In situ formation and processing of ultra high molecular weight polyethylene blends into precursors for high strength and stiffness fiber.
Polyethylene is a very useful thermoplastic because of its high calculated Young's modulus, abrasion, and wear resistance (1-3). Theoretical estimates of the Young's modulus for a single chain polyethylene crystal lie in the range 182.3 GPa (4) to 340 GPa (5). The highest modulus value achieved during the course of experimental work is 230-240 GPa (6). To achieve such high modulus values and tensile strengths, polyethylenes with high molecular weights are required (7). Nascent ultra high molecular weight polyethylene (UHMWPE) fulfills this criterion with weight average molecular weights in excess of 2 x [10.sup.6] kg/kmol. It is used as-polymerized with no thermal treatment above the melt transition temperature and is believed to have either a chain extended morphology (8) or the ability to undergo fast morphological reorganization when being heated (9). Unfortunately, UHMWPE in the melt is very difficult to process because of its high melt viscosity. So, alternative processing routes have been examined and assessed in terms of product quality and ability to produce industrial size quantities.
Gel spinning, developed by Smith and Lemstra (10-12) and the only commercial process, swell drawing (13) and die-free gel spinning (14) involve the swelling or dissolution of polyethylene in solvent at elevated temperatures to reduce the entanglement density, followed by crystallization or gelation and then orientation of the lower entanglement material. Examples of the tensile properties of films and fibers produced from UHMWPE materials using these processing techniques are shown in Table 1. The main disadvantage of these techniques is the use of harmful organic solvents, e.g. decalin and p-xylene, of the order of 10 kg solvent/kg polymer, which require removal and recycling. A number of solid state processing techniques have been investigated using nascent UHMWPE, e.g compression molding for films (15), coextrusion of crystal mats (6) and films, ram extrusion (16) and roll-drawing (17). Again, Young's modulus and tensile strength values for the products obtained from these processes are shown in Table 1. Generally, the Young's modulus is [approximately]100 GPa and the tensile strength appears to be process dependent, with the highest values for those processes which involve direct drawing of as-polymerized UHMWPE; the obvious exception to this being the drawing of crystal mats (6). In all of these processes the temperature does not exceed the melting temperature of UHMWPE. There are some clear disadvantages: low throughput rates for direct drawing of as-polymerized films (8, 18) and extensive preparatory work for coextrusion of films and crystal mats (6, 19).
[TABULAR DATA FOR TABLE 1 OMITTED]
Those processes involving compaction or compression of the UHMWPE powder particles are aiming to mechanically fuse the particles together. Zachariades (20) and Zachariades and Kanamoto (21) state that temperatures above 220 [degrees] C must be used during compression molding to produce coherent samples with reasonable tensile test stress-strain curves. However, above 220 [degrees] C the UHMWPE has had sufficient time to become a full, homogeneous melt (22), which retains no memory of its previous morphology. Also, the compressed nascent powder still shows a higher Young's modulus than the recrystallized material, whether or not the latter was heated above 220 [degrees] C (21). However, the compressed nascent powder samples are slightly more brittle, probably due to cohesive failure.
Truss et al. (23) found that compression at temperatures below the melting point of UHMWPE produced samples with higher fracture strengths relative to samples produced at above [T.sub.m] temperatures. They also found that commercial powders with connecting fibers, a sign of melting and recrystallization during polymerization, were stronger and more coherent after compaction than the commercial powders with nodular initial morphologies, e.g. Hifax 1900. This was attributed to the beneficial effects of a smaller effective particle diameter, the penetration of the fiber network into the particles during compaction, and the continuity provided by the fibers. In actual fact, the signs of melting and recrystallization indicate that a least a portion of the UHMWPE is no longer nascent (24). This portion probably has an increased entanglement density, thus removing one of the advantages of low entanglement density nascent UHMWPE. Rotzinger et al. (15) demonstrated that nascent UHMWPE powder can be compression molded below its [T.sub.m], at 80-120 [degrees] C, to produce samples with Young's modulus values of up to 100 GPa for EDR = 40. The actual values are dependent on the polymerization temperature, catalyst activity, and monomer pressure. These workers carried out their own controlled polymerizations, whereas the polymerization conditions of commercial samples are not always known. The compaction pressure used (15) was 49 MPa, significantly higher than in Refs. 20, 23, and 25 and lower than in Refs. 16 and 21. However, it is in full agreement with the compaction studies of Gao et al. (24), which showed an optimum compaction pressure at [approximately]30 MPa with Young's modulus = 5.3 GPa. They also state that initially the Young's modulus increases with compaction pressure because of void reduction, not inter-particle cohesion. Too high a compaction pressure reduces the volume available for chain movement and hinders inter-particle chain diffusion.
