Mechanical characterization of carbon fiber reinforced phenolic resol and epoxy condensate plastic.
Carbon fiber reinforced plastics composites are advanced materials that have high specific strength and stiffness while also possessing excellent high temperature resistance. Epoxy has been the workhorse matrix resin system used With carbon and other advanced fiber based composites (1, 2). A resin matrix recipe With phenol-formaldehyde resin, as well as with certain modifications, has frequently been used in carbon-carbon composites (3, 4). The matrix compositions of epoxy with phenolic resin were reported (5), wherein the novolac type of phenolic resin was used as a curator of the epoxy resin matrix. The advancement in the manufacturing techniques has led to the development of prepreg techniques, and the basic resin system should preferably be in a homogeneous liquid phase. From these considerations, the combinations of phenolic resin With medium viscosity epoxy resin system may become very attractive. The resol phenolic resin has not been used as a matrix resin because of its very short shelf life, the high reactive nature, and evolution of volatile matter during its curing reaction. Of late, suitable modifications in the production of resol type phenolic resin have been made, which have resulted in the development of the resin having long shelf life with moderate viscosity and which does not need any solvent. Although many engineering plastic materials have been made possible by the modification of epoxy with phenolic resin, worthwhile studies on carbon fiber reinforced composite with resol type phenolic and epoxy matrix have not appeared in the literature.
As part of a comprehensive program on the development of aerospace grade carbon fiber reinforced composite, a matrix consisting of phenolic (resol) and epoxy resin has been developed at the Macromolecular Research Centre, Jabalpur (India). The basic reaction scheme (5) of phenolic resin with epoxy system has been conveniently used in the matrix preparation. The intermediate reaction product, i.e., phenolic-epoxy (PR-EP) adduct, has been visualized during the prepreg preparation, wherein no solvent has been used in contrast with the earlier system. In this communication, the preparations of the prepreg via adduct formation and composite of the carbon fiber with this PR-EP adduct matrix are reported. Mechanical properties such as tensile strength and modulus, elongation at break, flexural strength and modulus for various composites of carbon fiber reinforced and PR-EP matrix are evaluated. The results are also compared with those composites prepared by one-shot lay-up techniques. The reaction mechanism of PR-EP adduct formation at prepreg preparation stage has been elucidated using FTIR spectroscopic analysis. The improvement in mechanical properties coupled with the favorable processability has been realized as a result of low volatile matter and readjustment in crosslink density.
High strength PAN based carbon fiber was supplied by Indian Petrochemicals Corporation Ltd., Vadodara, India. The carbon fiber type INDCARF-25 having 6K filament count was used. The fiber properties reported by the manufacturer are: tensile strength 2.5 GPa minimum, tensile modulus 215 to 240 GPa, ultimate elongation 1.05 to 1.40%, density 1.78 g/cc and the carbon fiber mass content 93%. The INDCARF-25 was containing 0.75 to 1.5% sizing material. Carbon fiber tows were used without removing the sizing material.
Diglycidyl ether of bisphenol-A type epoxy resin (LAPOX B-11) was obtained from Cibatul Ltd., Mumbai, India, which was basically unmodified epoxy resin. The LAPOX B-11 has epoxy value 5.2 to 5.5 equiv./kg. and its viscosity at 25 [degrees] C was 0.9 to 1.2 Pa.S.
The solventless resol type Phenol-formaldehyde resin (under the trade name of PR-100) was produced by ABR Organics Ltd., Hyderabad, India. The physicochemical characteristics of PR-100 resin are: dark brown liquid, viscosity (30 [degrees] C)-0.6 to 0.9 Pa.S., pH-7.5 to 7.8 and gel time (at 150 [degrees] C) 6 min.
Neat Matrix and Composite Preparation
Five matrix resin compositions were prepared by taking 100 parts of PR-100 resin with 0, 50, 100, 150, or 175 parts of LAPOX B-11, respectively. The matrix resin compositions having 200 or more parts of epoxy resin were not used as they showed tackiness after full curing. This mixed phenolic and epoxy resin system [TABULAR DATA FOR TABLE 1 OMITTED] was used in two different modes. In the first mode (i.e. prepreg method) the matrix resins were coated uniformly over the carbon fiber and were heated at 140-170 [degrees] C for 3-10 min and were immediately cooled to room temperature to get a carbon fiber prepreg. The prepreg was prepared in equipment designed and fabricated in-house. Phenolic-epoxy matrix resin was used directly for composite preparation in second mode wherein the intermediate PR-EP adduct formation as visualized during prepreg preparation was avoided. The composites were prepared in leaky type three parts split mold made up of mild steel. The composites were prepared by laying up in one direction (0 degree) and utmost care was exercised in the preparation of the unidirectional composites. The fiber volume fraction was kept close to 0.6 in all composites. The curing conditions employed were two step heating viz, 150 [degrees] C for 2h followed by 200 [degrees] C for 4h. The time-temperature conditions for two step prepreg method and one shot process are summarized in Table 1. No pressure was recorded for the mold as only manual tightening was followed. The neat resin consisting of PR-100 and LAPOX B-11 in different ratios as mentioned earlier were prepared to study the loss of volatiles during the cure, the change in chemical structure during prepreg, and full curing of the matrix resin.
