Impact fracture toughness of polyethylene/polypropylene multilayers.
In an increasing number of components, two thermoplastics are co-processed to form a multilayer structure. Two component polymers may be co-injected successively into a single cavity, or into a cavity that changes geometry to accommodate the second. Alternatively, as for auto panels, a thermoplastic may be injected over a preformed protective film layer previously loaded into the open cavity. A typical skin layer is < 1 mm thick. Unfortunately, as is well known, a skin layer can induce brittle behavior in an otherwise tough polymer core. In this work we studied the impact fracture behavior of solid specimens and double-skinned two-component multilayers, formed from three polyolefins. The parameters at issue are the impact fracture resistance [G.sub.c], tensile modulus, and impact fracture transition temperatures [T.sub.bt] of the components. The results of interest are [G.sub.c] and [T.sub.bt] of the multilayer structures.
Brittle fracture in layered polymers was first analyzed by Williams (1). He showed that whereas uncoated specimens of a tough polymer underwent yielding and drawing before fracture, the same polymer fractured in a brittle manner before yielding when coated with a brittle layer. Atkins et al. (2) studied the brittle fracture of polymer multilayers using the load P vs. deflection u diagram. The objective was to prevent generalized yielding prior to the fracture of tough polymers, so that their fracture resistance (critical strain energy release rate), [G.sub.c], could be determined using specimens of practical size. Atkins considered two separate bodies, with fracture toughnesses [G.sub.c1] and [G.sub.c2], loaded in parallel through common loading pins. He suggested that they would crack individually and the total load would be the sum of the separate loads, whereas if the two bodies strained together they would show a weighted average of the individual responses. Using Gurney's (3) quasi-static cracking analysis, Atkins described the fracture of the structure as presented in Fig. 1, in which the segmental area O[M.sub.1][N.sub.1] represents [G.sub.c1] ([a.sub.f] - [a.sub.i]) where [a.sub.i] and [a.sub.f] are respectively the initial and propagated crack lengths. Noting that the [DELTA]u for each layer must be the same, Atkins argued that the crack in the less-tough layer is suppressed (until O[M.sub.c][N.sub.c]) and cracking in the tougher layer would occur earlier (at O[M.sub.c][N.sub.c]) than in the case where the two materials were free. By conservation of energy, the total work required will be the same as before for the same increase in crack length, i.e.,
O[M.sub.1][N.sub.1] + O[M.sub.2][N.sub.2] = 2O[M.sub.c]
or from Gurney's work area
[G.sub.cT] = [G.sub.c1] + [G.sub.c2], (1)
where [G.sub.cT] is the apparent toughness of two equal layers.
Equation 1 is simply a law of mixtures. For simplicity the cracked stiffnesses were chosen to be the same, but the method is expected to work even if the brittle layer does not have the same cracked stiffness as the other layer. Whether or not the two materials have the same modulus or thickness, Atkins expected the toughness of the structure to follow the law of mixtures (4).
It is well known that the total work done during an impact test, as measured using a non-instrumented pendulum machine, includes both plane strain and plane stress components of fracture energy. Ward et al. (5) used Atkins' analysis (2) to reduce the plane stress component in a thin layer of a tough polymer. A pendulum impact machine was used to measure the impact fracture resistance of bend specimens in which this layer was sandwiched between two layers of a more brittle polyethylene. They argued, from Eq 1, that for a multilayer sandwich structure:
[G.sub.cT] = [[B.sub.1]/B] [G.sub.c1] + (1 - [[B.sub.1]/B])[G.sub.c2] (2)
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
where [B.sub.1] and B are the thicknesses of the middle layer and of the entire multilayer (sandwich) respectively (Fig. 2). A plot of [G.sub.cT] against [B.sub.1]/B, from tests on a range of multilayer plates of different proportions, was extrapolated to [B.sub.1]/B = 0 and to [B.sub.1]/B = 1 to determine [G.sub.c1] and [G.sub.c2] (Fig. 3). The same authors used a similar approach to analyze fracture in polymers where a plane strain shear lip layer could be distinguished on the fracture surface (5-8).
