Mechanics of Composite Materials. Vol. 31. No. 4, July-August, 1995
CHARACTERISTICS OF STRETCHING OF HIGH-IMPACT P O L Y S T Y R E N E AND P O L Y C A R B O N A T E C L O S E T O T H E GLASS TRANSITION TEMPERATURE
M. G. Tsiprin and L. A. Irgen (Jirgens)
A study has been made of the stress-strain characteristics of high-impact composites based onpolyso'renes (PS) or polycarbonates (PC) in uniaxial stretching at temperatures close to the glass transition temperature of the matrix material. Heterogeneity of the structure of the material at the microlevel, owing to the presence of a graft copolymer in the PS composites or the presence of polyethylene terephthalate (a reactive additive) in the PC, tends to lower the effective glass transition temperature as determined at high elongations. Weak interaction of the impact-strength modifiers with the matrix material (determining the heterogeneity of the structure at the macrolevel) results in a lower stability of uniform stretching in the region of the high-elastic state, but increases the values of the elongation at break in the transition region from the glassy state to the high-elastic state. From the standpoint of achieving high strength and deformability of high-impact composites during their processing, the structure shouM be organized so as to provide optimal heterogeneity at both the microlevel and macrolevel.
The use of polystyrenes (PS) and polycarbonates (PC) as materials of construction is often limited by the low impact strength of the PS  or the inadequate impact strength of the PC for certain applications ; the impact strength is highly sensitive to the shape and depth of the notch, the test temperature, and the specimen thickness. End-items made of PS or PC are subject to stress cracking. These disadvantages have been overcome by the development of high-impact polymer/polymer composites. In these composites, the polymer used to modify the impact strength may be formed as a disperse phase by modifying the copolymerization conditions ; if the modifier is incorporated by mechanical mixing, it may be either an additive that promotes compatibility  or it may be a reactive polymer . The compatibility indexes of the components of high-impact composites not only influence the service characteristics [4, 6], but also determine the conditions of stress transmission between the phases in processing the materials. Weak interfacial adhesion may lead to the formation of vacuoles during the stretching of elastomers ; when stretching plastic polymers, weak adhesion may lead to irreversible breakdown of the disperse phase, showing up in poorer service characteristics and reduced strength of joints after processing by pressure molding . Therefore, it has been of interest to investigate the features of uniaxial stretching of high-impact PS and PC composites in relation to interaction of the components at temperatures near the glass transition temperature, such that large strains are realized. We tested high-impact polystyrenes (HIPS) of two commercial grades: UPM-07031~ (Standard GOST 28250-89) and 475K (produced by BASF); these products are obtained by copolymerizing styrene and rubber. PC composites were obtained from a melt of commercial polymer PK-3S based on diphenylolpropane (subsequently known as KI) by mixing with impactstrength modifiers: Composite K2 contained 5 % low-density polyethylene (LDPE); composite K3 contained 5 % of the ionomer Serlin 1650, manufactured by Du Pont; composite K4 contained 30% polyethylene terephthalate (PET) and 10% butadiene-styrene copolymer. Certain physicomechanical characteristics of the PC composites have been reported in . The experiments on uniaxial stretching with a constant strain rate were performed in a tension viscometer  within a temperature range T = 75-125~ for the HIPS and 135-170~ for the PC composites. Cylindrical specimens were prepared Institute of Polymer Mechanics, Latvian Academy of Sciences, Riga LV-1006, Latvia. Translated from Mekhanika Kompozitnykh Materialov, Vol. 31, No. 4, pp. 509-517, July-August, 1995. Original article submitted April 18, 1995. 0191-5665/95/3104-0371512.50
Plenum Publishing Corporation
,, I |00
Fig. 1. Temperature dependences of stress at break ob (., o ), elongation at break eb ( , , V), and initial modulus of elasticity E 0 (A, /,) of HIPS specimens UPM-0703t~ (continuous curves) and 475K (broken curves). by extrusion in an instrument normally used to measure the melt index. The tests were performed on specimens immersed in a medium of PMS-200 silicone fluid; the test temperature was held within +0.5~ Tensile stresses were determined as o = crnX = FX/S 0, where a n is the nominal stress; F is the tensile force; X is the elongation ratio; SO is the initial cross section of the specimen. Simultaneous measurements were made of the instantaneous values of the diameter, from which we calculated Xs = S0/S, where S is the instantaneous cross section. The sampling frequency of the sensors was matched to the strain rate in such a manner that measurements were performed every 0.02 strain unit e = In X. Values