Materials and Design 32 (2011) 1129–1137

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Optimization of infrared radiation cure process parameters for glass fiber

reinforced polymer composites
P. Kiran Kumar a,b,⇑, N.V. Raghavendra b, B.K. Sridhara b
a
Department of Mechanical Engineering, B.N.M. Institute of Technology, Bangalore 560 070, India
b
Department of Mechanical Engineering, The National Institute of Engineering, Mysore 570 008, India

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

Article history: Elevated temperature post curing is one of the most critical step in the processing of polymer composites.
Received 8 July 2010 It ensures that the complete cross-linking takes place to produce the targeted properties of composites. In
Accepted 1 November 2010 this work infrared radiation (IR) post curing process for glass fiber reinforced polymer composite lami-
Available online 5 November 2010
nates is studied as an alternative to conventional thermal cure. Distance from the IR source, curing sche-
dule and volume of the composite were selected as the IR cure parameters for optimization. Design of
Keywords: experiments (DOE) approach was adopted for conducting the experiments. Tensile strength and flexural
Glass fiber reinforced polymer composite
strength of the composite laminate were the responses measured to select the final cure parameters.
Infrared curing
Central composite design
Analysis of variance (ANOVA), surface plots and contour plots clearly demonstrate that the distance from
Tensile strength the IR source and volume of the composite contribute nearly 70% to the response functions. This estab-
Flexural strength lishes that polymer composites cured using IR technique can achieve the same properties using only 25%
A. Polymer matrix composites of the total time compared to that of conventional thermal curing.
B. Laminates Ó 2010 Elsevier Ltd. All rights reserved.
E. Mechanical properties

1. Introduction temperature variation during the cure for the epoxy resin [3].
Ramakrishnan et al., have suggested reducing the curing time by
Fiber reinforced polymer composite finds application from head using internal resistive carbon mats without compromising with
gear to aircraft due to its lightweight, higher strength to weight ra- the composite quality [4]. Many researchers have worked on alter-
tio and adaptability to customize the composite according to the native methods of curing to overcome the disadvantages of con-
required strength and functionality for which it is employed. In ventional method, that is the hot air curing of composites,
the processing of polymer composites curing plays a vital role. Cur- known as the thermal curing. Different methods of curing are in
ing is the process of conversion of liquid resin into hard solid struc- practice apart from thermal curing such as microwave, radiofre-
ture and it takes place at molecular level. Curing is one of the quency, ultraviolet and infrared radiation cure. Rao et al., have
significant and complex processes which requires considerable worked on the microwave curing of composites. There is drastic
attention as it consumes lot of time and energy. Models developed reduction in curing time with higher mechanical properties of
by Loos and Springer for curing, stressed the importance of reduc- the composite [5]. Sabit and Arumugham have successfully utilized
ing the cure time and explained its importance on the final ultraviolet radiation for curing of composites and stressed the need
strength of the composite [1]. The process of polymerization is for better curing system [6]. Radio frequency (RF) was applied to
the joining of lower molecular weight reacting monomers to form cure the epoxy resin by Gourdenne and it was proved that the con-
a three dimensional network or polymer chain. The effect of cure ventional heating and RF has no structural difference in the
cycle on the final quality of the composite studied by Naji and strength of the composite [7]. Ribeiro et al., have demonstrated
Hoa, has revealed the effect of different cure cycles on varying the electron beam polymerization of epoxy resin. The cure rate in-
thickness and fiber volume fraction of the composite [2]. Zhang creased and the kinetic model was developed based on duration of
et al., have proposed three dimensional finite element models to exposure and degree of cure [8]. It is understood from various
analyze the temperature and degree of cure for epoxy resin. They works reported that improper curing leads to uncured resin
explained that final hardness depends on degree of cure and the patches inside the composite which leads to lower strength of
the composite. Infrared radiation (IR) curing is one of the efficient
⇑ Corresponding author at: Department of Mechanical Engineering, B.N.M.
methods of curing of polymer composites. Belhamra et al., have ex-
Institute of Technology, Bangalore 560 070, India. Tel.: +91 9243105364; fax: +91 plained the technology and applications of infrared heating [9].
080 26710881. Chern et al., in their work reported about IR heating of Hoop
E-mail address: kiran_1975@rediffmail.com (P.K. Kumar). wound cylinders, several process parameters were studied and

