Thermochimica Acta 623 (2016) 15–21

Contents lists available at ScienceDirect

Thermochimica Acta
journal homepage: www.elsevier.com/locate/tca

Isothermal curing kinetics and mechanism of DGEBA epoxy resin with

phthalide-containing aromatic diamine
Rong Ren a , Xuhai Xiong a,c,∗ , Xinghua Ma a , Siyang Liu b , Jing Wang a , Ping Chen b ,
You Zeng c
a
Liaoning Key Laboratory of Advanced Polymer Matrix Composites, Shenyang Aerospace University, Shenyang 110136, China
b
State Key Laboratory of Fine Chemicals, School of Chemical Engineering, Dalian University of Technology, Dalian 116012, China
c
Shenyang National Laboratory for Materials Science, Advanced Carbon Division, Institute of Metal Research, Chinese Academy of Sciences, Shenyang
110016, China

a r t i c l e i n f o a b s t r a c t

Article history: The curing reaction of diglycidyl ether of bisphenol A (DGEBA) with a novel aromatic diamine contain-
Received 1 September 2015 ing phthalide structure (BAPP) was studied by differential scanning calorimetery (DSC). The optimal
Received in revised form 3 November 2015 formulation of DGEBA/BAPP system was obtained in terms of the curing behavior and glass transition
Accepted 7 November 2015
temperature determined by dynamic DSC. The nth-order reaction model and autocatalytic model were
Available online 10 November 2015
used to analyze the curing kinetics and mechanism based on the isothermal DSC technique. The results
indicated that the Kamal autocatalytic model had a great agreement with the experimental data from
Keywords:
the start of curing reaction to the initiation of diffusion control stage. Diffusion-controlled effect was also
Epoxy resin
Aromatic diamine
taken into account and then the extended Kamal model could describe precisely the curing reaction of
Curing kinetics BAPP/DGEBA in the entire conversion range.
Isothermal DSC © 2015 Elsevier B.V. All rights reserved.
Diffusion control

1. Introduction polymer network. A chain-extended aromatic amine with

Epoxy resins are important and versatile thermosetting mate-
rials, which are effortlessly tailored to suit specific performance
characteristics, and thus widely used in diverse industries spanned
from electronic and electrical equipments to civil aviation and
space parts as coatings, adhesives, electrical insulations, matri-
ces for fibrous composites [1–4]. They could be cured by various
types of hardener to obtain cross-linked network with mechan-
ical properties ranging from extreme flexibility to high stiffness
and strength [5–7]. Aromatic amine-cured epoxy systems enjoy
more consideration and their applications are widened constantly
in large load-bearing and harsh conditions due to their excellent
corrosion and erosion resistance, high adhesive, good thermal sta-
bility and outstanding mechanical properties [6–9]. Unfortunately,
those rigid materials also easily suffer from brittle fracture as a
result of aromatic backbone and high cross-linked density. It is
a focus of interest to obtain a balance between heat resistance
and toughness of cured products by controlling microstructure of
∗ Corresponding author at: Shenyang Aerospace University, Liaoning Key Labo-
ratory of Advanced Polymer Matrix Composites, No.37 Daoyi South Avenue, Daoyi
Development District, Shenyang, 110136, China. Tel.: +86 24 89723970.
E-mail address: xiongxuhai@sau.edu.cn (X. Xiong).
phthalide cardo structure was found to be an ideal candidate hard-
ener for high-performance epoxy resins [10]. Its extended length of
molecular chain could lead to an increase in the distance between
the cross-linked points and then the toughness of epoxy network
was improved. In the other hand, the introduction of heat-resistant
multi-aryl skeleton might compensate for the loss of heat resistance
resulting from decrease in the cross-linked density.
Generally, the low molecular weight linear liquid epoxy resin
cannot meet end-use applications before it is converted to three-
dimensional crosslinked network through curing process [11].
As curing reaction proceeds, the epoxy system will experience
gelation and vitrification stages [9,12]. The crosslinked density
increases and glass transition temperature (Tg ) and mechanical
properties are subsequently enhanced. For the epoxy/amine sys-
tem, the curing reaction follows a complicated mechanism and
many reactions occur, such as ring-opening reaction of the EP ring
and primary amines for producing chain growth and secondary
amines for building chain branches, etherification of EP ring with
a pendant hydroxyl group and homopolymerization of EP group
at higher temperature in the absence of active N H functional-
ity [3–5,13,14]. These necessitate the investigation on kinetic and
mechanism of novel epoxy/amine system. Kinetic data can provide
information for optimizing curing cycles to reduce manufacture
cost or improve properties of cured products [5,8].

