Professional Documents
Culture Documents
amine-epoxy ring opening reaction which results in curing techniques. In addition, a substantial
the formation of a secondary amine and a hydroxyl amount of research work has been devoted to
group; (ii) the secondary amine-epoxy reaction identifying any differences in the physical and
results in the formation of a tertiary amine and a mechanical properties of microwave and thermally
second hydroxyl group; (iii) an etherification processed materials. We have previously reported
reaction, which occurs via a reaction between one analyzes of the kinetics of the cure of several epoxy-
of the pendant hydroxyl groups and the oxirane amine systems [615].
functionalities in the epoxy resin. The epoxy resin diglycidyl ether of bisphenol A
The conventional process of thermoset curing of (DGEBA) is one of the most commonly used in
composites involves subjecting the material to industry. An industrial formulation of this epoxy
temperature and pressure cycles in an autoclave. resin, Ciba-Geigy GY260, was used in this work.
Problems with this conventional curing technique These epoxy resins are typically cured with the aid
are related to the length, cost and complexity of the of amine crosslinking agents. In this work, two of
autoclave curing cycle, which, depending on the the most common amine based crosslinking agents
material being cured, can take several hours and were employed, 4,4'-diaminodiphenyl sulphone
involve several temperature ramping and de- (DDS) and 4,4'-diaminodiphenyl methane (DDM).
ramping steps [1]. The major difficulty with In order to gain an understanding of the cure of
temperature ramping in conventional curing tech- composite materials, it is necessary to first under-
niques is related to the slow response of the stand the kinetics of the reactions occurring during
autoclave. the cure. In this study we have examined the
The inherent difficulties and inefficiencies in- kinetics of the amine reactions for GY260 cured
volved in conventional curing techniques have led with DDS and DDM for both thermal and micro-
researchers to explore other ways of facilitating the wave processing, with a view to examining
cure process which could save time and energy, as whether there are any fundamental differences
well as possibly lead to the formation of materials between the two processes.
with unique and desirable mechanical properties.
A number of alternatives to the conventional
technique of autoclave cure have been suggested. EXPERIMENTAL W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W
One of the more promising of these is microwave
induced curing, which has a number of possible The experimental procedures adopted have been
advantages over conventional techniques and has described in detail previously [79]. The thermal
often been suggested as a cost-effective alternative and microwave cure of the samples were mon-
to conventional heating procedures [2]. itored using FT-NIR spectroscopy. The thermal
Microwave heating is a direct form of heating, cure reactions were carried out in 3 mm id pyrex
compared with conventional heating techniques tubes containing approximately 0.2 g of sample.
which are based on conduction of heat through a The samples were cured in a preheated aluminum
material. Since microwave radiation generates heat block thermostatt with temperature control to
within materials, sample temperatures are poten- within 0.1 C and the FT-NIR spectra were
tially able to be changed and controlled more obtained in real time by transmission using a
readily. Any increase in input power or improve- PerkinElmer 1600 spectrometer fitted with an
ment in the tuning of a microwave oven immedi- external bench on which the heater block was
ately results in an increase in temperature of mounted. The microwave cure reactions were
the material undergoing cure. Microwave energy carried out in a single mode, tunable, cylindrical
affects only the material being processed, and is cavity operating at a frequency of 2.45 GHz. The
therefore inherently more energy efficient than cavity was fitted with a Teflon, cylindrical,
thermal heating, which also heats the walls of the approximately 5 g sample compartment and silica
oven and the air surrounding the process material. fibre-optics, with a 200 mm fiber diameter, for
A definite potential advantage of microwave acquiring the real-time FT-NIR spectra in the center
heating over conventional techniques is thus lower of the sample using with an optical path length of
production costs as a consequence of its greater approximately 2 mm [9]. The temperature of the
energy efficiency. Another potential advantage is sample within the optical path was monitored by
related to the greater controllability of microwave an adjacent fiber-optic temperature probe and
heating, which offers the possibility of more easily controlled to within 0.2 C by variation of the
producing materials with unique physical and microwave power to the cavity [9]. The FT-NIR
mechanical properties. signal to noise ratios for the thermal cure reactions
Recently, several reports have suggested that were greater than those which were obtainable for
there are fundamental differences between the the microwave cure reactions, because of the fiber-
mechanisms of microwave and conventional heat- optic monitoring used in the latter experiments,
ing, and that these can affect their respective and hence the concentration and rate data for the
kinetics of polymerization or curing [35]. Workers former studies are subject to smaller experimental
studying these reactions have thus concentrated errors.
