Bull. Mater. Sci., Vol. 36, No. 7, December 2013, pp. 1217–1224.


c Indian Academy of Sciences.

Influence of curing agents on gelation and exotherm behaviour of an

unsaturated polyester resin

RAGHU RAJA PANDIYAN KUPPUSAMY† and SWATI NEOGI∗

Department of Chemical Engineering, Indian Institute of Technology, Kharagpur 721 302, India
† Birla Institute of Technology, Mesra, Ranchi 835 215, India

MS received 19 April 2012; revised 24 February 2013

Abstract. A judicious choice of curing agents such as initiator and promoter and their ratio to the resin can avoid
reduced gel-time and shortened exothermic reactions in applications such as liquid composite moulding processes. In
this study, effects of different ratio of initiator and promoter to the unsaturated polyester resin on curing of the resin
were investigated by measuring gel-time and peak exotherm using ASTM D2471 standards. Methyl ethyl ketone
peroxide (MEKP) was used as an initiator and a cobalt salt was employed as an accelerator for the free radical
polymerization of curing resin at ambient temperatures. It was observed that the resin gelation starts closely with
the initial rise in exotherm temperature and time of gelation decreases with the increase in initiator or accelerator
volume proportions. It was also found that the exotherm-peak and rate of temperature rise indicating that the curing
rate increases with the initiator or accelerator proportions also increased. A nonlinear regression analysis of all gel-
time and cure data were performed to quantify the dependence of curing parameters on the volume proportions of
accelerator and initiator. Thus, for this polymerization initiation system, the gel-time and cure parameters can be
predicted for any initiator and cobalt levels within the ranges studied.

Keywords. Unsaturated polyester resin; resin curing; gel time; exotherm behaviour; peroxide initiator.

1. Introduction transferred from liquid state into a rigid cross-linked mole-

cular structure that becomes insoluble and infusible.
Resin transfer moulding involves different phenomena such Physically, curing of resin is the overall transformation
as resin flow, heat transfer and polymerization reactions from liquid to gel, then from gel to rigid solid as a result of
simultaneously. The resin polymerization reaction leads to cross linking with a measurable exothermic effect. Gel time
phase transformation from viscous liquid to rigid solid with marks the formation of an infinite molecular weight 3 D net-
an exothermal effect. The gel time which marks the onset of work, which makes sudden irreversible transformation from
gelation or curing reaction associated with a sharp increase a liquid resin to a non-flowing gel. Therefore, all moulding
in viscosity of the resin system constitutes a crucial para- operations including resin application on to the reinforce-
meter for the mould fill time. The required component geo- ment fibre has to be completed before the point of gela-
metry needs to be filled well before the resin gels prior to tion (Agrawal 1988; Rodriguez 1991; Shailesh and James
the initiation of cure cycles. Hence, the prediction of resin 1991; Yang and Laurent 1991; Malkin et al 1997; Laza et al
gel times and exothermal behaviour enhance complete RTM 1998; Cook et al 2001; Ishida and Douglas 2001; Vilas et al
process cycle to manufacture a composite part. The magni- 2001; Abadie et al 2002; Rimduist and Ishida 2002; Han
tude of gel times and the exotherm behaviour for a parti- 2004; Lionetto et al 2004; Devillard et al 2005; Hossein
cular resin system can be altered with the ratio of initiators et al 2005; Xing et al 2007). The nature of exotherm raised
and promoters to the resin and the elevation of applied pro- from the exothermic effect of curing resin determines volume
cess temperatures. Unsaturated polyester (UP) resin is one shrinkage, warping, residual stress, degradation of polymer,
of the most popular thermoset matrices used in advanced smoke, resin cracking, etc. Hence, to achieve good quality
polymeric composite structures because of its optimized pro- product, the resin gelation and curing reaction has to occur
perties and cost parameters, good processing characteristics, in a controllable manner (Agrawal 1988; Rodriguez 1991;
specific physical properties, moderate price and its ability to Cook et al 2001; Abadie et al 2002; Hossein et al 2005).
cure under normal temperatures and pressures during ther- The phase transformations and the heat evolving nature of
mal and compression cycles makes it an attractive resin can- curing resin can be altered with the mixing proportions of
didate. The curing of UP resin is a free radical polymeriza- curing agents. Hence, a judicious choice of initiator and
tion, i.e. highly exothermic in nature in which the resin is promoter proportions can avoid reduced gel-time and short
time exothermic reactions (Rodriguez 1991; Shailesh and
James 1991; Yang and Laurent 1991; Laza et al 1998; Cook
∗ Author for correspondence (swati@che.iitkgp.ernet.in) et al 2001; Vilas et al 2001; Abadie et al 2002; Han 2004;

