Polymer International 41 (1996) 193-207

Adhesives in Demanding Applications

S. J. Shaw
Structural Materials Centre, Defence Research Agency, Farnborough, Hampshire, UK
(Received 7 May 1996; accepted 14 June 1996)

Abstract: Structural adhesives are nowadays employed extensively in structural

engineering applications where advantage is made of many outstanding attrib-
utes, most notably their ability to confer reduced weight, improved fatigue per-
formance together with reduced production and life cycle costs. Increasingly in
recent years, attempts have been made to employ bonded structures in harsh
environments. In this article an attempt is made to consider the demanding
environments which bonded joints are frequently required to resist, focusing
principally on hot/wet and high temperature conditions. With the former, the
ways in which water can adversely affect joint behaviour are highlighted and the
factors, such as nature of adhesive and surface treatment, crucial to optimum
joint performance are discussed. Adhesive joints in high temperature environ-
ments are discussed in terms of early high temperature adhesives, designed with
little thought to processability, together with more recent developments in high
temperature adhesives where processability has been 'engineered' into the
polymer structure.
K e y words: adhesives, adhesive joints, environmental affects, surface treatment,
organosilanes, interfacial stability, moisture stability, hydrophobicity, high tem-
perature adhesives.

INTRODUCTlON Increasingly in recent years, attempts have been made

to incorporate adhesively bonded joints into engineer-
ing structures in which the bonds are subjected to
The use of structural adhesives in engineering applica-
harsh, highly demanding environments. Most notably
tions has grown extensively in recent years. This can be
these include warm/moist atmospheres and temperature
attributed to the substantial benefits which bonding can
extremes (particularly elevated temperature), both being
provide in comparison to more traditional joining tech-
particularly severe when superimposed on cyclic fatigue
niques such as mechanical fastening and welding. These
loads. In many cases this has been achieved with a great
benefits include :
deal of success due primarily to the adhesive user being
1. Allowance of a more uniform stress distribution aware of the numerous factors vital to successful
resulting in potential for improved fatigue per- bonding. However, notable problems have occurred
formance ; with various simplifying assumptions on the part of the
2. the ability to join dissimilar materials with, in user, resulting in less than optimum performance.
particular, the dielectric nature of the adhesive In this paper an attempt will be made to consider the
minimising electrolytic corrosion between dis- most demanding environments that bonded joints can
similar materials; encounter, focusing primarily on warmbumid and high
3. the ability to join thin gauge metals to each other temperature conditions, and to outline the various
and in particular to honeycomb structures factors of crucial importance for successful exploitation
resulting in the availability of lightweight struc-
tures capable of exhibiting high strength to DEMANDING ENVIRONMENTS
weight ratios;
4. simpler and cheaper component construction; One of the most important requirements of an adhe-
5. avoidance of heat-affected zones which can arise sively bonded joint is that the structure in which the
from welding operations. joint is present should retain a significant proportion of
193
Polymer International 0959-8103/96/$09.00 0British Crown Copyright DERA/1996. Printed in Great Britain
194 S . J . Shaw

its load-bearing capability throughout the intended the environment indicated has a substantial effect on
service-life of the structure, under a variety of environ- the load-bearing capacity of the joint, with 1500h of
mental conditions. Several external environmental con- water immersion at 60°C resulting in just 15% retention
ditions have been known to cause severe bond of strength. In addition, with the adhesive formulation
weakening so as to pose a threat to the integrity of the (a diglycidyl ether of bisphenol A cured with a tertiary
bonded structure. As mentioned previously, possibly the amine curing agent) and substrate type employed,
best example include warm/moist environments, where immersion under these conditions was found to influ-
some catastrophic effects have been observed, and tem- ence the locus of failure with a pronounced change from
perature extremes, and much of this paper will focus on cohesive within adhesive to apparent interfacial (i.e.
these two environmental variables. However, depending between substrate and adhesive) with increased immer-
on the precise application, various other degradatory sion time. Surface-specific analytical techniques have in
effects have been observed relating to, for example, radi- fact confirmed interfacial failure in joints of this type,
ation, liquid contaminants in addition to water, e.g. thus agreeing with thermodynamic predictions based on
fuels, de-icing fluids and hydraulic oils in the case of interface stability considerations.
aircraft and, of course, stress. Although the accelerated test conditions employed to
produce the data in Fig. 1 can be regarded as extremely
and possibly unrealistically harsh, it is important to re-
M0ISTURE- RELATED EFFECTS cognise that substantial strength losses have been
observed with bonded joints subjected to long-term
General observations environment ageing in ‘real’ environments.’ The results
shown in Fig. 2 provide such an example. In this case,
Unfortunately, experience has frequently shown that the influence of long-term exposure on the strength of
one of the most hostile environments that a bonded aluminum alloy double overlap lap-joints is demon-
joint can encounter is one containing atmospheric mois- strated, with the adhesive employed being a polyamide-
ture.1*2To demonstrate the potential magnitude of this modified epoxy. Although initial joint strength can be
effect, Fig. 1 indicates the severe effect that a water- regarded as impressive, the long-term trend shows a
laden environment can exert on a bonded joint. Clearly, highly undesirable characteristic with, for unstressed
joints, just a two year exposure resulting in less than
10% retention of strength. In most applications, bonded
joints would, of course, be subjected to externally
applied stresses during service and the effects this can
have, combined with a hot/wet tropical environment, is
clearly apparent in Fig. 2. As indicated, stress superim-
posed on the tropical environment clearly provides a
further source of damage for this particular adhesive/
substrate system which, in the case of joints exposed to
20% stress, results in complete specimen failure within
two years of exposure. In fact stress, particularly when

Ao r

Stress. X
0 0
5 0
10 A
20 0

0 0 1 2 3 4 5 6 7
0 500 1.Ooo 1,500
Exoosure time, years
Immersion time, H20, 60°C (h)
Fig. 2. Strength of overlap joints of chromic acid-etched alu-
Fig. 1. Influence of water immersion on the strength of epoxy- minium alloy bonded with a polyamide-modified adhesive
bonded mild steel butt-joint. exposed to a tropical hot/wet climate.

