenol-aralkyl and silicone
composites give virtually constant permittivity and loss tangent values
over the temperature range 20 ° to 200°C. The overall chemical, solvent
and oil resistance of the phenol-aralkyl board is, however, very much
TABLE 4.6
COMPARISON OF ELECTRICAL PROPERTIES AND CHEMICAL RESISTANCE"
Glasscloth laminates
Property Phenolic SpeCIal Epoxy Epoxy- Acrylic Silicone Polyimide Xylok 210
phenolzc novolac
Electncal
Electric strength at 20°C (MV m - l)b 20 30 26 28 4 26 22 28-33 't)
Electric-strength life at 250°C (h) 144 216 700 430 o 750 144 1000--1400 :r:
m
Insulation resistance (Mil) 2·5 x 104 3·4 x 104 1·7 x 10 5 1·3 x 10 5 4·8 X 10 5 2'0 x 10 5 7 x 10 3 4·0 x 10 5 Z
Comparative tracking index 195 170 180 240 115 370 180 185 o
Permittivity at I MHz 7
:>
dry 5·73 5·77 5·05 5·34 4·42 3·80 5·77 4·77 :;>0
5·87 4·82 :>
wet 5·81 5·93 5·08 5·37 4-46 3·88 r
Loss tangent at 1 MHz ;:<:
0·0122 0·0261 0·0218 0·0107
-<
r
dry 0·0174 0·0165 0·023 0·003
wet 0·0380 0·0265 0·0208 0·0182 0·024 0·008 0·0243 0·0130 :>
Chemical resistance z
t:l
Change in weight, %, after exposure
for 168h in ~
r
water at 100°C 1·6 2·0 0·6 0·5 1·2 0·6 :>
10% NaOH at 90°C destroyed 2·8 24·0 -12'5 destroyed 2·5 rri
10% HCI at 90°C -8'2 4·9 -5,0 -14'3 -12·5 -5·7 t:l
't)
30% antifreeze at 90°C 9·4 2·0 0·9 1·4 0·9 0·9 or
engine oil at 150°C 1-4 0·2 0·2 2·3 1·3 0·1
0·5 0·0
-<
Skydrol 500B at 100°C 1·9 11·6 0·0 Delam
transformer oil at 100°C 0·2 -0,5 -0·6 1·9 0·0 -0'2 ~
toluene at 110°C 0·9 6·7 0·6 Delam -0·2 -0·2 ~
trichloroethylene at 85°C 3·3 13-8 1·3 Delam 0·2 0·0
a Reproduced from Harris, G. I., Edwards, A. G. and Huckstepp, B. G., FrIedel~rafts resins composites for hostile
environments, PlastiCS and Polymers, Dec. 1974; by permission of the publishers.
b The electric strength was measured on laminates 1· 5 mm thick in air, between 38 mm and 76 mm-diameter brass electrodes
under a voltage rise of 1 kVs -1. Proof testing was carried out at 11·8 MV m - 1 for 60 s when the electrIc-strength life was
being detennined. The insulation resistance was measured according to BS 2782 and the comparative trackIng index l;J
according to BS 3781. The dielectric measurements were made by Lynch's method (ERA 5183).
114 GL YN I. HARRIS
superior to that of the silicone. Components machined from glassc10th

reinforced boards have found applications in a number of chemical plant
applications including valve face plates of pumps handling highly corrosive
mixtures of chlorosilanes and benzene, end stops of flowmeters and
reducing pipe joints handling concentrated hydrochloric acid.
4.14. RADIATION AND FLAMMABILITY RESISTANCE
The radiation resistance of highly aromatic polymers is known to be

good, 50.51 so the excellent results obtained at the Rutherford High Energy
Laboratory on glasscloth reinforced phenol-aralkyl laminates are, per-
haps, not surprising. The changes in flexural strength of these laminates
on exposure to 10000 Mrad of gamma radiation at a dose rate of
2 Mrad h -1 are summarised in Table 4.7. The flexural strength shows
a fall of 8 %, while the flexural modulus increases by nearly 5 %.
TABLE 4.7
RADIATION RESISTANCE OF GLASSCLOTH REINFORCED PHENOL-ARALKYL
LAMINATE
Property Values ajter irradlGtion oj
0 1628 2500 4000 10000

rad Mrad Mrad Mrad Mrad
Flexural strength (MN m - 2) 605 635 670 675 555

Flexural mod ulus (G N m - 2) 31·1 31·2 32·2 31·9 32·6
The natural self-extinguishing character of the phenol-aralkyl com-

posites is quantitatively confirmed by a value of VE-O with the UL-94
test and an oxygen index of 62 %when examined against ASTMD 2863.
Both the phenol-aralkyl laminates and mOUldings are also characterised
by outstanding low smoke generating characteristics. Results of smoke
generation testing, performed in accordance with National Bureau of
Standards procedures, indicate times to reach total obscuration in excess
of 90 s for the phenol-aralkyl/glasscloth composites. These materials pass
current FAA requirements for smoke generation.
PHENOL~ARALKYL AND RELATED POLYMERS 115
4.15. EPOXIDE CURED PHENOL~ARALKYL RESINS
Phenol~aralkyl resins cured with hexamine are suitable for the manu-
facturing of thin, but not thick section glassc10th reinforced composites,
because of the processing difficulties arising from the release of ammonia
as a by-product of the cure. This shortcoming led to a search for an
alternative mechanism for curing phenol~aralkyl resins. Harris and
Edwards 52 found that on heating phenol~aralkyl pre-polymers with
selected cyc10aliphatic diepoxides and imidazole accelerators, an effective
cure can be achieved with the release of only negligible quantities of
by-products. The crosslinking occurs between the phenolic hydroxy groups
of the pre-polymer and the epoxide groups of the diepoxide, as shown
in Fig. 4.17.
~ /""
/""0
OH+CH~H-R~H~H2+HO
0 ~
heat
I
ImIdazole accelerator
m
,j..
OH OH ~
%--CH,--1H--R~H CH,-~
FIG. 4.17. Epoxide cure of phenol-aralkyl resins.
These resins are available commercially as two-pack systems, and are

characterised by rapid curing at temperatures above 140°C. They readily
wet a range of reinforcements to give composites with excellent high-
temperature mechanical strength coupled with sufficient heat stability
to satisfy the requirements of Class 'H' (l80°C) electrical insulation.
The superior high-temperature flexural strength retention of an un-
postcured glassc10th reinforced epoxide-cured phenol~aralkyl resin
laminate (Xylok 237) relative to postcured epoxide, epoxide novolac and
silicone composites is illustrated in Fig. 4.18.
Both the epoxide and epoxide novolac resins used for this comparative
study were cured with methyl nadic anhydride, in order to develop maxi-
mum high-temperature mechanical strength. The superior strength
retention of the glassc10th reinforced phenol~aralkyllaminate up to 180°C
116 GL YN I. HARRIS
800.------------------------------;
700
'"E
I
z
~
o~----~------~------~------~
20 60 100 140 180
Temperature °C
FIG. 4.18. Changes in flexural strength of various glasscloth reinforced laminates,
with temperature.
reflects the high degree of crosslinking. This is confirmed, also, by the

relative stable dielectric properties over the same temperature range, shown
in Figs. 4.19. The small changes in both permittivity and loss tangent
of the phenol-aralkyl resin, when measured over the temperature range
20-180°C, contrasts with the large changes occurring above 140°C for
the epoxide-novolac composite.
A measure of the wetting characteristics of an epoxide-curing phenol-
aralkyl resin (Xylok 237) is indicated by the ease with which it coats
and impregnates Nomex® paper, which is notoriously difficult. Further-
more, it bonds the plies to give composites with the physical properties
summarised in Table 4.8.
® Nomex is the registered trademark of DuPont de Nemours.
7 Or,------------------------------, 0·20
I J
Epoxy- novolac
tTl
L.. .... u"'l IIVYVI'--"'~ I
'"::z::Z
60 0·15i 0
t"'
I
>
::c
>- >
t"'
- ;:0::
:~ 5.0t eC' - ><
t"'
E Xylok I237
) >
Z
~ (unpostcured)
t01O~ ",Iun r...,.J'
t:)
::c
tTl
005 t- / t"'
40C Silicone? J >
--l
tTl
Silicone, t:)
\ ~----1
0
'"
t"'
3.020 180
60 100 140 180 20 60 100 140 s::><tTl
Temperature °C Temperature °C ~
(a) (b)
FIG. 4.19. Changes in (a) permittivity and (b) loss tangent of glasscioth reinforced epoxide-cured phenol-aralkyl resin, with
temperature (measured at 50 Hz).
- ..J
--
118 GL YN I. HARRIS
TABLE 4.8
PHYSICAL PROPERTIES OF
XYLOK 237/NOMEX 411 LAMINATES
Property Value
Specific gravity 1·26

Flexural strength (MN m- 2) 130·98
Compressive strength (MN m-2) 288·85
Izod impact strength (J/12·7 mm) 18·53
Dielectric strength (MV m -1) 60
Water absorption (%) 0·25
Punchability Good
Machinability Excellent
The high Tg value of the glasscloth reinforced epoxy-cured phenol-

aralkyl laminates, namely 220 e, coupled with good coating character-
D
istics have resulted in this resin being extensively evaluated by Phillips

and Murphy53 at the RAE in carbon fibre composites. They reported
that the resin in these composites continues to be characterised by very
rapid gelation and cure, and develops good hot strength retention without
postcure. This ability to produce high quality laminates after only a
short, single stage cure is unusual in the carbon fibre field. The adhesion
to carbon fibre is outstanding, while the room temperature properties
including flexural, tensile and interlaminar shear strength are higher than
is attainable with the widely used bisphenol A precondensate system,
Shell DXIO/BF 3 400.
4.16. CONCLUSION
To conclude, the phenol-aralkyl resins constitute an attractive new class
of high performance thermosetting resins which can be used alone or
in combination with phenolic resins. The opportunities offered, particu-
larly by the second use, will undoubtedly grow as demands for easily
processable high performance composites continue to increase. In that
respect the phenol-aralkyl resins offer to phenolic resins-the old 'war
horse' of the plastics industry-a new lease of life.
REFERENCES
1. FRIEDEL, C. and CRAFTS, J. M. Bull. Soc. Chim. Fr., 1881,25,52.
2. JACOBSON, R. A. J. Amer. Chern. Soc., 1932,54, 1513.
3. BEZZI, S. Gaz. Chim. tal., 1936, 66,491.

4. FUSON, R. C. and McKEEVER, C. H. Org. React., 1942, 1,63.
5. SHRINER, R. L. and BERGER, L. J. Org. Chem., 1941,6,305.
6. HAAS, H. c., LIVII\GSTON, D. I. and SAUNDERS, M. J. Polym. Sci., 1955, 15,503.
7. PARKER, D. B. V. European Polymer J., 1969,5,93.
8. ELLIS, B. and WHITE, P. G. J. Polym. Sci., 1973, Il, 801.
9. ELLIS, B., WHITE, P. G. and YOUNG, R. N. European Polymer J., 1969,5.307.
10. KENNEDY, J. P. and ISAACSON, R. B. J. Macromol. Chem., 1966, 1,541.
11. CONLEY, R. T. J. Appl. Polymer Sci., 1965,9,1107.
12. PARKER, D. B. V. RAE Tech. Note Chem., 1956, 1284.
13. PHILLIPSJ L. N. Trans. Plastics inst., 1964,32,298.
14. DOEBENS, J. D., US patent No. 2,911,380,1959.
15. GEYER, G. R., HATCH, M. J. and SMITH, H. B., US Patent No. 3,316,186,1967.
16. SPRENGLING, G. R., US Patent No. 3,405,091, 1968.
17. ANON. Modern Plastics, Oct. 1962, 110.
18. ANON. Chem. Eng., 1962, 69, 73.
19. ANON. Dow Diamond, 1963, 26, 5.
20. PHILLIPS, L. N. RAE Tech. Report, 1963, CPM 3.
21. GRASSIE, N. and MELDRUM, I. G. European Polymer J., 1969,5, 195.
22. GRASSIE, N. and MELDRUM, I. G. European Polymer J., 1970,6,499.
23. GRASSIE, N. and MELDRUM, I. G. European Polymer J., 1970,6,513.
24. GRASSIE, N. and MELDRUM, J. G. European Polymer J., 1971,7,17.
25. NIXON, B. RAE Tech. Note, 1964, CPM 71.
26. MOORE, B. J. C. RAE Tech. Report, 1965,65161.
27. HARRIS, G. I. and MARSHALL, H. S. B., UK Patent No. 1,099,123, 1968.
28. PHILLIPS, L. N., UK Patent No. 1,094,181, 1967.
29. HARRIS, G. I. (unpublished work).
30. EvISON, W. E. and KIPPING, F. S. J. Chem. Soc., 1931,2774.
31. YAKUBOVICH, A. YA. and MOTSAREV, G. V. J. Gen. Chem., Moscow, 1953,23,
1414.
32. HARRIS, G. I., UK Patent No. 1,127,122, 1968.
33. HARRIS, G. I. and EDWARDS, A. G. (unpublished work).
34. PAXTON, J. C. Procurement Executive, Ministry of Defence, D. Mat. Report,
1973,190.
35. PARKER, B. M. RAE Tech. Report, 1972, 72029.
37. PARKER, B. M. RAE Tech. Memorandum Mat., 1975, 217.
39. HUCK, P. J. and PRITCHARD, G. J. Polym. SCI., 1973, 11,3293.
40. HARRIS, G. I. and COXON, F., UK Patent No. 1,150,203, 1969.
41. HARRIS, G. I. Brll. Poly. J., 1970,2,270.
42. BENDER, H. L., FARNHAM, A. G. and GUYER, J. W., US Patent No. 2,464,207,
1949.
43. FRASER, D.A., HALL,R. W.andRAuM,J. L. J. J. Appl. Chem., London, 1957,7,
676.
44. HARRIS, G. I. and EDWARDS, A. G. (unpublished work).
45. ZINKE, A. J. Appl. Chem., London, 1951, 1,257.
46. CONLEY, R. T. and BIERON, 1. F. J. Appl. Polym. SCI., 1963,7,103.
120 GL YN J. HARRIS
47. HARRIS, G. I. (unpublished work).

48. HARRIS, G. I. and GOLLEDGE, J., UK Patent Application No. 15953/77, 1977.
49. BUCHI, G. and KULTzow, R. Paper presented at SPE Antech. Conference on
High Performance Plastics, Cleveland, October, 1976.
50. ALEXANDER, P. and CHARLESBY, A. Proc. R. Soc., A, 1955, 230, 136.
51. BAUMAN, R. G. and GLANTZ, J. A. J. Polym. Sci., 1957,26,397.
52. HARRIS, G. I. and EDWARDS, A. G., UK Patent No. 1,305,551,1973.
53. PHILLIPS, L. N. and MURPHY, D. J. RAE Tech. Memorandum Mat., 1979,322.
Chapter 5
INITIATOR SYSTEMS FOR UNSATURATED

POLYESTER RESINS
v. R. KAMATH and R. B. GALLAGHER

Pennwalt Corporation, Buffalo, New York, USA
SUMMARY
The basic principles of free radical initiators, their important parameters

and new developments as they relate to the selection of effective initiator
systems jor curing unsaturated polyester resins are reviewed.
Peroxide and azo type initiators continue to be the most popular initiators
jor the fibre reinforced plastics (FRP) industry. Certain peroxides can
be decomposed by specific promoters to provide good cures at ambient
temperatures. The mechanism of activation and particularly the detrimental
ejfects of excess promoter are described. The results of studies relating
the structure of ketone peroxides to their reactivity are reviewed.
The kinetics oj thermal decomposition of initiators are related to
important processing parameters such as shelj~life, rate ofcure and tempera-
ture of cure. For easy rejerence, specific initiators used for curing in
the lOOO~J60°C temperature range are tabulated along with their lOh
tt temperatures. The advantages oj peroxyketals over popular peroxyesters,
and their sensitivity to acidic fillers are discussed. The advantages of
peroxide blends in terms of jast cure systems and especially the long
shelf-life of blends containing unique azo compounds are described. Results
oj studies directed towards the reduction oj residual styrene in cured resin
are reviewed.
In addition to azo and peroxide compounds, other sources offree radicals
are reviewed including novel carbon~carbon initiator and photoinitiator
systems based on both UV and visible light.
121
122 v. R. KAMATH AND R. B. GALLAGHER
5.1. INTRODUCTION
The fibre reinforced plastics (FRP) industry continues to enjoy steady

growth in spite of cyclical economic movement. Over the period 1967-1976
the total yearly production grew from 544 million pounds (247000 tonnes)
to 1·86 billion pounds (844 000 tonnes), and is projected to enjoy significant
growth in the future. Contributing to such growth is the development
of new and improved initiators for unsaturated polyester resin curing.
Unsaturated polyester resins consist of an alkyd polymer containing
vinyl un saturation dissolved in a vinyl monomer, most commonly styrene.
Curing or crosslinking of these liquid resins into useful solid thermosets
occurs readily by a chain addition type of copolymerisation reaction.
Free radicals capable of initiating the cure reaction can be generated
in the resin system by the decomposition of organic initiators. Other
methods such as photoinitiation and thermal initiation have also been
used to some extent in specialised areas of application.
Organic initiators can be defined as compounds which, when subjected
to heat, decompose to yield highly reactive free radical species. Organic
peroxides and azo compounds are commonly used as initiators, but are
often incorrectly referred to as catalysts. In the traditional sense catalysts
are substances which are not consumed as they speed (promote) a chemical
reaction. In the case of unsaturated polyesters, the initiating molecule,
i.e. peroxide or azo, decomposes during the cure reaction. Therefore,
the term 'initiator' is more appropriate and as such will be used throughout
the discussion.
The phenomenal growth in the plastics industry during the past 10
to 20 years has led to a considerable research effort in the area of free
radical initiators. As a result, a number of new initiators as well as
improved initiator formulations have been developed. To meet the current
demands of highly diverse applications and processing techniques, more
than 50 different organic peroxide and azo compounds in over 75 different
formulations are produced and offered commercially.
In the past, the accepted practice has been for the compounder or
fabricator to depend solely on the resin and/or initiator producer to
select the proper initiator for his application. However, in the rapidly
advancing technology of today's polyester industry it has become impera-
tive that the user fully understands the properties and characteristics
of initiators. With such knowledge, he will not only be able to use initiators
more efficiently in existing processes but will also be able to select effective
initiators for newly developing processes. To fill this need, basic principles
INITIATOR SYSTEMS FOR UNSA TURA TED POLYESTER RESINS 123
of free radical initiators, their important parameters and new developments

as they relate to selection of the most effective initiator system for specific
processes and applications, are reviewed.
5.2. ORGANIC INITIATORS
The chemical bond in an organic compound can cleave either symmetri-

cally or asymmetrically, as shown below.
X-y~x·+·y
X-y~X+ + -:Y
In the first case neutral radicals result and in the second case the result
is charged species called ions. The term free radical refers to an atom
or group of atoms with unpaired electrons. Such radicals are generally
reactive species which can react very rapidly to initiate free radical chain
reactions.
By supplying sufficient energy, in the form of heat or radiation, chemical
bonds can be broken in a symmetrical manner to produce free radicals.
In order to serve as a convenient source of free radicals by thermal
homolysis, a potential initiator must contain bonds which will break
at relatively low temperatures. The energy required to break a chemical
bond is called bond energy and is usually expressed as kJ mole - I
(kcal mole-I). To break a normal carbon-carbon bond having a bond
energy of 380 kJ mole -I (90 kcal mole -I), temperatures of 350 to 550 DC
are required to attain sufficient thermal excitation. Several types of com-
pounds have bond energies of 105-150 kJ mole -I (25-35 kcal mole -I).
These bonds can be broken in the temperature range of 50 to 150 DC,
i.e. convenient temperatures for processing polyester resins.1
By far the most common initiators used in the curing of unsaturated
polyester resins are organic peroxides (R-O-O-R') and aliphatic azo
compounds (R-N N-R') since they decompose thermally to produce
free radicals at convenient temperatures. Peroxides decompose by initial
cleavage of the oxygen-oxygen bond to produce two free radicals.
R-O-O-R' ~RO·+·OR'
Azo compounds, on the other hand, decompose by simultaneous cleavage
of two carbon-nitrogen bonds to produce nitrogen and two alkyl radicals.
R--N N-R' -4 R· + N2 + ·R'

TABLE 5.1
COMMERCIAL INITIATOR CLASSIFICATION
Type Structure IOh tt

range (0C)
o 0
I I
Diacyl peroxide RC-OO-CR 20-75
o 0
I I
Dialkyl peroxydicarbonate ROC-OO-COR 49-51
o
II
tert-Alkyl peroxyester RC-OO----R' 49-107
o
OO-lerl-Alkyl,O-alkyl I
monoperoxycarbonate R-O O-C OR' 90-100
"'/
ROO R'
/'"
Di-(tert-alkyl) peroxyketal C 92-115
ROO R'
Di-tert-alkyl peroxide R-OO----R' 117-133
tert-Alkyl hydroperoxide R-OO--H 133-172

R' OOH
Ketone peroxide
"'/
C
/'"OOH
R'
+
'" '"
R' 00 R'
C/ C/
/1 I'"
R' 0 0 R'
o 0
H H
+
other structures
R' R'
I I
Symmetrical azonitrile R'-C-N N-C-R' 50-65
I I
CN CN
R Rl
I I
Asymmetrical azonitrile R-C-N N-C-R2 55-96
I I
CN R3
INITIATOR SYSTEMS FOR UNSATURATED POLYESTER RESINS 125
Certain classes of organic peroxides can also be decomposed by specific

promoters at temperatures well below their normal thermal decomposition
temperature. In this case radicals are produced by an oxidation-reduction
(redox) mechanism which represents an asymmetrical decomposition of
the peroxide. This phenomenon is utilised not only in curing unsaturated
polyester resins at room temperature but also in accelerating their cure
at elevated temperatures. Azo compounds, due to their high degree of
chemical inertness, are ordinarily not suited for use in promoted cure
systems.
The ten major different initiator types or classes that are used commer-
cially to cure unsaturated polyester resins are shown in Table 5.1.
The particular initiator selected for a specific resin formulation will
depend primarily on the shelf-life which is required of the formulation
TABLE 5.2
REDOX INITIATORS FOR AMBIENT TEMPERATURE CURE SYSTEMS
Chemical name Structure
o
oI II@
0
Dibenzoyl peroxide @ C-OO-C 0
Cumene hydro peroxide
2,4-Pentanedione peroxide
Methyl ethyl ketone peroxide

(mixture of structures)
126 V. R. KAMATH AND R. B. GALLAGHER
as a function of the moulding process. Shelf-life is defined as the useful

life time of a resin/initiator mixture. After an initiator has been added
to a resin system the curing process starts as soon as the initiator breaks
down into active free radicals. At some early stage in the curing process
the resin reaches a gelled state. At this stage the crosslinked network
will have formed sufficiently such that flow of the resin is no longer
possible. The time taken to reach the point at which the resin has lost
its useful flow properties is defined as the shelf-life of the resin/initiator
system. Naturally, other factors such as rate of cure desired, temperature
of cure, thickness of moulded parts, effect of fillers and initiator cost
will enter into selection of the most cost effective initiator system.
Initiators which decompose at low temperatures, either thermally or
through the action of promoters, may be selected for those processes
which require little or no shelf-life. Contact moulding processes such
as hand lay-up, spray-up and resin injection make use of initiators
(Table 5.2) which can be activated with promoters at ambient temperatures.
After the initiator has been added to the resin these processes require
only sufficient time for the resin to wet out the glass reinforcement before
the desired curing reaction occurs. Processes such as continuous pultru-
sion, rotational moulding, pressure bag moulding, etc., require shelf-lives
ranging from several hours to a few days. For these processes one may
select initiators which decompose at low to intermediate temperatures.
Specific compounds are shown in Table 5.3. These initiators will give
the desired shelf-life; however, some may require storage under controlled
temperature conditions.
Certain processes make use of compound formulations which require
TABLE 5.3
INITIATORS FOR ELEVATED TEMPERATURE CURE SYSTEMS: 10h tt, 80 C
D
Name IOh tt Approximate

O· 2 M (benzene) moulding
(DC) range (DC)
Di-2-phenoxyethyl peroxydicarbonate 41 70-120

Bis( 4-t-butyIcycIohexyl) peroxydicarbonate 42 70-120
2,5-Di(2-ethylhexanoylperoxy)-2,5-dimethyl-
hexane 67 85-125
Dibenzoyl peroxide 73 90-130
t-Butyl peroxy-2-ethylhexanoate 73 90-130
2-t-Butylazo-2-cyanopropane 79" 100-140
" Trichlorobenzene solvent.

TABLE 5.4
INITIATORS FOR ELEVATED TEMPERATURE CURE SYSTEMS: 10h, It> 80°C
Name JOh tt Approximate

0·2 M (benzene) moulding
(0C) range (0C)
2-I-Butylazo-2-cyanobutane 82 0 100-145
1,I-Di(t-butylperoxy)-3,3,5-trimethy1cyciohexane 92 130-160
1,I-Di(t-butylperoxy)cyciohexane 93 130-160
1-I- Butylazo-I-cyanocyciohexane 96 b 135-165
OO-t-Butyl, O-isopropyl monoperoxycarbonate 99 130-160
t-Butyl peroxybenzoate 105 135-165
Ethyl 3,3-di(t-butylperoxy)butyrate III 140-175
Dicumyl peroxide 115 140-175
o Mineral spirits solvent.

b Trichlorobenzene solvent.
a long shelf-life ranging from one week to several months. These include
sheet moulding compound (SMC), bulk moulding compound (BMC)
and the many recent variations of this form of compound. For these
processes one must select initiators which show a high degree of thermal
and chemical stability. Specific examples of suitable initiators are shown
in Table 5.4.
Very often moulders find it advantageous to combine a thermally
stable initiator with a small concentration of a less stable initiator. Such
a combination allows faster cure times with some sacrifice in shelf-life.
A detailed discussion of recent developments in this area is included
in this chapter.
5.3. INITIA nON BY REDOX MECHANISMS
The technique of producing radicals by electron-transfer oxidation-

reduction reactions is frequently used since peroxides which have high
thermal stability may be used to provide a source of free radicals at
low temperatures. The mechanism is based on the reaction of certain
classes of peroxides with specific promoters. The terms 'promoter' and
'accelerator' are often used interchangeably to refer to compounds which
are added to a resin system to speed the decomposition of the peroxide
into free radicals. The terms are different, however, in that promoters
can function alone while accelerators are only used in combination with
promoters.
Peroxides which are readily susceptible to activation include ketone
peroxides, hydroperoxides and diacyl peroxides (see Table 5.2). Ketone
peroxides and hydro peroxides both contain the -OOH grouping and
their decomposition is commonly promoted by transition metals such
as cobalt. For enhanced activity, doubly-promoted systems incorporating
an amine such as dimethylaniline (DMA) or diethylaniline (DEA) in
combination with the cobalt are often employed. In this case the amine
compound acts as an accelerator.
Methyl ethyl ketone peroxides (MEKP) are by far the most widely
used organic peroxides for room temperature curing applications. Over
90 % of the spray-up processes are conducted by combining polyester
resin and MEKP in the spray gun at the mould site. The reasons for
the popularity of ketone peroxides are many. They are low in cost and,
since they are liquids, they can be easily and accurately metered. They
are readily soluble in polyester resin and are available in a range of
activities. Finally, MEK peroxides can be handled safely when the manu-
facturer's recommendations are followed. 2
Ketone peroxides are normally promoted with transition metals such
as cobalt. A simplified redox mechanism for this reaction is shown below.
(A) R-OOH + Co+ 2 ---+ RO· + OH- + Co+ 3

(B) R-OOH + Co+ 3 ---+ ROO' + H+Co+ 2
(C) RO· + Co+ 2 ---+ RO- + Co+ 3
Equations (A) and (B) show that either oxidation state of the transition
metal can decompose the peroxide. Transition metals can also react with
free radicals to convert them to ions as shown in eqn. (C). The significance
of this reaction is that a free radical is destroyed and, therefore, can
no longer initiate curing reactions. This is why there is an optimum
level of transition metal that should be used for promoting efficient
cures. This etTect is illustrated in Table 5.5 where cobalt naphthenate
promoter levels were increased with three ditTerent levels of methyl ethyl
ketone peroxide in a typical orthophthalic type polyester resin. The data
show that, at all three peroxide levels, the cure (Barcol hardness) begins
to deteriorate- at excessively high levels of cobalt. The data also illustrate
that if one wants faster cures without sacrificing cure characteristics,
the best approach is to increase peroxide concentration. 3
A significant development has recently occurred which will atTect all
TABLE 5.5
ROOM TEMPERATURE CURING OF POLYESTER RESIN: EFFECT OF PEROXIDE
AND PROMOTER CON CENTRA nON
Cobalt naphthenate 0'5% MEKP 1-0% MEKP 1'5% MEKP

(6'0 % cobalt)
(phr) Gel Barcol Gel Barcol Gel Barcol
(min) hardness (mm) hardness (min) hardness
0·025 79-4 45-50 39·2 45-50 27·5 45-50

0·050 46·0 45-50 19·9 45-50 14·3 45-50
0·100 22·1 45-50 12·0 45-50 8·9 45-50
0·200 15·7 45-50 8·6 45-50 6-4 45-50
1·000 8·3 45-50 4·2 45-50 3-4 45-50
2·000 7·2 40-45 4·0 40-45 3·3 40-45
3·000 6·0 35-40 3·6 35-40 2·9 35-40
Gel tImes are measured at 30°C in general purpose orthophthalic resin
using the SPI exotherm procedure.
MEKP manufacturers as well as end users in the United States. The

US Department of Transportation (DOT) has initiated regulations which
will prohibit the shipment of MEKPs which contain more than 9 %
active oxygen. Active oxygen may be defined as the extra oxygen (bound
to oxygen) in peroxy compounds as compared to their nonperoxy
analogues. The major reason for this move is to provide an extra margin
of safety during shipment and storage of ketone peroxides. Previously
the most commonly used MEKP formulation was a clear solution of
ketone peroxides in dimethyl phthalate with about 11 % active oxygen.
This type of product is widely used for hand lay-up and spray-up applica-
tions since its activity provides suitable working times. In accordance
with the DOT regulations, MEKP producers will dilute these products
such that they will contain less than 9 %active oxygen in their commercial
form.
What does this mean to the fabricator who normally uses 1 part peroxide
per hundred parts of resin (1 phr) on a weight basis? Slower gel and
cure times will be observed since the diluted products contain less active
oxygen (9 % v. 11 %). To obtain equivalent activity the fabricator can
either increase the level of peroxide and/or promoter in the resin system
or select a more reactive MEKP formulation.
The cure rate is directly related to the concentration of both the peroxide
and the promoter. It is clearly shown in Table 5.6 that a 25 % increase
in the concentration of the 9 % MEKP (1,25 phr) will provide reactivity
TABLE 5.6
POLYESTER RESIN CURE DATA: 11 %v. 9 %ACTIVE OXYGEN MEK PEROXIDES
% ActIVe Concentration 30°C Cure data
oxygen (phr)
Gel Cure Peak
(min) (mm) (0C)
11 1·00 27·2 34·3 171
MEK peroxide A 9 1·00 32·5 41·9 167
9 1·25 25·8 39·3 167
9 1·50 21·7 29·2 176
11 1·00 14·3 23·1 173
MEK peroxide B 9 0·75 28·2 40·8 169
9 1·00 18·7 28·5 169
9 1·25 14·3 23·2 175
Cure activity measured in general purpose orthophthalic resin containing
0·065 %of 6 %cobalt promoter, using the SPI exotherm procedure.
equivalent to that previously obtained with 1·0 phr of the 11 % MEKP.

Alternatively, the concentration of peroxide can be maintained at the
1 phr level while the concentration of promoter is increased.
A third approach would be to select a more reactive MEKP formulation.
As shown in Table 5.6, peroxide B is inherently more reactive than
peroxide A at equivalent concentrations. Thus O· 75 phr of peroxide B-9 %
wiIl give cure activity which is approximately equivalent to that obtained
with 1·25 phr of peroxide A-9 %.
Although the active oxygen content of MEK peroxides does affect
their cure activity, other factors such as chemical structure and presence
of certain diluents can also cause significant effects. A number of studies
have been directed towards establishing the relationship between the
chemical structure of methyl ethyl ketone peroxides and their reactivity
in various resin systems. Ketone peroxides are generaIly manufactured
by reacting a ketone, R-C(=O)-R, with hydrogen peroxide (HOOH)
under acidic conditions. This results in a mixture of peroxy structures.
The structures of the major peroxide constituents of commercial MEKPs
are shown below.
HOOH
monomer dimer hydrogen

peroxide
INITIA TOR SYSTEMS FOR UNSA TURA TED POLYESTER RESINS 131
Recent studies 4 ,5 have shown that the cure characteristics of commercial

MEK peroxides are highly dependent upon the presence and the amount
of each of these components, As shown in Table 5,7, peroxides A and
B, which contain a relatively low level of dimeric species, give significantly
faster cures in conventional orthophthalic and isophthalic resins, In vinyl
ester resins however, peroxide C, which contains a relatively high level
of dimeric species, gives significantly faster cures than peroxides A and
B, These relationships hold true for resin systems which are singly
promoted with cobalt or doubly promoted with cobalt and dimethylaniline,
TABLE 5,7
MEK PEROXIDE: EFFECT OF COMPOSITION AND STRUCTURE ON CURE
ACTIVITY IN THREE RESIN TYPES
Iphr MEK Icl % 30 D

e Resin cure time (mm) Vinyl
peroxide dlmer HzO z ester
Orthophthahc Isophthalic
A 6A 0,3 41 34 50
B 6'2 1,9 29 31 56
C 19'3 l' 3 43 24 33
Peroxides contain 9 %active oxygen,

Cure times were measured using the standard SPI exotherm procedure,
Comparison of the performances of peroxides A and B illustrates that

the gellation and cure of conventional polyester resins is augmented by
the presence of hydrogen peroxide while, on the other hand, hydrogen
peroxide contributes little or nothing to the cure of vinyl ester type
resins,
The curing of highly reactive vinyl ester resins must be done carefully
using standard MtKP systems, Thick laminates must be built up slowly
to avoid high exotherms which can cause warping and cracking, It was
recently shown 4 that an MEKP with a high proportion of monomeric
species is capable of curing thick vinyl ester parts with lower exotherms
and avoids spli tting and warping which is characteristic of cures by MEK Ps
which have a high proportion of the dimeric product, Cumene hydro-
peroxide has also been found to give good cures with low exotherms
in highly reactive vinyl ester resins,6 These initiators can be used to
increase productivity since thick flanges or thick flat laminates can often
be fabricated in one step with minimal warpage when they are used
in place of standard MEKP systems,
Due to the hazards involved in their manufacture, all commercial MEK
132 V. R. KAMAlH AND R. B. GALLAGHER
peroxides are manufactured in the presence of solvents such as dimethyl

phthalate. Special ketone peroxide formulations are also available for
specific applications. Fire resistant formulations are desirable because
of their inherent safety features which often qualify a fabricator for
reduced fire insurance rates. Low assay or half strength formulations
are desirable for spray-up applications since they can be more accurately
metered.
Diacyl peroxides, especially benzoyl peroxide (BPO), are also frequently
used in redox type promoted systems. In this case the promoter is most often
a tertiary aromatic amine, such as dimethylaniline. A simplified reaction
mechanism is illustrated below.
CH 3 0 0
I + I _ I
C6HS-N· + .OCC6HS + OCC 6H S
I
CH 3
Here again, as in the case of ketone peroxides, overpromotion is detri-
mental to the cured properties of the polyester resin.
Although not as widely used as the ketone peroxides, BPO is used
extensively for applications such as auto body and marine repair kits
and for the fast growing mine bolt adhesive industry.
Nonseparating paste and pumpable liquid formulations of BPO have
been developed for safer and easier handling and to improve the solubility
of BPO in the resin. Typical products consist of 25-55 % BPO dispersed
in a plasticiser. Fire resistant formulations which do not require yellow
precautionary labelling are also available.
5.4. INITIATION BY THERMAL DECOMPOSITION OF

PEROXIDE AND AZO COMPOUNDS
5.4.1. Rate of Decomposition-Half-Life

Peroxides and azos decompose over a range of temperatures. The rate
at which they decompose increases with increasing temperature. Half-life
(t!) is a convenient means of expressing the rate of decomposition at

a particular temperature and is defined as the time required to decompose
50 % of the initiator.
Generally, the 10 h half-life temperature is used as a reference point.
This is the temperature at which 50 % of the initiator has decomposed
in 10 h. As a rule, 10 h t! temperatures are universally accepted as the
best means of comparing the relative reactivity of different initiators.
At any given temperature the most active initiator would be the one
with the lowest 10 h t! temperature.
The half-life of an initiator is experimentally determined by measuring
its concentration as a function of time at a constant temperature. This
is usually done in dilute solution using solvents such as n-decane, mineral
spirits, etc. Specific analytical procedures vary for each initiator. In the
case of peroxides, these generally involve the liberation of iodine from
sodium iodide under controlled conditions and then titration of the
liberated iodine with standard sodium thiosulphate. Infrared and vapour
phase chromatography may also be used. Gas evolution techniques are
particularly suited to azo compounds.
When using the 10 h half-life temperature or the kinetic measurements
for the decomposition of an initiator, it is important to consider the
limitations of the data. The rate of decomposition of peroxide initiators
varies markedly with the solvent used. Thus, while for the comparison
of kinetic data for the decomposition of various initiators in a given
solvent it is useful to place them in order of reactivity, the reported
kinetics may not always apply under actual polymerisation conditions.
In general the decomposition kinetics of azo compounds show little
solvent or concentration dependence, i.e. they are much less susceptible
to any form of induced decomposition. Hence, they may be used advan-
tageously in unsaturated polyester formulations containing fillers, pig-
ments, etc., which can interfere with the efficient decomposition of peroxide
initiators.
5.4.2. Decomposition Kinetics

The thermal decomposition of most peroxides in inert solvents has been
found to follow first order kinetics, i.e. the rate is directly proportional
to the initiator concentration.
d[l]
Rate = -~ = k[J]
dt
The term k is the first order rate constant and may be determined from
the slope of the straight line obtained by plotting the logarithm of residual
peroxide concentration (in moles of peroxide group/litre of solution)
as a function of time. By integration between limits it is possible to
define a half-life for the decomposition at a particular temperature.
1 0·693
I},=--
k
The activation energy (Ea) may be obtained by studying the variation

of reaction rate with temperature. By measuring k at different temperatures
the value of Ea can be obtained from the standard Arrhenius equation,
k = Ae- Ed /RT
which expresses the rate constant in terms of two parameters: activation
energy and a pre-exponential term A which is related to the probability
of the reaction occurring. The activation energy may be determined from
the slope of the line obtained by plotting log k as a function of the
reciprocal of absolute temperature. In general, the higher the activation
energy, the greater the change in rate of decomposition/unit change in
temperature. Thus, initiators with a lower activation energy will give
a more uniform rate of decomposition over a broader temperature range.
In practical terms this means that these initiators can be used over a
wide temperature range without substantial loss in efficiency.
5.5. SELECTING INITIATORS FOR CURING AT ELEVATED

TEMPERATURES
Common initiator types used in curing polyester resins in the temperature

range of 90° to 160°C are shown in Tables 10.3 and 10.4. Processing
techniques which utilise these compounds include pultrusion, compression,
injection, transfer moulding, etc. These processes generally require a shelf-
life ranging from several days to several months at ambient temperatures.
As described earlier, the primary consideration in selecting an initiator
is generally the shelf-life which is required as a function of the moulding
process. Other important parameters for initiator selection include the
rate of cure desired, the temperature of cure and part thickness.
Rate of cure is a function of initiator half-life, initiator concentration
and reaction temperature. Data in Table 5.8 show that at a given mould
temperature the cure time decreases with decreasing 10 h It temperature.
In general, one can select a moulding temperature which will be within
TABLE 5.8
EFFECT OF INITIATOR HALF-LIFE ON CURE CHARACTERISTICS AT 121°C
Initiator 10hti Cure Peak

0·2 M (benzene) time temperature
(OC) (mm) ( °C)
2,5-Di(2-ethylhexanoylperoxy)-2,5-imethyl-
hexane 67 2·1 186
2-t-Butylazo-2-cyanopropane 79" 2·7 212
I,I-Di(t-butylperoxy)cyciohexane 95 H 216
I-Butyl peroxybenzoate 105 4·7 230
" Trichlorobenzene solvent.

