Will - 2013

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Influence of Partial Cross-Linking Degree on Basic Physical Properties

of RTM6 Epoxy Resin


Judith Moosburger-Will, Michael Greisel, Markus G. R. Sause, Robert Horny, Siegfried Horn
Experimental Physics II, Institute of Physics, University of Augsburg, 86135 Augsburg, Germany
Correspondence to: J. Moosburger-Will (E- mail: judith.will@physik.uni-augsburg.de)
ABSTRACT: Thermo-physical and mechanical properties of partially and completely cross-linked RTM6 epoxy resin samples in the
glassy state have been investigated. A significant dependence of glass transition temperature, density, and modulus on the curing his-
tory and the curing degree is found. Density and modulus decrease with increasing curing degree and show a step-like irregularity in
the so called transition region, which is related to the transition from rubber to glassy state during cross-linking and the starting of
structural relaxation processes. The relationship between the thermo-physical and mechanical properties, which is important for the
development of new processing routes for fiber reinforced polymers, is addressed. VC
2013 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 130:
43384346, 2013
KEYWORDS: thermosets; crosslinking; thermal properties; mechanical properties
Received 10 April 2013; accepted 30 June 2013; Published online 17 July 2013
DOI: 10.1002/app.39722
INTRODUCTION
Tetra-functional epoxy resins are widely used in aerospace
industry for production of advanced composite materials, pri-
marily carbon fiber reinforced polymers (CFRP). Also in the
fields of transportation and construction a growing market with
numerous technical applications exists.
1,2
After infiltration of
the dry fiber fabric with the matrix polymer, the thermoset is
cross-linked in a thermal curing process. The reactivity of the
fully cross-linked epoxy resin is very low. Partial cross-linking
by modified curing cycles maintains a particular rest reactivity
of the resin and opens new options for CFRP production. For
example the joining of partially cured and fresh resin systems
can be realized. This allows an increased level of integral con-
struction of CFRP components, which is a highly relevant
aspect in aerospace industry.
3
A subsequent final curing cycle
guarantees the complete cross-linking of the new component.
The physical properties of a thermosetting resin in the cured
state depend on several factors, e.g. the chemical structure of
the polymer, the type and concentration of the hardener, and
the thermal cycle used for curing. For a given chemical compo-
sition in particular the curing parameters, namely the curing
temperatures, heating rates, and curing durations are important.
During the curing process a chemical reaction between epoxy
groups and reactive groups of the hardener takes place. Depend-
ing on the curing treatment the material will pass either from
liquid through rubber to glassy state, or from liquid to rubber
state or from liquid to glassy state.
4
Annealing of a glassy poly-
mer at temperatures below its glass transition temperature
results in physical ageing processes.
57
Such different curing and
annealing cycles will result in different materials properties of
the cured thermoset.
8
Thus, to consider variations of existing
curing cycles, it is essential to know the relation between basic
thermo-physical and mechanical properties, the curing history
and the cross-linking degree. In particular the interdependence
of glass transition temperature, curing degree, density, elastic
modulus, and water absorption of partially and fully cured
epoxy resins is of interest and subject to actual research activ-
ities.
813
However, because of the large variety of combinations
of resin, hardener and curing cycles, a general systematic assess-
ment of such interdependences is difficult.
In epoxy resins with functionality higher than two, a three-
dimensional network forms during curing. For such epoxy res-
ins a decrease of density and modulus with increasing cross-
linking of the polymer chains is observed in the glassy state
(above the gelation point), caused by the reduced possibility of
molecular rearrangements.
10,14
Also di-functional epoxy resins,
cured with hardeners of functionality higher than two form
three-dimensional networks. Here, a decrease of density and
modulus as a function of curing degree is also reported for
some systems,
912,15
in one case with a step-like irregularity.
6
For other di-functional network polymers an increase in density
and modulus is observed.
14
Linear systems, with a functionality
of epoxy and hardener of two, are characterized by an increase
of density with curing degree.
10
Generally, relaxation processes
VC
2013 Wiley Periodicals, Inc.
4338 J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 WILEYONLINELIBRARY.COM/APP
because of physical ageing of the cured epoxy resin result in
density and modulus increases.
57
Also, a low density of the
cured polymer is correlated to high water absorption.
12,13
In this work, the material properties of partially cross-linked
tetra-functional epoxy resin HexFlow VR
RTM6 samples are
investigated, which are produced by isothermal curing cycles at
reduced temperatures. Isothermal curing was used to obtain
comparable chemical reaction processes for all curing degrees.
12
Depending on the curing degree the chosen curing treatment
involves a liquidrubber or a liquidrubberglass transition dur-
ing curing. The kind of transition has a high impact on the
physical properties of the resulting partially cross-linked epoxy
resin samples in their glassy state and promises high potential
for different fields of applications. Thermo-physical and
mechanical properties of partially and completely cross-linked
RTM6 resin samples were investigated as a function of curing
degree focusing on glass transition temperature, density and
elastic modulus. To determine the elastic modulus the microme-
chanical method of nanoindentation is used and is compared to
results of macroscopic mechanical tests.
EXPERIMENTAL
Production of Partially and Fully Cross-Linked
RTM6 Epoxy Resin Samples
All resin samples analyzed in this work are manufactured by
the mono-component epoxy resin system HexFlow VR
RTM6
distributed by the Hexcel Corporation.
16
It is composed of the
tetra-functional epoxy resin tetraglycidyl methylene dianiline
(see Figure 1) and the hardeners 4,4-methylenebis(2,6-diethy-
laniline) and 4,4-methylenebis(2-isopropyl-6-methylaniline)
(see Figure 2).
Modified curing cycles are used to obtain resin samples with
defined curing degrees. The heating and cooling rates always
were set to 2

