Polymer Engineering and Science, Feb 15, 1993 v33 n3 p166(9) The use of cryogenically ground rubber tires as a filler in polyolefin blends. K. Oliphant; W.E. Baker. Author's Abstract: COPYRIGHT Society of Plastics Engineers Inc. 1993 The effects of cryogenically ground rubber tires (CGT) on some of the mechanical properties of blends with linear low density polyethylene (LLDPE) and high density polyethylene (HDPE) are presented. Precoating the CGT particles with an ethylene-acrylic acid copolymer is shown to overcome most of the deleterious effects of adding CGT to LLDPE, while still retaining composite processability. A blend of 40 wt% EAA coated CGT particles with LLDPE is shown to have impact and tensile strengths that are 90% of those for the pure LLDPE, representing increases of 60 and 20%, respectively, over blends with uncoated particles. Blends of LLDPE with ground tire bladders demonstrate that even better mechanical properties can be obtained with similar large rubber particle size but somewhat better adhesion. For HDPE, however, it is shown that with large rubber particles, moderate adhesion is not sufficient to produce useful composites. Full Text: COPYRIGHT Society of Plastics Engineers Inc. 1993 INTRODUCTION Increasing legislation restricting the disposal of used tires and greater environmental consciousness have heightened the search for economical and environmentally sound methods of recycling discarded tires. Although a number of uses have already been examined !1, 2^, they are either uneconomical or do not utilize sufficient quantities of tires to represent a complete solution, and new methods need to be explored. One area with the potential both to be economical and to utilize large volumes of tires is the use of ground tires as fillers in polymer composites. In these systems tires are cryogenically or ambiently ground to a fine powder (100 to 600!micro^m), separated from the metal and polyester cord, and then compounded with thermoplastics, elastomers, or thermosets. The addition of cryogenically ground rubber tires (CGT), however, generally results in significant deterioration in the mechanical properties of these composites. CGT is reported to have a detrimental effect !3-5^ on most of the physical properties of cured rubbers, the extent of deterioration increasing with the amount and size of the scrap rubber. Phadke, et al. !3^, reported that there seems to be little adhesion between the rubber phase and CGT in vulcanates with natural rubber, and this, in conjunction with the large particle size, is believed to be the reason for the poor properties. In blends of CGT with an unsaturated polyester resin, Rodriguez !6^ found that the addition of rubber particles decreased mechanical properties, concluding that the particles were too large to provide any toughening effects. The addition of CGT to a number of thermoplastics has also been reported to result in poor blend properties. Phadke and De !4^ reported that CGT shows poor adhesion to a polypropylene matrix and, therefore, decreases the impact strength. Poor mechanical properties were also found for blends of CGT with polyethylene !7^, polystyrene, and acrylonitrile-i butadiene-styrene !8^. In these systems it is again the poor adhesion and large particle size that are believed to be responsible for the poor blend properties. Although some applications may tolerate the loss in properties generally observed for polymers that have been blended with CGT, it is clearly desirable to find methods of overcoming the detrimental effect of using CGT as a filler. It is important, however, that solutions be economical in terms of total material costs and the costs of the processing steps. Because little or no reduction in the size of CGT particles is seen during normal melt blending operations, the particle size is limited by the cost of the grinding process. This has led to a focus on developing methods of improving adhesion in the hope of improving blend properties. Surface treatment of the CGT particles with a small amount of a liquid unsaturated curable polymer mixed with a curing agent has been reported to extend the level of CGT incorporation that can be tolerated in rubber compounds !9^. A similar approach has also been shown to improve the properties of polyethylene-CGT blends !7^. In a slightly different approach, grafting of styrene onto CGT particles using an aqueous slurry process !8^ was found to give a product superior to straight mechanical blends of CGT with polystyrene (PS), but properties were still below those of the pure polymer. Considerable research has also been carried out industrially, but little specific information related to chemical systems and adhesion is available !1, 10^. In this work, the relationship between particle/matrix adhesion and mechanical properties was studied for composites of different ground rubber particulates with LLDPE and HDPE. The influence of particle size on composite properties and