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 195

applied under cyclic conditions (i.e. fatigue) superim- comparison with epoxy-based adhesives, where inter-
posed on a hot/wet environment, is generally regarded facial type behaviour is a frequent occurrence, are
by many as the most severe environment a bonded joint unclear.
can experience in practice. Unfortunately, for a wide Detailed studies conducted within the author's own
range of applications just such a combination is fre- laboratory has shown that a wide range of proprietary
quently encountered, which therefore requires consider- adhesives, mostly epoxy based, exhibit characteristics
ation during the joint design and adhesive selection intermediate between the two extremes indicated in Figs
processes. 2 and 3.2 Within this study, exposure trials with joints
It is important to note that the adhesive system dis- subjected to both hot/dry (desert) and temperate (UK)
cussed above, a polyamide-modified epoxy, is not gener- climate were also conducted with, as would be expected,
ally recognised as being moisture resistant, having been hot/wet (tropical) climatic conditions resulting in the
developed initially to exhibit high levels of bulk tough- most severe degradation.
ness which would be translated into peel-resistant adhe- Numerous studies such as this have demonstrated the
sive formulations. Unfortunately, this approach to various factors which can strongly influence joint dur-
toughness enhancement also results in the formulated ability, and it is of interest to briefly consider what are
adhesive exhibiting a high affinity for water due to the now regarded as the most important, and to follow this
hydrophilic nature of the polyamide modifier. Thus the with a brief review of the various approaches which can
trends indicated in Fig. 2 cannot be regarded as entirely be employed to maximise joint durability.
unexpected. The following are now regarded as crucial in control-
Fortunately, other adhesive types have been shown to ling the long-term viability of a bonded joint.
exhibit superior performance, a notable example being
Environment. Although, from what has been
the vinyl-phenolic based adhesives, data for which are
stated so far, this should be entirely obvious, it is
shown in Fig. 3.2 With this system, exposure of vinyl-
worthwhile to emphasise that water will, in all
phenolic-bonded aluminium alloy lap-joints under
probability, be by far the most hostile environ-
hot/wet tropical conditions produced no significant
ment that an adhesive joint will encounter in
effect on joint strength over a six year period, the only
service.
exception to this being for joints subjected to 20% stress
Influence of substrate type. Bonded joints pre-
application, where a significant strength decline beyond
pared from most substrate types can exhibit sus-
two years' exposure was observed. Particularly note-
ceptibility to moisture-related degradation, with
worthy in this case was the locus of failure, with failure
the precise mechanism of attack frequently being
within the adhesive layer being the prime failure mode.
dependent upon the nature of the substrate.
Although, as will be discussed later, water usually
Major differences between metallic alloys and
exhibits a dominant effect within the interfacial regions
polymer composite substrates have been
of a bonded joint, the locus of failure noted for the
observed in this respect.
vinyl-phenolic system has been observed by other
Surface treatment. As will become apparent later
worker^.^,^ Factors responsible for this behaviour, in
in the discussion, the surface treatment
employed prior to bonding can dictate success
or failure in a warm/humid environment. In
addition, the use of primers has, with many
alloys, been found conducive to moisture resist-
I ance to the extent that primer addition to alu-
minium alloys is frequently employed prior to
bonding, particularly in various aspects of air-
craft construction.
Influence of adhesive type. As clearly demon-
strated from the above discussion, the chemical
nature of a structural adhesive can influence
quite substantially both the extent and rate of
v environmental damage.

0 1 2 3 4 5 6 Mechanisms of moisture attack

Exposure time. years
Before considering the approaches which can be
Fig. 3. Strength of overlap joints of chromic acid-etched alu- employed to maximise moisture resistance, it is worth-
minium alloy bonded witha nitrile-phenolic adhesive exposed while to consider the various mechanisms of attack
to a tropical hot/wet climate. which have been postulated to account for moisture

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

196 S. J . Shaw

sensitivity in bonded joints. Three main mechanisms with metal surface oxide stability in the presence of
have been proposed. moisture, has been shown to be relevant to bonded alu-
minium alloy structures. It is of interest to consider
Water can be absorbed into the adhesive layer
these areas in greater detail.
resulting in either chemical modification (e.g.
hydrolysis) or physical damage (e.g.
Interface stability. In a dry environment the stability of
microcracking). In addition, moisture uptake can
an interface between an organic adhesive and a sub-
result in a reversible plasticisation which can
strate material can be predicted thermodynamically by
cause a substantial reduction in glass transition
use of the Dupre equation:
temperature, T g . An approximate rule of thumb
frequently used to assess the effect of water on T, WA= ~s + Ya - Ysa (1)
is the so-called 20°C rule, i.e. for every 1% mois-
where W is the thermodynamic work of adhesion, ys
ture uptake Tg is suppressed by 20°C5 Although,
and ya are the surface free energies of the substrate and
in the author’s experience, there are numerous
adhesive phases, respectively, and ysa is the interfacial
exceptions to this guideline, it does indicate the
free energy. In the presence of a third encroaching
potential effect moisture could have on bond
liquid phase, for example water, this relationship can be
performance where, in particular, a substantial
modified to take into account the presence of the liquid
reduction in load-bearing capacity could be
thus :
experienced at elevated temperature.
Water can affect the substrate either chemically
or by physical modification. The former is often
where the suffix L denotes the liquid phase.
associated with certain metallic alloys such as
Calculations using appropriate surface and interfacial
aluminium and titanium where the metal oxide
free energies usually result in positive work of adhesion
layer is particularly prone to moisture attack,
values for most interface types, thus demonstrating
whereas the latter has been shown to be applic-
thermodynamic stability. In the presence of water,
able to polymer composite substrates such as
however, work of adhesion values are usually negative
carbon and glass-fibre-reinforced polymers.6
for metal oxide-adhesive interfaces, thus indicating
Water can influence the integrity of the interface
instability of such interfaces in the presence of moisture.
between adhesive and substrate causing an adhe-
As discussed previously for the data shown in Fig. 1,
sive de-bond which can in turn result in a vir-
practical observation has confirmed this prediction.
tually complete loss of load-bearing ~apacity.~,’
Theoretical predictions for polymer composite-adhesive
Although bulk modification and plasticisation of an interfaces, however, generally indicate thermodynamic
adhesive layer has been observed in practice, this mode stability in both dry and wet environments, predictions
of attack is not usually responsible for incidents of which have been confirmed in practice. Indeed, in terms
environmentally induced failure, where an adhesive of enhancing moisture resistance and extending the
having particularly enhanced hydrophilic characteristics service life of bonded joints, this remains the greatest
is usually a prerequisite for this mode of attack. The difference between these two types of construction
discussion and data relating to Fig. 2 are a typical material.
example of where this does occur in practice, the hydro-
philic nature of the adhesive resulting in a substantial Oxide layer stability. Although the thermodynamic
reduction in strength over a relatively short timescale, arguments outlined above should apply to aluminium
with failure of the joint being maintained within the alloy interfaces in addition to steel, experimental evi-
adhesive layer throughout.’ dence has suggested that the aluminium oxide layer
The interfacial regions of a bonded joint are usually within the alloy-adhesive interfacial zone is the most
recognised as the ‘weak link’ most vulnerable to the vulnerable to moisture attack. This has been shown to
effect of moisture. Thus although choice of adhesive be the case with alloy joints previously subjected to a
must always be considered carefully in relation to the chrome-sulphuric acid etch surface treatment.’ Joints
intended application, the nature of the adhesive- comprising aluminium alloy previously subjected to this
substrate interface and the various factors likely to surface treatment have been shown to be highly vulner-
impact upon interface integrity and stability are of fun- able to moisture attack, whereas treatments utilising
damental importance. anodising techniques have exhibited improved hot/wet
Within this context two primary mechanisms have performance, as indicated in Fig. 4.6 Improved oxide
been proposed to account for observations of environ- hydrolytic stability has been one of the factors proposed
mentally induced interfacial failure. The first, associated to account for this behaviour, whereas the fact that true
with thermodynamic considerations of adhesive- interfacial failure is not observed has been attributed to
substrate stability, has been shown to be applicable to oxide layers having the capacity to promote mechanical
joints with mild steel substrates. The second, associated interlocking with the adhesive.