C/min. All cycles start with an isothermal curing


period at a temperature of 120

C. This so called infiltration


cycle originates from the CFRP production process, where a cer-
tain time for resin infiltration of the dry fiber fabric is needed.
Afterwards the resin is cured for 2.5 hours at the standard heat-
ing temperature of 180

C to create a completely cross-linked


sample with a curing degree of nearly 100%. To obtain samples
with partial curing degrees between 55% and 80% the infiltra-
tion cycle is followed by an isothermal heat treatment at a
reduced temperature of 135

C. The different curing degrees


result from different durations of these isothermal heating peri-
ods. In Figure 3 the heating cycle for production of partially
cross-linked resin samples is shown schematically.
For final curing, the partially cross-linked samples are submitted
to a second curing cycle at 180

C for 2.5 hours. This second


curing step transforms the partially cross-linked resin, e.g. with
a curing degree of 60%, to a fully cross-linked resin further
nominated as 60/100. The other curing degrees are named
equivalently. The complete partial/final heating cycle with com-
bination of partial and final curing is shown schematically in
Figure 4.
Curing Degree, Glass Transition Temperature and Density
A 204F1 Phoenix differential scanning calorimeter (Netzsch)
was used to measure the heat flow during curing as a function
of temperature. To determine the degree of cure of the epoxy
resin, the residual heat flow of the cured resin DH
R
and the
total heat flow of the fresh, uncured resin DH
T
are needed. Both
are measured by an analyzing cycle, which comprises a heating
ramp from room temperature up to 260

C with a heating rate


of 5

C/min. The heat flow is determined from the resulting


DSC curves by integration of the exothermal reaction peak. The
curing degree a is determined by
a512
DH
R
DH
T
The glass transition temperatures T
g
of the partially and fully
cross-linked epoxy resin samples are determined from the
Figure 1. Chemical structure of epoxy resin tetraglycidyl methylene
dianiline.
Figure 2. Chemical structure of hardeners 4,4-methylenebis(2,6-diethyla-
niline) and 4,4-methylenebis(2-isopropyl-6-methylaniline).
Figure 3. Schematic heating cycle for preparation of partially cross-linked
resin samples. [Color figure can be viewed in the online issue, which is
available at wileyonlinelibrary.com.]
Figure 4. Schematic partial/final heating cycle with combination of partial
and final curing for production of finally cross-linked resin samples.
[Color figure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
ARTICLE
WWW.MATERIALSVIEWS.COM WILEYONLINELIBRARY.COM/APP J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 4339
modulated DSC curves as the midpoint of the endothermic shift
of the reversing heat flow, when the resin passes from the glassy
to the rubbery state. The heating rate was 5 K/min, the modula-
tion amplitude 1 K and the modulation period 60 sec.
Density of the partially and completely cross-linked polymer
samples is measured with a helium pycnometer AccuPyc II 1340
(Micromeritics).
For each curing degree at least three resin samples were
analyzed.
Micro- and Macromechanical Testing
Micromechanical elastic moduli of the partially and fully cross-
linked epoxy samples were measured by nanoindentation. It
allows a laterally resolved modulus analysis with a resolution in
the micrometer range. During the loadingunloading indenta-
tion cycles the loaddisplacement curves F(h) are recorded.
With the assumption of elasticplastic material behavior and
negligible viscoelastic effects, the micromechanical parameters
can be determined from the mainly elastic unloading curve, as
described in the following.
1719
The influence of viscoelastic
material behavior will be discussed afterwards.
The reduced modulus is dependent on the contact area A
between indenter and sample and the contact stiffness S and is
given by
E
r
5

p
p
2b
S

A
p
The contact area A is a function of the contact depth h
c
, which
is defined as
h
c
5h
max
2e
F
S
h
max
is the maximal indenter displacement and S 5dF/dh the
contact stiffness at the beginning of unloading. S is determined
by fitting the unloading curve by the power law F5a h2h
p

m
,
where a and m are fitting constants and h
p
is the final depths of
the indent after complete unloading.
18
For the Berkovich
indenter tip used the geometry constant e is 0.75 and
A 24:5 h
2
c
The shape correction factor b for a Berkovich indenter is
unity.
18
With the knowledge of the Poissons ratio ration of the
sample t and the indenter t
i
and the modulus of the indenter
E
i
, the modulus of the sample E can be determined by
1
E
r
5
12t
2