In the present work, a similar experimental approach has been used to estimate the effect of surface layers on a relatively thick core. The materials used are three structural polyolefins with a range of tensile modulus and fracture resistance.
Homogeneous plates and multilayer sandwich plates 200 mm square were prepared, by compression molding, from the two PE polymers ('PE-H', 'PE-M') and the ethylene/propylene copolymer ('PPco') listed in Table 1. Granules were poured into a picture-frame mold preheated to 180[degrees]C (PE-H), 165[degrees]C (PE-M) or 200[degrees]C (PPco) in a hot press, and left there for about 30 min to allow melting. Maximum pressure (20 tonnes) was then applied for another 30 min. Normally, about 20% of the material flashed, and a fairly flat plate was obtained with roughly the same thickness as the frame. The initial melting period, and the extra material added in order to flash during compression, were found to be the key parameters in eliminating voids.
[FIGURE 3 OMITTED]
Multilayer plates were prepared using similar procedures. First, the thin skin sheets were made at 200[degrees]C. In a separate operation, these sheets, together with the base material, were stacked in a mold on the hot press (at 150[degrees]C) and the same procedure described above was followed--i.e., 30 min to melt the three parts of the multilayer plate, followed by compression. About 10% of the base material was lost during compression. The skin material also flashed, losing thickness especially at the edges.
INSTRUMENTED IMPACT TEST
International Standard ISO/FDIS 17281 defines the preparation of a Charpy-like, sharp-notched bend specimen and a procedure for interpreting instrumented test records as [G.sub.c] results. [G.sub.c] is calculated, assuming linear elasticity (9), from the area [U.sub.p] under the load vs. displacement curve up to peak load:
[G.sub.c] = [U.sub.p]/[BW[PHI]([a.sub.i]/W)] (3)
where [PHI] is a dimensionless geometry factor that depends on the initial notch length [a.sub.i] and the extent W of the specimen in its direction (10). Nonlinear, viscoelastic, craze-forming materials such as PE present a severe challenge to the Linear Elastic Fracture Mechanics (LEFM) principles on which this method is based (10). ISO/FDIS 17281 sets down strict criteria for the validity (i.e., the geometry-independence) of its results, and these are often difficult to achieve without using very large specimens. Nevertheless, this method, particularly when supplemented by direct observations of the fracture surface, helps to quantify material properties associated with resistance to impact fracture and rapid crack propagation (RCP) even when its results are not strictly valid. Two such properties are the ability to grow a durable craze-tip craze, and the ability to develop wide, ductile shear lips. Since these properties rather than [G.sub.c] in itself are the focus of our interests, we have adopted a generally liberal and inclusive interpretation of load/displacement diagrams. More details are given in the next section.
Multilayer impact bend specimens of 75 X 10 X 10 mm size were cut from the compression molded plates and notched to a depth of 1 mm using a fly cutter. The notches were sharpened by razor blade. Each specimen was subjected to an instrumented impact test at 1 m/s using a high-speed Instron machine (Fig. 4). The three-point bend rig was mounted in an environment chamber cooled by liquid nitrogen. Test temperatures were chosen in the range -40[degrees]C to +40[degrees]C, and were controlled to [+ or -]1[degrees]C using a thermocouple embedded in a sample of the same size as the test specimens.
Load vs. time and displacement vs. time traces were recorded using a PicoScope 200 oscilloscope card directly interfaced to a computer. They were converted to load vs. displacement records using a Matlab program, which also found the peak load point, integrated the curve up to it to compute [U.sub.p], and integrated the entire curve to compute [U.sub.T]. From Eq 3, values of impact fracture resistance [G.sub.c] were computed using specimen geometry parameters.