When compaction is followed by extrusion, to orient the material, the samples produced are less brittle than when using compaction alone. The Young's modulus values are [approximately]15 GPa (16, 25) and inter-particle adhesion is provided as the explanation. The relative contributions of compaction and extrusion to this adhesion will depend on the initial compaction pressure prior to flow, according to the remarks of Gao et al. (24). Compacted and extruded UHMWPE powder is tougher than compacted alone because of increased contact surface area brought about by extrusion (26). The scanning electron microscopy photographs of compacted alone UHMWPE shown by Gao et al. (24) are similar to the one presented in Zachariades et al. (16). Micrographs of compacted and extruded UHMWPE powder (26) are, again, similar to those in Zachariades et al. (16). The micrographs demonstrate the closer packing and deformation of the particles and the presence of adhesion. Compaction and extrusion are not sufficient to bring about the amount of inter-particle penetration necessary to prevent or delay the onset of cohesive failure. Therefore, an additional technique is required. The most obvious is the use of a solvent to provide a low entanglement state of the UHMWPE, which is related to the high drawability in gel spinning (27). However, in solid-state extrusion, the authors believe that the continuity of structure at the macro-level is as important as, if not more so than, the low entanglement density.
In this paper we report the results of compaction and solid state ram extrusion through a "swelling" die of a 10 wt% mineral oil/UHMWPE blend below the melting point of the UHMWPE, i.e. cold extrusion. Precursors manufactured ill this manner may be hot drawn later. This method retains the ductility of the nascent UHMWPE by compaction below [T.sub.m] (24), and allows the UHMWPE to dissolve partially in the mineral oil (28), reducing the entanglement density prior to orientation and elongation during extrusion. Thus, the mineral oil has been shown to improve the processability characteristics of UHMWPE during cold extrusion and give Young's modulus and tensile strength values of [approximately]12 GPa and 0.1 GPa, respectively, at extrusion draw ratios of [less than] 10 (29, 30). The use of a "swelling" die aims to encourage the swelling of the polymer by mineral oil take-up; in contrast to PTFE paste extrusion, where the PTFE should not be soluble ill the liquid hydrocarbon extrusion aid (31). Hence, increasing chain mobility at the particle interface and producing better cohesion between particles. The novelty of the process described in this paper lies ill the combination of compaction and the simultaneous solid-state extrusion and swelling of UHMWPE in one processing step by the use of a specially designed die.
Materials and Sample Preparation
The UHMWPE, Hifax 1900, used in this study was supplied by Montell Polyolefins. It has a weight average molecular weight of 6 x [10.sup.6] kg/kmol and an average particle size of [approximately]200 [[micro]meter]. A light mineral oil, with a boiling point of 350 [degrees] C and a specific gravity of 0.838, was purchased from Fisher.
The 10 wt% mineral oil/UHMWPE blend was prepared from the as-received UHMWPE and mineral oil at ambient temperature using a 750 Watt Moulinex mechanical food mixer operated with a pulse action, 15 sec mixing followed by 45 sec relaxation, to minimize shear heating. This was continued until the mixture was macroscopically homogeneous. The blend material, [approximately]15 g, as fed into the 12 mm diameter, 200 mm length barrel of a capillary rheometer (Goettfert Raheograph 2003) preheated to 135 [degrees] C. The barrel was blocked off with a solid cylinder die block for compaction. Feeding of the material was carried out slowly with intermittent low-pressure compaction to remove air pockets. The packed barrel was left to equilibrate for 2-3 min. The blend material was then compacted at 135 [degrees] C for 15 min under a piston pressure of [approximately]9 MPa. After this time the solid die block was removed and replaced with a round hole capillary die of known dimensions in preparation for extrusion. Extrusion was carried out at constant shear rates ill the range 0.27 [s.sup.-1] to 10.8 [s.sup.-1].
For most of the experiments reported in this paper a "swelling" die was used. This differs from the usual round hole dies used for extrusion ill that the initial small diameter die is of short length, 5 mm, with an entrance angle of 90 [degrees], before a sudden expansion to a die diameter of between 2 mm to 2.5 mm larger than the initial diameter. The overall length of the die is 30 mm. This particular die design allows constrained swelling and annealing to take place simultaneously with extrusion. The extrudate temperature is maintained and with the increase in pressure due to the sudden expansion there should be orientation relaxation and increased contact area between particles. This will lead to improved inter-particle penetration.
All of the following characterization techniques were applied to samples of pure UHMWPE and blend materials. The mineral oil was not extracted from the latter.
Melting endotherms were recorded using a TA Instruments 2910 TA differential scanning calorimeter (dsc) with heating and cooling rates of 10 [degrees] C/min. The sample weight was [approximately]5 mg and indium ([T.sub.m] = 156.61 [degrees] C and [Delta][H.sub.f] = 28.4 J/g) was used for calibration purposes. The dsc experiments were carried out under a nitrogen atmosphere between 30 [degrees] C and 180 [degrees] C. The melting points quoted refer to the peak temperatures in the thermograms and are for first heatings, unless otherwise stated. Sample crystallinities were calculated from heats of fusion taken from the thermograms assuming a purely crystalline polyethylene has a heat of fusion of 289 J/g (32). These crystallinities are nominal as the presence of mineral oil has not been accounted for. Therefore, the blend [Delta][H.sub.f] values quoted are those measured for the mixture of mineral off and UHMWPE. Two samples were taken from the extrudates: skin and core. They were prepared by careful slicing of the extrudates at ambient temperature.