Cure Extent Study by FTIR Analysis
The FTIR spectra of the neat phenolic resin (PR) and epoxy resin (EP) were taken. A Perkin-Elmer FTIR spectrophotometer model 1720 x was used and data compilation and analysis were done with the help of PE 3700 data station. The FTIR spectra for the sample of neat PR and PR-EP were taken at different stages, i.e., at prepreg stage (130 to 170 [degrees] C for 3 to 10 min and fully cured stage (150 [degrees] C for 2h). The samples for FTIR studies were prepared by coating a very small quantity of the resin sample between two spectroscopic grade KBr pellets, and samples were cured at 150 [degrees] C for 2 h followed by 200 [degrees] C for 4 h. The extent of the reaction between resol resin and epoxy was monitored by the disappearance of the band at 916-914 [cm.sup.-1] characteristics of epoxide ring stretch. The ratio of the intensity of band at 916-914 [cm.sup.-1] to intensity of the band at 1508 [cm.sup.-1] (assigned to semicircle stretching of benzene ring) (6, 7) indicates that the original epoxy and its decrease due to curing will provide quantitative indication of the PR-EP adduct formation.
Mechanical Testing of Unidirectional Carbon Composite
Tensile tests for tensile strength, modulus, and elongation were performed on Universal Testing Machine of Instron Ltd., UK, according to ASTM test method No. D-638. The crosshead speed (initial strain rate) was 5 mm/min and grip length 80 mm. The flexural tests were performed on same Universal Testing Machine according to ASTM test method D-790. A crosshead speed of 1.3 mm/min and a gauge length of 25 mm were maintained. The average of the five results was reported.
Loss of Volatiles During Cure
The loss of volatiles (in percentage) was determined by the weight loss of the sample cast as thin film over glass plate and cured at different temperature and duration. The loss of volatiles was determined for two cure conditions, which are: 1) the prepreg preparation and 2) full cure condition (i.e., 150 [degrees] C for 2 h and followed by 200 [degrees] C at 4 h).
RESULTS AND DISCUSSION
The curing of resin matrix containing the PR (resol type) having varied amounts of EP resin (as mentioned in the Experimental section) may involve two independent or concurrent reactions. The likely cure reactions are:
1. The condensation of epoxy with the phenolic hydroxyl group of PR, which may result in PR-EP adduct.
2. The establishment of curing reaction of the resol type PR wherein the crosslinking is introduced mostly with the formation of methylene group (5, 8).
Phenolic resin prepared by the reaction of phenols and formaldehyde under alkaline conditions comprises complicated mixtures of mononuclear or polynuclear hydroxymethyl phenols (9, 10) and para-quinone methides (11). Since the resol type phenol-formaldehyde resin used in this study was procured from a commercial source, the actual phenol/formaldehyde ratio is not disclosed. Because of the multiplicity of the reaction products, it is very difficult to calculate the stoichiometry ratio of the phenolic and epoxy groups, and the extent of various reactions during PR preparation. In light of these unavoidable situations, traditional recipe compositions were followed, i.e., parts of epoxy resin per hundred parts of phenolic resin. Nevertheless, the effect due to the inadvertent variation in the stoichiometric ratio will be reflected in macroscopic properties.