the effectiveness of a reactive coating process were also examined. EXPERIMENTAL Three different cryogenically ground rubber samples were used: i) whole tires ground to 40 mesh (!approximately equal to^ 300/!micro^m), from which most of the metal and polyester cord had been separated (SR40); ii) the same whole tires ground to 80 mesh (!approximately equal to^ 150!micro^m), again with the cord separated; and iii) a 40 mesh ground tire bladder (TB40), supplied by Custom Cryogenics. The whole tire particulates are composed of 40 to 50% rubber (styrene-butadiene rubber, butadiene rubber, natural rubber, butyl rubber, etc.), 25 to 40% carbon black, and 10 to 15% low molecular weight additives, the exact composition depending on the specific type of tire and from where in the tire the particle originated. The tire bladder has a different composition, the rubber component being solely butyl rubber. For some samples (designated SR40C and SR80C) the tire particulate was precoated with an ethylene-acrylic acid (EAA) copolymer, Dow Primacor 3460. The coated or uncoated CGT was melt blended with linear low density polyethylene (LLDPE), Esso Escorene 5101, or high density polyethylene (HDPE), Du Pont Sclair 2907, The polymer properties are listed in Table 1. Both the precoating and blending processes were performed on a Haake-Buchler Rheomix (Model 600) batch mixer with roller blades. EAA precoated CGT was prepared by blending 10 wt% EAA copolymer with CGT for 3 min at 180!degees^C and 150 rpm. The coated or uncoated CGT was blended with the LLDPE or HDPE for 5 min at 180!degrees^C and 100 rpm in a separate operation. Test specimens were molded on a Chemical and Mining Co. MCP 100SA injection molding machine at a pressure of 60 psi and a temperature of 190!degrees^C. Dumbbell-shaped tensile specimens 3.4 x 3.1 x 60 mm (20 mm gage length) were tested at room temperature TABULAR DATA OMITTED on an Instron tensile tester at a rate of 125 mm/min. FIat disk specimens of 3.2 mm thickness and 27.7 mm diameter were tested for impact strength on a Rheometrics RDT-5000 Instrumented Drop Weight Impact Tester equipped with a 11.1 kN piezo-electric load cell tip-mounted to the high velocity dart. Tests were performed at room temperature at an impact speed of 6.5 m/s (256 in/s). In order to obtain an estimate of the adhesion in the composite systems studied, compounds with typical tire tread and tire bladder formulations (Table 2) were prepared for peel testing, The compounds were pressed into sheets and cured at 140!degrees^C for 45 min. The rubber (cloth backed) and polymer sheets were pressed together in a mold for 5 min at 180!degrees^C and 5 MPa. Peel tests (11) were performed at an angle of 180!degrees^ on an Instron tensile tester at an extension rate of 125 mm/min. The SR compounds are intended to represent a typical tire tread formulation at 1) normal cure levels (SR) and 2) at low cure levels (SRLC), while the TB blends represent a typical butyl tire bladder composition (minus the stabilizers and antioxidants). Melt flow index (MFI) data were obtained on a Tinius-Olsen Thermodyne Extrusion Plastometer according to ASTM D 1238 condition E. The blend morphology was analyzed using a Jeol model JSM 840 scanning electron microscope (SEM). Samples were prepared by fracturing the blends under liquid nitrogen, followed by etching in a concentrated solution of chromic acid (5 gm), sulfuric acid (100 ml), and phosphoric acid (30 ml) in water (30 ml) at 80!degrees^C for 3 min to remove the CGT particles. Etched samples were sputter-coated with gold before examination. RESULTS AND DISCUSSION CGT/LLDPE Blends There were no difficulties encountered in incorporating up to 70 wt% CGT in LLDPE, and the resulting blends could still be injection molded to produce consistent test specimens with no voiding or distortion. The melt flow index (MFI), however, was found to decrease with increasing filler content and with decreasing particle size, indicating some relatively high upper limit on rubber incorporation for the retention of broad thermoplastic processability. Table 2. Composition of Vulcanates for Peel Testing (Parts by Weight). SR SRLC TB Styrene-butadiene rubber(a) 70 70 -- Butadiene rubber(b) 30 30 -- Bromo-butyl rubber(c) -- -- 70 Carbon black(d) 30 30 30 Stearic acid 2 0.3 2 Zinc oxide 5 0.8 3 Sulfur 2 0.3 0.5 Vanax-NS(e) 1.5 0.3 -- Methyl tuads(f) 0.5 0.1 -- Morfax(g) -- -- 1.5 a Polysar Krylene 1502 b Polysar Taktene 1203 c Polysar Bromo-Butyl X-2 d Cabot Vulcan N299 e N-tert-butyl-2-benzothiazolesulfenamide f Tetramethylthiuram disulfide g 4-morpholinyl-2-benzothiazole disulfide/paraffinic oil The simple blends of CGT with LLDPE were found to show a large drop in mechanical properties even at low levels of tire particulate. Figures 2a and 2b show the load-displacement curves of the impact tests for blends of LLDPE with SR40 and SR80 rubber particulates. Table 3 lists the various impact parameters obtained from these