Materials Science and Engineering A

Materials Science and Engineering A 527 (2010) 4560–4570 Contents lists available at ScienceDirect Materials Science and Engineering A journal homep...
Author: Brett Williams
3 downloads 0 Views 1MB Size
Materials Science and Engineering A 527 (2010) 4560–4570

Contents lists available at ScienceDirect

Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea

Laminates based on vinyl ester resin and glass fabric: A study on the thermal, mechanical and morphological characteristics Minakshi Sultania a , S.B. Yadaw b , J.S.P. Rai a , Deepak Srivastava a,∗ a b

Department of Plastic Technology, H.B. Technological Institute, Kanpur 208 002, India Head, Composite Materials Division, D.M.S.R.D.E., G.T. Road, Kanpur, India

a r t i c l e

i n f o

Article history: Received 11 October 2009 Received in revised form 4 January 2010 Accepted 14 April 2010

Keywords: Vinyl ester resin Glass fabric DSC TGA DMA Mechanical properties SEM

a b s t r a c t The effects of variation of styrene content on the thermal, mechanical and morphological behaviours of epoxy novolac vinyl ester resin (EVER)/glass fabric laminate have been investigated. The vinyl ester resin matrix was synthesized indigenously using epoxy novolac resin and methacrylic acid catalyzed by triphenylphosphine at a temperature of 85 ◦ C in nitrogen atmosphere. Fourier-transform infra-red spectroscopic (FT-IR) analysis was used to see the structural changes during the synthesis of the EVER. Differential scanning calorimetric (DSC) technique was used to investigate the curing behaviour of the EVER matrix. Exothermic peaks (Tp ) appeared in the range of 117–127 ◦ C for all the samples of epoxy resin and EVERs. Thermal stability of the prepared samples was analyzed by dynamic thermogravimetric runs. The TG/DTG trace of epoxy showed two-step mass loss decomposition behaviour whereas that of vinyl ester resin exhibited a single step mass loss. Mechanical properties such as tensile, flexural and impact strengths of the prepared laminates were determined and it was found cured resin containing 40% styrene showed the best balance of properties and EVER/glass fabric laminates exhibited better properties as compared to the epoxy/glass fabric laminates Dynamic mechanical analyses of the samples were done to determine the viscoelastic properties. Tg decreased significantly with increased styrene concentration due to increase in cross-linking density. Cross-sections of the cured samples, which failed during impact testing, have been critically studied through scanning electron microscopic (SEM) analysis to gain insight into the phase morphology. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Glass fabric-reinforced polymer composites have found wide use for a range of structural and functional applications because of their superior performance combined with a high strength, stiffness and weight reduction. The low cost, high level of ultimate strain achievable, impact resistance, damage tolerance and general ease of processing, high specific strength and stiffness, good formability, corrosion and fatigue resistance makes the use of glass fabric reinforcement very attractive to the aerospace and automotive industries. Vinyl ester resins are widely used as thermoset matrix to fabricate a variety of reinforced structures [1–8] including pipes, tanks, scrubber and ducts. They are the prime candidates for use in composite for transportation and/or infrastructure. In addition to these applications, vinyl esters are also being used in coatings, adhesives, molding compounds, structural laminates, electrical applications, etc. Vinyl ester resins combine the best properties of epoxies and

∗ Corresponding author. Fax: +91 512 2533812. E-mail address: [email protected] (D. Srivastava). 0921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2010.04.038

unsaturated polyesters. They have better chemical resistance than cheaper polyester resins, especially hydrolytic stability and at the same time offer greater control over cure rate and reaction conditions than epoxy resins [9]. Vinyl ester resin based on epoxy novolac are used for chemical storage tanks, pipes and ducting, fume extraction systems and gas cleaning units as this resin shows superior chemical resistance at high temperatures. These have high tensile elongation along with better corrosion resistance, which makes them promising material for producing lining coating with outstanding adhesion to other types of plastics and conventional materials such as steel and concrete. Vinyl ester resins also find a variety of applications in optical fiber coating, topcoats for containers as well as printed circuit boards. Vinyl ester resins are based on the reaction product of an epoxy resin and an ethylenically unsaturated carboxylic acid, which results in a polymer with chain end unsaturation. Various epoxy resins are used, including the diglycidyl ether of bisphenol A, or higher homologues thereof, the diglycidyl ether of tetrabromo bisphenol A, epoxylated phenol–formaldehyde novolac and polypropylene oxide diepoxide. The most commonly used acids are acrylic and methacrylic acids. The acid–epoxide reaction is straightforward and is catalyzed by ternary amines, phosphines or

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570 Table 1 Sample designation. S. No.

