Curing kinetics, rheological and IR study of imide oligomers

Progress in Reaction Kinetics and Mechanism, 2016, 41(1), 1–13

doi:10.3184/146867816X14544154610684 Paper: 1400346

Non-isothermal curing kinetics, chemorheological

behaviour, and IR spectral study of two trifunctional
phenylethynyl-terminated imide oligomers compared with
the corresponding bifunctional structure
Changwei Liua, Chunyan Qua, Chengyang Wangb, Dianxue Caoc, Dezhi Wanga,
Wanbao Xiaoa, Hao Fenga, Ping Liua and Xuefeng Baia*
aInstitute of Petrochemistry, Heilongjiang Academy of Sciences, Harbin, 150040; bThe 49th Research

Institute, China Electronics Technology Group Corporation, Harbin 150001; and cCollege of Material
Science and Chemical Engineering, Harbin Engineering University, Harbin 150001, P.R. China
*E-mail: bxuefeng@163.Net

ABSTRACT
Two trifunctional phenylethynyl-terminated imide oligomers, m-TPEPA and p-TPEPA, were
systematically compared with the corresponding bifunctional phenylethynyl-terminated imide
oligomer, BPEPA. The non-isothermal curing kinetics, Master Plots method, and rheological
behaviour of the three high performance oligomers were systematically studied by dynamic
DSC, small amplitude oscillatory shear rheometry, and infrared spectroscopy. The results
show that the activation energy (Ea) depended on the extent of conversion evaluated with three
different methods, and the lower Ea values of the trifunctional oligomers, which may lead to
different curing reactions, compared with the sustained growth of Ea values of BPEPA. The
chemorheological properties of the oligomers were measured and fit numerically with the
dual Arrhenius model and gel model. Moreover, the degree of cure (agel) at the gel time (tgel)
was calculated by the value of glass transform temperature at different curing temperature
combined with rheological and isothermal DSC results. According to structural changes
during the curing reaction characterised using FTIR spectra, it was inferred that the structures
of the cured trifunctional imide resins have cis-configurations, which may change the part of
the configuration from cis to trimerisation with the lower activation energies found in the later
part of the conversion, while the difunctional imide structure is trans.

KEYWORDS: curing kinetics, resin matrix, polymides, rheological behaviours

1. INTRODUCTION
Polyimides are a class of thermally stable high performance polymers for a wide range of
applications, e.g. matrix for high-performance advanced composite materials, adhesives, and
membranes for gas separation [1–4]. Compared to the normal thermosetting polyimides,
phenylethynyl-terminated imide oligomers (PETIs) have recently received considerable
attention due to their unique thermal curing process to generate a network via chain extension
and chain crosslinking [5–9]. The initial work on phenylethynyl-terminated imide oligomers
(PETIs) and their cured polymers was performed in the early 1980s at The University of
Akron [10]. Based upon NASA research, a phenylethynyl-terminated polyimide designated
as LaRC PETI-5, with a calculated molecular weight of 5000, was developed as a matrix for
an autoclave process to produce high impact-resistance carbon fibre-reinforced composites

1
2 Changwei Liu et al.

