Effects of processing conditions and copolymer molecular weight on the mechanical properties and morphology of compatibilized polymer blends.
Immiscible polymer blends are believed to offer the chance of combining the most desirable properties of each of the blend components while minimizing the poorer properties. In particular, the possibility of combining the high modulus of polystyrene with the ductility of PE, together with the wide availability of these two polymers, has resulted in much interest in PS/PE blends (e.g., references 1-6). It is generally accepted that the morphology of an immiscible polymer blend is that of a matrix phase within which the secondary phase is dispersed, usually as droplets. However, an immiscible mixture of, say, oil and water, is known to separate into two layers. An oil/water emulsion can be produced by the use of surfactants and intensive mixing, although, if left for a sufficient length of time, this mixture will also separate into two layers. By analogy, it might also be expected that immiscible polymer blends will separate into separate layers. We have seen evidence of such behavior in high molecular weight PS/polyisoprene blends (both with and without diblock copolymer) when solution cast from toluene (7). The fact that the majority of observed immiscible polymer blend morphologies are not layered suggests that these are not equilibrium morphologies. This means that the observed morphology will depend upon the way in which the blend has been prepared. This could well explain the many discrepancies between results reported in the literature by different groups.
The role of surfactants in polymer blends is often filled by block copolymers. In an attempt to reduce the number of unfavorable contacts between each homopolymer and the dissimilar blocks of the copolymer, an A-B copolymer capable of phase segregation will localize at the phase interface when added to an immiscible A/B homopolymer blend. In so doing, it causes a decrease in the interfacial tension between the homopolymers. For particular blending conditions, this has generally been found to result in a reduction in the average size of the dispersed size.
Brown, Creton, Kramer and co-workers have shown that in glassy-glassy systems, diblock copolymers can significantly increase the adhesion between immiscible polymer interfaces (8-11). They have shown that the effect is dependent upon the molecular weight of the copolymer blocks and on the areal density of the copolymer at the interface. A significant change in the reinforcing behavior of the copolymer occurs when the length of the blocks exceeds the molecular weight of entanglement of the corresponding homopolymer. The effects of copolymers on glassy-semicrystalline and glassy-rubbery interfaces have been less widely investigated and the correlations between block molecular weight and interfacial strength are not as well understood as for glassy-glassy systems.
Although the factors influencing the mechanical properties of two-phase polymer mixtures are still under discussion, it is generally accepted that for a composite in which the matrix phase deforms mainly by crazing (such as a PS matrix blend), there must be good adhesion between the blend components, and the diameter of the dispersed phase particles must be an optimum for the matrix of interest (12). The average size of the minor phase obtained when two immiscible polymers are simply blended is generally much larger than this optimum size. Therefore, it may be expected that the addition of a copolymer, which will result in a reduction in surface tension and hence a reduction in the average size of the dispersed phase plus an increase in the adhesion between the phases, will result in an improvement in the mechanical properties of the blend. As the emulsifying effects and the improvement in adhesion have both been shown to depend upon the molecular weight of the copolymer, the improvement in the blend mechanical properties might also be expected to be dependent upon the molecular weight of the copolymer.
We have shown in a previous paper that a series of semicrystalline PS-b-hPB diblock copolymers can improve the adhesion between a well-defined PS/PE interface (13). We showed that when the thickness of the copolymer layer placed between the homopolymers was less than that needed to saturate the interface, the trend in interfacial strength was similar to that reported by Brown, Creton, Kramer and co-workers for glassy-glassy systems. We also showed that when the copolymer was premixed with the PS layer it migrated to the PS/PE interface. In this paper we now investigate the effect of these same copolymers on the morphology and mechanical properties of a blend of the PS with the PE. The aim of this work was to find a correlation between the effects of diblock copolymers on the morphology and the mechanical properties of immiscible blends.