their efforts on investigating the possibility of The experimental techniques used to obtain the
reduced processing time with the use of microwave time dependence of the functional group concen-
radiation compared with conventional thermal trations of the primary and secondary amine
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
Epoxy-amine Resin Systems / 355
during the cure of these epoxy systems have been MY721. In the case of GY260, there could be no
described in our previous papers [68]. Data for the differentiation of E0 and E1 groups, since the epoxy
formation of tertiary amine were obtained using the molecule is difunctional, with well separated
mass balance assumption below, as described by St oxirane groups. Therefore a modification to eq. (1)
John and George [1012]. was made in which the total epoxy concentration
was used for the hydroxyl catalyzed reaction,
PA0 PAt SAt TAt instead of being divided into the E0 and E1
components. This modified equation is shown in
where [PA]t, [SA]t and [TA]t are the concentrations eq. (2).
of primary, secondary and tertiary amine at time t,
respectively, and [PA]0 is the initial concentration d[PA]/dt k1 [EP][PA]2 k2 [PA][EP][OH] 2
of primary amine.
The experimental data were analyzed for the Equation (2) was used to fit the rate curves for
kinetic models using the least squares curve fitting primary amine consumption versus time for the
routine found in SigmaPlot for Windows version GY260/DDS/DDM systems. However, the fit
2.0, which is based on the MarquardtLevenberg afforded by this model was found to be poor for
algorithm [8]. In each case the fitted data were both the thermal and microwave processes. The
checked for the presence of a global minimum, so extent of the deviations between fitted and raw rate
that the values of the kinetic rate parameters curves for amine consumption using this model, in
obtained from the fit were best values. fact, suggested that the proposed mechanistic
scheme was incorrect.
RESULTS AND DISCUSSION W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W Kozielski et al. [12] have reported a detailed NIR
study of the cure kinetics of DDS with two
Industrial epoxy resins based on the DGEBA industrial epoxy resins, Dow Tactix 742 and Shell
structure have been widely studied due to their 1153, which have similar structures to the DGEBA
applicability to composite materials. Significantly, resin studied here. In their NIR study of the
the DGEBA/DDS system has been the system most reaction kinetics of the thermal cure of these resin
commonly studied in the past by workers compar- systems, they fitted the primary amine rate curve
ing thermal and microwave cure kinetics. In using an equation of the form of eq. (2), but with an
this respect, it is the system for which most of additional term to account for the possibility of
the disagreement regarding differences between tertiary amine catalysis, as shown in eq. (3).
microwave and thermal cure kinetics has been
reported. Hedrick et al. [3, 16] and Wei et al. [17] d[PA]/dt k1 PA2 EP k2 [PA][EP][OH]
3
both found a microwave induced enhancement of k3 [PA][EP][TA]
the cure reaction of this system, while Mijovic et al.
[4] observed a slight microwave induced retarda-
tion or no change [18, 19] when compared with the
conventional curing process. The objective of GY260 Cured with DDS
studying this system was to directly compare the
thermal and microwave cure processes and to Figure 1 shows selected FT-NIR spectra obtained
derive a model to explain the amine cure kinetics. during the course of the cure of GY260/DDS 24.9
We have recently reported a study of the wt% at 423 K. The peaks at 5067 cm 1 arise from the
kinetics and mechanisms of the thermal and primary amine groups, at 6060 cm 1 from the
microwave cure of the tetraglycidyldiaminodiphe- oxirane groups, at 6678 cm 1 from both the primary
nyl methane (TGDDM) epoxy resin, Ciba Giegy amine and secondary amine groups and from
MY721, with DDM and DDS [8]. The modeling of 7000 cm 1 from the hydroxyl groups. The concen-
the consumption of primary amine in these resin tration of the primary amine during the cure can be
systems was based in the use of eq. (1), which was conveniently obtained from the band at 5067 cm 1.
introduced by St John and George [11, 15] Figure 2 shows a typical curve for the time
dependence of the primary amine concentration for
the cure of GY260/DDS 24.9 wt% at 393 C obtained
d[PA]/dt k1 [EP][PA]2 k2a [PA][E0][OH] from the in situ monitoring of the FT-NIR spectrum.
1
k2b [PA][E1][OH] Differentiation of this curve with respect to time
then provides information about the rate of
where [E0] and [E1] represent the concentrations of consumption of primary amine across the cure
epoxy groups in TGDDM which are adjacent to process. Typical curves showing the rate of primary
unreacted (E0) and reacted (E1) epoxy groups, amine consumption as a function of time and
respectively. This recognizes the difference in the primary amine concentration, obtained from the
reactivity of these groups due to the intramolecular data in Fig. 2, are shown in Fig. 3(A) and 3(B),
catalytic effect of the adjacent hydroxyl group (OH) respectively.
in the case of the E1 functional groups. Information with respect to the rate of con-
The exact form of the model of St John and sumption of primary amine was then tested against
George could not be applied to the GY260 resin, the proposed rate equations for various models.
due to the structural differences between it and Samples of the quality of fit afforded by eq. (3) to
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
356 / Hill et al.