[11,12]. Subsequently, phenylethynyl groups have been placed terminal on the ends of molecules,
pendent along the backbone, and both pendent and terminal on the same molecules [13–16].
Many investigators have utilised a variety of analytical methods to examine the curing process
of phenylethynyl-terminated imide polymers: for example, Cho and Drzal investigated
the chemical changes occurring during reaction of LaRC™ PETI-5 through observation of
temperature- and time-dependent intensity variations of the characteristic FTIR absorption
bands [17]. The effect of peak temperatures on cure behaviour and thermal stability of the
material using differential scanning calorimetry (DSC) and Dynamic Mechanical Analysis
(DMA) were also reported [18]. Solid-state 13C NMR was also used to analyse the chemical
structure of cured acetylene-terminated polyimides or acetylene-containing model compounds
[19–21]. Besides that, Takekoshi and Terry [22] studied thermal reactions of two PEPA model
compounds by using HPLC and field-desorption mass spectroscopy.
Among many studies, the cure process of these end-capped polyimides has been widely
investigated. Pickard et al. [23] have investigated the kinetics and mechanism of bulk thermal
polymerisation of 3-phenoxyphenyl acetylene. The authors revealed that the soluble material
consists mostly of polyene having a trans-cisoidal configuration. In sharp contrast to the
studies of Pickard on the simple 3-phenoxyphenylacetylene compound, Fang et al. [7] studied
the mechanism of thermal cure of PETI-5 oligomer by thermal analysis using the modified
DiBenedetto equation, which suggested that complicated reaction mechanisms such as
intramolecular and bimolecular double bond addition reactions are involved to form a highly
crosslinked polyimide. These reaction kinetics suggest that only a simple ethynyl-to-ethynyl
addition reaction has occurred.
Recently, much effort has been made to explore the possibility to reduce PETI’s high cure
temperature in order to maximise the application of the high performance resins. Wright et al.
[24,25] investigated substituent effects of the ethynyl groups on the thermal curing kinetics,
and revealed that modified end-capping reagents with naphthylethynyl and anthracenylethynyl
groups could effectively lower the cure temperature. McGrath et al. [26] reported that electron-
withdrawing substituents on the backbone side of the ethynyl unit could accelerate thermal curing
rates. Yang [27] noted that PENA-m based on 4-(2-phenylethynyl)-1,8-naphthalic anhydride
showed a cure temperature of 50 °C lower than PEPA-m, based on 4-phenylethynylphthalic
anhydride. In addition, we have recently reported two new trifunctional phenylethynyl-terminated
imide oligomers, m-TPEPA and p-TPEPA[28]. Their curing kinetic parameters were calculated
using DSC and compared to the difunctional imide oligomer BPEPA, which has the same
backbone in an attempt to elucidate the influence of various structural aspects of phenylethynyl-
terminated oligomers on the thermal cure mechanism and to develop faster curing materials.
These data suggested the importance of stereoelectronic effects on reducing cure temperatures.
However, few studies have been reported concerning the mechanism of the rapid curing reaction
and the difference for the possible main structures formed during the curing process.
Based on our previous research, in this study, the non-isothermal curing kinetics, Master
Plots method, and rheological behaviour of the three high performance oligomers, m-TPEPA,
p-TPEPA, and BPEPA, were studied systematically by dynamic DSC, small amplitude
oscillatory shear rheometry, and IR spectroscopy. The dependence of the activation energy
(Ea) on the extent of conversion was evaluated with the Friedman method, the Flynn–Wall–
Ozawa method, and the Tang method. The chemorheological properties of the oligomers were
measured and numerically fitted to the dual Arrhenius model and gel model. Moreover, the

www.prkm.co.uk
 Curing kinetics, rheological and IR study of imide oligomers 3
degree of cure (agel) at the gel time (tgel) was calculated from the value of the glass transition
temperature at different curing temperatures combining with rheological and isothermal DSC
data.
2. EXPERIMENTAL
2.1 Materials
Two trifunctional phenylethynyl-terminated imide oligomers (m-TPEPA and p-TPEPA) and
a bifunctional phenylethynyl-terminated imide oligomer (BPEPA) were synthesised in our
laboratory according to reference [28]. Figure 1 represents the chemical structures of the three
kinds of phenylethynyl-terminated imide oligomers, which were stored in a freezer prior to use
and warmed to ambient temperature for only a couple of hours immediately prior to use.
2.2 Differential scanning calorimetric (DSC) measurements
Dynamic DSC measurements were carried out by using a TA Instruments Q-100 thermal
analyser to investigate the effects of cure temperature and time. A dry nitrogen flow was used
as purge gas at a flow rate of 50 mL min-1. The calorimeter was calibrated for temperature
and enthalpy using high-purity zinc and indium standards. Samples with a mass of about
5 mg in size were enclosed in aluminum pans. dynamic experiments were performed with
temperatures ranging from 50 to 450 °C at different heating rates OF 2.5, 5, 10 and 20 °C min-1.
The polymerisation process was observed, and the effects on the enthalpy by the change of
reaction rates were investigated.

2.3 Rheological behaviour measurements

Dynamic rheological measurements were performed using circular (25 mm diameter)
aluminium parallel plates within a TA Instruments AR 2000 Rheometer. Samples were
prepared by pressing about 1.5 g resin powder into 1.3 mm thick disks. A temperature ramp
measurement was performed at 5 °C min-1 from 200 to 350 °C in a nitrogen gas purged
environment. The oscillation frequency was 1 rad s-1 with a variable shear strain. The isothermal
rheology measurements were run in the range of 280–330 °C. Dynamic oscillation at 1 rad s-1

Figure 1 Chemical structures of phenyl-

ethynyl-terminated imide oligomers (BPEPA,
m-TPEPA and p-TPEPA).

www.prkm.co.uk
4 Changwei Liu et al.

began as soon as the specimen was installed. Reported data was not recorded until the fixture
temperature recovered to the testing temperature. The crossover of G’ and G” in a single
frequency experiment was defined as the gel point.