0261-3069/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.matdes.2010.11.001
1130 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137

the developed models agreed well with experimental results [10]. selected as the main factors in order to optimize the cure process.
IR heating using lamps was proposed by Labeas et al., in their work Preliminary trials confirmed the effect of main factors selected on
for heating of thermoplastic parts. They studied the effect of vari- the overall quality and strength of the composite laminate. Compo-
ous process parameters which influence the IR heating process, sition of all the laminates and power was maintained constant for
and concluded that as the thickness of the part increases the heat- all the trials during the experimentation process.
ing time also increases [11]. Direct electro-magnetic radiation is a
very good source of energy for processing of polymer matrix com-
3. Experiments
posites. Since polymeric resins are polar in nature, they react to the
electrical and magnetic field components of the EM radiation.
The experiments were conducted based on central composite
Depending on the intensity and the frequency of the incident radi-
design (CCD) approach of design of experiments (DOE). CCD is a
ation the polymer molecules get polarized, in the process generat-
powerful diagnosing experimental tool to study the large number
ing heat. Decker stated in his review article that the photo
of factors. Two or more factors with three or more levels require
polymerization is one of the fastest means of generating three
higher experimentation time and cost in terms of materials, power
dimensional network [12]. Radiation curing finds lot of scope in
and labor. In order to save the same without sacrificing the re-
curing of polymer composites and hence infrared radiation (IR) will
quired quality CCD approach is best suited. CCD is used extensively
be a suitable alternative to conventional curing. IR is part of elec-
for second order response surface models. Several papers have re-
tromagnetic (EM) spectrum, with wavelength in the range of
ported the application of CCD to optimize the process and proper-
106–104 m. Infrared is transmitted in three energy bands. Short
ties of the composites. Suresha and Sridhara [13], Ruijun and Kokta
wave (0.76–2.3 lm) used for complex part shapes. Medium wave
[14] and Onal and Adanur [15] have successfully employed CCD
(2.3–3.3 lm) used for curing coatings on objects. Broad wave (3–
approach to determine the optimum properties and process of
8 lm) used for laminates and packing machines. It is highly suited
composites. The factors, their levels and the range selected are pre-
for FRP composites curing as the technology is proved for curing of
sented in Table 2. The total number of experimental runs presented
paints, coatings, processing of food items and also for industrial
by both coded and actual values of factors along with the central
heating applications. The main advantage of infrared curing is, it
runs as per CCD approach [17–19] is indicated in Table 3.
heats only the composite and not the air present in between the
The thermal cure schedule is based on conventional thermal
heater and the product. The radiation is directly absorbed by the
curing as indicated in Table 4. The curing time schedule has differ-
composite and hence the losses are minimum. It is mainly suited
ent cure schedule of ramping and soaking. Ramping and soaking at
for flat surfaces and the efficiency depends mainly on the infrared
different temperatures is beneficial in achieving improved proper-
absorbance capacity of the material. Different materials have dif-
ties compared to room temperature and single high temperature
ferent absorbance capacity. Very few researchers have worked in
cure. Optimum properties were achieved by curing in steps [16].
the area of IR curing. In this context an attempt is made to develop
Curing process is more uniform and the fiber matrix bonding is im-
optimum IR cure process parameters for glass fiber reinforced
proved. Although at room temperature the epoxy matrix appears
polymer composite laminates.
to be solidified still there is still large amount of un-reacted resin