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
Epoxy-amine Resin Systems / 357
between the two process, as demonstrated in Fig. 6 the cure temperatures of the two samples which are
for the cure at 393 K. Again, similar observations cured in separate experiments. The good agree-
were reported for the MY721/DDS systems [8]. The ment between the two curves is indicative of there
small differences between the two curves in Fig. 6 being no fundamental difference between the
arise from unavoidable small differences between reaction mechanism for the two cure methods;
TABLE 1. Kinetic Rate Parameters for the Thermal Cure of the GY260/DDS 24.9 wt% System at 403, 423 and 443 K,
Together with the Corresponding Data for the Shell 1153/DDS 22 wt% System Reported by Kozielski et al.[12]
Rate parameter
(kg2 mol 2 min 1) GY 260 (403 K) GY260 (423 K) 1153 (423 K) GY260 (443 K) 1153 (443 K)
4 3 3 3 3
k1 6.56 10 1.77 10 2.51 10 1.08 10 6.80 10
3 3 2 2 2
k2 3.49 10 6.66 10 1.23 10 2.23 10 3.02 10
1 1 2 2 2
k3 1.09 10 1.59 10 4.18 10 5.20 10 1.98 10
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
358 / Hill et al.
FIGURE 6. The rate of consumption of primary amine with FIGURE 7. Arrhenius plots for the cure of GY260/DDS
primary amine concentration for GY260/DDS 24.9 wt% at 24.9 wt% by thermal (&) and microwave (*) cure.
393 K for thermal (&) and microwave (*) cure. Line shows
tted curve.
We have previously shown that for the thermal
and the microwave cure of the MY721/DDS system
namely conventional thermal cure and microwave [8], the kinetics for the reaction of the secondary
cure, across the whole range of cure temperatures amine can be modeled using a relationship based
studied. on the model proposed by St John and George
The apparent activation energy for the primary [12, 15], given in eq. (4).
amine reaction was measured by plotting the
maximum rate of cure versus the reciprocal of the
d[SA] d[TA] k4 [SA][EP][OH]
temperature, as shown in Fig. 7 for the thermal and
microwave cure processes. The activation energies dt dt 1 expC 1
4
obtained for the two cure methods were in good k5 [SA][EP][TA]
agreement, as demonstrated in Table 2, and also
1 expC 2
agreed with values reported by Kozielski et al. [12]
which were in the range 5973 kJ mol 1.
The DGEBA/DDS system has been the most In eq. (4), k4 and k5 are the rate parameters for the
common of those reported in the literature with reactions catalyzed by hydroxyl group and tertiary
respect to comparisons between microwave and amine, respectively, a is the degree of cure, a1 and
thermal cure kinetics. Some of this literature has a2 are the critical extents of cure for hydroxyl group
indicated that microwave radiation is able to and tertiary amine catalyzed reactions, respec-
significantly accelerate the reactivity of this tively, and represent the conversions above which
epoxy-amine system over conventional curing diffusion begins to become important and C is a
techniques, as discussed earlier. In a study of this fitted constant related to the degree of diffusion
system Marand et al. [5] reported that the reactiv- control. In the curve fit, a1 and a2 were constrained
ities of the secondary amine- and primary amine- to the range 01 and k4 and k5 were constrained to
epoxy were approximately the same for microwave positive values. C was left unconstrained.
cure process. In the light of the previous studies, it Figure 8(A) and 8(B) show the rate curves of
is important to study closely the secondary amine- tertiary amine formation versus time, and the
epoxy reaction and to compare its rate parameters corresponding curve fits obtained from eq. (4), for
with those of the primary amine-epoxy reaction. the thermal and microwave cure reactions of
GY260/24.9% DDS at 443 K. The same data plotted
against primary amine concentration are shown in
Fig. 9. These plots show the level of agreement
TABLE 2. Arrhenius Parameters for the GY260/DDS
24.9 wt% System for Thermal and Microwave Cure between the experimental data and the model.
Processes Although there are some small differences, most
likely due to uncertainties in the calculation of the
Parameter Thermal Microwave tertiary amine concentration, the agreement is
good. Certainly, the discrepancies observed be-
1
Ea (kJ mol ) 69.0 65.7 tween the rates of tertiary amine formation in the
1 1
Ln (Rate/mol kg min ) 16.5 15.5 thermal and microwave cure reactions of this resin
system are not sufficiently large as to indicate a
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
Epoxy-amine Resin Systems / 359
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
360 / Hill et al.