Vinyl ester resin (wt.%)

Styrene (wt.%)

Sample code

1 2 3 4

65 60 55 50

35 40 45 50

EVER-1 EVER-2 EVER-3 EVER-4

35 40 45 50

EVERL-1 EVERL-2 EVERL-3 EVERL-4

Vinyl ester glass laminates 5 65 6 60 7 55 8 50

alkalis [10,11]. The mechanical properties such as tensile strength and modulus, flexural strength and modulus, interlaminar shear strength, of fibrous composites, depends on the structure of matrix resins, properties of fibers, volume fraction of fibers (Vf ), and interphase. The structure of cross-linked network of vinyl ester resin can be altered by changing the backbone structure or end groups or the nature and concentration of reactive diluents (e.g., styrene) [12,13]. Styrene imparts good mechanical properties (i.e., tensile strength), heat distortion resistance, and dielectric properties to cured resins [14]. However, an increase in the styrene content resulted in a decrease of tensile modulus [12]. In the present article, an attempt was made to develop composites based on epoxy vinyl ester resin and glass fabric and to study the thermal, mechanical and morphological characteristics of laminates based on vinyl ester resin and glass fabric by varying the styrene concentration. A comparison was also made between the composites prepared from market procured epoxy novolac resin (EPN 1138) and in situ vinyl ester resin using glass fabric. 2. Experimental 2.1. Material The novolac-based epoxy resin used was EPN 1138 with an epoxide equivalent weight of 250 g/equiv., as determined by acid titration and the cure agent HY 951 (triethyl tetramine) were procured from M/s Ciba Specialty Chemicals Pvt. Ltd., Mumbai, India. Methacrylic acid and triphenylphosphine procured from M/s CDH Pvt. Ltd., New Delhi were used for the synthesis of vinyl ester resin. Styrene, benzoyl peroxide and hydroquinone were obtained from M/s E. Merck, Germany, which was used for the curing of vinyl ester resin. Glass fabric (aerial density 351 g/m2 ) obtained from M/s Unnati Corporation, Ahmedabad, India was used for the preparation of the laminates. 2.2. Synthesis of vinyl ester resin Epoxy vinyl ester resin was prepared by using 1:0.9 mole ratios of epoxy resin and methacrylic acid. The reaction was carried out in the presence of triphenylphosphine (TPP) catalyst (1% by weight of the epoxy resin) and hydroquinone (200 ppm as inhibitor) at 85 ◦ C in nitrogen atmosphere for about 5 h to obtain a product with desired acid value [15]. In order to remove the free methacrylic acid, the prepared resin was dissolved in benzene and treated with potassium carbonate, stirring for 2 h at 30 ◦ C. The acid, in the form of an acid salt, was extracted by water and benzene was evaporated using a RotovapTM evaporator under vacuum [16]. 2.3. Curing of vinyl ester resin The curing of vinyl ester resin was done by using varying concentrations of styrene (as per the formulations given in Table 1) viz. 35, 40, 45 and 50% (w/w) and free radical initiator benzoyl perox-

4561

ide (2% by weight). Half-of-the styrene was mixed with the resin and to the other half benzoyl peroxide was added. Both the flasks were sealed and kept under refrigeration to avoid premature polymerization prior to use. Then equal amount of the solutions were placed in small glass vial and stirred vigorously with a glass rod at room temperature. The mixture was poured into pre-heated iron mold and cured into hot air over at 120 ◦ C for 1 h and then postcured for 2 h at 150 ◦ C. Specimens for the entire test were cut from this block (square sheet) of cured material. 2.4. Composite preparation The prepregs for the glass fiber/vinyl ester composite were prepared by using vinyl ester resin, styrene and benzoyl peroxide in the ratios as discussed in Section 2.3 above. Each layer of fabric was well impregnated with a solution of this mixture by using the hand lay–up method. Glass fabrics (20 pieces each of size 25 cm × 30 cm) were stacked between two plain Teflon sheets. This system was placed in between the two mild steel plates between the two preheated platens of a hydraulic press. The laminates were molded at 120 ◦ C under a pressure of 50 kg/cm2 . Post-curing was done at 150 ◦ C for 2 h to release the residual strains in the laminates. 2.5. Characterization of the samples 2.5.1. Fourier-transform infra-red (FT-IR) spectroscopy The purified vinyl ester resin was subjected to Perkin Elmer Fourier-transform infra-red (FT-IR) (Model: RX-1) spectroscopic analysis, to monitor the formation or disappearance of various functional groups in the wavelength range of 400–4000 cm−1 . 2.5.2. Differential scanning calorimetric (DSC) analysis Cure temperatures of the prepared samples were observed by taking very little quantity of samples into shallow aluminum pan sealed by an aluminum cover of differential scanning calorimeter (DSC) (TA Instrument, USA; Modulated DSC 2920). This was placed in sample cell of the instrument. The starting temperature, programmed rate and final temperature were taken at heating rate of 10 ◦ C/min. Dynamic scans were obtained which were used for assuming the cure temperature. 2.5.3. Thermogravimetric analysis (TGA) The percent weight loss and thermal degradation characteristics of the prepared samples were evaluated by thermogravimetric analyzer (TGA) of TA Instrument (Model Hi. Res. 2950 TGA unit interfaced with Thermal Analyst 2100 control unit). About 5–10 mg of sample was taken in a platinum sample pan and nitrogen was purged at 60 ml/min during the dynamic runs. The heating rate in each run was kept at 10 ◦ C/min and the temperature range was ambient to 800 ◦ C. The relative thermal stability of the cured samples was evaluated by comparing the initial decomposition (Ti ), temperature of maximum rate of mass loss (Tmax ) and final decomposition temperature (Tf ). 2.5.4. Mechanical properties 2.5.4.1. Tensile, flexural and impact tests. The tensile tests were performed in Tinius Olsen machine designed as per Patent No-2, 784, 048 at a crosshead speed of 1 mm/min according to ASTM D-638 standards. The values were taken from the average of five specimens. Flexural tests were performed using Hounsfield Tensometer W-5236 testing machine fitted with a three-point-bending fixture at a crosshead speed of 1 mm/min according to ASTM D-790. The dimensions of the specimen were 75 mm × 15 mm × 3 mm and the span to thickness ratio was set at L/D = 32:1 in all the cases. The values were taken from an average of at least five specimens.