Table 1. Homopolymers Used in these Investigations. Polymer [M.sub.w] [M.sub.n] PS 278,300 104,800 PSd8 190,000 176,800 PE 43,700(*) 7800(*) * GPC values: LDPE equivalent molecular weights.
[TABULAR DATA FOR TABLE 2 OMITTED]
The molecular characteristics of the polymers used in these investigations are shown in Tables 1 and 2. Except for the deuterated PS, which was prepared by Professor R. W. Richard's group at the University of Durham, the polymers were provided by DSM Research, The Netherlands. The hydrogenous PS used is a general-purpose industrial polymer (Shell N7000). The PE is a low-modulus research grade similar to that described by Deblieck and co-workers (14). The copolymers were prepared by sequential anionic polymerization with the PS block being formed first. This resulted in "copolymers" containing a portion of terminated PS-blocks, i.e., homopolymer PS. We were not aware of this problem until the majority of the experiments had been completed and the copolymers were used as supplied (i.e., containing a proportion of homopolymer).
The PS:PE ratio was kept constant at 75:25 in all the blends. Most blends prepared contained 5% copolymer, i.e., the composition was 95% (75PS:25PE), 5% copolymer. The specimens were prepared by melt mixing and subsequent injection molding. The molecular weights of the blend components were found to be the same before and after processing within the error of GPC measurements, suggesting that no significant degradation had occurred.
The melt mixing was performed in a mini twin-screw extruder developed by Bulters, Elemans, and co-workers at DSM Research for the compounding of small batches of material (approximately 4 to 5 g). The extruder has conical co-rotating fully intermeshing twin screws approximately 100 mm in length and 10 to 15 mm in diameter. The material can be repeatedly passed through the extruder via a recirculation channel.
The extruder was filled at a low screw speed (approximately 50-100 rpm) and the filling procedure took approximately 30 seconds. Once all the polymer had been added to the extruder, the S/PE mixtures were blended for 35 [+ or -] 5 seconds at a temperature of 300 [+ or -] 2 [degrees] C using a screw speed of 250 [+ or -] 5 rpm. The recirculation times in the extruder are approximately 15 seconds and 40 seconds at 250 rpm and 100 rpm, respectively, so the blend will have recirculated at least three times before extrusion. Preliminary experiments on the PS/PE binary blend showed that under these conditions a melt temperature of 300 [degrees] C was necessary to ensure a good dispersion of the PE within the PS matrix. This was a consequence of the high molecular weight of the PS used, which was necessary to ensure that the matrix had reasonable mechanical properties. To minimize degradation, blending was performed under a nitrogen atmosphere. As will be shown below, these mixing conditions were sufficient to enable the copolymer to migrate to the PS/PE interface. As a further indication of the extent of mixing, these mixing conditions were found to produce a one-phase blend (as determined by the observation of a single glass transition temperature in the DSC trace) of PS and poly(phenylene oxide) when these were fed as powder/granules.
The extrudate was injection molded into the required shape using a prototype mini-injection molding apparatus developed at DSM Research. The blends were held in the melt chamber at 240 [degrees] C for approximately 90 seconds before being injected into the mold. The piston pressure of the injection molder is much lower than on most full scale machines. To enable the mold to be filled completely, a mold temperature of approximately 45 [degrees] C was necessary.
The specimens were not conditioned prior to testing, but were prepared at least one day before testing and were stored at normal room temperature (23 [degrees] C).
Transmission Electron Microscopy
The samples were trimmed parallel to the direction of injection molding and then stained for 48 hours using ruthenium tetroxide. The ruthenium tetroxide solution was prepared from ruthenium (III) chloride hydrate and sodium hypochlorite. The oxidation process responsible for staining was stopped prior to subsequent sectioning by washing the specimens thoroughly in an excess of vitamin C solution followed by rinsing with distilled water. The staining process causes stiffening of the low modulus PE phase, which allows the samples to be cut at room temperature.