Initiators compared on equal equivalents corresponding to I·Owt. % t-butyl
peroxybenzoate.
Cure time and peak exotherm temperatures were measured according to the SP!
exotherm procedure.
the optimum range for a given initiator by adding 40 ° to the 10 h It

temperature, in 0c. Using this as a starting point, it is then usually
necessary to experimentally determine the optimum process temperature
for the specific moulding compound. For any given resin/initiator system
there is an optimum temperature at which the resin can efficiently utilise
the free radicals formed by the initiator. When radicals are generated
too fast they tend to recombine to form inactive products or they can
terminate growing polymer chains (primary radical termination) resulting
in lower molecular weight polymers and a loss in physical properties. This is
illustrated in Table 5.9 where a temperature of 121°C is near optimum for
t-butyl peroctoate. Above 121°C cure times are not shortened and cure
characteristics begin to deteriorate, while below 121 °C too much time is
required to complete the cure.
Thickness of the part being moulded can play an important role in
selecting an initiator. As part thickness increases, heat transfer becomes
slower and cure time will usually increase because it takes longer to
reach reaction temperature, especially in the centre of the part. High-
temperature initiators coupled with reduced heat transfer produce higher
peak exotherms which can lead to cracking or warping of thick sections.
Use of a low-temperature initiator will result in reduced cure times,
however, for optimum efficiency one must be careful not to exceed the
optimum temperature range. High mould temperatures may also cause
pregel at the mould surface which can cause rejected parts. The best
answer is to use higher concentrations of a low-temperature initiator
TABLE 5.9
EFFECT OF TEMPERATURE ON INITIATOR EFFICIENCY
Bath ]'66%
lemperature t-Butyl peroxy-2-ethylhexanoate a
(0C)
Gel time Cure time Peak exotherm
(min) (min) (0C)
107 2·2 3·3 195

121 1·2 2·4 198
135 1·0 N 191
149 1·0 2·4 180
Cure activity measured in general purpose ortho-

phthalic resin using the SPI exotherm procedure.
a Often referred to as t-butyl peroctoate.
at reduced mould temperatures. This is particularly true, for example,

in the pultrusion process.
Pultrusion is the only process for converting continuously and automati-
cally glass reinforcement and resin into finished products and, thus, it
is enjoying steady growth and increasing importance in the FRP industry.
Although the process can be conducted over a wide range of temperatures,
temperatures around 100 e are most popular. This is especially true
0
when moulding profiles with thick cross sections are used, since lower
exotherms are obtained at lower moulding temperatures. Initiators which
give acceptable cure rates at these low temperatures are those with a
10 h ti below 80 0 e (Table 5.3). Effective initiators include benzoyl per-
oxide, 2,5-dimethyl-2,5-di(2-ethylhexanoyl-peroxy) hexane and bis( 4-t-
butyl cyclohexyl) peroxydicarbonate. The latter two, in their commercial
form, require shipment and storage at controlled temperatures.
Di-2-phenoxyethyl peroxydicarbonate is a relatively new initiator which
is now being test marketed in the form of a powdered solid. This initiator
is stable at ambient temperatures below 38°e and gives faster cures
than benzoyl peroxide.
Although each of the above initiators give acceptable cures, their per-
formance is enhanced when a low concentration of a higher-temperature
initiator such as t-butyl peroxybenzoate is added to the formulation.
For moulding at temperatures around l50 o e, t-butyl peroxybenzoate
has been considered for years to be the 'work horse' of the industry.
It offered the moulder acceptable reactivity and shelf-life, all at reasonable
cost. However, to keep pace with increasing product demand from many
application areas, the moulder found it necessary to increase the rate
of production. One way of doing this is to reduce mould cure time.
Shorter cure times can be obtained by several methods including higher
moulding temperatures, using faster (less stable) peroxides or by using
initiator blends.
Raising the mould temperature is not always feasible, nor always eco-
nomical because of equipment limitations or due to the high cost and
limited availability of energy.
A faster initiator, i.e. one which decomposes faster at a given tempera-
ture, can be used to reduce cure time. As noted in Section 5.4.1 the
10 h t! temperature is often used to compare the relative reactivity of
initiators. The initiator with the lower 10 h t! temperature is considered
the faster initiator. Using fast initiators to obtain short cure times is
possible, however, shelf-life is significantly reduced. This can be a serious
limitation in SMC ahd BMC applications where precatalysed moulding
compounds are often aged before use. One type of initiator which has
been used successfully to produce faster cures and at the same time
give good shelf-life is the peroxyketal group. 7
Peroxyketals are diperoxides that decompose to generate alkoxy radicals
upon cleavage of the peroxy bond.
R'OO Rl
~/
C ~ 2R'O· + other fragments
/~
R'OO R2
The mechanism of decomposition is not fully understood, however,

chemical analysis indicates that other decomposition products include Rl "
R 2· and CO 2,8 As with other classes of peroxide, the nature of the R' groups
and also the Rl and R2 groups plays a role in determining activity. WhenR 1
and R2 are the same, t-butyl peroxyketals are the most stable, followed by
t-amyl, t-octyl and t-cumyl peroxyketals, in descending order of stability.
The effect on 10 h t! temperature resulting from a change in Rl and
R2 groups is illustrated in Table 5.10. The first three compounds show
a decreasing stability as the R2 group increases in electron donating
capability, i.e. from C 2H s OC(O)CH 2- to C 2H s- to (CH3)2CHCH2-'
The cyclohexyl peroxides show a more significant lowering of stability
which may be due to a combination of steric strain relief and electronic
effects. 3
TABLE 5.10
COMMERCIAL DIPEROXYKETALS: EFFECT OF STRUCTURE ON THE TEN HOUR HALF-LIFE
TEMPERA TURE
t-C 4 H 9 00 Rl
"'"'/
C
t-C 4 H 9 00
/"'"'R2
Name lOh tt
0·2 M (benzene)
(0C)
0
Ethyl 3,3-di(t-butylperoxy)- \I
butane -CH 3 -CH 2COC 2H 5 111
2,2-Di(t-butylperoxy)-butane -CH 3 -C2H 5 104

2,2-Di(t-butylperoxy)-4-
methylpentane --cH 3 -CH 2CH(CH 3h 101
1,I-Di(t-butylperoxy)-
cyclohexane ----{CH 2} 5 - 95
1,I-Di(t-butylperoxy)-3,3,5-
trimethylcyclohexane -CH2CH(CH3)C(CH3)2CH2- 92
The decomposition rate of peroxyketals can be accelerated by using

certain quaternary ammonium salts. 9 This technique represents a practical
means for accelerating the cure of polyester resins while retaining service-
able shelf-life of the premix. Based on the known acid sensitivity of
peroxyketals it is also possible to effectively accelerate their decomposition
at elevated temperatures by using small concentrations of mineral acids
in the resin formulation. 3
The peroxyketals have certain limitations which can make them ineffi-
cient under specific conditions. The effect of resin formulation on the
efficiency of peroxyketals has been studied. 10 This study revealed that,
unlike t-butyl peroxybenzoate, the shelf-lives and cure activities of mould-
ing compounds precatalysed with peroxyketals are related to filler pH.
The higher the pH, the longer the shelf-life and the shorter the cure
time. Although filler particle size has no effect on shelf-life, the cure
activity of clay filled formulations is inversely related to particle size,
i.e. the smaller the particle size, the longer the cure time. The effects
of acidic fillers are minimal in SMCjBMC formulations due to the neutral-

ising effect of strongly basic thickening agents, such as Mg(OH)2.
Peroxyketals like other single initiator systems can be used efficiently
only over a relatively narrow temperature range and often the moulder
cannot achieve the desired cure time by using them alone. A number
of studies have shown that blends or mixtures of initiators provide versa-
tility which cannot be achieved by single initiators.
5.6. INITIATOR BLENDS
When a high-temperature initiator (high 10 h tt temperature) is blended

with a small concentration (l 0-25 wt %) of a low-temperature initiator
(lower lO h tt temperature), faster cures are obtained often with some
sacrifice in shelf-life. In addition to offering reductions in cure time,
it has been reported that the depth of sink marks in SMC mouldings
can be significantly reduced by using peroxide blends. I I Recent work
with novel azo/peroxide blends shows that faster cures are possible and
shelf-life can be improved. 12 A major advantage of azo initiators, com-
pared to peroxides, is that they are not subject to induced decomposition
by impurities or transition metals. Therefore, azos exhibit markedly
superior shelf-life as compared with peroxides of similar thermal stability.
In the past, azo initiators found limited acceptance for polyester mould-
ing since, under certain moulding conditions, they reportedly led to surface
porosity problems. Most recent work with these initiators has shown
that surface porosity is not a problem when they are used as the minor
component in initiator blends.
The activity characteristics and the advantages of initiator blends are
illustrated in Table 5.11. Adding a low concentration of t-butyl peroctoate
to either t-butyl peroxybenzoate or l,l-di(t-butylperoxy)cyclohexane re-
sults in significantly faster cure times, however, shelf-life is reduced in
each case. The advantage of azo compounds in blends can be seen by
comparing blends 4 and 5. Blend 5, containing the azo initiator, exhibits
faster cure time and, at the same time, longer shelf-life.
5.7. RESIDUAL STYRENE
The influence of initiator curing systems on the residual styrene content

of cured polyester resins has been the subject of several recent
c!3
TABLE 5.11
INITIATOR BLENDS: CURE ACTIVITY AND SHELF-LIFE DATA -<
?"
No. I phr Initiator blend Cure actIVity at 12I o C 3rC ~
- - - - - - - - Shelf-life >
Wt. fraction 'A' component Wt.fraction 'B' component Gel Cure Peak (days)
::
>
-l
(min) (min) (0C) :t
>
Z
I 1·0 t-Butyl peroxybenzoate 4·2 5·6 217 40 o
2 0·17 t-Butyl peroxy-2-ethylhexanoate 0·83 t-Butyl peroxybenzoate 2·3 3·1 209 12
3 1·0 1,I-DI(t-butylperoxy)-cycIohexane a 2·7 3·7 209 81 '"
tll
4 0·10 t-Butyl peroxy-2-ethylhexanoate 0·90 I, I-Di( t-butylperoxy)-cycIohexane a 2·2 3·1 204 40
5 0·10 2-t-Butylazo-2-cyano-4-methylpentane 0·90 I, I-Di(t-butylperoxy)-cycIohexane a 1·9 2·8 199 56
a
>
r
r
a80 % in phthalate solvent. >
a
:t
Cure activity and shelf-life measured in general purpose orthophthalic resin using modified SPI exotherm procedure. 1Tl
Shelf-life measured using 100 g of resin in a glass jar stored in a constant temperature oven. '"
studies. 4,13 - 15 Theoretically, the copolymerisation of the styrene monomer

with the un saturation in the resin molecule will cease when all of the
monomer has reacted. In actual practice, however, this state is seldom
achieved unless special measures are taken. The quantity of residual
styrene in a cured resin is recognised as being an important parameter
since optimum physical properties are achieved only at the lowest values.
This is important in all applications, and is particularly important for
high chemical resistance and food handling applications.
The technique most often used to determine residual styrene in cured
mouldings involves extraction of the finely divided sample with methylene
chloride followed by analysis of the extract by gas chromatography. 16
Since sample size. extraction time and extraction temperature can all
affect results. the exact procedure that one follows must be carefully
evaluated to establish the accuracy of results. Experience indicates that
for best reproducibility, the sample must be ground to a fine powder
and extracted with methylene chloride for at least 7 days at 23 ac.
In polyester resins cured at room temperature, all current studies indicate
that relatively high levels of residual styrene are found in laminates pre-
pared using standard cure systems. It has recently been reported 4 that
after extensive evaluations no MEKP-promoter-resin systems were found
which would give acceptably low residual styrene levels via room tempera-
ture cures. The lowest residual styrene levels are obtained when high
levels of peroxide are used, when high exotherms are obtained or when
post cure techniques are employed.
At the usual initiator concentratIOns of 1 wt %. reported residual styrene
levels are In the range of 5-11 %. These levels tend to decrease with
time, however, after six months at 20 ac levels are generally greater than
1·0 wt %. Peroxide concentrations as high as 10 wt % will give residual
styrene levels of < 0·01 %. however. such initiator concentrations are
not practical. Increasing the level of promoter is far less effective and,
in the case of benzoyl peroxide. can have adverse effects. 13 . 15
Ketone peroxide/cobalt systems give the lowest residual styrene levels
when a post cure treatment is used. For example. an initial residual
styrene content of 5·25 % is reduced to <0·01 % after a post cure at
80 a C for 8h.
In those cases where post cure is practical. the optimum time and
temperature must be experimentally determined for each resin. As a
general rule, however, it has been reported that 8 h at, or slightly above.
the heat distortion temperature of the specific resin will reduce the residual
styrene to acceptably low levels. 13
In polyester reSInS moulded at elevated temperatures, lowest residual

styrene levels are obtained with peroxyketals 13 . 14 and medium reactivity
peroxyesters such as I-butyl peroxyoctoate. 13 At any given moulding
temperature one can decrease residual styrene levels by increasing the
initiator concentration and/or increasing the moulding time. In BMC
formulations containing I-butyl peroxyoctoate moulded at 140°C, residual
styrene decreased from 0·3% to 0·2% when the initiator concentration
was increased from O· 5 % to 1·0 %.
The type of radical formed by the initiator can have a significant
effect on the level of residual styrene. For example, benzoyl peroxide
gives relatively high residual styrene values compared to I-butyl peroxy-
octoate, an initiator with a similar 10h Ii temperature. I-Butyl peroxy-
benzoate (tBPB) produces radicals similar to those produced by benzoyl
peroxide and is also found to give relatively high residual styrene values.
For example, in a BMC formulation moulded at 140°C and containing
1·0 % IBPB, the residual styrene was found to be 0-45 %.
Certain organometallic promoters are commercially available which
can reduce cure times when used with peresters and also significantly
reduce residual styrene. Promoter 301,t for example, at a concentration
of 1·0 %, reduced residual styrene from 0-45 % to < 0·1 % in a formulation
initiated with I-butyl peroxybenzoate. These pieces were moulded at 140°C
for 2 min. 13 Cerium compounds are also reported to give significant
reductions in residual styrene when added to formulations containing
peroxyesters. 17
5.S. OTHER SOURCES OF FREE RADICALS
5.S.1. Carbon-Carbon Initiators

Azo and peroxide compounds are convenient, low cost sources of free
radicals. However, considerable work has been done to develop other
sources of free radicals to cure polyester resins. A majority of them
are designed for specialised applications where cost considerations are
not very critical. These include a new class of compounds, commonly
known as carbon-carbon initiators. They are tetra substituted dibenzyl
compounds and can be represented by the general structure:
Rl Rl
©-t-t-@ R2 R2
t Air Products and Chemicals, Inc. USA.
where Rl and R2 are bulky substituents. The carbon-<:arbon bond here

is highly strained and, thus, undergoes homolytic scission readily at low
temperatures to yield two free radicals. The decomposition temperature
of these compounds is strongly influenced by the nature of the substituents.
A series of tetra phenyl ethane compounds were recently investigated 18
as potential initiators for curing polyester resins. Results of this study
indicate that compounds such as I ,2-dimethoxy-l, 1,2,2-tetraphenyl ethane
(0 MTPE) and 1,2-dichloro-l, 1,2,2-tetraphenyl ethane (OCTPE) will
effectively initiate the cure reaction at temperatures in the range of 100-
130°C. The properties of the cured resin, i.e. flexural strength, impact
strength and hardness were found to be equivalent to those obtained
with benzoyl peroxide. Although these compounds tend to be more stable
than peroxides, they have certain disadvantages which have, to date,
limited their acceptance on a commercial scale. They are solids which
can pose dispersion and resin solubility problems. In addition, their chemi-
cal structure does not lend itself to activation and, thus, they are limited
to elevated temperature curing systems. These disadvantages along with
their high cost will probably limit commercial utilisation.
5.8.2. Photoinitiators
The unimolecular decomposition of photoinitiators as a source of free
radicals has certain inherent advantages. Considerable effort has been
directed towards development of commercially viable systems using ultra-
violet and visible light as sources of energy. These systems are especially
important in the field of coatings where benzoin ether compounds are
popular UV photoillitlators. A recent review 19 of these systems indicates
that a number of new photoinitiators are being developed to keep pace
with projected growth m these applications.
Visible light curing of unsaturated polyesters using photoinitiators is
a commercially feasIble method which has recently been described. 20
The use of visible light offers the benefit of eliminating the hazards
associated with high energy UV radiation and thus increases the practi-
cality of the system for the FRP industry. Another major advantage
compared to UV technology is that thick parts can be cured quite easily.
Various photoinitiator-amine reducing agent combinations have been
evaluated in visible light activated systems. 21 The amine compound is
selected such that it will reduce the photoinitiator in its excited state
only. Of the various amines evaluated, shortest gel times were obtained
with I-methyl imidazole. In terms of photoinitiators, faster gel times
(e.g. 30 to 60 s) were obtained with benzoin ether as compared to benzil
( '" 3 mm). This study also identified vinyl ester resins, bisphenol A/fumaric
acid polyesters and isophthalic polyesters as having clear visible-light-

curing specificity. The technology is projected as being suitable for
electronic materials processing including potting, encapsulation and
prepreg processing applications.
REFERENCES
I. PRYOR, W. A. Free Radicals, 1966, McGraw-Hill Inc., New York.

2. GALLAGHER, R. 8. and KAMATH, V. R. Plastics Design and Processing,
1978, 18(7), 38.
3. SHEPPARD, C. S. and KAMATH, V. R. Polym. Eng. Sci., 1979, 19,597.
4. CASSONI, J. P., HARPELL, G. A., WANG, P. C. and ZUPA, A. H. 32nd SPI
Reinforced Plastics/Composites ConI, 1977, Paper 3-E.
5. THOMAS, A., JACYSZYN, 0., SCHMITT, W. and KOLCZYNSKI, J. 32nd SPI
Reinforced Plastics/Composites Conj., 1977, Paper 3-8.
6. MAxsTADT, A. KunststojJe, 1979,69,266.
7. THOMAS, A., ROSKOTT, L., GROENENDAAL, A. A. M. and KOLCZYNSKI, J. R.
33rd SP I Reinforced Plastics/Composites Conj., 1978, paper 5-E.
8. BUKATA, S. W., ZABROCKI, L. L., McLAUGHLIN, M. F., KOLCZYNSKI, J. R. and
MAGEL!, O. L. Ind. Eng. Chem. Prod. Res. Develop., 1964,3,261.
9. US Patent No. 4,032,596 issued to Air Products and Chemicals Inc., 28 June,
1977.
10. CASSONI, J. P., GALLAGHER, R. B. and KAMATH, V. R. 34th SPI Reinforced
Plastics/Conpo~tles Conj., 1979, Paper 14-F.
11. AMPTHOR, F. J. Pla~tics World, 1978,36(5),48.
12. KAMENS, E. R., GALLAGHER, R. B. and KAMATH, V. R. 34th SPI Reinforced
Plastics/Composites Conf., 1979, Paper 16-B.
13. RosKOTT, L. and GROENENDAAL, A. A. M. 33rd SPI Reinforced Plastic~,/
Composites Conf., 1978, Paper 5-B.
14. IZZARD, K. J. and NEWTON, G. P. 33rd SPI Reinforced Plastics/Composites
Conf., 1978, Paper 5-C.
15. VARCO, P. 30th SP I Reinforced Plastics/Composites Conj., 1975, Paper 6-C.
16. SHAPRAS, P. and CLOVER, G. C. Anal. Chem., 1964,36,2282.
17. German Patent No. 2,815,924 issued to Akzo G.m.b.H., 19 April, 1977.
18. BRAUN, D. and QUELLA, F. KunstojJe, 1979,69, 100.
19. DELZENNE, G. A. Makromol. Chem., 1979, Supp\. 2,169.
20. DIXON, B. G., LONGENECKER, D. M. and GRETH, G. G. 32nd SPI Reinforced
Plasllcs/Composites Conj., 1977, Paper 5-0.
21. KOLEK, R. L. and HAMMILL, J. L. Plastic~ Compounding, 1979,2(6),52.
Chapter 6
HIGH-TEMPERATURE PROPERTIES OF
THERMALL Y STABLE RESINS
G. J. KNIGHT
Royal Aircraft Establishment, Farnborough, Hants, UK
SUMMARY
This chapter deals with the chemistry and the thermal and mechanical
properties of a number of resin systems knov.'n to possess good high-
temperature properties. The resins examined include the poly(aryl ether
suIphones), poly(phenylene sulphide), phenol-formaldehyde and melamine-
formaldehyde resins, silicones, and the polyimides. In the Introduction
the term thermal stability is discussed, together withfactors to be considered
when looking for a heat resistant polymer.
6.1. INTRODUCTION
6.1.1. The Search for New Resins

Mainly due to the demands of the aerospace industry, there has been
a considerable amount of effort put into the search for temperature
resistant materials during the past 15 to 20 years. One of the first synthetic
polymeric materials prepared, phenol-formaldehyde resin, is, in fact,
reasonably thermally stable and is still used for such applications as
brake linings. Where good electrical properties are required together
with temperature resistance then melamine-formaldehyde or silicone resins
are used. Although these materials have served their purpose for many
years, and the silicones, for example, are suitable for prolonged service
at temperatures up to 250°C, the new demands for structural composites
145
146 G. J. KNIGHT
to be used in aircraft and missiles have stimulated the search for stronger
and yet more thermally resistant resins.
6.1.2. The Meaning of 'Thermally Stable'

Before going on to consider the chemistry and properties of the various
resin systems commercially available today one should consider exactly
what is meant by the term 'thermally stable'. It has been known for
an author, describing the synthesis of a potentially thermally stable poly-
mer, to make the statement that: 'the polymer was heated strongly and
showed no signs of decomposition', no mention being made of the tempera-
ture nor of the length of time the sample was heated. One of the first
techniques for assessing thermal stability systematically was that of thermo-
gravimetric analysis (TGA). This is a very useful tool and helps to indicate
the relative order of stability of various polymers, either in inert or
Or-------~~---- __~~--------~
20
..
c
~ 40
80
200 400 500

temprraturr·C
FIG. 6.1. Thermogravimetric analysis in air of (a) polyethylene, (b) polystyrene,
(c) epoxy resin, (d) polyethersulphone and (e) Kapton.
HIGH-TEMPERATURE PROPERTIES OF THERMALLY STABLE RESINS 147
oxidising atmospheres, as illustrated in Fig. 6.1. This illustrates the results

given by various common polymers when the temperature is raised by
2 DC/min. To a certain extent the results depend on the heating rate;
the higher the heating rates, the greater the apparent stability. (In this
chapter where a figure illustrates the weight loss behaviour of a polymer
the work has been performed on a Du Pont 951 Thermogravimetric
Analyzer attached to a 990 Thermal Analyzer, with a gas flow rate of
50 ml/min and a heating rate of 2°C/min.) The temperature of decompo-
sition, TD , the temperature at which the weight loss curve departs from
the baseline is subjecl to errors due to loss of volatiles from the polymer.
A commonly quoted value, therefore, is the temperature for 10 %weight
loss. Whatever value is quoted it must be emphasised that it does not
indicate the maximum temperature at which a material may be used.
It is not possible to give a single value for this temperature as it depends
on many variables including, for example, the local environment, the
length of time at a temperature and the type of mechanical stress to
which the polymer is subjected. In a missile to be used only once, with
a lifetime at temperature measured in minutes, much higher temperatures
are permissible than in an aircraft structure, where lifetimes of 10000
to 20000 h would be expected. Where an accurate assessment of the
life of a component is required the only way available at present is
to model the conditions and expose the article for the required time.
Attempts to shorten the test procedure by raising the temperature and
then extrapolating the results have to be treated with caution, as the
rise in temperature can provide the activation energy for reactions that
do not take place at lower temperatures. The Arrhenius equation for
the rate constant of a reaction is written
where k is the rate constant, A the Arrhenius constant, Ea the activation

energy for the reaction, R the gas constant and T is the temperature
in degrees Kelvin. By taking logs the equation becomes
Ea
Ink = - - + InA
RT
Isothermal weight loss experiments can be conducted at various tempera-
tures, then, if the reaction causing weight loss follows simple reaction
kinetic relationships, a plot of log (rate of weight loss) v. ljT should
give a straight line. Such a plot is shown in Fig. 6.2 for a polysulphone.
The experiments were performed over the temperature range 383 DC to
148 G.1. KNIGHT
~ -04
£!
.&:
a-
t -08 o
c
"' -1
c 2
-16
-20 132 140 150

11 temperolure (K) 103
FIG. 6.2. Arrhenius plot of log (rate of weight loss) at a constant temperature
v. I/temperature (K) for Udel PI700 polysulphone. Isothermal weight loss in
air.
493°C. By extrapolation, a weight loss of O' 5 %/year would be expected

at 200°C, but without knowing the actual mechanisms of the reactions
causing weight loss it is not possible to be certain of this prediction.
Isothermal weight loss experiments, like TGA, are better considered as
another means of comparing different polymers and of discovering more
about the chemistry of the compounds rather than of predicting their
useful life.
When used as a load bearing material the glass transition temperature
TABLE 6.1
THE GLASS TRANSITION TEMPERATURES Tg AND
DECOMPOSITION TEMPERA TURES TD OF THE
POLYMERS ILLUSTRATED IN FIG. 6.1
Polymer Tg °C TDoC
Polyethylene -175 1 200

Polystyrene 100 1 240
ICI Polyethersulphone 200P 230 2 4\0
Epoxy resina 2\03 250
Kapton type polyimideb 325 4 430
a The diglycidyl ether of bisphenol A hardened

by 4,4' diaminodiphenyl sulphone, post cured
for 4h at 200°C.
b The polypyromellitimide of 4,4'diaminodi-
phenyl ether.
(Tg) of a polymer gives a better indication of the upper temperature

limit for use. For example, polystyrene can be seen from Fig. 6.1 to
be stable up to about 240 o e, but its Tg is 100 °el and its melting point
is about 240°C.I Examples of the Tgs of various polymers are given
in Table 6.1 together with the decomposition temperatures, To, in air.
From Table 6.1 it can be seen that the thermoplastic materials soften
at temperatures well below their decomposition temperatures. A thermoset
resin can show a Tg but this represents the increased mobility of segments
within the crosslinked network, the overall structure remains fixed and
decomposition precedes any softening or melting. The value of 325°e
quoted as the Tg of the polyimide is an estimate and it is thought that
crosslinking reactions take place at temperatures below this.4
6.1.3. Fabrication
As mentioned above, this chapter is concerned with resins used as
reinforced pl.astics. A reinforced plastic can be considered as a composite
material consisting of continuous or discrete fillers bonded by a resin
matrix, each component contributing to the properties of the whole.
The reinforcement is the more important factor in the mechanical perform-
ance of the composite provided that there is good adhesion between
the fibres and the matrix. Thermoplastic resins being fusible and often
also soluble lend themselves to easy processing; no chemical reactions
take place during fabrication. An important factor with respect to thermo-
plastics is that the heat distortion temperature of a polymer can be
raised by the incorporation of a suitable filler (e.g. Nylon 6.6, resin
alone has a heat distortion temperature of 65°e but with 40 % glass
fibre it rises to 260°C).5 Other interactions are known between the resin
and the reinforcement but these will be discussed below with the resin
properties.
The thermoset materials have to be fabricated at some intermediate
prepolymer stage as, in the final crosslinked state, they are intractable.
Also most of the resins which show temperature resistance and are suitable
for laminating do not have properties that enable castings to be prepared,
e.g. they are very brittle and the cure produces samples with voids and
cracks. This means that few, if any, property measurements have been
made on the resins alone; the majority of values quoted in the literature
relate solely to composites.
The fact that the processor is, in effect, synthesising the final polymer
as he fabricates the product needs to be borne in mind. The curing
reaction may require the addition of a crosslinking agent and catalysts
as well as the application of heat. The mechanical properties of the
150 G. J. KNIGHT
final polymer will depend upon the chemical reactions that govern the
structure of the main chain and of the crosslink chain, and upon the
number of crosslinks. It is the variations which are possible that account
for the many resin systems that can be made. An increase in the crosslink
density of a resin causes a concomitant increase in modulus and hardness,
and in the resistance to creep, heat, and chemical reagents. At the same
time there is a decrease in impact strength, elongation at break and
reversible extensibility.
Chemically the crosslinking process can be brought about either through
addition or a condensation process. This is of practical importance as,
in the addition process, no volatiles are evolved. This means that processing
is simpler and, therefore, cheaper, lower pressures need to be used, void
formation is less likely and thicker sections are easier to make. Theoreti-
cally, after fabrication all the functional groups in the original resin
will have been used up but in practice this is rarely the case. Hence,
further cure may be possible when the article is exposed to elevated
temperatures causing changes in the mechanical properties. Also with
respect to electrical properties the number and type of polar groups
in the resin are of importance, for example OH, CO and COO groups
in the final structure may give rise to permanent dipoles in the material.
6.1.4. The Effect of Moisture

A further factor which needs to be considered when a composite is to
be used at elevated temperatures is the effect of moisture. It has been
shown 6 - 9 that water has a marked effect on the mechanical properties
of epoxy resins. It acts as a plasticiser, lowering the Tg of the resin
by as much as 20°C for every 1 % water content. As well as fostering
research into epoxy resin systems with low water pick-up it has also
focused attention on the fact that the high Tgs of the temperature resistant
resins makes them attractive candidates as the matrix for composites
to be subjected to severe environmental exposure. 8,10 -13 It can be seen,
therefore, that each resin system has its own advantages and disadvantages,
and the subsequent sections describing the various polymers will attempt
to draw attention to these.
6.2. THERMOPLASTIC RESINS

6.2.1. Introduction
A major factor in the development of reinforced thermoplastics is ease
of processing; this includes the ability to rework scrap materials. The
introduction of carbon fibre reinforcement together with the high-

temperature resins, poly(arylene ether sulphones) and poly(phenylene
sulphide) has opened up even further fields of application. The two poly-
imide resins, Du Pont's NR-150 and Upjohn's Polyimide 2080, are
thermoplastic materials and are considered in Section 6.4.
6.2.2. The Poly(Arylene Ether Sulphones)

The poly(arylene ether suI phones) are thermoplastic polymers, some of
which may be used under stress at temperatures up to 150°C. Synthetic
routes to these polymers were discovered independently in the laboratories
of the 3M Corporation 14 and Union Carbide Corporation 15 in the USA
and in the Plastics Division of ICI 16 UK. All three companies now
market different polysulphone plastics and the structures of these materials
and their Tgs are shown in Table 6.2.
Poly(phenylene sulphone) itself decomposes before melting, at > 500°C,
and the improvement in tractability of the polymers has been brought
about by the introduction of the aryl ether linkages. As can be seen from
the table, the more there are of these, the lower the Tg of the polymer.
There are two synthetic routes to these polymers:
(I} Reaction of a dihalo aryl compound with a diphenoxide of an alkali
metal. 15
Hal--@-S02-@-Hal + MO-Ar-OM -+
-f@-S02-@-O-Ar-of- + 2MHal
The sulphone group plays an essential part in the reaction as it acti-

vates the atoms to attack by the phenoxides.
(2) A polysulphonylation process in which sulphone linkages are
formed by reaction of arylsulphonyl chlorides with aromatic nuclei
H-Ar-H + ClS0 2--Ar'-SOzCI-+ [Ar-SO z-Ar'-SOzl + 2HCl
The reaction is performed in the presence of catalytic amounts of a
Friedel-Crafts reagent such as FeCI 3 , SbCl s or InCI 3 . z,17 The ArH2
can be a diphenyl, diphenyl ether or naphthalene but not a benzophenone
or a diphenyl sulphone as the electron withdrawing substituents reduce
the ability of the ring to provide an electron pair for bond formation.
The polymensation can be carried out in the melt but it IS preferable
VI
tv
TABLE 6.2
POL Y(ARYLENE ETHER SULPHONES)
Polymer Structure Tg °C
Astrel 360 (Carborundum Co.) Mainly --f-@-@-S02! 285
also some -f@-a-@-S02! o

:-
~
z
6
:c
..,
Udel PI700 (Union Carbide) 190
-{©Ji~o>-a-@-so,-@-+
Radel (Union Carbide) Not known 204"
Poly(ether sulphone) 200P (ICI Ltd) 230

t@-a-@-S02!
" Heat deflection temperature at 1·82 MPa, ASTM 0648.
to use an inert solvent. Melt polymerisation requires very high tempera-

tures to provide sufficient mobility for the growing polymer chains. The
preferred solvents are nitrobenzene or dimethyl sulphone; these are capable
of dissolving both the starting materials and the resultant polymer. They
are also compatible with the strong Lewis acid catalysts and stable to
electro phi lie attack by the sulphonyl chloride.
6.2.3. Thermal Stability of the Polysulphones

The results of thermogravimetric analysis of the polysulphones are illus-
trated in Figs. 6.3 to 6.6; the experiments were performed in air and
in nitrogen at a rate of temperature rise of 2°Cjmin. In addition a plot
of the first derivative is shown. It can be readily seen that all the polymers
show very high thermal and thermo-oxidative stability, the majority of
the degradation occurring between 450°C and 550 °C in each case. The
0 ,,
\
\
\
\0
\
\
\
20 \
I
I 015
I
I
I
\
.
c
I
\
I
.....
~ 40 \
\
\
0 -
....
....... "
o 10
"
.<!
'" c
; 60 " -.e
e'"
...
>-
/\
I
I \ 005
80 ,
b I
,,
I
b
I
I
I
," .- ... ----
.-
,,
100 0
300 400 500 600
Icmpcraturc·C
FIG. 6.3. Astrel 360-{a) weight loss and (b) rate of weight loss (dY). - - in
air; ---- in nitrogen.
0, ............ o ~ _______ ~ Vl
.j::..
I \
20 20 \ \
o 15 015
\ \
\
\
c c \
u
u
u
u
; 40 ~ 40
\ \
Q.
c.
~
~
" ,
....
~
o o 0
010 -- ......... ~ 0.10
., :-
~ ~
""'" c
; 60 = ; 60 E
Z
........ a co Ci
........... E 'E" ::c:
" .... j~ >-
-
>-
'0
...,
'" "0
005 0.05
80 80
j
------~ -,-'
100
~--I . 10 100 1 ~ \\ I
300 I a
300 400 500 600 400 500 600
temptratu .. ·C temperatur.·C
FIG. 6.4. Udel P1700--(a) weight loss and (b) rate of FIG. 6.5. Radel-(a) weight loss and (b) rate of weight
weight loss (dY). ~~- in air; ---- in nitrogen. loss (dY). ~-- in air; ---- in llltrogen.
Or-------~~------------------~
20
....
~
c
.....
.. .. 0
~
~
0
o 10
....
~
.c:
, c
=- 60
,
"
--- --..,...
e
"
E
>-
b
0.05
80
100 '----~----'-"""""''--'-----'----~---"-~--'--'O
300 .. 00 500 600
lemptralurc"C
FIG. 6.6. Poly(ether sulphone) 200P-(a) weight loss and (b) rate of weight loss
(dY). - - III air; ---- in nitrogen.
dY (rate of weight loss) plots show very clearly that in air the degradation
is a multistage process, in contrast, in nitrogen there is essentially a
single stage reaction with the formation of a fairly stable char representing
about 40 % of the starting material. The derivative plots also indicate
that the primary mode of breakdown is thermal rather than oxidative.
The activation energy for thermal degradation of Udel P1700 in an
argon atmosphere has been given as 293 kJ/moleY As a result of iso-
thermal experiments the values shown in Table 6.3 have been evaluated. 19
It has been shown 18 ,19 that the primary thermal breakdown route
of poly(arylene ether sulphones) in vacuo is by cleavage of the C-S bonds
to release S02 gas with a concomitant crosslinking process within the
polymer.
As a class the polysulphones are very stable chemically and they all
156 G. J. KNIGHT
show good stability with acids and alkalis, in addition to the good oxidative
stability described above. Resistance to hydrolysis is high but hot water
promotes stress cracking. Solvent resistance is good towards the non-polar
hydrocarbons, alcohols, Freons, silicone oils etc., but chlorinated hydro-
carbons, ketones and other polar solvents tend to swell or, in specific
cases, dissolve the polymers. Unfilled material has been shown to have
TABLE 6.3
ACTIVATION ENERGY E. FOR DEGRADATION IN AIR
AND NITROGEN
Sample Ea kJlmole
Air Nitrogen
Astrel360 172 ± 10 270 ± 12