Clutton and Channell (11) have identified the sequence of events during such a test. After impact, the load increases at a rate corresponding to the notched compliance of the specimen with a superimposed oscillation also determined by its mass and contact stiffness. During this period a craze grows from the notch tip. Then, if the impact speed is high enough and/or the temperature low enough, there is a burst of rapid crack propagation, which unloads the specimen more rapidly than the striker can load it. This is recorded on the load trace as a clearly defined peak value, followed by a sharp drop, and is recorded on the fracture surface as a smooth, brittle zone in the center of the specimen. There may also be regions of pronounced ductile deformation or shear yielding near the edges (shear lips, which indicate ductile drawing from the surface under plane stress, followed by a 45[degrees] shear separation) and at the rear edge of the specimen (a ductile hinge).
[FIGURE 4 OMITTED]
The craze size represents the ability of the material to retain structural integrity on impact: it decreases with increasing impact velocity, and increases with temperature. The energy-absorbing characteristics of crazes also depend upon polymer molecular weight, [M.sub.w] (12). Shear lips, on the other hand, provide an effective mechanism of crack-arrest. Their size is known to increase with [M.sub.w], and to decrease with decreasing temperature.
A brittle-to-tough transition in tough polyethylenes can be distinguished in impact fracture tests, when the strain rate (13, 14), temperature (15-18) or thickness (19) changes. As the impact speed decreases, for instance, to rates of about 0.1 m/s at room temperature, RCP no longer can be sustained, and the specimen instead fails by the extension of a craze across the fracture plane, and its subsequent rupture. At moderate impact speeds, increasing temperature--or decreasing thickness--enhances the ability of shear lips to arrest RCP, and the specimen fails predominantly in plane stress.
The focus here is on the transition temperature under sharp-notched impact bend loading. In these polymers, there is a temperature below which a crack jumps all the way through to the free surface, i.e., the material is no longer capable of arresting RCP within this distance. Using instrumented tests, Morgan (20) identified this point as that at which the post-peak load 'tail' seen in Fig. 4 appears as temperature increases. Whatever the dependence of peak energy [U.sub.p] (and thus, through Eq 3, [G.sub.c]) on temperature, the total energy [U.sub.t] begins to diverge upwards from it. As a result, RCP arrest in this type of test is well indexed by the quantity
U* = [([U.sub.t] - [U.sub.p])]/[U.sub.p].
For each material U* was plotted against temperature and the transition temperature [T.sub.bt] determined systematically as shown in Fig. 5. Below [T.sub.bt], U* is constant and essentially zero, while its increase above [T.sub.bt] can be fitted with a regression line. A third group of points often appears, as in Fig. 5, as an upper shelf. Data points are distributed between these three sets on a minimum-deviation criterion and the value of [T.sub.bt] is inferred from the intersection of the regression line with the lower shelf. This algorithm permits the possibility, as shown, that a point is shared between sets. All the results were double checked by direct observation of the fracture surface.
[G.sub.c] results from single-material specimens of PE-M, PE-H and PPco are shown in Figs. 6 to 8 respectively. As pointed out earlier, the load/displacement nonlinearity evident in Fig. 4 means that these results should be regarded as having a characterizing status rather than providing a basis for LEFM calculations. The fitting lines used to represent [G.sub.c] data points will appear on subsequent graphs without them. Also shown on these graphs are the transition temperatures determined using U* as in Fig. 5.
[FIGURE 5 OMITTED]
It can seen in Fig. 6 that the impact fracture resistance [G.sub.c] of PE-H increases slightly with temperature, but remains essentially constant between -20[degrees]C and room temperature. The transition temperature is -11[degrees]C, emphasizing that the RCP arrest transition does not necessarily correspond to a step change in [G.sub.c]; indeed, no such step-change is seen at any temperature within the range tested. The less tough PE-M material (Fig. 7) has a transition temperature of 6[degrees]C and its [G.sub.c] is more temperature dependent. Figure 8 shows results for PPco, in which the transition at 11[degrees]C does correspond to a step change in [G.sub.c]. Above 10[degrees]C this material is tougher than PE-H and there is no sign of RCP on the fracture surface--which shows considerable stress whitening--or in the load traces. Below 0[degrees]C, however, PPco suffers a dramatic loss in fracture resistance from about 12.5 kJ/[m.sup.2] to about 2.3 kJ/[m.sup.2]. The fracture surface shows some crazing, followed by RCP right across the specimen thickness; there is no apparent plane stress yielding at the edges at any temperature up to 40[degrees]C (Fig. 9).