A JEOL 6300F scanning electron microscope (SEM) was used to produce photographs of the morphologies of the blend and pure materials. The compacted and extruded samples were fractured ill liquid nitrogen and sputter-coated with a [approximately]100 [Angstrom] layer of gold to prevent charging. A low accelerating voltage of 5 kV and 150 mA was used to minimize surface heating effects. The micrographs were captured on Polaroid type 52 film. All the SEM photographs shown here are of extrudate cross sections, unless otherwise stated.
Mechanical property measurements, Young's modulus, volumetric energy to fracture and elongation, were performed using a Sintech Universal Testing Machine (UTM) with a maximum capacity of 50 kN at ambient temperatures and with a crosshead speed of either 2.54 mm/min or 5 mm/min. Extrudates of length 170 mm were loaded between the wedge action grips after both ends of the samples were capped by melting to prevent slippage during the course of stretching. An extensometer with a standard gauge length of 25 mm was used to measure the extension during draw. Elongation is calculated as the ratio of final length of sample to initial length of sample, i.e. 25 mm. The volumetric energy to fracture is taken as a measure of the toughness of the material.
RESULTS AND DISCUSSION
Pure UHMWPE Hifax 1900 Powder
To highlight the improvements brought about by processing nascent UHMWPE blended with mineral off and through a "swelling" die, it is necessary to present some results from the characterization and processing of the pure UHMWPE Hifax 1900 reactor powder. Figures 1a and b show SEM photographs of the as-received Hifax 1900 reactor powder at two levels of magnification. The reactor powder is made up of major particles with an average diameter of [approximately]200 [[micro]meter]. These major particles are in turn made up of minor particles with diameters of the order of 10 [[micro]meter]. The minor particles are aggregates of yet smaller, sub-micron particles. It is important to note the lack of connecting fibrillar bridges between these discrete particles, indicating very little, if any, partial melting during synthesis. When these materials are processed under different conditions the morphology changes, as shown in Figs. 2 a-c. After compaction at 13.7 MPa and 127 [degrees] C for 30 min, the major and minor particles have flowed into each other and interfacial cohesion is evident; this is discussed further by Gao et al. (24), who also point out that the tensile failure appears to be due to cohesive failure. The sub-micron particles constituting the minor particles are still visible, and it is assumed that, by compaction, interfacial cohesion has been initiated between these particles also. When compaction of the pure UHMWPE at [approximately]9 MPa and 135 [degrees] C for 15 min is followed by extrusion at [Mathematical Expression Omitted] and through a normal die of 8 mm diameter ([EDR.sub.apparent] = 2.25), interfacial cohesion is improved by increasing the surface contact area during extrusion. Figure 2b shows a typical boundary in the core of the extrudate between the minor particles after cryogenic fracture: the particles' boundaries are still very clear, but there are no features that might be associated with the sub-micron particles. There is also evidence of fibrillar bridges connecting the minor particles as a result of the partial melting taking place during compaction and extrusion. This partial melting has a greater effect on the surface of the extrudate, as shown in Fig. 2c. Here, distortion of sub-micron particles leading to the onset of fracture or failure at the surface during extrusion may be seen. Experiments have shown that these materials cannot be further drawn at elevated temperatures.
The mechanical properties of compacted pure UHMWPE are reasonable. The values for the Young's modulus range from 3.1 GPa at a compaction pressure of 11 MPa to a maximum of 5.3 GPa at a compaction pressure of 32 MPa. The volumetric energy absorbed to fracture increases from 2.5 kJ/[m.sup.3] to 4.85 kJ/[m.sup.3] over the same compaction pressure range and there is a maximum elongation ([L.sub.f]/[L.sub.i]) of 1.0014 (24). These materials are very brittle.
The results of the thermal characterization of the pure UHMWPE structures are given in Fig. 3 and Table 2. The as-received Hifax 1900 reactor powder shows a melt transition temperature of 143.0 [degrees] C and a crystallinity of 72.1% on the first heating and a melt transition temperature of 136.6 [degrees] C and a crystallinity of 46.7% on the second heating. The first heating values are similar to those reported by Ottani and Porter (33) and the decreases in [T.sub.m] and crystallinity are as expected. When compacted and extruded through a normal die of diameter 8 mm ([EDR.sub.apparent] = 2.25) both the skin, at 146.7 [degrees] C, and core, at 149.7 [degrees] C, of the pure UHMWPE extrudate show melt transition temperatures greater than that of the reactor powder. This is due to orientation-induced crystallization, which shifts the melting point upward, and the fact that larger crystals are relatively unaffected by the processing. The core sample has a crystallinity very similar to that of the reactor powder on first heating, whereas the skin sample has a crystallinity [approximately]17% less than that exhibited by the reactor powder on first heating and the extrudate core. The outer surface of the material experiences shear heating at the barrel wall and is at a higher temperature than the material in the core. Partial melting of small, imperfect crystals takes place in the skin region, evidenced by the initial shoulder or peak in the skin sample's thermogram that corresponds to the second heating melting peak of the pure UHMWPE powder. This initial shoulder is the molten material that has recrystallized as lower [T.sub.m] crystals (2). The core of the extrudate experiences orientation, but not the higher temperature at the barrel wall in contact with the skin as the thermal conductivity of polyethylene is low. This nonuniform melting, which gives rise to a skin-and-core structure, explains the large morphological differences observed between the core and skin SEM photographs in Figs. 2b and 2c.