2. Materials
Table 2
Factors and their levels as per CCD.
The polymer matrix used in this investigation is bifunctional
Factors Levels
diglycidyl ether of bisphenol A (DGEBA) type epoxy resin
(LY556). The curing agent used is 4–40 aminophenylmethylaniline (a) (1) (0) (1) (+a)
(HT972). The resin and the curing agents were procured from M/ Distance from 150 180 225 270 300
s. Huntsman Advanced Materials, USA The reinforcement used is heater (mm)
Curing time (min) 56 93 148 203 239
bidirectional, E-glass fabric with a fiber orientation of 0°/90° and
Volume of the 81,675 120,000 175,692 231,852 270,000
a fabric weight of 200 g/m2. Table 1 presents the properties of laminate (mm3)
the materials used for the preparation of composite laminates.
The fabric layers were pre-heated at 120 °C in an oven so as to re-
move adsorbed moisture, prior to lay-up. Composite laminates
with the weight ratio of 65:35 that is glass fibers to resin ratio Table 3
and the hand-layup method was adopted for preparing the lami- Coded and actual values of factors.
nates. The laminates were then allowed to cure at room tempera-
tc A B C Distance in Curing time Volume in
ture (RT) for 24 h. The fiber weight fraction of all the laminates was mm (A) in min (B) mm3 (C)
maintained at 65 ± 2%. The post curing process was carried out in a
1 1 1 1 180 93 120,000
custom designed infrared curing system having 400  400 stain-
2 +1 1 1 270 93 120,000
less steel chamber mounted with IR heater of 3–8 lm wavelength 3 1 +1 1 180 203 120,000
of 2 kW capacity. The chamber is insulated by ceramic board of 4 +1 +1 1 270 203 120,000
50 mm thick to avoid the heat loss to surroundings. 5 1 1 +1 180 93 231,852
6 +1 1 +1 270 93 231,852
Distance from IR source to laminate maintaining constant
7 1 +1 +1 180 203 231,852
power of 2 kW, IR cure schedule and the volume of composite were 8 +1 +1 +1 270 203 231,852
9 1.682 0 0 150 148 175,692
10 +1.682 0 0 300 148 175,692
Table 1 11 0 1.682 0 225 56 175,692
Properties of materials. 12 0 +1.682 0 225 239 175,692
13 0 0 1.682 225 148 81,675
Materials Density (g cm3) Tensile Young’s
14 0 0 +1.682 225 148 270,000
strength (MPa) modulus (GPa)
15 0 0 0 225 148 175,692
E-glass fibers 2.55 1750 70 16 0 0 0 225 148 175,692
Epoxy 1.25 55 3.5 17 0 0 0 225 148 175,692
 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137 1131

Table 4
Thermal cure schedule.

Temp in °C 28–120 (ramp) 120 (soak) 120–150 (ramp) 150 (soak) 150–28 (ramp) Total time in min
Thermal cure schedule in min. 20 60 06 120 30 236

Table 5
Various steps of IR curing schedule according to CCD.

Temp in °C Factor level indication of curing time

(a) (1) (0) (+1) (+a)
Schedule 1 in min Schedule 2 in min Schedule 3 in min Schedule 4 in min Schedule 5 in min
28–120 (ramp) 10 12 15 18 20
120 (soak) 15 24 38 52 60
120–150 (ramp) 03 04 06 08 09
150 (soak) 20 40 70 100 120
150–28 (ramp) 08 13 19 25 30
Total time in min 56 93 148 203 239

patches with in the composite. During elevated temperature pro- distance of 300 mm, indicating that the distance has significant ef-
cessing these resin patches will take part in the polymerization fect on the tensile strength of the composite. The cure time vs. ten-
process and hence the complete composite is cured which leads sile strength plots shows a saturation point at 93 min, due to
to higher strength of the composite. Polymerization takes place volumetric heating effect. Thus resulting in considerable reduction
slowly as the energy is initiated into the composite. As the infrared in process cycle time. The response of volume of polymer compos-
radiation passes through the composite the radiation energy is ite to tensile strength is constant between 81,000 and
converted to chemical energy which is absorbed by the multifunc- 175,000 mm3, and beyond this point there is drastic reduction in
tional monomer and the cross linking of the molecule begins. This tensile strength. The inference that can be drawn from this plot
means that polymerization starts as in case of thermal curing but is, uniformity and power of infrared radiation is a critical factor.
at a faster rate than thermal cure. Different IR cure schedules fol- Even-though infrared curing is a volumetric heating process, it is
lowed as per CCD are indicated in Table 5. This is derived from a line of sight process, therefore curing beyond the exposed area
thermal cure schedule. The objective is to minimize the curing occurs due to thermal diffusion and not due to IR activation.
time hence max curing time limit was equal to thermal cure The percentage contribution of each factor is calculated by
schedule. dividing the factor coefficient of sum of squares with its total value
Tensile test and flexural test samples were prepared and tested as indicated in Table 8.
according to ASTM D638 and ASTM D790 specifications respec- Fig. 2 represents the normal probability plot of standardized ef-
tively. Tests were conducted using Instron universal testing ma- fects of factors on tensile strength of the composite. The factors
chine with a 100 kN load cell. The crosshead speed was that are far from the normality line indicates nonzero means,
maintained at 2 mm/min for tensile test and at 1.3 mm/min for which have significant effect on the process. The other nonsignifi-
flexural test. Tensile strength is a fiber dominating property. The cant factors are normally distributed where mean and variance are
glass fibers are mainly responsible in achieving this. Flexural zero. The factors to the right of the normality line have positive ef-
strength is based on interfacial strength between the matrix and fect indicating increase in strength with increase in their levels and
the fiber. It is helpful in understanding the bonding of the matrix the factors to the left of the line have negative effect which indi-
with the fibers. Curing improves the interfacial strength and it is cates decrease in strength as their level increases. As observed
highly relevant to study the flexural strength along with the tensile from the plot, distance and volume are significant having negative
properties for determining the optimum cure process parameter.
Microstructure studies was taken up using Scanning Electron
Microscope, only for optimum IR cured sample and compared with Table 6
the thermal cured to analyze the bonding of fibers with matrix. Results of tensile and flexural strength tests.