1217
1218 Raghu Raja Pandiyan Kuppusamy and Swati Neogi

Figure 1. Measured exotherm of unsaturated polyester resin system at various catalyst and accelerator
volume proportion.

Table 1. Effect of accelerator and MEKP volume ratio on gel time.

2·5 2 1·5 1 0·5

Accelerator (phr) Gel time (S)

MEKP (phr) 2·5 227 238 249 384 1151

2 244 290 360 416 1213
1·5 248 305 372 448 2705
1 754 975 1210 1655 6422
0·5 1564 1776 2467 5028 –

Devillard et al 2005; Hossein et al 2005; Penczek et al 2005; inhibitor and applied cure temperature on gelation and
Xing et al 2007). exotherm behaviour (Rodriguez 1991; Shailesh and James
Experimental techniques such as the measurement of flow 1991; Yang and Laurent 1991; Cook et al 2001; Vilas et al
cessation time (as per ASTM D2471 test) (Agrawal 1988; 2001; Abadie et al 2002; Han 2004; Devillard et al 2005;
Rodriguez 1991; Cook et al 2001; Abadie et al 2002; Hossein et al 2005; Xing et al 2007; Laza et al 1998). There
Hossein et al 2005; Xing et al 2007), the measurement have been a few more detailed analyses of the effect of the
of infinity shear viscosity time (viscometers and rheome- proportions of MEKP and cobalt octoate on the gel time and
ters) (Yang and Laurent 1991; De La Caba et al 1997; curing behaviour of unsaturated polyester resins (Rodriguez
Malkin et al 1997; Laza et al 1998; Vilas et al 2001; 1991; Cook et al 2001; Vilas et al 2001; Abadie et al 2002;
Abadie et al 2002; Rimduist and Ishida 2002; Lionetto et Ling et al 2004; Lionetto et al 2004). In general, the mod-
al 2004; Lionetto and Maffezzoli 2008) and the measure- eling of gel time and exotherm parameters as a function of
ment appearance of equilibrium modulus time (dynamic levels of initiator, promoter, inhibitor and temperature can
mechanical analysers) (Shailesh and James 1991; Malkin be classified into two categories viz., mechanistic models
et al 1997; Ishida and Douglas 2001; Vilas et al 2001; and the phenomenological or empirical models. The mech-
Rimduist and Ishida 2002; Lionetto et al 2004) have been anistic models attempt to quantify the balance of chemical
utilized by many researchers to detect the time of resin species taking part in the cure reaction by using a mathe-
gelation. The measurements of resin cure exotherm have matical function that relate the gel time and other exotherm
been achieved by the use of thermocouples (as per ASTM parameters to concentrations of curing agents, inhibitors and
D2471 test) (Agrawal 1988; Rodriguez 1991; Cook et al applied cure temperatures (Agrawal 1988; Ramis and Salla
2001; Abadie et al 2002; Hossein et al 2005). Researchers 1999; Cook et al 2001; Han 2004). Phenomenological or
have qualitatively correlated the effect of initiator, promoter, empirical models are mathematical expressions which are
 Influence of curing agents on unsaturated polyester resin 1219

Figure 2. Gel time as a function of catalyst and accelerator volume proportions.

Table 2. Effect of accelerator and MEKP volume ratio on peak exotherm time.