http://dx.doi.org/10.1016/j.tca.2015.11.011
0040-6031/© 2015 Elsevier B.V. All rights reserved.
16 R. Ren et al. / Thermochimica Acta 623 (2016) 15–21
Scheme 1. Chemical structures of DGEBA and BAPP.
In our previous paper, we incorporated phthalide cardo struc-
ture into epoxy network by using a novel aromatic diamine
containing phthalide structure (BAPP) as curing agent of digly-
cidylether of bisphenol A (DGEBA) epoxy resin [10]. The curing
behavior and kinetics were investigated by differential scanning
calorimetry (DSC) and the kinetic analysis of non-isothermal cur-
ing reaction showed that autocatalytic model was suitable to
describe the curing mechanism. But it is actually rather difficult
and unfaithful to recognize the reaction model type under non-
isothermal conditions due to both rate constant, k, and reaction
model vary simultaneously with temperature, giving rise to sig-
moidal curves of the extent of conversion (˛) versus temperature
(T). On the contrary, autocatalytic profile is readily recognized
based on isothermal data because in this case k = const, so that the
kinetic curve shape is determined by the reaction model alone [15].
To continue previous study, isothermal DSC method was employed
to investigate curing kinetics of DGEBA/BAPP system in the present
paper.
2. Experiment
2.1. Materials
The epoxy compound used in the study was a diglycidylether
of bisphenol A (DGEBA) provided from Wuxi Resin Works,
with an epoxy equivalent weight (eew) of 185–210, and dried
at 100 ◦ C in vacuum for 1 h before use. The curing agent
3,3-Bis(4-(4-aminophenoxy)phenyl)phthalide (BAPP), was synthe-
sized according to the published procedures [16]. The chemical
structures of DGEBA and BAPP are shown in Scheme 1.
2.2. Preparation of DGEBA/BAPP blends
BAPP and DGEBA were completely mixed by mechanical stirring
at 70 ◦ C with BAPP/DGEBA stoichiometric ratios of 0.9, 1.0, 1.1, 1.2
and 1.3, respectively.
2.3. DSC characterization
To measure the total heat, H0 , evolved during the complete
cure, dynamic DSC analysis was performed under N2 protection
and run twice. The first scan was conducted from 30 ◦ C to 300 ◦ C at
the heating rate of 10 ◦ C/min to calculate H0 , and the second scan
was conducted from 30 to 200 ◦ C at a heating rate of 20 ◦ C/min to
obtain the glass transition temperature (Tg ).
To measure cumulative heat, Ht , isothermal DSC analyses were
carried out at 140 ◦ C, 150 ◦ C, 160 ◦ C and 170 ◦ C. And then the cured
samples were also scanned from 30 to 200 ◦ C at a heating rate of
20 ◦ C/min to obtain Tg . The temperature and heat is calibrated with
high-pure indium standard beforehand.
2.4. Fundamental theory on curing reaction kinetics of
DGEBA/BAPP
The basic assumption for the application of DSC technique to
the cure of the thermosetting polymers is that the reaction rate is
considered to be directly proportional to the heat flow [9,13,17,18].
d˛/dt = (dH/dt)/H0 = kf (˛) (1)
where, d˛/dt is the reaction rate, dH/dt is the heat flow, H0 is the
overall reaction heat, k is specific rate constant at temperature T,
f(˛) is the reaction model and ˛ is the extent of reaction, which is
usually calculated by the following expression:
˛ = Ht /H0 (2)
where, Ht is the heat evolved up to a specific time. In this work,
Ht was calculated by integrating the isothermal DSC peaks up
to a specific time and H0 was estimated by integrating the main
exothermic peak of non-isothermal DSC curve measured at a heat-
ing rate of 10 ◦ C/min.
The curing kinetics of epoxy/amine system has been frequently
described by both nth order and autocatalytic mechanisms and f(˛)
is usually taken in the form of (1 − ˛)n or of ˛m (1 – ˛)n , respectively
[10,13]. For nth order kinetics, whose maximum reaction rate is at
t = 0, d˛/dt is proportional to the fraction of unreacted material and
expressed as follow:
d˛/dt = k(1 − ˛)n (3)
where, n is the reaction order.
The characteristic of an autocatalytic kinetics is that the maxi-
mum rate of reaction is at some intermediate conversion, indicating
that d˛/dt is affected by unreacted material and resulted prod-
uct. The Kamal model is a more conventional autocatalytic model:
[17–22].
d˛/dt = (k1 + k2 ˛m )(1 − ˛)n (4)
where, k1 and k2 are the rate constants, and m and n are the kinetic
exponents that depend on the nature of the epoxy resin and the
amine hardener, and the curing temperature. Note that Eq. (4) is
transformed into nth-order model. When m = 0.
Generally, reaction rate constant for curing reaction of ther-
mosetting resin follows the Arrhenius law (Eq. (5)):