tests. The ultimate force, which has been shown to correspond to the yield point in the failure process (12), is seen to decrease with increasing CGT content. The values for blends with the smaller particle size (SR80) CGT are slightly higher (5 to 15%) than those of the large particle size (SR40) blends, the difference being more pronounced at higher rubber loadings. The relative stiffness (13) of the materials may be obtained from the initial slope of the load-displacement curve, defined as: Relative Stiffness (RS) = !F.sub.2^ - !F.sub.1^/!D.sub.2^ - !D.sub.1^ where !F.sub.2^ and !D.sub.2^ are the load and displacement at 50% of the ultimate force, and !F.sub.1^ and !D.sub.1^ are the load and displacement at 20% of the ultimate force. The results are given in Table 3. As expected, the addition of the lower modulus tire particles results in a reduction of the composite stiffness. Particle size is seen to have no effect on the composite stiffness over the range studied. For comparison, a blend of LLDPE with 40 wt% styrene-butadiene-styrene rubber has a relative stiffness of 1.0 x !10.sup.2^ kN/m. The impact energy is given by the area under the load-displacement curve. The impact strength, as listed in Table 3, is defined as the impact energy per unit thickness of material, and is observed to drop considerably upon incorporation of CGT into LLDPE. The smaller particle size of the SR80 blends does, however, result in impact strengths roughly 20% higher than the larger particle size SR40 blends. For both particle sizes, after a large initial drop at low filler loadings, the impact strength is seen to remain fairly constant with increasing CGT loading. In the impact failure of these systems the dart draws the material out in a ductile fashion as it passes through the specimen, with final failure occurring as a result of a tearing of the sample when the material is drawn too thin to support the load. As the filler loading increases, the amount of ductile deformation before tearing decreases, while the amount of elastic (or bending) deformation before the onset of ductile deformation increases. As a result, the total impact strength is seen to vary little with CGT content. The nominal tensile strength at yield of the blends decreases with increasing CGT content (Table 4). Particle size is seen to have no effect on the yield point of the composites, but has a pronounced effect on the post yield behavior, as shown by the differences in the total elongation (Table 4). For blends with both the SR40 and SR80 tire particles there is a dramatic decrease in elongation, but the smaller particle size SR80 blends have elongations 40 to 50% higher than the SR40 blends. Although the smaller particle size of the SR80 blends results in noticeably better properties, especially impact strength and tensile elongation, the properties still lie well below those of the pure LLDPE. The poor tensile and impact properties of CGT-LLDPE blends can be explained in terms of the extremely low adhesion in these systems. In peel tests of LLDPE with a tire compound (SR) that simulates the tire tread composition there is negligible adhesion (Table 5). This may be attributed to the high cross-link density of a typical tire, which does not allow for any interfacial interpenetration, resulting in a sharp interface !11^. That the poor adhesion is largely an artifact of the cross-link density can be seen in the significantly higher adhesion of LLDPE to a tire compound having a much lower concentration of curing agents (SRLC) and hence a lower degree of cross-linking. This low level of adhesion, coupled with the large particle size, results in little ductile deformation before tearing (see earlier discussion of impact failure) because the particles are unable to support any load and serve only to decrease the amount of matrix material. EAA Coated CGT/LLDPE Blends Precoating of the CGT particles with an EAA copolymer has a pronounced effect on the mechanical properties of the subsequent blends. This is clearly seen by comparing the load-displacement curves for SR80 coated and uncoated blends. The ultimate force and relative stiffness (Table 3) are seen to increase only slightly as a result of the coating process, but the impact strength (normalized area under the load-displacement curve) increases by 33 to 60%. The tensile strength is also seen to increase by 5 to 30% (Table 4), and the tensile elongation is increased by as much as 300% for 20% of SR40C. A blend of LLDPE with 40 wt% coated 80 mesh tire particulate (SR80C) has impact and tensile strengths that are 90% of those for the pure PE, representing increases of 60% and 20%, respectively, over the unmodified blend (SR80). The tensile elongation is 70% higher than for an unmodified blend, but still remains well below that of the pure LLDPE. Table 4. Tensile Properties of LLDPE/CGT Blends. Wt% Rubber Phase SR40 SR40C SR80 SR80C TB40 A. Tensile strength (MPa) 0 11.5 -- -- -- -- 20 10.1 10.5 