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
It should be mentioned that for PE materials, the high [G.sub.c] values reported above room temperature might include a contribution from plane stress yielding at the flanks of the crack front prior to fracture. As expected, PE-H clearly performed better than PE-M: not only is the transition temperature lower, but also the toughness and the ability to draw shear lips are undoubtedly superior. On the other hand, PPco, unlike the polyethylenes, does not suffer RCP at room temperature. Although this material does not develop shear lips, it does seem to form a craze which resists collapse. Comparing PE-M (Fig. 7) and PPco (Fig. 8), it would be difficult to define which material had the better overall impact fracture performance at a service temperature range appropriate to, for example, automotive interior components. Both have transition temperatures above 0[degrees]C, although they behave differently in terms of RCP. The dependence of toughness on temperature in PE-M and PE-H is similar: the toughness remains fairly constant. PPco, on the other hand, appears to be much tougher at temperatures above 20[degrees]C since it does not sustain RCP at all, but on decreasing temperature there is a sudden decrease in [G.sub.c] at the transition temperature. The material then becomes very brittle, sustaining plane strain RCP failure throughout the fracture surface (Fig. 9), and the initiation fracture resistance decreases considerably.
[FIGURE 9 OMITTED]
Impact specimens were prepared from 10-mm-thick sheets incorporating 1-mm skin layers, made using the method described above. The skin and the core were both edge notched as shown in Fig. 2.
PE-H as Core Material
Figure 10 shows that PPco skin layers significantly modified the fracture toughness of PE-H. A first impression confirms Atkins' view (2) that the toughness of the multilayer is that of a mixture. However the step change in [G.sub.c] of homogenous PPco appears to be represented out of proportion to the 20% presence of this material in the multilayer thickness. To confirm this, a composite [G.sub.c] has been calculated using the law of mixtures of Eq 2 and is overplotted on Fig. 10. Comparing the law-of-mixtures prediction and the multilayer results confirms that the polypropylene skin has a disproportionate effect. The skin appears to reduce the effective fracture resistance of the core to a level not seen at any temperature in tests on bulk material, since this core material did not show a conventional "toughness transition."
[FIGURE 10 OMITTED]
The dominance of skin properties is even more clear in the transition temperature [T.sub.bt]. The multilayered specimens show a [T.sub.bt] of about 10[degrees]C, whereas that of PPco is 11[degrees]C and that of the PE-H core -11[degrees]C. We will argue elsewhere that when decreasing temperature increases the modulus of the surface skin to a value significantly higher than that of the core, it constrains lateral contraction of the material around the crack front, holding it in a state of plane strain. In this sense the skin reproduces the effect of testing greater (or even infinite) thickness. Rather than merely reducing the toughening influence of plane stress regions near the free surfaces, a rigid skin can eliminate them without necessarily taking over the lost fracture resistance.
If this is the case, the core material may partially recover plane stress yield conditions only if it separates from the skin. At temperatures above 10[degrees]C, delamination does often occur near the back surface of the specimen (Fig. 11), and this is usually sufficient at this temperature range in polyethylene to arrest RCP and allow plastic hinge formation.
As noted above, PE-M is considerably less tough than PPco above 0[degrees]C, although PPco is stiffer. When a PE-M skin was bonded onto a PE-H core, it was not surprising to see that the fracture resistance of the multilayer was more or less the same as that of PE-H (Fig. 12), since both materials have similar fracture resistance. However, the -11[degrees]C transition temperature of the PE-H increased to 8[degrees]C, which is actually slightly greater than that of the PE-M skin.