4562

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

The impact strength of the specimen was determined by using Charpy Mandate Tensometer using rectangular specimen of 75 mm × 15 mm × 3 mm according to ASTM D-256 standards. The tests were carried out at room temperature and values were taken from an average of at least five samples. Samples for all the tests were cut from the cured sheet of 6 in. × 6 in. with the help of cutting machine. 2.5.4.2. Dynamic mechanical analysis (DMA). The glass transition temperatures of cured samples were determined by dynamic mechanical analyzer (DMA) with TA Instruments DMA 2980. Analysis by DMA is a well-known method for determining viscoelastic properties by applying a controlled sinusoidal strain to a sample and measuring the resulting stress. DMA gives both storage modulus and loss modulus characteristics as a function of temperature. 2.5.5. Scanning electron microscopic (SEM) analysis The fractured samples under mechanical analysis were sputtercoated with gold prior to scanning electron microscopy (SEM) examination. Jeol JSM 5800 model was used to view the specimen; several micrographs were taken for each sample. 3. Results and discussion 3.1. Fourier-transform infra-red (FT-IR) spectroscopic analysis FT-IR spectra of pure epoxy (EPN) and synthesized vinyl ester resin (EVER) have been shown in Figs. 1 and 2 respectively. The disappearance of band of epoxide group near 841–912 cm−1 in EVER sample (Fig. 2) and appearance of strong absorption due to carbonyl group of the ester linkage near 1713 cm−1 , confirmed the formation of EVER. The carbonyl stretching at 1700 cm−1 was absent (in EVER sample Fig. 2) indicating that the reaction of epoxy group with the acid was complete. The absorption bands at 1634 and 815 cm−1 appeared due to stretching and bending vibrations of the vinylic group. Band corresponding to acryloyl double bond (–C C–) appeared at 1607 cm−1 whereas the peak at 2929 cm−1 showed the presence of the methyl group in the sample. A broad absorption band observed in the region of 3400–3300 cm−1 , cen-

Fig. 1. FT-IR spectrum sample EPN.

tered at 3441 cm−1 was due to the hydroxyl group of the vinyl ester resin. The occurrence of a peak at 1163 cm−1 was probably due to the C–O–C stretching and the peak near 947 cm−1 depicted the out-of-plane bending of vinyl ester monomer. These findings were consistent with the previous work by various researchers [13,17,18]. 3.2. Curing studies Vinyl ester networks are based on polyfunctional acrylate or methacrylate monomers dissolved in styrene and are cured by free radical-initiated polymerization. Generally, the free radical polymerization consists of initiation, inhibition, propagation, and termination steps. The termination step may be neglected at the later stage of polymerization due to the formation of highly crosslinked microgels [19]. To prevent premature gelation and to inhibit the spontaneous polymerization of the resin, hydroquinone was added as an inhibitor [20]. The vinyl ester monomer provides crosslinking capacity and branch points for the developing network while the styrene monomer provides linear chain extension. A

Fig. 2. FT-IR spectrum of sample EVER-2.

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

4563

Fig. 3. DSC scan of sample EPN.

lower concentration of styrene increases the rate of cure for thermally cured VER systems [21]. This is perhaps due to a decrease in the termination rate constant with increased concentration of cross-linking species (i.e., the dimethacrylate) that rises the polymerization rate [21,22] or it may be caused by an increase in the propagation rate due to the replacement of the less reactive styryl radicals by the more reactive methacrylyl radicals, as claimed by Rey et al. [23]. Depending on the temperature and other curing conditions some of the species remain unaffected after curing in the form of residual monomers and soluble polymers and do not contribute towards the cross-linked network. Previous studies of curing behaviour of vinyl–divinyl systems show that the initial curing for such a system begins with an induction period, due to the presence of inhibitors and cage effects, followed by the formation of microgels [24]. Microgels have been defined as domains of high cross-link density dispersed in a pool of un-reacted monomers. The cross-linking of such microgels does not contribute to the global

network structure until they are incorporated into the gel phase. In the present work, we have used DSC measurements to evaluate the curing exotherm of the prepared vinyl ester resin samples with styrene curing agent. The dynamic DSC scan of epoxy resin (EPN) has been shown in Fig. 3 whereas Fig. 4 showed the results obtained by DSC scan of EVER-2 sample. The onset temperature of curing (Tonset ), the exothermic peak temperature or the temperature of maximum cure (Tp ) and the final temperature or the temperature at the end of the cure (Tstop ) for the samples were determined from DSC scans and the results have been tabulated in Table 2. It was evident from the data (Table 2), that the onset temperature decreased as the styrene concentration increased probably due to the reduction in resin viscosity and thus increased oxygen diffusion out of the resin during sample preparation in nitrogen atmosphere. The peak exothermic temperatures were 110, 114, 117 and 127 ◦ C respectively for samples EVER-1 to EVER-4. It was clear from these data that Tp shifted

Fig. 4. DSC scan of sample EVER-2.