E
1
12t
2
i

E
i
As polymers often are characterized by viscoelastic material
behavior, its influence on nanoindentation has to be considered.
The delayed viscoelastic material response results in a distortion
of the loaddisplacement curve.
17,19,20
In particular, because of
the gradually increasing deformation after loading and the grad-
ually decreasing deformation after unloading, the unloading
curve of a viscoelastic material is more convex than that of elas-
tic materials. This influences the resulting contact stiffness and
accordingly the determination of elastic modulus. The influence
of viscoelastic effects on the nanoindentation results can be
reduced by the insertion of a dwell time between loading and
unloading, by slow loading rates and by fast unloading
rates.
17,19
To check the impact of viscoelastic effects on the
results presented in this work, nanoindentation measurements
were performed with different loading rates between 2 and 10
mN/sec and with different dwell times between 0 and 20 sec.
Low loading rates and a dwell time of more than 2 sec ensure
undistorted, reproducible loaddisplacement curves, and reliable
modulus values.
The nanoindentation measurements were performed with a
NanoTest 600 nanoindenter (Micromaterials) and an indenter
in Berkovich geometry. The load-controlled mode was used
with loading and unloading rates of 2 mN/sec. In all measure-
ments a dwell time of 5 sec was inserted between loading and
unloading. Before analysis, the raw data of the loaddisplace-
ment measurement are corrected by the system compliance. To
determine the contact stiffness S, the unloading curve was fitted
with a power law between 0.8F
max
and F
max
. A Poissons ratio
of 0.35 is used.
21
Average modulus values were obtained by
investigating several comparable resin regions. All average values
are calculated from more than nine individual values. The
measurements were performed on plane surfaces of the polymer
samples, which were prepared by a grinding and polishing pro-
cess. Approximately 50 mm of the surface material was ablated
by this process.
During indentation compressive loading is applied to the sam-
ple surface. Therefore, the measured moduli should be corre-
lated with (macroscopically determined) compressive moduli.
For materials with pronounced tensile-compressive anisotropy
caution is recommended. For those cases comparing the moduli
measured by nanoindentation to macroscopically determined
tensile or flexural moduli might be error-prone.
The macroscopic mechanical testing was done with a Kappa 50
DS universal testing machine (Zwick). Tensile-compressive tests
in the elastic range were performed on dog bone-shaped resin
samples with 1 mm/min test speed. Tensile and compressive
moduli were determined for at least four specimens of each cur-
ing degree. The literature value for the macroscopic tensile
modulus of the completely cross-linked resin system is 2.89
GPa.
16,21
To our knowledge no literature data for the macro-
scopic compressive modulus exists. For comparable epoxy resin
systems different behavior of the compressive moduli is
reported. For some resin systems the compressive modulus
exceeds the tensile modulus by up to 50 %,
22,23
for others the
compressive and tensile moduli are comparable.
24
EXPERIMENTAL RESULTS AND DISCUSSION
In the following the thermo-physical and mechanical investiga-
tion of partially and fully cross-linked RTM6 epoxy resin sam-
ples is presented. First we describe the dependence of glass
transition temperature, density, and micro- and macromechani-
cal moduli on the curing degree. Then the influence of a sec-
ond, final curing cycle on the resin properties is investigated.
Dependence of Curing Degree on Isothermal Curing
Treatment
By isothermal heating cycles samples with defined partial curing
degrees between 55% and 95% are produced. Isothermal curing
ARTICLE
4340 J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 WILEYONLINELIBRARY.COM/APP
ensures comparable chemical reaction processes for all curing
degrees. The curing behavior is determined by an interrelation
of physical resin properties, e.g. the viscosity, and chemical reac-
tion kinetics, e.g. the progress of curing.
12
In Figure 5 the cur-
ing degree as a function of the isothermal heating time is
displayed for two curing temperatures, namely the reduced
heating temperature of 135

C and the standard heating temper-


ature of 180

C.
The curing degree is defined by the duration of isothermal
heating. With increasing heating time the curing degree
increases monotonically and for long heating periods saturates
at a defined maximum curing degree. At the standard heating
temperature of 180