2.4 FTIR analysis

About 0.2 g samples were cured in small glass vials in a preheated furnace. After curing at the
desired temperature (370 °C) for 1 h, the samples were cooled promptly and removed from
the vials for characterisation. FTIR analysis was used to monitor the change of related groups
before and after curing using a Bruker Vector22 spectrometer.

2.5 Theoretical approach

2.5.1 Methodological aspects of non-isothermal kinetics
An understanding of kinetic parameters, such as the kinetic model, activation energy and the
frequency factor, can be obtained by various approaches. The integral approach is generally
believed to be more convenient, reliable, and accurate than the differential method. It is usually
assumed that the basic kinetic equation for solid-state curing reactions under non-isothermal
conditions can be expressed as a function of the fractional conversion α (0 <α<1) in the
following form:

da E
= A0 exp(− α ) f (α ) (1)
dt RT
where α is the extent of conversion, t is the time, T is the temperature, A is the pre-exponential
factor, E is the activation energy, R is the gas constant, and f(α) is the differential reaction kinetic
model, in which A, E, f(α) are called the kinetic triplet of a reaction. A non-isothermal solid-
state reaction with a programmed linear heating rate ß is considered. As in all isoconversional
methods, it is assumed that the reaction model is independent of the heating rate, but both E and
A are dependent on the conversion α.
2.5.2 Friedman method [29]
This method is a differential isoconversional method and is directly based on Eqn (1) whose
logarithm is:

dα β da E
ln( ) = ln( ) ln( Af (α )) − (2)
dt dT RT
From this equation, it is easy to obtain values for E over a wide range of conversions by plotting
ln(ßda/dT) against 1/T for a constant a value.

2.5.3 Flynn–Wall–Ozawa method [30,31]

This method is derived from the integral isoconversional method. Using Doyle’s approximation
[32] for the integral which allows ln p ( E / RT ) ≅ −5.331–1.052E/RT
ln β = ln[ AE / Rg (α )] − 5.331 − 1.052 E / RT (3)
Thus, for a as a constant, the plot ln ß versus 1/T, obtained from thermograms recorded at
several heating rates, should be a straight line whose slope can be used to evaluate the apparent
activation energy.

www.prkm.co.uk
 Curing kinetics, rheological and IR study of imide oligomers 5
2.5.3 Tang method [33]
As Vyazovkin’s [34] suggestion of a modified integral non-linear isoconversional procedure
(MNLINT), this method also proposes an integral method that uses local heating rates to
evaluate the activation energy dependence on the extent of conversion.

β AE E
ln[ ] = ln[ ] + 3.63504095 − 1.89466100 ln E − 1.00145033 (4)
T 1.89466100
g (α ) R RT
From this equation, the values of E can be obtained by plotting ln(ß/T1.89466100) against 1/T for
a constant a value.
2.5.4 Master plots method [35]
If the value of the apparent activation energy is known, the kinetic model of the process can be
determined by the Master Plots method based on the integral form of the kinetic data.
The kinetic rate equation at infinite temperature is obtained by introducing the generalised time
q defined as [36]

t  −E 
θ = ∫ exp   dt (5)
0
 RT 
where q denotes the reaction time taken to attain a particular a at infinite temperature.
Differentiation of Eqn (5), combining with Eqn (1), and the integrated form of the kinetic rate
equation, leads to
α dα θ
g (α ) = ∫ = A∫ dθ = Aθ (6)
0 f (α ) 0

From the integral kinetic equation at infinite temperature in integral form, Eqn (6), we can
obtain the following equation using a reference point at a=0.5 [37]:

g (α ) θ p( x)
= = (7)
g (0.5) θ 0.5 p( x0.5 )
where x =E/RT. Several convergent series exist for the approximation of the function p(x). For
this purpose, the modified Starink empirical kinetic models can be applied, with a comparison
of both sides of Eqn (7) telling us which kinetic model describes an experimental reactive
process [38]. Table 1 indicates the algebraic expressions of g(a) for the kinetic models used.

P(u ) = exp(− Au + B) / u k (8)

From the model, the constant values in the range of 0.2≤a≤0.8 are as follows: A= 1.0008, B=
–0.312, K= 1.92.
2.6 Viscosity models
Several investigators, including Mussatti and Macosko [39,40], proposed semi-empirical
models to correlate viscosity data during cure. The models relating viscosity with reaction time
had the form.