of the initial modulus of elasticity. E0 were determined by approximating the first five measurements of a(e) by a quadratic parabola, using the least squares method. The uniformity of deformation was judged on the basis of coincidence of the values of the strain rate k = de/dt with the value of ks = d In Xs/dt. Any difference between the values of ~(e) and ~s(e) indicates localization of the deformation as a result of loss of stability of uniform elongation. We defined the failure point as the stress % and strain e b = In Xb at the moment of a visible break in continuity of the specimen, or the point of the prebreak maximum on the q(e) curve in plastic failure, since in this case the appearance of a maximum on a(e) indicates disruption of the uniformity of deformation . We also determined the reversible strain accumulated at the moment of break eeb = In Xeb. For these determinations, the broken specimen, or the specimen caught at the moment of reaching the maximum on a(e), was annealed at 130~ for 30 min, and its cross section Seb was measured (keb = S0/Seb ). In Fig. 1 we have plotted the temperature dependences of E 0, o b, and e b of the HIPS specimens that were investigatedl In the region T < 100~ the tests were performed at a strain rate ~ = 10 -2 sec -1. Data in the region T ~ 100~ were obtained from measurements at T = 100, 110, and 125~ in a range of ~ from 10 -3 to 1 sec -1, using the method of temperature-rate analogy. Judging from the plots of E0(T), the beginning of the transition from the glassy state to the high-elastic state, determined from the low-temperature break on log Eo(T) (minimum of the second derivative) is observed at a glass transition temperature Tg = 94 ~ and 95~ for the HIPS specimens 475K and UPM-0703t~, respectively; on the basis of dilatometric measurements, the respective values of Tg are 92 ~ and 93~ The differences between the values of Tg obtained by the two methods are evidently related to the dependence of Tg on e, which, on the basis of the temperature dependence of the temperature-time reduction factor, can be estimated at 2~ per decimal order of ~ at 95~ Copolymers of styrene with rubber typically exhibit a shift of the glass transition temperature relative to Tg of the PS matrix (100~ this shift being proportional to the content of rubber and amounting to about 1.5~ per 1% rubber . On this basis, the content of rubber in the HIPS test specimens is 5-6%, which is consistent with results reported in . In spite of the closeness of values of Tg of the two types of HIPS that were investigated, we can note (Fig. 1) that the passage through the glass transition temperature Tg is more drawn out for the 475K in comparison with the UPM-0703I~; this 372
o, M P a
tg Eo, lgS~ [MPa]
Fig. 2. Envelopes of breaks % vs Xb ( - * - , - o - ) and % vs Xeb ( - - - - - ) , and relationships o(X), of HIPS UPM-0703I~ ( - * ) and 475K ( - . - o ) at T > 100~ with ~ from 10 .3 to 1 sec -1. Tangents to or(X) determine the limits of stability of uniform stretching of elastoplastic bodies. Fig. 3. Temperature dependences of initial modulus of elasticity E 0 ( o ), stress at break % (~x), and elongation at break ,3b (o) for composites based on PC. Continuous curves are for K1, dashed curves for K2, dot-and-dash curves for K3, and double-dotand-dash curves for K4. difference is probably related to the greater degree of loosening of the polystyrene matrix by the rubber phase, in view of the higher content of the graft copolymer . The second break on the curve for log E0(T), which is the upper limit of the glass transition, occurs at Tg' = 103 and 102~ for the HIPS specimens 475K and UPM-0703t~, respectively; this corresponds to completion of the passage from the glassy state to the high-elastic state. The rate dependence of the temperature of this transition is defined by a value of 3~ per decimal order of k. Let us note that the relationship eb(T) has a maximum at T = Tg', whereas %(T) has a maximum approximately midway between Tg and Tg'. Thus, the results from a comparison of the relationships E0(T) obtained at small strains indicate greater heterogeneity in the structure of the HIPS grade 475K in comparison with UPM-0703t~. This conclusion follows from the slightly broader temperature range of the passage from the glassy state to the high-elastic state and the greater diffuseness of the transition near Tg. The differences in the relationships E0(T), however, are not too great. More significant are the differences observed in the relationships eb(T) and Ob(T) at T < Tg. Here we note a minimum on the curve for %(T) coinciding in temperature position with the break on the gb(T) curve, related to unfreezing of mobility of macromolecule fragments in the polymer matrix and realization of a mechanism of induced elasticity. At low test temperatures, the values of e b are small (% = 0.2 at T < 85~ and the true stresses are close to the nominal stresses. With increasing realization of the induced elasticity mechanism as the temperature is raised, the elongation at break increases sharply; with the nominal stresses at break decreasing monotonically, this gives an increase in values of %. As a result, a minimum is formed on the curve for %(T); the temperature at this minimum, the same as the temperature of the break on eb(T), can be defined as an effective glass transition temperature Tge at high stresses. For the UPM-0703t~, Tge = Tg; for the 475K, however, Tg - Tge = 5~ This result obviously indicates more effective utilization of the rubber in the 475K as a result of a greater degree of grafting of the PS to the rubber, as evaluated on the basis of the ratio of concentrations of gel fraction and rubber . The presence of graft copolymer hinders