[FIGURE 11 OMITTED]
[FIGURE 12 OMITTED]
These materials bond usually strongly and delamination is rarely seen. Only in the rare cases in which the skin did delaminate could the stress whitening that characterizes plane stress deformation be discerned. Meanwhile PE-M's own ability to draw shear lips is also relatively low so that RCP is easily sustained and the multilayer is capable of stopping RCP only above 6[degrees]C, where PE-M is capable of forming shear lips. If the skin were thicker, the shear lip might be able to develop enough volume to arrest the crack, and to bring the transition temperature of the PE-H/PE-M skinned multilayer closer to that of PE-M.
PE-M as Core Material
Figure 13 shows that a PPco skin is less detrimental to a PE-M core material than to PE-H, probably because the transition temperatures are similar and the low temperature toughness is not substantially different. Nevertheless, the transition temperature of this multilayer has increased from that of its PE-M core (6[degrees]C) to that of its thin PPco skins (11[degrees]C). Again, PPco present as only 20% of the multilayer thickness has increased the transition temperature, although not as significantly as for PE-H. It could be claimed that for temperatures around 0[degrees]C and above, the toughness of the multilayered wall based on PE-M was actually improved by the PPco skin. At low temperatures there is a drop in toughness, but overall the brittle PPco skin attached to the PE-M does not seem to have a very detrimental effect.
[FIGURE 13 OMITTED]
When a PE-H skin rather than a PPco skin is applied to PE-M, the stiffness is more closely matched to that of the core. Figure 14 shows the PE-H skin has barely influenced the initiation fracture resistance. Once again the skin appears to induce triaxial stress--favoring plane strain fracture--at the edges of the core material, but the PE-H skin itself is able to replace the plane stress layer as a source of crack resistance. Figure 15 shows the fracture surface of a PE-M/PE-H skinned multilayer, on which it is seen that shear lips formed at the free edges of the skin and extend towards the core, while no delamination occurred at the interface.
However, the ability of PE-H to draw shear lips contributes positively to the PE-M core. The transition temperature falls from 6[degrees]C to 0[degrees]C. It could be said that if the transition temperature of the skin material is considerably lower than the core material, [T.sub.bt] of the multilayer is likely to be lower than the [T.sub.bt] of the core on its own. However, since the transition temperature of each polymer appears to be determined by an interaction of plane stress and plane strain properties, it is better to say that replacing a PE-M surface layer by well-bonded PE-H enhances its RCP-arrest capacity.
[FIGURE 14 OMITTED]
[FIGURE 15 OMITTED]
PPco as Core Material
The importance of the shear lips to the transition temperature explains the high transition temperature in multilayers whose free edges do not develop shear lips, such as PPco-skinned PE-H. The role of the shear lips in determining the transition temperature was also clear when a PE-H skin was attached to PPco (Fig. 16). The thin layer of PE-H attached to a PPco core served to provide the shear lip layer needed to decelerate and arrest RCP, resulting in a [T.sub.bt] below -20[degrees]C.
Impact fracture resistance and transition temperature results are summarized in Table 2. The law of mixtures implies that the material that contributes most to the total volume has the greater influence. However, for the fracture resistance of these polymer multilayers, this does not always seem to be the case. This is particularly clear at low temperatures, where there is a large difference between the fracture resistance of PE-H and that of PE-H/PPco skinned multilayer. At low temperatures, the term [G.sub.c2] [1 - [B.sub.1]/B] makes only a small contribution in Eq 2 because the fracture toughness of PPco is low and it is multiplied by its small weight in the total multilayer. The reduction in multilayer fracture resistance [G.sub.t] is due mainly to that of the core, which is factored by [B.sub.1]/B. It seems that the fracture resistance of PPco was slightly increased when a PE-H skin was attached (Fig. 16), but there is still a dramatic loss in toughness below 0[degrees]C. The law of mixtures predicts a dominant effect of the tough polyethylene skin at low temperatures (e.g., -30[degrees]C)--which does not occur. For this material system our results are not consistent with those of Atkins (2, 4) and Ward (5).