4564

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

Table 2 Data obtained from dynamic DSC scans. S.No.

Samples

Ti (◦ C)a

Tonset (◦ C)b

Tp (◦ C)c

Tstop (◦ C)d

H (J g−1 )

tc (min)e

Tg f

1 2 3 4 5

EPN EVER-1 EVER-2 EVER-3 EVER-4

56.9 88.2 67.1 60.9 51.8

66.1 97.4 89.9 82.9 62.5

106.5 110.1 114.4 117.3 127.7

184.3 132.3 142.1 147.1 153.6

60.7 80.1 75.9 73.1 70.8

31.0 60.5 59.5 58.2 57.5

82.0 86.0 75.0 74.8 74.3

a b c d e f

Temperature of cure initiation. Onset temperature by extrapolation. Temperature of cure maximum. Temperature of end of cure. Cure temperature from isothermal run. Glass transition temperature obtained from DSC.

Fig. 5. TGA trace of sample EPN.

towards higher temperature side with increasing styrene concentration. This data was consistent with those presented by Scott et al. [21] and appeared to conflict with those presented by Varma et al. [13] where increased styrene concentration resulted in decreased exothermic temperature. On the basis of this, curing temperature of 120 ◦ C was selected to cure the samples in an air-circulatory oven (±2 ◦ C) for obtaining an appreciable rate of curing in all the samples.

This took about 60 min as pre-curing of vinyl ester resins. Finally, the sample was post-cured at 150 ◦ C for 2 h. The reduction in cure time (could be treated as a measure of reaction rate) with increased styrene content appeared to be due to the combined effect of the lower reactivity of the styryl radicals and the higher termination constant [25]. This may be due to the fact that the curing reactions led to increase in viscosity and, eventually, to gelation of the entire

Fig. 6. TGA trace of sample EVER-2.

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

4565

Fig. 8. DMA trace of EPN and EVER matrices (a) plot of storage modulus versus temperature and (b) plot of tan ı versus temperature.

3.3. Thermal stability

Fig. 7. Mechanical properties of the samples showing variations of: (a) tensile strength and modulus, (b) flexural strength and modulus and (c) impact strength and maximum dynamic strength.

resin mixture. The amount of heat liberated (H) was increased with increasing styrene content partly due to higher concentration of the vinyl groups in its mixture. For an EVER, an increase in styrene concentration which showed a reduction in cross-linked density (refer Table 2) would allow greater radical and monomer mobility enhancing the termination rate and therefore decreasing the maximum polymerization rate. Increase in styrene content resulted in lower Tg values (Table 2) as a result of its effect on the cross-link density of the cured material. Higher styrene levels reduced the Tg by plasticizing of the curing reaction with styrene monomer and by a reduction of the crosslink density, which allowed higher conversion prior to vitrification [26].

The thermal stability of the cross-linked networks plays an important role and is greatly influenced by the structure, chemical composition, kind and concentration of remaining polar groups, cohesive energy between molecular chains, molecular chain rigidity, different interaction parameters and other chemical structural factors like steric strain, conformational arrangements of groups, etc. [27–30]. The TGA curves of cross-linked epoxy (EPN) and vinyl ester resin (e.g. sample EVER-2) have been shown in Figs. 5 and 6, respectively. Relative thermal stability of cured samples was evaluated by comparing the initial decomposition temperature (Ti ), temperature of maximum rate of weight loss (Tmax ), final decomposition temperature (Tf ) and % char yield at 800 ◦ C. The TG/DTG trace of epoxy (EPN) (Fig. 5) showed two-step mass loss decomposition behaviour. A major mass loss of 55.6% was observed up to the temperature of 600 ◦ C (Table 3). A minor mass loss of 17.6% was observed above 600 ◦ C. A clear-cut single step decomposition behaviour was observed in all the thermograms of vinyl ester resins (Fig. 6 for EVER-2 sample). The results of the thermograms (Table 3) indicated that the presence of different concentrations of styrene did not affect the nature of cross-links formed during curing. In all the samples, initial weight loss of ∼1–3% was observed between 100 and 300 ◦ C, which was probably due to the volatilization of the residual solvents or diluents, by products of curing reactions as well as the entrapped moisture

4566

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

Table 3 Data obtained from TG/DTG traces of prepared samples. S.No.