It was found that for these specimens, the PE phase (identified by the clearly visible crystallites) was more heavily stained by the ruthenium tetroxide than the PS phase. This result is in contrast to those of Trent and co-workers, which suggest that PS should be more heavily stained than PE (15). The reason for this discrepancy is unclear. However, as the precise mechanism of staining with ruthenium tetroxide is still unclear, and the PE used in these investigations is different from that used by Trent and co-workers, there are many possibilities.
TEM micrographs of sections of the order 70 nm thick were taken with a Philips CM200TEM using an acceleration voltage of 120 kV.
Small Angle Neutron Scattering
The specimen morphologies were also investigated using small angle neutron scattering (SANS) in a similar manner to that described by Clark and co-workers (16). The experiments were performed on the LOQ small angle scattering apparatus at the Rutherford Appleton Laboratory, UK. The experiments and problems of data interpretation are discussed in detail elsewhere (7).
The distance scale probed by the LOQ apparatus is 30-1000 [Angstrom], which is small compared with the phase dimensions (of the order of half to several microns) seen in TEM micrographs of the SANS specimens. As a consequence, the information obtained [from standard Porod plots (16)] is confined to a value for the total interface between the two phases (17). If the shape of the domains is spherical, this surface area (or as it is often presented, surface to volume ratio) can be translated into a particle radius, r, but for random morphologies this is not possible. For the results presented here therefore, the SANS information is an indication of the relative amounts of interface for the different samples.
Except for the number of repeat specimens (which varied from 1(for the binary blend only) to 5), the Izod impact tests were performed in accordance with ISO 180 ([2.sup.nd] Edition, 1993). Specimens with dimensions 63.5 x 12.7 x 3.2 mm were used. A notch of type A (i.e. angle = 45 [degrees], radius = 0.25 mm) was introduced into each specimen. Impact testing was performed at room temperature using an impact speed of 3.46 m/s and a 2.75 J hammer. The impact strength was calculated using the instrument computer package.
Except for the number of repeat specimens (which was 4 or 5 as opposed to a minimum of 5), the tensile tests were performed in accordance with ISO R527 ([1.sup.st] Edition, 1966). Unnotched specimens of dimensions half those of specimen type 1 were used. The tensile tests were performed at room temperature at a strain rate of 1 mm/min. The instrument computer packages were used to obtain the modulus, maximum force, yield stress, elongation at yield, stress at break, and elongation at break of the blends. The values were also checked manually for accuracy.
RESULTS AND DISCUSSION
As will be discussed below, it was found that the addition of 5% copolymer resulted in some cases in an improvement in the mechanical properties of the 75:25 PS/PE blend under investigation. Given this potential property improvement, the investigations reported here concentrate on blends containing 5% copolymer. In order to introduce sufficient contrast to be able to extract useful information from the scattering data, the SANS samples were prepared using a deuterated PS in place of the standard PS homopolymer.
Mixing and Morphology
The SANS data showed that the "particle" surface area to volume ratio (as obtained from Porod plots of the data) for the binary dPS/PE blend and blends modified by 5% 8hPB or 5% 101hPB scaled as 1:2:2.9 i.e. the high molecular weight copolymer 101hPB resulted in a larger increase in the interfacial area than the lowest molecular weight copolymer 8hPB. A similar trend in the SANS data (1:2:3) was obtained for these blends when co-precipitated from xylene into methanol and subsequently compression molded (7). The increase in interfacial area is a result of the copolymer locating at the Ps/PE interface and reducing the interfacial tension. Mixing in a common solvent (in this case warm xylene) results in intimate mixing of the polymers, and the copolymers will have sufficient opportunity to migrate to the interface. As the relative change in interfacial area in the melt-processed samples is the same as in the co-precipitated samples, the melt-mixing conditions used were clearly adequate to allow the copolymer to reach the interface and locate there. As discussed in the introduction, all polymer blends exhibiting structures other than a bilayer are, in fact, not at equilibrium and will exhibit different morphologies depending upon their processing history.