UHMWPE/10 wt% Mineral Oil Blend
The advantages of the addition of 10 wt% mineral oil to UHMWPE reactor powder prior to cold extrusion have been explained previously (29, 30). In summary, the mineral oil acts as a lubricant, heat transfer agent, and solvent for the UHMWPE, By extrusion of the UHMWPE/10 wt% mineral oil blend through swelling dies of varying diameters, it is hoped that the solvating effect of the mineral oil on the UHMWPE will be enhanced by an increase in pressure due to the sudden expansion in die diameter, and that this in turn will lead to better inter-particle cohesion. The processability of the UHMWPE/10 wt% mineral oil blend through swelling dies of 5, 6, 7, 8, and 10 mm at 135 [degrees] C and with [Mathematical Expression Omitted] is shown in Fig. 4 as plots of extrusion pressure versus time. Apart from the peak extrusion pressure of 375 bar for a swelling die of diameter 5 mm, the extrusion pressures are [approximately]50-60 bar less than those reported for processing of this blend through normal (length = 30 mm) dies (29). After the initial peak extrusion pressure, there is a short period of steady extrusion pressure followed by a decrease. The rate of decrease increases with increasing die diameter, as would be expected. As will be shown in a subsequent section, this unsteady state extrusion does not detrimentally affect the mechanical properties of the extrudate products. Experimental values of the actual extrusion draw ratio (EDR) and die swell are given in Table 3. EDR is calculated as the ratio of [(barrel diameter).sup.2] to [(swelling die inlet diameter).sup.2] for an apparent value and as the ratio of [(barrel diameter).sup.2] to [(extrudate diameter).sup.2] for an actual value. The die swell increases to 23.5% at [EDR.sub.apparent] = 2.25 and decreases at [EDR.sub.apparent] = 4. From this point it increases again, with a rapid increase between apparent EDRs 5.76 and 9, but the actual EDR does not change greatly between die diameters 5 mm and 4 mm. This second increase is mainly a function of die geometry and the increase in die [TABULAR DATA FOR TABLE 2 OMITTED] swell at [EDR.sub.apparent] = 2.25 is believed to be caused primarily by favorable solvating and swelling conditions.
Experiments conducted with a D4L5 swelling die (D = diameter, L = length) over an apparent shear rate range of 0.9 [s.sup.-1] to 10.8 [s.sup.-1] show the die swell to be independent of [Mathematical Expression Omitted]. Therefore, average values may be used to represent global characteristics. The average die swell is 47.5% with a standard deviation of 1.5 percentage points. The actual EDR is 4.14, which is low when compared with the [EDR.sub.actual] = 8.32 for a blend extruded through a D4L30 normal die. The increased orientation experienced by the material extruded through a normal die inhibits die swell.
Tensile Property Measurements
Tensile properties are, like die swell, independent of [Mathematical Expression Omitted]. For blend extrudates processed using a swelling die with [EDR.sub.apparent] = 9, the average values and standard deviations are as follows:
Young's modulus 1.12 [+ or -] 0.11 GPa Elongation ([L.sub.f]/[L.sub.i]) 1.58 [+ or -] 0.07 Volumetric energy to fracture 15.16 [+ or -] 2.23 MJ/[m.sup.3]
The tensile tests were conducted at ambient temperatures with a crosshead speed of 2.54 mm/min. An UHMWPE/10 wt% mineral oil blend extrudate processed at 135 [degrees] C, [Mathematical Expression Omitted] with a D4L30 normal die and tested under the same conditions has the following tensile properties (30):
Young's modulus 8.3 GPa Elongation ([L.sub.f]/[L.sub.i]) 1.05 Volumetric energy to fracture 2.09 MJ/[m.sup.3]
Clearly, the blend material processed through the D4L5 swelling die is more ductile and tougher than when processed through a D4L30 normal die. Melt crystallized polyethylenes show Young's modulus values of [approximately]1 GPa, similar to the swelling die extruded blend materials. Therefore, the use of a swelling die allows for almost complete relaxation of oriented chains. The normal die-processed material is stiffer due to the chain orientation during extrusion. The swelling die-processed material is not as oriented, but compensates for this by possessing greatly enhanced ductility, [approximately]50% higher, and toughness, 625% higher. These superior properties are due to improved interfacial cohesion, [TABULAR DATA FOR TABLE 3 OMITTED] brought about by simultaneous swelling and extrusion causing diffusion of UHMWPE chains across particle boundaries.
The variations of tensile properties with swelling die diameter or apparent EDR are shown in Fig. 5. As would be expected, Young's modulus increases with increasing [EDR.sub.apparent]. However, the actual values are low when compared with those in Table 1 and those quoted for pure, compacted UHMWPE, implying that the blend extrudates processed using a swelling die are highly ductile, especially at low [EDR.sub.apparent] values. At higher [EDR.sub.apparent] the material is more oriented, with a higher concentration of aligned polyethylene chains per unit cross-sectional area. This makes the material stiffer. The elongation ([L.sub.f]/[L.sub.i]) decreases with increasing [EDR.sub.apparent], but the elongation values are in the realms of super-drawability, being [greater than] 50% higher than those of the pure, compacted UHMWPE and up to 150% higher than for the compacted and extruded blend, depending on apparent EDR value.