k = A exp(−E/RT ) (5)
where, A is the pre-exponential factor, E is the activation energy,
and R is the gas constant. Arrhenius parameters can be obtained
from the temperature dependence of the rate constants k, k1 and
k2 .
 R. Ren et al. / Thermochimica Acta 623 (2016) 15–21
Fig. 1. Non-isothermal DSC thermographs of epoxy systems with various
BAPP/DGEBA stoichiometric ratio.
3. Results and discussion
3.1. Non-isothermal DSC analysis of DGEBA/BAPP systems
Fig. 1 shows the dynamic DSC thermograms of DGEBA/BAPP sys-
tems with various stoichiometric ratios investigated at 10 ◦ C/min
under a nitrogen atmosphere. In the 100–250 ◦ C range, it is
observed there is one main exothermic peak for all systems, which
is associated with the addition reactions of epoxy ring with pri-
mary and secondary amines. Above 250 ◦ C, another exothermic
transition appears and its inducement is complicated. As well
known, some side reactions, including etherification of hydroxyl
groups and epoxy groups, epoxy ring-opening homopolymeriza-
tion and thermal decomposition of weak bonds, occur in this
high-temperature region [1,4,13]. For the epoxy systems with
the lack of curing agents, a relatively gentle exothermic transi-
tion is found, and it may be ascribed to the etherification and
homopolymerization of the residual epoxy groups. In contrast,
the thermal pyrolysis is probably a predominant factor for the
drastic exothermic transitions of the systems with the excess of
curing agents. Noteworthily, the optimal ratio of BAPP/DGEBA is
not the theoretical stoichiometric ratio but equals to 1.1 accord-
ing to the benchmark with respect to the disappearance of the
high-temperature exothermic transition.
The observation of the glass transition temperatures (Tg s)
evaluated from second scanning of samples also supports above-
mentioned conclusion. From Fig. 2, the system with BAPP/DGEBA
ratio 1.1 exhibits the highest Tg value, indicating more perfect net-
work structure of the resultant cured product. These phenomena
may result from the phthalide-containing structure of hardener,
which lowers the mobility of BAPP reactive sites trapped in the
gel-state network. Therefore, it is necessary for BAPP/DGEBA sys-
tem to add the extra dosage of hardener to achieve a full cure
of epoxy groups. In the sequential sections, the optimal system
with BAPP/DGEBA ratio 1.1 was chosen to carry out the analysis
of isothermal curing kinetics.
3.2. Isothermal curing behavior of the optimal DGEBA/BAPP
system
The DSC heat flow-time curves of the isothermal curing reac-
tions of DGEBA/BAPP mixture containing a 10th excess of amine at
the different curing temperatures (Tc ) are shown in Fig. 3. Allow-
ing for the presence of only one single exothermic peak without
shoulder for each isothermal curve, the side reactions such as
17
Fig. 2. Tg of the completely cured epoxy resins dependence on the BAPP/DGEBA
stoichiometric ratio.
Fig. 3. Isothermal DSC thermographs of epoxy system.
the etherification and homopolymerization of epoxide rings could
be neglected [13,17]. The maximum rate of the curing exotherm
increases with increasing isothermal temperature, whereas the
time to reach the maximum of curing exotherm is decreased.
Based on the heat flow data in Fig. 3, the conversion (˛) and
reaction rate (d˛/dt) with respect to time at different curing tem-
peratures were obtained by using Eqs. (1) and (2), respectively,
and illustrated in Fig. 4. We can see from Fig. 4a that the con-
version increases sharply with curing time at early stage of the
curing reaction, then decelerates, and eventually leaves off and
no longer changes upon reaction time. The possible cause for the
above-mentioned observation is that abundant concentration of
reactive sites and low viscosities at the beginning stage of reac-
tion benefit the proceeding of curing reaction, and then decreasing
concentration of reaction sites and increasing viscosity decrease
drastically the rate of reaction. As a consequence, the conversion
increases slowly as the curing degree increase.
The plots of the reaction rate versus time at different Tc s (Fig. 4b)
are similar to the heat flow-time curves. Initially, the reaction rate
increases with time, and then it gradually decreases, eventually
approaching to zero. For the isothermal cure at 140 ◦ C, lower values
of the conversion degree (˛ < 0.4) and reaction rate are observed. It
seems suspectable because ˛ is far below the theoretical conversion
at the gel point (˛gel = 0.577) calculated based on Flory theory [13].