Udel P1700 137 ± 11 270 ± 12
Radel 78 ± 4 310 ±4
Poly(ether sulphone) 200P 183 ± 10 311 ±8
poor resistance to ultraviolet radiation,21 - 24 but with high pigment load-

ings the weathering stability can be improved to an acceptable level
for outdoor applications. 17
A summary of the physical properties of the polymer is shown in
Table 6.4.
The polysulphones have been shown to have outstandingly good creep
properties. For example, a creep specimen of poly(ether sulphone) at
a stress of 50 MPa endured for more than two years, developed only
slight crazing and showed no inclination towards runaway creep.25 The
combustion characteristics and low smoke emission of these plastics make
them suitable, for example, for applications in electrical apparatus and
in the interior of aircraft.
When used in combination with glass or carbon fibres these materials
offer many attractive properties. Much effort has gone into developing
them in industrial, automotive and aerospace applications,26 e.g. in the
space shuttle 27 and the YC-14 aircraft. 28 Unfortunately the susceptibility
of the resins to attack by halogenated solvents, for example dichloro-
methane which is commonly used as a paint stripper, would seem, at
present, to preclude their use in aircraft structures. 10 Research is being
conducted on modification of the chemical structure of the resin to reduce
the effect of solvents. 28
:r:
Ci
:r:
TABLE 6.4 ~
3:
PHYSICAL PROPERTIES OF POL Y(ARYLENE ETHER SULPHONES) ;g
:;.:l
Properties Unit A~trel 360 Udel Radel 200P >
~
Density g/cm 3 1-36 1-24 1-29 1-37 ~
'"0
Tensile strength MPa 91 70 72 84
Elongation 13 50 to 100 7 30 to 80 ~
'"0
% m
Tensile modulus GPa 2-60 2-60 2-14 2A4 :;.:l
~
Flexural strength MPa 121 108 86 129
Flexural modulus GPa 2-78 2-60 2-28 2-57 ffi
RockwelI hardness MlIO M69, RI20 M88 Sil
Volume resistivity ohm _cm 3 x 10 16 5 X 10 16 9 X10 16 10 17 ~
Power factor at 10 6 Hz 0-0035 0-0010 0-0076 0-0035 ~

:;.:l
Service temperature °C 260 150 to 170 175 200 3:
Heat distortion temperature at 1-82 MPa (264 psi) °C 274 174 204 202 >
t""
t""
Coefficient of linear expansion cm/cm/oC 4-7 x 10- 5 5Ax 10- 5 5-5 x 10- 5 5-5 X 10- 5 -<
Flammability ASTM 0635 Self Self ~
extmguishing extmguishing '">
0::1
Water absorption (23°C, 24 h) % 1-8 0-02 1-I OA3 t""
m
~
Z
'"
Vl
--
-.J
158 G. J. KNIGHT
6.2.4. Poly(Phenylene Sulphide)

Poly(phenylene sulphide) resins were reported as early as 1948. 29 The
polymer being obtained by heating either an alkali metal sulphide or
an alkali metal salt and sulphur mixture with a dihalo-arylene compound.
The proposed reactions are:
+ Na 2 S2 0 3 + 4NaCI + 3C0 2 (1)
3Cl-@-Cl + 4Na 2 C0 3 + 4S -+
(2)
The product was described as a thermoplastic of high thermal stability,

soluble in molten sulphur and with a molecular weight greater than
9000. Phillips Petroleum Company developed a large scale synthetic
method of condensing para-dichlorobenzene with sodium sulphide 30 and
were able to offer trial quantities of the Ryton plastic in 1968.
The following characteristic properties are quoted for the material:
high strength, toughness and rigidity, high service temperature, good
resistance to chemicals, hydrolysis and oxidation, low water absorption,
good processability, low shrinkage, good dimensional stability and good
fire performance.
An illustration of the high thermal stability is given in Fig. 6.7 where
the TGA curves are shown for experiments conducted in air and in
nitrogen. From this it can be seen that the initial reactions causing weight
loss are purely thermal and that oxygen causes a certain amount of
crosslinking to take place. It has been shown that the thermal stability
can be increased by a form of post cure which causes branching and
chain extension to take place. 31 Heating in vacuo at 250-260 DC showed
that the volatile degradation products were dimeric and trimeric chain
fragments, dibenzothiophene and possibly thianthrene. 32 The polymer
shows excellent resistance to high energy radiation. 33
:r:
Ci
:r:
~
~
TABLE 6.5 ;:g
::<:'
PHYSICAL PROPERTIES OF POLY(PHENYLENE SULPHIDE) AND ITS COMPOSITES WITH GLASS AND CARBON FIBRE >
-l
C
Property ASTM Method Unlt5 Poly(phenylene With 30% With 30% 1:;
sulphide) glass fibre carbon '1:1
::<:'
fibre o
;:g
::<:'
Oensity g/cm 3 1·34 1·56 1-45 -l
Water absorptIOn in 24 h 0570 % 0·2 0·04 0·04 m
Vl
Tensile strength 0638 MPa 74 138 186
~
Tensile elongation % 3-4 3-4 2-3 -l
Flexural strength 0790 MPa 138 200 234 :r:
tTl
Flexural modulus 0790 GPa 4·13 11·02 16·88 ::<:'
~
Heat distortion temperature 0648 °C 137 260 260 >
Coefficient of thermal expansIOn cm/cmj"C x 10 ~ 5 5-4 2·34 1·08 r
r
Surface resistivity ohm/em 10 16 10 16 1-3 ...:
Vl
Flammability UL Subj. 94 94 v-a 94 v-a 94 v-a ~
t:=
r
tTl
::<:'
tTl
Vl
Z
Vl
Vl
\0
160 G.1. KNIGHT
0F=~=-----------------~
20
...,
~
.....
L 40
..
0 o 10
~
~
;; 60'
c
E
005"';;.
E
>-
80 '"0
100 L_--~-----=-,:-::--------=;;==:::Io
100 400 500 600
temperature ·C
FIG. 6.7. Poly(phenylene sulphideHa) weight loss and (b) rate of weight loss
(dY). - - in air; ---- in nitrogen.
The resin adheres well to filler materials 33 and both glass and carbon
fibre reinforced poly(phenylene sulphide) exhibit good mechanical proper-
ties.5.26.33 Examples of typical properties are listed in Table 6.5.
6.3. THERMOSETS
6.3.1. Introduction
The thermoset resins consist of three dimensional crosslinked networks
that do not show softening or melting points, hence there are more
of them, compared with thermoplastics, that can be used at elevated
temperatures. This section deals with those materials that show potential
at temperatures of 200 °C or higher; the Friedel-Crafts resins are not
included in this chapter as they are dealt with in Chapter 4.
6.3.2. Phenol-Formaldehyde Resins

The phenol-formaldehyde resins were the first wholly synthetic polymers
to be utilised, first going into production in 1910. Many papers and
books have been written on the subject so it will be mentioned only
briefly here. The formation of resinous products by the reaction of phenols
and aldehydes has been known for a long time, and was studied especially
by Baeyer 34 in 1872. There was, however, little interest in resins at that
time. The classical work of Baekeland 35 opened the way to the large
scale manufacture of articles made from these resins.
The preparation of phenol-formaldehyde resins is described in
Chapter 1. Briefly, the resins are prepared by the reaction of phenol
(or a mixture of phenols) with formaldehyde. Cresols, xylenols and
resorcinol are also used, but to a much lesser extent than phenol itself.
Several types of low molecular weight prepolymer are produced commer-
cially, but these can conveniently be divided into the so called resols
and novolacs.
Resols are produced when phenol and excess formaldehyde are reacted
under alkaline conditions to give a complex mixture of mono and poly-
nuclear phenols with methylol substituent groups.
The novo lacs are produced when formaldehyde and excess phenol
are reacted under acid conditions. This results in polynuclear phenols
linked by methylene groups, as shown:
OH OH OH
© @-CH'
1
-@ +HCHO
OH OH
~H'~H'OH ~ etc.
Thus novo lacs are mixtures of isomeric phenols of various chain lengths
but with an average of 5-6 benzene rings/molecule and a range of from
2 to 13 as shown by molecular weight determination. Some unreacted
phenol and water are also usually present. Unlike resols, novolacs contain
no reactive methylol groups and, therefore, do not condense on heating.
In order to obtain cured material they need a suitable crosslinking agent,
such as hexamethylene tetramine or paraformaldehyde. The mechanism
of the crosslinking process is not fully understood but the final structure
162 G. J. KNIGHT
is not very different from that obtained with the resols. In both crossJinking
reactions volatiles are evolved and, therefore, all manipulations have
to be performed under pressure to prevent the formation of voids.
6.3.3. Thermal Degradation of Phenol-Formaldehyde Resins
Pyrolysis of phenolic resins always gives higher homologues of the phenols
originally used. 36 37 The probable decomposition products of a novo lac
are:
OH : OH
@-CH2:-©r:CH2-©t\CH2-©tCH'
I : :
o-cresol
-OH
I
I
phenol
I
I
,
p-<:resol:
I
,
,
2,4-xylenol
OH, I
I
~~'-©-OH
2,4' -dlhydroxydiphenylmethane
It has been shown that when heated to 400 DC in air the initial degradation
t
reactions are: 38 . 39
OH OH OH OH
-t"'OH HOCH, + ~ --tH' ~H'1~

I/',.~OHI ~/',.
-rb
-l-
OH
[0]
+
T
ioI-~
~
A
~-9
OH
-CH 2
OH
-C$J--
I
"r ~ !
~ ~~
OH 0
I II
-~--C-OH
~
Or-~~----~--------------------.
20
..
c
.~ 40
Q.
0.1
c
E
o 05~
,.....,
80
FIG. 6.8. Phenol-formaldehyde resin-(a) weight loss and (b) rate of weight loss
(dY). - - in air; ---- In mtrogen.
An illustration of the TGA of a phenol-formaldehyde resin in air

and nitrogen is shown in Fig. 6.8. This shows the good thermal stability
of the resin. Electrical properties are not very good because of the presence
of polar groups, and the resins show relatively poor tracking resistance
under conditions of high humidity. In respect of chemical resistance
they are practically unaffected by water or organic solvents. They are
resistant to all except strong oxidising acids, but are readily attacked
by alkalis. Resins based on cresols have better acid and alkali resistance.
Some typical properties of phenol-formaldehyde composites are given
in Table 6.6.
6.3.4. Melamine-Formaldehyde Resins

The production of melamine-formaldehyde resins began in the mid 1930s,
the first patent being granted in 1935. 40 Melamine is a cyclic trimer
0'\
.j:>.
TABLE 6_6
TYPICAL PROPERTIES OF RESIN GLASS CLOTH LAMINATES
Property Units ASTM Phenol Melamine Silicone Furan u

test method formaldehyde formaldehyde
Laminating temperature DC 135-175 135-150 160-250 RT-80

Laminating pressure MPa 0-1-14 7-13 0-2-14 o upwards Cl
Tensile strength MPa D638 60-350 175-500 70-265 80
Tensile modulus GPa D638 8-17 14-17 10-14 6 :-
;>:
Flexural strength MPa D790 110-560 245-595 70-265 153 z
Compressive strength MPa D695 235-520 170-590 170-320 60 Ci
Impact strength Izod kJjm 2 D256 20-95 25-80 25-70 :I:
...,
Water absorption in 24 h % D570 0-12-2-70 0-20-2-50 0-07-0-65 1-3
Heat resistance, continuous DC 120-260 150 200-370 200-250
Dielectric strength 3 mm thickness kVjmm Dl49 12-28 8-24 7-19 4-7
Dielectric constant at 1 MHz Dl50 3-7-6-6 6-0-9-0 3-7--4-3
Dissipation factor at I MHz Dl50 0-005-0-050 0-011-0-025 0-005-0-010
Arc resistance s D495 Tracks 175-200 150-250
a Chopped strand glass cloth laminate_

of cyanamide. Like urea it forms colourless resins with formaldehyde,

but the melamine resins are superior to those from urea in heat and
water resistance.
The melamine reacts with formaldehyde under slightly alkaline condi-
tions giving methylol derivatives with up to six methylol groups/molecule.
NHCH 2 0H
I
C
/~
+HCHO N N +excess HCHO
-----+- I I )
C C
H2N
/ *'\N/ ""NH2
H OCH 2 -N---CH 2 OH
I
C
/~
N N
I I
HOCH 2 C C CH 2 0H
~/'\/~/
N N N
/ ~
HOCH 2 CH 2 0H
The presence of all possible methylolmelamines in a reaction mixture
has been demonstrated. 41
On heating, methylolmelamines condense to form resinous products.
The rate of resinification is strongly dependent on pH, but for practical
purposes sufficient crosslinking occurs under the action of heat alone.
At temperatures of 150°C the crosslinking takes place by the reaction
of the methylol groups, water and formaldehyde being evolved. A cured
resin may be represented as
166 G. J. KNIGHT
There will be some reacted methylol groups, some methylene and some
ether linkages. All three groups are present in a normal cured resin though
not necessarily in equal amounts; it is generally assumed that ether links
predominate over methylene links.
6.3.5. Thermal Stability of Melamine-Formaldehyde Resins

The methylol groups have been shown to decompose at 140 0c. 42 •43 form-
aldehyde being given off. The next bond to break is probably the ether
bond, followed by loss of formaldehyde and the formation of methylene
bridges. 43 Under inert atmosphere it is believed that the methylene bridges
do not break down until 380 °C is reached; the triazine ring itself is
stable up to 400°c.43 This means in practice that so long as inert fillers
are used and the material is carefully post cured at temperatures up
to 200°C, then the melamine-formaldehyde resins do have a useful service
o~--~------------------------~
20
...,
<: I a
I
\
,
~ 40 06
Co. " ....a
\
\
\
'"...
\
. 60
\
\
04
~
\ \ <:
\ e
~
E
..,
>-
80 0.2
I
'I
100e:=~~~~~~::::::--~:=~~--~0
200 400 600 800
Itmper5lure ·C
FIG. 6.9. Melamine resm-(a) weight loss and (b) rate of weight loss (dY).
- - in air; ---- in nitrogen.
life at elevated temperatures. The TGA curves for a melamine resm

cured at 200 °C are illustrated in Fig. 6.9; this shows that the resin
breaks down in a similar manner in air or in nitrogen, initial weight
loss occurring at about 180°C in air and 210°C in nitrogen. It can
be clearly seen that there are three phases to the degradation probably
corresponding to the breakdown of ether links between 200 and 300°C,
methylene links at 370°C to 380°C, followed by the residual char decom-
posing between 400 and 700°C. Some typical properties of these resins
are listed in Table 6.6. Laminated with glass cloth, they have been used
as arc barriers, switchboard panels, circuit-breaker parts and for other
high duty electrical purposes.
6.3.6. Silicone Resins

The polyorganosiloxanes are polymers in which the backbone is comprised
of alternate silicon and oxygen atoms, the silicon atoms being also joined
to organic groups.44.45 A representation of the structure would be
!
I
R 0
I I
~-Si-O-Si-C>---
6 ~
I ~
~-Si-O-Si-O-Si~-
~
I I I
R R 0
I
l
The resins are manufactured by first formulating an appropriate blend
of mono ethyl, dimethyl, monophenyl, diphenyl, methylphenyl, monovinyl
and methylvinylchlorosilanes with silicon tetrachloride. The final proper-
ties of the resin depend as much on the processing and cure conditions
as on the original composition, but it is possible to generalise and say
that trichlorosilanes produce hard resins, immiscible in other organic
polymers, whereas dichlorosilanes increase softness and flexibility and
phenylsilanes give resins that are more miscible in organic polymers,
that are less brittle and have superior thermal resistance. The chlorosilane
blend, commonly dissolved in inert solvents which modify the rate of
reaction, is then mixed with water. Because of the difference in the rates
168 G. 1. KNIGHT
of hydrolysis of the various chlorosilanes, problems can arise in producing

polymers incorporating a proper balance of all the constituents. It is
possible to control this by the choice of reaction solvent or, if necessary,
by sequential addition of the chlorosilanes. After hydrolysis the resin
incorporates a certain proportion of silanol groups, these serve to initiate
the final cure of the resin to a fully crosslinked structure. This final
cure is achieved by processing at elevated temperatures either with small
amounts of residual acid left in the resin as catalyst or by incorporation
of metal soaps such as tin or zinc octo ate, cobalt naphthenate or amines
such as triethanolamine.
6.3.7. Thermal Stability of Silicone Resins

Detailed reviews of the thermal stability of the silicones have been
written. 46 .47 Overall the silicones can be regarded as providing materials
0.-------------------------------.
20
------ ----
004
b"
I
--...\
\
c
E
, " -' o 02~

80 \
\ , .... , _
I >- ..,
---
100 '-----4....L
O-=-0--~---5::-'0L.0--~---,-60.L0,---~----=-'70 0
temperature ·C
FIG. 6.10. Silicone reSIn MS 84(}-(a) weight loss and (b) rate of weight loss (dY).
- - In air; ---- in nitrogen.
with good high-temperature resistance as illustrated in Fig. 6.10, the

TGA curves for a silicone resin in air and in nitrogen. It is known
that branched-chain polysiloxanes are stable at temperatures in excess
of 400°C and tetramethylsilane vapour is stable to above 600°C showing
the stability of the Si-C bond in non-oxidative conditions. Oxidative
attack takes place on the organic substituents, for example the methyl
groups giving rise to evolution of formaldehyde, water, carbon monoxide
and crosslinking through siloxane bridges. Thus the resin finally forms
a silica residue. The incorporation of phenyl groups improves both the
thermal and oxidative stability of a resin.
The resins are used as coatings, for laminating and, to a lesser extent,
as moulding compounds. Glass fibre laminates are used where the com-
bined properties of good electrical insulation and heat resistance are
required. The mechanical properties of the laminates, given in Table 6.6,
are inferior to those of epoxy or polyester resins. In sealed equipment
the gradual loss of volatiles from these resins can cause build up of
non-conducting silicone deposits on contacts or commutator brushes,
but this has been partially overcome by a special design of the brushes.
6.3.8. Furan Resins

The basic chemistry and properties of the early systems were reviewed
by Dunlop and Peters 48 in 1953. The intermediate used in the production
of furan resins are derived entirely from natural resources by the steam
distillation of such food by-products as corn cobs and oat husks. The
furfural and furfuryl alcohol so produced may be resinified by the action
of heat in the presence of an acid catalyst. The reaction is a condensation
one with the elimination of water to form a linear polymer.
n n
~OY-- CH 2 OH + ~OY--C H 2 OH --+!f1
~O;>- CH 2~Orc H 2OH--+n
n r~H
~~ 2~O/j;
f/ll CH n
2~O)\-C 2
H OH
The polymer is normally neutralised and further diluted with monomeric

furfuryl alcohol or furfural to give a low viscosity resin suitable for
use by the fabricator. Once stabilised the resins may be stored many
months without change in activity or viscosity.
The final reaction stage, again in the presence of an acid catalyst,
170 G.1. KNIGHT
produces a highly crosslinked infusible resin (the mechanism is uncertain).

Control of this reaction by suitable catalysts led to the development
of laminates with good mechanical properties. 49 - 5 3
The densely crosslinked structure and absence of polar groups gives
the resins a good resistance to all chemicals except concentrated sulphuric
and hydrofluoric acids, strong caustic soda and oxidising media. In
addition they have high heat distortion temperatures and good thermal
stability. They are also suitable for applications where fire retardancy
and low smoke generation are required. On heating, the furan resins
form a very stable char with little loss of volatile material. For example,
weight loss in air starts at 300°C, 10 % weight loss occurs at 380 °C
and 50 %weight loss occurs at 530°C. 51 Some examples of the properties
of glass fibre laminates are given in Table 6.6 and an illustration of
the strength retention at temperature is given in Fig. 6.11.
80
.......
o
.t 40 flexural strength
20
0L-~-----I~OO----------2~0~O----~--~300
Irmprrature·C
FIG. 6.11. Property retention of Furan resin chopped glass fibre laminates with
varying temperature.
HIGH- TEMPERA TURE PROPERTIES OF THERMALLY STABLE RESINS 171
6.3.9. Polyimide Resins-Introduction

The polyimides comprise the most thermally stable family of organic
resins currently made. They are available in several forms suitable for
different purposes. Films and varnishes for electrical uses were introduced
commercially in 1961, to be followed later by moulding powders, lamina-
ting resins and metal-to-metal adhesives. Fibres and foams have also
been evaluated.
The aromatic imide structure may be represented as
o
II
C
\
N-Ar
/
C
II
o n
and it is very stable both thermally and oxidatively. Polymers of this

basic structure have very high melting points, often above their decomposi-
tion temperatures, and the history of the polyimides is the development
of processable resin systems.
6.3.10. Condensation Type Polyimides

The first polyimides to be developed were prepared from aliphatic diamines
by a melt fusion of the salt formed from the diamine and tetra-acid,
or diamine and diacid/diester. 54
)§(
HOOC COOCH 3
+ NH,(CH,),NH, --+ Salt--+
H3COOC COOH
diacid--diester diamine
polyimide
For preparation of high polymers by these procedures, and for shaping,

172 G. J. KNIGHT
the polyimide had to be fusible. As a result, the synthesis of useful

polypyromellitimides required the use of aliphatic diamines that possessed
nine or more carbons in a normal chain, or seven in a branched chain. 55
These long chain aliphatic groups make the polymer susceptible to thermo-
oxidative degradation.
The method of preparation of fully aromatic polyimides was developed
about twenty years ago. 56- 581t involves the synthesis of a soluble polyamic
acid precursor and a final cyclisation step to the required polyimide.
The dianhydride and diamine are reacted at ambient temperatures
in polar solvents such as dimethyl formamide, dimethyl acetamide or
N-methyl-2-pyrrolidone. The cyclisation step can be accomplished by
heating or by a suitable chemical treatment.
o o
CII C"
0(:JQr. )0 +
C C
H,NAr· NH,
oII 0"
dianhydride 1
dlamme
o 0
" "
HN-~C-OH
HO---C C-NH-Ar
oII 0"
1heat
or
chemical reagents
o 0
<JQ(>-Ar
C C
" "
C C
o 0
" "
Pure reagents and solvents are necessary for the synthesis of high
molecular weight polymers and the ratio of reactants used and the method
of mixing are also critical factors. 59 .6o The intermediate polyamic acid
solutions are unstable, necessitating storage under dry refrigerated con-
ditions. The polyimides finally produced are usually insoluble and infusible
and since the temperatures for fusion of the intermediate and the cyclisation
reaction are similar, precipitation effects markedly influence the flow
characteristics and make for difficulties in moulding or laminating. Matters
are not helped by the evolution of water during the process. All these
factors limit the size and thickness of articles which can be made. Never-
theless, despite these disadvantages the condensation type polyimides
are being used very successfully. Du Pont's Pyralin and Monsanto's
Skybond resins are examples of the condensation type of polyimide.
The exact constitution of these resins is not known but essentially they
consist of the polyamic acids derived from either pyromellitic dianhydride,
benzophenone tetracarboxylic dian hydride or diesters of these acids with
diamines such as 4,4' -diaminodiphenyl methane or diaminodiphenyl ether.
By using the esters the stability of the polyamic acid intermediate is
improved, giving better shelf lives for the resin solutions at room tempera-
ture. The cure reaction is also slower giving better resin flow and allowing
the use of vacuum bag autoclave rather than the comparatively expensive
heated press. However, the volatiles generated during the cyclisation
reaction, ethanol or methanol, can give rise to voids in the final product.
6.3.11. Thermal Degradation of Condensation Polyimides

The condensation polyimides are very stable as can be seen from Figs.
6.12 and 6.13, the weight loss curves in air and nitrogen for Kapton
film and Skybond 700. These polymers are thought to have the following
structures:
o 0
I I
<)gt;- @o
c c
I I
o 0 n
Kapton film from Du Pont

174 G. J. KNIGHT
o 0
I 0 II
c I C
~-©Cl-@-O
II I
o 0 n
Skybond 700 from Monsanto
The figures show that the polymers are equally stable in air or in
nitrogen, weight loss not occurring until over 450°C, the dY plots indicate
that the maximum rate of weight loss occurs at about 570°C. Pyrolysis
experiments in vacuum have shown that the principal volatile decomposi-
tion products from these structures are carbon monoxide 59 %and carbon
o __
---- ............. a
"-
\
a \
\
\
\
\
\
20 \
\
,,
,
" '
c., ...
................ , ....
.....
u
.. 40
0.2
...
.c
c
£
O.l~
80 ...>-
--------
100 E:-=-:::-=~~:=:=--==--=-::'-::=::::~~_~:;-:-Ib~~-~70~
FIG. 6.12. Kapton film--{a) weight loss and (b) rate of weight loss (dY). - - in
O~==~~~------------------I
20
.
~
c
....
u
... 40
0.2
.....e
80 0.1 E
...
>-
100 '-----:4~0-:-0--~--~:7"'--~---:C607:0'---~~~70~~
t e mPfrQ t U r e·C
FIG.6.13. Skybond 700-(a) weight loss and (b) rate of weight loss (dY). - - - - in
air; ---- in mtrogen.
dioxide 35 %.61- 63 Pyrolysis of model compounds supports the view that

the carbon monoxide derives from the decomposition of the imide ring,
and the carbon dioxide is formed from the isoimide structure resulting
from a rearrangement reaction. 64 . 65 Other experiments have shown that
in air the reactions involved in weight loss are accelerated by moisture 66
and it has also been shown that considerable crosslinking and degradation
takes place in the Kapton film before any weight loss is observed. 67 . 68
In fact after five hours at 400°C, and only 2-3% weight loss, 50 %
of the diphenyl ether units had reacted and 30 % of the pyromellitimide
units had undergone modification. 69
6.3.12. Processing and Properties of Condensation Polyimides

The processing conditions recommended for producing glass cloth
laminates of Monsanto's Skybond 703 are shown in Table 6.7 and examples
176 G. J. KNIGHT
TABLE 6.7
TYPICAL PRESS CURE AND POST CURE CYCLE FOR
SKYBOND 703 70
B-stage prepreg at 121-149 DC

Preheat press to 204 DC
Prepreg into press-kiss contact for I min
Increase pressure to 1·72 MPa, leave for 30 min
Cool in press to 65 DC
Post cure to obtain optimum hot strength
raise to 204 OC hold for 4 h
raise to 260 DC hold for 4 h
raise to 315 DC hold for 4h
raise to 343 DC hold for 4 h
of typical properties and property retention at elevated temperatures

are shown in Table 6.8.
It will be realised that at the recommended post cure temperatures
a certain amount of oxidative crosslinking is taking place. Work has
also been done to prepare carbon and boron fibre laminates of these
resins and good retention of properties has been shown at temperatures
up to 315°C, although for long term use the maximum temperature
is nearer 250 DC. 71 - 74
TABLE 6.8
TYPICAL PROPERTIES OF POLYIMIDEj
GLASS CLOTH LAMINATE AS PRE-
PARED IN TABLE 6.7
Resin content 30-35 %

Void content - 5%
Flexural strength
at room temperature-489-545 MPa
after 500 h at 260 DC"-455 M Pa
after 1000h at 260 DC"-407MPa
after 100 hat 315 DC"-427 MPa
after 100h at 343 DC"_I 72 MPa
after 100h at 371 DC"_ 48MPa
" Values given for elevated temperature

ageing are the values that were obtained
on the test specimens at the exposure
temperature.
6.3.13. Addition Type Polyimides: Polyaminobismaleimides

One method of incorporating the imide structure into a polymer without
the problems caused by volatile evolution is to make use of the activated
maleimide double bonds. For example, l,4-dimaleimidobenzene when
irradiated at 240°C yields a highly crosslinked polymer structure. 75
heat
~
hy
Di(4-maleimidophenyl) methane will form a polymer in the same way

when heated. 76 Being so highly crosslinked these materials are very intract-
able and are very brittle. It is also known that amines will react with
a maleimide group. 77 By combining these two reactions workers at Rhone-
Poulenc have shown that a proper ratio of bismaleimide to aromatic
diamine (e.g. diaminodiphenylmethane, DDM) will give polyimides with
good physical properties as well as good thermal stability.78
O=C-." ~=o
N
I
R
I
l n
n
178 G. J. KNIGHT
These resins are available from Rhone-Poulenc in forms suitable for

moulding (Kinel) or for laminating (Kerimid 601). The latter is supplied
as a yellow powder which is dissolved in N-methylpyrrolidone for prepar-
ing prepregs. Kerimid 601 is readily soluble in many organic solvents
but an extensive study of polymer-solvent interactions has shown that
many of the solvents produce laminates with poor thermo-oxidative
stability, for example 1,2-dichloroethane affords laminates that blow apart
on post cure. 79 For ease of processing, prepregs are required with drape
and tack, this means leaving 10 % or more solvent in the resin, which
can cause considerable degradation in final laminate properties. The exact
reasons for this are not clear from the literature but it is possibly a
combination of several factors; the formation of voids in the composite
by the volatile solvent, the incorporation of the solvent into the polymer
structure during the cure reaction and the plasticisation effect of the
excess solvent left in the laminate. The possibility of solvent-polymer
interaction has been noted with respect to the effect of dimethylformamide
on poly-N,N'-(4,4'-diphenyl ether)pyromellitimide. 66
6.3.14. Thermal Degradation of Kerimid 601

An illustration of the thermal stability of the resin is shown in Fig. 6.14,
the TGA curves in air and in nitrogen. Initial breakdown occurs at
about 300°C in air and at 320 °C in nitrogen, 10% weight loss being
at 370°C in air and at 385 °C in nitrogen. The dY plots show that the
degradation is a two stage process in air with the first peak coinciding
with the nitrogen curve at 410°C. These values show how the changes
in structure of the polymer considerably reduce the overall thermal stability
when compared to the fully aromatic condensation type polximides.
6.3.15. Processing and Properties of Kerimid 601

An example of the suggested processing conditions necessary to produce
glass cloth laminates is shown in Table 6.9 and some typical mechanical
properties of such laminates are shown in Table 6.10.
The strength retention at temperature is illustrated in Fig. 6.15 which
shows that the polymer has good strength retention at temperatures
up to 250 o e, but that for long-term use 180 to 200 e is the upper
0
limit.
6.3.16. Kerimid 353

In the search for easier processing a very interesting resin system is
that developed by Technochemie GmbH and marketed by Rhone-Poulenc
o~--------~-----------------.
20
\
\
,,
."
C ,
,
....
L 40 "
'0
~
0.10
~
0
~
.,.
; 60
0.05-;;.
=
E
E
80
...>-
-'" _---b
100 0
200 400 600
temptrolureOC
FIG.6.14. Kenmid 601---(a) weight loss and (b) rate of weight loss (dY). - - - III
TABLE 6.9
PROCESSING CONDITIONS FOR KERIMID 601/GLASS CLOTH LAMINATES 80
Preheat press to 120 DC

Place prepreg in press and pressurise to 1'45-5'52 MPa
Increase temperature progressively to 180 DC over 30 min
Cure for I hat 180 DC
(The laminate may be removed hot from the press)
To obtain maximum mechanical properties a post cure of 24h at 250 DC or 48h
at 200 DC is recommended
TABLE 6.10
SOME TYPICAL MECHANICAL PROPERTIES OF
KERIMID 601/GLASS CLOTH LAMINATES 80
Property ASTM test method Result
Flexural strength 0790

at 25°C 480MPa
at 200°C 414MPa
at 250°C 345 MPa
Flexural modulus D790
at 25°C 27-5GPa
at 200°C 26·2 GPa
at 250°C 22·1 GPa
Tensile strength D638
at 25°C 345MPa
Compressive strength D695
at 25°C 345 MPa
ISO·C
.
_ 80
c
u
c
o
:; 60
.....
.c
..
-;; 40
.....
20
2000 4000 6000 SOOO 10000

hour 5
FIG. 6.15. Flexural strength retention of Kerimid 601 glass cloth laminates aged
at 180, 200, 220 and 250°C.
as Kerimid 353. It is reported to consist of a mixture of three bismale-

imides. 81 These form a eutectic mixture with an overall melting point
of 70-125°C. This is well below the curing temperature of 200-240°C
and means that prepreg can be prepared by means of a filament winding
technique using molten resin. As solvent is not used there are no problems
with removing solvent during lamination. The prepreg is stiff at room
temperature but acquires tack and drape if warmed to 50°C.
o 0
I I
C C
(
C
~N-@-cH2-@-N~
C
J m.p. 154-156°C
~ »
o 0
di( 4-maleimodophenyl)methane
o 0
I ~
C C
[)N--TcY-- N(
C ~£",ll C
J m.p.I72-174°C
\I CH 3 I
o 0
2,4-blsmaleimido-toluene
o 0
II CH CH II
C\ I 3 I 3 IC)
(I IN~H2~~H2~H~H2~H2-N\ m.p.70-130°C
C I C
II CH 3 I
o 0
I ,6-blsmaleimldo-2,2,4-trimethylhexane
6.3.17. Thermal Degradation of Kerimid 353

Figure 6.16 shows the TGA curves in air and in nitrogen of a sample
of cured Kerimid 353 resin and, illustrates the good thermal stability
of the resin. Decomposition starts at about 350°C in air and at about
400 °C in nitrogen, 10 %weight loss being at 420°C in air and at 460 °c
in nitrogen. Unlike Kerimid 601, which is equally stable in air or nitrogen,
181 G. 1. KNIGH1
o~~~--------------------------.
"0\
,,
\
o \
20
,
I
,
\
\
\.
"-
"- ,
"
'" "-
...... - ........... 020
c
E
010 ~
>-
."
80
---------
L---4~0-0--~---5~0-0--~----60~0--~----J70g
lempera!ure'C
FIG.6.16. Kerimid 353-(a) weight loss and (b) rate of weight loss (dY). - - in
there is a marked decrease in thermo-oxidative stability. This difference

is not altogether surprising considering the incorporation of the aliphatic
bismaleimide. In fact the surprise is that the polymer should he as stahle
as it apparently is. The dY plots also emphasise the difference between
the thermo-oxidative and thermal breakdown, with a three stage reaction
in air and just one period of rapid weight loss in nitrogen before the
development of a stable char.
Thermogravimetric analysis experiments have been reported on the
polymers derived from a series of aliphatic bismaleimides with different
numbers of methylene linking groups, also on several aromatic bismale-
imides, confirming the superior stability of the aromatic compounds. 82
Pyrolysis mass spectrometry was also done on the same set of polymers
showing that the aliphatic polymers decompose by cleavage of the C-C
bonds in the aliphatic bridge, mainly in the vicinity of the N--C bond.
The aromatic polymers decompose in a similar manner to the condensation

type polyimides by breakdown of the imide ring with evolution of CO
and CO 2 . 82
6.3.18. Processing and Properties of Kerimid 353

Descriptions of the preparation of unidirectional glass, Kevlar and carbon
fibre laminates are given in the literature. 83 - 85 An example of the autoclave
cure of the resin is given in Table 6.11 and the properties achieved
by the laminate are shown in Table 6.12.
The values in Table 6.12 illustrate the high strengths that can be achieved
with the resin system and the comparative ease of processing. The property
retention at 250°C is also high but it has been shown that appreciable
degradation of a carbon fibrejKerimid 353 laminate takes place during
isothermal ageing at 260°C in air for 125 hours. It was found that coating
the laminate with aluminium filled polyimide resin improved the resistance
of the laminate to degradation. 84
TABLE 6.11
RECOMMENDED CYCLE FOR AUTOCLA VE LAMINATION OF GLASS FIBRE KERIMID 353
PREP REG
Lay up prepreg at 45-60 oe

Apply full vacuum for lOmin, then heat to lOooe whilst mamtaining vacuum
Release vacuum, heat to 180°C, hold for I h
Apply 0'3MPa pressure and maintain at 180°C for 30min
Apply 0·7 MPa pressure and heat to 2lOoe, hold for 4h
Cool to room temperature
Post cure at 240°C for 15 h
TABLE 6.12
PROPERTIES OF UNIDIRECTIONAL KERIMID 353-E GLASS
COMPOSITE
Property At room temperature At 250°C
Fibre content 60%

Void content <2%
Density (g/cm 3 ) 2·1
Tensile strength (MPa) 1058 1028
Tensile modulus (GPa) 41 38
Flexural strength (MPa) 1215 837
Flexural modulus (GPa) 42 40
Short beam shear (MPa) 76 50
184 G.1. KNIGHT
6.3.19. Other Bismaleimide Polyimides