The morphologies of the melt-processed SANS samples were investigated directly using TEM (7). SEM of the fracture surfaces of the hydrogenous PS/PE blend extrudate had revealed that the blending conditions used resulted in the minor PE phase being well dispersed within the PS matrix (7). In contrast, the SANS binary blend specimen exhibited an almost co-continuous morphology. Addition of 5% 8hPB to the dPS/PE blend resulted in phase inversion such that the PE formed the matrix phase. Addition of 5% 101hPB resulted in a blend with a PS matrix. The PE phases were relatively small (of the order 1[[micro]meter]) and irregularly shaped. Careful inspection of the micrographs revealed micelles within the PE phase of both the copolymer modified blends (these would not be very evident in the SANS results because both the copolymer and PE phase contained no deuterium labeling). There are several reasons for believing that the dark circles observed in the micrographs are micelles (7): (i) such structures were not observed in any of the binary PS/PE blends, (ii) such regularly sized spherical structures are not the expected form of staining artefacts, (iii) the diameters of the spheres (approximately 200 to 800[Angstrom]) is in the range expected for copolymer micelles, and (iv) the diameter of the spheres can be shown to scale with the molecular weight of the copolymer.
Phase inversion in compatibilized blends has been investigated by Adedeji and co-workers for solvent east samples (18). They suggest that in ternary blends containing a diblock copolymer the overall interfacial curvature is determined by the balance between interfacial swelling of the block copolymer segments by the homopolymers on each side of the interface. The interfaces become curved when swelling on one side is more than on the other. It can easily be imagined that the more highly swelled side should form the matrix phase. Adedeji and co-workers control the relative degree of swelling on each side of the interface (and hence control which polymer forms the matrix phase) by varying the chemical composition of one of the homopolymers in an A/A-B/C ternary blend, thus varying the degree of exothermic mixing at the interface. For A/A-B/B ternary blends the degree of swelling will be determined by the relative molecular weights of the homopolymers and the blocks of the copolymer, i.e., how "wet" the brushes formed by the copolymer are. In Tables 1 and 2 it can be seen that the PS blocks of all the copolymers investigated here have molecular weights lower than the molecular weights of either of the PS homopolymers used. Therefore, the brushes formed by the PS-blocks of the copolymers may all be expected to be dry. The number average molecular weights of the hPB blocks of all the copolymers are greater than that of the PE homopolymer, whereas only the weight average molecular weights of copolymers 42hPB and 101hPB are greater than that of the PE homopolymer. Therefore, depending upon which molecular weight average determines the degree of swelling (because of the narrow molecular weight distributions generally used by researchers investigating this problem, this is not clear) it is expected that either all the copolymer modified interfaces will be swelled on the hPB side or only those interfaces modified by copolymers 42hPB and 101hPB. The copolymers that result in swelling on the PE side of the interface are expected to cause phase inversion such that the PE homopolymer forms the matrix phase. This would mean at least copolymers 42hPB and 101hPB, but perhaps all the copolymers, are expected to cause phase inversion. However, it has been seen above for the SANS specimens that copolymer 8hPB caused phase inversion and copolymer 101hPB did not. Also, TEM micrographs of the Izod specimens will illustrate below that addition of copolymer 8hPB or 21hPB result in phase inversion, copolymer 42hPB results in a co-continuous morphology, and copolymer 101hPB results in a blend with a PS matrix. These results indicate that the observed phase inversion is not a thermodynamic effect.