Some examples of engineering stress-strain curves for different swelling die diameters are shown in Fig. 6. At the smaller die diameters and higher [EDR.sub.apparent] values, a yielding point is clear, separating the elastic and plastic deformation zones. Fracture takes place at engineering strains of [approximately]32% for the D5L5 die and 56% for the D6L5 die. The volumetric energies to fracture, i.e. energy absorbed before fracture per unit deformation volume, are 13.1 MJ/[m.sup.3] and 24.6 MJ/[m.sup.3] for the D5L5 and D6L5 swelling dies, respectively. For the D8L5 and D10L5 swelling die extrudates, there are very long plastic deformation zones after the yield point and the D8L5 swelling die extrudate shows strain hardening at engineering strains [greater than] [approximately]120%. The D10L5 swelling die extrudate would probably also show strain hardening if tested to higher engineering strains. These two samples are extremely difficult to fracture due to their high ductility, but initial attempts 'indicate that the D8L5 swelling die ([EDR.sub.apparent] = 2.25) produces the highest volumetric energy to fracture, 31.2 MJ/[m.sup.3]. It should be noted that UHMWPE blended with mineral oil and extruded through either a normal or a swelling die will produce a precursor with volumetric energy to fracture at least 1000 times greater than extruded, pure UHMWPE. This is because the mineral oil dissolves some of the UHMWPE, allowing diffusion across particle boundaries.
The higher the value for volumetric energy to fracture, then the tougher the material. The toughness of a material is improved by increasing the number of physical entanglements present or by increasing the interfacial thickness available for entanglements. At low [EDR.sub.apparent] values the processing conditions allow for better and/or more entanglements due to interparticle diffusion of polyethylene chains whilst they are dissolved in the mineral off, The further the mineral oil penetrates into the particles, of all sizes, then the larger the thickness available for diffusion and inter-particle entanglements. This greater depth of penetration, caused by the constrained swelling, is believed to be the reason for the high drawability.
The results of the thermal characterization of the UHMWPE/10 wt% mineral oil blend extrudates processed at 135 [degrees] C and [Mathematical Expression Omitted] through swelling dies are shown in Figs. 7-9. The variation of dsc thermograms with [EDR.sub.apparent] for skin and core samples are shown in Figs. 7 and 8, respectively. An [EDR.sub.apparent] = 1 is for a sample that was compacted only. The variations of melt transition temperature and crystallinity with [EDR.sub.apparent] are shown in Fig. 9.
The thermograms in Fig. 7 and 8 show that a certain critical EDR exists below which twin melting peaks are apparent. The initial shoulder appearing at a temperature between 134 [degrees] C and 136 [degrees] C corresponds to the [T.sub.m] of the melt crystallized UHMWPE [ILLUSTRATION FOR FIGURE 3 OMITTED]. The critical EDR shown here is about 3. At higher EDRs only one melting peak is observable. These results agree qualitatively with our previous publication on extrudates obtained using normal dies (29). This shoulder is also found in pure, extruded UHMWPE and is attributable to shear heating at the barrel wall. The low [T.sub.m] shoulder is sharper in the swelling die extruded blend material, and this is attributed to the mineral oil acting as a heat transfer agent and some solvation-induced partial melting during the extrusion process. As the [EDR.sub.apparent] increases, the increased chain alignment and shear-induced orientation makes the complete melting of low melting temperature molecules more difficult. Therefore a serf-blend structure is not observed at the higher [EDR.sub.apparent] values. The skin and core thermograms are similar at identical [EDR.sub.apparent] values for the swelling die extruded blend: twin peaks are seen in both samples and the peaks are similar shapes. For the extruded, pure UHMWPE, this is not the case; only the skin sample shows two peaks.
The plot of major melting peak versus apparent EDR shown in Fig. 9 indicates that the melting peaks for both the skin and the core do not show a strong dependence on the apparent EDR. The skin shows an average [T.sub.m] of 146.2 [degrees] C with a standard deviation of [+ or -] 0.18 [degrees] C and the core has an average [T.sub.m] and standard deviation of 147.5 [+ or -] 0.64 [degrees] C. The ranges of the skin and core Tins are 0.46 [degrees] C and 2.24 [degrees] C, respectively. This gives an average melting temperature difference between the skin and core of 1.3 [degrees] C, with the core showing the slightly higher melting temperature. These results compare favorably with those for the pure UHMWPE extrudates, which had a 3 [degrees] C temperature difference between skin and core. It is also an improvement over the blend extrudates from the normal die, where the skin shows a temperature between 2 [degrees] C and 4 [degrees] C lower than that of the core (29). The use of a swelling die along with the mineral oil produces extrudates with more uniform melting temperatures progressing from skin to core. The minimum in melting temperature when plotted against EDR observed by Kanamoto et al. (19) and Chuah et al. (34) at low EDRs for hot drawn, coextruded UHMWPE films is not observed here.
