Epoxy modified with triblock copolymers: morphology, mechanical
properties and fracture mechanisms
Jing Chen, Ambrose C. Taylor*
Department of Mechanical Engineering, Imperial College London, South Kensington Campus, London
SW7 2AZ, UK.
* Corresponding author:
Tel. +44 2075947149, Fax: +44 2075947017, Email: a.c.taylor@imperial.ac.uk
Abstract
The morphology, fracture toughness, and mechanical properties of an anhydride-cured diglycidylether
of bisphenol A (DGEBA) epoxy polymer modified with poly(methyl methacrylate)-bpoly(butylacrylate)-b-poly(methyl methacrylate) (MAM) have been investigated. The addition of three
different MAM triblock copolymers (M22N, M52N, and M52) to the epoxy polymer gives two
different microstructures. An organised nanostructure with well-dispersed worm-like micelles was
obtained using M22N, and the addition of M52N or M52 gives dispersed micron-size particles in the
epoxy matrix or a co-continuous microstructure at higher MAM contents. These triblock copolymers
toughen the epoxy polymer significantly, with only a slight reduction of the mechanical and thermal
properties of the epoxy polymer. The maximum values of fracture toughness and fracture energy (1.22
MPam1/2 and 450 J/m2) were measured using 12 wt% M22N, which is an increase of 100% and 350%
respectively compared to the unmodified epoxy. The M52 and M52N modified materials show a
maximum toughness when a co-continuous microstructure is formed. The potential toughening
mechanisms are identified and discussed.
Keywords
Epoxy; Block copolymer; Nanostructure; Fracture toughness; Toughening mechanisms; Mechanical
properties
1
1. Introduction
Epoxy polymers are a class of high-performance thermosetting polymers, which are known for their
excellent engineering properties, such as a high modulus, low creep, high strength, and good thermal
and dimensional stabilities. However, epoxy polymers inherently have low toughness and impact
resistance due to their highly crosslinked structure. This phenomenon leads to brittle behaviour and
causes the polymer to suffer from poor resistance to crack initiation and growth. To improve the
toughness of epoxy polymers, one of the most successful conventional methods is to incorporate rubber
modifiers into the epoxy systems to form a multi-phase structure with a discrete rubbery phase [1-3].
These polymer modifiers that are typically added to the epoxy systems are linear homopolymers and
random copolymers, which generally give a second soft phase in the size of micrometres in the cured
epoxy polymers [3-5]. Recently, the use of block copolymers (BCPs) to generate micro- and
nanostructured phases has shown great potential to toughen epoxy polymers, e.g. [3, 5-12].
Amphiphilic block copolymers with epoxy-miscible blocks and epoxy-immiscible blocks can selforganize into nanostructures in the uncured epoxy polymer by using the epoxy as a selective solvent,
from which a variety of substructures can be obtained, e.g. vesicles, spherical micelles, and worm-like
micelles [3, 8, 13-16]. These self-organized nanostructures can subsequently be preserved by the
polymerisation process of the epoxy polymer. However the successful preservation of the nanostructure
in the cured epoxy polymer depends on the composition of the block copolymer, the miscibility of the
epoxy-miscible blocks and the epoxy precursor, the curing agent, and the curing process. It was
proposed that the key factor determining whether the nanostructuration can be preserved is that the
miscible block of the block copolymer needs to remain miscible with the epoxy polymer up to a very
high conversion ratio during the epoxy polymerisation process [3, 14, 15, 17, 18].
Due to the great potential of BCP toughening, considerable attention has been given to this area after
Hillmayer et al. [19] first reported the ordered nanostructure of an epoxy polymer modified with
poly(ethylene oxide)-b-poly(ethylene) (PEO-PEE) and poly(ethylene oxide)-b-poly(ethylene-altpropylene) (PEO-PEP). Since then, numerous studies have reported that incorporating amphiphilic
diblock and triblock copolymers into epoxy polymers can significantly improve the toughness of the
epoxy polymer, with only a minor reduction in the modulus and the glass transition temperature, Tg, [46, 9-13, 20, 21]. In contrast, a significant reduction in the modulus and the Tg is often caused by
conventional rubber toughening, e.g. [1, 22, 23]. However, most research on the effect of block
copolymers only focuses on the effect of the morphology on the mechanical properties of the cured
epoxy polymer. Few studies investigate the toughening mechanisms of the block copolymer toughened
epoxy polymer. Dean et al. [16] investigated the toughening mechanisms of a PEO-poly(1,2-butadiene)
(PB) and poly(methyl methacrylate-co-glycidyl methacrylate)-poly(2-ethylhexyl methacrylate) (MMAEH) modified epoxy system which used poly(bisphenol A-co-epichlorohydrin) (BPA348) as the epoxy
and 4,4’-methylenedianiline (MDA) as the hardener. They ascribed debonding and the subsequent void
growth and plastic deformation as the toughening mechanisms for their nanostructured epoxy polymer
with dispersed vesicles. Liu et al. [6] investigated the toughening mechanisms of a poly(ethylene-altpropylene)-b-poly(ethylene oxide) (PEP-PEO) toughened epoxy system which used diglycidyl ether of
bisphenol-A (DGEBA) epoxy and 1,1,1-tris(4-hydro-xyphenl)ethane (THPE) as the hardener. The
2
morphology of the resulting epoxy polymer contained dispersed worm-like micelles. The authors
proposed that the toughening mechanisms responsible were a combination of many mechanisms:
cavitation or debonding, shear yielding, crack tip blunting, crack bridging, and viscoelastic energy
dissipation [6]. Nevertheless, many studies either only briefly mention the toughening mechanisms of
the block copolymer modified epoxy systems, or the toughening mechanisms were the subject of
preliminary work only.
This study systematically investigates the microstructure, the mechanical properties, and the
toughening mechanisms of an epoxy polymer modified with block copolymers. Three different
commercial poly(methyl methacrylate)-b-poly(butylacrylate)-b-poly(methyl methacrylate) (MAM)
triblock BCPs were used. They were M22N, M52N, and M52 from Arkema, France, and possess
different epoxy compatibility. They were used to modify an anhydride cured DGEBA resin, which has
previously been used to investigate silica nanoparticle toughening [24, 25]. A set of mechanical tests
and microscopy work were carried out to study these MAM modified epoxy polymers. The blend
morphology, structure/property relationship, and thermal-mechanical behaviour of the modified epoxy
polymers are discussed. The toughening mechanisms involved are also identified and discussed.
2. Materials and Experimental Procedure
2.1. Materials
An anhydride-cured epoxy system was used. The epoxy resin was a standard diglycidyl ether of
bisphenol-A (DGEBA, Araldite LY556) with an epoxide equivalent weight (EEW) of 186g/eq,
supplied by Huntsman, UK. The curing agent was an accelerated methylhexahydrophthalic acid
anhydride (Albidur HE600) with an anhydride equivalent weight (AEW) of 170g/eq, supplied by
Nanoresins, Germany. Three different poly(methyl methacrylate)-bloc-poly(butylacrylate)-blocpoly(methyl methacrylate) (PMMA-b-PbuA-b-PMMA) block copolymers (M22N, M52, and M52N)
supplied as powders by Arkema, France, were used as modifiers. The molecular weight and the softphase content of the M52 and M52N are similar, both being categorised as ‘medium’ in the Arkema
datasheet [26]. The suffix N indicates that the MAM incorporated dimethylacrylamide (DMA) into the
PMMA blocks to increase the compatibility of the miscible block with more polar curing agents [26,
27]. The M22N has a higher molecular weight, and a slightly lower soft-phase content, than the M52
and M52N. M22N is also described as more polar than M52N [26].