An alternative explanation depends on the melt rheology. When melt mixed, it has been suggested, the blend composition at which phase inversion occurs is dependent upon the relative viscosities of the polymers making up an immiscible blend (19, 20). The copolymers investigated have a wide range of molecular weights and might hence be expected to modify the blend viscosity in different ways. One way of determining this is to note the forces involved in the mixing processes. If the volume of polymer and the screw [TABULAR DATA FOR TABLE 3 OMITTED] speed in the mini extruder are kept constant during melt mixing of the blends, the measured force on the screws is dependent upon the viscosity of the melt within the extruder for a particular extruder geometry. For a particular blend composition, the force was observed to be surprisingly constant (for a measured force of 575 a.u., variation = [+ or -] 26 a.u. for a series of batches with the same composition). Figure 1 illustrates the variation in force with copolymer molecular weight for the PS/PE blends compatibilized with 5% copolymer. Except for the lowest molecular weight copolymer (8hPB), the force can be seen to increase steadily with copolymer molecular weight. The relationship between blend rheology and morphology is complicated because each is dependent upon the other. From this and previous work (13), the copolymers used here as compatibilizers are known to be interfacially active, and as discussed above we have seen that excess copolymer forms micelles within the PE phase. So the copolymer may be affecting blend rheology in one of two ways: by modification of the interface or by modification of the viscosity of the PE phase. By which mechanism the addition of copolymer affects blend rheology and morphology will be discussed in more detail in a future paper (21). From the results presented here, it is, however, clear that the change in blend rheology is more complicated than merely a change in the interfacial tension and that this results in more complex morphologies than anticipated. A consequence of this is that each type of sample prepared for the mechanical tests described next may have a different morphology even for identical compositions.
The tensile results for the homopolymer PS, the binary PS/PE blend and the copolymer modified ternary blends are listed in Table 3 and illustrated in Figs. 2 and 3. In Fig. 2 it can be seen that addition of all the copolymers improves the elongation at break of the blends to a similar degree. Figure 3 illustrates that addition of all the copolymers except 8hPB results in an improvement in the maximum force withstood by the specimens prior to break of yielding. The modulus and stress at break appear to follow no obvious pattern (see Table 3). Assuming that the morphology of the tensile specimens are similar to those of the SANS specimens (which is reasonable as the molds used for injection molding were of similar dimensions), the trends in the various tensile parameters cannot be explained merely in terms of which polymer forms the matrix phase.
PE is a low modulus, ductile material, which may be expected to exhibit a low yield stress but very high elongation at break (of the order of several hundred percent). Addition of 25% PE to PS can be seen from Table 3 to have no significant effect on the elongation at break, suggesting that the matrix phase, i.e., PS, determines this property. However, addition of copolymer to the binary blend can be seen from Fig. 2 to increase the elongation at break. It was shown that the ternary blends have morphologies ranging from PE matrix to PS matrix. This trend cannot be linked to that observed in elongation at break. In a previous paper we showed that all the copolymers improve the adhesion between the PE and PS phase (13). Thus these elongation-at-break results indicate that for the secondary phase to influence this property, there must be sufficient adhesion between the two immiscible phases. As the high molecular weight copolymers are expected to increase the adhesion between the phases to a greater degree than the low molecular weight copolymers, and no corresponding trend is observed in the elongation at break with copolymer molecular weight, it may be assumed that the improvement in degree of adhesion that can be considered sufficient is no greater than that provided by a low molecular weight copolymer.