close packing of the molecular chains of both the matrix PS and that occluded in the gel fraction at T _< Tg ; this favors realization of the induced elasticity mechanism. In the region of passage from the glassy state to the high-elastic state of the matrix, the increase of density of the threedimensional network of entanglements (primarily in the gel fraction) that will result from increasing the concentration of graft copolymer may hinder the development of large strains. This follows from the data of Fig. 1 and also Fig. 2, where we have shown the envelopes of the limiting characteristics % vs k b and o-b VS ~keb of these two types of HIPS, obtained as a result of tests at T > 100~ over a wide range of variation of e. Since in the region of the high-elastic state the relationships ab(/h T), kb(k, T) and E0(~, T) are all subject to a single temperature-rate reduction function, the envelopes and the relationships a(X) that are "tied" to them (see Fig. 2) proved to be invariant with respect to the temperature-rate conditions of test. It can be seen that in the region of the maxima of %(Xb), where the relationships o(k) are similar to those for cross-linked elastomers, the increased content of graft copolymer in the HIPS 475K does actually lead to a decrease of the stresses % in comparison with those for the UPM-0703I~. For some of the relationships a(k), it proved to be possible to pass a tangent starting from the point o = k = 0. At the point of tangency, with k = k*, & r / O k = a / k , or 0(ln o)/O(ln k) = 1, or 0F/0t = 0. For elastoplastic materials, these conditions determine the limits of stability of uniform stretching [12, 13]. As can be seen from Fig. 2, in the region of the high-elastic state (on the branches of the envelopes % vs Xb to the right of the maximum %), the magnitude of X* is equal to ~keb, within the experimental error. This fact can be interpreted to mean that at k = X* -- )keb there is a breakdown of stability of deformation of those structural elements of the medium that correspond to reversible strain. In the present case, these are primarily the gel fraction of the rubber. Since the values of )keb in the region of the high-elastic state for the HIPS specimens in the comparison are virtually equal, we can conclude that with an increase of temperature or a decrease of the strain rate, the graft copolymer has less influence on the ductility of the HIPS, which is then determined primarily by the properties of the matrix and the concentration of the rubber. Here, a substantial contribution to the overall strain is made by flow deformation, and the specimen failure proceeds not as a result of exceeding the ultimate strength, but rather as a consequence of loss of stability of uniform stretching when Xb > k*. In the region of the quasiglassy state (on the branches of % vs Xb to the left of the maximum %), with k _> k*, a neck is formed in the course of stretching the specimen. From this discussion it follows that the deformability of HIPS in the temperature region close to Tg is largely determined by the dual-level organization of their heterogeneous structure. The presence of a graft copolymer, creating heterogeneity at the level of molecular structure (microlevet), predetermines realization of large strains in the glassy state of the matrix, but lowers the strength in the region of passage from the glassy state to the high-elastic state, thus limiting the deformability of the macroelements of the structure, i.e., the disperse phase in the form of a gel fraction. The main difference between the PC composites and the HIPS is that the PC is a far more plastic polymer than the PS at T < Tg. The impact strength indexes and the values of eb for these two types of materials differ by a factor of more than 10. In addition, the disperse phase of the HIPS is a chemically cross-linked structure, whereas the disperse phases of the PC composites within this range of temperatures (below Tg) are capable of unlimited flow, at least in principle, since their melting point or glass transition temperature is lower than the corresponding values for the matrix PC. All of this predetermines the features of the stress--strain curves of the PC composites. In Fig. 3 we show the temperature dependences of % and e b of the matrix PC, and also the temperature dependence of E0, for all of the PC composites that we have examined. Attention is drawn to the decrease of the modulus of elasticity