η (t ) = η0 exp(kt ) (9)

www.prkm.co.uk
6 Changwei Liu et al.

where η0 is the initial melt viscosity and k is the reaction rate constant. Roller [41] extended this
model by assuming an Arrhenius temperature dependence for η0 and k.
Eη
η0 = η∞ exp( )
RT (10)
and
Ek
k = k∞ exp( ) (11)
RT
where Eη is the viscous activation energy, Ek is the kinetic activation energy, η∞ is the fully
developed viscosity, k∞ is the kinetic rate constant, R is the gas constant, and T is the isothermal
cure temperature. Therefore, the dual Arrhenius viscosity model, considering the cure
temperature and time, can be modified as follows:
Eη − Ek
ln η (T , t ) = − ln η∞ + + tk∞ exp() (12)
RT RT
At high isothermal temperature, the initial lower melt viscosity of the low-molecular-weight
oligomer may have some deviations. An empirical approach proposed that the values of Eg and
tgel∞ may also be determined via analysis of an Arrhenius plot of ln(tgel) versus 1/T [42]. The
data in this plot can be calculated from Eqn (13).

∆Eg
∞
t gel = t gel exp( ) (13)
RT

3. RESULTS AND DISCUSSION

3.1 Characteristics of the oligomers
The high glass transform temperatures and excellent thermal oxidative stabilities of m-TPEPA
and p-TPEPA in the ranges of 456–500 °C, 71–69%, respectively, which were higher than
for the normal phenylethynyl-terminated imide oligomers [28,43], resulted from the higher
crosslink densities suggested by Argyropoulos and Bolker [44], shows that the introduction of
a branched structure could improve Tg and thermal stability without sacrificing too much melt
processability of the phenylethynyl-terminated oligomer. Moreover, based on the kinetic analysis
by isothermal and ramp temperature DSC and calculation of crosslink densities, the stable
aromatic ring structures formed in the trifunctional oligomers were more thermooxidatively
stable than the main polyene structures present in monofunctional or bifuntional phenylethynyl-
terminated model compounds. Consequently, in order to investigate the different curing
mechanism for the three oligomers in depth and optimize the rapid cure cycle, the change of
rheological behaviours and the non-isothermal curing kinetics of the resin, investigated by the
dynamic rheological measurements, ramp temperature DSC measurements, and FT-IR spectra,
must be fully understood.
3.2 Differential scanning calorimetry measurements
3.2.1 Non-isothermal curing kinetics
An accurate curing kinetics model helps to predict the cure behaviour of the resin for process
design and cure process optimization. DSC is a prevalent analytical technique for studying
crosslinking reactions of thermosetting polymers, primarily due to its high accuracy and

www.prkm.co.uk
 Curing kinetics, rheological and IR study of imide oligomers 7
easy sample preparation. A ramp temperature DSC study consists of performing a kinetic
analysis which includes curing curves obtained at different heating rates for BPEPA
(a), m-TPEPA (b), and p-TPEPA (c) in order to deduce the dependence of the kinetic
parameters on the degree of conversion (Figure 2). The DSC curves show that, in general,
the wide exothermic peaks, attributed to the thermal curing reactions of the trifunctional
phenylethynyl oligomers, m-TPEPA and p-TPEPA, were obviously lower than that of the
bifunctional oligomer BPEPA in the range of 2.5 to 20 °C min-1. Moreover, it can be noted
that the exothermal peak temperature of the curing reaction showed a large shift to lower
temperature on decreasing heating rate, which is caused by the more adequate crosslinking
reaction time. The exothermal peak temperatures of m-TPEPA were on average 20 °C
lower than those of BPEPA, which may due to a change in the curing mechanism.
It is well known that many solid-state reactions show a dependence of the activation energy
on the degree of conversion. The occurrence of such a dependence could be interpreted in
terms of a complex reaction mechanism. In a conversion range from 2 to 98%, Figure 3
shows the dependence of the activation energy evaluated on the degree of conversion by
means of the Friedman, Flynn–Wall–Ozawa, and Tang methods for the non-isothermal
data.
For m-TPEPA and p-TPEPA, the activation energy does not vary significantly with the
extent of conversion. In this case, the Friedman, Flynn–Wall–Ozawa, and Tang methods
lead to an almost identical Ea dependence in the range of 0.2< a <0.8, which means the
heat of polymerisation of the phenylethynyl groups was not significantly affected by the
substituent effect. For BPEPA, from Figure 3, it is clear that Ea presents a sustained increase
over the whole region. This dependence of Ea on α is an indication of a different reaction
with the trifunctional oligomer. Otherwise, the dependence of the Ea values of BPEPA on
a yielded by the Tang and FWO methods which are based on the integration of the rate
equation for constant Eα , is higher than that estimated by the Friedman method [45].
At the beginning of the curing reaction, the lower Ea values of the two trifunctional
oligomers may cause the relatively higher level of conversion rate compared with the values
of the bifunctional oligomer, which was consistent with the ramp DSC data and confirmed
that the trifuntional end-capped oligomers raised the activity of the reaction. In the later
part of the conversion, the lower Ea values of the trifunctional oligomers, especially
p-TPEPA, may lead to different curing reactions, compared with the sustained growth of
the Ea values of BPEPA. The theoretically predicted values for cyclotrimerisation based on
bond energy calculations for the thermal polymerisation of the acetylenic group and the
heats of polymerisation for the phenyl and vinyl-substituted acetylenic monomers were 189,
338 and 288 kJ mol-1, respectively, suggesting that trimerisation of the acetylenic groups
might be the preferred reaction pathway [46]. Because of the lower steric factor restrictions
of the main chains and the benzene ring next to the ethynyl group, the structure in the
bifunctional phenylethynyl-terminated model compounds was replaced by simple ethynyl-
to-ethynyl addition reactions to form linear and crosslinked polyene structures. However,
topological and steric factors restrictions by the branched structure may result in a higher
cis-form content (Scheme 1), and continued with a major curing reaction of trimerisation
or tetramerisation to form substituted aromatic ring structures with the lower activation
energies occurring in the later part of the conversion.