tc Tensile strength (MPa) Flexural strength (MPa)

4. Results and discussion 1 320.0 432.0
2 289.0 365.0
The experimental results of tensile and flexural tests conducted 3 312.0 409.0
as per CCD experimental plan is present in Table 6. The ANOVA re- 4 318.0 424.0
5 318.0 424.0
sults of tensile strength are presented in Table 7. A confidence limit
6 279.0 310.0
of 95% or P-value less than or equal to 0.05 is considered to decide 7 295.0 377.0
the significance of factors. The factors A, C and AB are significant 8 285.0 333.0
and the factor B has to be included in the analysis due to the 9 323.0 436.0
requirement of model hierarchy as it has interaction effect with 10 280.0 319.0
11 293.0 376.0
factor Onal and Adanur, have successfully employed ANOVA to 12 303.2 379.0
determine the influence of factors for the compression molding 13 310.0 407.0
process [15]. Fig. 1 shows the variation of each factor and the 14 285.0 333.0
corresponding tensile strength response obtained. The tensile 15 310.0 419.0
16 315.0 398.0
strength drops significantly with respect to distance, it is
17 309.0 422.0
320 MPa at a distance of 150 mm and drops to 280 MPa at a
1132 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137

Table 7
Analysis of variance for tensile strength, using adjusted SS for tests.

Source DF Seq SS Adj SS Adj MS F P

A 1 1567.65 1567.65 1567.65 34.39 0.000
B 1 32.77 32.77 32.77 0.72 0.418
AB 1 544.50 544.50 544.50 11.95 0.007
C 1 792.66 792.66 792.66 17.39 0.002
AC 1 72.00 72.00 72.00 1.58 0.240
BC 1 180.50 180.50 180.50 3.96 0.078
ABC 1 8.00 8.00 8.00 0.18 0.685
Error 9 410.24 410.24 45.58
Total 16 3608.32
S = 6.75145 R  Sq = 88.63% R  Sq(adj) = 79.79%

Main Effects Plot (data means) for Tensile strength, MPa

Distance, mm Curing Time, Min

320

310
Mean of Tensile strength, MPa

300

290

280
150 180 225 270 300 56 93 148 203 239

Volume, mm3
320

310

300

290

280
81675 120000 175838 231820 270000

Fig. 1. Main effect plot of factors for tensile strength.

effect. The tensile strength decreases as the distance or volume or where A = (a  225)/45, B = (b  150)48, C = (c  175837.5)/
both increases. The interaction effect of curing time and distance is 55982.5. A, B and C are coded values and a, b and c are actual values
significant with positive effect indicating as their value increases of the factors.
the tensile strength increases.
Regression analysis is performed to verify the nonlinear effects
of the factors and the result is shown in Table 9. It is concluded by
regression analysis that the nonlinear effects are not present hence Normal Probability Plot of the Standardized Effects
further regression analysis is performed for significant factors. The (response is Tensile Strength, MPa, Alpha = .05)
final equation to predict the model is represented by Eq. (1). Fig. 3 99
Effect Type
represents normal probability effect plot of the residuals for tensile Not Significant
strength. The points lie close to the normality line implying that 95
Significant
the errors are distributed normally and the model represented in 90
AB
Eq. (1) is adequate. Fig. 4 shows the plot of fitted value vs. deleted 80
residuals of tensile strength. The points do not form any particular 70
Percent