The bulk samples were prepared by gently mixing the DGEBA epoxy resin with the relevant amount of
the BCPs at room temperature to avoid agglomeration of the powders. The temperature was increased
to 120°C, and then the mixture was stirred for at least 60 minutes using a mechanical stirrer to ensure
all the powder was fully dissolved in the epoxy resin. The resin mixture was subsequently put into a
vacuum oven and degassed at 90°C and -1 atm for at least 15 minutes until the mixture was free of
bubbles. A stoichiometric amount of the anhydride curing agent was then added, and stirred thoroughly
for 15 minutes. The resin mixture was degassed again at 70°C and -1 atm. After degassing, the mixture
was poured into release-agent (Frekote 770, Henkel) coated steel moulds to produce plates of bulk
polymer from which samples for testing could be machined. The steel moulds were placed in an oven
3
and the resin mixture was cured at 90°C for 60 minutes, plus a post-cure of 160°C for another 120
minutes.
Various formulations were prepared in this study. They are named according to the following rule, viz.
unmodified epoxy (termed ‘control’), epoxy with MAM block copolymers (termed XY), where X
refers to the amount of BCP contained in the total formulation by percentage weight and Y refers to the
MAM type. Up to 12 wt% of each BCP was used.
2.2. Bulk Thermal and Mechanical Tests
2.2.1. Dynamic Mechanical Analysis
The glass transition temperature, Tg, of all the bulk samples was measured using dynamic mechanical
analysis (DMA) with a Q800 DMA machine from TA Instruments in double cantilever mode at 1 Hz,
on specimens of 60 x 10 x 3 mm3. The temperature range was -80 to 200°C with a heating rate of
4°C/min. The value of Tg was determined at the peak value of tan δ. The number average molecular
weight between cross-links, Mnc, was also calculated from the equilibrium modulus, Er, in the rubbery
region by using
𝑀𝑛𝑐 = 𝑞𝜌𝑅𝑇/𝐸𝑟
(1)
where T is the temperature at which the value of Er was taken, ρ is the density of the epoxy at
temperature T, R is the universal gas constant, and q is the front factor (a value of 0.725 was used to
obtain reasonable result, because only density of the epoxy at 23°C was measured) [22, 28]. The
density of the epoxy samples was measured according to BS ISO 1183-1 method A [29].
2.2.2. Uniaxial Tensile Tests
Uniaxial tensile tests were conducted in accordance with the BS ISO 527 standard [30], using an
Instron 5584 universal testing machine. Dumbbell specimens with a gauge length of 25 mm were
machined from the bulk plates. The tests were performed at a constant displacement rate of 1 mm/min
and a test temperature of 21°C. The displacement over the gauge length of the samples was accurately
measured using an Instron 2620-601 dynamic extensometer. The maximum tensile stress for each
sample was recorded, and the elastic modulus, E, was calculated between strains of 0.0005 - 0.0025. At
least five samples were tested for each formulation.
2.2.3. Plane Strain Compression Tests
Plane-strain compression (PSC) tests were performed after Williams and Ford [31], using bulk polymer
samples to obtain the yield stress and the high strain behaviour, neither of which can usually be
obtained from tensile tests because of the brittleness of the samples. The tests were performed at 21°C
using an Instron 5585H using a constant displacement rate of 0.1 mm/min to approximately match the
strain rate from the tensile tests. Samples with a size of 40 x 40 x 3 mm3 were used, and loaded in
compression between two parallel 12 mm wide dies. A minimum of two specimens were tested for
each epoxy polymer formulation. The results of the tests were corrected by subtracting the machine and
rig compliance. Based on the von Mises criteria, the true compressive stress, σc, was calculated using:
4
√3
𝜎𝑐 = ( 2 ) 𝜎𝐸
(2)
where σE is the engineering stress. The true compressive strain, εc, was calculated using:
2
𝐵
𝜀𝑐 = ( ) ln ( 𝐵𝑐)
(3)
√3
where Bc is the compressed thickness and B is the initial thickness [25].
2.2.4. Single-Edge Notch Three-Point Bending (SENB) Tests
The single-edge notch three-point bending (SENB) tests were conducted in accordance with the BS
ISO 13586 standard [32], using an Instron 3369 universal testing machine equipped with an Instron
2620-601 dynamic extensometer to obtain the fracture toughness, KIc, and the fracture energy, GIc, of
the samples. The SENB samples were machined from the bulk plates, and the pre-cracks were
produced by tapping a liquid nitrogen chilled razor blade into the notch. The lengths of the pre-cracks
were measured using a Nikon SMZ800 stereo microscope. The tests were conducted at a constant
displacement rate of 1 mm/min and a test temperature of 21°C. All the samples failed by unstable crack
growth, and the fracture toughness was calculated using:
𝑃
𝐾𝐼𝑐 = 𝐵𝑊 1/2 𝑓(𝛼/𝑊)
(4)
where P is the critical load, B is the sample thickness, W is the specimen width, a is the average precrack length, and f(a/W) is the non-dimensional shape factor [32]. The plane strain fracture energy was
calculated from KIc using [32]:
𝐺𝐼𝑐 =
2
𝐾𝐼𝑐
(1 −
𝐸
𝑣 2)
(5)
where υ is the Poisson’s ratio. A value of υ = 0.35 was used, as is typical for epoxy polymers [33].
2.2.5. Double-Notched Four-Point Bending (DN-4PB) Test
Double-notched four-point bending (DN-4PB) tests were conducted to investigate the plastic
deformation zone ahead of the sub-critically loaded crack tip, to understand the contribution and the
sequence of the toughening mechanisms. The DN-4PB tests were performed as described by Sue and
Yee [34]. An Instron 5584 universal testing machine was used to load the specimens in four-point
bending, at a constant displacement rate of 1 mm/min and a temperature of 21°C. The samples were
machined from the bulk plates and the pre-cracks in the samples were produced by tapping a liquid
nitrogen chilled razor blade into the notches. Care was taken to ensure the four loading points contacted
the specimens simultaneously in the tests. After fracture occurred from one pre-crack, the plastic zone
at the tip of the other, sub-critically loaded, crack was examined using optical microscopy.
2.3. Microscopy studies
2.3.1. Atomic Force Microscopy
Atomic force microscopy (AFM) was performed, using a MultiMode scanning probe microscope from
Veeco equipped with a NanoScope IV controller and an ‘E’ scanner, to obtain the polymer morphology.
The smooth surface of the samples was prepared using a PowerTome XL ultramicrotome from RMC.
Silicon probes were used in tapping mode. Both height and phase images were captured at 512 x 512
pixel resolution and at a scan speed of 1 Hz. For the phase images, which are sensitive to viscoelastic
5
properties, the apparent hardness of the material is shown by the colour of the phase images, in which
the harder phases are brighter [35].
2.3.2. Field Emission Gun Scanning Electron Microscopy
A Leo 1525 (Zeiss) scanning electron microscope equipped with a field emission gun (FEG-SEM) was
used to obtain high resolution images of the fracture surfaces. The accelerating voltage used was 5 kV.
All the samples were sputter-coated with a thin layer of chromium to prevent charging.
2.3.2. Optical Microscopy
Optical microscopy was carried out using an AXIO microscope (Zeiss) to investigate the plastic
deformation zone ahead of the crack tip in the DN-4PB specimens. Samples were cut out from the
central plane strain region perpendicular to the fracture plane and parallel to the crack direction using
an Accutom-5 precision cutter (Struers) equipped with an E0D15 diamond blade. The samples were
then mounted to glass microscope slides using an optically transparent adhesive (Araldite 2020,
Huntsman). After mounting, these samples were ground and polished to a nominal thickness of 100 µm
for observation. Samples were observed using transmission optical microscopy (TOM), with white and
cross-polarised light.
3. Results and Implications
3.1. Introduction
The commercial MAM block copolymers used in this study are non-reactive pure acrylic symmetric
triblock copolymers which consist of a centre epoxy immiscible poly(butylacrylate) (PbuA) block and
two epoxy miscible end blocks of poly(methyl methacrylate) (PMMA). These block copolymers were
designed to achieve nanostructuration in thermoset systems via self-assembling before curing and
polymerisation-induced fixation after curing [3, 26]. In this study, the epoxy polymer modified with the
three different BCPs can be sorted into two different categories - transparent samples and opaque
samples. All of the M22N modified samples are optically transparent before and after the curing
process, but the optical transparency of the M52N and M52 modified samples can only be observed
before gelation. This indicates that, after the curing process, no macroscopic phase separation that
exceed the wavelength of visible light (about 100 µm) was happened in the M22N modified epoxy
polymers, while macroscopic phase separation occurred in the M52N and M52 modified samples
which gave rise to the opacity. This finding shows that M22N is more compatible to the selected
anhydride-cured epoxy system compared to the M52N and M52.