To illuminate the above findings, the Tg s of samples, which
had been cured during isothermal DSC experiments at various
18 R. Ren et al. / Thermochimica Acta 623 (2016) 15–21
Fig. 4. Conversion degree (a) and rate of conversion (b) as a function of time.
Fig. 5. Plots of Tg versus ˛.
temperatures, were obtained by performing dynamic DSC
measurement at a heating rate of 20 ◦ C/min. The scatter diagram
of Tg against ˛ is shown in Fig. 5. The sample cured at 140 ◦ C has a
Tg of 151 ◦ C and exhibits the largest span between curing tempera-
ture and Tg . This may be the reason for the lower conversion of the
isothermal curing reaction at 140 ◦ C. As well known, with the pro-
gressing of curing process the crosslink density of resin systems
increases gradually, simultaneously leading to a steady enhance-
ment of glass transition temperature (Tg ). The mobility of growing
molecular chains will be lowered and frozen eventually if the sys-
tem vitrify, which occurs when the glass transition temperature
of the system reaches the actual curing temperature [12,14]. Con-
sequently, the curing reaction moves from chemical to diffusion
control [12,14]. The curing mechanism of DGEBA/BAPP system at
140 ◦ C may be that the addition of epoxy groups and primary amine
is predominant for the initial stages of the reaction [13]. As curing
reaction progresses, the Tg of growing linear polymer exceeds the
curing temperature, which results in an early termination of curing
reaction.
In addition, the dependence of Tg , which was evaluated from
the middle point of the glass transition process in the DSC curve,
on ˛ presents a good linear correlation. According to the fitting for-
mula (Tg = 20.6˛ + 143.5), the Tg value of completely cured sample
achieves 164.1 ◦ C, which is slightly lower than the experimental
value (Tgm = 169.5 ◦ C) shown in Fig. 2. The believable cause is that
the occurrence of a small amount of side reactions at higher temper-
ature during dynamic DSC experiment, such as self-polymerization
of DGEBA, led to an increase in the crosslink density and Tg of cured
network.
3.3. Isothermal curing kinetic analysis of the optimal
DGEBA/BAPP system
3.3.1. nth order model
To carry out the kinetic analysis based on the nth order model,
Eq. (3) should be transformed into Eq. (6).
ln(d˛/dt) = ln k + n ln(1 − ˛) (6)
The kinetic parameters k and n can be calculated from linear
regression of ln(d˛/dt) versus ln(1−˛) and the results were listed
in Table 1. The plots obtained by using Eq. (6) are shown in Fig. 6. It
can be found from the inset in Fig. 6 that ln(d˛/dt) versus ln(1 − ˛)
shows a good linearity in the 0.1 < ˛ < 0.4 region, and thus this con-
version interval was selected for the computation of ln k and n. The
experiment at 140 ◦ C was not used in the calculations because the
conversion degree and reaction rate is too low to obtain convincing
results.
According to Eq. (5), the activation energy (E) and the log-
arithm of preexponential factor (ln A) can be calculated by
plotting ln k versus 1/T in the linear region. The obtained val-
ues are E = 53.87 kJ/mol and ln A = 13.02 s−1 , with a correlation
coefficient r = 0.9994. The E value is comparable to the results
(47.5–60.9 kJ/mol), with the average value of 56.1 kJ/mol, cal-
culated by model-free isoconversional method with the more
accurate Starink equation based on non-isothermal DSC experi-
ments [10].
3.3.2. Autocatalytic model
The determination of the kinetic parameters in Eq. (4) is a
challenge in spite of the fact that several numerical and graphi-
cal procedures have been developed [9,19]. In this study, Kenny
analytical–graphical method, which is known as the best calcula-
tive approach, was used for calculating parameters m, n and k2 . The
reaction rate constant k1 was determined as the initial reaction rate.
At the start of the curing reaction (t = 0, ˛ = 0), Eq. (4) simplifies to:
[d˛/dt]t=0 = k1 (7)
The k1 values were obtained directly from isothermal reaction
rate curves by extrapolating to zero time and listed in Table 1.
Eq. (4) can be rewritten in the following form:
ln(d˛/dt) = ln[k1 + k2 ˛m ] + n ln(1 − ˛) (8)
Fig. 7a represents a typical plot of ln(d˛/dt) as a function
of ln(1 − ˛) for the DGEBA/BAPP system at 150 ◦ C. Linearity is
observed for the interval 0.1 < ˛ < 0.4 which was used to estimate n.
It can be concluded from Eq. (8) that the slope of the linear fitting
observed is the kinetic exponent n when ˛ is close to 0.
 R. Ren et al. / Thermochimica Acta 623 (2016) 15–21 19