Other addition type polyimides based on bismaleimides are available
commercially, examples of these are Kerimid 711, another product of
Rhone-Poulenc, and Hexel F178, sold by the Hexel Corporation but
only available in the form of prepreg.
Kerimid 711 has also been designed as a filament winding resin and
consists of a mixture of components that give the resin a very low melting
point of '" 60°C. The TGA of the cured resin is illustrated in Fig. 6.17
which shows the weight loss in air and in nitrogen. The initial weight
loss starts at about 300°C in air and in nitrogen, 10 %weight loss occurring
at 390°C in air and in nitrogen. The lower melting point and the lower
stability when compared to Kerimid 353 suggests that Kerimid 711 might
contain a greater proportion of aliphatic material. The recommended
processing conditions for filament winding are to melt the resin by heating
\
\
20 \
\ a
\
\
\
,,
, ...
40 "
0·10
............
c:
E
o 05~
80
...
>-
10 o'-----~--:-'40:-:0-~----,5~0-=-0-~-~-::---'----'0
FIG. 6.17. Kerimid 71 I-(a) weight loss and (b) rate of weight loss (dY). - - in
TABLE 6.13
PROPERTIES OF CURED KERIMID 711 RESIN 80
Property At room temperature At 200 D

e
Density (g/cm 3 ) 1·3
Glass transition temperature ( DC) 280
Flexural modulus (GPa) 2·84 2·26
Aiter 2000 hours at 200 D
e
Flexural modulus (GPa) 2·75 1·96
to 115 DC, and to wind the impregnated fibres onto a mandrel heated
to 120-150°C. The component is maintained at ISO °C for approximately
I h then the temperature is raised to 200°C over 2 h and cured for 4 h
at 200°C. The whole is then cooled slowly to room temperature. To
achieve maximum mechanical properties a post cure of 15 h at 240 °- 250°C
is recommended. Some typical properties of the cured resin are given
in Table 6.13. From the figures in Table 6.13 it can be seen that 200°C
is probably the upper temperature limit for use.
Hexel Fl78 can be obtained only in prepreg form. The resin is known
to be of the bismaleimide type and it contains triallylisocyanurate. 86
This compound will homo polymerise to give thermally stable materials

and presumably it is also able to take part in the free radical polymerisation
reactions of the bismaleimide. Its low melting point, 24°C, means that
it helps to make the F 178 a solvent free system with good drape but
no tack at room temperature; heating to 60-90°C is sufficient to give
good ply-to-ply tack. If extreme drape and tack are required then very
small amounts of N-methylpyrrolidone can be added. 87 The suggested
186 G. J. KNIGHT
autoclave cure is 30 min at 149 DC and 0-4 MPa minimum pressure. A

post cure is said not to be required. 87 But a post cure of 64 h at 204 DC
is recommended for a carbon fibre reinforced component after an initial
cure of 2 h at 177 DC and 0·6 MPa pressure. 86 The cured resin has a
density of 1·28 g/cm 3 and a glass transition temperature of about 275 DC.
Weight loss experiments at 5 DC/min show 3 % weight loss at 375 DC
and 7% at 400 Dc. 87
6.3.20. Norbornene Systems

An alternative method that has been developed for obtaining addition
type polyimides is the use of norbornene end groups rather than the
maleimides described above. The method of synthesis originally adopted
was the preparation of low molecular weight amic acid prepolymers end
capped with the reactive norbornene rings.88
ax
o 0 0
II II 0 II~r
C- OH HO--{:)g:rllQ{C~NH-Q;-jH2
C-NH~H2 I(5\-HN--{: C-OH
II '\:::5 ~ ~ II II
0 0 0
~NHX:xD
amic acid prepolymer HO--{:
II
o
o
II
XD
C
O-N/
""CII
imide prepolymer o
The postulated crosslinking reaction is considered to involve a reverse
Diels-Alder reaction which leads to the formation of cyclopentadiene
and a maleimide structure. These immediately react to form an adduct

(1), and this adduct then initiates homopolymerisation of the norbornene
species. 89
""'I~:.e(hate
~ctlO~
O=C"'" /C=O
N
I
R (I)
(1)+
a
.n_ ~
-----r
O=C "'" /C=O O=C" /C=O

N N
I I
R R
These polymers were successfully developed by TR W Inc. and marketed

as P13N resin by Ciba-Geigy Corporation. Further developments were
the use of pyromellitic dianhydride in place of benzophenone tetra-
carboxylic dianhydride and the incorporation of 4,4'-diaminodiphenyl-
sulphone in addition to 4,4' -diaminodiphenylmethane. These changes gave
resins with improved long term thermo-oxidative stability, but it was
observed that the prepolymer solutions exhibited a very limited shelf
life at room temperature. 90 In addition, the preferred solvents, dimethyl
formamide or N-methylpyrrolidone pose the same problems during pro-
cessing as they do with the condensation polyimides. Many of these
problems have been overcome by a method of polymerising the monomeric
reactants in situ, i.e. the laminate or moulding being manufactured. The
basic chemistry of these PMR resins is the same as the Pl3N type,
but in this case the fibres are impregnated using a methanol solution
188 G. J. KNIGHT
contaInIng the correct proportions of 4,4' -diaminodiphenylmethane

(DDM), the dimethylester of benzaphenone tetracarboxylic acid (BTDE)
and the monomethyl ester of 5-norbornene-2,3-dicarboxylic acid (NE).
Using the molar proportions three parts DDM, two parts BTDE and
two parts NE the prepolymer formed has a formulated molecular weight
of 1500, hence the designation PMR_15. 91 Other reactant ratios have
been examined giving PRM -II, 13 and 19. A recommended cure cycle
is shown in Table 6.14.
TABLE 6.14
RECOMMENDED PRESS CURE FOR PMR-15 COMPOSITES 92
Collimation and impregnation from 50 %solids solution

Air or IR dry, oven stage to desired tack and drape
Imidise by heating, in oven or in tool at 120°-200°C for 1-3 h
Place in mould preheated to 230-315°C, dwell for 30s-lOmin
Apply pressure of 1-7MPa and cure at 315T for I h
To obtain optimum mechanical properties at elevated temperatures it is necessary
to post cure at 343°C for 16 h
In addition to the press cure it is possible to use the PMR resins

in autoclave procedures and also to produce mouldings with chopped
fibre. 92 More recently, improved compositions employing the PMR con-
cept have been developed 93 . 94 which exhibit significantly improved high-
temperature performance while retaining the processing ease of PMR-15.
The improved formulations contain p-phenylene diamine, the dimethyl-
ester of hexafluoro-isopropylidene-bis-( 4-phthalic acid) and the mono-
methyl ester of 5-norbornene-2,3-dicarboxylic acid.
Some properties of HM-S carbon fibrejPMR-15 composites are shown
in Table 6.15.
TABLE 6.15
PROPERTIES OF HM-S CARBON FIBRE/PMR-15 COMPOSITES 92
Property Without post cure After post cure of

16h at 343°C
At room temperature At 316°C At 316°C
Flexural strength (MPa) 1262 483 1103

Flexural modulus (GPa) 185 104 173
Shear strength (MPa) 59 22 44
The density of the cured PMR-15 is 1·30g/cm 387 and it has a Tg

of about 335°C after the 343 °C post cure. Higher temperature post
cures can give higher Tgs, but with a concomitant lowering of interlaminar
shear strength. 95 This lowering was interpreted as being due to micro-
cracking within the composite, partly due to the residual stress that
originates from the large differences between the matrix and fibre coeffi-
cients of thermal expansion. Such cracking is observed with epoxy resins
and, consequently, can be expected to be greater with the polyimides
as they require even higher temperatures for cure. But in this case there
appears to be an additional factor in that the cracking was greater with
some varieties of carbon fibre. It has also been observed that the carbon
fibre can affect the thermo-oxidative degradation of the resin to a consider-
able extent, as measured on a weight loss basis. For example, PMR-II
o~------~~--------------------,
~
,,
\
,0
\
\
20 \
\
o ,
,,
\
\
,,
-..
" "-,
,,
.
"~O
....
L
<>.
"- ...
o o 10
"
o 05~
E
80 >- ..,
---------
b
IOOL---2~0·~0-------4~O~O--~--~6~OO~~--~80~
Itmperolurt·C
FIG.6.18. PMR-15-(a)weight loss and (b) rate of weight loss (dY).--in air;
---- in nitrogen.
190 G.1. KNIGHT
resin alone loses 13 % by weight after 2100 h at 316°e, whilst incorpora-

tion of HTS-I carbon fibre causes a weight loss of 24--26 % under the
same conditions. This increase in degradation of the resin was interpreted
as being due to impurities in the carbon fibre-probably sodium salts. 96 . 97
The TGA curves for cured PMR-15 are illustrated in Fig. 6.18 which
show the resin to be initially equally stable in air or in nitrogen, initial
weight loss occurring at about 350°C. Ten per cent weight loss is not
reached until 420 e in air or 445°e in nitrogen. This is very similar
0
to the weight loss curves for Kerimid 601, Fig. 6.14, although PMR-15
gives a much higher residual char in nitrogen.
The bismaleimide and norbornene methods of crosslinking the poly-
imide do give resins with easier processability. It is possible to fabricate
composites with void contents < I % so that the mechanical properties
are as good as those achieved with epoxy resins, but because of the
alicyclic entities introduced into the polymer structure the thermal stability
is reduced compared to that of the condensation type polyimides.
6.3.21. Thermid 600

Another variety of thermosetting polyimide that cures by an addition
polymerisation process has been developed. This is marketed by Gulf
Oil Chemicals and is known as Thermid 600. The structure of the
prepolymer is shown below.
o 0
II 0 II
O-N", ~
/ rry-c-rry
C
~
'" -@-C=oCH
I
0
C
/N
C C
II II
o 0
Thermld 600
The compound is prepared from I mole 1,3-bis(3-aminophenoxy)benzene,

2 moles benzophenone tetracarboxylic dian hydride and 2 moles 3-amino-
phenylacetylene. 98
Polymerisation and cure occurs via a process which is believed to

involve the trimerisation of the terminal acetylenic groups into aromatic
rings. This theory has not been proven, but NMR studies on a model
compound, N-(3-ethynylphenyl)phthalimide showed that new aromatic
C-H groups were formed as the acetylenic groups disappeared. The high
thermal stability of the cured resins also supports the formation of aromatic
linking groups.
R CH R
"',C# /
HC
~
C
C
III
CH -+ ¥
RWR
R
I
R
The prepolymers are readily soluble in dimethylformamide or N-methyl-

pyrrolidone and these solutions can be used to apply the resin to fibres
or fabrics in the manufacture of prepregs. As mentioned above such
solvents are very difficult to remove and can cause trouble during pro-
cessing and degrade the properties of the laminates. To obviate these
problems a method of hot melt prepregging has been devised, which
provides prepregs that are easily processed into composites. 99 The hot
TABLE 6.16
RECOMMENDED CURE CYCLE FOR GLASS OR CARBON FIBRE/THERMID 600 LAMINATES
Place prepreg mto press prooeated to 252°C

Close to contact and maintain for 60 s
'Bump' lay-up (momentarily release the pressure) every 10 s for 30 s
Slowly increase pressure to 1·4 MPa
Cure for I to 2 h at 1·4 MPa and 252°C
Cool under pressure to below 93°C and remove from press
To obtain optimum high-temperature mechanical properties a post cure is
recommended either in air or mert atmosphere
Place laminate in cold oven
Raise temperature to 232°C rapidly
Increase from 232 to 343°C at 14°Cjh
0
Hold at 343°C for 4 h

Increase from 343 "C to 371°C at 7"Cjh
Hold at 371°C for 4 h
Cool slowly in oven to below 93°C before removing
192 G.1. KNIGHT
melt process takes place at 310°C and the resin cures rapidly at this
temperature. To prevent the cure going too far, the prepreg has to be
chilled rapidly as it exits from the hot melt zone.
A further problem is that the resin will cure without melting if the
heating rate of the lay-up is insufficient. 99 Heat up rates of the order
of 8°C/min are required; such heating rates are readily achieved in a
press but require specially modified autoclaves. Allowing for this it is
possible to obtain good quality laminates by either press or autoclave
cure. I t is also possible to use the resins to prepare mouldings and as
high temperature adhesives.
A typical cure cycle is shown in Table 6.16, examples of the properties
of this resin and its laminates with glass and carbon fibre are shown
in Table 6.17.
The TGA curves for a sample of cured Thermid 600 are illustrated
TABLE 6.17
TYPICAL PROPERTIES OF CURED THERMID 600 RESIN AND ITS
LAMINATES WITH GLASS· AND CARBON FIBREb 100
Property ReSin Glass Carbon
Flexural strength (MPa)

at room temperature 131 479 1344
at 316°C 29 310 1020
Flexural modulus (GPa)
at room temperature 4-48 32-4 103
at 316°C 20·7 83
Short beam shear (MPa)
at room temperature 64 83
at 316°C 45 55
Tensile strength (MPa) 83
Tensile modulus (GPa) 3·93
Compressive strength (MPa) 172
Effect of ageing at 316°C
%weight loss-after 500h 2·89
-after I 000 h 4·04 9·72 10·00
Flexural strength after I 000 h at 316°C
at room temperature 92 162 600
at 316°C 18 160 572
• Glass lam mates are 3·18mm thick, 7628 glass cloth W/CS 290
fimsh. Resin content 35 %.
b Carbon fibre laminates are 1·59 mm thick, unidirectional,
Hercules HTS fibre. Resin content 35 %.
HIGH-TEMPERA lURE PROPERTIES OF THERMALLY STABLE RESINS 193
in Fig. 6.19, where the weight losses in air and in nitrogen are shown.
From this it can be seen that weight loss starts at about 350 °C in both
air and nitrogen, 10% weight loss occurring at 460 OC in air and 525 °C
in nitrogen. On a weight loss basis this makes Thermid 600 almost as
stable in nitrogen as Kapton or Skybond 700, good evidence, as mentioned
O~~~~--------------------------.
20
..
c
.~
Q.
40
020
c
E
o 10~
80 >-
."
---- -----~ --------

500 600 700
trmporature"C
FIG.6.19. Thermid 600--(a) weight loss and (b) rate of weight 10ss(dY). ~~ in
air; - --- in mtrogen.
above, for the highly aromatic character of the cured resin. The great
reduction in oxidative stability shows that there are some fundamental
differences between the resins, possibly related to the difficulty in getting
all the acetylenic end groups to react fully during the cure. Any olefinic
or acetylenic un saturation left in the molecule would presumably act
as a site for the initiation of thermo-oxidative degradation.
194 G.1. KNIGHT
6.4. THERMOPLASTIC POLYIMIDES
6.4.1. Introduction
As has been shown above, the addition type polyimides although easier
to process than the fully aromatic condensation type polyimides suffer
a marked reduction in their overall thermal stability. An alternative method
of improving processability is to introduce fiexibilising groups into the
polyimide molecule so that after the initial cycJising reaction the applica-
tion of suitable temperatures and pressures can eliminate any voids formed.
Two such products will be described here, one is Du Pont's NR-150
resin system, and the other is Polyimide 2080 marketed by the Upjohn
Corporation.
6.4.2. NR-1S0 Polyimides

A considerable amount of work has gone in to developing these resins
for use either as matrix resins for composites, as chopped fibre moulding
compounds, or as high-temperature adhesives; several papers have been
published on the subject over the past eight years.101-110 The resins
were originally sold as solutions in N-methylpyrrolidone or in N-methyl-
pyrrolidonejethanol mixtures, of a free tetrafunctional acid and a diamine
which could be coated onto the reinforcement. These solutions have
recently (January 1980) been withdrawn from the market, and Du Pont
will now sell only fabricated products.
The basic chemistry of the NR-150 binder solutions and their cured
products is shown below. When sold as solutions, they had a shelf life
of at least nine months when stored at 4 DC or below. 11 0 The hexafiuoro-
propylidene group fiexibilises the molecule, lowering the Tg of the poly-
imides to 280-300 DC in the case of NR-150A2 and to 350-371 DC for
NR-150B2.
The NR-150B material has a better thermo-oxidative stability than
the NR-150A. An illustration of the thermal stability is given in Fig. 6.20
which shows the TGA curves obtained by heating cured NR-150A2 in
air and in nitrogen. From this it appears that the polymer is equally
stable in air or nitrogen, weight loss starting at about 300°C in inert
and oxidising atmospheres, but 10% weight loss not being reached until
485°C in air and 495 °C in nitrogen.
Because of the thermoplastic nature of the resin, prepregs based on
NR-150 systems are easier to process than any other aromatic condensation
type polyimides. High quality mouldings can be produced using a variety
of processes. This processing versatility is indicated by the four methods
CF 3
HOOC I COOH
HOOC~~~COOH Tetra-acid
hexaftuoroisopropylidene-bis( 4-phthalic acid)
+ +
NR-JSOA2 NR--JSOB2
H2 N-@-O-@-NH 2 Diamine
4,4' -diaminodiphenyl ether

solution in a 3:1 mixture of
m- and p-phenylenediamine
solution in a J: 3 mixture of
1
Binder
N-methylpyrrolidone/ethanol N-methylpyrrolidone/ethanol solutions
heat
1
Prepregging
ethanol
N-methylpyrroJidone
1
Cure
o 0
water
1
Shaped
II CF 3 I object
C I C
N( iQr?-1(3Y ;N-JQ\-O
C~CF3~C ~
I I
o 0
polYlmide
o 0
I CF 3 I
>
C I C
N( iQrT--fc3Y
C~ CF3~C
I I
o 0
polyimide
196 G. J. KNIGHT
--- ---
0r===~~~~~------------~
.....
\
\
\
\
\
20 \
\
\
\
\
\
a ,
,
"
' ......0.........
,,
,
80
/
I--....-.:=====----....~,/ 0
100L-~Z~O~0-------4~O~O------~60~O~----~800
8---~- ---
temprrature'c
FIG. 6.20. NR-lSOA2-(a) weight loss and (b) rate of weight loss (dY). - - in
air; ~~~~ In nitrogen.
illustrated in Table 6.18 and some typical properties of the resins and
composites made from them in glass and carbon fibre are shown in
Table 6.19.
The effect of long term ageing on the mechanical properties of four
different laminates from NR-ISO solutions is shown in Table 6.20. It
can be seen that there is a minimal effect of exposure on the mechanical
properties at elevated temperature for all the laminates, though there
is evidence that the room temperature properties are falling. There was
no sign of any surface cracking or erosion. The reasons for this minimal
reduction in mechanical properties may be associated with the small
weight change of the samples after exposure and the glass transition
data. The Tg for all samples increased during exposure, presumably
as a result of oxidative crosslinking reactions as was seen to occur in
Kapton films. Any increase in crosslink density would be seen as an
TABLE 6.18
RECOMMENDED METHODS OF PROCESSING NR-150 POLYIMIDE PREPREGS 111
Method A:" Vacuum bag/autoclave at high temperatures

Precure
I. Heat from room temperature to 200°C at 2°C/min under full vacuum
2. Hold 30 min at 200°C under full vacuum
Removal of volatiles and consolidation
3. Heat to 343°C over 2 h under full vacuum
Mould
4. Hold I h at 343°C under full vacuum and 1·4 M Pa
5. Cool under pressure
Method B:b Vacuum bag precurmg fol/Olred by matched die mouldmg
Precure
I. Heat from room temperature to 200°C at 2 OC/min under full vacuum
2. Hold 30 min at 200°C under full vacuum
Mould
3. Preheat matched die to 427 °C and insert precured laminate
4. Hold IOmm at 427°C and l7MPa
5. Cool to below the Tg under pressure no greater than 1-4 MPa
Method C:' Vacuum bag/autoclave mouldmg at 100r temperature~ fol/Olred by
vacuum bag/oven po~t cure
Precure
I. Heat from room temperature to 175°C at 2°C/min under full vacuum
Consohdate and mould
2. Hold 3 h at 175°C under full vacuum and 1-4 MPa pressure
Removal of volatiles m oven
3. Heat from 23°C to 316°C over 24-48 h under full vacuum
4. Cool under full vacuum
Method D:d Vacuum bag/oven only
Precure
I. Heat from room temperature to 200°C at 2°C/mm under partial
vacuum
2. Hold 30 mm at 200°C under full vacuum
Consohdate and removal of volatIles
3 Heat from 200°C to 316 °C over I h under full vacuum
4. Hold I h at 316°C under full vacuum
5. Heat to 343°C and hold for 2 h under full vacuum
U Produces a 3·18 mm thick lammate usmg NR-150A2j181 style E-glass fabric.

VOId content < I 10.
b Produces a 3·18mm thick laminate using NR-150B2/181 style E-glass fabric.
Void content < 1%.

, Produces a 1·78mm thick lammate using NR-150B2/S-2 glass fabnc. Void
content 3-6 %
d Produces a 3·l8mm thick lammate usmg NR-150A2j181 style E-glass fabric.
VOId content 7-10 %.
TABLE 6.19
PROPERTIES OF NR-150 RESINS AND THEIR COMPOSITES WITH GLASS AND CARBON FIBRE!!!
'-C>
Propert)" Cured resin Composites Iritll E-glass jabric Carbon fibre composites 00
Magnamite Magnamite
AS HMS
NR-J50A2 NR-J50B2 NR-J50A2 NR-J50B2 NR-J50A NR-J50B2
Laminate thickness (mm) 2·54 2·54 3·05 3·18
3·05 3·05
Volume % fibre 53 52 55 52
Volume % voids I I I 2
T (0C) 280-300 350-371 257 360
Ffexural strength (MPa)
room temperature 97 117 510 558 1530 869
250°C 641
260°C 407 Cl
288°C 641 ~
316°C 262 682 ;.:

343°C
z
496 Ci
Flexural modulus (GPa) ::r:
-I
room temperature 23-4 30·3 110 145
250°C 96
260°C 28·3
288°C 124
316°C 25·5 138
343°C 117
Tensile strength (MPa)
room temperature 110 110
Short beam shear strength (MPa)
room temperature 67·6 74·5 120·0 51·0
250°C 63-4
260°C 37·2
288°C 37·2
316°C 28·3 31·7
343°C 31·7
TABLE 6.20
THERMO-OXIDATIVE STABILITY OF NR-150 LAMINATES EXPOSED IN AN AIR OVEN 111
CompositIOn resin remjorcement NR-J50A NR-J50B2 NR-J50B2 NR-J50B2

::t:
47% quart:: fibre 53 % E-glass fabnc 52% umdlrectlOnal 52 % umdll'ectional Ci
Magnamlte HMS Magnamlte HMS ::t:I
carbon fibre carbon fibre -I
m
Agemg conditions
s::"t:I
m
temperature °C 260 316 316 343 :::0
time at temperature (h) 20000 1000 3000 500 >
-I
C
Properties After After After After :::0
m
Control exposure Control exposure Control exposure Control exposure "t:I
:::0
Flexural strength (MPa) 0
"t:I
m
room temperature 613 434 345 869 869 :::0
-I
250°C 365 358" m
316°C 241 296 682 565 '"0.,.,
343°C 496 517
Flexural modulus (GPa) -I
::t:
room temperature 24·1 33·1 27·6 145 145 m
:::0
250°C 22·1 20·0" s::
316°C 23A 26·9 138 131 >
r
343°C 117 117 r
-<
Short beam shear strength
-I
'"
(MPa) >
t:I:l
room temperature 93·8 68·9 30·3 51·0 51·0 r
m
250°C 29·0 33·8 :::0
316°C 22·1 22·1 31·7 40·7 m
343°C 31·7 40·0 '"Z
Weight loss (%) 3·0 1·7 2·5 1·5 '"
Tg (0C) 270 295 343 356 350 360 350 381
'-0
" After 10 000 h at 260°C. '-0
200 G. J. KNIGHT
increase in flexural strength but would probably cause a decrease in

impact properties.
By extrapolation of this data it can be shown that the time taken
to lose 50 %of the original flexural strength measured at the test tempera-
ture is approximately 50000 h at 260°C, 5000 h at 316°C and 1500 h
at 343°C. This extrapolation is made for low-void laminates reinforced
with fibres such as E-glass, quartz or high modulus carbon fibre. Compo-
sites based on the high strength, or type A, carbon fibres would show
a lower level of stability due to the poorer oxidative resistance of such
fibres.
6.4.3. Polyimide 2080

Polyimide 2080 is unique compared to all the polyimides discussed above
in that it is a polymer with a fully imidised structure, which still remains
fusible and soluble in solvents such as dimethyl formamide or N-methyl-
pyrrolidone. The polymer is prepared 112 by reacting benzophenone tetra-
carboxylic dianhydride with toluene di-isocyanate (a mixture of the 2,4-
and 2,6-isomer) in dimethyl sulphoxide solution. On completion of the
first part of the reaction 4,4'-diisocyanate diphenyl methane is added
and the reaction taken to completion. The aim is to produce a block
copolymer with the approximate structure shown below.
20%
and
o 0 0
I I I
<c~©(>
C C
80%
I /I
o 0
The polymer with this constitution has a Tg of 305°C. The thermal
stability of the polymer is illustrated in Fig. 6.21 which shows the TGA
.."' 40
~
..
u
L
0.
~
~
0 1·0
o·s
80
0·2
100 L....~4;;;O;;:0;:;;;~:::::::5~0'-0-~~~60"0=~:=:=::l0
tem perature·C
FIG. 6.21. PolYlmide 2080-{a) weight loss and (b) rate of weight loss (dY).
- - in air; ---- in nitrogen.
curves obtained in air and in nitrogen. The initial weight loss is at 375°C
in air and 410°C in nitrogen, 10% weight loss is at 490°C in air and
at 530°C in nitrogen.
6.4.4. Processing and Properties of Poly imide 2080

It is claimed that it is possible to obtain good quality laminates from
Polyimide 2080 using either high pressure platen press, low pressure
TABLE 6.21
PRESS CURE OF POL YIMIDE 2080 LAMINATES
1. 30 mill at 250 DC to reduce solvent level to less than 4 %

2. Raise temperature of press to 350 DC and bump repeatedly to release the
remaining volatilised solvent
3. Hold for one hour in the press at 350 DC and 2·1 to 3-4 MPa
202 G. J. KNIGHT
vacuum bag, or moderate pressure autoclave vacuum bag methods of

processing. No B-staging is required and no water is given off due to
ring closure, curing consists simply of solvent removal. The recommended
processing conditions are given in Table 6.21 and some typical properties
are shown in Table 6.22.
TABLE 6.22
POLYIMIDE 2080, RESIN AND LAMINATE PROPERTIES
Property Resm!!3 Type 181E Fortaftl5-T

glass cloth, unid/l'ectlOnal carbon
30% fibre, 30% resin!!4
resm!!3

room temperature 198 379 1027
288°C 34 303
room temperature 3·3 27-3 116
288°C 1·1 24·8
Interlaminar shear strength
(MPa)
room temperature 15·9
The process is essentially compression moulding, utilising the thermo-

plastic properties of the polymer. As in a conventional moulding process
the laminate can be formed in direct contact with the pressure surfaces,
eliminating the waste associated with bleeder cloths and resin flow.
6.4.5. Polyamide-Imides
Another method of introducing flexibilising groups into a polyimide is
via the ester or amide groupings. The polyester- and polyamide-imides
have been available for some time. Being soluble in polar solvents and
capable of forming films with good mechanical properties, they have
been used as varnishes and wire coating enamels. As with the other
linking groups described above, the overall thermal stability of the polymer
is reduced compared with that of the aromatic polyimides.
In addition to their use as insulating coatings, the polyamide-imides
have been developed for use as thermoplastic moulding compounds, which
may be used unfilled or reinforced with, for example, carbon fibre. liS -117
The amide-imides are based on the combination of trimellitic anhydride

with aromatic diamines.
o 0 0 0
~ I I ~
N/ ~-NH-R-NH-C~ ~N-R
~C~ ~C/
I I
o 0 n
These materials may be moulded, extruded or machined to produce the

required articles. For compression moulding, temperatures of 325-350°C
are pressures of 21-28 MPa are required. To obtain optimum engineering
properties the moulded parts require to be heat treated for 50-200 h
at temperatures of up to 260°C.
Examples of the engineering properties of a polyamide-imide, Torlon,
and carbon fibre reinforced Torlon are given in Table 6.23.
TABLE 6.23
PROPERTIES OF TORLON RESIN AND RESIN CONTAINING CHOPPED CARBON FIBRE
Property Torlon Torlon Carbon fibre reinforced

2000 4000 Torlon
25 % Fibre 50 % Fibre
Tensile strength (MPa)

room temperature 92 117 172 165
149°C 81 79 124 103
260°C 61 28 13 8
Tensile modulus (GPa)
room temperature 19 31
149°C 10 26
260°C 1·6 1·8
149°C 127 133 150 237
260°C 98 62 66 60
room temperature 4·8 3·6 17 24
149°C 3·7 2·8 11 23
260°C 3·1 1·9 6 10
Shear strength (MPa)
204 G.1. KNIGHT
The maximum temperature for use is about 250°C because of the

loss of mechanical strength at that temperature, but the retention of
properties with time at temperature is very good, for example, 94 %
retention of strength after 2000 h at 260°C.
6.5. PSP RESIN
6.5.1. Synthesis
Another resin capable of forming crosslinked networks without evolution
of volatiles is the PSP resin. I 18 The oligomeric prepolymer is prepared
by condensation of aromatic dialdehydes with methylated derivatives
of pyridine, in particular 2,4,6-trimethylpyridine.
D
CH 3
+ OHC--@-CHO -->
H3C N CH 3
The synthesis can be arrested at various points producing either liquid

resins, solids soluble in ethanol or acetone/ethanol mixtures, solids soluble
in polar solvents such as dimethylformamide or N-methylpyrrolidone,
or materials more suitable for use in injection or compression moulding.
The prepolymer has been shown to have a good shelf life, over six months
at 5°C. The resin undergoes initial cure by heating at 200°C, but to
obtain optimum engineering properties a post cure of at least 2 h at
250°C is required. I 19 The resin so obtained has good thermal stability,
TGA in air or in argon shows initial weight loss occurring at 300°C
and 10 %weight loss at 435 °C in air, or 460°C in argon. In inert atmosphere
the pyrolytic residue is of the order of 65 % at 1000 0C. 120
6.5.2. Processing and Properties of PSP Resin

After impregnation the prepregs are dried at about 100°C to a volatile
content of about 10 %, this gives good drape and tack but not too much
TABLE 6.24
RECOMMENDED PRESS CURE CYCLE FOR PSP LAMINATES
1. Place prepreg in mould, raise temperature to 200T at about 10°C/min,

hold for 1 h
2. Apply a pressure of 0,5-1,0 MPa, hold for two hours, maintaining the
temperature at 200°C
3. Post cure:
2hat250°C
or 16 h at 250°C
or 3 h at 300°C
The first post cure gives material suitable for use at temperatures up to 150°C;
the latter permits the use of the material at temperatures up to 300°C.
flow when moulding. Such prepregs may then be moulded in a press

or subjected to vacuum bag-autoclave cure. A recommended cure cycle
is given in Table 6.24.
The materials prepared in this way have lifetimes of > 10000 h at
200°C, 1000--1500 h at 250°C and 120--150 h at 300°C. Some of the
properties of cured resin and its laminates with glass and carbon fibre
are shown in Table 6.25. 119
TABLE 6.25
PROPERTIES OF CURED PSP RESIN AND ITS LAMINATES WITH GLASS
AND CARBON FIBREl19
Property Re.l/n Glass fibre UnidIrectiOnal

£-181 jabl'lc carbon fibre
HTS HMS

200°C 530
250°C 1250
room temperature 2·5 30 110 150
200°C 30
250°C 110
Shear strength (MPa)
room temperature 48 90 50
200°C 31 40
250°C 70
206 G.1. KNIGHT
6.6. FIRE RESISTANCE
One property of these resins ·that has not been discussed above is their
flammability. The high-temperature resins are all fire resistant. The poly-
imides in particular are very resistant, being self extinguishing and having
very good limiting oxygen index figures. Where fire and smoke emission
are known to be of consequence then any of the above polyimide resins
could be recommended for use.
6.7. CONCLUSIONS
Overall, it can be said that a considerable effort has been expended
on the synthesis of polymers with high-temperature capabilities. That
the synthetic work has paid off can be seen in the number of commercially
available resins especially the polyimides. Some of the latter are capable
of service at temperatures over 300°C and even at 400 °C for very short
times. There is one major drawback to all these resins when the construction
of large structural items for aircraft is being considered, i.e. the high
temperatures and extended times required for post cure; 250°C for a
minimum of 15 h. Where composites are concerned, and especially carbon
fibre reinforced composites, the internal stresses set up by the high tempera-
ture cure are very severe. Prolonging the time taken over the post cure
can help to alleviate the stresses but then the costs of manufacture escalate
rapidly.
It would appear that the processor would like a resin with a long
shelf life at room temperature, yet capable of being cured at temperatures
of no more than 50-100°C, then with no post cure the structures should
show good property retention at temperatures of 300°C or even higher.
It seems unlikely that the polymer chemists will ever develop such a
resin system, and the processor will have to learn to come to terms
with the exacting processing conditions if he wants to make use of the
excellent high temperature properties that can be achieved with carbon
fibre laminates.
REFERENCES
l. LEE, W. A. and RUTHERFORD, R. A. Glass transitIOn temperatures of

polymers, chapter in Polymer Handbook 2nd ed., eds. Brandrup, J. and
Immergut, E. H., 1975, Wiley-Interscience, New York.
2. ROSE, J. B. Polymer, 1974, 15,456.

3. BARTON, 1. M. Unpublished work.
4. BARTON, J. M. Polymer, 1970, 11,212.
5. THEBERGE, J. Nat. SAMPE Symp., 1975,20,72.
6. BROWNING,C. E. 28th SPl Reinjorced Plastics/Composites Conf., 1973, Paper
15-A, pp. 1-16.
7. BROWNING, C. E., HUSMAN, G. E. and WHITNEY, J. M. Compo Mat. Test. and
Design, 4th Con}., ASTM STP 617,1977, pp.481-96.
8. HER1Z, J., NASA-CR-124290, 1973.
9. MORGAN, R. J. and O'NEAL, J. E. Polym. Plast. Technol., Eng., 1978, 10,49.
10. HUSMAN, G. E. Nat. SAMPE Symp., 1979,24,21.
11. HANSON, M. P., NASA-TN D-6604, Dec. 1971.
12. HASKINS, J. F., NASA-CP-OOI Part 2, Nov. 1976, p. 799.
13. SCOLA, D. A., Nat. SAMPE Symp. 1978,23,950.
14. VOGEL, H. A., UK Patent No. 1,060,546, March 1967.
15. JOHNSON, R. N., FARNHAM, A. G., CLENDINNING, R. A., HALE, W. F. and
MERRIAM, C. N. J. Polym. Sci., 1967,5, Part AI, 2375.
16. JONES, M. E. 8., OK Patent No. 979,111, Jan. 1965.
17. VOGEL, H. A. J. Polym. Sci., 1970,8, Part AI, 2035.
18. HALE, W. F. ACS Polymer Preprlnts, 1966,7,503.
19. KNIGHT, G. J. Unpublished work.
20. DAVIES, A. Makromol Chem., 1969, 128,242.
21. ALVINO, W. M. J. Appl. Polym. Sci., 1971, 15,2521.
22. ABDUL-RASOUL, F. Europ. Polym. J., 1977, 13, 1019.
23. BROWN, J. R. Polym. Letters, 1970, 8, 121.
24. GESNER, B. D. J. Appl. Polym. Sci., 1968, 12, 1199.
25. GOTHAM, K. V. and TURNER, S. Polymer, 1974, 15,665.
26. THEBERGE, J. 29th SPl Reinforced Plastics/Composites Con/., 1974, Paper
20-D, pp. 1-14.
27. HOUSTON, A. M. Materials Engineering, 1975,81,28.
28. HOUSE, E. E. Nat. SAMPE Symp., 1979,24,201.
29. MACALLUM, A. D. J. Org. Chem., 1948, 13, 154.
30. EDMONDS,1. T. and HILL, H. W, OS Patent Nos. 3,354,129, Nov. 1967 and
3,524,835, Aug. 1970.
31. BLACK, R. M., LIST, C. F. and WELLS, R. J. J. Appl. Chem., 1967,17,269.
32. EHLERS, G. F. L., FISCH, K. R. and POWELL, W. R. J. Polym. SCI., 1969,7,
Part AI, 2955.
33. VERNON JONES, R. 30th SPI Relnjorced Plasflcs/Composlfes Con}., 1975,
Paper 12-B, pp. 1-5.
34. BAEYER, A. VON. Chem. Ber., 1872,5,25,280, 1094.
35. BAEKELAND, L. H. Ind. Eng. Chem., 1909, 1, 149, 545; 1910, 2,478; 1911,3,
932; 1912,4,737; 1913,5,506; 1916,8,568; 1925, 17,225; 1935,27,538.
36. MEGSON, N. J. L. Trans. Faraday Soc., 1936,32,336.
37. WATERMAN, H.1. and VELDMAN,A. R. Brit. Plast., Mould. Prod. Trans., 1936,
8, 125, 182.
38. CONLEY, R. T. ACS DIV. Org. Coatings and Plasllcs Chem., 1966,26, 138.
39. SHULMAN, G. P. and LOCKTE, H. W. ACS DIV. Org. Coatings and Plastics
Chem., 1966,26, 149.
208 G. J. KNIGHT
40. UK Patent No. 455,008 issued to Henkel and Cie, 1935.