The two reactions mentioned above may compete with each other, and it is essential to identify the products resulting from these reactions at the prepreg stage. Precise literature on such factors is not available; however, IR spectroscopic characterization (6, 8) has been used to qualitatively study the curing reaction. The infrared absorption band at 1400 - 1300 [cm.sup.-1] (assigned to -OH deformation of phenol) and at 1034 [cm.sup.-1] (assigned to out of plane -CH deformation of an aromatic ring) could be used (6, 8); however, the absorption band at 916-914 [cm.sup.-1] was used to predict the formation of a PR-EP adduct. Figure 1 shows the FTIR spectra for pure epoxy resin (spectrum [A]), uncured pure PR Resin (spectrum [B]), cured PR Resin (spectrum [C]), uncured PR-EP mix (w/w 100: 100) as (spectrum [D]), PR-EP at prepreg preparation stage (spectrum [E]), and PR-EP prepreg fully cured state (spectrum [F]). It can be seen from Fig. 1 that the bands other than 916-914 [cm.sup.-1] could not be found suitable for qualitative as well as quantitative studies because of the inability to observe isolated bands. The disappearance of the epoxide bands at 916-914 [cm.sup.-1] for the phenolic-epoxy matrix at the prepreg stage (i.e. spectrum [E]) itself indicates that the reaction (i.e., PR-EP adduct formation) has taken place. This band also disappeared when the same sample was cured to full stage (spectrum [F]). It can be inferred from this FTIR study that this reaction takes place at prepreg preparation stage itself. Figure 2 describes the reaction scheme for PR-EP adduct formation. The introduction of Bisphenol-A diglycidyl ether moiety of epoxy resin in PR-EP matrix may lead to the reduction in crosslink density. Extensive investigations are under way in our laboratory that may shed light on the mechanisms of these reactions.
The major drawback of the resol type PR matrix is the evolution of volatiles as a product of cure reaction, which leads to high degree of shrinkage and void in the composites. The loss of volatiles for different composite compositions and for two different techniques of the composite preparations is given in Table 2. It can be seen that the loss of volatiles during the curing of neat PR by the one-shot technique was found to be as high as 23.6% and to the EP level of 175 php, the [TABULAR DATA FOR TABLE 2 OMITTED] loss of volatiles was reduced to 4.7%. This clearly indicated that the curing of PR-EP combination has a positive effect on the volatility parameter. The prepreg technique for composite making involves two steps and the major volatile loss was observed at the prepreg preparation step. A loss of 16.2% was observed for neat PR; it was reduced to 8.1% and to 2.5% in the case of 50 and 175 parts of epoxy php, respectively. The volatile loss in the case of prepreg curing was reduced to a very low level, as can be seen in Table 2. The extent of volatiles was reduced to 4.2% and to 1.4% for 50 and 175 parts of epoxy php, respectively; therefore, the advantage of prepreg over the one-shot technique is obvious in terms of low volatile loss, and the composite prepared by the prepreg technique may have improved mechanical properties.
High specific strength and stiffness together with the lightness of the carbon fiber composites are the basic characteristics of importance for structural applications. The matrix plays an important role in determining the performance properties of high strength carbon fiber composites. It has been observed that the prepreg techniques used in this study impart process advantages and consequently make the matrix most suitable technologically for composite preparations. It became imperative to evaluate the mechanical properties of these composites prepared by both the techniques. Five parameters viz., tensile strength and modulus, elongation at break and flexural strength and modulus are evaluated to have the comparison of the mechanical properties of the composites having PR-EP resin matrices and prepared by one shot and prepreg techniques. Figure 3 shows the plot of tensile strength of the unidirectional carbon fiber composites against the epoxy php in the resin matrix. All the composites except for neat phenolic matrix show improved tensile strength for prepreg techniques. The improvement was quite appreciable for the matrix resin having 50, 100, and 150 epoxy php. The composite prepared by the prepreg technique exhibited poor tensile strength in comparison to its counterpart prepared by the one-shot method. The observed deterioration was attributed to the very high loss of volatiles. The voids were observed on visible inspection. The improvement of tensile strength up to 125% was observed at the epoxy level of 100 php. The substantial improvement in tensile strength at epoxy level of 50, 100, and 150 php could be explained by the formation of PR-EP adduct.
The tensile moduli determined for these composites are shown in Fig. 4 and the trend on the effect of epoxy php on tensile modulus was more or less similar to the one observed for the tensile strength parameter. The highest modulus was observed at epoxy php of 50 and modulus values at 50 epoxy php were 136.3 GPa and 120.4 GPa, respectively, for the prepreg technique and one-shot method. The reduction in tensile modulus values was seen in pure PR matrix when the one-shot technique was adopted. The improvement in tensile modulus for the composites having the PR-EP matrix can be rationalized on the basis of adduct formation and the low loss of volatiles in the prepreg technique.
The results of the elongation values are seen in Fig. 5, and the improvement in elongation at break was observed in almost all PR-EP compositions except at an EP level of 175 php. The enhanced elongation could be attributed to the flexibility incorporated in the matrix and the reduction in crosslink density. The compositions prepared by the prepreg technique showed improved elongation at epoxy php of 50 in comparison with the composite prepared by one-shot technique.