www.prkm.co.uk
8 Changwei Liu et al.

Figure 2 DSC scans of the cure exotherms Figure 3 Dependence of Ea on a for non-
of phenylethynyl-terminated imide oligomers isothermal data evaluated by different
at different heating rates. iso-conversional procedures: Friedman
Method (■),Flynn–Wall–Ozawa method (◆)
and Tang method (○) for (a) BPEPA,
(b) m-TPEPA and (c) p-TPEPA.
3.3 Master Plots method
In order to find the kinetic model of the curing reaction, the Master Plots method based on
the integral form of the kinetic data was chosen as it involves the curing reaction mechanism.
The modified Starink empirical kinetic models can be applied, and the expressions of the
associated functions g(a) are shown in Table 1. The master curve plots g(0.5)/g(a) versus a
for different mechanisms according to the Master Plots method for the three phenylethynyl-
terminated model compounds are illustrated in Figure 4. It is clearly seen that, irrespective
of the temperature profiles of the kinetic data, the experimental master plots for BPEPA and

www.prkm.co.uk
 Curing kinetics, rheological and IR study of imide oligomers 9

trans

cis A

Scheme 1 Proposed curing reactions of phenylethynyl-terminated imide oligomers.

Table 1 Algebraic expressions for g(a) functions for the commonest reaction mechanisms in solid-state reactions
No. Mechanism model Integral g(a) function
1 Avram-Erofe’ev, m=4 (A4) [–ln(1–a)]1/4
2 Avram-Erofe’ev, m=3 (A3) [–ln(1–a)]1/3
3 Avram-Erofe’ev, m=2 (A2) [–ln(1–a)]1/2
4 Avram-Erofe’ev, m=1.5 (A1.5) [–ln(1–a)] 2/3
5 Phase boundary reaction, n=1 (R1) a
6 Phase boundary reaction, n=2 (R2) 1– (1– a)1/2
7 Phase boundary reaction, n=3 (R3) 1– (1– a)1/3
8 One-dimensional diffusion (D1) a2
9 Two-dimensional diffusion (D2) a+(1–a)ln(1–a)
10 Ginstling-brounshtein (D4) 1– (2a/3) – (1–a) 2/3
11 Jander’s type diffusion (D3) [1– (1–a)1/3] 2
12 Power law n=1/4 (P4) a1/4
13 Power law n=1/3 (P3) a1/3
14 Power law n=1/2 (P2) a1/2
15 Power law n=3/2 (P0.5) a3/2
16 First-order (F1) –ln(1–a)
17 Second-order (F2) (1–a) -1–1
18 Third-order (F3) [(1–a) -2–1]/2

p-TPEPA are in good agreement with the theoretical master plot corresponding to the Avram-
Erofe’ev (n=1.5) mechanism. Meanwhile, a comparison of the experimental master plots
revealed that the kinetics of m-TPEPA were best describable by the phase boundary reaction
(n=2 or n=3) models. This showed that the trifunctional oligomer, especially the trifunctional
imide oligomer with a meta position (m-TPEPA), has a different curing mechanism from the
normal bifunctional or monofunctional phenylethynyl end-capped oligomers in the range of
0.2< a <0.8, although it has similar structures and functional groups. Because, with knowledge
of the value of Ea, the experimental master plot in a differential form can be constructed from the
experimental kinetic data under any temperature profile, it is very useful to check the constancy

www.prkm.co.uk
10 Changwei Liu et al.