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 197

50 degradation associated with plasticisation and Tg

reduction will be reduced. Secondly, water absorption
into the bulk of an adhesive will clearly play a promi-
nent role in interfacial failure processes since it will be
by a diffusion mechanism that water reaches the critical
40 interfacial region. It is therefore intuitively reasonable
Chromic acid anodired to assume that hydrophobic adhesives, i.e. those
exhibiting relatively low values of both equilibrium
moisture uptake and diffusion coefficient, would impart
I
better moisture-resistance characteristics to a bonded
10
joint than their relatively hydrophilic counterparts. Evi-
-
$30
dence in support of this view is discussed later.
s
W
It would now be of interest to briefly discuss the
E
c approaches which have been considered to improve the
c.
.-E0 moisture sensitivity of both adhesives and adhesive-
.-
g 20 substrate interfaces.
-1

Interface stability enhancement. In order to enhance the

durability of bonded mild steel alloys, where failure at
the adhesive-substrate interface occurs on atmospheric
10 exposure, work has generally focused on the potential of
coupling agents to overcome interfacial weakness.
Although several types of coupling agent have been
considered for this purpose, most research has been
devoted to organosilanes. Exploited initially as a means
1 I I
0 of overcoming the pronounced moisture susceptibility
500 1 .Ooo 1,500
Time in water at 50°C (h). of glass-reinforced polymers, their ability to substan-
tially enhance the stability of the glass-polyester inter-
Fig. 4. Effect of surface pretreatment on the durability of alu-
face indicated their potential for use in adhesively
minium alloy/epoxy joints subjected to accelerated ageing in
water at 50°C. bonded structures. Following this recognition, a
number of investigations have confirmed this poten-
tial.’-’
Optimisation of moisture stability Most commercially available organosilanes are based
upon the following generalised structure:
Bearing in mind the nature of the failure processes R-Si-(X),
which have been observed in bonded joints subjected to
moisture-laden atmospheres, two main approaches can where X is a hydrolysable group designed for inter-
be considered as a means of diminishing the impact of action with a hydrolysed surface layer such as on glass
this type of environment. The first, based upon the pre- or a metal oxide, and R is an organofunctional group
viously discussed observation of interfacial type failure, capable of interaction with a polymer. For metal alloy
is, quite logically, concerned with enhancing interfacial adhesion, most of the studies in this area have been
stability. The manner in which this is conducted devoted to y-glycidoxypropyltrimethoxy silane, which
depends entirely on the precise location of the ‘weak has the following structure:
link’. For example, with the previously discussed mild
steel bonded joint (Fig. l), where moisture-induced /O\
C H~-CH-CH~-O-(CHJ~-S~-(OCH~)~
failure is located precisely at the adhesive-oxide inter-
face, remedial action must center upon countering the Organosilanes of this type have been developed with
thermodynamic tendency for adhesive displacement. one specific mechanism of action in mind, namely the
With bonded aluminium joints, where the oxide layer chemical bonding theory. Although several different
frequently provides the failure zone, enhancing the theories have been proposed to account for the actions
moisture resistance characteristics of this oxide provides of organosilanes, it is worthwhile considering this
the best option. theory in more detail since this, and a variant of it, are
The second approach concerns the development and now considered the most relevant to practical observ-
use of adhesive formulations which exhibit hydrophobic ations.’ In simplistic terms, the alkoxy groups are, in
characteristics. Two potential benefits can arise from many practical applications, hydrolysed by addition of
systems of this kind. Since the potential for water water which thus allows, according to the theory, inter-
uptake is greatly diminished in such systems, property action and reaction of the resulting silanols with

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

 198 S . J . Shaw

60 Clearly, for an equilibrium condition such as this to be

Primer GPMS: ?-Glycidoxypropyltrimethoxysilane maintained, the adhesive contributing to the interfacial
Primer APES: 7-Aminopropyltrimethoxy sibne zone must have the required level of rigidity to restrict
Primer SAHS: 1styrene functional amine hydrochloride silane
movement of silanol groups away from the interface.
Equilibrium conditions would be lost if silanol groups
50
resulting from hydrolysis were to be physically removed
from the interface. Indeed, examples exist in the liter-
ature where silanes have been shown not to exhibit
-2 40
beneficial effects where one component of the interface
is elast~meric.~
E
b-
Primer GPMS Metal oxide stability. As previously mentioned, with
ui
In some alloy types, e.g. aluminium, the nature of the
F!
c
ul surface oxide layer and in particular its tendency to
30 hydrate in the presence of moisture has been seen by
a
c
0 many as the primary cause of environmentally induced
F
r
c
failure in bonded joints.8 Joints prepared from alu-
.- minium substrates originally pretreated using a
.-0 chrome-sulphuric acid etch treatment and bonded with
5 20
rn an epoxy adhesive have in fact been shown to be partic-
ularly vulnerable to moisture, where oxide hydration
has been suggested as the primary cause. Interestingly,
cleaned) this surface treatment has been shown to promote per-
10 fectly acceptable moisture-resistant bonds with
phenolic-based adhesives, and although this is a well
recognised phenomenon, the factors responsible are still
unclear.6
To promote improved moisture sensitivity, two anod-
0 ising surface treatments are now generally employed for
0 500 1,000 1,500
Time in 6O0C/H2Oenvironment (h) pretreating aluminium alloys.’ Both have been shown
to provide superior hot/wet performance to the etching
Fig. 5. Influence of water immersion on the strength of
procedure, as shown previously in Fig. 4, where both
epoxy-mild steel joints employing various silane-based
primers. improved oxide hydrolytic stability and topographical
features conducive to mechanical interlocking have
been proposed to account for this observation. More
hydroxyl groups present on the substrate surface. Like-
recently, however, hydrolytic stability has lost favour,
wise, by judicious choice of the organofunctional group,
with the morphology of the oxide layer allowing a
this can react with the polymeric component of the
‘required’ degree of penetration of adhesive into the
adhesive formulation thus resulting in the formation of
oxide topographical features now appearing to play a
a ‘covalent bridge’ linking together the organic and
leading role.I3
inorganic components of the joint. Although the forma-
tion of such an interface would be expected to provide
The influence of adhesive hydrophobicity. Although, as
far greater moisture resistance than secondary force
indicated above, several routes to strength deterioration
interactions alone, several anomalies have been noted
exist for joints subjected to humid environments, water
with this approach, relating most notably to the observ-
absorption into the bulk of the adhesive will play a
ation that such a highly polar M-0-Si oxane bond
prominent role in the degradation process, since at the
should be vulnerable to hydrolysis. To counter this
very least it can be viewed as the primary mechanism by
problem, a modified version of the chemical bond
which water reaches the critical interfacial region.
theory has been proposed by Plueddemann.’ Known as
However, as mentioned previously, absorbed water can
the reversible hydrolysable bond theory, chemical bond
cause a significant reduction in bulk adhesive proper-
formation is once again the main feature. However, the
ties. The substantial strength reduction previously dis-
inherent vulnerability of the oxane bond to moisture is
cussed and demonstrated in Fig. 2 for a
accounted for by the apparent ability of the hydrolysis
polyamide-modified epoxy adhesive subjected to a
reaction products to reform and reproduce oxane
tropical hot/wet environment can be partly attributed
bonds, as indicated in the following equilibrium reac-
to bulk adhesive effects.
tion :
However, whatever the eventual mechanism of
M-0-Si + H , O e M - O H + HO-Si- failure, it does appear intuitively reasonable to assume