C curing degrees of maximal 95% can be


reached, which is in agreement with the literature.
25
Up to a
curing time of 30 minutes a increases rapidly with an average
rate of approximately 1.1% per minute, at longer heating times
the curing degree approaches slowly to 95%. Heating at a
reduced temperature of 135

C allows a maximal curing degree


of only 81%. Up to a heating time of 150 minutes a increases
monotonically with an average rate of approximately 0.5% per
minute, thus by a factor of two slower than at 180

C. At
higher heating times the curing degree approaches slowly to
the maximum value of 81%. The lower reaction rate for the
reduced heating temperature of 135

C allows a good control of


the curing reaction. A defined adjustment of partial curing
degrees is possible.
Glass Transition Temperature and its Impact on the Resins
Physical State
Generally, the glass transition temperature of thermosetting
polymers is directly correlated to the curing degree, as a highly
cross-linked network needs more energy to allow large-scale
molecular motions. In Figure 6 the glass transition temperatures
of the partially cross-linked epoxy resin samples cured isother-
mally at 135

C are shown as function of the curing degree.


A nonlinear increase of the glass transition temperature is
observed, which can be described by the empirical Di Benedetto
equation. The fitting parameter k is 0.4848, which is in good
agreement with values reported for comparable epoxy resin sys-
tems by Refs. 26 and 27. The correspondent fit line is included
in Figure 6. The glass transition temperature data agree well
with the reported measurements on tetra-functional epoxy resin
during curing at reduced temperatures.
28
For a curing degree of
76.5%62.5% the glass transition temperature crosses the cur-
ing temperature, T
c
, of 135

C. This has an important impact on


the physical properties of the resin and its curing behavior.
4,5
The gelation point is reached at curing degrees between 40%
and 50%, as can be extrapolated from gel times and curing tem-
peratures given in literature.
16
Thus all partially cured samples
discussed in this work have transformed from liquid to rubber
state during curing.
As long as T
c
>T
g
, namely for samples with a smaller than
76.5%, the samples are still in their rubber state at the end of
the isothermal heating procedure. The curing reaction is mainly
chemically controlled, in particular for curing degrees below
60%.
29
When cooling down to room temperature, at T5T
g
the
transition to the glassy state with a freezing of chain motions
takes place.
For samples with a higher than 76.5%, T
c
<T
g
applies at the
end of isothermal curing and the samples are in their glassy
state. Because of this vitrification during cure the reaction now
is mostly diffusion controlled and the polymerization rate slows
down or essentially stops, limiting the curing degree to 81%
and the glass transition temperature to 165

C. Additionally,
because of annealing at T
c
<T
g
, physical ageing of the nonequi-
librium glassy material with structural relaxation processes
starts. The molecular rearrangements result in a gradual
decrease of the specific enthalpy of the system, which can be
attributed to a decrease of free volume and/or a decrease in
molecular conformational energy with time and influence the
resins physical properties.
57
When cooling to room tempera-
ture, the sample stays in its glassy state and the structural rear-
rangements remain.
Figure 5. Dependence of curing degree on isothermal heating time for the
curing temperatures 135

C and 180

C.
Figure 6. Glass transition temperatures in dependence of curing degree
for a heating temperature of 135

C together with di Benedetto fit line.


The transition region, where the glass transition temperature crosses the
heating temperature, is marked.
ARTICLE
WWW.MATERIALSVIEWS.COM WILEYONLINELIBRARY.COM/APP J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 4341
In the following, for the partially cross-linked samples the
region of 76.5%62.5% curing degree will be called transition
region. The region with curing degrees higher than 76.5% will
be nominated as ageing region. In Figure 7 the described physi-
cal states of the resin system and the different regions of curing
degrees are visualized schematically.
The influence of the different physical states on the properties
of the epoxy resin, in particular in the transition and ageing
regions, is of high importance for the development of new cur-
ing and processing concepts and will be analyzed in the
following.
Fully cured samples with a curing degree of about 95%, which
are heated at a temperature of 180