P
α

Figure 4 Theoretical master curves Figure 5 Experimental and calculated isothermal

in integral form representing g(a)/g(0.5) as a viscosity curves of m-TPEPA at different
function of a for the different kinetic models isothermal temperatures.
describing curing reactions of BPEPA (○),
m-TPEPA (●) and p-TPEPA (▲).

Table 2 Parameters in viscosity models Eqns (12) and (13) for phenylethynyl-terminated imide oligomers
Resin lnη∞ (Pa.s) Eη (kJ mol-1) lnk∞ (s-1) Ek (kJ mol-1) ln tgel∞ (s) Eg (kJ mol-1) tgel(s) a agelb
BPEPA 10.2 35.76 22.9 145.22 –34.5 208.28 2176 0.25
m-TPEPA 9.7 47.29 19.3 120.95 –27.5 164.09 351 0.19
p-TPEPA 11.7 55.02 22.5 140.30 –32.6 193.75 970 0.20
atgel was detected during isothermal tests at 320 °C.
ba
gel was calculated by the value of glass transition temperature with isothermal DSC results.

and/or variation of the reaction mechanism of the solid-state reactions under different types of
temperature conditions.

3.4 Chemorheological
Based on our latest research and theoretical analysis, isothermal rheology measurements were
run at the following temperatures: 280, 290, 300, 310, 320, and 330 °C. Using these isothermal
viscosity data and the dual Arrhenius method, the natural logarithm may be taken of both
relationships to provide a linear relationship between lnη0 and 1/T, as the relationship between
lnk and 1/T. From these plots, the corresponding viscosity model Eqn (12) could be established,
and the constants and activation energies in the model are listed in Table 2. The calculated and
experimental viscosity curves of m-TPEPA at different isothermal temperatures are depicted in
Figure 5, and excellent agreement between experimental and predicted values are obtained in
the beginning at all measured temperatures. Discrepancies between the model and experimental
results at later stages of cure were primarily due to the difficulty of measuring experimentally
the viscosities of the resin at the gel point. The same phenomenona were also observed with
p-TPEPA and BPEPA.
In order to evaluate the gelation behaviours of the oligomers, according to the Hsich method,
the values of Eg and tgel∞, which were determined via analysis of an Arrhenius plot of ln(tgel)
versus 1/T (Figure 6), are also listed in Table 2. These values may be used to predict the time to
reach the gel point at any isothermal temperature within the experimental temperature range.
The difference between the values of Eg and Ek may be caused by the different kinetic equations.
But the values of both Eg and Ek for p-TPEPA and m-TPEPA were lower than that of BPEPA,

www.prkm.co.uk
 Curing kinetics, rheological and IR study of imide oligomers 11

Figure 6 Arrhenius plot of ln (tgel) versus 1/T for Figure 7 FTIR spectra of (a) uncured imide
BPEPA (▲), m-TPEPA (■) and p-TPEPA (○). oligomers and (b) heat-treated at 350 °C for 1 h.

suggesting that the greater number of radicals and the higher collision probabilities, caused the
trifunctional oligomers to react more readily.
Table 2 also reports the gel times for all the oligomers during isothermal tests at 320 °C, which
confirms that the gel reaction rates of m-TPEPA and p-TPEPA were much higher than that
of BPEPA. Moreover, the degree of cure (agel) at the gel time was calculated by the value of
glass transition temperature combined for different curing levels with rheological (tgel) and
isothermal DSC data [47]. No significant differences were observed for the values of agel for
m-TPEPA (0.19) and p-TPEPA (0.20), which were both lower than that obtained for BPEPA
(0.24), confirming that the topological and steric factor restrictions for the trifunctional
oligomers by the branched structure, compared with the values of bifunctional oligomer at
the beginning of the curing reaction, resulted in radicals becoming embedded or difficultly
in forming resonance-stabilised radicals at relatively higher levels of conversion. Otherwise,
these comparisons between activation energies and gelation rates determined via different
viscosity data and kinetic equations were useful for the purpose of validating the accuracy of
the results. The above combined evaluation methods for resin processability can be used for
other phenylethynyl-terminated imide oligimer systems.