2·5 2 1·5 1 0·5

Accelerator (phr) Peak exotherm time (S)

MEKP (phr) 2·5 750 765 825 1055 2055

2 780 840 915 1020 2220
1·5 810 870 1005 2040 4155
1 1410 1665 2055 2730 9765
0·5 2550 2865 4170 8265 –

framed to fit the dependence of gel time and other cure volatile content. A thermal source or reduction–oxidation
parameters in terms of curing aids proportions (Shailesh and (redox) initiation system is often used for the curing of UP
James 1991; Devillard et al 2005) and these empirical models resin (Agrawal 1988; Rodriguez 1991; Shailesh and James
are easier to handle in comparison with mechanistic models. 1991; Yang and Laurent 1991; Ramis and Salla 1999; Cook
The main objective of this work is to quantify and cor- et al 2001; Vilas et al 2001; Lionetto et al 2004; Hossein
relate the effects of varying initiator and promoter volume et al 2005). For applications such as resin transfer mould-
levels on gel time and exotherm behaviour of an unsatu- ing (RTM), UP resins are usually cured at ambient tempera-
rated polyester resin using experimental and empirical curve tures through the use of a redox initiation system composed
fitting technique. This article is structured as follows: first, of initiators/catalysts and promoters/accelerators (Shailesh
the exotherm obtained from the experiments on varying cur- and James 1991; Cook et al 2001; Vilas et al 2001; Hossein
ing aids concentrations are presented. Then the shifts of gel et al 2005). Methyl ethyl ketone peroxide (MEKP) is one of
time with change of initiator and accelerator proportions the most frequently used peroxide based initiator (Rodriguez
have been correlated using nonlinear regression analysis. 1991; Yang and Laurent 1991; Cook et al 2001; Vilas et al
Finally, the exotherm parameters such as peak exotherm 2001; Abadie et al 2002; Lionetto et al 2004; Hossein et al
temperature, peak exotherm time and rate of rise in tem- 2005). Cobalt salt such as cobalt octoate is employed as the
perature are profiled with the levels of cure ingredients accelerator to decompose peroxide based initiators in induc-
dependence. ing free radical co-polymerization during resin curing pro-
cess (Rodriguez 1991; Yang and Laurent 1991; Cook et al
2001; Vilas et al 2001; Abadie et al 2002; Lionetto et al
2. Materials 2004). In this work, methyl ethyl ketone peroxide along with
A commercially available medium reactive unsaturated cobalt octoate was used as redox initiation system for curing
polyester resin with a grade name ISO RTM was used in this the unsaturated polyester resin used. MEKP has 8% active
study. It is tailored with a mixture of isophthalic acid, maleic oxygen content whereas cobalt octoate has 3% cobalt con-
anhydride and propylene glycol to an acid value of 15 ± 2 tent. UP resin and curing agents were supplied by Mechemco
mg KOH/g. It has the density of 1·080 g/cc with 43% (w/w) Resins Pvt. Ltd., Mumbai, India.
1220 Raghu Raja Pandiyan Kuppusamy and Swati Neogi

Figure 3. Peak exotherm time as a function of catalyst and accelerator volume

proportions.

Table 3. Effect of accelerator and MEKP volume ratio on peak exotherm temperature.

2·5 2 1·5 1 0·5

Accelerator (phr) Peak exotherm temperature (◦ C)

MEKP (phr) 2·5 167 167 168 169 169

2 159 160 160 160 160
1·5 157 156 156 156 155
1 152 152 151 152 150
0·5 147 147 146 145 –