Table 3 shows that the binary blend is able to withstand a much lower force before yielding than the pure PS. This is a usual disadvantage of adding a low modulus secondary phase. Figure 3 illustrates that the effect of copolymer addition on this tensile property of the blend is dependent upon the copolymer molecular weight. Again, the polymer forming the matrix phase appears not to be the determining factor, and the observed trend may be linked to the improvement in interfacial adhesion expected when a copolymer is added to the blend. If it is remembered that the hPB blocks of all the copolymers are much longer than Me(hPB) and that Me(PS) [approximately equal to] 20,000, Fig. 3 suggests that for stress to be transferred from the weaker PE phase to the "stronger" PS phase (thus increasing the force that the blend can withstand before failing), good adhesion between the PS and PE phases is required. The PS block of 8hPB, the lowest molecular weight copolymer, has a molecular weight approximately half Me and is expected to pull out of the PS homopolymer phase relatively easily because of the lack of entanglements between the PS blocks and the homopolymer chains. The PS blocks of the higher molecular weight copolymers 42hPB and 101hPB have molecular weights twice or five times Me and are expected to be entangled with the PS homopolymer chains. The interface is expected to fail only if the copolymer chains fracture. This requires much more energy than chain pull-out and is expected to result in a much improved interfacial adhesion. Copolymer 21hPB, whose PS block has a molecular weight approximately equal to Me(PS), is expected to have PS blocks that are "just entangled" with the PS homopolymer chains, so the interface will be intermediate in strength to those reinforced by the low or high molecular weight copolymer extremes. The trend in force appears to fit this expected trend in interfacial strength well. Such a clear trend was not observed for the peel tests samples reported previously (13). However, the peel test specimens were prepared by placing a thin copolymer film of constant thickness between the PS and PE phases so the number of copolymer "stitches" across the interface was inversely proportional to the copolymer molecular weight. Thus the degree of saturation in the PS/PE peel test interfaces was dependent upon the molecular weight of the copolymer used. In contrast, in the ternary blends investigated in these tensile tests, an excess of copolymer is available, so the interface can become saturated with copolymer.
As shown in Fig. 1, the variation in blend modulus with copolymer molecular weight is similar to that seen for the effect of copolymer molecular weight on the force measured on the extruder screws during mixing. This similarity suggests that similar morphological factors (e.g. phase size, degree of Interfacial adhesion) affect both these blend properties. If the modulus of a blend is considered as its resistance to applied force, it becomes clear why the two trends are similar in shape.
At present, there appears to be no explanation for the variation in stress at break as a function of copolymer molecular weight. For materials with similar morphologies it might be reasonable to expect the stress at break of the material to be dependent upon similar morphological parameters as the maximum force. From these results this appears not to be the case.
Izod Impact Tests
Specimens with similar compositions to the above SANS and tensile specimens were prepared for Izod impact testing. In Table 1 it can be seen that the deuterated PS used in the SANS experiments has a lower weight average molecular weight and lower polydispersity than the PS used for mechanical tests. This will result in the dPS having a different, most probably lower, viscosity, and may also influence the interactions between the copolymers and the PS phase. The SANS (and tensile) specimens are also much thinner than the Izod specimens (1.5 mm as opposed to 3.2 mm), which will result in a different shear field being applied to the blend during injection molding.
The morphology of the Izod specimens was investigated using TEM. It was not possible to use SANS as the samples were too thick and there was insufficient contrast between the hydrogenous polymers. Figure 4 illustrates the morphology of the PS/PE binary blend Izod specimen. The PE homopolymer forms the disperse phase. Figures 5 to 8 illustrate the effects on PS/PE blend morphology of adding 5% copolymer of various molecular weights. Addition of copolymers 8hPB and 21hPB can be seen to result in phase inversion such that the PE minor phase forms the matrix. Addition of copolymer 42hPB results in a co-continuous morphology and addition of copolymer 101hPB results in a morphology similar to that of the binary blend except that the PE phases are more highly elongated. Careful inspection of the micrographs again reveals micelles of the copolymer within the PE phase.
The dimensions of the PE phases illustrated in Figs. 4 to 8 are much larger than those observed in the TEM micrographs of the SANS samples. The changes in morphology with sample dimensions and major component molecular weight observed between the Izod and SANS samples clearly demonstrate the strong influence of shear rate and relative viscosity on the sample morphology.
Figure 9 illustrates the effect of various concentrations of either copolymer 8hPB or copolymer 101hPB on the Izod impact values measured for the ternary blends. The point on the y-axis (copolymer concentration = 0) is that measured for the binary PS/PE blend. It can be seen that addition of up to 5% 101hPB has no significant effect on the impact properties of the blends. Addition of 1% 8hPB also has no significant effect on blend impact strength; however, addition of 5% 8hPB results in an improvement in impact strength.