C, reach a glass transition


temperature of 209

C. As T
c
<T
g
at the end of cure, these sam-
ples are in the glassy state and ageing processes start. Because of
the vitrification a is limited to 95%.
Density and Modulus of Partially Cross-Linked Resin
Density changes during cross-linking of glassy thermosetting
network polymers are caused by different, partly opposing
effects. On the one hand, the increase of the covalent bond
fraction with increasing cross-linking causes volume shrinkage
and densification of the polymer. On the other hand the more
and more restricted packing of polymer chains because of net-
work formation results in a decrease of density with conver-
sion.
10,11
Besides, physical ageing during curing at T
c
<T
g
causes a volume decrease and thus a densification of the glassy
material because of structural relaxation processes.
57
Also an
emission of volatile specimen can occur, which causes a mass
loss of the sample.
9
Additionally, also the cooling process from
the end-of-curing state (rubbery or glassy) to the room tem-
perature state (glassy) influences the room temperature density
of the glassy thermoset. If the cooling process involves a long
cooling period within the rubbery state, an efficient chain
packing and a high density will result. In contrast, a long cool-
ing period within the glassy state allows only minimal molecu-
lar rearrangements and thus causes an inefficient chain
packing and a low density.
11
Dependent on the chemical struc-
ture of polymer and hardener and the processing conditions,
the individual contributions of the described effects determine
the final density of the glassy thermoset as a function of the
curing degree.
The density is directly correlated to the molecular structure of
the polymer, e.g. the crosslinking density and the molecular
weight. Therefore it is also closely connected to the modulus of
the resin sample, which is determined by the molecular struc-
ture, intermolecular forces, intramolecular forces, and relaxation
processes. For epoxy systems a similar behavior of modulus and
density is reported.
9
Figure 7. Schematic illustration of different physical states of resin system in dependence of curing degree and temperature.
Figure 8. Density of the partially and fully cross-linked polymer samples
(T
c
5135

C and 180

C, respectively) in dependence of the curing degree,


together with linear fit lines.
ARTICLE
4342 J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 WILEYONLINELIBRARY.COM/APP
In Figures 8 and 9 the densities and moduli, respectively, of par-
tially and fully cross-linked epoxy resin samples are shown as a
function of the curing degree.
To determine the moduli of the epoxy resin samples microme-
chanical testing by nanoindentation and macromechanical ten-
sile and compressive testing was performed. The high lateral
resolution of nanoindentation can be important for interface
investigations, e.g. of joined partially cross-linked resins sam-
ples, and therefore, represents a promising micromechanical
method. A direct comparison of all mechanical results is per-
formed, to evaluate the reliability of micromechanical testing
for polymer samples.
For curing degrees below 76.5% a linear decrease of density and
micro- and macromechanical moduli with curing degree is
observed, as visualized by the linear fit lines in Figures 8 and 9.
For low curing degrees a rubbery, three-dimensional network
with relatively low crosslinking density and low glass transition
temperature develops. During cooling to room temperature,
rapid molecular rearrangements of not cross-linked molecules
occur as long as T>T
g
, and result in relatively low specific vol-
umes and high density and modulus values. With increasing
curing degree an increase of crosslinking and glass transition
temperature takes place. The growing three-dimensional net-
work and the increase of T
g
reduces the possibility of molecular
arrangements in the rubber state during cooling, increases the
specific volume and thus decreases the density and modulus, in
agreement with literature.
8,1012,14,15
The described effects domi-
nate over the opposing trend of volume shrinkage.
In the transition region an irregularity with a step to higher
density and modulus values is found. Here the transition to the
glassy state takes place during the curing process once T
c
crosses
T
g
. After the crossing, T
c
<T
g
applies and annealing processes
start. They allow structural relaxation processes and molecular
rearrangements and cause the observed step-like increase of
density and modulus.
57,10
For curing degrees above the transition region again a linear
decrease of density and modulus is observed with increasing
curing degree. Here different, partly opposing effects influence
the values. Curing in the glassy state induces further slowly pro-
ceeding three-dimensional cross-linking. This is the dominating
effect and causes a decrease of density and modulus. Smaller
effects are produced by volume shrinkage and structural relaxa-
tion processes because of physical ageing, which result in a den-
sification of the material and an increase of modulus.
5,7,9
Considering the absolute density values, the literature value for
a fully cured RTM6 sample is 1.14 g/cm
3
.
16
This is slightly
higher than the measured value of 1.1360 60.00129 g/cm
3
, pos-
sibly because of minor differences in the curing cycle.
Overall a density decrease of 1.01% is observed with increasing
curing degree in the range between 55% and 95% curing
degree. The result is in agreement with literature values, where
for other three-dimensional epoxy resin systems also a decrease
of density with increasing curing degree is described for a com-
parable range of curing degrees in the glassy state. The density
decreases reported are between 0.10% and 0.41% and are
smaller than our value, but of the same order of magnitude.
912
An irregularity in the density values is described in Ref. 6, how-
ever, at much higher curing degrees and within the glassy state.
Focusing on the absolute values of micro- and macromechanical
moduli, some points should be noted. First, the agreement of
macroscopic tensile and compressive moduli is very good. No
tensile-compressive anisotropy is observed for the RTM6 resin
system. Consequently, a comparison of moduli from nanoin-
dentation to both, tensile and compressive moduli is possible.
Comparison to literature data shows a very good agreement of
the macroscopic tensile modulus of the fully cross-linked resin
sample (95% curing degree) of 2.893 60.0563 GPa to the litera-
ture value of 2.890 GPa.
16,21
Secondly, macromechanical testing shows that the modulus
decreases by 22.9% with increasing curing degree in the range
between 55% and 95%. Micromechanical testing exhibits a
decrease of modulus of 15.9% in the same range of curing
degrees. These results are in qualitative agreement with litera-
ture, where also a decrease of modulus with curing degree is
described.
8,11,14,15
The reported values lie between 9.3% and
41.8%.
Finally, the agreement of the absolute values of micro- and
macromechanical moduli differs for the two regimes of curing
degrees. For curing degrees below the transition region, the
absolute modulus values of micro- and macromechanical testing
are in very good agreement. In contrast, the step height in the
transition region and the modulus values in the ageing regime
differ clearly. The macromechanical moduli show an absolute
step height of 0.10 GPa, which is by a factor of three smaller
than that of the micromechanical moduli of 0.29 GPa. Conse-
quently, for high curing degrees the absolute micromechanical
moduli are clearly higher than the macromechanical moduli.
They differ by about 0.25 GPa. One possible reason for the dif-
ferences between micro- and macromechanical moduli in the
ageing regime above the transition region could be an enhanced
Figure 9. Micro- and macromechanical modulus of partially and fully
cross-linked polymer samples (T
c
5135