The flexural strength values for the composites do not show very specific trends [ILLUSTRATION FOR FIGURE 6 OMITTED]; however, the flexural strength for the composite prepared by the prepreg technique exhibits high values in comparison with the one prepared by the one-shot technique. The expected reduction in flexural strength was observed for the compositions having pure PR matrices and prepared by the prepreg technique.
The flexural modulus versus epoxy php are displayed in Fig. 7. It can be seen from Fig. 7 that the extent of improvement in flexural modulus due to the incorporation of epoxy in phenolic resin was much smaller in order as compared with that observed for other parameters (tensile strength, tensile modulus, and elongation). Nevertheless, this enhancement was more or less of the same order as recorded for flexural strength. In general, the addition of epoxy resin was found to improve the flexural modulus, and the composites prepared by the prepreg method displayed higher values than those prepared by the one-shot technique.
The overall improvement in the mechanical properties of the carbon fiber composites was observed at 50 and 100 parts of epoxy php of phenolic resin in its matrix resin. The prepreg technique as adopted here has been demonstrated to be better than the one-shot technique, and this could be attributed to the formation of PR-EP adduct during prepreg preparation. As mentioned earlier, the stoichiometric ratio for PR-EP could not be calculated as it was not possible to presume the specific structure of the PR resin because of the multiplicity of the products. If one presumes that the resol PR has only polynuclear hydroxyl-methylphenol, the stoichiometric composition lies between 100 to 150 parts of epoxy php of phenolic resin. The maximum improvement around 50 epoxy php indicates that the stoichiometric composition may be in this epoxy php range. The loss of phenolic hydroxyl functionality may be due to the formation of hemiformals (9, 12) and quinone methides; however, these theories could not be confirmed and need further investigation.
A matrix resin comprising resol type phenolic resin (PR) and bisphenol-A diglycidyl ether type epoxy resin (EP) could be synthesized and unidirectional carbon fiber reinforced composites were prepared with PR-EP matrix wherein the epoxy resin was varied from 0 to 175 parts per hundred part of phenolic resin (php). A substantial improvement in mechanical properties was observed for composites with PR-EP matrix at epoxy php between 50 to 150 when the composites are prepared by prepreg techniques. The improvement of mechanical properties may be attributed to the formation of PR-EP adduct at the prepreg preparation stage, and the same was established by FTIR spectral analysis. The accomplishment of a wide spectrum of mechanical properties shall be useful for finding many applications for these carbon fiber composites.
The financial assistance of the Indian Petrochemicals Corporation Ltd., Vadodara, India, is gratefully acknowledged. The authors express their gratitude to the testing section of Central Institute for Plastic Engineering and Technology (CIPET), Bhopal, India, for mechanical characterization.
1. C. E. L. Reader and J. M. Morley, Proc. 4th London, Intern. Carbon and Graphite Conf. Soc. Chem. Ind., p. 238, London (1976).
2. G. Dorey, Phys. Technol, 11, 56 (1980).
3. E. Fitzer, K. H. Gelgl, and W. Huttner, 14th Bienn Conf. on Carbon, Pennsylvania State University, University Park, Pa., extended abstracts, p. 236 (1974).
4. L. M. Manocha, H. Bhatt, and S. M. Manocha, Carbon. 34, No. 7, 841 (1996).
5. S. Paul, Comprehensive Polymer Science, Vol. 6, Ch. 6, p. 149, Pergamon Press, Oxford, England (1989).
6. Y. Zaks, J. Lo, D. Render, and E. M. Pearce, J. Appl. Polym. Sci, 27, 913 (1982).
7. D. W. Schiering and J. E. Katon, J. Appl. Polym. Sci., 34, 2367 (1987).
8. A. Knop and L. A. Pilato, Comprehensive Polymer Science, Vol. 5, Ch. 35, p. 621, Pergamon Press, Oxford, England (1989).
9. R. Perrin, R. Lamartine, J. Vicens, M. Perrin, A. Thozet, D. Haton, and R. Fugies, Nouv. J. Chim., 10, 179 (1986).
10. B. Mechin, D. Hanton, J. Le. Goff, and J. P. Tanneur, Eur. Polym. J., 20, 333 (1984).
11. R. T. Jones, J. Polym Sci., Polym. Chem. Ed., 21, 1801 (1983).
12. P. W. Kopf and E. R. Wagner, J. Polym. Sci., Polym. Chem. Ed., 11, 939 (1973).