E0 at T < Tg with increasing polarity of the modifying additive and increasing capability for chemical interaction with the matrix polymer. When speaking of reactivity in this case, we are referring to a reaction of interchain exchange between the PC and the PET, which could have been realized only in part, since the blending of the polymers was performed at T < 260~ For these composites, in the temperature interval between 45~ and Tg, the values of E 0 for the PC/PET blend are lower than for the PC, but higher than for the PET; if T > Tg, the reverse picture is observed, with values of Eo for the PET higher than for the PC ; in the region close to the glass transition temperature, the value of E o of the blend may be smaller than the modulus of elasticity of its components . These relations are reflected in the unusual temperature dependence E0(T) for composite K4, where we observe a tendency to form a minimum at the glass transition temperature of the PC. For all of the PC composites, the same as for the PS, at T -- Tg' there is a maximum on eb(T); but in contrast to the HIPS, the relationships %(T) are monotonic in the vicinity of Tg -- behavior that is obviously related to the relatively high level of deformability of PC at T < Tg.
211gl5 [MPa] 1
lg(5~ ~lPal 1,5
Fig. 4. Stress-strain relationships for PC composites K1 (continuous curves), K2 (dashed curves), K3 (dot-and-dash curves), and K4 (double-dot-and-dash curves) at T = 140 (1), 150 (2), or 160~ (3). The straight lines correspond to 0(ln cr)/0e = 1; the points indicated on these lines denote the points of breakdown or loss of stability. Fig. 5. Envelopes of failure of PC compositions K1 (continuous curve), K2 (dashed curve), K3 (dot-and-dash curve), and K4 (double-dot-and dash curve). In Fig. 4 we show curves for a(e) of the PC-based composites with T < Tg, with T > Tg', and in the region between Tg and Tg'; also, lines have been drawn corresponding to the condition 0(ln a)/Oe = 1 and defining the stability of stretching of elastoplastic bodies. With T < Tg, stretching of the matrix PC is accompanied by the formation of a neck that appears when 0(ln o)/Oe = 1 in the region of the local maximum on a(e), and stabilized under this same condition in the region of the local minimum on a(e). The ratio of the specimen cross section on the defect-free section to that of the neck (ratio of natural drawing) is proportional to the ratio of stresses at the sites of the local extrema. Analyzing the influence of the impact-strength modifiers from this point of view, we can note that the additives lower the natural drawing ratio in the series LDPE > ionomer > PET and that they tend to improve the stability of uniform deformation. The elongation of the K4 specimens is stable over the entire range of temperatures investigated. The c~(~) curves for T < Tg are positioned lower, and those for T > Tg higher, than the analogous curves for the K1, K2, and K3 composites. This behavior can be attributed to features of the structural organization of the K4 composite: The partially compatible mixture of PC and PET, which is heterogeneous at the microlevel of the matrix, also contains at the macrolevel an impact-strength modifier in the form of a disperse phase. For the high-elastic state with T > Tg', uniformity of deformation typically breaks down in the region of the maximum on a(e). As can be seen from Fig. 4, additions of impact-strength modifiers in the form of LDPE or the ionomer, while they do not change the basic character of G(e) in comparison with that of the matrix PC, do bring about a lowering of the maximum strain for stable deformation. This evidently takes place as a consequence of the high mobility of these additives and their weak bonding with the matrix PC, where they play the role of defects. What we have said regarding the PC-based composites is illustrated by the envelopes of failure shown in Fig. 5. It will be seen that additions of the LDPE or ionomer give substantial increases of deformability of the K2 and K3 composites in the region of the transition from the glassy state to the high-elastic state, but they lower the stability of uniform stretching as the transition is made to the viscous flow state, in comparison with the matrix PC. For the K4 composite, the matrix of which is essentially a PC-PET melt, we find that in comparison with the PC it has a higher level of stresses % at comparable strains eb, indicating active participation of all components of the mix in processes of stress transmission. Thus, from an analysis of the entire set of stress-strain curves for high-impact PS and PC composites in the region of temperatures close to the glass transition temperature of the base polymer, we can conclude that the desirability of achieving high strength and deformability does impose somewhat contradictory requirements on the structure of the composites for the 375
region of temperatures above and below the glass transition temperature. The optimal structural organization, apparently, will be that providing heterogeneity at both the microlevel and macrolevel.