10.1 10.6 10.3 40 8.5 9.1 8.5 10.1 8.9 60 6.5 7.6 6.4 8.1 6.8 B. Elongation (%) 0 470 -- -- -- -- 20 68 256 150 315 376 40 52 112 88 147 191 60 44 85 70 90 149 Table 5. Peel Test Results. Peel Force (N/m) LLDPE-SR 35 EAA-SR 35 LLDPE-SRLC 900 LLDPE-TB 1100 LLDPE-SBS 1800 HDPE-TB 1150 Impact failure in these composites is again due to a tearing of the specimen after some initial ductile deformation, but the extent of deformation before tearing is significantly greater than for simple blends of CGT and LLDPE. The blends of the EAA coated 40 mesh particles (SR40C) show similar improvements in properties (Table 3). The effect of particle size is similar. however, to that in the unmodified blends discussed previously, and properties are below those for the coated 80 mesh particles. The adhesion of EAA copolymer to a tire tread compound (Table 5) is negligible, and given the scatter in the peel test results, not significantly different from that of LLDPE. This is again believed to be a result of the high cross-link density of the tire compound, which prevents any interpenetration of the two phases at the interface. The poor adhesion also indicates that there are no specific interactions between the carboxylic acid groups on the EAA copolymer and any functional groups that might be present on the CGT surface. This system does not, however, exactly represent the surface of the scrap rubber particles, in that it has not been exposed to the environment or the cryogenic grinding process, which, through oxidation or the formation of free radicals on fracture, may introduce functional groups onto the CGT particle surface. The lower MFI's (higher viscosity) of the SR80C blends, compared with those with uncoated CGT (Fig. 1), may be an indication that specific interactions between the EAA copolymer and the CGT particles do in fact occur. However, straight blends of EAA, LLDPE, and CGT with no precoating step were found to have similar MFI's, so the MFI decrease may be due to some rheological phenomenon rather than specific interactions of the EAA copolymer with the CGT particles. An infrared analysis of a CGT surface has been reported to show the presence of double bonds !10^ but does not mention the presence (or absence) of other functional groups. More recent work has shown that a number of functionalized polymers (with similar functionality) are effective in improving composite properties, suggesting that some interaction with the CGT particles is occurring. A more detailed analysis of the CGT surface would be valuable in determining whether any specific interactions resulting in increased adhesion are in fact occurring. Because of the observed influence of particle size on the composite properties, scanning electron micrographs (SEM) of the fracture surfaces of blends with EAA coated and with uncoated particles were compared to see if particle size reduction during the high shear coating process might be responsible for the observed mechanical property increases. In all blends, however, there was no detectable decrease in particle size due to the coating process. It was also seen, as was expected, that the highly cross-linked rubber particles are not broken down significantly on straight mechanical blending with LLDPE, although some smaller particles are visible. Typical micrographs for blends of LLDPE with 40 wt% coated and with uncoated particles are shown in Fig. 3. To further examine the effect of the coating process, blends of EAA copolymer, CGT, and LLDPE were prepared in which there was no precoating step. Figure 2d shows the load-displacement curves for blends of LLDPE with 40 wt% CGT; unmodified CGT (a), EAA coated CGT (b), and the EAA/CGT/LLDPE blend with no precoating (c). The presence of the EAA copolymer in blends with no precoating (c) results in minimal improvements in the impact properties over unmodified blends, clearly showing the importance of the coating process. This also provides evidence that the EAA coating on the CGT particles remains at least partially intact on blending the coated particles with LLDPE. Table 6 lists the impact properties for blends of LLDPE with the EAA copolymer. Considering the blends of EAA coated CGT with LLDPE, if the EAA coating on the particles were dispersed into the LLDPE phase instead of remaining on the particle surface, this would result in EAA concentrations of 5 to 10% in the LLDPE (depending on the CGT concentration). Over this concentration range the impact properties of LLDPE/EAA blends (Table 6) are lower than those of pure PE. It can be concluded then that the EAA coating on the CGT particles is not increasing the composite properties by simply dispersing into, and toughening, the LLDPE phase. Rather this provides further evidence that the EAA coating is remaining, at least to a certain extent, intact on the CGT particle surface. The remnant sharp metal reinforcement may also be playing a role in the failure process that the EAA coating of particles