3.5 Infrared spectra of the uncured and cured imide resins

In order to better understand the curing process of the three imide oligomers, FTIR spectra
were used to monitor the chemistry of the various reactions taking place. The most significant
differences between the IR spectra of the various structures are shown in Figure 7. The structure
of cured BPEPA resin (1b) shows a very weak band at 924 cm-l, which was assigned by other
authors to the trans C–H out-of-plane deformation vibration [46,48]. However, for the cured
trifunctional resins, m-TPEPA (2b) and p-TPEPA (3b), the band at 924 cm-l is very weak. In
addition to the IR spectra of the structures, the cured m-TPEPA and p-TPEPA resins show two
bands at 748 and 700 cm-1 as a function of the cis content, while at the same time the bands
are absent for BPEPA, indicating that the structures of cured trifunctional imide resins are cis
configurations, while the difunctional imide structure is trans. Under these conditions, the
simplest scheme representing the different curing reactions in multifunctional phenylethynyl-
terminated imide oligomers can take the form of Scheme 1. In the first step for the trifunctional
oligomers, the ethynyl-to-ethynyl addition reactions lead to a cis configuration containing

www.prkm.co.uk
12 Changwei Liu et al.

polyene structures, which may change the part of the configuration from cis to trimerisation
in the later stages of the reaction. However, the difunctional imide oligomer may just form the
structure with a trans configuration, which can explain the sustained growth of the Ea values
over the full range of conversion that have been discussed previously.

4. CONCLUSIONS
A series of high performance phenylethynyl-terminated imide oligomers used as RTM resin
matrices were studied systematically by dynamic DSC, small amplitude oscillatory shear
rheometry, and IR spectra. For the non-isothermal curing kinetics, the results shows that the
activation energy (Ea) depended on the extent of conversion evaluated with the Friedman
method, Flynn–Wall–Ozawa, and Tang methods, and the Ea values of the trifunctional
oligomers continued to decline compared with the sustained growth of Ea values of BPEPA.
The experimental master plots for BPEPA and p-TPEPA were in good agreement with the
theoretical master plot corresponding to the Avram-Erofe’ev (n=1.5) mechanism. Meanwhile,
comparison of the experimental master plots revealed that the kinetics of m-TPEPA was most
probably describable by the Phase boundary reaction (n=2 or n=3) models. Moreover, the
degree of cure (agel) at the gel time (tgel) was calculated from the value of the glass transition
temperature at different curing temperature combined with rheological and isothermal DSC
data. The chemorheological properties of the oligomers were measured and numerically fit
with the dual Arrhenius model and gel model. According to structural changes during curing
characterised using FTIR spectra, it was inferred that the structures of cured trifunctional imide
resins have cis configurations, which may change the part of the configuration from cis to
trimerisation in the later stages of the reaction, while the difunctional imide structure is trans.

5. ACKNOWLEDGEMENTS
This work was supported by Youth Science Foundation of Heilongjiang Province, China (Grant
No. QC2014C008) and Youth Science Foundation of Heilongjiang Academy of Sciences
(No. 2014-YQ-01).

Published online: 8 February 2016

6. REFERENCES
[1] Hsiao, S.H., Yang, C.P. and Chu, K.Y. (1997) Macromolecules, 30, 165–170.
[2] Wang, J.H., Jiao, Q.Z., Li, H.S. et al. (2014) Polym. Compos., 35, 1952–1959.
[3] Li, X.B. and Coleman, M.R. (2014) Polym. Compos., 35, 1473–1485.
[4] Khazaka, R., Locatelli, M.L., Diaham, S. and Bidan P. (2013) Polym. Degrad. Stab., 98, 361–367.
[5] Rao, X., Zhou, H., Dang, G. et al. (2006) Polymer, 47, 6091–6098.
[6] Iqbal, M., Norder, B., Mendes, E. and Dingemans, T.J. (2009) J. Polym. Sci. Part A Polym, Chem., 47, 1368–1380.
[7] Fang, X., Rogers, D.F., Scola, D.A. and Stevens, M.P.X. (1998) J. Polym. Sci. Part A Polym. Chem., 36, 461–470.
[8] Miyauchi, M., Ishida, Y., Ogasawara, T. and Yokota, R. (2013) React. Funct. Polym. 73, 340–345.
[9] Chae, B., Lee, S.H., Kim, S.B. et al. (2012) Vib. Spectrosc., 60, 137–141.
[10] Harris, F.W., Pamidimukkala, A., Gupta, R. et al. (1984) J. Macromol. Sci. Part A Pure Appl. Chem., 21, 1117–1135.
[11] Hergenrother, P.M. and Smith, J.J.G. (1994) Polymer, 35, 4857–4864.
[12] Hergenrother, P.M., Connell, J.W. and Smith, J.J.G. (2000) Polymer, 41, 5073–5081.
[13] Zuo, H.J., Chen, J.S., Yang, H.X. et al. (2007) J. Appl. Polym. Sci., 107, 755–765.
[14] Jayaraman, S., Srinivasan, R. and McGrath, J.E. (1995) J. Polym. Sci. Part A Polym. Chem., 33, 1551–1563.
[15] Connell, J.W., Smith, J.G., Hergenrother, P.M. and Criss, J.M. (2003) High Perform. Polym., 15, 375–394.
[16] Smith, J.G., Connell, J.W., Li, C.J. et al. (2006) High Perform. Polym., 18, 341–354.
[17] Cho, D. and Drzal, L.T. (2000) J. Appl. Polym. Sci., 76, 190–200.