1. 2. 3.
4. 5. 6. 7. 8. 9. 10.
12. 13. 14.
A. Ya. Malkin, S. A. Vol'fson, V. N. Kuleznev, and G. I. Faidel', Polystyrene [in Russian], Moscow (1975). T.N. Andreeva, A. S. Kolerov, T. N. Vakhtinskaya, I. I. Solov'eva, and T. N. Yudakova, "High-impact materials based on polycarbonate," Plast. Massy, No. 2, 24-27 (1993). G . D . Ballova, T. A. Stepanova, and A. D. Chegodaeva, "Production of high-impact styrene copolymers by the block-suspension method," Obz. Inf. Nauchno-Issled. Inst. Tekh.-Ekon. Issled. Minist. Khim. Promst. SSSR, Moscow (1977). J.W. Barlow and D. R. Paul, "Mechanical compatibilization of immiscible blends," Polym. Eng. Sci., 24, No. 8,525533 (1984). M. Xantos and S. S. Dagli, "Compatibilization of polymer blends by reactive processing," Polym. Eng. Sci., 31, No. 13, 929-935 (1991). D.M. Bigg, "Mechanical properties of particulate filled polymer," Polym. Compos., 8, No. 2, 115-122 (1987). R.J. Farris, "The influence of vacuole formation on the response and failure of filled elastomers," Trans. Soc. Rheol., 12, No. 2, 315-324 (1968). Ya. Ya. Indulevich, L. A. Irgen, G. M. Kerch, D. I~. Pablak, and M. G. Tsiprin, "Automated system for determining rheological properties of polymer solutions and melts," Mekh. Kompoz. Mater., No. 1, 109-113 (1985). V . E . Dreval and E. K. Borisenkova, "Peculiarities of rheological behavior of filled polymer melts in uniaxial stretching," Rheol. Acta, 32, No. 4, 337-351 (1993). B . I . Shapiro, Yu. V. Nikitin, V. L. Budtov, N. G. Podosenova, V. M. Belyaev, V. A. Zinchenko, L. L. Sul'zhenko, and A. F. Nikolaev, "Influence of properties of polystyrene matrix on characteristics of high-impact polystyrene," Plast. Massy, No. 2 43-44 (1977). I . I . Perepechko and V. V. Nizhegorodov, "Relaxation processes in block copolymers of butadiene rubber and polystyrene," in: Mechanical Properties of Rubbers and Rubber/Cord Composites: Test Methods (Proceedings of International Conference IRC-94), Moscow (1994), Vol. 4, pp. 483-489. I.M. Ward, Mechanical Properties of Solid Polymers, Wiley, New York (1971). M.G. Tsiprin and L. A. Irgen, "Stability of longitudinal flow and stretchability of polymer melts," Mekh. Kompoz. Mater., No. 1, 105-114 (1992). V.A. Danilov, I. I. Perepechko, V. V. Nizhegorodov, Yu. A. Godovskii, E. V. Konyukhov, and S. I. Belousov, "Component compatibility and relaxation processes in mixtures of polycarbonate and polybutylene terephthalate," Vysokomol. Soedin. A, 36, No. 1, 69-74 (1994). I . I . Perepechko, V. A. Danilov, V. V. Nizhegorodov, N. P. Bessonova, and E. V. Konyukhova, "Viscoelastic behavior and relaxation processes in mixtures of polycarbonate and polybutylene terephthalate," Vysokomol. Soedin. A, 35, No. 2, 174-178 (1993).