3.2. Viscoelastic Properties
The storage modulus, E’, and Tg values measured using DMA are summarised in Table 1. The glass
transition temperature of the control was 161°C, and the number average molecular weight between
crosslinks, Mnc, of the control, estimated using the theory of rubber elasticity, is 359-461(410) g/mol,
which indicates that the control is a intermediately crosslinked epoxy polymer [22, 28]. The addition of
MAM only causes a slight reduction in the Tg of the epoxy polymer, to 158 ±1°C, and the reduction is
approximately the same for the three BCPs. There is also no trend in the reduction for increasing
6
content of BCP, see Table 1. For example, Figure 1 shows E’ and tan δ versus temperature for the
control, 7 wt% M22N, 7 wt% M52, and 7 wt% M52N epoxy formulations. For the MAM modified
samples, the storage modulus was lower relative to the control over the whole temperature range,
indicating that the MAM is less stiff than the epoxy. The storage modulus of the 7 wt% M22N samples
was higher relative to the 7 wt% M52N or M52 samples until a temperature of 127°C was reached.
These results show that MAM modified samples with an ordered nanostructure retained their rigidity
better than the MAM modified samples with the larger phase separation.
From Figure 2, one large relaxation peak was observed for the control and all the MAM modified
epoxy polymers, which is the relaxation peak of the epoxy polymer. However, a small shoulder on the
tan δ curves was also distinguished next to the main relaxation of the epoxy. These observations
suggest that some PMMA was phase separated from the epoxy matrix and the relaxation peak of the
PMMA miscible blocks largely overlapped with the main epoxy relaxation. It is also can be seen in
Figure 2 that different extents of phase separation of PMMA occurred in the MAM modified epoxy
polymers. Phase separation of PMMA in the 7 wt% M22N samples is higher than the 7 wt% M52N and
7 wt% M52 samples. This result corroborates the previous suggestion that M22N may have a higher
PMMA/PbuA ratio compared to M52N and M52.
Table 1. Glass transition temperature, storage moduli, elastic moduli, and tensile strength of the control and MAM modified
epoxy samples at 20°C.
M22N
Tg (˚C)
E' (GN/m2)
E (GN/m2)
σt (MN/m2)
Control
161
3.38
2.90(±0.02)
81(±4)
2% M22N
158
3.35
2.87(±0.05)
76(±5)
3% M22N
158
3.07
2.85(±0.03)
75(±3)
5% M22N
158
2.84
2.80(±0.02)
75(±5)
7% M22N
158
3.15
2.80(±0.04)
74(±4)
10% M22N
154
3.06
2.82(±0.03)
76(±1)
12% M22N
159
3.13
2.69(±0.05)
76(±0)
M52
Tg (˚C)
E' (MN/m2)
E (GN/m2)
σt (MN/m2)
Control
161
3.38
2.90(±0.02)
81(±4)
2% M52
157
3.21
2.78(±0.08)
74(±5)
3% M52
157
3.00
2.81(±0.10)
77(±3)
5% M52
156
2.86
2.74(±0.03)
74(±3)
7% M52
157
2.65
2.60(±0.03)
68(±4)
10% M52
159
2.56
2.33(±0.04)
29(±1)
12% M52
158
2.59
2.25(±0.07)
23(±2)
M52N
Tg (˚C)
E' (GN/m2)
E (GN/m2)
σt (MN/m2)
Control
161
3.38
2.90(±0.02)
81(±4)
2% M52N
159
3.37
2.81(±0.03)
68(±5)
3% M52N
156
2.91
2.81(±0.08)
76(±1)
5% M52N
159
2.68
2.69(±0.05)
73(±2)
7
7% M52N
158
2.79
2.63(±0.05)
72(±1)
10% M52N
159
2.54
2.32(±0.05)
29(±1)
12% M52N
159
2.90
2.18(±0.02)
24(±1)
Control (E')
6000
1.4
7%M22N (E')
5000
1.2
7%M52 (E')
Control (tan δ)
1
7%M22N (tan δ)
4000
7%M52N (tan δ)
0.8
7%M52 (tan δ)
3000
0.6
Tan δ
Storage Modulus, E' (MN/m2)
7%M52N (E')
0.4
2000
0.2
1000
0
0
-0.2
-100
-50
0
50
100
150
200
250
Temperature (°C)
Figure 1. DMA data, storage modulus (E’) and loss factor (tan δ) versus temperature, for the control sample and samples
modified with 7 wt% of M22N, M52N, and M52.
8
Control (E')
6000
0.1
7%M22N (E')
0.08
7%M52 (E')
Control (tan δ)
7%M22N (tan δ)
4000
0.06
7%M52N (tan δ)
7%M52 (tan δ)
3000
0.04
2000
0.02
1000
0
0
Tan δ
Storage Modulus, E' (MN/m2)
7%M52N (E')
5000
-0.02
-100
-50
0
50
100
150
200
250
Temperature (°C)
Figure 2. DMA data, storage modulus (E’) and loss factor (tan δ) versus temperature, for the control sample and samples
modified with 7 wt% of M22N, M52N, and M52 with magnified tan δ axis.
3.3. Morphology
The AFM phase images show the morphologies of the MAM modified epoxy polymers, and selected
images are shown in Figure 3. The morphology of the control was homogeneous and featureless, as
expected for a single phase thermosetting material, see Figure 3a. Heterogeneity was introduced by
adding MAM. For the epoxy polymer containing M22N, wormlike micelles with a clear core/shell
structure and a size ranging from 6 to 13 nm in width were observed in all the formulations containing
from 2 wt% to 12 wt% of M22N, as shown in Figure 3(b-d). The exception was the co-existence of
wormlike micelles and spherical micelles that was partly observed in the epoxy polymers containing 12
wt% M22N, although these were rare. By considering the volume fraction of each block in the MAM
and the differences in viscoelastic properties between the epoxy matrix and the MAM, the core of the
dispersed wormlike micelles was ascribed to the immiscible PbuA block, and the harder shell to the
locally phase separated PMMA blocks. These findings are corroborated by the DMA results. The
morphology of the M22N modified epoxy polymers shown in Figure 3(b-d) suggests that these epoxy
polymers have either a three-dimensional bicontinuous gyroid microstructure, as shown in Figure 4, or
a microstructure with irregular wormlike micelles well dispersed in three dimensions.
For the opaque epoxy polymer samples containing M52N and M52, micro-phase separation was
confirmed. Micron size phase-separated MAM particles are seen in the cured epoxy polymers, for
example the polymers containing 7 wt% M52N and 7 wt% M52 are shown in Figure 3(e-f). Within the
MAM particles, an ordered nanostructure with wormlike micelles can be seen, which shows the
immiscibility of the M52N and M52 in this epoxy system. The occasional partly phase-inverted particle
9
was also observed using 7 wt% of M52, see Figure 3e, although these were relatively rare. It was also
noted that the size of the phase separated MAM particles increased linearly as the content of the M52N
or M52 increased. AFM also indicated that there was relatively good interfacial adhesion between the
epoxy matrix and the particles.
The use of 10 wt% of M52 or M52N gave a co-continuous microstructure, in which epoxy rich matrix
with BCP particles plus phase-inverted BCP rich matrix with epoxy particles was found, as may also be
seen on the fracture surfaces discussed below. Phase inversion of the cured MAM modified epoxy
samples was observed when 12 wt% of M52N and M52 were added to the epoxy polymer. Here the
microstructure showed a MAM matrix containing particles of epoxy which were in the range of 1.743.73 μm and 1.94-4.38 µm in radius respectively. However, no phase inversion was observed for any
cured M22N modified epoxy samples even up to a high MAM content of 12 wt%. This is consistent
with the findings of other studies where phase inversion will not occur if full nanostructuration of the
block copolymer modified thermosetting polymers is achieved, even up to very high BCP
concentrations [3, 11, 12, 14, 15, 17, 18].