The crystallinity data shown in Fig. 9 show a different dependence on [EDR.sub.apparent] between the skin and the core. The crystallinity of the skin sample decreases steadily from 66.3% to 59.6% as the EDR increases from 1.44 to 2.94. Further increase in apparent EDR results in an increase in crystallinity. There is a tendency for the crystallinity to level off at about 65% at higher apparent EDRs. Owing to the limited data presented here no further comments will be made. The difference between skin and core crystallinities is limited to [approximately]2%, whereas for normal die-extruded, blend material the difference is [approximately]5% (29) and for normal die-extruded, pure UHMWPE it is [approximately]12%. For core samples, the swelling die extrudate crystallinities are lower than those for normal die extrudates. This trend is repeated in the skin samples, except for [EDR.sub.apparent] = 1.44, and is a reflection of the orientation relaxation occurring when the back pressure increases due to the sudden expansion in the swelling die diameter.
The core sample, apparently, shows a very weak dependence on EDR except initially and for the apparent EDR = 9 sample. If the datum for the compacted-only blend sample ([EDR.sub.apparent] = 1) is put aside, the core shows a small increase in crystallinity from about 62.7% to about 65.2% as apparent EDR increases from 1.44 to 2.25. It then stays almost constant for apparent EDRs up to 6. A sudden jump to 73.6% is observed at an apparent EDR of 9. The initial increase at low apparent EDRs is expected because of the realignment of crystals along the flow direction. The dramatic increase in crystallinity for the [EDR.sub.apparent] = 9 extrudate sample is due to two factors: (i) increased chain alignment giving rise to a larger proportion of orientation-induced crystallization; and (ii) the increase in solvation effects from the swelling process. The solvation of molecules in the mineral oil increases chain mobility at particle surfaces. The highly mobile molecules are then able to diffuse more deeply into the particles. Upon cooling these molecules are locked into their new locations and positions by crystallization. The fact that the melting point of the [EDR.sub.apparent] = 9 extrudates is not higher than those of the lower [EDR.sub.apparent] extrudates also appears to support this argument. Further, the large die swell exhibited by the [EDR.sub.apparent] = 9 samples is also indicative of enhanced solvation effects. (All the crystallinities are based on the blend sample weight and so are lower than would be expected for pure UHMWPE. When the amount of mineral oil present in the precursors can be quantified it can be allowed for in the crystallinity calculation.) However, both the melting peak and crystallinity show very weak dependencies on apparent EDR when compared with extrudates from normal dies. This indicates that at low apparent EDRs swelling of powder particles and solvating of the polyethylene chains is more effective with swelling dies than with normal dies, and that these two effects dominate over that of shear-induced orientation.
For comparative purposes, Figs. 10a and b show SEM photographs of an UHMWPE/10 wt% mineral oil blend extrudate processed at 135 [degrees] C and [Mathematical Expression Omitted] through a D8L5 normal die. The major particles have melded together and have been compressed and compacted. There are, however, still clear particle boundaries, which are delineated by dark borders. The dark color is thought to be concentrations of mineral oil. During extrusion, some of the UHMWPE chains are dissolved in this mineral oil and diffusion from one particle to another takes place. Those major particles that underwent fracture are noticeable because of the clustering of dark mineral oil spots. Again, we suggest that these are predominantly at the interfaces or boundaries between minor particles. This enhances the interfacial cohesion between these smaller particles. At higher magnification, Fig. 10b, showing a boundary between the two minor particles, the slightly swollen sub-micron particle structures may be observed. This should be compared with Fig. 2b, which shows melted and recrystallized structures with clearer, straighter boundaries. So, extrusion through a normal die produces some swelling of UHMWPE particles as the mineral oil penetrates in and dissolves some of the polyethylene chains.
When the UHMWPE/10 wt% mineral oil blend is extruded through a swelling die the swelling is greater and the boundaries between major and minor particles are less obvious. Figures 11a-c show cryogenically fractured extrudates processed through a D8L5 swelling die at 135 [degrees] C and [Mathematical Expression Omitted]. Figures 11a and b show core cross sections, The major particle boundaries are fewer, but still delineated by mineral oil filled zones, as in Fig. 10a. At the higher magnification the boundaries between the minor particles are all but lost and the sub-micron particles are swollen to structures with micron-order dimensions. Most of these appear to be in intimate contact, allowing the mineral oil-dissolved polyethylene chains a greater surface area for diffusion. The surface of this extrudate, at high magnification, is shown in Fig. 11c. Here, again, swelling of the sub-micron particles to micron dimensions is observed. They appear to be flattened or smudged into two-dimensionality, unlike the particles in Fig. 11b, which still retain an impression of three-dimensionality, if not always sphericity. This flattening or smudging at the surface is probably a result of partial melting due to shear heating at the barrel wall. However, there is no evidence of fibrillation or fracture of melted structures, as there is in Fig. 2c. Certainly, in the darker regions associated with pockets of mineral oil, there is less fusing together of particles; they are more plainly discrete particles. In the lighter areas particles flow into one another, an effect caused by melting and the mineral off acting as a solvent of the UHMWPE and allowing inter-particle diffusion, the lack of fibrillation and early tensile failure is also due to the heat transfer properties of the mineral off.