Table 1
Kinetic parameters of isothermal curing for DGEBA/BAPP system.

Temperature/◦ C nth order model Kamal autocatalytic model

n k/10−3 s−1 m n m+n k1 /10−3 s−1 k2 /10−4 s−1 k2 /k1 ˛c C

150 1.77 1.655 0.187 1.992 2.18 0.965 0.18 0.0187 0.628 15.24
160 1.36 2.443 0.150 1.572 1.72 1.438 0.26 0.0181 0.811 22.34
170 1.39 3.298 0.195 1.640 1.84 1.932 0.355 0.0184 0.854 38.92

Fig. 6. Plots of ln(d˛/dt)versus ln(1 − ˛) for nth order model analysis.

Eq. (8) can be further rearranged in the following form:

 
d˛/dt
ln − k1 = ln k2 + m ln ˛ (9)
(1 − ˛)n

A typical plot of the first term of Eq. (9) as a function of ln ˛ is

shown in Fig. 7b. The first term of Eq. (9) has been calculated using
initial values of k1 and n. The kinetic exponent, m and the kinetic
constant, k2 , were calculated, respectively, from the slope and inter-
cept of this plot. A first set of parameter values was obtained using
the procedure described above. However, to obtain more values, an
iterative procedure can be applied. Eq. (8) can be rearranged again
to give:
 
d˛/dt
ln − = n ln(1 − ˛) (10)
k1 + k2 ˛m

After k2 and m have been evaluated using Eq. (9), the left-side
term of Eq. (10) can be calculated and plotted as a function of
ln(1 − ˛). The corresponding plot at 150 ◦ C is shown in Fig. 7c. A
new value of the reaction order n obtained from the slope of the
linear portion of this plot is close to the one obtained previously.
Eq. (9) can be used again to obtain a second set of parameter val-
ues of m and k2 that are close to the previous ones. These new
values are applied again in Eq. (10). This iterative procedure above-
mentioned was repeated until the acquired values for n, m and k2
revealed less than 1% difference between subsequent calculations.
Therefore, the four kinetic parameters (k1 , k2 , m and n) of the Kamal
equation obtained by this procedure were reported in Table 1.
The values of power exponents (m and n), as shown in Table 1,
follow the restriction of 1 ≤ n ≤ 2 and m ≤ 1, which was found
by Talbot based on the simple mechanistic model of the epoxy-
amine addition [23].The overall reaction order (m + n) in a range
of 1.72–2.18 approximates to the restriction that: m + n = 2, and is
also comparable to the results obtained from the DGEBA cured with
aromatic diamine systems [8,9]. The value of m for the present sys- Fig. 7. Graphic representation of (a) Eq. (8), (b) Eq. (9) and (c) Eq. (10) for the
tem seemingly signifies a low variation on the curing temperature isothermal cure at 150 ◦ C.
20 R. Ren et al. / Thermochimica Acta 623 (2016) 15–21