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 199

that a hydrophobic adhesive would, everything else I

being equal, contribute to improved moisture resistance
and thus enhanced joint durability. A number of invest-
igations which have attempted to study the relationship
between joint strength and water content within the
adhesive layer have in fact confirmed that a link * *
between hydrophobicity and joint durability does
appear to This is highlighted by the data
demonstrated in Figs 6 and 7 from work by Parker and Temoerafe a
Shaw.' The latter shows the relationship between the 20°C/65XRH 0

strength of aluminium alloy lap-shear joints in which t

the alloy was previously subjected to a chromic acid I"

etch treatment, as a function of adhesive calculated 0 0.5 1.0 1 .s 2.0

moisture content, the adhesive in this case being a Calculated joint moisture content, %
polyamide-modified epoxy. The strength values indi- Fig. 7. Relationship between strength of overlap joints of
cated relate to joints which had been subjected to chromic acid-etched clad aluminium alloy bonded with an
ageing in a tropical environment. Adhesive layer water epoxy-nitrile film adhesive and estimated adhesive moisture
contents were estimated from knowledge both of the content.
water diffusion characteristics of the bulk adhesive
under various temperature/humidity environments and
meteorological data from the tropical test site. As decline, the amount of water absorbed and hence the
clearly shown in Fig. 6, although the degree of scatter is influence on joint strength is much reduced in compari-
high, data for a range of climatic and accelerated ageing son with the results in Fig. 6. This is primarily due to
conditions fit reasonably onto a single curve, with the the enhanced hydrophobic characteristics of the nitrile-
influence of water content on joint behaviour being phenolic adhesive (Fig. 9) in comparison with the pre-
clearly apparent. The substantial decline in strength viously discussed polyamide-modified epoxy.
with water content is what would be expected for a On a more fundamental level, one approach to
polyamide-modified epoxy adhesive which has been hydrophobic enhancement adopted by the current
shown to absorb considerable quantities of water, as author and others has been to consider the concept of
indicated in Fig. 8. Clearly, at relative humidities molecular ha10genation.l~Although a number of such
approaching saturation, water contents are exceedingly attempts have been made in recent years, as far as the
high, which obviously contributes greatly to the very author is aware, these have not yet resulted in com-
poor results demonstrated in Figs 2 and 6. mercially exploited materials. Halogenated epoxies
Figure 7 shows results obtained from the nitrile- were developed by Johncock et al. in the early to
phenolic-based adhesive system previously discussed in mid-1980~.'~*'~ Initial studies were concerned with
Fig. 3. Although an increase in calculated adhesive layer halogen containing diglycidylamines. Various degrees of
moisture content results in a moderate joint strength halogenation (as low as 0.8%) were found to result in
significant increases in hydrophobic character, with
both bromine and chlorine modification having greater
effects than fluorination.
Outoaw climate
uoi/ws~ a Johncock and co-workers have also considered the
uot/CYy a halogenation of tetraglycidyl 4,4-diaminodiphenyl-
TemDerate a
methane (TGDDM) resins cured with 4,4-
diaminodiphenylsulphone (DDS), the currently pre-
ferred resin for many composite aerospace applications.
Using such an approach, moderate improvements in
hydrophobic character were obtained with, however,
significant reduction in glass transition temperature
with some of the modifications.
Goobich and Marom have investigated the effect of
introducing bromine into epoxy resin formulations by
0 employing a brominated co-reactant with TGDDM/
0 1 2 3 4 DDS systems." This type of modification has been
Estimated water content, % shown to significantly reduce water uptake, whilst
Fig. 6. Relationship between strength of overlap joints having only a limited effect on the high temperature
bonded with a polyamide-modified adhesive and estimated properties and glass transition temperature of dry
adhesive moisture content. resins.