Samples

Ist step ◦

1 2 3 4 5

EPN EVER-1 EVER-2 EVER-3 EVER-4

Percent char yield at 800 ◦ C

IInd step ◦









Ti ( C)

Tmax ( C)

Tf ( C)

ML (%)

Ti ( C)

Tmax ( C)

Tf ( C)

ML (%)

368 250 252 351 260

467.0 490.8 490.0 488.0 488.1

600 600 700 700 750

55.6 77.4 79.5 76 79.6

600 – – – –

700 – – – –

760 – – – –

17.6 – – –

24 20 18 19 17

Tq : initial degradation temperature; Tmax : peak degradation temperature; Tf : stop degradation temperature; ML: mass loss.

present in the sample [31,32]. However rapid decomposition was observed in the temperature range of 350–600 ◦ C and ∼76–80% of the sample was lost during the major decomposition step. This might correspond to the advanced fragmentation of the macromolecules formed in the second stage, secondary reactions of dehydrogenation, thermal cracking, and disproponation and gasification process. The Ti value of epoxy resin was found to be more than that of vinyl ester resin (Table 3) samples. This might be due to the presence of bulky methyl group of methacrylic acid (MA) which might have started early thermal decomposition and made vinyl esters less stable as compared to epoxy due to the steric congestion in the three-dimensional networks. The Tf values also showed the lower reactivity of the vinyl ester reins in comparison to epoxy resin. The concentration of styrene marginally affected the values of Ti and percent char yield.

3.4.2. Dynamic mechanical analysis Dynamic mechanical analysis (DMA) is an extremely versatile thermal analysis technique used to study the viscoelastic behaviour of polymeric materials and yields quantitative results for tensile storage modulus (E ), loss modulus (E ), and loss factor (tan ı). The value of E is the measure of the mechanical energy stored under load whereas tan ı values compare the amounts of dissipated and stored energy. The maximum of the tan ı curve corresponds to the glass transition temperature (Tg ) above which significant chain motion takes place [36]. When Tg of a polymer is exceeded, certain mechanical properties may be compromised severely. Also, the Tg gives an indication about cross-linking density of a polymer. DMA provides a direct method for determining the crosslink density () of cross-linked materials by measuring the modulus at

3.4. Mechanical properties 3.4.1. Tensile, flexural and impact tests Tensile strength, tensile modulus, flexural strength, flexural modulus, impact strength, and maximum dynamic strength, plotted for different epoxy and vinyl ester matrices and laminates have been shown in Fig. 7(a–c). The tensile strengths for the neat epoxy and vinyl ester resin (sample EVER-2) matrices were found to be 16 and 35 MPa, which increased to about 19 and 11 times for their laminates, respectively. An increase was also observed in tensile modulus, flexural strength, flexural modulus, impact and maximum dynamic strengths (Fig. 7a–c) for the neat resins and the corresponding laminate systems. The incorporation of styrene (35–50%) into vinyl ester laminates enhanced the values of tensile strength and tensile modulus up to 45% and thereafter decreased (Fig. 7a). This might be due to the increased flexibility caused by the uncross-linked styrene monomer present in the laminates. A similar trend was also observed with flexural strength, flexural modulus and impact strength (Fig. 7b and c). This implied that increasing styrene content in the vinyl ester system, at high styrene concentration, the styrene chains formed as cross-linked through the EVER molecules and, thus, the inhomogeneous shear deformation changed from very localized to more diffuse region which led to the observed behaviour [33]. Such behaviour can be related to the existence of residual styrene in the cross-linked resin. The influence of curing time and temperature affected the presence of styrene in the resin. This styrene might act as a plasticizer and lowered the stiffness and increased the strain-to-failure. The opposite effect was observed with epoxy resins [34]. The impact strength of the composite reinforced with fibers was much higher than the impact strength of the vinyl ester resin matrix (Fig. 7c). This result was expected and was in accordance with literature data [35]. The fracture of the vinyl ester composites reinforced with glass fibers showed a fiber pullout mechanism and total rupture of the fiber and matrix after the impact test.

Fig. 9. DMA trace of EPNL and EVERL laminates (a) plot of storage modulus versus temperature and (b) plot of tan ı versus temperature.

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

4567

Fig. 10. SEM micrographs of (a) neat EPN, (b) neat EVER containing 40% styrene, (c) EPNL specimen EVERL-1, (d) EVERL-2, (e) EVERL-3 and (f) EVERL-4.

Table 4 Comparative data of DMA of prepared samples. DMA data for matrices S.No.

Samples

Tg (◦ C)

E (at 50◦ C, MPa)

E (at Tg + 50 ◦ C, MPa)

 × 103 (g cm−1 )

1 2 3 4 5

EPN EVER-1 EVER-2 EVER-3 EVER-4

125.7 105.4 103.7 96.9 99.1

6558.6 6139.9 5315.8 4958.9 6507.6

33.6 43.9 27.4 25.6 43.7

2.46 4.19 2.57 2.39 4.15

DMA data for laminates S.No.