As is illustrated in Figs. 5 and 8, the 5% 8hPB modified blend comprises a PE matrix with PS inclusions, while the 5% 101hPB modified blend has a PS matrix and PE secondary phase similar to the PS/PE binary blend. PE is much tougher than PS [a similar PE to the one used for these investigations has been found to exhibit an Izod impact strength greater than 60kJ/[m.sup.2] (22), while the PS used has an impact strength of approximately 2.1kJ/[m.sup.2]]. Hence a material with a PE matrix might be expected to be tougher than a material with a PS matrix. TEM micrographs of the 1% 8hPB modified blend illustrated that no phase inversion occurred, i.e., PS forms the matrix. Therefore, from the above arguments the impact strength of the 1% 8hPB modified blend might be expected to be similar to that of the binary blend as is the case. A similar argument can be applied to the 16% 8hPB modified blend; it is not to be expected that further modification of the blend viscosity will result in a reversion of the morphology to PS matrix. The impact strength might therefore be expected to be similar to that of the 5% 8hPB modified blend if the impact strength is dominated by the blend morphology in these blends. This is indeed the case.
Figure 10 illustrates the effect of varying the copolymer molecular weight on the impact strength of ternary blends containing 5% copolymer and compares the values obtained with those measured for pure PS and for the PS/PE binary blend (point on vertical-axis; block molecular weight = 0). The observed trend in impact strength is the exact reverse of that expected if the copolymer molecular weight is determining the impact strength via an increase in interfacial strength. The trend can again be explained in terms of blend morphology. In Figs. 5 to 8, the morphology of the blends can be seen to change from PE matrix with relatively small PS inclusions for the 8hPB modified blend to PE matrix with large PS inclusions for the 21hPB modified blend to co-continuous for the 42hPB modified blend and finally PS matrix for the 101hPB modified blend. The trend in impact strength suggests that as soon as the PS forms a continuous phase in the blend, the impact strength of the blend becomes similar to that of the pure PS. An explanation for this might be that catastrophic cracks can then travel relatively unhindered through the material, resulting in brittle behavior.
SUMMARY AND CONCLUSIONS
It has been shown that as a result of their interfacial activity, semicrystalline PS-b-hPB copolymers are capable of compatibilizing the PS/PE blends investigated, thus resulting in an increase in the average specific area between the phases. It has been shown, however, that rather than a mere change in the average dimensions of the disperse phase, depending on molecular weight, addition of the copolymers can result in complete changes in blend morphology, including phase inversion. It is suggested that this is a consequence of the effects of the copolymers on blend rheology during processing. The differences in phase sizes observed for blends with similar compositions but injection molded into specimens of different thickness (1.5 mm and 3.2 mm) adds weight to this argument. Generally it was found that very little has been published on the rheology of copolymer modified immiscible blends. As this has a large influence on the morphology of melt-mixed blends, this is an important area that warrants much more attention.
For these blends, the elongation at break and the maximum force withstood by the blends prior to failure as measured by low rate tensile tests were found to be controlled by the expected strength of the interface and fairly independent of blend morphology. In contrast, the impact strength has been shown to be independent of interfacial strength, but rather controlled by gross phase morphology. The dependence of tensile and impact properties on different morphological properties of such blends is interesting in itself and has not been widely discussed in the literature.
It is clear that a multiplicity of variables must be considered if the properties and morphologies of these blends are to be fully understood. This work has given only a preliminary idea of the complexity of the effects possible in the processing of compatibilized blends. What is clear from this work is that the much-quoted improvement of mechanical properties with increased interfacial strength is not as simple as often assumed and that before the effects of copolymers on immiscible blend morphology and mechanical properties can be fully understood, the effects of melt processing conditions on immiscible blend morphology must be better understood.