C and 180

C, respectively) in
dependence of curing degree, together with linear fit lines.
ARTICLE
WWW.MATERIALSVIEWS.COM WILEYONLINELIBRARY.COM/APP J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 4343
surface ageing of the polymers, as described in literature.
30,31
However, this description, which would result in an increased
surface modulus and surface density, cannot explain our experi-
mental results as will be explained in the following. As nanoin-
dentation probes the surface properties, the micromechanical
moduli would be affected by such surface degradation. In con-
trast, bulk properties as macromechanical modulus and density
should only be marginally influenced by an aged surface layer
and should show a smaller step in the transition region and
smaller values in the ageing regime. However, a similar relative
step height (defined as ratio between the step height in the
transition region and the total decrease between 55% and
76.5% curing degree) of micromechanical modulus (55.1%) and
density (50.0%) is observed. It is clearly higher than that of the
macromechanical moduli of 13.5%. Thus, the differences
between micro- and macromechanical testing cannot be caused
by surface ageing effects. The good agreement between density
and micromechanical modulus suggests that micromechanical
testing by nanoindentation is more sensitive to the molecular
arrangement of the polymer, in particular to the molecular
relaxations because of annealing, than macromechanical testing.
The mechanical testing conditions seem to allow an emphasis of
different modulus contributions, i.e. the molecular structure or
molecular forces.
Nanoindentation of polymer surfaces thus represents a very val-
uable tool for modulus and density investigation of polymers
with high lateral resolution that expands the possibilities of
macroscopic mechanical testing.
The irregularity of density and modulus observed in the transi-
tion region is a particular feature of our partially cross-linked
samples, which to our knowledge is not described in literature
so far. It is caused by the specific isothermal curing cycles,
which allow a transition from rubber to glassy state during cur-
ing. This transition influences the properties of the partially and
fully cross-linked thermosetting samples even in the glassy state
after cooling down to room temperature and represents a
promising effect for defined adjustment of physical resin prop-
erties. The relatively high moduli above the step within the
transition region are typical for much lower cross-linking
degrees and could be advantageous during further curing treat-
ments or processing.
Investigation of Partially/Finally Cured Resin Samples
All partially cross-linked resin samples are submitted to a sec-
ond, final curing cycle at 180

C. In the CFRP production pro-


cess this final heating cycle is responsible for the complete
curing of partially cross-linked resin samples. The physical
properties of the partially/finally cured samples are analyzed, to
check the influence of the partial curing cycle on the final resin
properties. In Figures 10 and 11 the curing degrees and the glass
transition temperatures of the partially/finally cured samples
(called 58/100, 62/100, 71/100, and 79/100) and the fully/finally
cured sample (called 94/100) in comparison to that of the cor-
responding partially (called 58, 62, 71, and 79) and fully (94)
cured resin samples are shown, respectively.
All partially/finally cured resin samples achieve very similar cur-
ing degrees between 93% and 94% and very similar glass transi-
tion temperatures between 209.2

C and 212.4

C. As the glass
transition temperature of all partially cured samples lies below
180

C, these samples transform to rubber state during final cur-


ing. This allows further cross-linking and causes the observed
strong increases of curing degrees and glass transition tempera-
tures. All partially/finally cured resin samples thus seem to be
equivalent in terms of curing degree and glass transition tem-
perature, independent from their partial curing degree. Besides,
they exhibit values similar to that of the fully cross-linked 94-
sample (a 594.3%, T
g
5209.0

C). This is an important point


for application of partially/finally cured epoxy resins, as it guar-
antees conformance of thermo-physical properties of partially/
finally cured samples with conventionally fully cured resin
samples.
The fully cross-linked 94-sample has a glass transition tempera-
ture above the final curing temperature of 180