41. KOEDA, K. J. J. Chem. Soc. Japan, Pure Chem. Sec., 1954,75,571.
42. LADY, J. H., ADAMS, R. E. and KESSE, I. J. Appl. Polym. Sci., 1960,4,65.
43. MANLEY, T. R. J. Polym. Sci., Symp. No. 42, 1973, 1377.
44. FRIEDEL, C. and CRAFTS, J. M. Ann., 1865, 136,203.
45. ROCHOW, E. G., US Patent No. 2,380,995, Aug. 1945.
46. WRIGHT, W. W. and LEE, W. A. The search for thermally stable polymers,
chapter in Progress in High Polymers Vol. 2, eds. Robb, J. C. and Peaker, F.,
1968, Iliffe Books Ltd, London.
47. GOLDOVSKII, E. A. Kauchuk i Rezina, 1979,24.
48. DUNLOP, A. P. and PETERS, F. N. Chapter 18 in The Furans, ACS Monograph
Series, 1953, Reinhold Publ. Corp., Washington.
49. BOZER, K. B., BROWN, L. H. and WATSON, D. 26th SPI Reinforced
Plastlcs/Composites Conf., 1971, Paper 2-C, pp.I-6.
50. BOZER, K. B. and BROWN, L. H. 27th SPI Reinforced Plastics/Composites
Conj., 1972, Paper 3-C, pp.I-5.
51. SELLEY, J. E. 29th SPI Reinforced Plastics/Composites Con!, 1974, Paper
23-A, pp. 1---4.
52. RADCLIFFE, A. T. Furane resins, Chapter 5 in Developments with
Thermosetting Plastics, eds. Whelan, A. and Brydson, J.A., 1975, Applied
Science Publishers Ltd, London.
53. DOWNING, P. A. Chem. Eng. (London), 1978,331,272.
54. EDWARDS, W. M. and ROBINSON, I. M. US Patent Nos. 2,710,853, 1955 and
2,867,609, 1959.
55. EDWARDS, W. M. and ROBINSON, I. M., US Patent No. 2,880,230, 1959.
56. IDRISJONES, J., OCHYNSKI, F. W. and RACKLEY, F. A. Chem.lnd., 1962, 1686.
57. EDWARDS, W. M., US Patent Nos. 3,179,614,1965 and 3,179,634,1965.
58. ENDREY, A. L. Canadian Patent No. 659,328, 1963 and US Patent Nos.
3,179,631,1965 and 3,179,633,1965.
59. BOWER, G. M. and FROST, L. W. J. Po/ym. Sci., 1963, 1, Part A-I, 3135.
60. DINE-HART, R. A. and WRIGHT, W. W. J. App/. Po/ym. Sci., 1967, 11,609.
61. BRUCK, S. D. Polymer Preprints, 1964,5, 148.
62. HEACOCK, J. F. and BERR, C. E. SPE Trans., 1965,5, 105.
63. JOHNSTON, T. H. and GAULIN, C. A. J. Macromo/. Sci., Chem. A., 1969,3,
1161.
64. COTTER, J. L. and DINE-HART, R. A. Chem. Commun., 1966,809.
65. COTTER, J. L. and DINE-HART, R. A. Organic Mass Spectrometry, 1968, 1,
915.
66. DINE-HART, R. A. Bra. Polym. J., 1971,3, 163.
67. DINE-HART, R. A., PARKER, D. B. V. and WRIGHT, W. W. Brit. Polym. J.,
1971,3,222.
68. DINE-HART, R. A., PARKER, D. B. V. and WRIGHT, W. W. Bra. Po/ym. J.,
197I, 3, 226.
69. DINE-HART, R. A., PARKER, D. B. V. and WRIGHT, W. W. Bm. Polym. J.,
1971,3,235.
70. Skybond 703, heat resistant reSin, Monsanto Co. Plastics Products and Resins
division, technical bulletin No. 6236, Jan. 1970.
71. PETKER, I. Nat. SAMPE Symp., 1976,21,37.
72. PIKE, R. A., MAYNARD, C. P. Nal. SAMPE Tech. Con}., 1969,331.

73. STUCKEY, J. M. SAMPE Quarlerly, 1972,3,22.
74. WOLKOWITZ, W. and POVEROMO, L. M. 28th SP/ Reinforced Plastics/
Composites Conf., 1973, Paper 15-C.
75. IVANOV, V. S., KOPTEVA, I. P. and LEVANDO, L. K., USSR Patent No. 164,678,
Aug. 1964.
76. SAMBETH, J. and GRUNDSCHOBER, F. French Patent No. 1,455,514, 12 Nov.,
1965.
77. KOVACIC, P., US Patent No. 2,818,407, Dec. 1957.
78. BARGAIN, M., COMBET, A. and GROSJEAN, P., US Patent No. 3,562,222,1971.
79. DARMORY, F. P. Nat. SAMPE Symp., 1973, 18,317.
80. Kerimid, high temperature polyimide resin, technical data sheet, Rhone-
Poulenc UK, London.
81. PETROVICKI, H., STENZENBERGER, H. D. and HARNIER, A. VON. Proc. 32nd
SPE Tech. Con}., 1974,88.
82. STENZENBERGER, H. D. J. Polym. SCI., Polym. Chem. Ed., 1976, 14,2911.
83. ALVAREZ, R. T. Nat. SAMPE Symp., 1975,20,253.
84. JONES, J. S. Nat. SAMPE Symp., 1976,21,438.
85. STEl\ZENBERGER, H. D. Appl. Polym. Symp., 1973,22, 77.
86. HARRUFF, P. W. Nal. SAMPE Tech. Conj., 1979, 11, 1.
87. LORENSEN, L. E. SAMPE Quarterly, 1975,6, 1.
88. BURNS, E. A., LUBOWITZ, H. R. and JONES, J. F., NASA-CR-72460, Oct. 1968.
89. SERAFINI, T. T. Appl. Polym. Symp., 1973,22,89.
90. SERAFINI, T. T. J. Appl. Polym. Sci., 1972, 16,905.
91. CAVANO, P. J. and WINTERS, W. E., NASA-CR-135,113, Dec. 1976.
92. WINTERS, W. E. Nat. SAMPE Symp., 1975,20,629.
93. DELVIGS, P.,ALSTON, W. B. and VANNUCCI, R. D. Nal. SA MPE Symp. , 1979,
24, 1053.
94. WINTERS, W. E. Nat. SAMPE Tech. Con}., 1978, 10,661.
95. PETKER, I. Nal. SAMPE Symp., 1978,23,775.
96. ALSTON, W. B. Nal. SAMPE Symp., 1979,24, 72.
97. McMAHON, P. E. Nal. SAMPE Symp., 1978,23, 150.
98. BILOW, N. Nat. SAMPE Symp., 1975,20,618.
99. ApONYI, T. J. Nat. SAMPE Symp., 1978,23, 763.
100. Thermid 600, technical data sheet, Gulf Oil Chemicals Co., Houston, Texas.
101. GIBBS, H. H. Nat. SAMPE Symp., 1972, 17, Section III-B-Six, pp.I-9.
102. GIBBS, H. H. Nat. SAMPE Symp., 1973, 18,292.
103. GIBBS, H. H. 29th SPI Reinforced Plastics/Composites Conj., 1974, Paper
II-D, pp. 1-15.
104. GIBBS, H. H. Adv. Chem. Ser., 1975, 142,442.
105. GIBBS, H. H. Nal. SAMPE Symp., 1975,20,212.
106. GIBBS, H. H. Nal. SAMPE Symp., 1976, 21, 592.
107. GIBBS, H. H. Nal. SAMPE Tech. Con}., 1978, 10,211.
108. GIBBS, H. H. Nal. SAMPE Symp., 1978, 23, 806.
109. GIBBS, H. H. Nat. SAMPE Symp., 1979, 24, II.
110. HOEGGER, E. F. Nat. SAMPE Symp., 1979,24,533.
Ill. Du Pont NR-150 po Iyim ide precursor solutions, technical data sheet, E. I. Du
Pont de Nemours and Co., Wilmington, Delaware.
210 G. J. KNIGHT
112. ALBERINO, L. M., US Patent No. 3,708,458, Jan. 1973.

113. RAUSCH, K. W., FARRISSEY, W. J. and SAYIGH, A. A. R. 28th SPI Remjorced
PlastIcs/Composites Con}., 1973, Paper II-E.
114. FARRISSEY, W. J.,ALBERINO, L. M.,RAUSCH, K. W. and SAYIGH,A.A.R. Nat.
SAMPE Tech. Conf., 1972,4,29.
liS. CHEN, Y. T. Nat. SAMPE Symp., 1978,23,826.
116. ANON. Kunststofje, 1977, 67, 17.
117. WALKER, R. H. Nat. SAMPE Symp., 1974, 19, 186.
118. ROPARS, M. and BLOCH, B. French Patent No. 2,261 ,296,1974 and US Patent
No. 3,994,862, 1976.
119. BLOCH, B. Nat. SAMPE Symp., 1978,23,836.
120. ROPARS, M. and BLOCH, B. La recherch aerospatiale, 1977(2), 103.
Chapter 7
STRUCTURE-PROPERTY RELATIONSHIPS AND

THE ENVIRONMENTAL SENSITIVITY OF EPOXIES
ROGER J. MORGAN
Lawrence Livermore Laboratory, University of California, USA
SUMMARY
The structure, dejormation andjailure process, and the mechanical proper-

ties, oj amine-cured epoxies are discussed, together with the question of
how jabrication and environmental factors can ajject resin behaviour.
The nature oj the chemical reactions that produce amine-cured epoxide
networks, and the chemical and physica/features controlling these reactions,
are described, together with the physical structural parameters which deter-
mine the mechanical response oj epoxies in terms of their crosslinked
network morphologies and microvoid characteristics. The author discusses
critically the mechanisms responsible jor heterogeneous crosslinked mor-
phologies, and the techniquesJor detectmg them.
Evidence is presented for the deformation oj the epoxies in terms oj
crazing and shear banding, and consideration is given to the ability oj
these crosslinked glasses to undergo flow, in terms oj incomplete network
jormation, network morphology, and bond breakage.
M odiftcation oj the structure-property relationships by combinations
of stress, thermal environment and humidity is discussed.
7.1. INTRODUCTION
The increasing use of high-performance light-weight fibrous composites,

particularly in the aircraft and automobile industries, has led to a need
to predict the lifetimes of these materials in service environments. The
211
212 ROGER J. MORGAN
durability of the epoxy matrices (primarily amine-cured epoxy thermosets)

used in these high-performance composites and of the overall composite
has been a <cause for concern. A number of studies have indicated that
the combined effects of stress, sorbed moisture, and thermal exposure
can cause significant changes in the mechanical response of the compo-
site. 1.2 The structural and mechanical integrity of both the epoxy matrix
and the fibre-matrix interfacial region can be modified by these environ-
mental factors.
Predicting the durability of the epoxies in service environments with
confidence requires knowing the structure, modes of deformation and
failure, mechanical-response relationships, and the possible modification
of such relationships by fabrication procedures and environmental
exposure. The structure-property relations of epoxies (and thermosets
in general), however, have received little attention compared to other
commonly used polymer glasses. Because of the dependence of their
chemical and physical structure on fabrication conditions and because
of their infusible, insoluble nature, thermosets are more difficult to study
than are noncrosslinked polymeric glasses.
This chapter reviews the fundamental areas necessary for making mean-
ingful durability predictions for epoxies. These are: (1) their physical
structure, which includes the crosslinked network morphology and the
microvoid characteristics, (2) their chemical structure, (3) their modes
of deformation and failure and (4) the effect of environmental factors
on their structural and mechanical integrity. Although amine-cured epoxies
are addressed primarily, the discussion also applies to the entire glassy
thermoset family.
7.2. MATERIALS
Several specific epoxy systems are mentioned in this chapter. Most of

them are amine-cured. Reference is made to a number of amine-cured
bisphenol-A-diglycidyl ether (Dow; DER 332) (DGEBA) epoxies, to a
diaminodiphenyl sulphone (Ciba-Geigy; Eporal)-cured tetraglycidyI4,4'-
diaminodiphenylmethane(Ciba-Geigy, MY 720)(TGDDM-DDS) epoxy,
and to the following curing agents used in conjunction with DGEBA
epoxides:
(I) Jeffamine T-403, an aliphatic polyether triamine (Jefferson),
(2) Diethylene triamine (DETA) (Eastman) and
(3) 2,5-Dimethyl,-2,5-hexane diamine (DMHDA) (Aldrich).
STRUCTURE-PROPERTY RELATIONS OF EPOXIES 213
Another curing agent mentioned is the polyamide, Versa mid 140 (General
Mills). Some of the references to TGDDM-DDS systems are to 'in-house'
preparations, while others are to the commercial systems Narmco 5208,
which contains no catalyst, and Fiberite 934, which contains about
O· 5 wt % BF 3 catalyst. 3
7.3. CHEMICAL STRUCTURE
The chemical structure of an epoxy is often complex. The structure depends

on specific cure conditions because more than one reaction can occur
during cure and the kinetics of these reactions exhibit different temperature
dependencies. In addition, the structure is affected by factors such as
steric and diffusional restrictions of the reactants during cure,4 - 9 the
presence of impurities that act as C<:!lalysts, I 0 the reactivities of the epoxide
and curing agent, II isomerisation of epoxide groups, II -13 nonhomo-
geneous mixing of the reactants 9.14 and cyclic polymerisation of the
growing chains. II These factors can lead to heterogeneous network
structures.
In amine-cured epoxides, networks are generally assumed to result
from addition reactions of epoxide groups with primary and secondary
amines, II as illustrated in Fig. 7.1. Epoxides and amines with more
FIG. 7.1. Epoxide-amine addition reactIOn.
than three functional groups can form highly crosslinked network struc-
tures. However, amine-cured epoxies often exhibit considerable duc-
tili ty8 9.14 - I 7 and microscopic flow, 8 9 14 - 18 and they exhibit yield stresses
that show thermal-history and strain-rate dependencies very similar to
those of noncrosslinked glasses. 17 Such observations suggest either that
the epoxies are not as highly crosslinked as expected if only simple epoxide-
amine addition reactions had taken place or that the reactions themselves
did not proceed completely. (Further possibilities discussed in later sections
214 ROGER J. MORGAN
are that (1) crosslinks are broken under stress and (2) the microscopic
flow processes are controlled by regions of low crosslink density where
the network morphology is heterogeneous.)
In addition to the steric and diffusional restrictions that limit the
cure reactions and the crosslink density, a ring-opening homopolymerisa-
tion of the epoxide groups can also occur, 10 -12.19 as illustrated in Fig. 7.2.
(
-R- C r-
OH
/0"
CH2 ) n
I H2
R-C-C
I
°
I H2
R-C-C
H I
°I
H2
R-C-C
H I
°
I H2
R-C-C
H I
R-C--
°I
H
FIG. 7.2. Homopolymerisation of epoxides.
As a consequence of this trans-etherification reaction to form polyether

linkages, fewer epoxide groups are available during the later stages of
cure to react with the secondary amine groups and, subsequently, to
form crosslinks. Although the epoxide homopolymerisation reaction by
itself can lead to crosslinked networks in many amine-cured epoxide
systems, a lower crosslink density generally results in these systems when
the trans-etherification reaction occurs. The epoxide-amine reactions are
controlled by the presence of H-bond donors, such as OH groups, that

are necessary to open the epoxide rings. 5.6.1 0 The trans-etherification
reaction, however, requires a tertiary amine as a catalyst and an H-bond
donor as a cocatalyst. 1o ,12,19 Hence, the final chemical structure of the
epoxy network can be complex because it depends on such parameters
as (1) the relative rates of the chemical reactions at room temperature,
(2) the final postcure temperature, (3) the concentrations of catalysts
such as sorbed moisture in the system and (4) steric restrictions that
inhibit reactions at secondary-amine sites. Fourier-transform infrared
spectroscopy studies using spectral stripping, which reveals differences
in the spectra recorded at different stages of cure, should enhance our
knowledge of the chemical reactions involved in the network formation
of amine-cured epoxy systems.
In the case of the TGDDM-DDS epoxy systems, there is significant
evidence that these networks do not form exclusively from epoxide-amine
addition reactions. The structure of the unreacted TGDDM epoxide
and DDS amine monomers are illustrated in Fig. 7.3. Table 7.1 illustrates
the wt % of DDS required for reaction of (1) all primary and secondary
CH -CH-CH
2, / 2
o
tetraglycidyl 4, 4'-diaminodiphenyl methane
TGDDM
4, 4'-diaminodiphenyl sulfone
DDS
FIG. 7.3. The TGDDM-DDS epoxy system.
216 ROGER J. MORGAN
TABLE 7.1
TGDDM-DDS
THEORETICAL REACTION MIXTURES FOR EPOXY SYSTEM
50% TGDDM epoxide 100% TGDDM epoxide

groups react groups react
100% DDS primary and

secondary amines react 23wt % DDS 37wt% DDS
100% DDS primary amines
react 37wt% DDS 54wt% DDS
amines in the DDS and (2) only the primary amines in the DDS with
50 % and 100 % of the epoxide groups in the tetrafunctional TGDDM
molecules. The glass transition temperature (Tg) of the TGDDM-DDS
epoxies rises with increasing DDS concentration (up to 25 wt % DDS)
because of corresponding increases in molecular weight and/or crosslink
density.9 The Tg , however, reaches a maximum of about 250°C at about
30 wt %DDS and decreases, subsequently, for higher DDS concentrations
because of plasticisation by unreacted DDS molecules. However, 37 wt %
DDS is required to consume half the TGDDM epoxide groups when
only epoxide-primary amine reactions occur (Table 7.1). Hence, the Tg
maximum value at about 30 wt % DDS suggests that, assuming only
epoxide-primary amine reactions occur, less than half of the TGDDM
epoxide groups react because steric and diffusional restrictions inhibit
further reaction. It seems doubtful whether networks in which only half
of the epoxide groups have reacted would exhibit the respectable mechani-
cal properties shown by the TGDDM-DDS epoxies (20-35 wt %DDS).9
Evidently, other cure reactions, in addition to the epoxide-primary amine
reactions, are occurring and possibly involve (I) epoxide homopolymerisa-
tion, (2) epoxide-secondary amine reactions and (3) internal cyclisation
within the TGGDM epoxide as a result of hydroxyl and/or secondary
amino groups reacting with adjacent unreacted epoxides.
Infrared spectroscopy also suggests that TGDDM-DDS networks do
not form exclusively from epoxide-amine addition reactions. In Fig. 7.4
the percentage of epoxide groups consumed during cure, as determined
by the disappearance of the epoxide band at 910 cm -1 in the infrared
spectrum, is plotted as a function of cure conditions for both a BF 3-
catalysed TGDDM-DDS epoxy (Fiberite 934; 25 wt % DDS) and a
noncatalysed TGDDM-DDS epoxy (Narmco 5208; 20 wt % DDS). For
a standard l7rC cure lasting 2·5 h, all the epoxide groups are consumed
(within experimental error) for the BF 3-catalysed system despite the fact
STRUCTlJRE~PROPERTY RELATIONS OF EPOXIES 217
~
!II
C.
:J
...
0
Cl
CI)
~ 50
x
0
c.
..
CI)
al
CJ
III
CI)
ex:
Oo---~---L--~--~----~--~--~
(0 C) 23 120 177 200 225 250 275 300
(h) 0 1 2.5 1 1 1 1 1
Cure conditions
FIG. 7.4. Percentage of reacted 1::poxide groups as a function of cure conditions

for TGDDM-DDS-based epoxies.
that the wt % of DDS in this system is well below the stoichiometric

quantity necessary to consume all the epoxide groups by normal epoxide~
amine addition reactions. In the case of the noncatalysed system about
35 % of the epoxide groups remain unreacted after curing at 177 °C for
2·5 h and all such groups reacted only after exposure at 300°C for I h.
7.4. PHYSICAL STRUCTURE
The primary physical and structural parameters that control the modes
of deformation and failure as well as the mechanical response of epoxies are
the crosslinked network structure and micro void characteristics. 8,9, 14 - 16,18
The chemistry of the cure processes and the final network structure
of epoxies have been deduced from the chemistry of the system. This
knowledge is based on the assumptions that the cure reactions are known
and that they go to completion. The experimental techniques used to
determine the cure processes and epoxy network structure include IR
and 13C NMR spectroscopy and swelling, ultrasonic, thermal conductivity,
dynamic mechanical, and differential scanning calorimetry measurements
(see references in Ref. 8). However, as discussed in the previous section,
in many epoxy systems the chemical reactions are diffusion controlled
218 ROGER J. MORGAN
and do not proceed to completion. Furthermore, a heterogeneous distribu-

tion in the crosslink density can occur. Hence, a variety of crosslinked
network structures can be produced.
Figure 7.5 schematically illustrates possible network topographies of
epoxies. 8,20 - 22 An ideal uniform crosslinked network structure is illus-
trated in Fig. 7.5(a). In reality, however, networks contain loops and
dangling chain-ends as illustrated in Fig. 7.5(b). Such networks can exhibit
la)
Ideal network
Ib)
Loops; dangling
chain ends
leI
Uniform crosslink
density
(d)
Non·uniform crosslink
density
FIG. 7.5. Epoxy network morphologies.

an essentially uniform crosslink density in a low molecular weight or

crosslink density matrix (Fig. 7.5(c». Also, nonuniform crosslink density
networks in which regions of high crosslink density form either a con-
tinuous (Fig. 7.5(d» or a discontinuous phase in a low molecular weight
or crosslink density matrix are other possible network morphologies.
High crosslink density regions, from 6000 to 10 000 nm in diameter,
have been observed in crosslinked resins. 3,8,15,16,18,23-49 The formation
of a heterogeneous rather than a homogeneous system depends on poly-
merisation conditions, i.e. temperature, solvent, and chemical composi-
tion. The high crosslink density regions have been described as agglo-
merates of colloidal particles 28 ,29 or floccules 31 in a lower molecular
weight interstitial fluid. Funke 5o .51 suggested a number of factors that
can be responsible for heterogeneous network formation: (1) difference
between the reactivities of different functional groups, (2) unreacted func-
tional groups, (3) intramolecular cyclisation reactions and (4) phase
separation. Phase separation in the form of microgelation is generally
believed to be the primary mechanism in the formation of heterogeneous
crosslinked networks. Solomon et al. 30 originally suggested that a two-
phase system is produced by microgelation prior to the formation of
a macrogel. Kenyon and Nielsen 34 indicated that the highly crosslinked
microgel regions are loosely connected during the latter stages of the
curing process. Karyakina et al. 41 suggested that microgel regions originate
in the initial stages of polymerisation from the formation of microregions
of aggregates of primary polymer chains. More recently, Luettgert and
Bonart 47 discussed the morphology of epoxies in terms of the relative
rates of microgel formation and the subsequent growth rate of these
gel particles. At low cure temperatures, only a small number of gel particles
are nucleated and, hence, large nonhomogeneities are produced; at higher
cure temperatures, the rate of nucleation of microgel particles is faster, and
larger numbers are produced, which, therefore, limits their growth in
size. Finally, the high crosslink density regions have been reported to
be only weakly attached to the surrounding matrix,28.29,31 and their
size also varies with the proximity of surfaces 31 ,45 and the presence of
solvents. 30 ,52
Most of the evidence for heterogeneous regions of crosslink density
in epoxies is derived from electron microscope investigations. These micro-
scopy studies involve carbon-platinum replication of etched and nonetched
free surfaces and fracture surfaces. However, artefacts can often result
from replication techniques. Scepticism of evidence of nodular morphology
based on these techniques arises for the following reasons: (1) similar
220 ROGER J. MORGAN
nodular structures that are observed by replication of epoxy surfaces

can also be produced by replication of inorganic glass slides; (2) blisters
that are produced as a result of the etching of epoxy surfaces can be
interpreted as nodular regions of high crosslink density, (3) the fracture
topographies can exhibit a nodular appearance in the initiation region
as a result of fractured craze fibrils,8.9.14 .18 and (4) the surface structure
implied by carbon-platinum replication techniques is not always observed
by scanning electron microscopy studies of gold-coated surfaces.
More confidence can be placed in bright-field transmission electron
microscopy studies of the morphology of thin epoxy films strained directly
in the electron microscope. For example, in DGEBA- DETA epoxies,
it was observed that particles 6 to 9 nm in diameter remain intact and
flow past one another during flow processes. Figure 7.6 illustrates a
network of these particles . It was suggested that the 6 to 9 nm diameter
75 nm
FIG. 7.6. Bright-field transmiSSIOn electron micrograph of strained network

structure of particles of 6-9 nm in diameter in DGEBA- DETA epoxy.
particles were molecular domains that were mtramolecularly crosslinked

and that formed during the initial stage of polymerisation. 8.18 The inter-
connection of molecular domains by regions of either low or high crosslink
density allows two types of network structure: (l) regions of high crosslink
density embedded in a low or noncrosslinked matrix or, (2) noncrosslinked
or low crosslink density regions embedded in a high crosslink density
matrix. In the case of DGEBA- DETA epoxies, both types of network
morphology were seen, with the first type more prevalent. 8.18 Deformation
of the first type of network involves preferential deformation of the
regions of low crosslink density without causing cleavage of the high
crosslink regions. For the second type of network, the deformation

process is more complex. Local affine deformation requires that network
cleavage and flow occur simultaneously in the high crosslink density
regions while at the same time flow occurs with little network cleavage
in the neighbouring low crosslink density regions. This deformation pro-
cess results in progressively larger regions that are poorly crosslinked.
In other amine-cured epoxies, such as TGDDM-DDS systems,9 no
evidence for heterogeneous crosslink density distributions was found from
observation of strained films in the electron microscope. However, under
polarised light, a network of larger I mm sized nodules was seen in
BF 3-catalysed TGDDM-DDS epoxy systems (Fiberite 934).3 These bi-
refringent networks are permanently destroyed above Tg (or at 25°C
below Tg when stress is applied). The birefringence originates from the
preferential alignment of the macromolecules within the birefringent net-
work. The network alignment is caused by biaxial shrinkage stresses
imposed on the epoxy sheet during cure. (The TGDDM-DDS epoxies
are cured between glass plates, and the shrinkage stresses are not com-
pletely relieved by the release agent present on the surface of the glass
plates.) The nonhomogeneous distribution of the catalyst within the
TGDDM-DDS system produces a heterogeneous structure because
regions of high catalyst concentration polymerise more rapidly than sur-
rounding regions. The shrinkage stresses, in conjunction with the hetero-
geneous polymerisation conditions during gelation and glass formation,
produce the birefringent network of aligned molecules; the more highly
polymensed regions, which contained high concentrations of catalyst,
align more readily under stress than do surrounding regions of lower
molecular weight.
The microvoid characteristics of the epoxy are also important in control-
ling the mechanical response of the glass. Microvoids can have a deleterious
effect on the mechanical properties of epoxies by acting as stress concentra-
tors and by accumulating sorbed moisture. Microvoids can result when
air is trapped in the system during cure and from trapped low molecular
weight material that is subsequently eliminated from the glass during
postcure. This low molecular weight material results from either nonhomo-
geneous mixing of epoxide and curing agent or from the aggregation
of unreacted constituents. In polyamide-cured DGEBA epoxies, crystals
of DGEBA epoxide monomer trapped in the partially cured resin at
room temperature can produce microvoids by melting and volatilising
under certain postcure conditions. 14 Thermal-anneal, moisture-sorption,
and mechanical-property studies also indicate that in TGDDM-DDS
222 ROGER J. MORGAN
epoxies, the melting and volatilisation ofunreacted DDS crystallites during

cure produces microvoids. 9
7.5. MODES OF DEFORMATION AND FAILURE
Localised plastic flow has been reported to occur during the deformation
and failure of epoxies; in a number of cases, the fracture energies were
two to three times greater than the expected theoretical estimates for
purely brittle fracture. 8 However, no systematic studies have been made
to elucidate the microscopic flow processes occurring during the deforma-
tion of epoxies and to determine the relation of such flow processes
to the network structure.
Recent investigations revealed that both DGEBA-DETA and TGDDM-
DDS epoxies deform and fail by a crazing process. 8.9,14 -16,18 Crazes
were observed in films either strained directly in the electron microscope
or strained on a metal substrate. The fracture topographies of these
epoxies, fractured as a function of temperature and strain rate, were
interpreted in terms of a crazing process. The TGDDM-DDS epoxies
also deformed to a limited extent by shear banding, as indicated by
multiple right-angle steps present in the fracture-topography initiation
region (Fig. 7.7). Shear-band propagation in these partially crosslinked
glasses produces structurally weak planes because of bond cleavage caused
during molecular flow. Hu1l 53 and Mills 54 both noted that the intersection
of shear bands, which occurs at right-angles, causes a stress concentration.
This stress concentration is sufficient to cause a crack to propagate through
the structurally weak planes caused by shear-band propagation. These
phenomena can produce the multiple right-angle steps observed in the
fracture topography. Mixed modes of deformation that involve both
crazing and shear banding were also observed in the fracture topography
of TGDDM-DDS epoxies. More recently, in the case of DGEBA-
DMHDA epoxies, macroscopic shear bands were observed under polarised
light.
The ability of amine-cured epoxies to undergo considerable microscopic
flow can be explained by either, (1) incomplete cure reactions and/or
epoxide homopolymerisation resulting in a lower crosslink density net-
work (see Section 7.3), (2) regions of low crosslink density controlling
the flow processes (see Section 7.4) or, (3) bond breakage under stress,
which results in a lower overall crosslink density. Two pieces of evidence
were recently published which suggest that bond breakage readily occurs
STRUCTURE-PROPERTY RELA nONS OF EPOXIES 223
10IJ. m
FIG. 7.7. Scanning electron micrograph illustrating multiple right-angle steps in

the fracture-topography initiation regIOn of a TGDDM-DDS epoxy.
in epoxy networks under stress. Firstly, Levy and Fanter 55 reported

enhanced chemiluminescence of TGDDM-DDS epoxies under stress.
This chemiluminescence is associated with bond fracture and subsequent
reaction of the macro-radicals with oxygen. Secondly, Gledhill et a/. 56
reported that an amine-cured DGEBA epoxide is appreciably toughened
after it is subjected 10 an applied load. This observation suggests that
stress causes bond breakage, which lowers the crosslink density; this
decreased crosslink density allows an increase in molecular flow at the
crack tip, which results in increased toughness.
7.6. DURABILITY
The durability of epoxy matrices depends on many complex interacting

phenomena. The factors that control the critical path to ultimate failure
or unacceptable damage depend specifically on the particular prevailing
environmental conditions. These environmental factors include service
224 ROGER J. MORGAN
stresses, humidity, temperature, and solar radiation. The combined effects

of thermal history, moisture exposure, and stress have a deleterious effect
on the physical and mechanical integrity of epoxies. Morgan et al. 57
have recently studied the effect of specific combinations of moisture,
heat, and stress on the physical structure, failure modes, and tensile
mechanical properties ofTGDDM-DDS epoxies. The main findings from
these studies are outlined in the following paragraph.
Sorbed moisture plasticises TGDDM-DDS epoxies and lowers their
tensile strengths, ultimate elongations, and moduli. The fracture topo-
graphies of the initiation cavity and mirror regions of these epoxies indicate
that sorbed moisture enhances the craze initiation and propagation pro-
cesses. The crazing process is more susceptible to sorbed moisture than
is the Tg; this can be explained in terms of local moisture concentrations
enhancing the local cavitation and flow processes. Hence, modification
of Tg by sorbed moisture cannot be utilised alone as a sensitive guide
to predict deterioration in the mechanical response and, hence, the dura-
bility of epoxies.
The effect of increasing stress levels on the subsequent moisture-sorption
characteristics of initially dry TGDDM-DDS epoxies was investigated.
In Fig. 7.8 the equilibrium moisture-sorption levels after about 40 days
exposure to 100 % relative humidity at room temperature are plotted
versus the stress levels that were applied to the epoxies before moisture;
sorption. All data points fall within the shaded areas. Stresses in the
5.5 ,---r----r-----r--,........-..,---,-----,
4.0 '--_...L-_-L._--L_ _- ' - - _......._ - - ' - _ - - '

o 10 20 30 40 50 70
1h room· temperature stress IMPa)
FIG. 7.8. Equilibrium wt %mOIsture sorbed by a TGDDM- DDS epoxy at 100 %

relative humidity and 23 °C versus I h constant-stress levels that were applied
before exposure to moisture.
oto 38 M Pa range had no detectable influence on the subsequent moisture-

sorption levels. However, moisture-sorption levels increase sharply by
up to about II % in the 38 to 43 MPa stress range. At higher stress
levels in the 43 to 65 M Pa range, in which a few specimens actually
broke, there is only a slight trend towards higher moisture-sorption levels
with increasing stress.
The data in Fig. 7.8 indicate that the initial stages of craze-crack
growth enhanced the accessibility of moisture to sorption sites to a greater
extent than the later stages of growth. (The primary sorption sites within
the TGDDM-DDS epoxy are the hydroxyl, sulphonyl, and primary and
secondary amine groups-all of which are capable of forming hydrogen
bonds with water molecules.) The TGDDM-DDS epoxy specimens that
fractured under constant load were found to exhibit fracture topographies
similar to specimens previously studied that fractured in shorter times
in the 10 - 2 to 10 1 min - 1 strain-rate region. 9 Such topographies have
been interpreted in terms of a craze-crack growth process 9 in which
crazing followed by crack propagation predominates in the initial stages
of failure, and crack propagation alone predominates during the later
stages of failure. The dilatational changes produced in the epoxy glass
by the crazing process enhance the accessibility of moisture to sorption
sites within the epoxy to a greater extent than does crack propagation
alone. Hence, the initial stages of failure in TGDDM-DDS epoxies enhance
the accessibility of moisture to sorption sites to a greater extent than
do the later stages of failure.
One of the more extreme environmental conditions experienced by
an epoxy composite matrix occurs during a supersonic dash of a fighter
aircraft. The aircraft dives from high altitudes (where outer surface
temperature is - 20 to - 55°C) into a supersonic, low-altitude run during
which aerodynamic heating raises the surface temperature to 100-150°C,
in a matter of minutes. On reduction of speed, the outer surface tempera-
ture drops extremely rapidly at rates up to about 500 °Cjmin, thus exposing
the epoxy composite to a thermal spike. Simulation of such thermal
spikes has been shown to increase the amount of moisture sorbed by
the epoxy or epoxy composite. 58 -u 4 However, after a certain number
of consecutive thermal spikes, the amount of moisture sorbed ceases
to increase. Browning 60 •63 suggested that such increases result from micro-
cracks caused by the moisture and temperature gradients present during
the thermal spike. McKague 62 recently noted that damage does not occur
unless the thermal-spike maximum temperature exceeds the Tg of the
moist epoxy.
226 ROGER J. MORGAN
The amount of moisture sorbed by TGDDM-DDS epoxies was en-

hanced by about 1·6 wt % after exposure to a 150°C thermal spike. 5 7
The surfaces of the thermally-spiked epoxies were examined for the
presence of surface microcracks using scanning electron microscopy. No
significant areas of microcracking were observed in any of the specimens
when examined under magnifications of up to 30000 times. Hence, the
sorbtion of additional moisture by the epoxies after exposure to thermal
spikes is not primarily caused by microcracking.
The primary mechanism by which thermally spiked epoxies sorb
additional moisture can be explained in terms of moisture-induced free
volume changes. The molecular mobility of the epoxy is enhanced at
the high temperatures experienced during the thermal spike as the Tg
of the epoxy-moisture system is approached. This molecular mobility
is sufficient to enhance the dissociation of H-bonds between the water
molecules and active sites within the epoxy. Although the ruptured
H-bonds can re-form at active sites, there is an overall decrease in the
amount of hydrogen bonding and a corresponding increase in the mobility
of the water molecules. The more mobile water molecules with fewer
H-bonds require a greater free volume because H-bonding generally causes
a volume decrease. The molecular mobility of the epoxy-moisture system
during a thermal spike is sufficient to allow configurational changes
to occur within the epoxy network, which accommodates the greater
free volume required by both the more mobile water molecules and the
normal moisture-induced swelling stresses imposed on the epoxy. Such
free volume increases, which involve permanent rotational isomeric
population changes within the epoxy network, are frozen into the epoxy
glass during the rapid cool down portion of the thermal spike. The
additional free volume allows water molecules access to previously inacces-
sible active sites within the epoxy.
To a lesser extent, the rupture of crosslinks, crazing and/or cracking,
and the loss ofunreacted material can also contribute to enhanced moisture
sorption after thermal spike exposure. Thermal spike exposure can cause
surface crazing and/or cracking of epoxies if the moisture-induced swelling
stresses, together with those stresses that result from temperature gradients
and relaxation of fabrication stresses, exceed the craze-initiation stress at
the maximum thermal spike temperature. Thicker epoxy specimens are
more susceptible to the growth of permanent damage regions during
thermal spike exposures because they are exposed to larger temperature
gradients and shrinkage stresses during cure that, in turn, produce larger
fabrication stresses and strains.
7.7. ACKNOWLEDGEMENT
This chapter refers to work performed by the author under the auspices
of the US Department of Energy, by the Lawrence Livermore Laboratory,
under contract number W-7405-ENG-48. Neither the US nor the US
Department of Energy, nor any of their employees, nor any of their
contractors, subcontractors, or their employees, makes any warranty,
express or implied, or assumes any legal liability or responsibility for
the accuracy, completeness or usefulness of any information, apparatus,
product or process disclosed, or represents that its use would not infringe
privately-owned rights.
Reference to a company or product name does not imply approval
or recommendation of the product by the University of California or
the US Department of Energy to the exclusion of others that may be
suitable.
REFERENCES
1. Air Force durabilllY workshop, September 1975, Battelle Columbus