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

200 S. J. Shaw
-
w r

1
c 0

0
0 o'6

1.6 l.? 1.8 1.9 2.0

2
o -0.4
l -
1.6 1.7 1.8 1.9 2.0
log,o (Relative humidity. %)
logfo (Relative humidity, %)
Fig. 8. Saturation moisture contents at 50°C for a polyamide-
modified adhesive. Fig. 9. Saturation moisture contents at 50°C for a epoxy-
nitrile film adhesive.
Probably the greatest success at hydrophobic
enhancement has been achieved by Grifith and co- HIGH TEMPERATURE ENVIRONMENTS
w o r k e r ~ . ' ~ -In~ ~an extensive programme, fluorinated
epoxies together with compatible curing agents were Most structural adhesives are formulated to yield high
developed in the 1970s. The synthetic process developed temperature limits only marginally greater than 100°C.
was capable of producing resins having the following A typical example of this is demonstrated in Fig. 10
structure : which shows the retention of room temperature
strength as a function of test temperature for a proprie-
tary rubber-modifier epoxy adhesive system. As indi-
cated, this adhesives load-bearing capacity is greatly
diminished with increasing temperature, with less than
10% room temperature strength being retained at
150°C. This behaviour can usually be related to T g ,
with the CnFznflgroup ranging from zero (in effect, a which, to a first approximation, is often considered the
hydrogen atom on the 5 position of the benzene ring) to high temperature limit for any structural adhesive. For
numbers in excess of 10. The hexafluoroisopropylidene long-term applications in a high temperature environ-
groups were common to all the resins developed. ment, the thermal stability characteristics of the poly-
One major disadvantage which was realised early in meric components on which the adhesive is based also
the development of the above fluorinated epoxies con- become important, to the extent that T, can be a mis-
cerned compatibility with common curing agents. The leading pointer to high temperature capability.
incorporation of large quantities of fluorine, although The historical development of high temperature poly-
likely to enhance hydrophobic behaviour, was sufficient mers has essentially occurred in two relatively distinct
to render the resins incompatible with most curing phases. Beginning shortly after World War Two, the
agent types. Two main types of curing agent were there- synthesis of new polymers having high temperature
fore developed by Griffith in an attempt to alleviate this capabilities reached a peak of activity in the late 1950s
problem, these being silicone-amines and fluoro- and early 1960s. During this time, numerous polymer
anhydrides. Work conducted by Shaw et al. on various types and variations within a particular polymer theme
formulations derived from the above resin resulted in were developed, with the prime goal being thermal sta-
equilibrium water concentrations as low as O-2%." bility maximisation. Although this approach resulted in
Although a thorough mechanical property study has some exceptionally thermally stable polymers, these
been conducted to characterise bulk resin behaviour, no were almost invariably of little use in the real world.
attempt has been made, to the author's knowledge, to This is because the type of organic chemical structures
determine whether the hydrophobic characteristics that can convey good thermal and thermo-oxidative
exhibited by the bulk polymers are translated into stability to polymer molecules are also those which lead
highly moisture-resistant adhesive joints. This is also to infusible, insoluble materials. The seemingly competi-
true of the other hydrophobic resin systems outlined tive drive to supreme thermal stability therefore resulted
above, and it would be of major interest to determine in a wealth of polymer types which were impossible to
the existence, or otherwise, of such a correlation. Such a fabricate.
study is about to begin in the author's laboratory. In recognition of this problem, since the mid- to late

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 201

1960s, the development of heat-resistant polymers and at the Narmco Materials Division of the Whittaker
adhesives has been based on a compromise between Corporation, polybenzimidazoles (PBI), having the
thermal capability and processability. Indeed, much of structure shown below, have demonstrated useful
this effort has been devoted more to the modification of potential as high temperature structural adhesive^.'^
existing polymers for improved processability than to W H
the synthesis of novel material^.^^-^^
In this necessarily limited account, examples of high
temperature behaviour will be considered from both
types of adhesive, i.e. the early high temperature systems Intractability problems generally dictate that adhesive
exhibiting processability problems, and the more recent- formulations based on PBIs be employed in low molec-
ly developed systems with improved processing charac- ular weight prepolymer form, with further polymeris-
teristics. ation to high molecular weight polymer being
conducted within the bond-line during joint prep-
Early high temperature adhesives aration. Since this occurs via a polycondensation reac-
tion involving the liberation of copious quantities of
The early commercially available structural adhesives phenol and water, this reaction mechanism can result in
capable of operating at temperatures in excess of 150°C porous adhesive layers having low mechanical strength.
for both short- and long-term applications were of three To moderate such effects, high bonding pressures
principal types, i.e. phenolics, polybenzimidazoles and during cure at 320"C, together with controlled venting
condensation polyimides. procedures during the bonding process, are generally
Phenolic resins, i.e. those based on the initial reaction considered obligatory. Even with such precautions
of phenol and formaldehyde, are rarely used alone as being taken, the fabrication of large bonded areas can
structural adhesives, since they undergo a high degree of remain particularly difficult.
shrinkage during cure which, when combined with pro- In spite of these deficiencies, PBI adhesives offer
nounced brittleness, can generate stress concentrations excellent high temperature capabilities for short-term
at a bond line sufficient to cause debonding. Thus phe- applications with, as indicated in Fig. 12, the ability to
nolics are usually modified with other polymeric retain 50% room temperature strength at approx-
materials yielding, most notably, vinyl, nitrile and epoxy imately 450"C.28 Note, in particular, the capability of
modification^.^^,^^ operating for short times (minutes) at temperatures
Although vinyl- and nitrile-modified phenolics cannot approaching 600"C, where it is likely that they are
be considered for use above approximately W C , superior to all other structural adhesive types in this
epoxy-phenolics offer a quite substantial high tem- respect. Unfortunately, this capability is not maintained
perature capability, providing excellent strength reten- under long-term high temperature conditions owing to
tion for short periods up to approximately 300"C, as the susceptibility of PBI polymer to oxidative degrada-
indicated in Fig. ll.27 This shows tensile lap-shear tion at temperatures in excess of 250°C.28 For this
strength data as a function of test temperature. Also reason, post-cure operations at about 400°C require
included for comparison purposes are data for a stan- inert atmospheres, thus adding further complexity and
dard epoxy adhesive. Although it is clear that signifi- cost to the processing operation.
cantly higher strengths can be obtained at relatively low Although PBIs offer a virtually unique combination
temperatures for the epoxy, temperatures in excess of of properties, they have not enjoyed the reasonable
about 100°C cause a precipitous strength decline. Since success of other high temperature adhesives, such as the
the epoxy-phenolic is relatively unaffected by such tem- polyimides and epoxy-phenolics. A major reason has
peratures, it clearly offers significant strength advan- undoubtedly been associated with the monomers
tages in excess of 100°C. Note, in particular, its ability employed in their preparation, most notably aromatic
to retain in excess of 50% room temperature strength at tetra-amines. These have proved costly and difficult to
250-300°C. obtain in the required purity. In addition, doubts con-
For long-term usage, epoxy-phenolics are limited to cerning carcinogenic activity have been expressed,
a ceiling of about 260°C since severe oxidative degrada- which, together with a difficult processing requirement,
tion results in pronounced mechanical property decline has restricted utilisation. For these reasons, commercial
at higher temperatures. availability has to date been extremely limited.
The poor processability of phenolics stems from the Commercially available condensation polyimides are
condensation reaction mechanisms responsible for generally prepared from dianhydrides and diamines via
polymer formation, which can render adhesive bond- the formation of an intermediate polyamic acid.23Since
lines vulnerable to considerable void formation in the the initial reactants have molecular structures which
absence of high processing pressures, which for some inevitably yield highly intractable polymers, processing
applications can pose serious difficulties. is almost invariably conducted at the polyamic acid
Initially developed in the 1960s under USAF control solution stage, followed by conversion to the polyimide

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

202 S. J . Shaw

Test Temperature " C

Fig. 10. High temperature strength retention for proprietary modified epoxy adhesive.

within the bond-line. Since this occurs by a conden- Unfortunately, owing to their molecular structure, even
sation mechanism, the evolution of water during this the most highly crosslinked epoxies are unable to toler-
conversion, together with the need to remove solvent, ate long-term service at temperatures at or above
provides a serious potential processing difficulty, with 175°C. However, their major advantage concerns their
severely voided bond-lines being highly likely. As with labile processability, i.e. they can be processed without
PBI adhesives, the use of high bonding pressures solvents and without the evolution of volatile by-
together with controlled venting techniques can mini- products during cure. Indeed, most of the recent devel-
mise these difficulties. However, there are usually quite opments with high temperature adhesives have had
stringent limitations on the area which can be suc- epoxy-like processability as a major goal; it is therefore
cessfully bonded using these adhesives. of interest to briefly consider these developments, and
Despite these problems, condensation polyimides can indicate the high temperature capability that bonded
exhibit excellent properties. For short-term application joints can exhibit whilst enjoying reasonable adhesive
they can retain about 50% room temperature strength processability.
at approximately 300"C, with long-term capability close For convenience, the following discussion will con-
to this temperature being attainable with a suitably sider high temperature adhesive developments under
compounded adhesive form~lation.'~ two broad categories relating to whether the organic
constituent of the formulation is thermoset or thermo-
Recent developments in high temperature adhesives plastic in nature.