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
Epoxy-amine Resin Systems / 361
CONCLUSIONS W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W
The kinetics of the polymerizations of GY260 with
DDS and DDM have been studied. The rates of the
reactions of primary amine and secondary amine
have been found to be similar for microwave and
thermal cure processes. Best values of the kinetic
rate parameters for the amine reactions have been
determined based upon the model proposed by St
John and George [11, 15] for the MY721/DDM
system. The kinetic parameters revealed no evi-
dence for any specific effects of microwave radia-
tion on either the primary amine or secondary
amine reactions. The GY260/DDS and GY260/
DDM systems both displayed evidence for catalysis
by tertiary amine and a negative substitution effect
for the secondary amine reactions, with the reduc-
tion in the rate constants of the secondary amines to
about 20% of the values of the primary amine. The
activation energy for the polymerization of the
GY260/DDS system was determined, being in the
range 6669 kJ mol 1 and within experimental
uncertainty, the same for both cure processes.
ACKNOWLEDGEMENTS W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W
The financial support for this project was obtained
from the Australian Research Council and the
FIGURE 13. The rate of formation of tertiary amine versus University of Queensland.
time for GY260/DDM 20.94 wt% at 373 K for (A) thermal
and (B) microwave cure. Line shows tted curve. REFERENCES W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W
1. Metaxas AC, Meredith RJ. Industrial Microwave
Heating. Peter Peregrinus Publishers: London, 1983;
acceleration in the reaction rate for this epoxy- 305.
amine system. 2. Cabibsky I. Symp. Proc. Matl. Res. Soc. 1988; 124: 17.
3. Hedrick JC, Lewis DA, Ward TC, McGrath JE. Polym.
The rate of formation of tertiary amine groups Prepr. 1988; 29: 363.
versus time for the thermal and microwave cure 4. Mijovic J, Fishbain A, Wijaya J. Macromolecules 1992;
reactions of GY260/DDM 20.94 wt% at 373 K, along 25: 986.
with their corresponding fits using eq. (4), are 5. Marand E, Baker KR, Graybeal JD. Macromolecules
shown in Fig. 13(A) and 13(B). The deviation 1992; 25: 2243.
between the fitted curves and the experimental 6. George GA, Cole-Clarke P, St John NA, Friend G. J.
data was larger than observed when fitting the rate Appl. Polym. Sci. 1991; 42: 643.
curves for primary amine consumption, which is 7. Rogers DG, Marand E, Hill DJT, George GA. High
attributed to the greater level of uncertainty Perform. Polym. 1999; 11: 27.
involved in calculating the concentration of tertiary 8. Rogers DG, Hill DJT, George GA. Polym. Adv. Technol.
2001; 12: 169.
amine groups. The shapes of the curves are 9. Rogers DG, Bialkowski ME, Hill DJT, George GA.
fundamentally alike, which suggests that the High Perform. Polym. 1998; 10: 341.
mechanisms of reactions occurring during thermal 10. St John NA, George GA, Cole-Clarke PA, Mackay
and microwave cure of this epoxy-amine system ME, Halley PJ. High Perform. Polym. 1993; 5: 21.
are identical. There is no evidence in these plots to 11. St John NA, George GA. Progr. Polym. Sci. 1994; 29:
support the proposal that the secondary amine- 755.
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)
362 / Hill et al.
12. Kozielski KA, George GA, St John NA, Billingham 17. Wei J, Hawley MC, De Long JD. Polym. Eng. Sci. 1993;
NC. High Perform. Polym. 1994; 6: 263. 33: 1132.
13. George GA, Cole-Clarke PA, St John NA, Friend G. J. 18. Mijovic J, Andjelic S, Winnie Yee GF, Bellucci F,
Appl. Polym. Sci. 1991; 42: 643. Nicholas L. Macromolecules 1995; 28: 2797.
14. George GA, Cole-Clarke PA, St John NA. Materials 19. Mijovic J, Corso WV, Nicholas L, d'Ambrosio G.
Forum 1990; 14: 203. Polym. Adv. Technol. 1998; 9: 231.
15. St John NA, George GA. Polymer 1992; 33: 2679. 20. Beldjouri N, Bouazizi A, Douibi D, Gourdenne A.
16. Hedrick JC, Lewis DA, Lyle GD, Wu SD, Ward TC, Eur. Polym. J. 1988; 24: 53.
McGrath JE. Polym. Mat. Sci. Eng. 1989; 60: 438.
Copyright 2002 John Wiley & Sons, Ltd. Polym. Adv. Technol., 13, 353362 (2002)