Two possible reasons can be suggested. First, the crack initiation process itself is not uniform across the surface and internal regions of the crack front. In these polymers the plane strain internal region develops a craze, which, in thickening rapidly at its mouth during impact loading, is believed to self-destruct by means of a local adiabatic heating process (21). Surface layers, of a thickness that depends on the material, cannot develop the triaxial stress necessary for craze initiation. Their effect on toughness may originate merely from their role in bearing stress (up to the magnitude of the constrained yield stress), which would otherwise, if applied to the internal craze regions, thicken them more rapidly.
[FIGURE 16 OMITTED]
Second, the two materials may interact in some way to an extent that depends, nonlinearly, on their relative proportions. Any such interaction automatically invalidates a law of mixtures. We believe that the mechanism by which a skin polymer interacts with a polymer like PE, which forms tough plane-strain surface layers, is triaxial constraint. This constraint, as we will argue analytically elsewhere, depends nonlinearly on the skin thickness.
Figure 17 shows modulus data for PE-H and PPco, measured using a DMTA method at 10 Hz. The modulus of both materials decreases substantially with increasing temperature, but that of the PPco decreases more rapidly. There is a point of intersection at around 10[degrees]C, where the moduli of both materials are the same. This correlates well with the temperature at which (Fig. 10) PPco surface layers forming only 20% of the multilayer induced loss of toughness from PE-H. It should be noted that similar changes in toughness are observed in PPco/PE blends (22). However, the toughness of the multilayered PE-H and the modulus do not follow the law of mixtures normally seen in the mechanical properties of blends (22, 23).
[FIGURE 17 OMITTED]
The modulus of PE-H decreases from nearly 2 to less than 1 GPa in a temperature range of about 100[degrees]C, while that of the PPco copolymer undergoes a similar change in less than 10[degrees]C. In the same temperature region, this dramatic decrease in modulus is accompanied by the glass transition relaxation peak in the tan [delta] curve. The decrease in fracture toughness at low temperatures is sometimes attributed to a decrease in molecular mobility of network chains in the region of glass transition temperature. However, it has more recently been argued that plane strain impact fracture in thermoplastics results from adiabatically activated decohesion (21). This model predicts an impact fracture resistance [G.sub.c] that depends principally on bulk properties and varies inversely with the 1/3 power of modulus; the present results are in good qualitative agreement.
Impact fracture tests to ISO 17281 have been used to determine both the fracture resistance [G.sub.c] and the "crack arrest transition" temperature [T.sub.bt] of various homogeneous and multilayer polyolefin plates. The transition temperature of PE-H increased when PPco skin was attached, and this system did not resist impact fracture in the way that would be expected from a simple law of mixtures. The results suggest instead that the stiffness of the skin has an important role in determining the fracture resistance and transition temperature of the multilayer. If the skin of a multilayer has a considerably higher modulus than the core material, the fracture toughness of the structure would be lower than that of the core on its own. This might be expected for any combination of skin/core materials unless a rigid skin material is tough enough to replace the toughness it has removed from the core.
Table 1. Experimental Materials. Code Comments PE-H High-density PE, melt index 0.4 (190[degrees]C, 5 kg), density 958 kg [m.sup.-3] PE-M Medium-density PE, melt index 0.9 (190[degrees]C, 5 kg), density 943 kg [m.sup.-3] PPco PP/PE copolymer, melt flow rate 0.3 (230[degrees]C, 2.16 kg), density 905 kg [m.sup.-3] Table 2. Fracture Resistance and Transition Temperature of Multilayer Specimens. Impact Fracture Resistance Transition Materials (kJ [m.sup.-2]) Temperature, Core Skin Low Temperature High Temperature [degrees]C PE-H 9.5 at -40[degrees]C 11.0 at +40[degrees]C -11 PE-M 4.5 at -40[degrees]C 8.0 at +40[degrees]C 6 PPco 2.0 at -40[degrees]C 13.0 at +40[degrees]C 11 PE-H PPco 3.1 at -40[degrees]C 10.5 at +40[degrees]C 10 PE-H PE-M 9.5 at -30[degrees]C 12.5 at +30[degrees]C 8 PE-M PPco 4.0 at -20[degrees]C 10.0 at +20[degrees]C 11 PE-M PE-H 5.0 at -30[degrees]C 8.5 at +30[degrees]C 0 PPco PE-H 3.0 at -30[degrees]C 13.0 at +30[degrees]C -22