pattern and hence the model represented in Eq. (1) is adequate to 60

50
predict the response. 40
30 C
Tensile strength ¼ 303  10:7A þ 1:55B  7:62C þ 8:25AB ð1Þ
20
A
10
5
Table 8
Percentage contribution of factors for tensile strength. 1
Source A B AB C Error and others -6 -4 -2 0 2 4

Percentage 43.44 0.90 15.09 21.92 13.58

Standardized Effect
contribution (%)
Fig. 2. Normal probability plot of standardized effects for tensile strength.
 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137 1133

Table 9 Contour Plot of Tensile strength, MPa vs Volume, mm3, Distance, mm

Regression analysis for tensile strength.
260000 Tensile
Predictor Coef SE Coef T P strength,
MPa
240000
Constant 311.095 3.783 82.23 0.000 < 280
A 10.713 1.777 6.03 0.000 280 - 290
220000 290 - 300
B 1.549 1.777 0.87 0.406

Volume, mm3
Distance, mm = 215.940 300 - 310
AB 8.250 2.321 3.55 0.006 200000 Volume, mm3 = 157176
Tensile strength, MPa = 307.312 310 - 320
C 7.618 1.777 4.29 0.002 320 - 330
180000 > 330
AA 2.653 1.955 1.36 0.208
BB 3.854 1.955 1.97 0.080 160000 Hold Values
Curing Time, Min 147.5
CC 4.067 1.955 2.08 0.067
S = 6.56585 R  Sq = 89.2% R  Sq(adj) = 80.9% 140000

120000 Distance, mm = 151.744

Volume, mm3 = 85822.0
Tensile strength, MPa = 332.467
100000
Normal Probability Plot of the Residuals
(response is Tensile Strength, MPa) 150 175 200 225 250 275 300
99
Distance, mm
95
Fig. 6. Contour plot of volume vs. distance for the tensile strength.
90
80
Percent

70
60 Surface plots with tensile strength as response is plotted for vol-
50
40 ume vs. distance and curing time vs. distance as shown in Fig. 5.
30 The surface plot is a plane surface with rising ridges because of ab-
20
sence of curvilinear effects of the factors. The optimum points of
10
5
interest can be obtained and further analysis can be carried out
with contour plot for the same as shown in Fig. 6. Since the re-
1 sponse surface is a plane the contour plot contains inclined straight
-2 -1 0 1 2
lines [19]. These inclined straight lines in the contour plot of vol-
Deleted Residual ume vs. distance indicate, only linear effects are present which af-
fect the final strength of the laminate. Distance of 151 mm and
Fig. 3. Normal probability effect plot of the residuals for tensile strength.
volume of 85,822 mm3 has tensile strength higher compared to
the distance of 215 mm and volume of 15176 mm3 as noticed from
the plot. As the distance or volume or both increases the strength
Residuals Versus the Fitted Values
decreases.
(response is Tensile Strength, MPa)
2.5 Surface plots with tensile strength as response is plotted for
curing time vs. distance as shown in Fig. 7. The curvilinear surface
2.0
of the surface plot is due to the presence of interaction effect of dis-
Deleted Residual

1.5 tance and curing time. The curved contour lines present in the con-
tour plot of curing time vs. distance in Fig. 8 is due to interaction
1.0
effect between these two factors. The optimized region is shown
0.5 in the graph. From Figs. 6 and 8 the tensile strength is maximum
0.0 when distance is 150 mm and volume is 85822 mm3 and curing
time is 57 min.
-0.5
The distance and volume have negative effect on the tensile
-1.0 strength, which may be explained by the IR radiation phenomenon.
That is observance of IR radiation depends on the geometric rela-
280 290 300 310 320 tionship between the IR source and the composite laminate sur-
Fitted Value face. It is directly dependent on the orientation and area of two
emitting surfaces and inversely proportional to the distance.
Fig. 4. Residuals vs. the fitted values for the tensile strength. Therefore amount of energy received by IR increases as the