(a)
(b)
(c)
(d)
10
(e)
(f)
(g)
(h)
Figure 3. AFM phase images of epoxy polymers: (a) control, (b) 3 wt% M22N, (c) 10 wt% M22N, (d) 12 wt% M22N,
(e) 3 wt% M52, (f) 7 wt% M52, (g) 3 wt % M52N, (h) 7 wt% M52N.
Epoxy network
PMMA block
PbuA block
11
Figure 4. 3D bicontinuous gyroid microstructure of the M22N modified epoxy polymers, after [36]. (For the ball-stick
model in the schematic diagram, blue network denotes epoxy; green chain end denotes PMMA block; red chain component
denotes PbuA block.)
3.4. Tensile Properties
The mean values of the elastic modulus and tensile strength of the MAM modified epoxy polymers are
summarised in Table 1, and representative tensile engineering stress-strain curves are plotted in Figure
5. A modulus of 2.9 GPa and a tensile strength of 81 MPa were measured for the control epoxy. The
addition of MAM slightly reduces the modulus and tensile strength, but the presence of the MAM
allows more elongation relative to the control, indicating greater ductility of the material, see Figure 5.
The data indicate that the MAM has a lower modulus and strength than the epoxy. This is clearly
demonstrated for 10 or 12 wt% of M52 or M52N when there a co-continuous or phase inverted
microstructure with dimensions of hundreds of microns formed, as the measured modulus and tensile
strength drop dramatically (to about 2.3 GPa and 26 MPa respectively). There is no significant
difference between the values of the elastic modulus and the tensile strength of the three different
MAM modified epoxy polymers with low BCP contents.
90
Engineering stress, σt (MPa)
80
70
60
50
Control
40
7%M22N
7%M52N
30
7%M52
20
10
0
0
0.01
0.02
0.03
0.04
0.05
0.06
Engineering strain, ε
Figure 5. Tensile engineering stress versus strain response of unmodified epoxy polymer and MAM modified epoxy
polymers.
3.5. Compressive Properties
Mean values of the compressive modulus, Ec, compressive true yield stress/strain, σyc/εyc, and
compressive true fracture stress/strain, σfc/εfc for the control and MAM modified epoxy samples are
summarised in Table 2. The values of Ec obtained were slightly smaller than the E from the tensile tests
12
shown in previous section, due to the compliance correction and frictional effects in the plane strain
compression test [31]. Nevertheless, good agreement was found between the results of the plane strain
compression tests and the uniaxial tensile tests. The addition of MAM reduces the modulus, and this
effect is more pronounced in the M52 and M52N modified samples than for the nanostructured M22N
samples. The compressive yield stress is notably higher than the tensile yield stress due to the
constraint in the PSC test and the pressure dependence of yielding [37]. However, a tensile yield stress
could not be measured due to the brittleness of the epoxy polymers, see Figure 5. The compressive
yield stress data in Table 2 show that the addition of MAM reduces the yield stress, due to the relative
softness of the MAM. Note that the M22N modified epoxies tend to have a higher yield stress than the
corresponding M52N or M52 modified materials. This indicates that either there is better adhesion
between the MAM and the epoxy for the M22N than the other BCPs, or that the morphology of the
M22N modified samples leads to less reduction in the yield stress than the particulate structure seen for
M52N and M52. The first explanation is unlikely because analysis of the fracture surfaces indicates
very good adhesion between the M52N or M52 particles with the epoxy matrix, and thus it appears that
the small magnitude of the reduction is due to the nanostructured bicontinuous gyroid microstructure.
Representative compressive true stress-strain curves for control, 7 wt% M22N, 7 wt% M52N, and 7 wt%
M52, are shown in Figure 6. These curves clearly show three different stages of deformation. Firstly,
there is a relatively linear (elastic) region with a steep rise in stress at relatively small strains, until the
yield point is reached. With continued loading, strain softening is observed. Thereafter strain hardening
is observed until the specimens fracture.
In Figure 6, the degree of strain softening (i.e. the magnitude of the stress drop after yield) was reduced
by the addition of M22N relative to the control. This suggests that the M22N suppresses the
macroscopic inhomogeneous yielding and strain localisation in the epoxy matrix [38]. Further, based
on the preceding evidence, it can be inferred that shear band yielding will occur but may not be the
only toughening mechanism contributing to the toughness improvement for the MAM modified epoxy
samples in this study. A more significant reduction of strain softening is found for the 7 wt% M52N
and 7 wt% M52 samples. For these samples, only a very small amount of strain softening can be seen
in the compressive true stress-strain curves, see Figure 6, which suggests that the addition of M52N or
M52 gives more significant suppression of the macroscopic inhomogeneous yielding and strain
localisation in the epoxy matrix than the M22N. Moreover, the disappearance of strain softening of the
7 wt% M52N and 7 wt% M52 samples was consistent with the optical microscopy results, in which
shear banding was not observed in the plastic deformation zone ahead of the sub-critically loaded crack
tip for the 7 wt% M52N and 7 wt% M52 DN-4PB samples.
The presence of the MAM does not significantly reduce the fracture stress and strain for the M22N
modified material, see Table 1. However, the fracture stress is reduced for the M52 and M52N
modified epoxy polymers, although the fracture strain is relatively unchanged, indicating that less
strain hardening occurs. In addition, the fracture stress and strain are significantly reduced when phase
inversion occurs for the M52 and M52N modified epoxy polymers, which will be discussed in section
3.8.
13
Table 2. Compressive modulus, Ec, compressive true yield strength, σyc, compressive true yield strain, εyc, compressive true
fracture strength, σfc, compressive fracture strain, εfc, and fracture properties for the control and MAM modified epoxy
samples.
M22N
Ec (GPa)
σyc (MPa)
σfc (MPa)
εfc
KIc (MPam1/2)
GIC (J/m2)
Control
2.49 (±0.07)
107 (±0)
264 (±3)
0.90 (±0.02)
0.60 (±0.03)
102 (±8)
2% M22N
2.55 (±0.3)
106 (±0)
253 (±36)
0.87 (±0.02)
0.73 (±0.01)
162 (±12)
3% M22N
2.35 (±0.23)
106 (±0)
264 (±23)
0.88 (±0.00)
0.73 (±0.04)
182 (±2)
5% M22N
2.40 (±0.07)
104 (±1)
282 (±5)
0.91 (±0.01)
0.91 (±0.07)
245 (±38)
7% M22N
2.25 (±0.06)
104 (±1)
275 (±39)
0.86 (±0.02)
1.05 (±0.05)
340 (+16)
10% M22N
1.79 (±0.03)
101 (±1)
288 (±14)
0.90 (±0.03)
1.22 (±0.05)
407 {±32)
12% M22N
1.82 (±0.02)
99 (±1)
278 (±12)
0.95 (±0.01)
1.22 (±0.04)
450 (±19)
M52
Ec (GPa)
σyc (MPa)
σfc (MPa)
εfc
KIc (MPam1/2)
GIC (J/m2)
Control
2.49 (±0.07)
107 (±0)
264 (±3)
0.90 (±0.02)
0.60 (±0.03)
102 (±8)
2% M52
2.53 (±0.07)
105 (±0)
264 (±3)
0.90 (±0.02)
0.69 (±0.02)
136 (±3)
3% M52
2.74 (±0.07)
103 (±1)
221 (±13)
0.85 (±0.01)
0.79 (±0.03)
179 (±13)
5% M52
1.80 (±0.14)
98 (±1)
193 (±12)
0.84 (±0.00)
0.87 (±0.05)
234 (±33)
7% M52
1.60 (±0.07)
94 (±0)
236 (±5)
0.85 (±0.02)
0.92 (±0.06)
278 (±32)
10% M52
1.52 (±0.02)
96 (±0)
115 (±5)
0.69 (±0.02)
2.16 (±0.25)
1796 (±92)
12% M52
1.24 (±0.03)
87 (±3)
42 (±8)
0.50 (±0.04)
0.59 (±0.12)
109 (±3)
M52N
Ec (GPa)
σyc (MPa)
σfc (MPa)
εfc
KIc (MPam1/2)
GIC (J/m2)
Control
2.49 (±0.07)
107 (±0)
264 (±3)
0.90 (±0.02)
0.60 (±0.03)
102 (±8)
2% M52N
2.53 (±0.31)
105 (±0)
225 (±4)
0.86 (±0.02)
0.74 (±0.02)
165 (±12)
3% M52N
1.88 (±0.09)
103 (±1)
217 (±13)
0.85 (±0.01)
0.85 (±0.03)
233 (±24)
5% M52N
1.70 (±0.10)
99 (±0)
212 (±2)
0.83 (±0.04)
0.84 (±0.02)
218 (±2)
7% M52N
1.48(±0.11)
95(±1)
233(±22)
0.83(±0.07)
0.93 (±0.04)
303 (±18)
10% M52N
1.44 (±0.00)
94 (±1)
127 (±1)
0.73 (±0.01)
2.01 (±0.20)
1466 (±294)
12% M52N
1.17 (±0.09)
88 (±3)
44 (±6)
0.54 (±0.00)
0.57 (±0.02)
89 (±18)
14
Compressive true stress, σc (MPa)
140
120
Strain softening region
100
Strain hardening region
80
Compressive yield point, σyc
60
Control
7%M22N
40
7%M52N
20
7%M52
0
0
0.1
0.2
0.3
0.4
0.5
0.6
Compressive true strain, εc
Figure 6. Plane strain compression plots of true stress versus true strain for control, 7 wt% M22N, 7 wt% M52N, and 7 wt%
M52, at 21°C and a constant displacement rate of 0.1 mm/min.