Samples

Tg (◦ C)

E (at 50 ◦ C, MPa)

E (at Tg + 50 ◦ C, MPa)

 (g cm−1 )

1 2 3 4 5

EPNL EVERL-1 EVERL-2 EVERL-3 EVERL-4

141.1 111.6 109.5 104.9 183.2

40421.8 33742.3 36586.9 28228.9 32035.7

6482.2 13706.2 11244.5 9732.4 10092.4

0.51 1.38 1.04 0.92 0.87

4568

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

Scheme 1. Mechanism of synthesis of vinyl ester resin.

a temperature well above the Tg . The values of E in the rubbery region at T > Tg were taken to calculate  by using Eq. (1) [37,38]. =

E 3RT

(1)

where E is the storage modulus at Tg + 50 ◦ C, R be the gas constant and T be the absolute temperature at Tg + 50 ◦ C. Fig. 8(a) and (b) showed the variation of storage modulus (E ) and loss factor (tan ı) with temperature (T) for EPN and EVER samples whereas Fig. 9(a) and (b) showed the variation of dynamic properties of epoxy laminate (sample EPNL) and vinyl ester resin laminates (EVERL samples). Fig. 8(a) clearly evidenced that the values of E dropped with increasing temperature indicated that all the formulations gradually passed from stiff-hard-solid to soft and flexible material which agreed well with previous findings [39–41]. It is well known that the cross-linking density is proportionate to E at temperature (Tg + 50 ◦ C) according to the rubber elasticity theory [42] which well suited to our results in the case of matrices and laminates. Fig. 8(b) revealed the structure of the different networks formed in the polymer. Increasing the amount of styrene in the system not only reduced the temperature of the glass–rubber transition, but also reduced the width of the peaks and increased its height [43]. For vinyl ester system, it has been suggested [44] that phase separation occurs during cure producing highly crosslinked, methacrylate-rich region possessing a high Tg , surrounded by lightly cross-linked, styrene-rich region, presumably possessing a low Tg . Such a two-phase network should be observable in dynamic mechanical tests [45]. The single, quite symmetrical glass transition observed in the tan ı curve (Fig. 8b) does not support the existence of two-phase structure; however, it has been suggested [46] that the phase-separated domains were too small to be detected by DMA. The Tg of EVER sample was found to be lower than that of its epoxy counterpart (Table 4). This might be due to the presence of styryl radicals in the vinyl ester matrix. Also, the increased con-

centration of styrene in the EVER matrix reduced the Tg , which might be due to the decrease in the crosslink density [47] (Table 4). However, the Tg does not consistently decrease with increased styrene concentration. In contrast, Auad et al. [18] found that the Tg decreased significantly with increased styrene concentration for bisphenol A-based VERs and Li et al. [48] found that the Tg decreased with increased styrene concentrations for VERs with a low molecular weight dimethacrylate oligomer while it was almost invariant with styrene concentration for VERs with a high molecular weight dimethacrylate oligomer. The difference may be due to the effect of competing factors on Tg . In this series of VER/styrene blends, the ␣-relaxation or Tg reflects the motion of large chain segments between the cross-linking sites and shifts to lower temperatures as the concentration of styrene is increased. If these polymers had not been miscible on a molecular level, separate glass transitions would have been observed close to the temperatures for the pure components. These results thus indicated a good miscibility of styrene and EVERs and resulted in a single-phase morphology, which well accorded, with the SEM results, as discussed in the forthcoming section. The Tg values associated with fabric-reinforced composites, regardless of the curing schedule, were higher than those of the resin samples (Table 4). The difference in Tg between the fabricreinforced composites and resin materials may have been due to difference in the dynamic behaviour, determined in parts by the stiffness, mass and damping properties of the material. For example, because the fabrics are stiffer and can carry more load than the matrix, the properties observed in samples loaded in threepoint bending may not be appreciably sensitive to variations in the mechanical properties of the matrix [49]. Other possibilities for the difference in the Tg may be a greater amount of curing or cross-linking found in the fabric-reinforced polymeric materials. For example, the heat transfer through the fabric-reinforced polymers may have been different than that in the resin during DMA temperature ramp because of the fabric reinforcement and acted as a thermal energy conduit. It also is possible that the surface chem-