With epoxy polymers and adhesives, improvements in Thermosetting adhesive developments. The processability
high temperature capability can be obtained by increas- difficulties inherent in many high temperature polymers,
ing crosslink density via a suitable choice of resin which rely on condensation reaction cure chemistry, has
and/or curing agent. The use of polyfunctional epoxy resulted in research programmes directed towards the
resins and curing agents can produce cured products development of low molecular weight, terminally reac-
having glass transition temperatures in excess of 2WC, tive polymers. With many of these systems, poor pro-
resulting in substantial maintenance of mechanical cessability is minimised by the existence of both the
properties at 150°C and beyond. The structure below required molecular high temperature 'architecture' in
shows one particular polyfunctional epoxy which has the polymer backbone together with terminal function-
enjoyed commercial acceptance as the basis of some ality capable of reaction by addition, rather than con-
adhesive formulations :29 densation mechanisms. Typical of such oligomeric
species are adhesives based on bismaleimides,
norbornene-terminated oligomers and acetylenic com-
pounds, most, if not all, being based on an imide back-
bone structure.
The earliest attempt at developing such materials

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 203

30

0
-100 -50 0 50 I 00 I50 200 250 0
Terl Temperature “C

Fig. 11. Lap-shear strength as a function of test temperature for an epoxy-phenolic adhesive and a typical epoxy adhesive.

’.B;i 0 0 0 0 +.*b, 0

LARC- 13
dates back to the 1960s with the concept of norbornene removed prior to bond formation, thus substantially
terminal chemistry, where polymerisation and cross- enhancing processability. Some typical results which
linking is possible via a free-radical mechanism involv- demonstrate the high temperature capability of this
ing the norbornene groups.30 Although early resins adhesive system are given in Tables 1 and 2. Owing to
based on this approach were found not to be acceptable the brittle nature of LARC-13, attempts have been
for structural bonding applications, a major research made to toughen this system with elastomeric agents
programme was initiated by the National Aeronautics with a view to joint strength enhancement, with suc-
and Space Administration (NASA) in the United States cessful results being achieved with peel-type joint
in the 1970s, aimed at the development of norbornene geometries.
systems for both adhesive and composite applications. An alternative route to imide prepolymers using
With regard to adhesives, these efforts culminated in the acetylenic termination was developed during the 1970s
development of a system as LARC-13, the chemical by the Hughes Aircraft Company under the sponsor-
structure of which is shown a b ~ v e : ~ ~ . ~ ’ ship of the United States Air F o r ~ e . ~The ~ , ~ ~
The major advantage of this and indeed other programme was successful in producing
systems based on this molecular approach is that, in acetylene-terminated imide oligomers, typically with the
relation to adhesives based on condensation polyimides, structure shown below, having potential for high tem-
both solvent removal and condensation products are perature use for both adhesive and Composite applica-
tions.
TABLE 1. Typical properties of LARC-13 (Ti
adherends)
TABLE 2. Typical properties of LARC-13
Property Value (composite adherends)

Lap-shear strength (MPa) Substrate Lap-shear strength (MPa)

Room temperature 22.1
232°C 17.9 RT 260°C 31 6°C
After 1000 h at 232°C 17.9
Glass transition temperature, T, (“C) 300 Graphite/PMR-I 5 15.9 - 13.3
Decomposition temperature, T , (“C) 450 Graphite/N R - 150B2 34.5 20.7 11.7

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

204 S . J. Shaw

I I 1 I 1

0 100 200 300 400 500 600

Test Temperature C

Fig. 12. Lap-shear strength as a function of test temperature for a polybenzimidazole adhesive.

Acetylene-terminated imide oligomers

Various studies have demonstrated the excellent proper- enhance thermal ageing stability is also clearly demon-
ties possible using acetylenic i m i d e ~ . ~Glass
~ - ~ tran-
~ strated in Fig. 13.
sition temperatures of 350°C can be achieved, Attempts at modifying the adhesive characteristics of
depending upon molecular structure and post-cure con- acetylene-terminated imides (ATIs) has been carried out
ditions. In addition, a decomposition temperature of via the development of semi-interpenetrating networks
approximately 520°C is close to values obtained from in which the AT1 is allowed to react and crosslink in the
various condensation polyimides. presence of a relatively high molecular weight thermo-
In spite of exhibiting impressive thermal and mechan- plastic p~lymer.~’In particular, St. Clair and Hanky
ical properties, the early acetylenic imide adhesives suf- were able to generate lap-shear strengths of about
fered from a major disadvantage in having close melt 20MPa following ageing for 1000h at 232°C for an
and cure temperatures, thus leading to a very narrow adhesive system based on polyimidesulphone chemistry
processing ‘window’ which can be measured in minutes in both thermoplastic and acetylene-terminated form.40
and seconds at the customary cure temperature. In addition to imide-based oligomers, acetylenic
Although acetylene-terminated imides have been groups have also been used to terminate other species
found suitable for bonding materials such as titanium, such as quinoxalines, with useful adhesive character-
aluminium, copper and various composites, the istics being generated using this a p p r o a ~ h . ~ ~ * ~ ~ * ~ ~
extremely short gel time has been the cause of disap- Another way of introducing improved processability
pointingly low joint strengths at both ambient and ele- into polyimides is to utilise the reactivity of the malei-
vated temperature^.^^.^' Work has shown that the mide group, with resins employing this chemistry being
addition of various cure inhibitors such as hydro- known as bismaleirnide~.~~ A simplified structure is
quinone, reduces cure rate resulting in a greatly shown below:
increased gel time. This has allowed significant
improvements in strength both at ambient and elevated
temperatures, as shown in Fig. 13. As indicated, in the
absence of hydroquinone, an ambient tensile lap-shear
strength of 15.9 MPa is reduced to 5.7 MPa at 285°C.
However, hydroquinone incorporation, particularly with R being predominantly aromatic in nature.
when combined with the use of a primer, increases lap- With such systems, heating at temperatures greater
shear strength considerably, both at ambient and ele- than approximately 150°C results in reaction of the ter-
vated temperatures. The ability of hydroquinone to minal maleimide groups by a free-radical mechanism,