www.prkm.co.uk
 Curing kinetics, rheological and IR study of imide oligomers 13
[18] Cho, D., Choi, Y. and Drzal, L.T. (2001) Polymer, 42, 4611–4618.
[19] Fang, X., Xie, X.–Q., Simone, C. D. et al. (2000) Macromolecules, 33, 1671–1681.
[20] Gong, C., Luo, Q., Li, Y. et al. (2011) J. Polym. Sci. Part A Polym. Chem., 49, 4476–4491.
[21] Nakamura, K., Ando, S. and Takeichi, T. (2001) Polymer, 42, 4045–4054.
[22] Takekoshi, T. and Terry, J.M. (1994) Polymer, 35, 4874–4880.
[23] Pickard, J.M., Jones, E.G. and Goldfarb, I.J. (1979) Macromolecules, 12, 895–902.
[24] Wright, M.E. and Schorzman, D.A. (2001) Macromolecules, 34, 4768–4773.
[25] Wright, M.E., Schorzman, D.A. and Berman, A.M. (2002) Macromolecules, 35, 6550–6556.
[26] Ayambem, A., Mecham, S.J., Sun, Y. et al. (2000) Polymer, 41, 5109–5124.
[27] Yang. Y., Fan, L., Ji, M. and Yang, S.Y. (2010) Eur. Polym. J., 46, 2145–2155.
[28] Liu, C.W., Zhao, X.G., Yu, X.H. et al. (2013) Polym. Degrad. Stab., 98, 230–240.
[29] Friedman, H.L. (1964) J. Polym. Sci. Part C, 6, 183–195.
[30] Flynn, J.H. and Wall, L.A. (1966) J. Polym. Sci. Part B Polym. Lett., 4, 323–328.
[31] Ozawa, T. (1965) Bull. Chem. Soc. Jpn., 38, 1881–1886.
[32] Doyle, C.D. (1961) J. Appl. Polym. Sci., 5, 285–292.
[33] Tang, W., Liu, Y., Zhang, H. and Wang, C. (2003) Thermochim. Acta, 408, 39–43.
[34] Vyazovkin, S. (2001) J. Comput. Chem., 22, 178–183.
[35] Gotor, F.J., Criado, J.M., Malek, J. and Koga, N. (2000) J. Phys. Chem. A, 104, 10777–10782.
[36] Ozawa, T. (1986) Thermochim. Acta., 100, 109–118.
[37] Koga, N. and Criado, J.M. (2005) J. Am. Ceram. Soc., 81, 2901–2909.
[38] Starink, M.J. (2003) Thermochim. Acta., 404, 163–176.
[39] Mussatti, F.G. and Macosko, C.W. (1973) Polym. Eng. Sci., 13, 236–240.
[40] Tajima, Y.A. and Corzier, D. (1983) Polym. Eng. Sci., 23, 186–190.
[41] Roller, M.B. (1986) Polym. Eng. Sci., 26, 432–440.
[42] Raghavan, S.R., Chen, L.A., McDowell, C. and Khan, S.A. (1996) Polymer, 37, 5869–5875.
[43] Simone, C.D. and Scola, D.A. (2003) Macromolecules, 36, 6780–6790.
[44] Argyropoulos, D.S. and Bolker, H.I. (1987) Macromolecules, 20, 2915–2922.
[45] Alvarenga, L.M., Xavier, T.P., Barrozo, M.A.S. et al. (2012) Procedia. Eng., 42, 113–122.
[46] Tan, D., Shi, T. and Li, Z. (2011) Res. Chem. Intermed., 37, 831–845.
[47] Monti, M., Puglia, D., Natali, M. et al. (2011) Compos. Sci. Technol., 71, 1507–1516.
[48] Simionescu, C.I., Percec, V. and Dumitrescu, S. (1977) J. Polymer Sci. Polymer Chem. Ed., 15, 2497–2509.

www.prkm.co.uk