Fig. 5. Surface plot of volume vs. distance for the tensile strength.
1134 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137

Fig. 7. Surface plot of distance vs. curing time for the tensile strength.

distance decreases. During the hand-layup process it is difficult to strength. Change in volume leads to change in the area of exposure
eliminate void content and hence as the volume increases the per- to IR of composite laminate whereas the IR emitter size is constant.
centage of void content also increases which leads to reduced The change in shape factor leads to reduced absorbance of IR and
hence affecting the strength of the composite laminate. The contin-
ued exposure to IR in case of increased curing time has resulted in
overriding the negative effect of distance contributing to the posi-
Contour Plot of Tensile strength vs Curing Time, Mi, Distance, mm
tive effect as observed by the interaction effect of distance and cur-
Tensile
strength,
ing time. The IR radiation absorbance is a complex phenomenon
220 MPa particularly for composites which depends on the type of matrix
< 260
200 260 - 280
and reinforcement. The peak absorbance capacity also changes
280 - 300 with wave length. These combined effects have resulted in the var-
Curing Time, Min

180 300 - 320

320 - 340 iation of results.
> 340
160 Similar analysis was carried out for the values of flexural
Hold Values
140 Volume, mm3 175838
strength. The results of ANOVA showed similar significant factors
as for tensile strength. Percentage contribution of each factor is
120 shown in Table 10. Curing time is insignificant whereas it has the
100 interaction effect with the distance and the other two factors are
Distance, mm = 150.102 significant.
80 Curing Time, Min = 57.4254
Tensile strength, MPa = 340.543 Regression analysis was carried out considering the nonlinear
60 effects and there were no nonlinear terms contributing to the flex-
150 175 200 225 250 275 300
ural strength. Hence only significant factors as indicated in Table
Distance, mm 10 were included and the mathematical model is represented in
Eq. (2).
Fig. 8. Contour plot of curing time vs. distance for the tensile strength.

Flexural strength ¼ 386  29:8A þ 1:25B  22:7C þ 19:0AB ð2Þ

Table 10 where A = (a  225)/45, B = (b  150)48, C = (c  175837.5)/

Percentage contribution of factors for flexural strength. 55982.5. A, B and C are coded values and a, b and c are actual values
Source A B AB C Error and others
of the factors.
Normal probability effect plot and normal probability plot of
Percentage 43.96 0.07 10.48 25.60 19.89
contribution (%)
residuals were similar to tensile strength, where factors distance
and volume have negative effect and the interaction of distance

Fig. 9. Surface plot of volume vs. distance for the flexural strength.
 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137 1135

Contour Plot of Flexural Strength vs Volume, mm3, Distance, mm indicating as the values of distance and volume increase the
Flexural strength decreases. Figs. 11 and 12 show the surface plot and con-
260000 Strength tour plot of flexural strength for curing time vs. distance. The sur-
< 300
240000 300 - 325 face is curvilinear indicating presence of interaction effects. The
325 - 350
220000 350 - 375
contour plot has curved lines indicating the interaction effect of
Volume, mm3