C. Therefore this
Figure 10. Curing degrees of partially and fully cured resin samples and
corresponding partially/finally and fully/finally cured samples.
Figure 11. Glass transition temperatures of partially and fully cured resin
samples and corresponding partially/finally and fully/finally cured
samples.
ARTICLE
4344 J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 WILEYONLINELIBRARY.COM/APP
sample stays in its glassy state during final curing and only little
increase of curing degree and glass transition temperature is
possible. The resulting 94/100-sample exhibits a curing degree
of 96% and a glass transition temperature of 217

C, which are
slightly higher than the values obtained by combined partially
and finally curing.
Figure 12 shows the micromechanical moduli of the partially/
finally and fully/finally cured resin samples, in comparison with
the corresponding values of the partially or fully cured samples.
For all partially cured resin samples a decrease of modulus with
final curing treatment at 180

C is observed. As described, heat-


ing at T
c
>T
g
results in further cross-linking and the corre-
spondent reduction of modulus.
8
The strongest effect is
detected for the 58/100-sample. With its low partial curing
degree it exhibits the strongest increase in curing degree and
thus the strongest built up of three-dimensional network. The
smallest change of modulus is observed for the 71/100-sample,
which has a relatively small modulus already in its partially
cured state. The micromechanical moduli of all partially/finally
cured samples (58/100, 62/100, 71/100, 79/100) range between
3.11 and 3.22 GPa, and are very similar to that of the fully
cured 94-sample of 3.16 GPa. The combination of partial and
final curing cycle thus creates elastic properties of the epoxy
resin that are comparable to that after conventional full curing,
which is in agreement with literature data.
11
This is of high
relevance for application of partially cross-linked epoxy resins,
as the combined curing treatment guarantees mechanical prop-
erties similar to that of a conventional single curing step. Only
small mechanical mismatch is expected for samples with differ-
ent heating histories, e.g. for joining of partially/finally and fully
cured (94) resin regions.
The step-like increase of modulus, observed for the partially
cross-linked samples, does not exist in the partially/finally cured
state. No irregularity of modulus is observed in the transition
region. The second heat treatment removes the distinctions cre-
ated by the different ageing conditions of the partial curing
cycles. All finally cured samples are in their glassy state at the
end of final curing and thus are aged by similar heat
treatments.
For the fully cross-linked sample (94) an increase of modulus is
observed after the final heat treatment (94/100). The modulus
increases from 3.16 to 3.41 GPa. Because of the high glass tran-
sition temperature of the 94-sample of 209.0