3. Experimental the reacting material no longer adhered to the end of a clean

probe, it was recorded as gel time. The temperature with time
The medium reactive unsaturated polyester resin was mixed was recorded continuously until the temperature started to
with MEKP which acts as catalyst (initiator) and cobalt drop. Highest temperature reached was recorded as the peak
octoate which acts as accelerator (promoter) and allowed to exothermic temperature and the elapsed time from the start
cure. During curing, the important curing characteristics of of mixing to reach the maximum temperature was recorded
this resin system such as gel time, peak exothermic tempera- as the peak exothermic time. The volume shrinkage due to
ture and peak exothermic time were measured at room tem- resin cure and exotherm heat was measured using the change
perature according to the specification ASTM D2471. Peak in volume between initial reacting resin volume and the final
exotherm of the resin system was measured using various resin cast. The measurement of volume shrinkage helps to
volume proportions of MEKP and cobalt octoate, which were predict the shape distortions of a composite structure during
performed to quantify effect of different concentration le- resin cure. All the experiments were conducted at an ambient
vels of initiator and promoter. Accelerator volume was var- temperature of 25 ◦ C.
ied from 0·5 to 2·5 mL and the initiator volume was varied
from 0·5 to 2·5 mL/100 mL of resin. Initially, 100 mL of the
resin sample was added with desired proportion of acceler- 4. Empirical gel time and exotherm model
ator and the mix was agitated slowly with a stirrer avoiding
air entrapment. The timer was started when the initiator was Gel time and exotherm data obtained from the performed
added. A thermocouple was inserted into the geometric cen- experiments were correlated to the volume proportions of ini-
tre of the reacting mass and the temperature with time was tiator and promoter. A nonlinear regression analysis of all
recorded till the end of experiment. An applicator stick was gel time and exotherm data was performed using datafit soft-
inserted in the centre of the reacting mass for every 15 s ware to quantify the variation of these parameters with the
and the flowability of the reacting mass was checked. The volume proportions of accelerator and initiator. This correla-
time elapsed from the start of mixing of initiator and when tion can be used as an important tool to predict the gel time,
 Influence of curing agents on unsaturated polyester resin 1221

Figure 4. Peak exothermic temperatures as a function of catalyst and accelerator

volume proportions.

Table 4. Effect of accelerator and MEKP volume ratio on rate of rise in temperature.

2·5 2 1·5 1 0·5

Accelerator (phr) Rate of temperature rise (◦ C/min)

MEKP (phr) 2·5 15·22 14·42 13·8 11·83 9·75

2 12 11·56 11·3 10·83 9·45
1·5 11·04 10·96 10·72 9·16 5·55
1 10·38 9·76 8·34 6·1 2·7
0·5 5·32 3·16 2·19 1·66 –

peak exothermic temperature and peak exotherm time for tabulated in the subsequent sections. In all subsequent tabu-
medium reactive unsaturated polyester resin, MEKP initiator lations, the varying concentrations of accelerator and initia-
and cobalt octoate accelerator. tor to the resin parts are presented as “phr”, which means
the parts of accelerator and catalyst to hundred parts of
resin.
5. Results and discussion
5.2 Effect of accelerator and MEKP volume ratio on gel
5.1 Effect of initiator and accelerator volume ratios time
on resin gel time and exotherm
The empirical gel time data obtained from the measured
Figure 1 shows change in temperature with time during cur- exotherm are given in table 1. Figure 2 shows gel times
ing of the resin system at various initiator and accelerator measured for varying proportions of initiator and acceler-
volume proportions. These time–temperature curves show ator from 0·5 to 2·5 mL/100 mL of resin. From the same
the nature of rise in temperature and the minimum time figure, it is clearly observed that, for a particular initiator
required for the resin cure with specific volume proportion of proportion, more the accelerator part, lesser the gel time.
initiator and accelerator. From the measured exotherm data And for a particular accelerator proportion, more the initia-
as shown in figure 1, it is evident that the use of higher tor part, lesser the gel time. This fact is chiefly attributed to
proportions of initiator and accelerator yields reduced cure the increase in decomposition of peroxides due to increase
cycle time. Similar exotherm graphs can be produced for in initiator concentration to produce highly reactive free rad-
initiator proportion varying from 1·5 to 0·5 mL with acce- icals which in turn react with polymer molecules to pro-
lerator proportion varying from 2·5 to 0·5 mL/100 mL of duce cross links. Also, the gel time difference appears to be
the resin. The gel time and other curing parameters such as approximately low with high levels of accelerator and initia-
peak exotherm, peak exotherm time, rate of rise in tempe- tor. Using nonlinear regression analysis, gel times were fit as
rature and volume shrinkage during curing reaction can a function of initiator and accelerator volume proportions to
be determined from the measured exotherm, which are the form as given by (1) and it is seen that the experimental
1222 Raghu Raja Pandiyan Kuppusamy and Swati Neogi

Figure 5. Rate of rise in temperature during reaction as a function of catalyst and

accelerator volume proportions.