3.6. Fracture Properties
The fracture toughness, KIc, and fracture energy, GIc, for the control and MAM modified epoxy
samples are summarised in Table 2. The graphs of the GIc versus the concentration of the MAM in the
MAM modified epoxy polymers, with the corresponding inclusion size (in terms of radius), are plotted
in Figures 7-9. The values of KIc and GIc for the control were measured as 0.60 MPam1/2 and 102 J/m2
respectively. These are in good agreement with the values reported by Johnsen et al [39] and Blackman
et al [40] for this epoxy polymer.
The addition of MAM gave a significant improvement in the toughness. For the M22N modified epoxy,
the values of KIc and GIc increased linearly with the increasing concentration of MAM, see Figure 7.
Maximum values of KIc = 1.22 MPam1/2 and GIc = 450 J/m2 were measured for the 12 wt% M22N
sample, which are 100% and 350% higher than the control respectively. For the M52N and M52
modified epoxies, the values increased linearly up to a content of 7 wt%, see Figures 8 and 9. For
example, adding 7 wt% of MAM, the maximum of KIc = 0.93 MPam1/2 and GIc = 303 J/m2 were
measured for the M52N modified samples. The data shown in Table 2 show that the toughness
enhancement of all the MAM modified samples is approximately equal, despite the M52N and M52
MAM copolymers may having a higher effective PbuA ratio.
The exception was for the partly phase inverted samples, where MAM rich domains and epoxy rich
domains were present. When a co-continuous microstructure was formed in the M52N and M52
modified materials, at 10 wt%, there was a large increase in the measured toughness, to KIc = 2.16
MPam1/2 and GIc = 1796 J/m2 for the M52 modified epoxy for example. The experimental scatter in the
measured values also increases due to the variability of the microstructure at the crack tip. At the
15
highest content used, of 12 wt%, the toughness dropped dramatically, to KIc = 0.59 MPam1/2 and GIc =
109 J/m2 for the M52 modified material. These values are approximately equal to those for the control.
This type of behaviour has been reported in the literature. Pascault and Williams [41] stated that “the
presence of this maximum seems to be related to systems in which incipient phase-inverted structures
exhibit lack of adhesion between both phases and a consequently low fracture toughness value. In these
cases, bicontinuous ... morphologies lead to a better fracture resistance”. Indeed, Yoon et al [42]
showed such a maximum toughness for an epoxy polymer modified using poly(ether sulfone) (PES)
with poor adhesion, but a monotonic increasing toughness for PES with good adhesion between the
phases. McGrail & Street [43] and Zheng et al [44] showed similar effects with other thermoplasticmodified thermoset polymers.
500
6-13nm worm-like micelles
450
6-13nm (width) worm-like
micelles
400
350
GIc (J/m2)
6-13nm (width) worm-like micelles
300
250
6-13nm (width) worm-like micelles
200
6-13nm (width) worm-like micelles
6-13nm (width) worm-like micelles
150
100
50
0
0
2
4
6
8
10
12
14
Concentration of M22N (wt%)
Figure 7. Fracture energy of M22N modified epoxy polymers versus the concentration of M22N.
16
2000
Co-continuous morphology
Particle radius in the epoxy
phase 0.06-0.28 µm; particle
radius in the MAM phase
0.44-1.23 µm (10 wt%)
1800
1600
GIc (J/m2)
1400
1200
1000
800
600
Particle radius
0.17-0.60 µm
(5wt%)
Particle radius
0.16-0.40 µm
(3wt%)
400
200
Particle radius
0.12-0.63 µm
(7wt%)
Complete phase
inversion
Particle radius
1.74-3.73 µm
(12wt%)
0
0
2
4
6
8
10
12
14
Concentration of M52N (wt%)
Figure 8. Fracture energy of M52N modified epoxy polymers versus the concentration of M52N.
2000
Co-continuous morphology
Particle radius in the epoxy phase
0.09-0.47 µm; particle radius in the
MAM phase 0.38-0.621 µm (10 wt%)
1800
1600
GIc (J/m2)
1400
1200
1000
800
600
Particle radius
0.05-0.40 µm
Particle radius
0.05-0.30µm (3wt%)
(2wt%)
400
Particle radius
0.08-0.50µm
(5wt%)
200
Particle radius 0.08-0.90 µm,
some particles have occlusion
like a doughnut (7wt%)
Complete phase
inversion
Particle radius 1.944.38 µm (12wt%)
0
0
2
4
6
8
10
12
14
Concentration of M52 (wt%)
Figure 9. Fracture energy of M52 modified epoxy polymers versus the concentration of M52.
3.7. Fractography
FEG-SEM micrographs of the fracture surface of the control and MAM modified epoxy samples, taken
from the deformation zone ahead of the crack tip, are shown in Figures 10-15. The fracture surface of
the control sample was relatively smooth and mirror-like, see Figure 10. The small scale river line
marks and step changes of the crack plane on this fracture surface are caused by the excess stored
17
elastic energy that was released by fast crack propagation [45]. This multi-planar nature of the fracture
surface is a way of absorbing excess energy in brittle thermosets [24]. These river lines and steps were
observed in all the samples examined in this study.
3.7.1. M22N Modified Samples
For the M22N modified epoxy samples, although the KIc and GIc values were increased significantly by
adding M22N to the epoxy polymer, the fracture surfaces of the M22N modified samples still had a
relatively brittle appearance in the low magnification micrographs. This may due to the small size and
nanostructuration of the M22N phase, as the toughening mechanisms occur on the nanometre scale.
The topographic features of the M22N modified samples could only be seen in high resolution FEGSEM micrographs, see Figure 11. The fracture surface of the 3 wt% M22N samples showed more
plastic deformation and a rougher surface, see Figure 11a, compared to the control epoxy. Moreover,
many small cavities with a size ranging from 7 to 16 nm in radius were also found well dispersed on
the fracture surface of these samples. In Figure 11b, the fracture surface of the 10 wt% M22N samples
was considerably rougher than that of the 3 wt% M22N samples, which suggests more plastic
deformation occurred. This is consistent with the higher KIc and GIc values of the 10 wt% M22N
samples. Cavities were also observed on the fracture surface of the 10 wt% M22N sample, but the size
and density of these cavities were almost unchanged compared to the 3 wt% M22N sample. It is worth
noting that these cavities were not an artefact of the coating process used prior to FEG-SEM analysis,
because they were not observed on other coated non-M22N modified samples and the presence of these
cavities was independent of the coating material used. Similar nano-cavities were found in other studies,
e.g. [24, 39], where nanometre-sized particles debonded from an epoxy matrix.