might act to suppress. Removal, however, of the ultra-fine particle fraction (which is believed to contain the majority of the metal particulate) through sieving results in no difference in mechanical properties for either coated or uncoated blends. From the instrumented impact tests it is seen that failure is similar for both coated and uncoated composites, but that the coated systems undergo more ductile deformation before tearing, resulting in an increase in impact energy. It is also seen that tensile elongation is greatly increased by the coating process, and that yield stress is moderately increased. Although no increase in adhesion was observed for the model systems, the mechanical property improvements and more recent work with similar functionalized polymers seem to suggest that there is some interaction between the EAA copolymer and functional groups on the CGT surface, resulting in increased adhesion. It is this increased adhesion that is believed to be responsible for increased ductility and hence the observed mechanical property improvements. The results also demonstrate that quite reasonable mechanical properties can be obtained for blends of CGT with LLDPE using a simple coating process that still maintains composite processability. Ground Tire Bladder/LLDPE Blends To determine the effect that the composition of the ground rubber particles has on mechanical properties, blends of LLDPE with ground tire bladder (TB40) were prepared. The tire bladder, received from Custom Cryogenics, was a carbon black filled butyl rubber vulcanate ground to about the same particle size as the SR40 blends described previously. The instrumented impact properties are shown in Fig. 2e and listed in Table 3. Although the ultimate force and relative stiffness are similar to those for blends with the mixed tire scrap (SR40), the impact strength is seen to be considerably higher, and is even greater than that of the pure PE at higher TB40 loadings. The tensile elongation (Table 4) is also much higher than for the SR40 blends, although it is still well below that of the pure PE. The tensile strengths are similar for both TB40 and SR40 blends. Table 6. Impact Properties of LLDPE/EAA Blends. Ultimate Impact Force Strength (N) (kJ/m) LLDPE 1350 5.19 95% LLDPE/5% EAA 1334 4.19 90% LLDPE/10% EAA 1342 4.91 83.40% LLDPE/16.6%.EAA 1370 5.26 75% LLDPE/25% EAA 1472 6.17 EAA copolymer 1592 7.81 The impact failure is similar to the systems described earlier, but the amount of ductile deformation before tearing is considerably higher. The composite material is drawn out by the impact dart to higher elongations than a sample of pure LLDPE, resulting in an actual increase in the impact energy over pure LLDPE at 40 wt% TB. This difference in failure behavior compared with the composites with whole tire particulates may be explained in terms of the difference in cross-link densities between the tire bladder and mixed tire particles. Butyl rubber (normally a brominated butyl rubber), a random copolymer of isobutylene with a minor amount of isoprene (0.5 to 2.5%), is used for the elastomeric component of the tire bladder. Unlike the typical rubber components in a tire (natural rubber, polybutadiene, SBR, etc.), which have reactive sites on every monomer unit, the unsaturation in butyl rubber is widely spaced along a saturated, flexible hydrocarbon chain (14). The tire bladder, therefore, has a lower degree of cross-^inking than the rest of the tire components. As shown by the comparison of peel force for typical tire tread and tire bladder compounds with LLDPE, this results in substantially better adhesion (Table 5). It is this difference in adhesion that is believed to be the major factor contributing to the differences in mechanical properties of the TL40 and SR40 blends. The level of adhesion in these systems is, however, still quite low (compare with the adhesion of a styrenebutadiene-styrene block copolymer with LLDPE--Table 5). The lower cross-link density and differences in the carbon black filler loads may also result in a lower modulus for the TB particles compared to the SR particles. This could also contribute to the mechanical improvements. However, no difference in modulus of the particles is suggested upon comparison of the relative stiffness (Table 3) of the SR40 and TB40 blends. SEM analysis of TB40 blends again revealed no breakdown of the particles on blending. This indicates that quite reasonable mechanical properties can be obtained for blends of ground tire bladder in LLDPE, even with 200 to 300 !micro^m particles and only moderate adhesion. CGT and Ground Tire Bladder/HDPE Blends Blends of HDPE with CGT and with ground tire bladder were prepared in order to assess the effect of a less ductile matrix on the composite properties. As shown by the load displacement curves in Fig. 4 (data in Table 7), the impact properties are sharply decreased upon the addition of CGT to the HDPE. The deleterious effect of the