Laboratories, Columbus, Ohio.
2. Air Force conjerence on the effects oj relative humidllY and temperature on
composite structures, March 1976, University of Delaware, AFOSR-TR-77-
0030 (\ 977).
3. MORGAN, R. J. and MONES, E. T. Composiles Tech. Rev., 1979, 1(4), 18.
4. FRENCH, D. M., STRECKER, R. A. H. and TOMPA, A. S. J. Appl. Polym. SCI.,
1970,14,599.
5. HORIE, K., HIURA, H., SAWADA, M., MITA, I. and KAMBE, H. J. Polym. SCI.,
1970,8, Part AI, 1357.
6. ACITELLI, M. A., PRIME, R. B. and SACHER, E. Polymer, 1971, 12,335.
7. PRIME, R. B. and SACHER, E. Polymer, 1972, 13,455.
8. MORGAN, R. J. and O'NEAL, J. E. Polym. Plast. Technol. Eng., 1978, 10,49.
9. MORGAN, R. J., O'NEAL, J. E. and MILLER, D. B. J. Mater. SCI., 1979, 14, 109.
10. WHITING, D. A. and KLINE, D. E. J. Appl. Polym. SCI., 1974, 18, 1043.
11. LEE, H. and NEVILLE, K. Handbook of Epoxy Resins, 1967, McGraw-Hili, New
York.
12. SIDYAKIN, P. V. Vysokomol. Soedin., 1972, A14, 979.
13. BELL, J. P. and McCAVILL, W. T. J. Appl. Polym. SCI., 1974, 18,2243.
14. MORGAN, R. J. and O'NEAL, J. E. J. Macromol. SCI. Phys., 1978, BI5(1), 139.
15. MORGAN, R. J. and O'NEAL, J. E. Structural parameters ajjecting the brittleness
oj polymer glasses and composites in Advances in Chemistry No. 154: Toughness
and Brittleness oj Plastics, eds. Deanin, R. D. and Crugnola, A. M., 1976,
American Chemical Society, Washington, D.C.
228 ROGER J. MORGAN
16. MORGAN, R. J. and O'NEAL, J. E. in Chemistry and Properties of Crosslmked

Polymers, ed. Labana, S. S., 1977, Academic Press, New York, p. 289.
17. MORGAN, R. J. J. Appl. Polym. SCI., 1979,23,2711.
18. MORGAN, R. J. and O'NEAL, J. E. J. Mater. SCI., 1977, 12, 1966.
19. NARRACOTT, E. Brit. Plast., 1953,26, 120.
20. NIELSEN, L. E. J. Macromol. Sci., Rev. Macromol. Chern., 1969, C3(I), 69.
21. NIELSEN, L. E. Crosslmkmg-effect on physical propertles oj polymers, 1968,
Monsanto/Washmgton University/ONR/ARPA Association, Report HPC-
68-57.
22. MANSON, J. A., KIM, S. L. and SPERLING, L. H. Influence oj crosslinkmg on the
mechanical properties oj high Tg polymers, 1976, Technical Report AFML-TR-
76-124, for US Air Force Materials Laboratory, Wright-Patterson AIr Force
Base, Ohio, by Materials Research Center, Lehigh University, Bethlehem, Pa.
Available from National TechnIcal InformatIOn Service, ArlIngton, Va.,
Document AD-A033078.
23. CARSWELL, T. S. Phen op lasts , 1947, Intersclence, New York.
24. ROCHOW, T. G. and ROWE, F. G. Anal. Chern., 1949,21,261.
25. SPURR, R. A., ERATH, E. H., MYERS, H. and PEASE, D. C. Ind. Eng. Chern.,
1957,49, 1839.
26. ERATH, E. H. and SPURR, R. A. J. Polym. SCI., 1959,35,391.
27. ROCHOW, T. G. Anal. Chern., 1961,33, 1810.
28. ERATH, E. H. and ROBINSON, M. J. Polym. SCI.. ;1963, 3, Part C, 65.
29. WOHNSIEDLER, H. P., J. Polym. SCI., 1963,3, Part C, 77.
30. SOLOMON, D. H., LOFT, B. C. and SWIFT, J. D. J. Appl. Polym. SCI., 1967, 11,
1593.
31. CUTHRELL, R. E. J. Appl. Polym. SCI., 1967, 11,949.
32. NEVEROV, A. N., BIRKINA, N. A., ZHERDEV, Yu. V. and KOZLOV, V. A.
Vysokomol. Soedm., 1968, AIO, 463.
33. NENKOV, G. and MIKHAlLOv, M. Makromol. Chern., 1969, 129, 137.
34. KENYON, A. S. and NIELSEN, L. E. J. Macromol. SCI., Chern., 1969, A3(2), 275.
35. TURNER, D. T. and NELSON, B. E. J. Polym. SCI., Polym. Phys. ed., 1972, 10,
2461.
36. BASIN, V. YE., KORUNSKll, L. M., SHOKALSKAYA, O. Y. and ALEKSANDROV,
N. V. Poly. Sci. USSR, 1972, 14,2339.
37. KESSENIKH, R. M., KORSHUNOVA, L. A. and PETROV, A. V. Polym. SCI. USSR,
1972, 14,466.
38. BOZVELlEV, L. G. and MIHAJLOV, M. G. J. Appl. Polym. Sci., 1973, 17, 1963; J.
Appl. Polym. SCI., 1973, 17, 1973.
39. MORGAN, R. J. and O'NEAL, J. E. Polym. Preprmts, 1975, 16(2),610.
40. SELBY, K. and MILLER, L. E. J. Mater. Sci., 1975, 10, 12.
41. KARYAKINA, M. I., MOGILEVICH, M. M., MAIOROVA, N. V. and UDALOVA, A.
V. Vysokomol. Soedm., 1975, AI7, 466.
42. MAIOROVA, M. V., MOGILEVICH, M. M., KARYAKINA, M. I. and UDALOVA, A.
V. Vysokomol. Soedm., 1975, A17, 471.
43. SMARTSEV, V. M., CHALYKH, A. YE., NENAKHOV, S. A. and SANZHAROVSKll, A.
T. Vysokomol. Soedm., 1975, A17, 836.
44. MORGAN, R. J. and O'NEAL, J. E. Polym. Plast. Technol. Eng., 1975,5(2),173.
45. RACICH, J. L. and KOUTSKY, J. A. J. Appl. Polym. Sci., 1976,20,2111.
46. MANSON, J. A., SPERLING, L. H. and KIM, S. L. Influence oj crossllnking on the

mechanical properties oj high Tg polymers, 1977, Technical Report, AFML-
TR-77-109, for US Air Force Materials Laboratory, Wright-Patterson Air
Force Base, Ohio by Materials Research Center, Lehigh University,
Bethlehem, Pa. Available from National Technical Information Service,
Arlington, Va., Document AD-A054431.
47. LUETTGERT, K. E. and BONART, R. Prog. CollOid. Polym. SCI., 1978,64,38.
48. DUSEK, K., PLESTIL, J., LEDNICKY, F. and LUNAK, S. Polymer, 1978, 19,393.
49. SCHMID, R. Prog. Colloid. Polym. SCI., 1978,64,17.
50. FUNKE, W. Chimw, 1968,22, Ill.
51. FUNKE, W., BEER, W. and SEITZ, U. Prog. CollOid. Polym. SCI., 1975,57,48.
52. BELL, J. P. J. Polym. SCI., 1970,8, Part A2, 417.
53. HULL, D. Acta. Met., 1960,8, II.
54. MILLS, N. J. J. Mater. SCI., 1976, 11,363.
55. LEVY, R. L. and FANTER, D. L. Polymer Preprlnts, 1979,20(2),543.
56. GLEDHILL,R.A., KINLOCH,A. J. and SHAW, S. J. J. Mater. Sc·i., 1969, 14, 1769.
57. MORGAN, R. J., O'NEAL, J. E. and FANTER, D. L. J. Mater. SCI., (in press).
58. VERETTE, R. M. Temperature/humidity ejjects on the strength oj graphite/epoxy
laminates, 1975, AIAA Paper No. 75-1011.
59. McKAGUE, E. L., JR., HALKIAS, J. E. and REYNOLDS,J. D. J. Compos. Mater.,
1975,9,2.
60. BROWNING, C. E. Ph.D. thesis, Umverslty of Dayton, Dayton, Ohio, 1976.
61. HEDRICK, 1. G. and WHITESIDE, J. B. Effects of Environment on Advanced
Composite Structures in AIAA Conjerence on mrcrajt composlte~. the
emerging methodology oj ~tructural assurance, 1977, San Diego, Calif., Paper
No. 77-463.
62. McKAGUE, E. L. Chapter 5 in Proceedings oj Conjerence on EnVironmental
DegradatIOn oj Engineering Materw/~', eds. Louthan, M. R. and McNitt,
R. P., 1977, Vlrgmla Polytechnic Inst. Pnntmg Dept., Blacksburg, Virginia,
p.353.
63. BROWNING, C. E., The mechanism of elevated temperature property losses in
high performance structural epoxy resm matnx materials after exposure to high
humidity enVironments, 22nd National SAMPE Symposium and Exhibition,
San Diego, Calif., 1977, 22, 365.
64. Advanced Composite Matenals-Envlronmental Effects, ASTM STP 658, ed.
Vmson, J. P., 1978, American Society for Testing and Materials, Philadelphia,
Pa.
Chapter 8
SOME MECHANICAL PROPERTIES OF

CROSSLINKED POLYESTER RESINS
W. E. DOUGLAS and G. PRITCHARD
Kingston Polytechnic, Surrey, UK
SUMMARY
This chapter is concerned with (a) the dynamic mechanical behaviour

and (b) the fracture toughness of crosslinked unsaturated polyester resins.
Four dynamic mechanical relaxations are normally seen. The reasons
for assigning these relaxations to particular motions oj the network struc-
ture are given, although there is still some uncertainty about the origin
oj the p-transition.
The jracture behaviour oj crosslinked polyesters is also considered.
The jracture surjaces are similar to those oj glass in many respects.
Fracture toughness is low, and only slightly improved by the addition
oj a dispersed elastomeric phase.
8.1. INTRODUCTION
This chapter is concerned chiefly with topics which, although not especially
novel, have recently become of special interest to those concerned with
unsaturated polyester resins. The main subjects discussed are:
(I) the relationship between polyester network structure and dynamic
mechanical properties
(2) fracture toughness and fracture surface morphology.
It has already been established by those concerned with other mechani-
cal properties how resin formulations affect the tensile strength, hardness,
231
232 W. E. DOUGLAS AND G. PRITCHARD
elongation at break, heat distortion temperature, Izod impact strength,

flexural modulus, etc., and the results have been reviewed by Boenig.!
The effect of degree of cure on mechanical properties has also been
given considerable attention. 2 ,3
8.2. CROSSLINKED NETWORK STRUCTURES
Polyester networks may be described in chemical terms, by reference

to the dibasic acids, diols, and crosslinking agents used, and also in
physical terms, by considering the crosslink density, chain length, chain
length distribution, average bridge length and sol fraction. The chemicals
used in polyester formulations are discussed in Chapter 3, and the abbrevia-
tions for chemicals mentioned in this chapter are listed in Table 8.1.
TABLE 8.1
POLYESTER RESIN COMPONENTS: ABBREVIA nONS USED IN THIS CHAPTER
Acid Code DIOI Code Crosslink mg Code

agent
o-Phthalic PA 1,2- Propane Methyl

Isophthalic IPA Diol PG Methacrylate MMA
Terephthalic TPA Diethylene Styrene S
Maleic MA Glycol DEG Bromostyrene BS
Fumaric FA 1,6-Hexane
Adipic AA Diol HD
Succinic SA Ethane
Sebacic. SBA Diol ED
Methyl succinic MSA
The importance of the correct selection of resin components is obvious,

but many other considerations have to be taken into account. Some
of these may be evident from Fig. 8.1, which shows a typical crosslinked
network, from which the branched chain polyester molecules have been
omitted. In particular, the following should be noted:
(1) the polycondensation chains (unbroken lines),
(2) the crosslinking bridges (-S-S-S),
(3) the polyester segments between bridges,
(4) unreacted crosslinking sites, @(non-terminal maleic un saturation
is less reactive than the fumaric isomer),
MECHANICAL PROPERTIES OF CROSSLINKED POLYESTER RESINS 233
R mged letters represent

diluent monomer
FIG. 8.1. Typical crosslinked polyester network.
(5) uncrosslinked polyester molecules,

(6) trace quantities of 'diluent' molecules, e.g. PG, and
(7) water bridges (dotted lines, ----).
The polycondensation chains are usually much shorter than the cross-
linking chains, which pass through several reactive sites. However, different
crosslinking agents behave in very different ways. For instance, MMA
forms a few very long bridges while S forms many short ones. It is
partly for this reason that S is overwhelmingly preferred, and is always
assumed to be the crosslin king agent, unless otherwise stated. Vinyl toluene
is occasionally used; this resembles S in many ways.4
8.3. DYNAMIC MECHANICAL PROPERTIES
Dynamic mechanical testing involves subjecting a specimen of suitable

geometry to regular periodic deformations, usually sinusoidal. The internal
234 w. E. DOUGLAS AND G. PRITCHARD
friction within the material (caused by molecular or other movements)

dissipates energy as heat and causes the strain to lag behind the stress
input. The phase lag angle (b) can be measured by various means. For
freely damped oscillations, as produced by a torsional pendulum,s
tan 15 1
= -.~ = 1
-·In (An) =
-- 1
-In (An)
--
n n An+l nk An+k
where An' An+k are the amplitudes of the nth and (n + k)'h oscillations.
~ is called the logarithmic decrement. Other classes of instrument include
forced vibration, non-resonant devices, 6 forced vibration devices utilising
a resonant frequency characteristic of the material and its geometry 7
and pulse propagation instruments. 8
These methods give a measure of the internal energy dissipation as
a function of temperature or, less commonly, as a function of frequency.
They also give a storage modulus (shear storage modulus G', or Young's
FIG. 8.2. Nonius torsional pendulum.

storage modulus E', in the appropriate case). A corresponding loss

modulus Gil, or E" can be obtained when tan J < O· 3, by putting
tan J = G"/G' or E"/E'
or, more accurately,
G" 4n~
G = 4n 2 + ~2
Relaxation processes in polymers and the methods of studying them
are discussed in several texts. 9 . 10 This chapter centres chiefly around
studies carried out by means of the torsional pendulum, at frequencies
of around 1 Hz, over a wide temperature range. This frequency is very
suitable for the study of relaxations in polymers. A torsional pendulum
is shown in Fig. 8.2.
8.4. THE GLASS TRANSITION TEMPERATURE (Tg)
The temperature at which an amorphous glassy plastic material becomes

rubberlike is classically determined by dilatometry, but dynamic mechani-
cal methods also give an estimate of Tg • Crosslinked polymers have
broad transitions, indicated by a fall in storage modulus (see Fig. 8.3).
A peak in tan J or in ~ is associated with, but not always coincident
with, the fall in storage modulus. As a result, different ways of identifying
Tg have emerged. Cook and Delatycki 11 list three ways of determining
T g , by selecting either
(1) the temperature at which log G' is midway between the glassy and
rubbery plateaux, or
(2) the temperature at which G' reaches 10 8 Pa, or
(3) the temperature at which tan tan () is a maximum.
These values do not differ very much at low frequencies, but high frequen-
cies increase the Tg determined by the position of the damping peak.
The glass transition temperature is a major parameter, and frequently
the relaxation causing the damping peak is known as the a-transition.
It is believed to be caused by relaxation of substantial parts of the cross-
linked network, and therefore depends on the inherent flexibility of both
condensation and polyaddition chains as well as on the crosslink density.
Lenk and Padget 12 studied the Tg of unsaturated polyesters as a function
G'
FIG. 8.3. G' and tan b in the region of the glass transition temperature of a
crosslinked thermosetting resin.
TCMPCIi'ATI.III£C ' C J
FIG. 8.4. The effect ofS concentration in the resin on the tan b peak, in the region
of the a-transition (courtesy of John Wiley and Sons Inc. and the authors of
Ref. 13).
of unsaturated: saturated acid radio, and found that increasing this ratio
caused:
(1) a rise in Tg
(2) an increase in G' at Tg
(3) a decrease in tan b (max.)
(4) a broadening of the transition range.
The temperature of the main transition (the (X-relaxation) is dependent

on polyester chain length, and falls with Mn (number-average molecular
weight). It also depends on S bridge length, and was found to first
rise, in the case of one resin (1: 1 PA : FA), from 50°C at 15 % w/w
S to 120°C at 40 %w/w S, and then decrease slightly to 115°C at 70 %w/w S.
This is demonstrated by the tan c5 peaks in Fig. 8.4.13 The increasing
S concentration produces an increase in the fumarate utilisation and,
therefore, a higher crosslink density. This restricts the mobility of the
polyester chains, and results in an increase in the temperature of the
relaxation. Above 40 %S, however, the fumarate un saturation is probably
already well used, and further increases simply augment average bridge
length, thus reducing crosslink density.!!
8.5. CROSSLINK DENSITY AND Tg
Crosslink density can be determined from the stiffness of crosslinked

polymers in the region of the rubbery plateau (T> Tg + 50 K). The
dynamic shear storage modulus is given by
G' = 2</JpdRT
where </J = a front factor
p = the molar concentration of cross links per unit mass
d = the density
R = the gas constant
T = the temperature (K).
According to Graessley,!4 the equation can be modified to the form
G' = </JdRT(2C4 + C3 )
= </JdRTpe
where C3 , C4 are the concentrations in mole kg-! of effective tri- and

tetra-functional crosslinks and Pe is the concentration of elastically active

network strands, in mole kg - 1. Blanchard and Wooton 15 proposed
a further modification to allow for the space taken up by the strands
themselves, so that
G' = 4>1 dRT

(1 - KPe) Pe
where K is a steric hindrance term, and 4>1 is the ideal network front
factor, i.e.
It should be remembered that the polyester chains are relatively short

and, as indicated in Fig. 8.1, there is a sol fraction unconnected to
the gel. Cook has pointed out 16 that in order to determine the effective
crosslink density, Pe, a consideration of the varieties of crosslink found
in the three-dimensional structure is required. Eight types are listed:
(1) ordinary S-fumarate crosslinks in which the polyester chain is con-
nected to other crosslinks on both sides,
(2) those crosslinks in which the polyester is connected to the gel on one
side but at the other side it is a 'loose end',
(3) those crosslinks which form the only reacted site in a given polyester
chain.
These crosslinks do not all make equivalent contributions to network
properties. In low-S systems, where fumarate-fumarate homopolymerisa-
tion occurs, there may also be fumarate-fumarate links in which the
two conjoined polyester chains have:
(4) no other links with the gel, or
(5) one link only and three loose ends, or
(6) two links and two loose ends, or
(7) three links and one loose end, or
(8) four links and no loose ends.
It is found that G'jT passes through a maximum at a certain optimum
concentration of S, and then falls towards zero as the polyester mole
fraction diminishes. Cook and Delatycki 11 also express the effective cross-
link density in terms of the parameters illustrated in Fig. 8.1, i.e.
where: M 1 = number-average molecular weight of the polyester conden-

sation chain (typically between 700 and 2500),
M 2 = number-average molecular weight of the S chain (estimated
(I) to be of the order of 8000 to 14000),
Me, Me are the molecular weights between crosslinks of poly-
1 2
ester segments and of S segments, respectively,

pi = number of moles of reacted unsaturation per unit mass.
In practice, the crosslink density of thermosetting resins is non-uniform,

and evidence for this has been submitted not only in the form of electron
microscopic observations but also as a consequence of uneven swelling
observed when the resins are placed in suitable liquids. 1 7 The structure
of crosslinked resins may be conceived as approximating to one of the
models in Fig. 1.1. Maizel et al. 18 determined the crosslink density of
a series of polyesters, made from MA, PA and DEG (MA:PA molar
ratios 1: 0, 4: 1, 2: 1, 1: I and I: 2), from the theory of high elasticity
and from the measurement of initial Young's modulus in the region
T ~ Tg + 50 K. The density of cross1inks was found to increase almost
linearly with mole fraction of MA until the proportion of PA became
very small, whereupon this effect was reduced. The Tg fell from 106°C
for the 1:0 resin to 36°C for the 1:2 resin. Since the polyester:S ratio
was constant, a reduction in the MA: PA ratio should have given longer
S bridges (i.e. higher Me ).
Cook 19 also studied the effect of polyester crosslink density on Tg ,
using DEG/SA/FA resins crosslinked with S. An expression was obtained
by combining the Di Marzio equation 20 i.e.
where T~ = the Tg of the uncrosslinked, infinite molecular weight polymer

KDM = a crosslinking constant
M 0 = molecular weight in g mole - 1 of a structural unit
a = number of rotatable bonds/unit
p = crosslink density
with a Fox-Loshaek expression.
This gave:
where M l ' M 2 are the molecular weights of polyester and S chains respec-
tively, C is the weight % of S in the network, T; is the Tg of a hypo-
thetical blend of the two kinds of chain, both infinitely long. The other
terms are constants. In practice, the S-fumarate crosslinking chains were
considered long enough for the term
(~:)(.~o)
to be neglected, and a plot of 1/ Tg against 1/ M 1 was found to be linear.
The slope of the lines was similar to that derived from the work of
Grieveson. 22 Values obtained for the crosslinking constant K; were in
the region of 4·8 to 6·7 x 10- 4 kg mole -1 K -1.
The following are some of the factors which may give rise to the
variability of crosslinked polyester properties. Crosslinks can be irregularly
spaced, especially in IPA resins. Also, fumaric unsaturation is more easily
utilised than the non-terminal maleic form, and the proportion of maleic
to fumaric depends partly on the nature of the diol used. Chain length
depends both on reaction time and on diol: diacid stoichiometry. Finally,
residual diluent substances such as excess diol can plasticise the resin
with resultant lowering of modulus. As with all polymers, rate or frequency
of loading can influence properties.
8.6. SECONDARY RELAXATIONS
Many polymers, including crosslinked unsaturated polyester resins, show

secondary transitions below Tg • These are customarily denoted by {3,
y, .5, etc., in order of decreasing temperature. Sometimes the transitions
detected by dynamic mechanical methods are also observed by analogous
a.c. electrical measurements. 23 The existence of secondary transitions
gives the material some capacity for energy absorption below Tg and
increases the impact strength. Figure 8.5 shows schematically the tran-
sitions found in unsaturated polyester resins.
Perepechko et ai.24 studied the viscoelastic properties of polyesters
made from 1 mole DEG to 0·17 mole FA and 0·83 mole saturated acid.
The saturated acids consisted of the series HOOC(CH2)nCOOH where
n ranged from 2 (SA) up to 8 (SBA). In addition, MSA and thiovaleric
acid were tried. An acoustic technique allowed measurements to be made
of modulus E', tan.5, and velocity of sound, V, by means of a forced
resonance cantilever beam. Frequencies were 150 and 900 Hz.
T ___
FIG. 8.5. Dynamic mechanical relaxations of a crosslinked unsaturated polyester

resin.
Generally, V fell from about 2 x 10- 4 mms- 1 at -100°C to 1 x 10- 4

mm s - 1 at about O°C and reached a plateau of less than 2 x 10- 5 mm s - 1
above + 100°C. Four transitions were noticed in the case of the succinic
resin: at - 46 DC, - 25°C, + 12 °C and +98°C. The transition at - 25 °C
was assumed to be equivalent to Tg • The AA polyester resin was found
to have an additional transition at - 16°C and the SBA at + 12°C.
The relationship between the various transitions and the acid chain length
was observed but not found to be particularly simple. A liquid-liquid
transition (Til transition) (> Tg) in amorphous homopolymers and block
copolymers of S has been studied by torsional braid analysis, and is
thought to arise from a molecular rather than a macroscopic relaxation. 2S
8.7. RELAXATIONS IN THE CONSTITUENT CHAIN
The polyester network is made up of two interlocking chain types, and

it may be useful to consider what transitions are found in the linear
chain structures when these are examined separately.
Homopolystyrene has a Tg of 100°C. The transitions in S-maleic an-
r
hydride alternating copolymers have been studied by Block et al.,26 but
1©
these copolymers do not contain opened maleic rings, but rather have
the structure
H2 -CH-CH----C H
I I I
C~ /CO
o "
242 W. 1::. DOUGLAS AND G. PRITCHARD
Measurements with a Nonius pendulum over the temperature range 140 K

to 460 K showed a prominent f)-transition at about 375 K, and a loss
peak at about 170 K which could be correlated with an analogous dielectric
relaxation observed at 240 K and 10 3 Hz (the frequency difference would
be sufficient to shift the peak considerably). This latter peak was attributed
to the twisting of the unopened ring in the amorphous region. A y-transi-
tion, attributed to the same motion in the crystalline phase, could be
found only by electrical measurements.
More information is available about transitions in linear polyesters.
Four or more methylenic groups have been found to co-operate to produce
a low-temperature, f)-relaxation which is not very pronounced in semi-
crystalline, linear polyesters such as poly(hexamethylene sebacate) but
t
much more prominent in the completely amorphous isomer, poly(2-
f
methyl,2-ethyl,I,3-propylene sebacate), i.e.
O(CH2)8CO' O· CH2 '?'

CH3
CH2· 0
CH 2CH 3 n
The methylenic groups were believed to rotate by 180 ° or 360 ° about

the chain axis. 27 The IX-relaxation for this polyester occurred at - 60°C.
The f)-transition temperature falls as the diollength increases through
the series of linear poly(alkyl terephthalates).28 Recently, the results of
a study of the effect on the f)-relaxation of structural variation of several
aromatic polyesters have been published. 29 It was found that the relaxation
was controlled by the structure of the diol. When Cook and Delatycki 30
measured Tp for a series of crosslinked FA/DEG/SA resins of differing
FA: SA ratios, the values obtained were found to approach the temperature
of the Tg of poly(diethylene succinate) (- 30 0c) as the FA proportion
approached zero. Therefore, these investigators concluded that the
f)-transition temperature in crosslinked polyesters was closely related to
the Tg of the corresponding linear polyesters.
8.8. THE P-TRANSITION IN CROSSLINKED RESINS
The f)-transition temperature has been studied systematically as a function

of polyester molecular weight, of diol, of diacid, and of water content.
It was found that this relaxation could be related to the motion of
the polycondensation segments; its temperature decreased as diollength

increased. 3 0
The polycondensation chains are made more flexible by addition of
ether-containing diols such as DEG, but this is not very effective in
lowering T{i because, although the ether groups have low rotational
energy barriers, polar interactions reduce the mobility of the segments.
There was also very little change in the value of G'/J (max.) as the diol
was progressively lengthened from ED to HD. This suggests that the
phthalate group is mainly responsible for the j3-relaxation. All three
phthalic isomers produce a j3-transition, with the order of increasing
T{i being PA < IPA < TPA. However, the groups causing the relaxation
must be able to relax at temperatures below the Tg , but in some environ-
ments this is not possible. For instance, Isaoka et al. 31 found no relaxation
at all below 219°C for crosslinked poly(diallyl phthalate), but according
to Roshchupkin et al.,32 the more open structure of the crosslinked
polymer of bis-methacrylato-triethylene glycol phthalate
gave a transition at 66°C and 1 Hz.

The ether-type diols had little effect in lowering Tp but it was found
that increasing the length of such diols did increase the specific j3 loss
(i.e. G'/J (max.) per mole of saturated dibasic acid).
A j3-transition is also found when there is no phthalic isomer present;
T{i increases as the SA: FA ratio increases in the series of phthalic-free
resins.
Lenk and Padget 12 drew a different conclusion from investigations
of a series of maleic/bis-acid A2 resins: that the j3-transition was associated
with S crosslink relaxation. This was because the loss tangent is increased
by a reduction in maleic proportion (and hence by an increase in S
bridge length). However, Cook and Delatycki pointed out 30 that this
proposal is not consistent with the observed dependence of the j3-relaxation
on the S concentration. For FA/DEG/SA resins with various SA mole
fractions, these authors explained results similar to those obtained by
Lenk and Padget, by suggesting that the j3-relaxation is the result of
motion of the polyester segments between crosslinks.
8.9. Y AND y' TRANSITIONS IN CROSSLINKED RESINS
Transitions have also been found at much lower temperatures than those
of the f3-relaxation. The fourth transition for crosslinked polyesters is
called y',33 perhaps to avoid confusion with the loss angle b. It wa~
found that the low temperature region is affected by water contained
in the resins, and since all polyesters contain water when first made,
desiccation is necessary.
A low-temperature relaxation zone was found by Witort et al. 34 in
the range - 93°C to - 133°C. The resin used was made by reacting
MA (sometimes partly replaced by PA) with 2,2-di(4-hydroxypropoxy-
phenyl)propane and 1 to 2 mole % polypropylene glycol (Mn = 3000).
The aromatic diol was sometimes partly replaced by PG.
The low-temperature, y-transition was studied by Cook and Delatycki. 33
Conventional FA/PG resins with 40% w/w S were found to exhibit
y-transitions at about -100°C. On desiccation, the tan b peak indicating
this transition was progressively reduced, but a still lower temperature
transition (y' between -190°C and -140°C) increased in prominence.
Complete drying left hardly any sign of the y peak; at the other extreme
it approached a constant value as the water content reached 3.0 %.35
Ty was found to decrease as S concentration increased from 25 % to
70% (for FA/PA/PG resins). The size of the peak was linearly related
to fumarate concentration. This led the investigators to the conclusion
that the y peak requires water to be present and to form bridge structures
with the fumaric units, e.g.
""-/ ""-/
C C
II II
0 0
H H
/ ""-0
0 or
""-H /
H
o CH 0
""-/""-/~
C CH 2
""-/""-/
CH CH
2 2
II
0
These fumarate-water complexes can relax, provided that they are not
sterically hindered. 1~ increases, for instance, when the ester groups

are close enough to induce interchain interactions between fumarate
groups. Insertion of PA results in random structures with greater oppor-
tunities for intrachain ester interactions. (The y'-transition was attributed
to motion of the diol section of the polyester chain.)
Addition of 4 %w/w PG (or xylene) to the resin lowered Tg by plasticisa-
tion, but did not affect the y-relaxation. Polyester chain length was not
found to affect either the 1'- or y'-transition significantly in the crosslinked
reSIn.
The same authors examined the y'-transition further 36 by using resins
of various diol and diacid units. The y'-transition was most prominent
in plots of Gil versus T, especially with dry resins. However, drying did
not affect T" so much as it affected peak height. The peak became
less prominent as the S concentration increased and the polyester
proportion fell.
The 1" peak is due to motions of diol segments. It becomes more
prominent on drying, because water restricts these motions, although
whether it does so by hydrogen bonding or simply by filling spaces
is not known. The y' peak also increases as diol length increases (PG
to HD) and is shifted to lower temperatures.
8.10. THE FRACTURE TOUGHNESS OF CROSSLINKED

POLYESTER RESINS
In this section some recent studies of the fracture of polyester resins

are discussed and reference is made to experiments with pre-notched
specimens such as the single edge notched tensile (SEN), centre-notched
tensile (CN) and double cantilever beam (DC B) types. The propagation
of cracks in thermosetting resins is discussed in Chapter 9 as a separate
topic.
Extensive research has been carried out to determine the fracture
behaviour of metals, ceramics, glass and linear organic glassy thermo-
plastics. Less has been achieved in the case of thermosetting resins, partly
because of their variable and uncertain composition. Another factor is
that thermosetting resins are generally toughened by fibre reinforcement,
which completely alters their mode offailure. Nevertheless, there is growing
interest in the relationship between chemical structure and toughness
in crosslinked polymers. 37 . 38
The fracture surface energy, 1', of a material containing either a through-
thickness edge crack of length a, or a through-thickness central crack
of length 2a, is given by
(plane stress)
or
GeE 2Ey
(if = -----''----- = -----'--- (plane strain)
na( I - 1l2)t na( I - 1l2)t
where (if = stress at fracture

Ge = critical potential energy release rate with respect to crack
length at fracture
E = Young's modulus
11 = Poisson's ratio
Berry 39 calculated that y should be about 0·45] m - 2 for linear poly-
(methyl methacrylate), whereas experimental values were found to be
several hundred times greater. (See Table 8.2.) The fracture surfaces
showed coloration under reflected light. Crosslinking gradually reduces
y to a value very much smaller than before, which suggests that viscous
flow is responsible for the additional work of fracture. The values given
in Table 8.2 for unsaturated polyester resins are great enough to suggest
that some flow still occurs, despite crosslinking.
TABLE 8.2
FRACTURE SURFACE ENERGY OF POLy(MMA) AND OF
UNSATURATED POLYESTER RESINS
Polymer Fracture Rejerence

sUljaee energy
}'
(Jm- 2 )
Poly(MMA) [uncrosslmked] 300 39

Poly(MMA) [uncrosslinked] 188 40
Poly(MMA) [uncrosslinked] 500 41
Poly(MMA) [lightly crosslinked] 80 42
Poly(MMA) [crosslinked] 46 43
Poly(MMA) [densely crosslinked] 23 42
Unsaturated polyester 42 44
The flow occurs at the tip of the propagating crack, but it is uncertain
whether this flow involves movement of uncrosslinked (or lightly cross-
linked) regions with respect to denser micelles, or whether it involves
more extensive movement within a single, dense and highly interconnected
structure. (See Fig.!.!.)
Owen and Rose 46 determined several fracture toughness parameters
for mixtures of a conventional low reactivity polyester resin (molar ratio
MA = I, PA = 2, PG = 3) with a flexibilising polyester resin of identical
alkyd: S ratio (the alkyd was made from MA, AA and PG). Measurements
were made in tension using CN plates. It was found that Gc was hardly
changed by the addition of 15 % w/w flexibilising resin, but it increased
tenfold on addition of 30 %. Young's modulus declined steadily, but
Poisson's ratio passed through a minimum at 15 % addition. As the
resin blend became more flexible, the difference between values of the
critical stress intensity factor, KIt' obtained using thick specimens, and
the corresponding values (KJ obtained from thin specimens, increased.
This was attributed to the increasing tendency for the thin specimens
to fail by a mixture of modes rather than by 'mode I' opening.
The poly(propylene maleate adipate) flexibiliser was reactive and
capable of being incorporated into the network, so apart from unspecified
differences in chain length, acid number, etc., the main difference between
the various blends was the ratio of aromatic phthalic to aliphatic adipic
sequences in the polyester condensation chains. (The ratio of MA to
AA in the flexibiliser was not stated.)
This indicates that increasing the mobility of polyester segments greatly
increases the work of fracture. A similar increase in mobility can be
achieved by large increases in crosslink spacing, leading to a change
from brittle to ductile behaviour. 47
Christiansen and Shorta1l 48 carried out rather similar investigations
using a PA-based flexibiliser. This additive did not lower the modulus
so drastically and was probably more akin to the base resin than was
the flexibiliser used by Owen and Rose. Fracture energy increased from
22·8 J m - 2 to 28· 5 J m - 2 at 20 % addition of flexibiliser. These workers
also used the expression,49.so
D = ~(Kc)2
8 (Jy
to obtain the value of the Dugdale plastic zone length D in terms of

the compressive yield stress(Jy. Assuming /1 = 0·35, the value of D for
the medium reactivity orthophthalic base resin was calculated to be
5· 26 11m. An estimate of the radius of the assumed plastic zone r y , obtained

from the expression, 51
ry = 6~ (~cr
was 0·8I1m. These values can be compared with the considerably larger
estimated sizes of dense globular micelles commonly reported to be con-
tained in thermosetting resins. The critical flaw size (c) obtained by
applying the Griffith equation
2 2Ey
O"f = nc(l _ 112)
to polished tensile specimens was about 2211m, which is similar to the

size of the smallest globular aggregates mentioned above. Linear polymeric
glasses have higher critical flaw sizes; for example, polystyrene has
c = lOOOl1m, and polymethyl methacrylate has c = 50 11m. Crosslin king
reduces these values considerably. Abeysinghe et af.52 obtained the value
c = 3811m for a polyester resin formulated from equimolar quantities
of MA, PA, PG and DEG.
In practice, flaws are often holes within materials and not through-
cracks. Polyester resins sometimes develop internal disc-shaped cracks
after prolonged immersion in hot water. The distribution of stress in the
neighbourhood of a disc-shaped internal crack, of diameter a, within
an elastic solid is given by,53
By use of this expression, a critical diameter for these osmotically generated

disc cracks was found to be 90 11m for the resin mentioned above. 52
The fracture toughness of unsaturated polyester resins depends on
several structural factors. Rhoades 54 found that K,c falls progressively
as the maleic: phthalic ratio is increased from I : I to 1: O. I t had previously
been found that KIc falls after prolonged immersion in hot water 55 but
Rhoades concluded that the extent of this susceptibility to water is itself
structure-dependent, being more noticeable at high MA: PA ratios. The
cure temperature affects the crosslink density, and it is not surprising
that a low cure temperature favours a high fracture energy.48 Some
typical values of the fracture toughness of unsaturated polyester resins
are given in terms of KIc in Table 8.3.
TABLE 8.3
CRITICAL STRESS INTENSITY FACTORS OF UNSATURATED POLYESTER
RESINS
K/c Specimen MA:PA Reference

(MN m- 3 / 2 ) geometry mole ratio
0-49 DeB (probably) I: I 48

0·72 eN 1:2 46
0·84 SEN 3: I 55
0·71 eN I: I (contained PG and DEG) 52
0·71 SEN 1:1 54
8.11. TOUGHENING WITH ELASTOMERIC PHASES
Attempts have been made to increase the fracture toughness of unsaturated

polyester resins by the same method as that used for polystyrene, i.e.
by addition of an elastomeric phase. The use of carboxyl-tipped butadiene
acrylonitrile rubbers has been successfully established for the toughening
of thermosetting epoxide resins, but compatibility problems arose when these
elastomers were first blended with unsaturated polyester resins. Whereas
epoxy resins can be toughened fifty-fold, the improvements achieved
with polyesters are more commonly three-fold or less. Tetlow et al. 56
added butadiene-acrylonitrile rubbers with various end-groups (carboxyl,
vinyl, amine) to a highly reactive I PA polyester resin and measured
the fracture toughness by using DeB fracture toughness specimens. In
general, addition of vinyl terminated elastomers led to the greatest increase
in fracture toughness, although even in these cases the enhancement
was less than 100 % at 8 % rubber addition. The combination of both
rubber and added flexibilising resin had no greater effect. The theory
of rubber toughening described by these investigators is that rubber
particles of at least l.um in diameter induce crazing with void formation,
whilst particles of size ~O·I.um induce deformation by shear banding.
The optimum size distribution is therefore bimodal, with the two
mechanisms operating jointly, so that shear banding prevents crazes from
developing into cracks. According to this theory, very highly crosslinked
resins are the ones with least molecular mobility and are thus the least
able to be toughened by such mechanisms. Therefore, the resins most
in need of toughening are those for which the improvement is most
difficult to achieve.
Butadlene- S block copolymers have been used to toughen polyester

resins , with better effects than those obtained by use of uncured
S-butadiene rubber. 5 7
8.12. FRACTURE SURFACES
The fracture morphology of crosslinked polyester resins resembles that

of glass in many respects .48 The fracture surfaces of simple tensile speci-
mens show small regions of smooth, mirror-like appearance and small
mist regions, followed by hackle and, eventually, crack branching. The
initiation site is often at the surface and, from this site, long narrow
'river lines' are occasionally found , passing through the mirror into the
mist region. These river lines are caused by crack propagation at different
levels.
Different features are seen according to whether the specimen geometry
favours stable, i.e. not self-propagating, or unstable crack growth. Owen
and Rose 58 examined the surfaces of eN fracture toughness specimens,
which contained only a small region of river markings adjacent to the
crack tip, followed at a distance by a smooth surface characteristic of
slow, unstable propagation and, eventually, by a region of greater rough-
ness. This region contained conic markings, caused by the spread of
secondary fractures, in a plane close to , and parallel to, the main fracture
mirror ron
,
notch
FIG. 8.6. The position of the mIrror zone III the fracture surface of a SEN sample
of an unsaturated polyester resin, broken in tension.
plane. With the addition of increasing proportions of flexibiliser, the

river markings became more widespread and regular.
The surfaces of specimens broken in tension after crack tip production
by cyclic loading were similar to those of CN fracture toughness specimens.
Some of the features found on polyester resin fracture surfaces are
indicated schematically in Fig. 8.6. The rough region is in most cases
related to increased propagation velocity. Flexible resins were found to
have larger smooth regions of slow, unstable crack propagation. 58
Christiansen and Shorta1l 48 found that the rough hackle region did not
appear in tapered DeB specimens.
Rhoades 54 examined the fracture surfaces of three specimens of a
2:1 MA: PA resin (Fig. 8.7). The top, relatively uncomplicated surface
FIG 8.7. TenSIle fracture surfaces of samples of a 2:1 MA:PA unsaturated

polyester reSIn (top) SEN; (mIddle) unnotched; (bottom) unnotched but contaInIng
a promInent flaw 54
was that of an SEN specimen broken in tension, whilst the middle surface
was of a similar specimen without the notch; this surface appears much
more complex. The bottom surface was also an unnotched tensile specimen,
but the fracture initiation site can be seen to be a void inadvertently
introduced during casting.
As mentioned earlier, polyester resins immersed in water for long periods
frequently develop disc cracks. Figure 8.8 shows the fracture surface
FIG. 8.8. Fracture surface of half of a eN unsaturated polyester resin specimen,

broken in tenSIOn after 6000 h immerSIOn III distilled water at 65°C. The figure shows
the central regIOn ( x 215).
of a eN specimen, broken in tension after 6000 h immerSIOn in distilled

water at 65°C. The figure shows a close-up of one part of the region
midway between the notch and the far edge of the specimen. The develop-
ment of disc cracks releases strain energy and is accompanied, at least
initially, by a small increase in the fracture toughness.
Mushkatel and Marom 59 found that boiling of fracture toughness
specimens in water before fracture was carried out had the effect of
converting the surface morphology to a smooth, mirror-like form, but
if part of the S was replaced by ring-substituted BS, the water-boiling

treatment no longer had this effect. BS appeared to suppress the develop-
ment of water-induced microcracks about 0·02 mm wide. Scanning electron
micrographs showed a little crazing at low BS concentrations but none
at all at higher levels despite 107 h boiling. Also, the retention of fracture
initiation energy (del ermined by three-point bending of notched bars)
was improved.
This was explained by the theory that large crosslinking monomers
such as BS, chlorostyrene and vinyl toluene produce a relatively open
structure, allowing water molecules to penetrate without generating exces-
sive internal stresses. Craze resistance has long been thought to be inversely
proportIOnal to density. 60
Improved resistance to microcracking was also shown by the BS resins
after combined humidity and ultravIOlet treatment, but photo-oxidative
degradation was increased.
Fracture surfaces of polyester resins broken in the three-point bending
mode are also discussed by Kitoh et a/. 61 Resin formulations ranged
from FA:DEG molar ratios of 1: I to FA:AA:DEG of 1:4:5. Agam,
mirror and rough regions were observed. A very wide range of crosshead
speeds were employed (3 to 120000mm/min).
8.13. CONCLUSIONS
Crosslinked unsaturated polyester resins exhibit four dynamic mechanical

relaxations. The IX-transition is believed to involve the whole network,
while the p-transition is probably caused by polyester mobility although
present views are not unanimous on this point. The y-transition is caused
by water-bridges linking ester groups in neighbouring polyester chains,
and the lowest temperature, y' -transition is caused by movement of diol
segments.
Most polyester resins, except those containing high proportions of
linear aliphatic acid, or high proportions of diols containing ether groups,
undergo brittle fracture under normal conditions. The fracture toughness
falls as the degree of crosslinking increases. Characteristic fracture surfaces
are glass-like. The effect of water immersion on these surfaces has been
studied. Addition of an elastomeric phase has very little beneficial effect
on the toughness of polyester resins.
REFERENCES
1. BOENIG, H. V. Unsaturated Polyesters, 1964, Elsevier, Amsterdam.