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 205

20

--------__---__ 41

-P .
-
.-al
'm

-+Aged at 257°C
Adhesive + 5 % hydroquinone aged at 257°C

I I 1 I
0
0 200 400 600 800 1,000
Ageing Time (hours)

Fig. 13. Lap-shear strength as a function of ageing time for acetylene-terminated polyimide adhesive joints

thereby eliminating the processing difficulties associated Thermoplastic adhesive developments. As discussed pre-
with many of the other high temperature polymers. viously, the early condensation polyimide-based adhe-
Owing to the presence of aliphatic groups in the sives were generally derived from anhydride and amine
structure of crosslinked bismaleimides (derived prin- reactants which restricted thermoplastic character and
cipally from the aliphatic nature of the original malei- solubility in the polymer, thus requiring processing at
mide terminal groups), adhesives based on BMIs are the intermediate polyamic acid solution stage. Recent
usually only considered for service applications developments have attempted to diminish the pro-
requiring a 200-230°C long-term capability. In this cessing difficulties which accompany this process by
respect they can be regarded as essentially 'filling the concentrating on introducing molcular species into the
gap' between epoxies and other polyimides, such as the backbone of the polymer, which would enhance
condensation and acetylenic-terminated types. Their thermoplastic character. If this could be successfully
short-term high temperature capabilities are, however, achieved then bonding, using for example a fully imid-
very reasonable, as indicated by the data in Table 3, ised film, would be possible, thus removing the need for
which show over 50% retention of room temperature both solvent evaporation and polyamic acid to poly-
strength at 260°C for a proprietary adhesive system. imide conversion within the bondline.
In addition, the low bonding pressure of 0.2MPa One of the most successful attempts made using this
shown in Table 3, relative to that required for more tra- approach has been the development of a polyimide
ditional high temperature adhesives such as conden- adhesive known as LARC-TPI by NASA in the
sation polyimides and epoxy-phenolics, can be USA.31*44 The enhanced thermoplasticity of this system
attributed to the addition curing mechanism which in comparison to more transitional condensation poly-
occurs with bismaleimides and which allows the forma- mers has been attributed to the presence of flexibilising
tion of bond-lines of very low void content.

TABLE 4. Typical properties of LARC-TPI (with Ti

TABLE 3. Lap-shear data for bismaleimide (paste ad herends)
adhesive) bonded a Iurniniurn joints
Property Value
Test temperature ("C) Lap-shear strength (MPa)
Lap-shear strength ( M Pa)
25 20.0 Room temperature 36.5
232 18.6 232°C 13.1
260 12.7 After 3000 h at 232°C 20.7
Glass transition temperature, T, ("C) 250
Cure conditions: 1 h at 177"C, 0.2MPa+2h at 246°C Decomposition temperature, T , ("C) 520
post-cure

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

206 S . J . Shaw

carbonyl groups spaced periodically along the molecu- polymer had demonstrated a titanium lap-shear
lar backbone, as shown in the structure strength of 33.3 MPa at room temperature, falling
to 21MPa at 200°C whilst similar work on a
polyaryleneetherbenzimidazole demonstrated equally
impressive characteristics.

Some typical adhesive joint properties obtained from OTHER DEMANDING ENVIRONMENTS
LARC-TPI adhesive on titanium adherends as shown
in Table 4.31 In addition to enhancing processability, In addition to atmospheric moisture and elevated tem-
the thermoplastic character of the polymer can be seen perature, other potentially harsh environmental condi-
to enhance joint strength substantially, with a very tions would include stress, radiation and various other
impressive 36.5 MPa at room temperature. However, liquid contaminants, e.g. fuels, de-icing fluids, hydraulic
owing to a glass transition temperature of 250"C, adhe- oils, etc. In addition, mention must be made of low tem-
sive strength drops quite substantially above 200°C-an perature conditions where, for example, adhesives
inevitable result of molecular flexibilisation. employed in aircraft construction would be required to
More recent work on the LARC-TPI system has resist the effects of temperature down to -50°C.
focused on attempts at promoting water solubilisation However, of major importance would be the combined
as a means of diminishing reliance on organic solvents, influence of stress, particularly cyclic, superimposed on
in addition to cost reduction attempts via changes in the two primary environments discussed above, i.e.
the types of initial reactants employed in the original moisture and elevated temperature. Evidence is grad-
polymer ~ y n t h e s i s . ~ ~ * ~ ~ ually being accumulated that demonstrates the severe
Various other types of thermoplastic polymer have effects such combinations can have on adhesive joint
been developed with a view to high temperature performance and which must be considered in the initial
bonding application^.^*-^^ Harris and co-workers design of a bonded structure.
recently developed a thermoplastic polyimide, having
the structure shown below, which exhibited a room
temperature lap-shear strength with titanium adherends ACKNOWLEDGEMENT
of 54 MPa. Although a substantial strength reduction
down to approximately 28 MPa was observed at 120°C, Published with the permission of the Controller of Her
the room temperature value can be regarded as Britannic Majesty's Stationery Office. British Crown
extremely impressive. Copyright DERA/1996.

REFERENCES

1 Butt, R. K. & Cotter, J. L., J . Adhesion, 8 (1976) 11.

Continuing with polyimide systems, Hergenrother 2 Parker, B. M. & Shaw, S. J., Defence Research Agency report,
and Havens synthesised a range of polyimides contain- DRS/SMC/CR951184, Oct. 1995.
3 Bolger, J. C., Adhesion and Adhesives Vol 3, ed. R. L. Patrick.
ing carbonyl and other connecting groups.55 One Marcel Dekker, New York, 1973.
system in particular, having the structure below, when 4 Buck, B. I. & Hockney, M. G. D., Aspects of Adhesion-7, eds K. W.
employed to bond titanium substrates exhibited lap- Allen and D. J. Alner. Transcription Books, London, 1973.
5 Wright, W. W., Composite, 12 (1981) 201.
shear strength values of 43 MPa at room temperature 6 Kinlock, A. J., Adhesives and Adhesion, Science and Technology.
and 31.1 MPa at 177°C. However, as is common with Chapman and Hall, London, 1987.
thermoplastic-based adhesives, a significant strength 7 Gledhill, R. A. & Kinloch, A. J., J . Adhesion, 6 (1974) 315,
8 Noland, J. S., Adhesion Science and Technology, ed. L. H . Lee.
reduction was observed as temperature was increased Plenum, New York, 1975.
towards the polymer's glass transition temperature, only 9 Plueddemann, E. P., Silane Coupling Agents. Plenum, New York,
4.1 MPa being observed at 232°C. 1982.
10 Gledhill, R. A., Shaw, S. J. & Tod, D. A., Int. J . Adhesion Adhe-
sives, 10 (1990) 192.
11 Mittal, K. L. (ed), Silanes and Other Coupling Agents. VSP,
Utrecht, 1992.
12 Maddison, A., Handbook of Adhesion ed. D. E. Packham.
Longman, London 1992, p. 51.
13 Bishopp, J. A. & Thompson, G. E., Surf: Interface Anal., 20 (1993)
Recently, a significant effort has been devoted to 485.
incorporating heterocyclic units into polyarylene ethers 14 Comyn, J., Developments in Adhesive-2, ed. A. J. Kinloch. Applied
with a view to the development of adhesives, composites Science, London, 1981, p. 279.
15 Shaw, S. J., Tod, D. A. & Griffith, J. R., Adhesives, Sealants and
matrices, coatings and Work by Hergenro- Coatings for Space and Harsh Environments, ed. L. H. Lee. Plenam,
ther et al. on a polyaryleneetherphenylimidazole New York, 1988, p. 45.