1. J. G. Williams, private communication (1970).
2. A. G. Atkins and Y-W. Mai, Int. J. Fracture, 12, 923 (1974).
3. C. Gurney and J. Hunt, Proc. Roy. Soc. London, A299, 508 (1967).
4. A. G. Atkins and Y-W. Mai, Elastic and Plastic Fracture, Ellis Horwood, Chichester, U.K. (1988).
5. R. W. Truss, R. A. Duckett, and I. W. Ward, Polym. Eng. Sci., 23, 708 (1983).
6. G. L. Pitman and I. M. Ward, Polymer, 20, 895 (1979).
7. R. A. W. Fraser and I. M. Ward, Polymer, 19, 220 (1978).
8. P. J. Hine, R. A. Duckett, and I. M. Ward, Polymer, 22, 1745 (1981).
9. E. Plati and J. G. Williams, Polym. Eng. Sci., 15, 470 (1975).
10. J. G. Williams, Fracture Mechanics of Polymers, Ellis Horwood, Chichester, U.K. (1984).
11. E. Q. Clutton and A. D. Channell, Impact and Dynamic Fracture of Polymers and Composites, ESIS Publication 19, p. 215, MEP, Bury St. Edmunds, U.K. (1995).
12. E. Q. Clutton, V. Rose, and G. Cappaccio, J. Macromol. Sci. Phys., 629 (1996).
13. B. Z. Jang, O. R. Uhmann, and J. B. VanderSande, J. Appl. Polym. Sci., 29, 409 (1984).
14. P. I. Vincent, Plastics, 27, 115 (1962).
15. I. G. Zewi, W. J. Rudik, and R. D. Corneliussen, Polym. Eng. Sci., 20, 622 (1980).
16. P. I. Vincent, Impact Test and Service Performance of Thermoplastics, The Plastic Institute, London (1971).
17. E. Plati and J. G. Williams, Polymer, 16, 915 (1975).
18. M. Parvin and J. G. Williams, J. Mat. Sci., 10, 1883 (1975).
19. M. K. V. Chan and J. G. Williams, Polym. Eng. Sci., 21, 1019 (1981).
20. R. E. Morgan, Ductile-Brittle Transitions in Pipe Grade Polyethylene, PhD Thesis, University of London (1994).
21. P. S. Leevers and R. E. Morgan, Engng. Fracture Mech., 52, 999 (1995).
22. U. Niebergall, J. Bohse, H. Sturm, S. Seidler, and W. Grellmann. Appl. Macromol. Chem. Phys., 253, 27 (1997).
23. M. J. Folkes and P. S. Hope, Polymer Blends and Alloys. Blackie-Academic Professional (1993).
Samuel Lunenfeld Research Institute
Mount Sinai Hospital
600 University Avenue
Toronto M5G 1X5, Canada
Imperial College London
Department of Mechanical Engineering
London SW7 2AZ, U.K.
*To whom correspondence should be addressed.
|Printer friendly Cite/link Email Feedback|
|Author:||Moreno, Luisa; Leevers, Patrick|
|Publication:||Polymer Engineering and Science|
|Date:||Sep 1, 2004|
|Previous Article:||Surface texture characterization of injection-molded pigmented plastics.|
|Next Article:||Removal of ungrafted monomer from polypropylene-graft-maleic anhydride via supercritical carbon dioxide extraction.|