The morphology of a D10L5 swelling die extrudate is shown in Figs. 12a and b at different magnifications. At this lower [EDR.sub.apparent] of 1.44 the major particle boundaries have almost completely disappeared leaving a "terraced" topography. The mineral oil is now in small circular regions, rather than as boundary lines. The sub-micron particles are swollen to micron dimensions and there is loss of clear boundaries between these particles. We also suspect that originally [10.sup.-8] and nanometer-sized particles are also swollen by the mineral oil forced into them during extrusion through the swelling die. Figure 12c shows the morphology of the blend extruded through a D4L5 swelling die, i.e. high [EDR.sub.apparent] of 9. This should be compared with Figs. 10a, 11a and 12a, all of which are magnified 250x. Here boundaries are clear because there are groups of striations and these groups show different orientations. It is similar to a schematic showing the variously oriented directors of an anisotropic material. These striated regions have characteristic dimensions, which may associate them with swollen minor particles. This aspect requires further investigation. At this higher apparent EDR the material is more oriented under shear and the regions of striation are probably a reflection of this. Preliminary wide angle X-ray diffraction patterns indicate that orientation has started by [EDR.sub.apparent] = 5.76. However, the degree of orientation is low and extrusion of the blend material through a normal die provides a product with greater orientation at the same apparent EDR.
The SEM photographs described above and shown in Figs. 10-12 demonstrate that the major, minor, and sub-micron particles of the UHMWPE reactor powder undergo swelling when blended with mineral oil and extruded through a swelling die of novel design. This swelling is much greater than that produced by extrusion through a normal die of the same diameter. The swelling is also most obvious at lower [EDR.sub.apparent] values. We suggest here that on exiting the 5 mm length of die the extrudate experiences an internal pressure resulting from the sudden expansion. This allows the UHMWPE to undergo orientation relaxation, enhances the contact surface area, and forces mineral oil further into the particles than extrusion through a normal die would. More UHMWPE chains inside the particles dissolve in the mineral oil; the number of entanglements is reduced and the chains diffuse across particle boundaries. Entanglements are then able to reform in different locations, further toward the particles' cores, enhancing the inter-particle cohesion.
Mechanistic Model for Swelling and Solvating
A model has been devised to explain the differences in the tensile and thermal properties of precursors manufactured from UHMWPE reactor powders in the three following ways: (i) compaction only for pure UHMWPE; (ii) compaction and extrusion through a normal die for an UHMWPE/10 wt% mineral oil blend; and, (iii) compaction and extrusion through a swelling die for an UHMWPE/10 wt% mineral oil blend. A schematic diagram of this swelling and solvating model is shown in Fig. 13. When the reactor powder is compacted, surface-to-surface contact is established, which provides some degree of cohesion. This is the adhesion referred to by Zachariades et al. (25) that produced samples with Young's modulus values of 15 GPa at an [EDR.sub.apparent] = 24. However, under tensile test conditions, this compacted-only material is very brittle, and SEM photographs reveal cohesive failure (24). When 10 wt% mineral oil is added to the UHMWPE reactor powder and this blend is compacted and extruded, the mineral off is forced into the powder particles through the pores within the aggregates making up the major, minor, and possibly, the sub-micron particles. With the mineral oil in intimate contact with the UHMWPE, some of the polyethylene chains dissolve in it. The polyethylene chains now have greater mobility and start to diffuse. It will be easier for them to diffuse into locations that also contain mineral off, which means that they will probably travel across inter-particle boundaries. On cooling and crystallization, the polyethylene chains are locked into position and a better cohesion than surface-to-surface contact (adhesion) is achieved. We believe that this sort of behavior is possible because of the inherently low entanglement density of nascent UHMWPE powder,
When the blend is extruded through a swelling die of novel design, the back pressure increase experienced by the material at the sudden expansion within the swelling die forces more of the mineral off into the powder particles. There is now a greater thickness of each particle in which polyethylene chains may be solvated in the mineral oil and diffuse. As can be seen from Fig. 13, this produces a larger volume of material in continuous, entangled contact. This is how the impressive ductility of swelling die-extruded blend material is brought about.
Future Directions - Two-Stage Drawing
This evidence of improved inter-particle cohesion, brought about by the use of mineral oil and a swelling die, leads to the postulation that on further drawing, materials prepared in the manner described above, should have improved tens fie properties. The direct extrudates are already known to be very ductile and tough. They are soft materials with a high degree of flexibility, which is good for handling during hot draw. The modulus values are low, but are expected to increase with increasing draw ratio, When they are oriented, by hot drawing, the precursors should also prove to be very strong in both the longitudinal and lateral directions. Work on this aspect has already begun and initial results would seem to provide support for this postulation. Some of the direct extrudates were hot drawn at 120 [degrees] C with a crosshead speed of 5 mm/min and then subjected to ambient temperature tensile tests. SEM photographs of the extrudate core morphologies are shown in Figs. 14a and b. For the D10L5 swelling die extrudate drawn to [EDR.sub.actual] = 3.63, Fig. 14a, it can be seen that a fine, lattice structure develops between the clumps of swollen particles. This structure is thought to originate from particles that have fused together successfully. This is shown more clearly in Fig. 14b, a D8L5 swelling die extrudate drawn to [EDR.sub.actual] = 3.91. Here, elongated particle structures may be observed between the clumps of more spherical particles. The result of drawing these direct extrudates further may be seen in Fig. 14c. This shows the morphology of a longitudinally fractured direct extrudate hot drawn to [EDR.sub.actual] = 19. The swollen structures have been drawn into long, continuous, oriented fibers with diameters of [less than] 0.1 [[micro]meter]. There are horizontal connecting fibers, which we believe will improve the transverse tensile strength of the drawn products. This requires further investigation and testing. These transverse fibers are actual structures and not merely debris or other fracture artifacts. They occur at different depths within the SEM photograph, joining different longitudinal fibers. Some have broken during drawing and curled back on themselves to their point(s) of origin. As previously, the dark regions we associate with mineral oil and voids when the oil is removed. This extraction is the subject of a separate project and will allow these precursors to be used as materials for membranes.