2. iMAI, T. J. Appl. Polym. Sci., 1967, 11, 575.
3. PRITCHARD, G., Ph.D. thesis, The crossllnking oj polyesters: a study oj network
formation by physical methods, 1968, University of Aston in Birmingham,
England.
4. HEBERT, N. T. Plast. World, 1975,33,47.
5. SEEFRIED, C. G., JR. and KOLESKE, J. V. J. Test. and Eval., 1976,4,220.
6. DEKKING, P. Determination oj dynamic mechanical properties oj high polymers
at lowjrequencles, 1961, Report from TNO Laboratory, Delft, Netherlands.
7. LEARMONTH, G. S. and PRITCHARD, G. Soc. Plast. Eng. J., 1968, 24,47.
8. SOFER, G. A., DIETZ, A. G. H. and HAUSER, E. A. Ind. Eng. Chem., 1953,45,
2743.
9. WARD, I. M. Mechanical Properties oj Solid Polymers, 1971, Wiley, London.
10. HAWARD, R. N., ed. The PhysIcs oj Glassy Polymers, 1973, Applied Science
Pubhshers Ltd, London.
11. COOK, W. D. and DELATYCKI, O. J. Polym. SCI., Polym. Phys. Ed., 1974, 12,
1925.
12. LENK, R. S. and PADGET, J. C. Europ. Polym. J., 1975, 11,327.
2111.
14. GRAESSLEY, W. W. Macromolecules, 1975,8,865.
15. BLANCHARD, A. F. and WOOTON, P. M. J. Polym. SCI., 1959,34,627.
16. COOK, W. D. Europ. Polym. J., 1978, 14, 721.
17. FUNKE, W. J. Polym. SCI., 1967, Part C3, 1497.
18. MAIZEL, N. S., MOZZHECHKOVA, N. I., TSVETKOVA, 0 S., GURINOVICH, L. N.
and MIKHAILOVA, Z. V. Vysokomol. Soedln., 1978, A20, 636.
19. COOK, W. D. Europ. Polym. J., 1978, 14,715.
20. DI MARZlO, E. A. J. Res. Wat. Bur. Stds., 1964, 68A, 611.
21. Fox, T. G. and LosHAEK, S. J. Polym. SCI., 1955, 15,371.
22. GRIEVESON, B. M. Polymer, 1960, 1,499.
23. POCHAN, J. M. and HINMAN, D. F. J. Polym. SCI., Polym. Phys. Ed., 1975, 13,
1365.
24. PEREPECHKO, I. I., SVETOV, A. YA., SEDOV, L. N. and SAVICHEVA, O. I.
PlastlcheJkle Massy, 1974,1,42.
25. GILLHAM, J. K. Polym. Eng. SCI., 1979, 19,749.
26. BLOCK, H., COLLINSON, M. E. and WALKER, S. M. Polymer, 1973, 14,68.
27. FROIX, M. F. and GOEDDE, A. O. J. Macromol. SCI., Phys., 1975, B11, 345.
28. FARROW, G., McINTOSH,J. and WARD, I. M. Makromol. Chem., 1960,38,147.
29. URALIL, F., SEDEREL, W.,ANDERSON, J. M. and HILTNER,A. Polymer, 1979,20,
51.
30. COOK, W. D. and DELATYCKI, O. Europ. Polym. J., 1978, 14,369.
31. ISAoKA, S., MORI, M., MORI, A. and KUMANOTANI, J. J. Polym. SCI., 1970,8,
Part A I, 3000.
32. ROSHCHUPKIN, V. P., KOCHERVINSKII, V. V, SEL'SKAYA, O. G. and KOROLEV,
G. V. Vysokomol. Soedin., 1971, B13, 497.
1049.
34. WITORT, I., SLUPKOWSKI, T., CHOMKA, W. and JACHYM, B. Polimery Tworzywa
Wlelkoczasteczkowe, 1976, 21, 155.
1967.
1953.
37. PRITCHARD, G. and RHOADES, G. V. Mater. SCI. Eng., 1976,26, I.
38. YOUNG, R. J. Chapter 6 m Developments In Polymer Fracture-l, ed. Andrews,
E. H., 1979, Applied Science Publishers Ltd, London.
39. BERRY, J. P. J. Polym. SCI., 1961,50, 107.
40. WEICHERS, W., M.Sc. thesis, The fracture properties of thermosetting resins,
1969, Loughborough Ulliverslty, England.
41. DAVIDGE, R. W. and TAPPIN, G. J. Mater. SCI., 1968,3,165.
42. BROUTMAN, L. J. and MCGARRY, F. J. J. Appl. Polym. SCI., 1965,9,609.
43. BERRY, J. P. J. Polym. SCI., 1963, 1, Part AI, 993.
44. DIGGWA, A. D. S. Polymer, 1974, 15, 101.
45. PRITCHARD, G. and RHOADES, G. V., unpublished data, Kmgston Polytechlllc,
England.
46. OWEN, M. J. and ROSE, R. G. J. Phys. D. Appl. Phy~., 1973,6,42.
47 MATVEYEVA, N. G., ZEMSKOVA, Z. G., SIVERGIN, Yu. M. and BERLIN, A. A.
Vysokomol. Soedln., 1974, A16, 588.
48. CHRISTIANSEN, A. and SHORTALL, J. B. J. Mater. SCI., 1976, 11, 11l3.
49. WILLIAMS, J. G. into J. Fracture Mechs, 1972,8,393.
50. MARSHALL, G. P., COUTTS, L. G. and WILLIAMS, J. G. J. Mater. SCI., 1974,9,
1409.
51. KNOTT, J. F. Mater. SCI. Eng., 1971,7, I.
52. ABEYSINGHE, H. P., PRITCHARD, G. and ROSE, R. G. unpublIshed data,
Kmgston Polytechnic, England.
53. SNEDDON, I. N. Proc Roy. Soc. Lond., 1946, A187, 229.
54. RHOADES, G. V., Ph.D. theSIS, The ejject of the structural parameters oj
polyester reszns on the mechanzcal properties oj polyester mouldzng compounds,
1979, Kmgston Polytechnic, England.
55. PRITCHARD, G., ROSE, R. G. and TANEJA, N. J. Mater SCI., 1976, 11,718.
56. TETLOW, P. D., MANDELL, J. F. and MCGARRY, F. J. 34th SPi Reznjorced
Plasllcs/Composlles Conj., Ne~r Orleans, La., USA, 1979, Paper 23-F.
57. HAYASHI, Y., IBATA, J., TOYOMOTO, K. and UTA, B. Japan. Kokal 7430,480.,
Chem. Abst., 81: 107084e.
58. OWEN, M. J. and ROSE, R. G. J. Mater. SCI., 1975, 10, 1711.
59. MUSHKATEL, M. and MAROM, G. J. Appl Polym. SCI., 1976,20,2979.
60. HAWARD, J. B. Soc Plas!. Eng. J., 1959, 15,379
61. KITOH, M., MIYANO, Y. and SUZUKI, K. Kobumhl Ronbunshu Eng. Ed., 1976,
5, 122.
Chapter 9
CRACK PROPAGATION IN THERMOSETTING

POLYMERS
ROBERT J. YOUNG
Queen Mary College, University oj London, UK
SUMMARY
Thermosetting polymers are widely used as matrix materials in fibre-

reinjorced composites. They are generally brittle and their jracture
properties are of considerable interest. Recent developments in the study
oj crack propagation in thermosetting polymers have been reviewed using
a linear elastic jracture mechanics (LEFM) approach. It has been shown
that considerable advances have recently been made In understanding the
criteria controlling crack propagation. It appears that for both continuous
and stick/slip propagation the criterion jor failure is that a critical stress
oj the order oj 3-4 times the Yield stress of the material must be achieved
at a crillcal distance ahead oj the crack. The possibility oj crazing and
the ejfect oj microstructure upon crack propagation have also been
discussed.
9.1. INTRODUCTION
Over recent years the subject of crack propagation in thermosettmg resins

has received considerable interest. 1 - 17 Thermosetting polymers are being
used increasingly as adhesives and matrix materials in reinforced plastics.
Because of this, knowledge of their mechanical properties is extremely
important. In particular it is desirable to know how, and under what
conditions, cracks propagate in these materials so that the properties
of the adhesive joint or composite can be predicted. Although the ultimate
257
258 ROBERT J. YOUNG
properties of reinforced polymer composites are dominated by the strength

of the reinforcing medium (usually fibres), it is highly likely that the
long-term strength and performance of these materials may be controlled
by crack propagation in the matrix.
The general subject offracture in thermosetting resins has been reviewed
recently by the author in another book in this series.! It was shown
that considerable advances had been made in the understanding of fracture
in these materials. The fracture of typical thermosets such as epoxy
resins and polyester resins was reviewed and it was shown that although
a considerable amount of experimental data had been collected there
was a notable lack of understanding of the mechanisms of crack propaga-
tion. This chapter is concerned with advances that have recently been
made in explaining how and why cracks propagate in thermosetting poly-
mers and, in particular, understanding the mechanisms of crack propaga-
tion in epoxy resins.
9.2. FACTORS AFFECTING CRACK PROPAGATION
9.2.1. Fracture Mechanics Testing

The most convenient way of studying crack propagation in thermosetting
resins is through the use of linear elastic fracture mechanics (LEFM).18
For LEFM to be applied to a particular material it is necessary that
the stress-strain curve of the material is linear to fracture, and that
any plastic deformation is confined to regions close to the crack tip.
Since thermosetting polymers are brittle materials showing only very
limited plastic deformation during tensile loading their mechanical proper-
ties normally meet the requirements of LEFM. It has been shown l that
a variety of test pieces can be used to monitor crack propagation in
brittle polymers. However, the most versatile specimens are those for
which the stress intensity factor1.8 is independent of crack length. Such
specimens are the double torsion (DT) and tapered double cantilever
beam (TDCB) and they have been used widely for the study of crack
propagation in thermosets. 15 - I 7 They are shown schematically in Fig. 9.1.
One of the most characteristic aspects of crack propagation in thermo-
sets is that, under certain circumstances, cracks propagate in a stable,
continuous manner and that at other times they propagate in a stick/slip,
discontinuous way. These types of behaviour are shown in Fig. 9.2 for
cracks propagating in DT or TDCB specimens. In these test pieces a
constant cross-head speed leads to propagation at a fixed load (or stress
CRACK PROPAGATION IN THERMOSETTING POLYMERS 259
(a)
(b)
FIG. 9.1. Schematic diagrams of two of the vanous fracture mechanics specimens
that have been used with thermosetting polymers. (a) Double torsIOn (DT). (b)
Tapered double cantilever beam (TDCB).
intensity factor, KIc ) when the specimen is undergoing continuous

propagation. They allow excellent control of the way in which cracks
propagate. For example, if the cross-head speed is reduced the crack
slows down and if the cross-head speed is increased the crack speeds
up. Also slow crack growth can be followed in some materials when
the cross-head is held at a fixed displacement. This has been done for
PM MA (polymethyl-methacrylate)19 and it has enabled the crack velocity
to be measured as a function of the stress intensity factor (KIc ), over
a wide range of velocity, using only one specimen. If the material is
Load Load
p - - - -
a b
V Displacement
or Time
Displacement
or Time
FIG. 9.2. Schematic load/displacement or load/time curves for crack propagation

in thermosetting resins. (a) Continuous propagation. (b) Stick/slip propagation.
260 ROBERT 1. YOUNG
exhibiting stick/slip propagation the OT and TOCB specimens are again

very useful. In this case the load/displacement curves have characteristic
saw-tooth shape. Propagation takes place at the peak load and ceases
at the minimum load as the crack grows in a stick/slip manner. This
explanation has recently been shown 1.20 to be slightly incorrect and it
is now thought that propagation ceases not at the minimum load but
at some point up to the reloading part of the curve. The use of stable
test pieces has greatly improved our understanding of crack propagation
in brittle polymers and most of the results discussed in this chapter
were obtained using OT or TOCB specimens.
9.2.2. Material Variables

It is now well established I that there are a variety of material variables
which affect the stability of crack propagation and the stress levels at
which it occurs in thermosetting resins. These include the type of resin
and curing agent used, the amount of curing agent, the temperature
of the curing treatment and time of curing. The effect of the type and
quantity (in parts per hundred (phr)) of curing agent used upon the
fracture energies (GIc) of a series of epoxy resins is shown in Table
9.1. The values of Glc quoted are the initiation values when the mode
of propagation is stick/slip and all the measurements have been made
at room temperature. The details of cure schedule, testing geometry and
rate can be found by consulting the original references. It is difficult
to make generalisations from the data in Table 9.1 since they have been
obtained from such a wide variety of resins and curing agents, using
different testing geometries and rates. It can be seen that the values
of Glc are typically of the order of 100 J m - 2 and in some cases much
larger. They are considerably higher than values of GIc obtained for
other thermosetting polymers. I ,II This shows that epoxy resins are some
of the toughest thermosets available and hence are used in the most
critical applications such as in the matrices of high-performance
composites.
The detaIled effect of changmg the amount of curing agent and cure
schedule upon crack propagation in a given resin can only be found
by altering one particular variable at a time. The effects of changing
curing agent content and post-cure period upon KICI (critical stress
intensity factor for initiation) for a OGEBA (bis phenol-A-diglycidyl
ether) epoxy resin cured with triethylene tetramine (TETA) are shown
in Fig. 9.3. It can be seen that the value of Klcl increases as the amount
of curing agent is increased. A similar effect has been found with an
TABLE 9.1
Glc VALUES MEASURED FOR CRACK PROPAGATION IN DGEBA
EPOXY RESINS CURED WITH VARIOUS HARDENERS (AFTER REFS.I
AND 21)
----- ----- --_. ---------
Resin Hardener G h (Jm- 2 ) Ref.

(phr)
Epikote 828 10DETA 86 22

95 MNA + 0·5 BDMA 154 3
27DDM 340 4
14·6 MPD 110 4
4DMP 180 4
27DDM 312 5
CT200 13PA 220 6
DER332 5 PIP 121 23
Various TEPA 52-227 24
Various HHPA 158-262 25
MY750 8·3 EDA 329 9
12·2TDA 489 9
16·IHDA 575 9
11·5DETA 130 9
II·OTETA 141 9
15·0TEPA 136 9
Key to hardeners
BDMA = benzyl dimethylamine
DDM = diphenyl diaminomethane
DETA = diethylene triamine
DMP = tris(dimethylaminomethyl)phenol
EDA = ethylene dJamine
HDA = hexamethylene diamine
HHPA = hexahydrophthalic anhydnde
MNA = methyl nadic anhydride
MPD = m-phenylene diamine
PA = phthalic anhydnde
PIP = piperidine
TDA = tetramethylene diamine
TEPA = tetraethylene pentamine
TET A = triethylene tetramine.
increase in the post-cure temperature.!7 It appears, therefore, that, at

least for this particular system, Klcl increases as the amount of crosslinking
in the resin (characterised by the degree of cure) is increased. At first
sight these observations would appear to be in conflict with those of
other workers 2 ! who studied crack propagation in similar DETA-cured
262 ROBERT 1. YOUNG
K1 MNm-3/2
15
10
(a)
05
phr
0
0 5 10 15 20
(b) 05
Tlme(h)
5 10 15 20 25
FIG. 9.3. (a) Variation ofthe critical stress mtensity factor for initiation, Klcp with
parts per hundred (phr) of TETA curing agent for a DGEBA epoxy resin (data
taken from Ref. 26). (b) Variation of KIc' with post-cure period for a TETA cured
epoxy resm post-cured at 100°C (data taken from Ref. 17).
1.0
(a) 0.6
08
~ (b)
1.0
•
I
•
••
04
0.2 phr Time(h)
0
6 8 10 0 20 40 60 80
FiG. 9.4. (a) Variation of critical strain energy release rate for crack initiation,
G,c" with phr ofDETA curing agent for a DGEBA epoxy resin (after Ref. 20. (b)
Variation of G1c1 with post-cure time for a DETA cured epoxy resin post-cured at
106°C (after Ref. 21).
systems and found that Glcl peaked at particular curing agent contents
or post-cure periods. The results of Mijovic and Koutsky21 are shown
in Fig. 9.4. They have employed considerably longer cure schedules than
those for the specimens used in Fig. 9.3 and it must be remembered
that the relationship 18 between Klc and Glc (K2:::: EG) involves the
Young's modulus, E, which can also vary with cure. 26 These factors
may help to reconcile the apparent discrepancies between the data in
Figs. 9.3 and 9.4.
9.2.3. Testing Variables

Although it is clear from the previous section that crack propagation
behaviour in different systems varies with the composition of the resin,
the most significant advances in understanding the mechanisms of crack
propagation have been made by following the effect of changing testing
KI MNm-3/2 50°C
0.8 ••• •
0 0-0
i • •
0 0
0.6 0 0
.
100°C
0.8
~~
0
0.6 0-0-0--
t:J
1.2 ~ 150°C
1.0
0.8
~
0 0
0.6 -0-0 0
0
dy/dt mm/mln
0.05 0.5 5
FIG. 9.5. Variation of K,,, (e) and KlLd (0) with cross-head speed for a TETA
cured epoxy resin post-cured for 3 h at different stated temperatures, all tested at
room temperature (after Ref. 17).
264 ROBERT J. YOUNG
variables such as rate 2 . IS and temperature lS -I 7 and keeping the com-

position of the resin and curing conditions constant. The variation of
the critical stress intensity factors for initiation and arrest, KIc' and
K,w for the same composition of resin, cured at different temperatures
for the same length of time and tested at different cross-head speeds
(i.e. rates), is given in Fig. 9.5. It can be seen that there is a tendency
for the size of the jumps to decrease as the cross-head speed is reduced,
and for propagation to be continuous at high cross-head speeds and
stick/slip at low rates of testing. This effect has been shown to be similar
for a wide variety of thermosetting resins when they are tested at different
rates.1.2· l s -17 This behaviour is clearly unusual and any theoretical ex-
planations of crack propagation in thermosetting polymers must be able
to account for this effect.
10 K MNm-3/2 71.phr
Ie
o'--o _ _ el=O
05
O~--~--~ __~--~----~
15
10
05
O ~--~--~--~--~----~
15
10
o •
05
'---o-o-o~
OL---~--~--~--~~--~
100 -50 a 50 100 ISO
FIG. 9.6. Variation of K,c with testmg temperature for epoxy resins con taming
different amounts ofTETA . • K,C>' 0 K,w 0 K,c continuous. Data obtained using
a DT specimen at a cross-head speed of O· 5 mm min - I .
It is possible to shed more light upon the mechanisms of crack propaga-

tion by looking at the effect of changing the temperature of testing.
Figure 9.6 shows the variation of Klcl and Klcd for three different formula-
tions of an epoxy resin cured under identical conditions but tested at
different temperatures. It can be seen that in all cases propagation is
continuous at low temperatures but stick/slip at higher temperatures.
This behaviour has been shown to be applicable to other formulations
of epoxy resins 9 . 16 . 1 7.27 and is consistent with the effect of testing rate.
Since epoxy resins are viscoelastic solids it would be expected that reducing
the rate of testing would be similar to increasing the testing temperature.
Both of these factors appear to promote stick/slip behaviour.
Some recent observations by Scott el al. 27 have shown that the picture
may be somewhat more complex than was previously thought. They
extended measurements of KIP as a function of testing temperature,
down to liquid nitrogen temperature ( - 196 QC). They found that below
about 0 QC propagation became continuous, which is consistent with
previous observations. 9.15 - 17 It remained continuous down to about
- 100 QC but became discontinuous (stick/slip) again below this temperature.
Some of their results are given in Fig. 9.7. This low-temperature stick/slip
propagation has been observed below - 100 QC in a series of amine-cured
epoxy resins . 27 It appears to be different from the stick/slip propagation
at high temperatures as the size of the jump increases as the temperature
IS reduced. The origm of this low-temperature stick/slip behaviour is
•
250
200
150
100
50~ ______~______~____~______ ~__°c

TEMPERATURE ~
-200 -150 -100 -50 0
FIG. 9.7. Variation of Gic with testing temperature for an EDA-cured epoxy res III
tested at different temperatures (after Ref. 27).
266 ROBERT J. YOUNG
not known. The cooling agent used by Scott et al. 27 was nitrogen, N 2 ,
and this could be responsible for the observed effect. It is well known
that gases such as N 2 can affect the crazing and fracture of thermoplastics
at low temperatures 28 and it is possible that they have a similar effect
upon the epoxy resins. A critical test to prove this is to look at the
effect of cooling with helium gas upon the crack propagation. This gas
does not affect thermoplastics and so testing epoxy resins in a helium
environment would be of great interest.
It has been suggested 29 that the high-temperature stick/slip propagation
in epoxy resins may also be due to an environmental effect such as
the absorption of water vapour from the atmosphere. It appears that
this is not the case 30 since stick/slip behaviour is found when cracks
are propagated under vacuum. I 5 It will be shown later that recent investiga-
tions have indicated that this type of stick/slip propagation is due to
crack blunting which is controlled by the plastic deformation character-
istics of the resin.
9.3. MECHANISMS OF CRACK PROPAGATION
It has been recognised for several years that crack propagation in thermo-
plastics, such as PMMA, takes plal:e through the breakdown of crazes. 31
There have also been suggestions that crazing occurs in thermosetting
resins such as epoxies.32.33 The evidence for this has been discussed
in a previous publication 1 and it seems unlikely that in fully cured resins
the amount of tensile drawing required to form a craze would be possible.
Recent experimental observations 21 have tended to suggest that crazing
does not normally occur. It has also been suggested 21 that glassy polymers
such as epoxy resins have a nodular structure and that this affects crack
propagation.
9.3.1. Fractography
Examination of the fracture surfaces of epoxy resin samples using optical
and electron microscopy has enabled both the possibility of crazing and
the effect of the nodular structure to be investigated. Previous publica-
tions 9 . 17 have shown that there are several characteristic features that
can be seen on the fracture surfaces of epoxy resins. When crack propaga-
tion is continuous the surfaces tend to be smooth and relatively featureless.
On the other hand, when stick/slip propagation occurs there are features
on the surface in the vicinity of the crack arrest lines. These features
(al
(b)
FIG. 9.8. Fracture surfaces of an epoxy resin cured with 9'8phr ofTETA. The
specimen was post -cured at 50°C for 3 h and tested at 22°C. (a) Optical micrograph
of surface. (b) EM replica of crack arrest line. (Crack growth direction indicated by
arrows.)
268 ROBERT 1. YOUNG
fall into three main categories: (I) triangular features, (2) fine arrest lines
and (3) broader rough areas. 17 The occurrence of each type of feature
depends upon the state of cure of the resin and on the temperature
of testing.
The triangular markings are found typically in under-cured resins,l 7
i.e. due to low curing temperature and/or small quantities of curing
agent. They appear along rather indistinct arrest lines as shown in
Fig. 9.8(a). The electron micrograph in Fig. 9.8(b) was obtained from
the same type of area and it can be seen that the surface features are
streaked in the direction of crack propagation. There is also a broad
band along the arrest line and an underlying nodular structure on the
scale of ~ 1000 A.
Fine arrest lines are obtained on the surfaces of specimens just under-
going the transition from stable to unstable propagation. 26 Figure 9.9
shows an optical micrograph and an electron micrograph of the fracture
surface of a specimen at the transition. The fine arrest line can be seen
in Fig. 9.9(a), and Fig. 9.9(b) shows the same line at a higher magnification.
I t can be seen that the structure is again streaked and the arrest line
corresponds to an abrupt change in the direction of the streaks. An
underlying nodular structure on the scale of ~ 500 A can also be
resolved.
A typical example of the broader type of crack arrest line is given
in Fig. 9.10(a). This type of feature is typical of well-cured specimens
fractured at high temperature. 1 7 The fracture surface up to the crack
arrest line is relatively featureless. After crack arrest there is a slow-growth
region 9 of closely spaced striations parallel to the crack growth direction.
There is, then, a rough hackled area where the crack accelerates during
the 'slip' process. Examination of the fracture surface through electron
microscope replicas has shown that the smooth areas of such specimens
are relatively featureless, in contrast to the slow growth region which
is shown in Fig. 9.IO(b). In this area there are V-shaped features which
appear to be caused by the crack propagating on different levels. It
is not normally possible to resolve any underlying nodular structure
in well-cured specimens such as that used in Fig. 9.10.
It is clear that there is no evidence of any craze debris on the surfaces
of the epoxy resins shown in Figs. 9.8-9.10. Mijovic and Koutsky21
came to a similar conclusion for DETA-cured resins. It seems likely,
therefore, that the crazing in epoxy resins suggested by Morgan and
O'Nea1 33 from the examination of thin films, may not occur in bulk
samples.
(a)
lmm
(b)
5!Jm
FIG. 9.9. Fracture surface of an epoxy resin cured with 9·8 phr of TETA. The
specimen was post-cured at !OO°C for 3h and tested at 22°e. (a) Optical
mIcrograph of fracture surface. (b) EM replica of crack arrest line. (Crack growth
direction mdicated by arrows.)
270 ROBERT 1. YOUNG
(a)
100jJm
(b)
I
'/
FIG. 9.10. Fracture surface of an epoxy resin cured with 9·8 phr of TETA. The
specimen was post-cured at 150 a C for 3h and tested at 22 a e. (a) Optical
micrograph offracture surface. (b) EM replica of slow-growth region following the
crack arrest line. (Crack growth direction indicated by arrows.)
Koutsky and co-workers 21 . 34 have recently suggested that the mechani-

cal properties of epoxy resins may be controlled by the nodular structure.
This is an attractive proposition as there is a well established dependence
of the properties of metals upon the microstructure. However, even the
existence of nodules in glassy polymers is a matter of some controversy,
with Uhlmann 35 suggesting that they are artefacts, on one hand, and
Yeh 36 affirming that they represent the true structure, on the other.
Supporters of the existence of nodular structure have problems explaining
the lack of small-angle X-ray scattering that should occur if the nodules
are present in glassy polymers, but even Uhlmann suggests 35 that the
X-ray scattering evidence is ambiguous for epoxy resins. It would seem
reasonable that thermosetting polymers, which are normally made by
the addition of a curing agent to a pre-polymer, could have a non-uniform,
two-phase structure. The micrographs in Figs. 9.8-9.10 and the work
of Mijovic and Koutsky 21 suggest that the nodule size in a given resin
curing agent system varies with the state of cure. Since the properties
of the resin also depend upon the cure it is tempting to relate the mechanical
properties to the nodular microstructure. Mijovic and Koutsky have
recently done thiS and, although they find variations of properties with
microstructure, they have not been able to explain why a particular
nodule size gives rise to particular properties.
9.3.2. Plastic Deformation at the Crack Tip

It is clear that plastic deformation most occur at the crack tip during
crack propagation in thermosetting resins. However, it is extremely diffi-
cult to show directly that plastic deformatIOn has taken place. It has
been pOSSible to see blunt cracks with tiP radii of the order of several
microns 37 but observation of plastiC zones, which must be present, has
proved extremely difficult. Phillips and co-workers 9 have shown that
when stick/slip propagation occurs in epoxy resills there is a characteristic
slow-growth region after the crack arrest line. They observed the growth
of a crack in an epoxy sample and found that after the 'slip' process
the crack became stationary at the arrest line. They found that prior
to the next 'slip' step it grew slowly through a small region similar to
that shown in Fig. 9.10 before bursting through and jumpmg ahead.
It is found that the length of the slow-growth region (I,) increases as
the temperature of testing is raised. This effect can be clearly seen in
Fig. 9.11 for a TETA-cured resin, where micrographs are given for speci-
mens which have been fractured at two different temperatures. It was
shown III Fig. 9.6 that for this formulation the value of KIc\ also increases
272 ROBERT J . YOUNG
(a)
200~m
---~---------------------~--------
200~m
FIG. 9.11. Optical micrographs of the crack arrest region on fracture surfaces of
an epoxy resin hardened with 14·7 phr of TET A and post-cured for 3 h at 100 °C.
The length of the slow-growth region IS given by I,. (a) Specimen fractured at 22 0C.
(b) SpecImen fractured at 60 °C.
2.0
K MNm-3/2
lei
1.5 •
1.0
.•,. •• •
•
0.5
•
lr IJm
a
a 50 100 150 200
FIG. 9.12. Plot of K,,, against I, for the specimens used III Fig. 9.11 and other
formulatIOns of resill rested at different rates and temperatures.
with increasing temperature. Figure 9.12 shows that there IS a unique

correlation between the length of the slow-growth region I, and KIcI
for different epoxy formulations tested under a variety of experimental
conditions. In fact, it is found that Ir is approximately the same as
the radius of a Dugdale plastic zone (rp) calculated using the equation: 18
r =
p
!:(K1C )2
8 (J y
(1)
where (Jy is the yield stress of the resin. Figure 9.13 is a plot of
(K1c ,/(Jy)2 against Ir using some of the data from Fig. 9.12 and values
of (Jy from earlier publications. 16 . 26 The straight line has a slope of 8/n
(~2·55) and represents the relationship between rp and (KIc,/(Jy)2 if a
Dugdale plastic zone is present at the crack tip. The proximity of the
experimentally determined points to the theoretical line strongly suggests
that Ir is closely related to r p.
A clear picture of stick/slip propagation is now emerging. It appears
that during loading after crack arrest a plastic zone forms at the tip
of the crack. Propagation then takes place by slow growth through the
plastic zone followed by rapid propagation through virgin material. The
slow-growth region therefore defines the plastic zone at the crack tip.
Although this is only indirect evidence of the presence of the plastic
274 ROBERT 1. YOUNG
500 (Krc Jc1y)2 IJm

400
• •
300
81n
200 •
100
50 100 150 200

FIG. 9.13. Plot of (K. u /rr)2 against I" using some of the data from Fig. 9.12. The
straight line is drawn with a slope of 8/71. according to eqn. (I), assuming that If ~ rp'
zone it gives an important new insight into processes taking place at

the crack tip.
9.4. CRACK GROWTH CRITERIA
Criteria controlling the propagation of cracks in metals and thermoplastic

polymers 18 have been developed over the past few years and recent ex-
perimental and theoretical investigations have allowed criteria to be
developed to account for crack propagation in thermosets, particularly
in epoxy resins. Slightly different theories must be applied to the two
different types of crack growth encountered in thermosets (continuous
or stick/slip propagation) although it will be shown that the basic criterion
controlling fracture is the same in both cases.
9.4.1. Continuous Propagation

The continuous propagation of cracks in epoxy resins that occurs at
low temperature is somewhat similar to the type of propagation that
takes place in brittle thermoplastics such as polymethyl-methacrylate,
PMMA. 18 . 38 It has been clearly demonstrated that in PMMA propagation
takes place through a constant crack-opening displacement (b) cri-

terion. 18 .38 There is now strong evidence that a similar criterion can
be applied to continuous propagation in thermosetting polymers such
as epoxy resins. 16 The value of b is given by: 16.18.38
b=
(JyE
(K
K~ ~ ey 1C )2
(Jy
(2)
where (Jy is the yield stress, e y the yield strain and E the Young's modulus
of the material. The criterion for continuous propagation is that crack
growth takes place when b reaches a critical value, be' Calculations have
shown that be for continuous propagation in epoxy resins is remarkably
constant,16 but as soon as stick/slip propagation ensues, the value of
be rises rapidly and the constant b criterion no longer applies. The effect
of this criterion upon KJc for continuous propagation can be seen in
Fig. 9.6. In the continuous regions KIc decreases slightly as the temperature
is increased. This is because the yield stress and modulus (Jy and E
are also falling but L\ is remaining constant (eqn. (2)).
In thermoplastics such as PMMA the constant b criterion is associated
with the growth of a crack through a single craze at the tip of the
crack, which can be accurately modelled by a Dugdale plastic zone. 18 .38
The evidence for the presence of crazes at the tips of moving cracks
in fully cured thermosets is not strong 1.21 and it has been shown in
Section 9.3.1 that examination of the fracture surfaces implies that crazing
has not occurred.
9.4.2. Stick/Slip Propagation

There have been recent important developments in the understanding
of stick/slip propagation in thermosetting polymers and it is now thought
that this type of behaviour can be explained in terms of blunting at
the crack tip and a quantitative theory has been developed very
recently37.39 to explain stick/slip propagation. It has been known for
several years 16 that if the resin has a high yield stress then crack propaga-
tion is continuous, whereas if the material has a low yield stress, due
to changes in composition or testing variables, propagation tends to
be stick/slip in nature. Since epoxy resins are brittle, measurements of
yield stress can normally only be made in compression. This behaviour
can be quantified as shown in Fig. 9.14 where K1e, is plotted against
(Jy for a whole range of TETA-cured epoxy resins tested at different
rates and temperatures. The values of KJc' and (Jy have been taken from
this and previous publications. 16 .26 It can be seen that all the data fall
276 ROBERT J. YOUNG
K]C,
•
m-3/2 •
15
10
•
,
05
. 7~ phr
,•
98 ..
123··
• 11.7 ..
O'y -1Po
00 50 100 150 200
FIG. 9.14. Plot of Klc> against (Jy for different formulations of resin tested at a
vanety of rates and temperatures. The values of (Jy have been taken from previous
pubhcatlOns. 16 . 25
approximately upon a master curve indicating a unique relationship

between Klc1 and (Jy. Kinloch and Williams 37 have obtained a similar
correlation between KIc' and (Jy for a series of resins hardened with different
curing agents. This correlation has allowed stick Islip propagation to be
explained quantitatively.39
It can be shown 40 that for a crack under an applied stress of (Jo,
the stress «(Jyy) normal to the axis of the crack at a small distance (r)
ahead of the crack is given by:
(JoJa (l + plr)
= --
(J
YY fo (I + p12r)312
(3)
where p is the radius of the crack tip and a is the crack length. If
it is postulated that the failure criterion is that fracture occurs when
a critical stress «(Jc) is reached at a distance r = e, then eqn. (3) can
be rewritten as:
(1 + pI2e )3/2
(4)
(1 + pic)
The term rrc~ can be considered to be the critical stress intensity

factor for a 'sharp' crack, Klc ,20 and rrfo
as the stress intensity
factor for a blunt crack, KIB • 20 Hence it follows that:
KIB (1 + p/2C)3i2
(5)
+ pic)
-----~
K 1c (l
This equation relates KIB to the radius of a blunt crack and the theory
can be checked by measuring the variation of KIB with p. However,
direct measurement of p is difficult as it tends to be small ('" 10 /lm)
for natural cracks in thermosetting polymers. Kinloch and Williams 37
have overcome this problem by measuring KIB as a function of p for
a series of epoxy resin samples containing pre-drilled holes of large,
known diameter. They showed that a relationship of the form of eqn. (5)
held for these materials.
However, there is still the problem of determining p for natural cracks
in specimens undergoing stick/slip propagation. Kinloch and Williams 37
o ./
1L.7phr
...
/'"
./
(P/2d /2
o ./
o 2 3
" The value of6
FIG. 9.15. (KIBI K Ic ) as a function of (p /2c) 1 2.
5 7
p has been taken as b as
given by eqn. (2). The experimental points have been fitted to the theoretical (solid)
curve by assuming the values of c given in Table 9.2. The curve is asymptotic to the
dashed line of slope = ~.
278 ROBERT J. YOUNG
have shown that extrapolation of the relationship between KIB and p

for artificially drilled holes to natural cracks indicates that p is approxi-
mately the same as the crack opening displacement, 6. Moreover, it
has been shown in Section 9.3.2 that there is strong evidence for a Dugdale
plastic zone at the crack tip during stick/slip propagation, and so p
can be estimated from 6 in eqn. (2) which can then be used to predict
KIB/ KJc using eqn. (5). The theory has been tested by measuring the
ratio KIB/ Klc as a function of (p/2e)1 2 as shown in Fig. 9.15. The theoretical
line is fitted to the experimental points by choosing suitable values of
the critical distance, c, which is the only fitting parameter. Klc has been
taken arbitrarily as the value of K\c at the lowest temperature of measure-
ment and KIB as the value of Klcl at higher temperatures. The agreement
has been found to be equally good for epoxy resins cured with different
curing agents 37 and even for rubber-modified epoxy resins. 37 The values
of critical distance (e ) for these materials and the data in Fig. 9.15
are given in Table 9.2. The values of critical stress (oJ can be deter-
mined from the relationship Kic = (fc~ and they are also given in
TABLE 9.2
DERIVED VALUES OF CRITICAL STRESS ac AND CRITICAL DISTANCE C FOR
DGEBA EPOXY RESINS OF DIFFERENT FORMULATIONS
FormulatIOn a, (MPa) c (J1m) a.. (25°C) a,i a) Ref.