The cavities observed in the nanostructured epoxy samples with M22N suggest that cavitation may be
one of the toughening mechanisms that contribute to the increased toughness of these nanostructured
samples. Alternatively, these cavities may be caused by the ligaments of the bicontinuous gyroid
structure fracturing below the surface, and then pulling out. A similar pull-out effect is observed using
short fibres or carbon nanotubes, e.g. [46]. However, the independence of the size and the density of
the cavities on the nanostructured samples also implies that cavitation or pull-out will not be the chief
toughening mechanism responsible for the increased toughness of the M22N modified samples. The
uneven fracture surfaces of the M22N modified samples corroborate the earlier suggestion that the
M22N modified epoxy samples have a bicontinuous gyroid microstructure, see Figure 4. This
microstructure explains the relatively unchanged size and density of the cavities irrespective of the
concentration of the M22N. The reason is if the M22N modified epoxy samples have the bicontinuous
gyroid microstructure, then any cavitations probably only occur in the junction of the PbuA gyroids,
where the concentration of the PbuA rubber phase is relatively high. Hence, the size and density of the
cavities are relatively independent of the concentration of the M22N, as epoxy samples with higher
concentrations of M22N will only increase the whole density of the bicontinuous gyroid rubber phase
and have a comparatively smaller effect on the junction PbuA size.
Identifying the toughening mechanisms in such a bicontinuous (or co-continuous) structure is difficult.
However, the size and number of cavities is insufficient to provide such a large increase in the
18
measured toughness. Previous work using thermoplastic modification of thermoset polymers has
reported that in a co-continuous structure, deformation of the two phases occurs as the crack must pass
through both phases during fracture. The phases will experience relatively low constraint due to the
structure and the relative softness of the MAM phase. Thus they will be able to deform, absorbing
energy and hence increasing the toughness of the material.
3.7.2. M52N and M52 Modified Samples
The fracture surfaces of the microphase separated M52N and M52 modified samples are similar to
those of the traditional non-reactive thermoplastic modified epoxy polymers reported in the literature,
e.g. [42-44, 47-49]. Figures 12 and 13 show that the fracture surfaces of the 5 wt% M52N and 5 wt%
M52 samples are similar. The surfaces are very rough and multi-planar, with evenly dispersed micronsize cavities. These cavities show that cavitation of the MAM particles has occurred. The size of the
cavities was found to increase with the increasing concentration of M52N and M52. Comparing the
mean radius of the cavities with that of the particles observed using AFM shows that the MAM cavities
are larger than the original particle size. These findings show that plastic void growth followed the
cavitation process. Hence, based on the FEG-SEM observation, cavitation and subsequent void growth
may be the toughening mechanisms responsible for the enhancement in fracture toughness for M52N
and M52 modified epoxy samples up to and including the 7 wt% formulations.
At 7 wt% of M52 or M52N, occluded particles were observed using AFM, see Figure 3e. These
particles show a doughnut-like structure, with an epoxy inclusion within the MAM phase. These
occluded particles also showed cavitation and void growth, as shown in Figure 14 where two cavitated
occluded particles are visible. In this Figure, the lump within each particle is the epoxy inclusion. This
confirms that there is good adhesion between the epoxy and the MAM phases, and that the formation of
such occluded particles does not prevent cavitation.
In addition, from the high magnification insets of Figure 12 and 13, the inside surface of the dilated
cavities of the M52N and M52 shows that the dispersed MAM particles appear to have a bicontinuous
gyroid microstructure, which further corroborates the results of the AFM studies, i.e. the wormlike
micelle dispersed morphology is equivalent to a bicontinuous gyroid microstructure in 3D.
Phase inversion was observed for the epoxy samples with >7 wt% of M52N and M52. Figure 15 shows
the fracture surface of the partly phase inverted 10wt% M52N, on which both the MAM rich domain
and epoxy rich domains can be clearly seen. The microstructure is co-continuous on a macro-scale.
From the higher magnification inset of Figure 15, it can be seen that brittle epoxy particles were evenly
dispersed in the MAM rich domain, while well dispersed MAM particles were observed in the epoxy
rich domain. When complete phase inversion occurs, the crack will propagate through the relatively
low strength MAM matrix, and hence a low toughness is measured. Here the fracture surfaces show the
features typical of a phase-inverted structure, of deformation of the soft matrix phase with no effect on
the harder epoxy particles.
19
Crack direction
Crack direction
100 µm
Figure 10. FEG-SEM micrograph of fractured SENB specimen of the control, taken near the pre-crack arrest plastic
deformation zone.
Crack direction
200 nm
(a)
Crack direction
200 nm
(b)
Figure 11. FEG-SEM micrograph of fractured SENB specimen of (a) 3 wt% M22N, and (b) 10 wt% M22N taken near the
pre-crack arrest plastic deformation zone. White arrows point to some cavities.
20
200 nm
Crack direction
Crack direction
1 µm
Figure 12. FEG-SEM micrograph of fractured SENB specimen of 5 wt% M52N, taken near the pre-crack arrest plastic
deformation zone.
200 nm
Crack direction
1 µm
Figure 13. FEG-SEM micrograph of fractured SENB specimen of 5 wt% M52, taken near the pre-crack arrest plastic
deformation zone.
21
Crack direction
1 µm
Figure 14. FEG-SEM micrograph of fractured SENB specimen of 7 wt% M52N, taken near the pre-crack arrest plastic
deformation zone.
1 µm
Crack direction
100 µm
Figure 15. FEG-SEM micrograph of fractured SENB specimen of 10 wt% M52N, in which both the MAM rich domain and
epoxy rich domain are shown.
3.8. Sub-surface Analysis of the DN-4PB Specimens
Based on a knowledge of linear elastic fracture mechanics (LEFM), the fracture toughness of a material
is closely related to the corresponding plastic deformation zone in the immediate vicinity of the crack
tip of the material. Hence, it is important to study the plastic deformation zone ahead of the crack tip. In
this study, the investigation of the plastic deformation zone of the MAM modified epoxy samples was
conducted by examining the sub-critically loaded DN-4PB specimens using transmission optical
microscopy (TOM). The TOM images for a set of formulations of DN-4PB samples are shown in
Figure 16. It can be seen in Figure 16(a-b) that only a very small plastic deformation zone was
observed for the control sample using both transmitted and cross-polarised light. The small plastic
deformation zone for the control is consistent with the low measured toughness of the control samples.
22
For the 7 wt% M22N samples, a large feather-like damage zone was observed in front of the crack tip
using transmitted light, see Figure 16c. This feather-like damage zone has been reported previously
from other block copolymer modified epoxy polymers in the literature [6, 7]. It can be ascribed to the
presence of crack-branching and dilation bands. Shear band yielding was also observed in the featherlike damage zone under cross-polarised light, shown in Figure 16d. This observation shows that shear
banding must have occurred during fracture and contributed to the improvement in fracture toughness.
This agrees with the results of the plane strain compression tests, where strain softening was observed
after yield, which indicates that shear band yielding would be expected. Hence, crack-branching and
shear band yielding may be the toughening mechanisms which contribute to the toughness
improvement of the M22N modified epoxy polymers.
For the 7 wt% M52N samples and the 7 wt% M52 samples, large circular damage zones were observed
using transmitted light, see Figures 16e and 16g. The size of both damage zones was similar, which
agrees well with the similar measured values of fracture toughness for these samples. These dark
circular damage zones suggest that cavitation must have occurred in front of the crack tip to dissipate
the excess strain energy [50]. Further, little or no birefringence was observed in front of the crack tip
under cross-polarised light for both samples, see Figures 16f and 16h. This is consistent with the
disappearance of strain softening observed from the plane strain compression tests shown in the
previous section for the 7 wt% M52N and 7 wt% M52 samples. These findings confirm that cavitation
must be one of the toughening mechanisms responsible for the toughness improvement in these
microphase separated MAM modified epoxy polymers.