375 - 400 curing time and distance. The distance of 153 mm and curing time
200000 400 - 425
425 - 450 of 61 min has higher flexural strength compared to distance of
180000 > 450 222 mm and curing time of 148 min as shown in Fig. 12. Behaviour
160000
Hold Values
similar to tensile strength is also observed in case of flexural
Curing Time, Min 147.5
strength. The factors have similar effects.
140000 Overlaid contour plot shown in Fig. 13. It is drawn in order to
120000 Distance, mm = 153.152
Distance, mm = 277.859
Volume, mm3 = 86298.2
compare both tensile strength and flexural strength with respect
Volume, mm3 = 86273.5 Flexural Strength = 387.150 to significant factors holding the curing time at minimum. The do-
100000 Flexural Strength = 470.436
main of higher tensile and flexural strength is shown in the plot.
150 175 200 225 250 275 300 The optimum values observed from the overlaid plot are, distance
Distance, mm
is 150 mm, volume is 83179 mm3and the curing time at 56 min.
Confirmatory test were conducted to ascertain the accuracy of
Fig. 10. Contour plot of volume vs. distance for the flexural strength. prediction for the models proposed in Eqs. (1) and (2). The confir-
matory test results with chosen factor levels are indicated in Table
11. The predicted results are reasonably accurate as the error lies
and curing time has positive effect on the overall strength of the within 10%, hence confirming the validity of the model.
composite, the same is also proved in the regression analysis in Figs. 14–17 show the micrographs of infrared and thermally
Eq. (2). cured polymer composite laminates. Three distinct bands are visi-
Figs. 9 and 10 respectively show the surface plot and contour ble in both the micrographs. The 1st and 3rd band are having fibers
plot of flexural strength for distance vs. volume. The surface is a perpendicular to the direction of cutting, and the 2nd band is par-
plane with rising ridges and the inclined lines present in contour allel to the direction of cutting, clearly showing 0/90 fiber orienta-
plot indicate absence of nonlinear effects. The distance of tion. The inference that can be drawn from these micrographs is
153 mm and volume of 86273 mm3 has higher flexural strength that the fiber–matrix interface and the inter-ply bonding are
compared to distance of 277 mm and volume of 86298 mm3 similar in IR cured and thermally cured laminates.

Fig. 11. Surface plot of distance vs. curing time for the flexural strength.

Contour Plot of Flexural Strength vs Curing Time, Min, Distance, mm

Overlaid Contour Plot of Tensile strength, MPa, Flexural Strength
Flexural
Strength Tensile
220 < 300
260000 strength,
300 - 350 Distance, mm = 151.245 MPa
Distance, mm = 222.371
200 Curing Time, Min = 148.627 350 - 400
240000 Volume, mm3 = 267512
200
Tensile strength, MPa = 327.751
Curing Time, Min

Flexural Strength = 387.822 400 - 450 340

Flexural Strength = 447.761
220000
Volume, mm3

180 > 450 Flexural

Strength
Hold Values 200000 Distance, mm = 298.084 300
160 Volume, mm3 175838 Volume, mm3 = 267512 450
180000 Tensile strength, MPa = 247.655
140 Flexural Strength = 246.562 Hold Values
Curing Time, Min 56
160000
120
140000
100 Distance, mm = 152.912
Distance, mm = 150.718
Distance, mm = 298.084
Curing Time, Min = 61.3138
120000 Volume, mm3 = 83179.3
Volume, mm3 = 83839.9
Tensile strength, MPa = 353.142
80 Flexural Strength = 479.911
100000 Flexural Strength = 523.389 Tensile strength, MPa = 272.669
Flexural Strength = 321.200

60
150 175 200 225 250 275 300 150 175 200 225 250 275 300
Distance, mm Distance, mm

Fig. 12. Contour plot of curing time vs. distance for the flexural strength. Fig. 13. Over laid contour plot of tensile strength and flexural strength.
1136 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137

Table 11
Validation results of the suggested model.

Factor levels Experimental tensile Predicted tensile %Error Experimental flexural Predicted flexural %Error
strength strength strength strength
A B C
310 56 187,500 287.2 281.5 1.74 263 253.15 3.74

Fig. 16. Microstructure of thermal cured specimen for 236 min curing time
Fig. 14. Microstructure of IR cured specimen for 56 min cure schedule (200 lm). (200 lm).

Fig. 15. Microstructure of IR cured specimen for 56 min cure schedule (100 lm). Fig. 17. Microstructure of thermal cured specimen for 236 min curing time
(100 lm).

5. Conclusions

In conventional thermal curing process the outer layer receives posite laminate transparent to infrared radiation as the curing pro-
more heat than the inner layers and hence the conversion rate is cess is complete.
completed at a faster rate compared to the core layers which lead IR utilizes only 25% of total time as compared to conventional
to nonuniformity in curing and hence the stresses are developed curing method. IR curing process has drastically reduced the curing
within the composite laminate. time.
IR curing results in volumetric heating and the entire composite Increase of volume or distance or both have negative effect con-
laminate will be uniformly heated resulting in uniform curing of all tributing to the extent of 70% on both the tensile and flexural
the layers and hence the stresses are reduced. More uniform cross- strength. Positive interaction effect is observed in case of curing
linking takes place and hence there is no uncured resin patches time with distance affecting to an extent of 10–15%, even though
with in the laminate. Further duration of IR curing makes the com- the main effect of the curing time is insignificant.
 P.K. Kumar et al. / Materials and Design 32 (2011) 1129–1137 1137