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996

Adhesives in demanding applications 207
16 Johncock, P. & Tudgey, G. F. Br. Polym. J., 15 (1983) 14. 39 Pascal, T. & Sillion, B., Advances in Interpenetrating Polymer Net-
17 Bade, J. A., Sagoo, P. S. & Johncock, P., Polymer, 26,(1985) 1167. works iv, eds. D. Klemprier and K. C. Frisch. Technomic 1994,
18 Goobich, J. & Marom, G., Polym. Eng. Sci., 22 (1982) 1052. p. 141.
19 G f l i t h J. R. & Quick, J. E., Adv. Chem. Ser. 92 (1970) 8. 40 St. Clair, T. L. & Hanky, 0. A., 30th Natl. SAMPE Symp., (1985)
20 Griffth, J. R., Sands, A. G. &. Cowling, J., Adv. Chem. Ser., 99 p. 912.
(1971) 471. 41 Kovar, R. F., J. Polym. Sci., Polyn. Chem., Edn., 15 1081 (1977).
21 Gmith, J. R., ORear, J. G. & Reines, S. A,, Chemtech, (1972) 311. 42 Maximovich, M. G., Adhesives Age, Oct (1977) 40.
22 Griftith, J. R., Chemtech, (1982) 290. 43 Stenzenberger, H. D., Herzog, M., Romer, W., Scheiblich, R. &
23 Critchley, J. P., Knight, G. J. & Wright, W. W., Heat Resistant Reeves, N. J., Br. Polym. J., 15 (1983) 2.
Polymers. Plenum, New York, 1983. 44 St. Clair, A. F. & St. Clair, T. L., SAMPE Q., 13 (1981) 20,
24 Hergenrother, P. M. (ed.), Advantages in Polymer Science, 117, 45 St. Clair, T. L. & Progar, D. J., Adhesion Science and Technology,
High Performance Polymer. Springer-Verlag, Berlin. 1994. ed. L. H. Lee.Plenum, New York, 1975, p. 187.
25 Lin, S.-C. & Pearce, E. M., High Performance Thermosets, Chem- 46 Hendricks, C. L. & Hill, S. G., Polyimides-Synthesis, Character-
istry, Properties, Applications. Hanser, Munich, 1994. isation and Applications, Vol. 2, ed. K. L. Mittal. Plenum, New
26 Tredwell, S., Handbook of Adhesion, ed. D. E. Packham. Longman, York, 1984, p. 1103.
London, 1992, p. 305. 47 Pike, R. A,, Polyimides- Synthesis, Characterisation and Applica-
27 Wake, W. C., Adhesion and the Formulation of Adhesives, 2nd ed. tions, Vol. 1, ed. K. L. Mittal. Plenum, New York, 1984, p. 15.
Applied Science, London, 1982. 48 St. Clair, T. L. & Yamake, D., Polimides-Synthesis, Character-
28 Litrak, S., Adhesives Age, 11,(1968) 17. isation and Applications, Vol. 1, ed. K. L. Mittal. Plenum, New
29 Morgan, R. J., Developments in Reinforced Plastics-1, ed. G. Prit- York, 1984, p. 99.
chard. Applied Science, London, 1980, p. 211. 49 Dezern, D. & Young, P. R., Int. J. Adhesion Adhesives, 5 (1985)
30 Lubowitz, H. R., US Patent 3528980,1970. 183.
31 St. Clair, A. K. & St. Clair, T. L., Polyimides-Synthesis, Character- 50 Berger, A., Polyimides- Synthesis, Characterisation and Applica-
isation and Applications, vol. 2, ed. K. L. Mittal. Plenum, New tions, Vol. 1, ed. K. L. Mittal. Plenum, New York, 1984, p. 67.
York, 1984, p. 977. 51 Choudhary, B., Polyimides- Synthesis, Characterisation and Appli-
32 St. Clair, T. L. & Progar, D. J., Proc. 24th Natl. SAMPE Symp. 24 cations, Vol. 1, ed. K. L. Mittal. Plenum, New York, 1984, p. 401.
(1979) 1081. 52 Johnson, R. D. & Burlhis, H. S., J . Polym. Sci., Polym. Symp., 70
33 Bilow, N., Landis, A. L. & Miller, L. J., US Patent 3845018,1974. (1983) 129.
34 Landis, A. L., Bilow, N., Boscham, R. H., Lawrence, R. E. & 53 Serfaty, I. W., Polyimides- Synthesis, Characterisation and Applica-
Aponyi, T. J., Polym. Prepr., 15 (1974) 537. tions, Vol. 1, ed. K. L. Mittal. Plenum, New York, 1984, p. 149.
35 Bilow, N. & Aponyi, I. J., Proc. 20th Natl. SAMPE Symp., 20 54 Hams, F. W., Beltz, M. W. & Hergenrother, P. M., SAMPE J.,
(1975) 618. Jan/Feb (1987) 6.
36 Bilow, N. & Landis, A. L., Proc. 8th SAMPE Tech. ConJ, 8 (1976) 55 Hergenrother, P. M. & Havens, S. J., J. Polym. Sci., A : Polym.
94. Chem., 27 (1989) 1161.
37 Boscham, R. J. & Landis. A. L., Proc. 21st Natl. SAMPE Svmn., - - 21 56 Connell, J. W., Hergenrother, P. M. & Wolf, P., Polymer, 33 (1992)
(1976) 356. 3507.
38 Bilow, N., Landis, A. L., Boscham, R. H. & Fasold, J. G., SAMPE 57 Hergenrother, P. M., Smith, J. G. & Connell, J. W., Polymer, 34
J.,Jan/Feb (1982) 8. (1993) 856.

POLYMER INTERNATIONAL VOL. 41, NO. 2, 1996