Table 5. Effect of accelerator and MEKP volume ratio on volume shrinkage.

2·5 2 1·5 1 0·5

Accelerator (phr) Volume shrinkage (%)

MEKP (phr) 2·5 10·4 10·2 10 7·6 6

2 8·6 8·2 8·4 7·6 7
1·5 8·5 8·4 8·6 7·6 7
1 7·8 7·6 8 7 6
0·5 7 6·8 6·8 6 –

dependence fits accurately well to the empirically modelled exotherm time data correlating accelerator and initiator le-
equation, as shown in figure 2: vels was performed and a comparison of actual and mod-
elled results are given in figure 3. The empirical form given
G = 1694 ∗ A−1·785 ∗ C −1·682 , (1) in (2) gives a better definition in relating peak exotherm time
to the volume proportions of the curing aids:
where G the gel time in seconds, A the accelerator volume in
mL and C the initiator volume in mL. P Et = 2832 ∗ A−1·534 ∗ C −1·612 , (2)

5.3 Effect of accelerator and MEKP Volume ratio on peak where PEt is the peak exotherm time in seconds.
exotherm time
5.4 Effect of accelerator and MEKP volume ratio on peak
The empirical peak exotherm time data obtained from the exotherm temperature
measured exotherm are given in table 2. Figure 3 shows
dependence of peak exotherm time measured on varying The empirical peak exotherm temperature data obtained from
proportions of initiator and accelerator from 0·5 to 2·5 mL the measured exotherm are given in table 3. The effect of
per 100 mL of resin. From the same figure, it is evident that, varying the levels of initiator and promoter on the exotherm
for a particular initiator proportion, more the accelerator is displayed in figure 4. From the same figure, it can be
part, lesser the peak exotherm time. And for a particular inferred that for varying initiator level, the peak exotherm is
accelerator proportion, more the initiator part, lesser the peak unaffected on cobalt salt level from the accelerator. This may
exotherm time. Comparing figures 2 and 3, it can be noticed be caused by the non-consumption of metal salts during cur-
that the gel time and the time to peak exotherm shows simi- ing reaction, since the consumed metal salts for peroxides
lar behaviour with respect to the variations of accelerator decomposition can be regenerated with the reaction progres-
and initiator levels. A nonlinear regression analysis of peak sion. It can also be noted that, for a particular accelerator
 Influence of curing agents on unsaturated polyester resin 1223

Figure 6. Volume shrinkage during curing reaction as a function of catalyst and

accelerator volume proportions.

proportion, lesser the initiator part, lower the peak exother- 5.6 Effect of accelerator and MEKP volume ratio
mic temperature. This trend is due to the depletion of ini- on volume shrinkage
tiator at the lower volume levels before full reaction. The
peak exothermic temperature varies from 169 to 145 ◦ C with The empirical cure volume shrinkage data obtained from
the initiator volume differing from 2·5 to 0·5 mL/100 mL the measured exotherms are given in table 5. The impact of
of resin. The empirical form given in (3) gives a better cobalt octoate and methyl ethyl ketone levels on cure volu-
prediction of peak exothermic temperature using volume me shrinkage of a medium reactive unsaturated polyester
proportions of the curing aids: resin is depicted in figure 6. As can be seen from the same
figure, the volume shrinkage during curing reaction varies
PET = 152·5 ∗ A−0·003 ∗ C −0·083 , (3) from 6 to 11% for the volume proportions of accelerator and
initiator used. It may be mentioned that the volume shrink-
age increases notably with the increase in the initiator in
where PET is the peak exothermic temperature in ◦ C. comparison with the accelerator part. The volume shrinkage
can be correlated to the volumes of accelerator and initiator
5.5 Effect of accelerator and MEKP volume ratio on rate by fitting the experimental data to an empirical form using
of rise in temperature nonlinear regression analysis as given by (5)