Over laid contour plot shows the optimal region for both the [5] Rao RMVGK, Rao Sandhya, Sridhara BK. Studies on tensile and interlaminar
shear strength properties of thermally cured and microwave cured glass–
tensile and flexural response for the selected process parameters.
epoxy composites. J Reinf Plast Comp 2006;25:783–95.
The scanning electron micrographs further prove that there is [6] Adanur Sabit, Arumugham Yuvaraj. Characteristics of ultraviolet cured glass
similar interfacial bonding between the fibers and the matrix in epoxy composites: experimental procedures and testing. J Ind Text
both thermal and IR curing process. 2002;32:93–106.
[7] Gourdenne Albert J. Radiofrequency (27.12 MHz) processing of thermosetting
It can be concluded from the analysis infrared radiation curing DGEBA epoxy resins. High Perform Polym 2005;17:425–48.
can be effectively employed for curing of epoxy matrix glass fiber [8] Ribeiro Rahul, Morgan Roger J, Bonnaud Leila, Lu Jim, Sue Hung- Jue.
reinforced composites. Characterization of the electron beam curing of cationic polymerization of
diglycidylether of bisphenol a epoxy resin. J Compos Mater 2005;39:1433–47.
[9] Belhamra Ali, Diabi Rabah, Moussaoui Abdallah. Technology and applications
of infrared heating in the industrial area. J Eng Appl Sci 2007;2:1183–7.
Acknowledgements
[10] Chern Bih-Cherng, Moon Tess J, Howell John R. On-line processing of
unidirectional fiber composites using radiative heating: I. Model. J Compos
The authors wish to acknowledge the support provided by the Mater 2002;36:1905–34.
Principal and Management of B.N.M. Institute of Technology, Ban- [11] Labeas GN, Watiti VB, Katsiropoulos ChV. Thermomechanical simulation of
infrared heating diaphragm forming process for thermoplastic parts. J
galore and The National Institute of Engineering, Mysore and Head Thermoplast Compos 2008;21:353–70.
of the Departments of Mechanical Engineering of these Institutes [12] Decker C. Photointiated crosslinking polymerization. Prog Polym Sci
for extending their support in this Project work. The authors pro- 1996;21:593–650.
[13] Suresha S, Sridhara BK. Effect of addition on graphite particulates on the wear
fusely thank Dr L. Krishna Murthy and Prof S. Suresha for their behaviour in aluminum–silicon carbide–graphite composites. Mater Des
valuable suggestions and guidance in the statistical analysis. 2010;31:1804–12.
[14] Gu Ruijun, Kokta Bohuslav V. Maximization of mechanical properties of birch-
polypropylene composites with additives by statistical experimental design. J
References Thermoplast Compos 2010;23:239–63.
[15] Onal L, Adanur S. Optimization of compression molding process in laminated
[1] Loos Alfered C, Springer George S. Curing of epoxy matrix composites. J woven composites. J Reinf Plast Compos 2005;24:775–80.
Compos Mater 1983;17:135–69. [16] Cao Y, Cameron J. The effect of curing conditions on the properties of silica
[2] Naji Malak I, Hoa Suong V. Curing of thick angle-bend thermoset composite modified glass fiber reinforced epoxy composite. J Reinf Plast Compos
part: curing cycle effect on thickness variation and fiber volume fraction. J 2007;26:41–50.
Reinf Plast Comp 1999;18:702–23. [17] Barker Thomas B. Quality by experimental design. New York: Marcel Dekker
[3] Zhang J, Xu YC, Huang P. Effect of cure cycle on curing process and hardness for Inc.; 1985.
epoxy resin. Express Polym Lett 2009;9:534–41. [18] Kuehl Robert O. Design of experiments. USA: Duxbury; 2000.
[4] Ramakrishnan B, Zhu L, Pitchumani R. Curing of composites using internal [19] Montogomery Douglas C. Design and analysis of experiments. New
resistive heating. J Manuf Sci E–T 2000;122:124–31. Delhi: Wiley India (P) Ltd.; 2007.