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

istry of the fabrics reacted in such a way with the resin to produce a higher exotherm and, therefore, a higher degree of cross-linking than that found in the resin alone. 3.5. Morphological studies Fig. 10 showed the morphology of fractured surfaces of the neat epoxy, neat vinyl ester resin having 40% styrene, and laminates of epoxy and vinyl ester resins. The SEM micrograph of neat epoxy resin showed only one phase that may be regulative cracks in the fracture surface, indicating a brittle fracture surface, which might be accounted for the poor toughness of the neat epoxy resin, as shown in Fig. 10a. From the SEM micrograph of neat EVER (Fig. 10b), it is evident that there appeared fine cracks in the fractured sample and it exhibited a single-phase morphology. The pattern is known as line morphology that depicted the propagation of energy during impact loading. These lines also indicated the multidirectional nature of propagation of energy (wave) within the matrix, which meant that the dissipation of energy within the bulk was not unidirectional. Thus the pattern of morphology was characteristic of brittle system having smooth, glassy fractured surfaces with cracks in different planes. In the case of composite laminates, the shear fracture occurred in the matrix-dominated interface area between the piles [50]. For typical glass fabric/vinyl ester laminates, the fracture micromechanism primarily occurred as interfacial de-bonding of the matrix from the fibers. When the styrene content in the laminate was 35% (sample EVERL-1), complete detachment of matrix from the fabric surface was observed in the SEM micrograph (Fig. 10c), which might be due to weak adhesion as evidenced by smooth and clean fabric surface. A high degree of fabric/matrix de-bonding is the dominant mechanism for shear failure of laminates fabricated with vinyl ester system. So, glass fabric/vinyl ester laminates displayed direct matrix de-bonding from the fabric surface. Moreover, interlaminar shear failure involved several fracture mechanism that occurred during delamination, such as matrix deformation and fracture, fabric pullout, fabric bridging and fabric/matrix interface de-bonding [51]. SEM micrograph of samples EVERL-2 and EVERL3 (Fig. 10d and e) exhibited better bonding between matrix and fabric, which means that the samples can bear more impact and tensile loading than sample EVERL-1 with lesser styrene content and, hence, possessed better mechanical properties. In these cases the transfer of impact energy from fabric to matrix within the laminate might be excellent as compared to that in sample EVERL-1. The phases get separated in the case of EVERL-4, as it is clear from the SEM micrograph (Fig. 10f). 4. Conclusions The following conclusions are based on the findings reported in this paper: 1. FT-IR spectra showed that the chemical reaction occurred in between the oxirane group of epoxy novolac resin and carboxylic group of methacrylic acid in the presence of basic catalyst to produce vinyl ester resin, which is established as shown in Scheme 1. 2. Vinyl ester resin matrices and laminates exhibited superior thermal, mechanical and morphological properties in comparison to epoxy novolac resin. 3. Thermal stability was almost the same for the vinyl ester resin samples cured using varying concentrations of styrene but less than epoxy novolac resin. 4. The results of the mechanical properties showed that vinyl ester resins containing styrene in different proportions indicated that the cured resin containing 40% styrene showed the best bal-

4569

ance of mechanical properties in both the matrices as well as the laminates. Also vinyl ester resins exhibited better properties as compared to epoxy resin. 5. Increasing the amount of styrene in the system not only reduced the temperature of the glass–rubber transition (Tg ), but also reduced the width of the peaks and increased its height. Increased styrene concentration increased the final conversion due to decreased cross-linked density, which reduced topological constraints. Laminates also showed similar results. 6. SEM photomicrographs of EVER matrix displayed a single-phase morphology and glass fabric/vinyl ester laminates displayed direct matrix de-bonding from the fabric surface. Moreover, interlaminar shear failure involved several fracture mechanism that occurred during delamination, such as matrix deformation and fracture, fabric pullout and fabric bridging and fabric/matrix interface de-bonding. Acknowledgements Authors are thankful to the Director, Defense Materials Stores Research and Development Establishment (DMSRDE, a unit of Defense Ministry, Government of India) Kanpur, India, who had permitted to do the work in the organization. Special thanks to Mrs. Punam Awasthi for helping us in DSC analysis. References [1] M.E. Kelly, P.E. In Bruins (Eds.), Unsaturated Polyester Technology, Gordon and Breach, New York, 1976, pp. 343–349. [2] H.Y. Yeh, S.C. Yang, J. Reinf. Plast. Compos. 16 (1997) 414–418. [3] S.S. Sonti, E.J. Barbero, J. Reinf. Plast Compos 15 (1996) 701–707. [4] N. Hag, P. Harrison, Corrosion Preventing and Control 43 (1996) 162–165. [5] B.P. Singh, R.C. Jain, I.S. Bharadwaj, J. Polym Sci 2 (1994) 941. [6] K. Liao, R.I. Altkom, S.M. Mikovich, J.M. Fildes, J. Gomez, J. Adv. Mater 28 (1997) 54–63. [7] J.R. Brown, Z. Mathys, Compos. Part A: Appl. Sci. Manuf 28 (1997), 675-671. [8] U. Sorathia, T. Dapp, Int. Sample Symp. Exhib. 42 (1997) 1020–1031. [9] D.K. Dwivedi, B.S. Kaith, A.S. Singha, Inter. J. Plast. Tech 7 (2003) 119–125. [10] B. Lubin, Handbook of Composites, Van Nostrand Reinhold, New York, 1984. [11] C.A. May, Epoxy Resin Chemistry and Technology, Marcel Dekker, New York, 1988. [12] M. Malik, V. Choudhary, I.K. Varma, J. Fire Sci 20 (2002) 329–342. [13] I.K. Varma, B.S. Rao, M.S. Choudhary, V. Choudhary, D.S. Varma, Die Angew. Makromol. Chem 130 (1985) 191–199. [14] R.D. Patel, J.R. Thakkar, R.G. Patel, V.S. Patel, High Perform. Polym 2 (1990) 261–265. [15] C.L. Ogg, W.L. Porter, C.O. Willitis, Ind. Eng. Chem. Anal. Edn 17 (1945) 394–397. [16] G. Malucelli, G. Gozzelino, F. Ferrero, R. Bongiovanni, A. Priola, J. Appl. Polym. Sci 65 (1997) 491–497. [17] E.L. Rodriguez, Polym. Engg. Sci 31 (1991) 1022–1028. [18] L.M. Auad, M. Aranguren, J. Borrajo, J. Appl. Polym. Sci 66 (1997) 1059–1066. [19] L. Suspene, D. Fourquier, Y.S. Yang, Polymer 32 (1991) 1593–1604. [20] S. Asean, B. Immirzi, P. Laurienzo, M. Malinconico, E. Martuscelli, M.G. Volpe, J. Mater. Sci 32 (1997) 2329–2336. [21] T.F.W.D. Scott, J.S. Cook, Forsythe, Euro. Polym. J 38 (2002) 705–716. [22] G.L. Batch, C.W. Macosko, J. Appl. Polym. Sci 44 (1992) 1711–1729. [23] L. Rey, J. Galy, H. Sautereau, Macromol 33 (2000) 6780–6786. [24] K. Dusek, H. Galina, J. Mikes, Polym. Bull 3 (1980) 19–25. [25] T.F. Scott, W.D. Cook, Forsythe, J.S., Polymer 44 (2003) 671–680. [26] F. Cardona, D. Rogers, S. Davey, G. Van Erp, J. Compos. Mat 4 (2007) 137–152. [27] D.K. Chattopadhyay, D.B. Rohini Kumar, B. Sreedhar, K.V.S.N. Raju, J. Appl. Polym. Sci 91 (2004) 27–34. [28] D.K. Chattopadhyay, S.S. Panda, K.V.S.N. Raju, Prog. Org. Coat 54 (2005) 10–19. [29] R. Narayan, D.K. Chattopadhyay, B. Sreedhar, K.V.S.N. Raju, N.N. Mallikarjuna, T.M. Aminabhavi, J. Appl. Polym. Sci. 97 (2005) 1069–1081. [30] D. Rosu, C.N. Cascavaf, C. Ciobanu, L. Rosu, J. Ana. Appl Pyrolysis 72 (2004) 191–196. [31] S. Agrawal, R. Singhal, J.S.P. Rai, J. Macromol. Sci. Pure Appl. Chem A36 (1999) 741–757. [32] S. Agrawal, A. Mishra, J.S.P. Rai, J. Appl. Polym. Sci 87 (2003) 1952–1956. [33] C. Baley, Y. Perrot, P. Davies, A. Bourmaud, Y. Grohens, Appl. Compos. Mater 13 (2006) 1–22. [34] C. Baley, P. Davies, Y. Grohens, G. Dolto, Appl. Compos. Mater 11 (2004) 96. [35] I.M. Low, M. McGrath, D. Lawrence, P. Schmidt, J. Lane, B.A. Latella, K.S. Sim, Compos. A 38 (2007) 963–974. [36] Luckenbach, T.A., Rheometrics, Inhttp://www.sealseastern.com/PDF/DynamicMechThermalAnal.pdf (accessed April 2004).