23
(a)
5 µm
(b)
0
5 µm
(c)
5 µm
(d)
5 µm
(e)
5 µm
(f)
5 µm
(g)
5 µm
(h)
5 µm
Figure 16. TOM images of the sub-critically loaded crack tip in the plane strain region of the DN-4PB specimens, (a-b)
control under transmitted light and cross polarised light respectively, (c-d) 7 wt% M22N under transmitted light and cross
polarised light respectively, (e-f) 7 wt% M52N under transmitted light and cross polarised light respectively, (g-h) 7 wt%
M52 under transmitted light and cross polarised light respectively.
24
3.9. Discussion
3.9.1. M22N Modified Epoxy Polymers
Based on the results from the present study, considerable knowledge about the microstructure, the
mechanical properties, and the toughening mechanisms of M22N modified epoxy polymers was gained.
From the results of the AFM and the FEG-SEM analysis, the microstructure of the M22N modified
epoxy polymers may be attributed to a three dimensional bicontinuous gyroid microstructure with a
nanostructured PbuA-PMMA (core/shell) network, see Figure 4. The glass transition temperature and
the mechanical properties of the epoxy polymer (Tg, E, Ec, σt, and σyc) were found only slightly reduced
compared to the control, which is consistent with other studies on block copolymer modified epoxies
[5-7, 10, 13]. Here it is suggested that the preservation of the Tg, E, Ec, σt, and σyc of the M22N
modified nanostructured epoxy polymers could be that the plasticisation effect of the PMMA blocks of
the M22N was weakened by the highly localised phase separation because of the steric confinement
from the covalent bond between the PMMA block and the PbuA block, and that the nanostructure
distributed the local PMMA plasticisation effect down to the nanoscale to further minimise the effect of
the phase separated PMMA.
The principal toughening mechanism responsible for the toughness improvement of the nanostructured
M22N modified epoxy polymers appears to be shear band yielding, and the increased matrix ductility
caused by the plasticisation effect of the addition of M22N. Although the addition of MAM suppresses
the strain softening after yield, and hence the macroscopic inhomogeneous yielding and strain
localisation in the epoxy polymer, there is still significant strain softening observed in the PSC tests.
TOM of the DN-4PB specimens using cross-polarised light showed a birefringent zone, indicating that
shear band yielding occurs. Other mechanisms caused the cavities seen on the fracture surfaces and the
crack branching at the crack tip. These mechanisms may involve local plastic deformation and
cavitation of the PbuA rich network junctions, or pull-out of fracture ligaments of the bicontinuous
gyroid network, plus causing the principal crack to divide into many secondary cracks. However, the
contribution of these mechanisms to the overall toughening effect will be small.
3.9.2. M52N and M52 Modified Epoxy Polymers
The microstructure, the mechanical properties, and the toughening mechanisms of the M52N and the
M52 modified epoxy polymers were also investigated. These MAM block copolymers were found to
phase separate in the epoxy matrix. Evenly dispersed micron-sized spherical particles, which have an
ordered nanostructure, were observed in the epoxy matrix from the AFM images, see Figure 3(e-f).
This microstructure is similar to the traditional thermoplastic modified structure reported in the
literature, as mentioned above. From the DMA results, the Tg of the M52N and M52 modified epoxy
polymer was found to be almost unchanged after the addition of M52N and M52. This suggests that the
relatively immiscible M52N and M52 fully phase separate during the polymerisation process of the
epoxy polymer. The values of E, Ec, σt, and σyc for the M52N and the M52 modified epoxy polymers
from the uniaxial tensile and the plane strain compression tests showed that the addition of the
relatively soft MAM reduced these mechanical properties. However, large decreases in the fracture
25
stress and strain to failure were only seen when a phase-inverted microstructure was formed using 12
wt% of M52 or M52N.
The addition of the M52N or the M52 was found to suppress strain softening, and hence
inhomogeneous yielding and strain localisation, in the plane strain compression results, see Figure 6.
This finding was confirmed the results of the plastic deformation zone examination using OM under
cross-polarised light after the DN-4PB tests, where no shear band yielding was observed. Hence, shear
band yielding is not a toughening mechanism for these materials. However, well dispersed MAM
cavities were observed on the fracture surfaces of the M52N and the M52 modified epoxy polymers.
Based on the subcritical plastic deformation zone study and the FEG-SEM study, the major toughening
mechanisms in these MAM modified epoxy polymers were attributed to cavitation of the MAM
particles and subsequent plastic void growth for epoxy samples containing up to 7 wt% of M52N or
M52. This void growth absorbs energy, hence increasing the toughness. Indeed, the toughening
increment due to this mechanism can be predicted using the work of Kinloch and co-workers [24, 39,
46]. This analysis shows that this mechanism can indeed give such a toughening effect.
At higher concentrations of M52N or M52, at 10 wt% and above, phase inversion occurs to give a cocontinuous or fully phase-inverted microstructure. This results in a sharp increase and subsequent
decrease in the measured toughness. Previous work using thermoplastic modification of thermoset
polymers has shown a similar effect of a maximum toughness when a co-continuous structure is
formed, as discussed above. However, there has been less discussion of the toughening mechanisms for
such a microstructure, as these are not readily observed. Based on the results of the current study, the
significant toughness enhancement provided by the co-continuous morphology comes from the
interconnected epoxy rich and MAM rich domains which form a hard and soft composite-like structure.
When the crack propagates through this structure, see Figure 17, the brittle epoxy phase fractures.
However, the MAM phase was observed to span across the fracture surfaces. The MAM can be
expected to possess a low yield stress and relatively high ductility. Hence as the crack opens, the MAM
will deform and absorb energy, thus increasing the measured toughness, before the bridging ligaments
fracture. Note that the epoxy matrix regions in the co-continuous structure also contain MAM particles
which will undergo cavitation and void growth, and these processes will also contribute to the
toughening effect. In addition, due to the interconnected structure of the two domains and the low yield
stress of the MAM, the plastic deformation zone at the crack tip may be greatly increased in size, with
a consequent increase in toughness.
When complete phase-inversion occurs, the low strength of the MAM leads to premature failure and a
low measured toughness. As the toughness is so much less when the structure is completely phaseinverted, the presence of the epoxy domains in the co-continuous structure must stabilise the material.
Although there is poor adhesion between the two domains, as discussed above, the interpenetrating
nature of the microstructure prevents the premature failure seen when phase-inverted.
26
100 µm
V notch
Epoxy rich domain
Interconnected epoxy and MAM domain
Epoxy rich domain
Crack tip in MAM domain
Crack propagation in the MAM domain was constrained
Figure 17. TOM image of the pre-cracked DN-4PB sample modified with 10 wt% M52N under transmitted light.
4. Conclusions
The microstructure, the mechanical properties, and the toughening mechanisms for an anhydride cured
DGEBA epoxy modified with three different types of MAM block copolymers were investigated. The
addition of M22N to the epoxy polymer was found to form a three dimensional bicontinuous gyroid
nanostructure at all contents used. The addition of ≤ 7 wt% of M52N or M52 gave dispersed micronsized spherical MAM particles in the epoxy matrix, but phase inversion begins with ≥ 7 wt% of these
block copolymers. The Tg, elastic modulus, and yield strength decreased slightly when MAM was
added, due to the relative softness of the MAM.
A significant increase in the fracture toughness, KIc, and fracture energy, GIc, was measured due to the
addition of M22, and maximum values of KIc = 1.22 MPam1/2 and GIc =450 J/m2 were measured for the
12 wt% M22N sample, which are 100% and 350% higher respectively than the control. For the
nanostructured M22N modified epoxy polymers, the principal toughening mechanism was identified to
be shear band yielding. For the microphase separated M52N and M52 modified epoxy polymers with ≤
7 wt% of BCP, the toughness enhancement were ascribed to cavitation of the particles and subsequent
void growth. When 10 wt% of M52N or M52 was added, a co-continuous morphology was formed,
and the principal toughening mechanism may be the bridging of the extensive interconnected MAM
domains. Beyond the addition of 10 wt% of M52N or M52, phase inversion was observed, and the
toughness was reduced to less than that of the control epoxy.