CGT particles is more pronounced than for blends of the CGT with LLDPE, although the observed trends are similar. Even for the TB40 blends, which show moderate adhesion (Table 5), there is still a substantial drop in mechanical properties, whereas there is an increase in impact strength in blends with LLDPE. Precoating the particles with EAA copolymer results in much smaller mechanical property improvements than in the LLDPE blends. For pure LLDPE, impact failure is seen to be a ductile yielding process in which the dart draws the material out as it passes through (12). In contrast, the failure of the pure HDPE, although it involves some plastic deformation, is observed to occur through catastrophic propagation of a crack through the impact zone. This difference in impact failure is believed to be responsible for the poorer properties of the CGT-HDPE composites. For LLDPE, where failure is ductile, large particles with moderate adhesion are easily tolerated (see TL40-LLDPE blends). In order to induce a transition from brittle to ductile failure in rubber toughening of semi-brittle materials, Wu !15^ has suggested that the distance between particles must be below a critical value. In HDPE/CGT composites, however, the failure remains semi-brittle because particles are too large to induce a brittle-to-ductile transition. Failure then occurs largely through crack propagation, and the large particles, even with moderate adhesion, act as serious flaws, providing an easy path for the crack to follow. The addition of CGT to a semi-brittle matrix is therefore believed to require much higher levels of adhesion (to retard crack growth at the particle/matrix interface), or much lower particle sizes (to lower the brittle-ductile transition temperature). CONCLUSIONS It is seen that the addition of CGT to LLDPE and HDPE results in a substantial drop in mechanical properties. Reducing the CGT particle size improves mechanical properties somewhat, but generally is associated with higher grinding costs. For LLDPE, however, the reduction in properties can be largely overcome by precoating the CGT particles with an EAA copolymer. In this manner, blends of LLDPE with 40 to 50 wt% CGT are obtained that have impact properties approaching those of the pure PE while retaining adequate processability. These property improvements are believed to be due to an interaction between the carboxylic acid groups on the EAA copolymer and functional groups on the CGT surface, which result in increased adhesion and greater ductility. For blends with a ductile (LLDPE) matrix it was found that with a large particle size only moderate adhesion is necessary to produce composites with useful mechanical properties. In a semi-brittle (HDPE) matrix, however, with the large particle sizes and moderate adhesion, poor mechanical properties are obtained. This believed to be a result of the large particles' providing an easy path for cracks to propagate through the composite, and it is thought that either higher levels of adhesion (to control crack propagation) or much smaller particles (to lower the brittle-ductile transition temperature) are necessary to obtain worthwhile properties in these systems. TABULAR DATA OMITTED The utilization of these findings requires consideration of the incremental costs of the grinding and the blending processes, and the added compatibilizer. Commercialization would be dependent on balancing these costs against the material performance properties. ACKNOWLEDGMENTS The authors are grateful to the Ontario Centre for Materials Research (OCMR) and Imperial Oil Ltd. for financial support of this project and Custom Cryogenics (Waterford, Ont., Canada) for providing ground rubber samples. REFERENCES 1. F. G. Smith and W. B. Klingensmity, Presented before Rubber Div., Amer. Chem. Soc., Washington, D.C. Oct. 9-12, 1990. 2. J. Paul, Encyclopedia of Polymer Science and Engineering, Vol. 14, 2nd ed., H. Mark, ed., p. 787, John Wiley & Sons, New York (1986). 3. A. A. Phadke, S. K. De, and S. K. Chakraborty, Rubber Chem. Technol., 57, 19 (1984). 4. A. A. Phadke and S. K. De, Polym. Eng. Sci., 26, 1079 (1986). 5. N. C. Hilyard, S. G. Tong, and K. Harrison, Plast. Rubber Proc. Applic., 3, 315 (1983). 6. E. L. Rodriguez, Polym. Eng. Sci., 28, 1455 (1988). 7. M. Duhaime and W. E. Baker, Plast, Rubber Compos. Proc. Appl., 15, 87 (1991). 8. D. Tuchman and S. L. Rosen, J. Elast. Plast., 10, 115 (1978). 9. F. J. Stark, Jr., and A, Leigton, Rubber World, 12, 36 (1983). 10. G. Koski, SPE ANTEC Tech. Papers, 34, 1799 (1988). 11. S. Wu, Polymer Interface and Adhesion, M. Dekker, New York (1982). 12. T. Liu and W. E. Baker, Polym. Eng. Sci., 31, 753 (1991). 13. T. Liu and W. E. Baker, Polym. Eng. Sci., 32, 944 (1992). 14. H. C. Wang, R. H. Schatz, and E. N. Kresge, Encyclopedia of Polymer Science and Engineering," Vol. 8, 2nd ed., p. 423, H. Mark ed., John Wiley & Sons, New York 1986). 15. S. Wu, Polymer, 26, 1855 (1985). Article A13796879