4570

M. Sultania et al. / Materials Science and Engineering A 527 (2010) 4560–4570

[37] R.B. Prime, in: E.A. Turi (Ed.), Thermal Characterisation of Polymeric Materials, Academic Brooklyn, 1981, p. 1380. [38] R. Narayan, D.K. Chattopadhyay, B. Sreedhar, K.V.S.N. Raju, J. Mater. Sci 37 (2002) 4911–4918. [39] D. Verchere, H. Sautereau, J.P. Pasecualt, S.M. Mosechain, J. Appl. Polym. Sci 41 (1990) 467–485. [40] N. Chikki, S. Fellahi, M. Bakar, Euro. Polym. J 38 (2002) 251–264. [41] L.T. Manzione, J.K. Gillham, C.A. McPherson, J. Appl. Polym. Sci 26 (1981) 889–905. [42] A.V. Tobolsky, Properties and structure of polymers, Wiley, New York, 1960. [43] L. Rey, J. Duchet, J. Galy, H. Sautereuu, D. Vougner, L. Carrioon, Polymer 43 (2002) 4375–4384.

[44] [45] [46] [47] [48] [49] [50] [51]

S. Dua, Mc.R. Cullough, G.R. Palmese, Polym. Compos 20 (1999) 379–391. L. Nielson, J Macromol Sci.-Rev Macromol Chem C3 (1969) 69–103. J Zhang, Mow. Richardson, Polymer 41 (2000) 6843–6849. V. Bellenger, J. Verdu, M. Genem, B. Mortaigne, Polym. Polym. Compos 2 (1994) 9–16. H. Li, E. Burts, K. Bears, Q. Ji, J.J. Lesko, D.A. Dillard, J.S. Riffle, P.M. Puckett, J. Compos. Mater 34 (2000) 1512–1528. M. Akay, Compos. Sci. Technol. 47 (1993) 419–423. R. Roy, B.K. Sarkar, N.R. Bose, Bull. Mater. Sci 24 (2001) 137–142. M.F. Hibbs, M.K. Tse, W.L. Bradley, in: J. Johnston Norman (Ed.), Toughened composites, ASTM STP 937. American Society for Testing and Materials, 1987, pp. 115–130.