Fig. 8. Arrhenius plots for DGEBA/BAPP system.

and is smaller than these reported by other researchers when they

investigated the curing kinetics of epoxy/aromatic amines systems
[8,9,22].
According to Eq. (4), it is known that the kinetic exponent (m)
represents the quantitative evaluation of autocatalytic effect, that
is, a smaller m value indicates weaker autocatalysis. Therefore,
the autocatalysis plays a weak role during the curing process of
DGEBA/BAPP system. This may be ascribed to the huge molecu-
lar volume and non-coplanar molecular structure BAPP hardener,
which is bad for the formation of low energy cyclic transition state
between amine–epoxide–hydrogen donor groups.
Seeing from Table 1, the parameters k1 and k2 increase some-
what with increasing isothermal curing temperatures, whereas the
ratio of the reaction rate constants k1 /k2 is independent of the reac-
tion temperature. The difference in the values of k1 and k2 is quite
noticeable, k1 » k2 , and the value of k1 is approximately equal to that
of k obtained from above nth order model analysis. Both observa-
tions also demonstrate that the autocatalysis is not dominant and
that the curing mechanism is different from the previous results
reported for DGEBA cured with aromatic diamine [8,9,22].
The activation energy (E1 , E2 ) for the Kamal autocatalytic model
reactions can be calculated from the slope of the linear plot of
ln k1 and ln k2 against 1/T (Fig. 8), respectively. E1 = 54.16 kJ/mol is
greater than E2 = 52.97 kJ/mol, which is consistent with the results
reported in the literatures [13,14]. Additionally, the value of E1 is
almost identical with that obtained from nth model. This is because
both values are associated with the same reaction mechanism that
amine addition is initiated by hydrogen-bond donor molecules,
such as moisture or impurities, whereas E2 is in relation to the
amine addition internally catalyzed by the hydroxyl group of sec-
ondary amine [14]. The E1 ≈ E2 also indicates that the autocatalytic Fig. 9. Comparison of calculated data from reaction model with experimental data.
effect is secondary.