The empirical rate of rise in temperature data obtained from VS = 6·82 ∗ A0·204 ∗ C 0·196 , (5)
the measured exotherm are given in table 4. The plot rate of where VS is the volume shrinkage in %.
rise in temperature after resin gelation with respect to vary-
ing initiator and accelerator volume levels is portrayed in
figure 5. It can be observed from the same figure that there 6. Conclusions
exists a quick rise in temperature with the use of higher pro-
portions of initiator and accelerator volumes. The reason for The impact of cobalt octoate and methyl ethyl ketone volume
this observation is a higher decomposition rate of initiator ratios on gelation time and exotherm behaviour of an unsatu-
with higher proportion of accelerator. The effect of accelera- rated polyester resin were studied. The resin gelation was
tor and initiator levels on the rate of rise in temperature dur- found to correspond with the onset of temperature rise during
ing resin curing can be defined with an empirical form given resin curing and also the gelation time was found to decrease
by (4). Figure 5 shows comparison of the experimentally with increase in concentrations of accelerator and initiator.
obtained temperature rise per unit time with that predicted by Results showed that the peak exotherm and rate of rise in
the empirical model: temperature increase with increasing curing aids as a cause of
higher decomposition rate of initiator. The use of nonlinear
RTR = 6·17 ∗ A0·31 ∗ C 0·72 , (4) regression analysis helped to correlate the gel time data and
exotherm data with volume proportions of cobalt octoate
where RTR is the rate of rise in temperature ◦ C/min. and methyl ethyl ketone, qualitatively and quantitatively.
1224 Raghu Raja Pandiyan Kuppusamy and Swati Neogi

References Ling Li, Xia Cao and James Lee 2004 Polymer 45 6601
Lionetto F and Maffezzoli A 2008 Adv. Polym. Technol. 27 63
Abadie M J M, Mekhissi K and Burchill P J 2002 J. Appl. Polym. Lionetto F, Rizzo R, Luprano V A M and Maffezzoli A 2004 Mater.
Sci. 84 1146 Sci. Eng. A 370 284
Agrawal J P 1988 Eur. Polym. J. 24 93 Malkin A Ya, Kulichikhin S G, Kerber M L, Gorbunova I Yu and
Cook Wayne D, Lau Michelle, Mehrabi Mansour, Dean Katherine Murashova E A 1997 Polym. Eng. Sci. 37 1322
and Zipper Marcus 2001 Polym. Int. 50 129 Penczek P, Czub P and Pielichowski J 2005 Adv. Polym. Sci. 184 1
De La Caba, Guerrero P, Eceiza A and Mondragon I 1997 Eur. Rimduist S and Ishida H 2002 Rheol. Acta 41 1
Polym. J. 33 19 Rodriguez Ernesto L 1991 Polym. Eng. Sci. 31 1022
Devillard Antoine Laut and Suresh G Advani 2005 Polym. Compos. Salla R and Salla J M 1999 J. Polym. Sci. B: Polym. Phys. 37 751
26 74 Shailesh V Muzumdar and James Lee 1991 Polym. Eng. Sci. 31
Han Tong-chun 2004 J. Univ. Sci. 5 928 1647
Hossein Beheshty M, Hassan Nasiri and Mehdy Vafayan 2005 Vilas J L, Laza J M, Garay M T, Rodriguez M and Leon L M 2001
Iranian Polym. J. 14 990 J. Appl. Polym. Sci. 79 447
Ishida H and Douglas J Allen 2001 J. Appl. Polym. Sci. 79 406 Xing Cheng, Zhang S Y, James Deng and Siqun Wang 2007 J. Appl.
Laza J M, Julian C A, Larrauri E, Rodriguez M and Leon L M 1998 Polym. Sci. 103 1566
Polymer 40 35 Yang Y S and Suspene L 1991 Polym. Eng. Sci. 31 321