(MPa)
9·8 phr TETA 360 0·60 112 3·2 45

12·3 phr TETA 300 1·10 91 3·3 45
14·7 phr TETA 270 1·60 86 3·2 45
9-4 phr 3 ° Amme 360 0·38 83 4·3 37
IOphr TEPA 495 0·12 117 4·2 37
Rubber-modified 220 8·30 70 3·1 37
Table 9.1. The yield stress (fy of each formulation at 25°C is also given
and it can be seen that the ratio (fcl(fy for each system is approximately
the same ( '" 3-4). The temperature of 25°C is only an arbitrary reference,
but the constant ratio implies that the failure criterion is that a stress
of the order of three or four times the yield stress must be reached
in the plastic zone, regardless of the resin composition or yield stress.
The critical distance, c, varies from 0·1 to lO,um and its significance
is not yet clear.
It is important to mention at this stage that the constant 6 criterion
(Section 9.4.1) for continuous propagation can also be considered as

requring a critical stress at a critical distance. In this case (Jc will be
virtually constant as KIc falls only slightly as the temperature is increased.
Since the yield stress also drops it is also likely that (Jc/(Jy will, again,
be invariant.
9.5. TIME-DEPENDENT FAILURE
Many materials fail under prolonged loading at stresses which they can
sustain, without failure, for short periods of time. This phenomenon
is known as static fatigue and is exhibited by ceramics, glasses 41.42 and
thermoplastics such as PMMA.19 It is thought that static fatigue in
ceramics and glasses is due to environmental attack 41 .42 whereas in PMMA
it is known to be caused by the slow propagation of cracks which takes
place because of the time dependence of the Young's modulus of the
material. 38 It is known that, under certain conditions, thermosetting
polymers are prone to failure at constant load by static fatigue, but
the results of different groups of workers appear to be in conflict. Some
people have observed static fatigue in epoxy resins 43 and others have
not. 9 .44 It seems that this conflict may be resolved by examination of
the careful work of Gledhill and Kinloch 43 .44 who have recently shown
that, at least at room temperature, certain formulations of epoxy resins
appear to show static fatigue while others do not. They looked at the
failure of TDeB aluminium adhesive joints bonded with a thin layer
of epoxy (0·5 mm). This particular test piece is useful because similar
ones made from the pure resin are prone to creep over long periods
of loading which complicates the stress distribution.
Gledhill and Kinloch 43 found that, if the epoxy was cured with a
tertiary amine, the critical strain energy release rate at failure, GIe , dropped
dramatically as the loading time was increased, as shown in Fig. 9.16.
The Gle measured after 10 8 s (~3 years) was only 25 % of that needed
to propagate a crack during short-term loading (over a few seconds).
Using exactly the same testing geometry and testing procedure they found
that when the same epoxy resin cured with tetraethylenepentamine
(TEPA), was used in the adhesive joint, static fatigue was not observed. 44
Moreover, they found that when specimens had been held under a load
corresponding to 86 % of the failure load required to cause crack growth
during rapid loading, there was an increase in failure load measured
on subsequent rapId loading of the specimens. This effect is shown in
280 ROBERT 1. YOUNG
GIe Jm
-2
loglO(follure tl me. 5)
o ~--~--~--~~~--~--~--~--~--~
o 2 3 4 5 6 7 8 9
FIG. 9.16. Relationship between Glc and failure tIme for TDCB adhesIve joints
consistIng of an epoxy resin cured with a tertiary amine (after Ref. 43).
Fig. 9.17 where GIcl (measured on retesting the specimen following static
loading) is plotted against the static loading period. This means that
the epoxy resin is actually becoming tougher on static loading and this
effect has been termed a 'self-toughening mechanism'. 44
The question that must be answered is, why does the same resin hardened
with different curing agents show such a difference in behaviour? The
answer almost certainly lies in the behaviour of the two systems during
testing at a constant loading rate. The tertiary amine-cured resin shows
stick/slip propagation at room temperature, 43. whereas, there is continuous
150
•
100
50
loglO(loadlng tlme.s)
o~--~--~--~--~--~--~--~--~
o 2 3 4 5 6 7 8
FIG. 9.17. Increase in the value of GIc WIth testIng tIme for TDCB adhesIve JOInts
conSIsting of an epoxy resin cured WIth TEPA (after Ref. 44).
propagation in the TEPA cured resin.43 This means that cracks are
relatively blunt even during short-term loading in the tertiary amine-cured
material, and this is reflected in the significantly higher values of GIc
that this material can sustain over short periods of time. On the other
hand since the TEPA-cured resin undergoes continuous propagation at
room temperature the cracks are relatively sharp leading to low values
of GIc during short-term loading. However, as the loading period increases
blunting starts to take place and so the value of GIc increases. Indeed,
it is thought that the blunting only occurs for the initial crack since
after the initial 'jump', propagation takes place in a continuous way
as for an unprestressed specimen. Increasing the period of loading is
equivalent to either reducing the testing rate or increasing the temperature,
both of which have the effect of causing stick/slip propagation and hence,
promote blunting. The drop in Gic observed for the tertiary amine cured
resin has been explained in terms of propagation taking place when
a critical plastic zone size or crack opening displacement is achieved.
The yield stress and modulus of the resin will drop as the loading time
increases, and so the critical conditions can be reached at a lower value
of G,c'
It is worth considering at this stage if the two types of behaviour
can be reconciled. Examination of Figs. 9.16 and 9.17 shows that the
lowest value of G,c found for both systems is about the same ( - 50 ] m - 2),
and it may be that the two plots are representing similar types of behaviour.
Clearly measurements over longer periods of time are required to see
if, eventually, a limiting value of G,c will be reached for the tertiary
amine-cured resin and if static fatigue will occur in the TEPA-cured
material after a sufficiently long loading period.
9.6. CONCLUSIONS
I t is clear that there has been a considerable advance in the understanding

of the factors controlling crack propagation in thermosetting resins. The
most important recent advance has been in developing a theory which
quantitatively explains stick/slip propagation in terms of a crack blunting
process. It also follows from this theory that for both stick/slip and
continuous propagation, the criterion for crack propagation is that a
critical stress (typically three or four times the Yield stress) must be
reached at a critical distancc ahead of the crack.
It has been shown that there IS no strong evidence for crazing in
282 ROBERT J. YOUNG
fully cured thermosetting resins and the crack propagation appears to

take place without the involvement of crazes. Attempts have been made
to relate the nodular microstructure of thermosetting polymers to their
mechanical properties. It has been shown that although both the micro-
structure and mechanical properties vary with resin formulation and cure
schedule there is no strong evidence for any direct correlation between
the two.
It is evident that more work must be done before there can be a
full understanding of crack propagation in thermosetting polymers. By
far the greatest proportion of the work discussed in this chapter has
been concerned with epoxy resins and it is important that similar investiga-
tions should be carried out on other thermosets to see how far the
conclusions that have been drawn concerning epoxy resins, can be extended
to other thermosetting polymers. Also, more work on epoxy resins is
required, particularly concerning the effect of microstructure upon
mechanical properties. The long-term aims of such investigations should
be to enable the properties of thermosetting polymers to be tailored
by controlling their chemical and physical microstructure.
REFERENCES
1. YOUNG, R. J. Chapter 6 in Developments in Polymer Fracture-I, ed. Andrews,

E. H., 1979, Applied Science Publishers Ltd, London.
2. YOUNG, R. J. and BEAUMONT, P. W. R. J. Mater. Sci., 1976, 11,766.
3. DIGGWA, A. D. S. Polymer, 1974, 15, 101.
4. MEEKS, A. C Polymer, 1974, 15,675.
5. SELBY, K. and MILLER, L. E. J. Mater. Sci., 1975, 10, 12.
6. GRIFFITHS, R. and HOLLOWAY, D. G. J. Mater. Sci., 1970,5,302.
7. EVANS, W. T. and BARR, B. I. G. J. Strain Anal., 1974,9, 166.
8. PHILLIPS, D. C and SCOTT, J. M. J. Mater. Sci., 1974, 9, 1202.
9. PHILLIPS, D. C, SCOTT, J. M. and JONES, M. J. Mater. Sci., 1978, 13,311.
10. YOUNG, R. J. and BEAUMONT, P. W. R. J. Mater. Sci., 1977, 12,684.
11. PRITCHARD, G. and RHOADES, G. V. Mater. Sci. Eng., 1976,26, I.
12. OWEN, M. J. and ROSE, R. G. J. Phys. D. Appl. Phys., 1973,6,42.
13. CHRISTIANSEN, A. and SHORTALL, J. B. J. Mater. Sci., 1976, 11, 1113.
14. PRITCHARD, G., ROSE, R. G. and TANEJA, N. J. Mater. Sci., 11, 1976, 718.
IS. YAMINI, S. and YOUNG, R. J. Polymer, 1977, 18, 1075.
16. GLEDHILL, R. A., KINLOCH, A. J., YAMINI, S. and YOUNG, R. J. Polymer, 1978,
19,574.
17. YAMINI, S. and YOUNG, R. J. J. Mater. SCI., 1979, 14, 1609.
18. WILLIAMS, J. G. Adv. Polym. SCI., 1978,27,69.
19. BEAUMONT, P. W. R. and YOUNG, R. J. J. Mater. SCI., 1975, 10, 1334.
20. HAKEEM, M. and PHILLIPS, M. G., Private communicatIOn.

21. MIJOVIC, J. and KOUTSKY, J. A. Polymer, 1979,20, 1095.
22. BROUTMAN, L. J. and MCGARRY, F. J. J. Appl. Polym. Sci., 1965,9,609.
23. BASCOM, W. D., COTTINGTON, R. L., JONES, R. L. and PEYSER, P. J. J. Appl.
Polym. Sci., 1975, 19, 2545.
24. MOSTOVOY, S. and RIPLING, E. J. J. Appl. Polym. Sci., 1966, 10, 1351.
25. MOSTOVOY, S. and RIPLING, E. J. J. Appl. Polym. SCI., 1971, 15,641.
26. YAMINI, S., Ph.D. thesis, Crack propagation in epoxy resins, 1979, University
of London.
27. SCOTT, J. M., WELLS, G. and PHILLIPS, D. C. J. Mater. SCI. (to be published).
28. IMAI, Y. and BROWN, N. J. Mater. SCI., 1976, 11,417.
29. HAKEEM, M. I. and PHILLIPS, M. G. J. Mater. SCI., 1978, 13,2284.
30. YAMINI, S. and YOUNG, R. J. J. Mater. SCI., 1978, 13,2287.
31. KAMBOUR, R. P. J. Polym. SCI., Macromol. Rev., 1973,7, I.
32. LILLEY, J. and HOLLOWAY, D. G. Phil. Mag., 1973,28,215.
33. MORGAN, R. J. and O'NEAL, J. E. J. Mater. SCI., 1977, 12, 1966.
34. RAClCH, J. L. and KOUTSKY, J. A. J. Appl. Polym. Sci., 1976,20,2111.
35. UHLMANN, D. R., Disc. Farad. Soc., 1979,68, (to be published).
36. YEH, G. S. Cm. Rev. Macromol. Chem., 1972, 1, 173.
37. KINLOCH, A. J. and WILLIAMS, J. G. J. Mater. SCI. (to be published).
38. YOUNG, R. 1. and BEAUMONT, P. W. R., Polymer, 1976, 17,717.
39. YOUNG, R. J. and YAMINI, A. Paper presented at IUPAC Symposium on
Macromolecules, Mainz, 1979.
40. WILLIAMS, J. G. Stress Analysis oj Polymers, 1973, Longmans, London.
41. WIEDERHORN, S. M. J. Amer. Ceram. Soc., 1967,50,407.
42. EVANS, A. G. J. Mater. SCI., 1972,7, 1137.
43. GLEDHILL, R. A. and KINLOCH, A. J. Polym. Eng. SCI., 1979, 19,82.
44. GLEDHILL, R. A., KINLOCH, A. J. and SHAW, S. J. J. Mater. SCI., 1979, 14, 1769.
45. YAMINI, S. and YOUNG, R. J. J. Mater. SCI. (to be published).
INDEX
Activation energy, 134, 156 Bismaleimide polyimides, 184

Alkylene oxide, 60-3 Bisphenol A, 14, 70, 118
Alumina trihydrate, 75-6 epoxy, 32
Aluminium trihydrate, 22 fumaric acid condensation
Amino resins, 11-13 polyester, 36
Ammonium Bisphenol-A-diglycidyl ether. See
dimolybdate, 76 DGEBA
persulphate, 80 Bromostyrene, 71
Antimony trioxide, 71 Bulk moulding compounds, 15, 127,
Arrhenius equation, 134, 147 137, 142
Asbestos-filled mouldings, 108 Butadiene-acrylonitrile rubbers, 21
Astrel-360, 153 Butadiene-S block copolymers, 250
Atlac-580, 36 t-Butyl perbenzoate, 39
Azo compounds t-Butyl peroxybenzoate, 136, 142
decomposition rate, 132-3 t-Butyl peroxyoctoate, 142
half-life, 132-3
13C NMR
Benzaphenone tetracarboxylic acid neutron scattering, 24
(BTDE),188 spectroscopy, 217
Benzoin ethers, 35 Carbon--carbon initiators, 142-3
Benzophenone, 35 Carbon fibre reinforced composites,
Benzoyl peroxide (BPO), 39, 132, 99, 118
136, 141 Carbonic acid ester anhydrides, 82
fJ-transition in crosslinked resins, Cerium compounds, 142
242-3 Chemical resistance of phenol-aralkyl
Bis(4-t-butyl cyclohexyl) resins, 112-14
peroxydicarbonate, 136 Chemiluminescence, 223
Bis-diene resins, 20 Chlorine content, 22
Bis(2-hydroxyethyl ether), 70 Cobalt naphthenate, 128
285
286 INDEX
Cold curing, 9 DGEBA~DMHDA system, 222

Condensation reactions, 13 Diacyl peroxides, 132
Contact moulding processes, 126 ct.rl-Dialkoxy-p-xylenes, 95
Corrosion resistance of vinyl ester Diaminodiphenylmethane, 177
resins, 32, 44, 48~51 4,4' -Diaminodiphenylmethane, 187,
Costs, 26 188
Crack 4,4' -Diaminodiphenylsulphone (DDS),
opening displacement, 278 187,212,215-17
propagation Dibromo neopentyl glycol (DBNPG),
continuous, 259, 274, 275 70,73
criteria controlling, 274-9 4,4' -Di(chloromethyl)diphenylether, 89
curing and cure schedule, effect Dichlorosilanes, 13
of,260 I ,2-Dichloro-1 ,2,2-tetraphenyl ethane
epoxy resins, in, 258 (DCTPE), 143
factors affecting, 258~66 ct.ct.' -Dichloro-p-xylene, 91
mechanisms of, 266-74 Dicyclopentadiene (DCPD), 64-7
stick/slip, 259, 265-6, 273, 275~9 acrylate, 44
testing variables, 263-6 Diels~Alder
thermosetting polymers, in, addition, 65
257~83 reaction, 186
unstable, 251 Diethylaniline (DEA), 128
tip', plastic deformation at, 271~4 Diethylene glycol, 67
Craze-crack growth process, 225 Diethylene triamine (DETA), 212, 268
Crazing in epoxy resins, 268 4,4' -Dihydroxy-2,2' -diphenyl propane,
Crosslink density, 214, 219~22, 15
237--40, 248 Dl Marzio equation, 239
Crosslinked networks, 214, 218 I ,2-Dimethoxy-I,1 ,2,2-tetraphenyl
Crosslinked polyester resins, 231~55 ethane (DMTPE), 143
Crosslinking, 9, II, 14, 18~20, 25, 27, ct.ct.' -Dimethoxy-p-xylene, 95~ I 0 I
89, 116, 149~50, 186 Dimethylaniline (DMA), 128, 132
Curing N,N-Dimethylaniline, 39
agents, 212, 260 2,5-Dimethyl-2,5-di(2-
vinyl ester resins, of, 39 ethylhexanoylperoxy)
Cyclisation reactions, 14 hexane, 136
1:4 Cyclohexane dimethanol 2,5-Dimethyl-2,5-hexane diamine
(CHDM),69 (DMHDA), 212
Cyclohexyl peroxides, 137 Diols,21
Di-2-phenoxyethyl peroxydicarbonate,
136
Diphenyl ether, 97
Decachlordiphenyl, 73 Disc cracks, 251~2
Decomposition temperature, 148 Doryl resins, 91
DEG/SA/FA resins, 239 Dough moulding compounds, 15,23,
Derakane-41I, 32 76-80, 84
Derakane-470, 42, 44 Dow XD-8084, 38
Derakane-510-A-40, 34 Dow XD-9002, 35
DGEBA, 212, 221, 260, 261 Dugdale plastic zone, 273
DGEBA~DETA system, 220, 222~3 Durestos boards, 91
INDEX 287
Electrical properties Fox-Loshaek expression, 239

phenol-aralkyl resins, 112-14 Fractography, 266
phenol-formaldehyde resin, 163 Fracture
Electron-transfer oxidation-reduction mechanics testing, 258
reactions, 127 surface(s)
Energy energy, 245
conservation, 26 epoxy resins, 266
dissipation, 234 polyester resins, 250-3
requirements, 26 toughness
Epichlorhydrin, 15 crosslinked polyester resins,
Epoxide-amine addition reaction, 213 245-8
Epoxide(s) elastomeric phases, with, 249-50
cure of phenol-aralkyl resins, 115 Free radicals, sources of, 142-4
homopolymerisation of, 214 Friedel-Crafts
Epoxy Novolac vinyl ester resins, 34 catalyst, 97, 100-1
Epoxy resins, 15-17 condensation, 87
amine-cured,211 reaction, 99, 103-5
chemical structure, 213-17 resins, 88, 90-9
composite, 99 Furan resins, 17-18
crack propagation, 258 Furane resins, 169-70
crazing, 268 Furfuryl alcohol, 17
deformation modes and failure,
222-3
durability, 223-6 y and y' transitions in crosslinked
fracture surfaces, 266 resins, 244-5
mechanical properties, 271 Gel permeation chromatography, 24
microvoid characteristics, 221 Glass transition temperature, 148,
network structure, 3, 217-18 235-7
physical structure, 217-22 Glycols, 67-70
stresses in, 25 Griffith equation, 248
structure-property relationships
and environmental
sensitivity, 211-29 Halogenated additives, 22, 74
ESCA 13C NMR, 24 Halogenated formulations, 71
Estercrete, 80-1 Halogen-containing derivatives, 21
Ethanolamine, 73 Hardeners, 7, 261
2-Ethoxyethanol, 101 Health and safety measures, 27
Ethylene glycol, 67 Heat distortion temperature of vinyl
ester resins, 44
HET acid, 73, 74
Filament wound pipe, 41 Hexamethylene tetramine, 7, 9, 100
Fire retardancy, 15, 22, 71-6, 170, Hexamine, 115
206 Hexel F178, 184, 185
Flame retardant resins, 34-5 High-temperature properties, 145-210
Flammability resistance of phenol-aralkyl resins, 110, 115
phenol-aralkyl resins, 114 thermoplastic resins, 150-60
Flexibilisers, 247 thermoset resins, 160
Foaming agent, 82 vinyl ester resins, 34
288 INDEX
HM-S carbon fibrejPMR-15 Maleic anhydrides, 63, 66

composites, 188 Matrix function, 2
Hydrogen peroxide, 130 Maturation agent, 77
Hydroperoxides, 128 Mechanical properties
Hydroquinone, 39 crosslinked polyester resins, 231-55
Hydroxyl content of thermosetting epoxy resins, 271
resins, 25 Melamine, 11-13
Melamine-formaldehyde resins, 4, 5,
163-7
Impact properties of vinyl ester resins, thermal stability, 166-7
41 Methyl ethyl ketone peroxide
Infra-red spectroscopy, 24 (MEKP), 39,128-31
Inhibitors, 39-40 I-Methyl imidazole, 143
Initiator(s) Methyl methacrylate, 70
blends, 139 Methylol derivatives, 165
efficiency, temperature effect, 136 Methylolmelamines, 165
half-life effect on cure N-Methylpyrrolidone, 185
characteristics, 135 Microcracking, 226
high temperature curing, for, 126, Moisture effects, 150
134-9 Moisture-sorption characteristics,
organic, 123-7 224-6
unsaturated polyester resins, for, Molecular weight distribution
121--44 investigations, 101
Injection moulding, 10, 23, 26 Molybdenum
Injection-compression, 11 compounds, 76
Isophthalic aCId, 63, 64 trioxide, 76
Monochlorosilanes, 13
Moulding
compounds, 76-80, 106, 108
leffamine T-403, 212
powders, 12
Kapton film, 173, 175

Kerimld-353, 178-83 Neopentyl glycol (NPG), 67-8
autoclave lamination, 183 Network structures, 3, 4
processing and properties, 183 Nexus polyamide fibres, 44
thermal degradation, 181-3 Nodular structure, 271
Kerimid-601 Norbornene systems, 186-90
processmg and propertIes, 178 5-Norbornene-2,3-dicarboxylic acid
thermal degradation, 178 (NE), 188
Kerimid-711, 184 Novolaks, 7,17,51,161
Ketone peroxides, 128, 129, 132, 141 NR-150 polyimides, 194-200
Nuclear magnetic resonance studies,
101
Laminated structures, 40-2
Laser-Raman spectroscopy, 24
Linear elastIC fracture mechamcs Organic initiators, 123-7
(LEFM), 256, 258 Organometallic resins, 94-5
INDEX 289
Peroxides Pipe, filament wound, 41

decomposition Plastic deformation at crack tip, 271--4
kinetics, 133 PMR-15, 188, 190
rate, 132-3 Pollution control, 27
half-life, 132-3 Polyalkyl acrylate polymers, 30
Peroxyketals, 137-9, 142 Polyamide-imides, 202-4
Phase lag angle, 234 Polyaminobismaleimides, 177-8
Phenol-aralkyl pre-polymer, 102 Poly(arylene ether sulphones), 151-3,
Phenol-aralkyl resins, 87-120 157
chemical reactivity, 105 Polybenzyl, oxidative degradation, 89
chemical resistance, 112-14 Polybenzyl polymers, 88
electrical properties, 112-14 Polydiphenyl ether resins, 89-91
epoxide cured, 115-18 Polyester
examples of, 103 foam, 81-2
flammability resistance, 114 resins, 21, 22, 24, 59-86
Friedel-Crafts reaction, 103-5 alkylene oxide route, 60--3
hexamine cured, II S 'built-in' halogen, 73
hIgh-temperature bulk handling, 23
properties, 115 crosslink density, 237-40
stability and strength, 110 crosslinked, mechanical
moulding compounds, 106 properties, 231-55
physical properties, 116-18 crosslinked network structures,
preparation and hexamine cure, 232-3
100--2 DCPD modified, 66-7
radiation resistance, 114 dough moulding compounds, 80,
range of, 103 84
structure and properties, 105-8 dynamic mechanical properties,
upgrading of phenolic novolacs, 233-5
108 flame retardant, 71-6
wetting charactenstics, 116 foams, 81-2
Phenol-formaldehyde fracture
composites, typical properties, 163 morphology, 250
condensates, oxidative degradation, surfaces, 250--3
106 toughness, 245-8
resins, 7-11, 24, 161-3 future trends, 84-5
electncal properties, 163 glass transition temperature, 235-7
thermal degradation, 162-3 glycols in, 67-70
Phenolic inhibitors, 39 halogenated additives, 74
Phenolic nuclei, substitution in, 102 high performance unsaturated, 64
Phenolic resins, 23, 105, 108-10 manufacturing processes, 60--3
injection moulding, 10 maturation rate, 76-8
Phenolic resol reSInS, 108-10 monomers, 70--1
Phenyl compounds, 21-2 moulding compounds, 76-80
Phosphorus based additives, 75 refinement of existing processes,
Phosphorus compounds, 72 63
Photodegradation, 22 room temperature cunng, 129
Photoinitiators, 35, 143 selecting initiators for curing at
Phthalic anhydrides, 63 elevated temperatures, 134-9
290 INDEX
Polyester-contd. Radiation
resins-contd. curing, 35, 53
sheet moulding compounds, 23, resistance of phenol-aralkyl resins,
80, 84 114
shrinkage, 76, 78-80 Recycling, 23
styrene emission, 83-4 Redox
toughened DMC and SMC, 80 initiators for ambient temperature
unsaturated, 14-15,23,59 cure systems, 125
initiator systems, 121--44 mechanisms, 127-32
Polyesterification process, 63 Relaxation(s)
Polyethersulphones, 21 constituent chain, in, 241-2
Polyimide-2080, 200-2 processes, 235
Polyimides, 18-20,24,171-204,206 Resole formation, 9
addition type, 177-8, 184
aluminium-filled, 183
condensation type, 171-3 Scrap additions, 23
NR-150, 194-200 Secondary relaxations, 240
preparation, 172-3 Secondary transitions, 240
processing and properties, 175-6 Self-toughening mechanism, 280
thermal degradation, 173-5 Sheet moulding compounds, 23, 33,
thermal stability, 153-6, 171 52, 55, 76-80, 84, 127, 137
thermoplastic, 194-204 Shrinkage and shrinkage stresses, 76,
Polymer cement, 80 78-80,226
Polymerisation in situ, 20 Silicone resins, 13-14, 167-9
Po1y(methyl methacrylate), 246, 259, thermal stability, 168-9
274, 275, 279 Siloxane/phosphate composites, 95
Poly(phenylene sulphide) resins, Skybond-700, 173, 175
158-60 Smoke
physical properties, 159 generation, 22, 170
Poly(phenylene sulphone), 151 suppressants, 76
Po1y(propylene maleate adipate), 247 Stannic chloride, 101
Premix compounds, 15 Stick/slip propagation, 259, 260, 265,
Prepreg processing, 23 266, 273, 275, 277-8
Price increases, 26 Storage modulus, 234
PRM-ll, 188 Strain energy release rate at failure,
Propylene glycol, 66 279-81
1,2-Propylene glycol, 67 Stress
Proton NMR, 24 concentration, 222
PSP resin, 204-5 epoxy resins, in, 25
press cure cycle for laminates, 205 intensity factor, 249, 258-61, 264,
processing and properties, 204-5 275,277-9
synthesis, 204 Styrene
Pyrolysis gas chromatography, 24 content of cured polyester resins,
139-42
emission in polyester resins, 83-4
Quality control monomer, 42-3
finished products, 25 Supramo1ecular structure, 2-3
thermosetting resins, 23-5 Swelling stresses, 226
INDEX 291
Tensile elongation of resin castings, Thin layer chromatography (TLC), 24

40,42 Threshold limit value (TLV), 83
Terephthalic acid, 63-4 Time-dependent failure, 279
Tetrabromo phthalic anhydride, 73,74 Toxicological aspects of vinyl ester
Tetrachlorophthalic anhydride, 73 resins, 55-6
Tetraethylenepentamine (TEPA), 279, Toxicological propertIes, 27
281 Triallylisocyanurate, 185
Tetraglycidyl 4,4' -diaminodiphenyl Trichlorosilanes, 13
methane, 212, 215 Triethylene tetramine (TETA), 260,
TGDDM-DDS systems, 212, 213, 271
215, 221-6 2,2,4-Trimethyl-l ,3-pentanediol
theoretical reaction mixtures, 216 (TMPD), 68-9
Thermal degradation
Kerimid-353,181-3
Kerimid-601, 178
Urea, 11-13
phenol-formaldehyde resins, 162-3
Urea-formaldehyde, 22, 24
polyimides, 173-5
Thermal spike exposure, 225, 226
Thermal stability, 146
melamine-formaldehyde resins, 166-7 Vinyl ester resins, 17, 29-58
polyimides, 171 aIrcraft industry applications, 53
polysulphones, 153-6 applications, 47-54
silicone resins, 168-9 ballistic armour, 54
Thermally stable resins, 145-210 basic structure, 32-3
Thermid-600, 190--3 brominated, 34
Thermogravimetric analysis (TGA), corrosion resistance, 32, 44
146,148,153,163,182, 192 curing, 39
Thermoplastic resins dental filling material, 53
fabrication, 149 dicyclopentadiene acrylate-diluted,
high-temperature properties, 150--60 44
Thermosetting resins, 1-28 early history, 30--1
crack propagation, 257-83 effects of cast resin high tensile
fabrication, 149 elongatIon on performance
filled, 5 of laminated structures, 40-2
future prospects, 25- 7 electrical lllsulation applications,
handling and processing, 22-3 52-3
hardening reaction, 22 filament wound laminate, 47
high-temperature properties, 160 filament wound pipe, 41
hydroxyl content, 25 flame retardant, 34-5
limitations, 7 generalised structure, 31
matrix function, 2 glass fibre laminates, 44
properties of, 3-7 hand lay-up laminate, 47
property improvements, 21-2 heat distortion temperature, 44
quality control, 23-5 high-temperature applications, 34
reground, 23 HSMC,52
reinforced, 26 impact properties, 41
supramolecular structure, 2-3 inhibitors, 39-40
unreinforced, 5 laminate properties, 44-7
292 INDEX
Vinyl ester resins-contd. Vinyl toluene, 44

land transportation applications, Viscosity, 4
51-2 ViSCOSity-pressure characteristics, 21
marine applicatIOns, 53 Viscosity-temperature relationships, 4
matched die laminates, 47
mechanical properties, 44
monomers, 42-4
new developments, 54-5 X-ray
physical properties, 37 diffraction, 25
radiation scattering, 271
curable, 35 Xylok-21O, 110
curing applications, 53 Xylok-225, 108, 109
rubber-modified, 38 Xylok-237, 116
SMC resins, 33, 55
structure-imparted characteristics,
31-2
structures, 32-8 Yield stress, 278, 279
synthesis, 31 Young's modulus, 2
tensile elongation, 40, 42
toughness, 32
toxicological aspects, 55-6
urethane-based, 36 Zinc borate, 71

G. Pritchard - Developments in Reinforced Plastics - Resin Matrix Aspects-Springer Netherlands (1980)

Uploaded by

Copyright:

Available Formats

G. Pritchard - Developments in Reinforced Plastics - Resin Matrix Aspects-Springer Netherlands (1980)

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

G. Pritchard - Developments in Reinforced Plastics - Resin Matrix Aspects-Springer Netherlands (1980)

Uploaded by

Copyright:

Available Formats

DEVELOPMENTS IN

Developments in many fields of science and technology occur at such a pace

Resin Matrix Aspects

APPLIED SCIENCE PUBLISHERS LTD

British Library Cataloguing in Publication Data

WITH 69 TABLES AND 87 ILLUSTRATIONS

© APPLIED SCIENCE PUBLISHERS LTD 1980

All rights reserved. No part of this pubhcatlOn may be reproduced, stored In

Plastics are complex materials. Attempts to use them without employing

The subject of reinforced plastics is a fast-developing one. It is hoped

1. Thermosetting Resins for Reinforced Plastics

2. Vinyl Ester Resins 29

3. Polyester Resin Chemistry 59

4. Phenol-Aralkyl and Related Polymers 87

5. Initiator Systems for Unsaturated Polyester Resins 121

6. High-Temperature Properties of Thermally Stable Resins 145

7. Structure-Property Relationships and the Environmental

8. Some Mechanical Properties of Crosslinked Polyester Resins 231

9. Crack Propagation in Thermosetting Polymers 257

School of Chemical and Physical Sciences, Kingston Polytechnic,

BP Chemicals Limited, Research and Development Department,

Dow Chemical USA, Texas Division, Freeport, Texas 77541,

THERMOSETTING RESINS FOR REINFORCED

This book is mainly about resins-their synthesis, structure and properties.

in later volumes, so the emphasis here is on chemical and physical or

1.2. THE FUNCTION OF THE MATRIX

Every component of a composite material has its function. The reinforce-

1.3. SUPRAMOLECULAR STRUCTURE

Thermosetting resins are chemically reactive substances which undergo

crosslinked regions dispersed in non-bound matter. Figure l.l(c) shows

1.4. PROPERTIES OF THERMOSETTING RESINS

Prior to crosslin king, thermosetting resins may be viscous liquids (in

FIG. 1.2. Scanning electron micrograph of the surface of a pigmented, decorative

Property Units Unsaturated Glycidyl ether Phenol

Tensile strength MNm- 2 25-80 30-100 20-65

for a given resin, because of differences in precise formulation and

Property Units Material

Compressive strength MNm- 2 65-90 100-130 55-70 130-180 140-300 =l

are ostensibly similar. However, certain limitations are apparent in all

1.5. PHENOL-FORMALDEHYDE RESINS

Phenol condenses with aldehydes to give products which can be either

This compound decomposes on heating to generate ammonia and

These intermediates undergo further reaction with formaldehyde and

Crosslinking with hexamethylene tetra mine brings about the formation

These molecules undergo self-condensation to form insoluble, infusible

Resin 35 to 45 parts by weight

Alternatively resins can be dissolved in solvent so as to impregnate

60 %of phenolic moulding powder consumed in some European countries

1.6. AMINO RESINS

The reaction of aldehydes (again, usually formaldehyde) with amines

Amino resin synthesis is similar to that for phenolic resins. Excess

Urea resins are widely used for adhesives, in chipboard manufacture,

1.7. SILICONE RESINS

(3) Trichlorosilanes give network polymers:

1.8. UNSATURATED POLYESTER RESINS

Unsaturated polyester (U P) resins constitute an important and growing

and propane-l,2-diol, although 'bisphenol A' and its hydrogenated,

Premix compounds have been developed for the compression, transfer

1.9. EPOXY RESINS