3.4. Prediction of curing kinetics model between chemically controlled and diffusion-controlled regions
[24].
Fig. 9 presents the comparisons of the experimental results To predict precisely the rate of curing reaction in the whole
with the prediction values of nth order model (Eq. (3)) and Kamal range of ˛, an amended Kamal equation was proposed by consid-
autocatalytic model (Eq. (4)), respectively. As expected, Kamal ering diffusion effect on the later period of curing reaction. Eq. (4)
autocatalytic model exhibits better prediction effect than nth was evolved into [8,17]:
order model at initial stages of the reaction. At high conversion
region, however, one can observe significant deviations between 1
the experimental data and the calculated ones based on both mod- d˛/dt = (k1 + k2 ˛m )(1 − ˛)n (11)
1 + ec(a−ac )
els, and these discrepancies decrease with increasing isothermal
curing temperature (Tc ). As seen from Fig. 9, the deviation points where, C is a constant value, ˛c is the critical conversion at which
successively appear at ˛ = 0.42, 0.68 and 0.81 with increasing Tc . It the rate constant of chemically controlled kinetics equals to that of
is believed that the deviation is associated with kinetics transition diffusion-controlled kinetics.
 R. Ren et al. / Thermochimica Acta 623 (2016) 15–21
Additionally, f (˛) is known as the diffusion factor, which is
expressed as:
1 [3] T. Dyakonov, Y. Chen, K. Holland, J. Drbohlav, D. Bums, Thermal analysis of
f  (˛) = (12)
1 + ec(a−ac )
According to the Eq. (12), when ˛ < < ˛c , f (˛) approximates to
1, which suggests that the chemically controlled kinetics is dom-
inant and the diffusion effect is negligible. When ˛ approachs ˛c ,
the influence of diffusion increases gradually and the f (˛) values
are rapid decreasing; when ˛ = ˛c , the value of f (˛) is equal to 0.5,
and then the diffusion control is dominant. However, when ˛ » ˛c ,
the value of f (˛) reaches zero which implies that the vitrification
happens at higher conversion and further results in chain segment
motion is frozen and curing reaction is ceased.
Eq. (12) can be further expressed in the following form:
ln(1/f  (˛) − 1) = c(˛ − ˛c ) (13)
On the basis of Eq. (11), f (˛) can be obtained from the ratio
of experimental rate to that calculated by the Kamal model. And
then the values of C and ˛c at 150, 160 and 170 ◦ C can be deter-
mined according to Eq. (13) by linear fitting ln(1/f (˛)−1) versus
˛ at the latter stage of curing reaction. The resulting parameters
were listed in Table 1. From the table, critical conversion (˛c ) shows
temperature dependence, and increases gradually with tempera-
ture. It can be observed from Fig. 9 that the amended Kamal model
describes the curing reaction of DGEBA/BAPP precisely in the entire
conversion range.
4. Conclusions
The curing behavior and kinetics of diglycidyl ether of bisphe-
nol A (DGEBA) epoxy system, cured by a novel aromatic diamine
containing phthalide structure (BAPP), were studied in detail using
DSC method. Dynamic DSC experiments showed the DGEBA/BAPP
system with 10% excess of curing agent was optimal formulation
and its cured product had the highest Tg . The study on curing kinetic
and mechanism of DGEBA/BAPP based on the isothermal DSC tech-
nique exhibited that autocatalytic Kamal model could describe
the isothermal curing process better than nth order model in the
non-diffusion-controlled region. The amended Kamal model with
a diffusion term could describe the reaction rate in the whole con-
version range.
Acknowledgments
The financial supports from National Natural Science Founda-
tion of China (no. 51303107), Liaoning Key Laboratory Fundamental
Research Project (no. LZ2015058), Aerospace Science Foundation
of China (no. 2014ZF54030), National Defense 12th Five-Year
Fundamental Research Program (no. A352010****) and Shenyang
National Laboratory for Material Science (no. 2015RP13) are grate-
fully acknowledged.
References
[1] B.A. Rozenberg, Kinetics, thermodynamics and mechanism of reactions of
epoxy oligomers with amines, Adv. Polym. Sci. 75 (1985) 113–165.
21
[2] G.R. Saad, E.E.A. Elhamid, S.A. Elmenyawy, Dynamic cure kinetics and thermal