Acknowledgements
The authors would like to thank S. Sprenger (Nanoresins) and T. Fine (Arkema) for supplying
materials. Some of the equipment used in this work was provided by Dr. Taylor’s Royal Society
Mercer Junior Award for Innovation.
References
1.
Bagheri R, Marouf B, and Pearson R. Polymer Rev 2009;49(3):201-225.
27
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Petrie EM. Epoxy adhesive formulations, 1st ed. New York: McGraw-Hill, 2006 (chapter 8)
Zheng S. Nanostructured epoxies by the use of block copolymers. In: Pascault JP and Williams
RJJ, editors. Epoxy Polymers: New Materials and Innovations. Weinheim: WILEY-VCH
Verlag GmbH & Co. KGaA, 2010. pp. 79-108.
Larranaga M, Gabilondo N, Kortaberria G, Serrano E, Remiro P, and Riccardi C. Polymer
2005;46(18):7082-7093.
Hydro RM and Pearson RA. J Polym Sci, Part A: Polym Chem 2007;45(12):1470.
Liu J, Thompson ZJ, Sue HJ, Bates FS, Hillmyer MA, Dettloff M, Jacob G, Verghese N, and
Pham H. Macromolecules 2010;43(17):7238-7243.
Liu J, Sue H, Thompson Z, Bates F, Dettloff M, and Jacob G. Macromolecules
2008;41(20):7616-7624.
Kishi H, Kunimitsu Y, Imade J, Oshita S, Morishita Y, and Asada M. Polymer 2011;52(3):760768.
Könczöl L, Döll W, Buchholz U, and Mülhaupt R. J Appl Polym Sci 1994;54(6):815-826.
Bacigalupo LN and Pearson RA. On The Use of Triblock Copolymers to Toughen Epoxy
Resins. 32th Annual Meeting of the Adhesion Society, 2010.
Oldak RK, Hydro RM, and Pearson RA. On the use of triblock copolymers as toughening
agents for epoxies. 30th Annual Meeting of the Adhesion Society. Tampa, Florida, 2007. pp.
153-155
Pearson RA, Bacigalupo LN, Liang YL, Marouf BT, and Oldak RK. Plastic Zone Size Fracture Toughness Correlations in Rubber-Modified Epoxies. 31st Annual Meeting of the
Adhesion Society. Austin, Texas, 2008. pp. 27-29.
Barsotti R. Nanostrength Block Copolymers for Epoxy Toughening. 2008 Annual Meeting of
The Thermoset Formulator Association. IIIinois, USA, 2008.
Grubbs RB, Dean JM, Broz ME, and Bates FS. Macromolecules 2000;33(26):9522-9534.
Lipic PM, Bates FS, and Hillmyer MA. J Am Chem Soc 1998;120(35):8963-8970.
Dean JM. J Polym Sci, Part A: Polym Chem 2003;41(20):2444.
Ritzenthaler S, Court F, David L, Girard-Reydet E, Leibler L, and Pascault JP. Macromolecules
2002;35(16):6245-6254.
Ritzenthaler S, Court F, Girard-Reydet E, Leibler L, and Pascault JP. Macromolecules
2002;36(1):118-126.
Hillmyer MA, Lipic PM, Hajduk DA, Almdal K, and Bates, FS. J Am Chem Soc
1997;119(11):2749-2750.
Dean J, Grubbs R, Saad W, Cook R, and Bates F. J Polym Sci, Part B: Polym Phys
2003;41(20):2444-2456.
Thio Y, Wu J, and Bates F. J Polym Sci, Part B: Polym Phys 2009;47(11):1125-1129.
Pearson RA and Yee AF. J Mater Sci 1989;24(7):2571-2580.
Yee AF and Pearson RA. J Mater Sci 1986;21(7):2462-2474.
Hsieh TH, Kinloch AJ, Masania K, Sohn Lee J, Taylor AC, and Sprenger S. J Mater Sci
2010;45(5):1193-1210.
Masania K. Ph.D dissertation, London: Imperial College London, 2010 (chapter 4).
Nanostrength, Epoxy Application - Technical Datasheet of the Commercial Range. Colombes:
Arkema France, 2011. pp. 1-4.
Maiez-Tribut S, Pascault JP, Soulé ER, Borrajo J, and Williams RJJ. Macromolecules 2007;
40(4):1268-1273.
Ferry JD. Viscoelastic Properties of Polymers, 3rd ed. London: John Wiley & Sons, Inc., 1980.
pp. 224-263
28
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
BS-EN-ISO-1183-1 (2004) Plastics-Methods for determining the density of non-cellular
plastics, Part 1: Immersion method, liquid pyknometer methodand titration method. BSI,
London.
BS-EN-ISO-527-2 (1996) Plastics-Determination of tensile properties, Part 2: Test conditions
for moulding and extrusion plastics. BSI, London.
Williams JG and Ford H. J Mech Eng Sci 1964;6:405-417
BS-ISO-13586 (2000) Plastics-Determination of fracture toughness. Linear elastic fracture
mechanics (LEFM) approach. BSI, London.
Kinloch AJ. Adhesion and Adhesives: Science and Technology, 1st ed. London: Chapman &
Hall, 1987 (chapter 6).
Sue HJ and Yee AF. J Mater Sci 1993;28(11):2975.
Eaton P and West P. Atomic Force Microscopy. Oxford: OUP Oxford, 2010. pp. 99-100, 14851.
du Sart GG, Vukovic I, Vukovic Z, Polushkin E, Hiekkataipale P, Ruokolainen J, Loos K, and
ten Brinke G. Macromol Rapid Commun 2011;32(4):366-370.
Boyce MC and Haward RN. The post-yield deformation of glassy polymers. In: Haward RN
and Young RJ, editors. The Physics of Glassy Polymers. London: Chapman & Hall, 1997
(chapter 5).
Crist B. Yield processes in glassy polymers. In: Haward RN and Young RJ, editors. The
Physics of Glassy Polymers. London: Chapman & Hall, 1997 (chapter 4).
Johnsen BB, Kinloch AJ, Mohammed RD, Taylor AC, and Sprenger S. Polymer
2007;48(2):530-541.
Blackman B, Kinloch AJ, Sohn Lee J, Taylor AC, Agarwal R, Schueneman G, and Sprenger S.
J Mater Sci 2007;42(16):7049-7051.
Pascault JP and Williams RJJ. Formulation and characterization of thermoset-thermoplastic
Blends. In: Paul DR and Bucknall CB, editors. Polymer Blends: Formulation, vol. 1. New York:
John Wiley & Sons, Inc., 2000. pp. 379-415.
Yoon TH, Priddy DB, Lyle GD, and McGrath JE. Makromol Chem, Macromol Symp
1995;98(1):673-686.
McGrail PT and Street AC. Makromol Chem, Macromol Symp 1992;64(1):75-84.
Zheng S, Wang J, Guo Q, Wei J, and Li J. Polymer 1996;37(21):4667-4673.
Hull D. Fractography: observing, measuring and interpreting fracture structure topography.
Cambridge: Cambridge University Press, 1999.
Hsieh TH, Kinloch AJ, Taylor AC, and Kinloch IA. J Mater Sci 2011:46(17):5724-5735.
Yee AF, Du J, and Thouless MD. Toughening of epoxies. In: Paul DR and Bucknall CB, editors.
Polymer Blends: Performance, vol. 2. New York: John Wiley & Sons, Inc., 2000. pp. 225-267.
Bucknall CB and Gilbert AH. Polymer 1989;30(2):213-217.
Kim J and Robertson R. J Mater Sci 1992;27(1):161-174.
Kim DS, Cho K, Kim JK, and Park CE. Polym Eng Sci 1996;36(6):755-768.
29