C, final curing
takes place at T
c
<T
g
. This results in an annealing process with
structural relaxations and the corresponding increase in modu-
lus of the 94/100-sample. A not negligible mechanical mismatch
of moduli of 0.25 GPa between the 94 and 94/100 samples is
observed. A combination of fresh resin with completely cured
resin and following final curing, e.g. for joining, thus is not
favorable. Also CFRP processing steps, which involve a thermal
impact on a conventional fully cross-linked resin sample, e.g.
the application of functional layers, will cause a change of
mechanical properties in the heated resin regions. The resulting
embrittlement of the surface regions can weaken the CFRP
component.
SUMMARY
Partial cross-linking of tetra-functional thermosetting RTM6
epoxy resin retains particular chemical reactivity of the polymer
and marks a promising way for new curing and processing
routes for CFRP production, e.g. the development of energy
efficient curing cycles, joining of CFRP components or applica-
tion of functional layers. A second final curing step guarantees
the complete cross-linking of the resin.
Knowledge of the interrelation between basic thermo-physical
and mechanical properties of the partially and finally cross-
linked resins is necessary to develop specifically adapted curing
cycles for different applications. With this objective the investi-
gation of curing degree, glass transition temperature, density,
and micro- and macromechanical modulus of partially and
finally cured RTM6 epoxy resin samples in the glassy state was
performed. The results show a significant dependence of the
physical properties on the curing history and the curing degree.
Density and modulus show on the one hand a decrease with
increasing curing degree that is caused by the cross-linking reac-
tion of epoxy resin and hardener and the built-up of a three-
dimensional network. On the other hand an irregularity in the
so called transition region between 74% and 79% curing degree
is observed, which can be correlated to the transition from rub-
ber to glassy state during the thermal treatment. The resulting
ageing in the glassy state is connected to structural relaxation
processes, which enhance the density and modulus of the poly-
mer. The step-like irregularity of physical properties is a typical
feature of samples manufactured with our isothermal, low tem-
perature curing cycle and demonstrates the options of adapted
curing treatments for defined adjustment of physical properties.
The relatively high modulus and density above the step is typi-
cal for resins with much lower curing degrees and offers new
possibilities, e.g. for joining of partially cross-linked and fresh
resin parts. After final curing of the partially cross-linked resins,
the resulting curing degree, glass transition temperature and
Figure 12. Micromechanical moduli of partially and fully cured resin sam-
ples and corresponding partially/finally and fully/finally cured resin
samples.
ARTICLE
WWW.MATERIALSVIEWS.COM WILEYONLINELIBRARY.COM/APP J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 4345
modulus are similar to that of a conventional fully cured sample
and only negligible mismatch of the thermo-physical and
mechanical properties exists.
The investigations demonstrate the impact of the thermal his-
tory on the basic physical properties of tetra-functional epoxy
resins systems and the wide field of potential applications of
partially cross-linked epoxy resin systems.
ACKNOWLEDGMENTS
Thanks to G unter Obermeier and Andreas Monden for valuable
scientific contributions.
REFERENCES
1. May, C. A. Epoxy Resins: Chemistry and Technology, 2nd
ed.; M. Dekker: New York, 1988.
2. Mohan, P. Polym. Plast. Technol. Eng. 2013, 52, 107.
3. Sause, M.; Llopard Prieto, L.; Scholler, J.; Horn, S.;
Moosburger-Will, J.; Horny, R. Patent DE 10 2011 010 384
A1.
4. Yoon, S. S.; Yu, W. J.; Kim, H. C. J. Mater. Sci. Lett. 1992,
11, 1392.
5. Montserrat, S. J. Appl. Polym. Sci. 1992, 44, 545.
6. Wang, X.; Gillham, J. K. J. Appl. Polym. Sci. 1993, 47, 447.
7. Odegard, G. M.; Bandyopadhyay, A. J. Polym. Sci. Part B:
Polym. Phys. 2011, 49, 1695.
8. Ellis, B.; Found, M. S.; Bell, J. R. J. Appl. Polym. Sci. 2001,
82, 1265.
9. Venditti, R. A.; Gillham, J. K. J. Appl. Polym. Sci. 1995, 56,
1687.
10. Wang, X.; Foltz, V. J. Macromolecules. 2001, 34, 681.
11. Detwiller, A. T.; Lesser, A. J. J. Appl. Polym. Sci. 2010, 117,
1021.
12. Enns, J. B.; Gillham, J. K. J. Appl. Polym. Sci. 1983, 28, 2831.
13. Wang, X.; Gillham, J. K. J. Appl. Polym. Sci. 1993, 47, 425.
14. Morel, E.; Bellenger, V.; Bocquet, M.; Verdu, J. J. Mater. Sci.
1989, 24, 69.
15. White, S. R.; Mather, P. T.; Smith, M. J. Polym. Eng. Sci.
2002, 42, 52.
16. Hexcel Company. Product literature to HexFlow RTM6VR
.
17. Menck, J.; He, L. H.; Nemecek, J. Polym. Test. 2011, Bd. 30,
101.
18. Oliver, W. C.; Pharr, G. M. J. Mater. Res. 2004, 19, 3.
19. Briscoe, B. J.; Fiori, L.; Pelillo, E. J. Phys. D Appl. Phys.
1998, 31, 2395.
20. Cheng, Y.; Cheng, C. Appl. Phys. Lett. 2005, 87, 111914.
21. Brauner, C.; Block, T. B.; Purol, H.; Herrmann, A. S. J.
Compos. Mater. 2012, 46, 993.
22. Behzadi, S.; Jones, F. R. J. Macromol. Sci. Part B: Phys. 2005,
44, 993.
23. Ciba Speciality Chemicals. Product literature to Araldite MY
721.
24. Kozey, V. V.; Kumar, S. J. Mater. Res. 1994, 9, 2717.
25. Navabpour, P.; Nesbitt, A.; Degamber, B.; Fernando, G.;
Mann, T.; Day, R. J. Appl. Polym. Sci. 2006, 99, 1658.
26. Karkanas, P. I.; Partridge, I. K. J. Appl. Polym. Sci. 2000, 77,
2178.
27. Sadeghinia, M.; Jansen, K. M. B.; Ernst, L. J. Int. J. Adhes.
Adhes. 2012, 32, 82.
28. Cizmecioglu, M.; Gupta, A.; Fedors, R. F. J. Appl. Polym.
Sci. 1986, 32, 6177.
29. Karkanas, P. I.; Partridge, I. K. J. Appl. Polym. Sci. 2000, 77,
1419.
30. Monney, L., Bole, J.; Dubois, C.; Chambaudet, A. Polym.
Degrad. Stab. 1999, 66, 17.
31. Moosburger-Will, J.; Jager, J.; Horn, S.; Wellhausen, C.
Polym. Test. 2012, 31, 1008.
ARTICLE
4346 J. APPL. POLYM. SCI. 2013, DOI: 10.1002/APP.39722 WILEYONLINELIBRARY.COM/APP

You might also like