degradation of brominated epoxy resin–organoclay based nanocomposites,
Thermochim. Acta 524 (2011) 186–193.
some aromatic amine cured model epoxy resin systems-I: materials synthesis
and characterization cure and post-cure, Polym. Degrad. Stab. 53 (1996)
211–242.
[4] X.H. Xiong, P. Chen, J.X. Zhang, Q. Yu, B.C. Wang, Preparation and properties of
high performance phthalide-containing bismaleimide modified epoxy
matrices, J. Appl. Polym. Sci. 121 (2011) 3122–3130.
[5] A. Catalani, M.G. Bonicelli, Kinetics of the curing reaction of a diglycidyl ether
of bisphenol A with a modified polyamine, Thermochim. Acta 438 (2005)
126–129.
[6] M.F. Mustafa, W.D. Cook, T.L. Schiller, H.M. Siddiqi, Curing behavior and
thermal properties of TGDDM copolymerized with a new pyridine-
containing diamine and with DDM or DDS, Thermochim. Acta 575 (2014)
21–28.
[7] T. Maity, B.C. Samanta, S. Dalai, A.K. Banthia, Curing study of epoxy resin by
new aromatic amine functional curing agents along with mechanical and
thermal evaluation, Mater. Sci. Eng.: A 464 (2007) 38–46.
[8] E.S. Balcerzak, H. Janeczek, B. Kaczmarczyk, H. Bednarski, D. Sek, A. Miniewicz,
Epoxy resin cured with diamine bearing azobenzene group, Polymer 45
(2004) 2483–2493.
[9] M. Ghaemy, M. Barghamadi, H. Behmadi, Cure kinetics of epoxy resin and
aromatic diamines, J. Appl. Polym. Sci. 94 (2004) 1049–1056.
[10] X.H. Xiong, R. Ren, S.Y. Liu, S.W. Lu, P. Chen, The curing kinetics and thermal
properties of epoxy resins cured by aromatic diamine with hetero-cyclic side
chain structure, Thermochim. Acta 595 (2014) 22–27.
[11] D. Rosu, F. Mustata, C.N. Cascaval, Investigation of the curing reactions of
some multifunctional epoxy resins using differential scanning calorimetry,
Thermochim. Acta 370 (2001) 105–110.
[12] N. Sbirrazzuoli, S. Vyazovkin, A. Mititelu, C. Sladic, L. Vincent, A study of
epoxy-amine cure kinetics by combining isoconversional analysis with
temperature modulated DSC and dynamic rheometry, Macromol. Chem. Phys.
204 (2003) 1815–1821.
[13] N. Sbirrazzuoli, A.M. Mija, L. Vincent, C. Alzina, Isoconversional kinetic
analysis of stoichiometric and off-stoichiometric epoxy-amine cures,
Thermochim. Acta 447 (2006) 167–177.
[14] S. Vyazovkin, N. Sbirrazzuoli, Mechanism and kinetics of epoxy-amine cure
studied by differential scanning calorimetry, Macromolecules 29 (1996)
1867–1873.
[15] S. Vyazovkin, A.K. Burnham, J.M. Criado, L.A. Pérez-Maqueda, C. Popescu, N.
Sbirrazzuoli, ICTAC Kinetics Committee recommendations for performing
kinetic computations on thermal analysis data, Thermochim. Acta 520 (2011)
1–19.
[16] X.H. Xiong, P. Chen, Q. Yu, N.B. Zhu, B.C. Wang, Synthesis and properties of
chain-extended bismaleimide resins containing phthalide cardo structure,
Polym. Int. 59 (2010) 1665–1672.
[17] C. Li, H. Fan, J.J. Hu, B.G. Li, Novel silicone aliphatic amine curing agent for
epoxy resin: 1,3-Bis(2-aminoethylaminomethyl) tetramethyldisiloxane. 2.
Isothermal cure, and dynamic mechanical property, Thermochim. Acta 549
(2012) 132–139.
[18] L. Barral, J. Cano, J. LoApez, I. López-Bueno, P. Nogueira, Cure kinetics of
amine-cured diglycidyl ether of bisphenol A epoxy blended with poly(ether
imide), Thermochim. Acta 344 (2000) 127–136.
[19] C.S. Wang, C.H. Lin, Novel phosphorus-containing epoxy resins. Part II: curing
kinetics, Polymer 41 (2000) 8579–8586.
[20] V.L. Zvetkov, R.K. Krastev, S. Paz-Abuin, Is the Kamal’s model appropriate in
the study of the epoxy-amine addition kinetics? Thermochim. Acta 505
(2010) 47–52.
[21] C. Zhang, W.K. Binienda, L.M. Zeng, X.C. Ye, S.W. Chen, Kinetic study of the
novolac resin curing process using model fitting and model-free methods,
Thermochim. Acta 523 (2011) 63–69.
[22] V.L. Zvetkov, V. Calado, Comparative DSC kinetics of the reaction of DGEBA
with aromatic diamines. III. Formal kinetic study of the reaction of
DGEBA with diamino diphenyl methane, Thermochim. Acta 560 (2013)
95–103.
[23] J.D.R. Talbot, The kinetics of the epoxy amine cure reaction from a
solvation perspective, J. Polym. Sci.: Part A: Polym. Chem. 42 (2004)
3579–3586.
[24] J.E.K. Schave, A description of chemical and diffusion control in isothermal
kinetics of cure kinetic, Thermochim. Acta 388 (2002) 299–312.