HEH's work was funded by EPSRC, UK, and DSM Research, The Netherlands, in the form of a EPSRC Research Studentship. The SANS experiments were performed at the ISIS facility with the assistance of Steve King. Many thanks are also due to the following people at DSM Research: Monique Walet for performing the TEM work, Lizette van der Vondervoort and Luc Leemans for preparing the copolymers, Wiel Leunissen and his colleagues for their help with the mechanical testing, Margot van Wunnik, Pierre Elemans, and Markus Bulters for their help with sample preparation, and Rolf Scherrenberg for useful discussions.
1. W. Barentsen and D. Heikens, Polymer, 14, 579 (1973).
2. S. D. Sjoerdsma, M. E. J. Dekkers, and D. Heikens, J. Mater. Sci, 17, 2605 (1982).
3. R. Fayt, R. Jerome, and Ph. Teyssie, J. Polym. Sci.: Par tB: Poly. Phys., 27, 775 (1989).
4. J. W. Teh and A. Rudin, Polym. Eng. Sci., 32, 1678 (1992).
5. T. Li, V. A. Topolkaraev, A. Hiltner, E. Baer, X. Z. Ji, and R. P. Quirk, J. Polym. Sci.: Part B: Polym. Phys., 33, 667 (1995).
6. P. Cigana, B. D. Favis, and R. Jerome, J. Polym. Sci.: Part B: Polym. Phys., 34, 1691 (1996).
7. H. E. Hermes, PhD thesis, University of London (1996).
8. C. Creton, E. J. Kramer, C.-Y. Hui, and H. R. Brown, Macromolecules, 25, 3075 (1992).
9. J. Washiyama, E. J. Kramer, and C.-Y. Hui, Macromolecules, 26, 2928 (1993).
10. J. Washiyama, C. Creton, E. J. Kramer, F. Xiao, and C.Y. Hui, Macromolecules, 26, 6011 (1993).
11. J. Washiyama, E. J. Kramer, C. F. Creton, and C.-Y. Hui, Macromolecules, 27, 2019 (1994).
12. A. A. Collyer, ed., Rubber Toughened Engineering Plastics, Chapman and Hall (1994).
13. H. E. Hermes, D. G. Bucknall, J. S. Higgins, and R. L. Scherrenberg, in print Polymer, 1998.
14. R. A. C. Deblieck and V. B. F. Mathot, J. Mater. Sci. Letters, 7, 1276 (1988).
15. J. S. Trent, J. I. Scheinbeim, and P. R. Couchman, Macromolecules, 16, 589 (1983).
16. J. N. Clark, M. L. Fernandez, P. E. Tomlins, and J. S. Higgins, Polym. Eng. Sci., 32, 49 (1992).
17. J. S. Higgins and H. C. Benoit, Polymers and Neutron scattering, Clarendon Press (1994).
18. A. Adedeji, A.M. Jamieson, and S. D. Hudson, Macromolecules, 28, 5255 (1995).
19. D. R. Paul and J. W. Barlow, in Polymer Blends, D. R. Paul and S. Newman, eds., Academic Press, New York (1978).
20. G. M. Jordhamo, J. A. Manson, and L. H. Sperling, Polym. Eng. Sci., 26, 517 (1986).
21. H. E. Hermes and J. S. Higgins, in preparation.
22. DSM Research, unpublished results.
|Printer friendly Cite/link Email Feedback|
|Author:||Hermes, H.E.; Higgins, J.S.|
|Publication:||Polymer Engineering and Science|
|Date:||May 1, 1998|
|Previous Article:||Effect of processing conditions on shrinkage in injection molding.|
|Next Article:||Conductive elastomeric foams prepared by in situ vapor phase polymerization of pyrrole and copolymerization of pyrrole and n-methylpyrrole.|