Initial tensile testing of these hot drawn products shows retention of the good ductility at relatively high total draw ratios; see Fig. 15. The values of Young's modulus in this plot are based on the blend material. When the mineral oil content is quantified, the modulus values, and eventually the tensile strengths, per weight of polyethylene, will be much higher. According to the normalized modulus versus draw ratio plot of Wang et al. (1), the Young's modulus values of the present materials are equivalent to those of solid state coextrusion products. The highest total draw ratio achieved so far is 36, with a Young's modulus of 67.6 GPa. Referring to Table 1, the processes producing similar draw ratios - die-free gel spinning (14), swell drawing (13), ram extrusion (16), compression molding (15), and gel spinning plus hot draw (12) - produce materials with higher Young's moduli and which fracture at the given draw ratio. The UHMWPE precursors produced using the swelling die and hot drawing have not yet fractured under tensile test and are still very ductile. These swollen extrusion products are difficult to fracture because of this ductility and equipment load limitations. When fracture does occur it is probably the result of polyethylene fibers' slipping over each other.
This study has shown that the use of a novel swelling die has enabled us to produce precursors which exhibit very high room temperature ductility and toughness. The DSC results show the extrudates have more uniform thermal properties when compared with both the pure UHMWPE extrudates and those extruded using the blend and normal dies. The SEM micrographs indicate the extrudates show swollen structures, especially those from shallow extrusion dies.
The increase in toughness is attributed to the increase in chain penetration depth across the powder particle boundaries. The swelling effect of the mineral oil enabled molecules at the particle surfaces to become highly mobile and diffuse easily across particle boundaries. These chains recrystallize upon cooling, giving rise to enhanced cohesive strength or toughness. This diffusion and recrystallization takes place at all particle size levels.
The large room temperature ductility (up to 150%) indicates that the precursor is highly porous. This is further demonstrated by the Final drawn products. There is also visual evidence of transverse connecting fibrils in the drawn products, which suggests that the transverse tensile strength of the products should be better than in previous materials,
This project was funded by the Research Grants Council of Hong Kong with an Earmarked Grant for Research, grant number HKUST 580/94E. The authors acknowledge the generosity of Montell Polyolefins in supplying the UHMWPE used in this project.
1. L.-H. Wang, R. S. Porter, and T. Kanamoto, Polym. Commun., 31, 457 (1990).
2. P. J. Lemstra and R. Kirschbaum, Polymer, 26, 1372 (1985).
3. A. E. Zachariades, W. T. Mead, and R- S. Porter, Chem. Rev., 80, 351 (1980).
4. L. R. G. Treloar, Polymer 1, 95, 279 (1960).
5. R. F. Schantele and T. Shimanouchi, J. Chem. Phys., 47, 3606 (1967).
6. T. Kanamoto and R. S. Porter, in L. A. Kleintjens and P. J. Lemstra, eds., Integration of Fundamental Polymer Science and Technology, Applied Science, London (1989).
7. R. S. Porter and T. Kanamoto, Polym. Eng. Sci., 34, 266 (1994).
8. P. Smith, H. D. Chanzy, and B. P. Rotzinger, J. Mater. Sci., 22, 523 (1987).
9. Y. M. T. Tervoort-Engelen, and P. J. Lemstra, Polym. Commun., 32, 343 (1991).
10. P. Smith and P. J. Lemstra, Polymer, 21, 1341 (1980).
11. P. Smith and P. J. Lemstra, Makromol. Chem., 180, 2983 (1979).
12. P. Smith and P. J. Lemstra, J. Mater. Sci., 15, 505 (1980).
13. M. R. Mackley and S. Solbai, Polymer, 28, 1115 (1987).
14. M. R. Mackley and S. Solbai, Polymer, 28, 1111 (1987).
15. B. P. Rotzinger, H. D. Chanzy, and P. Smith, Polymer, 30, 1814 (1989).
16. A. E. Zachariades, M. P. C. Watts, and R. S. Porter, Polym. Eng. Sci., 20, 555 (1980).
17. L.-H. Wang and R. S. Porter, J. Polym Sci.: Polym. Phys. Ed., 28, 2411 (1990).
18. P. Smith, H. D. Chanzy, and B. P. Rotzinger, Polymer Commun., 26, 258 (1985).
19. T. Kanamoto, T. Ohama, K. Tanaka, M. Takeda, and R. S. Porter, Polymer, 28, 1517 (1987).
20. A. E. Zachariades, Polym. Eng. Sci., 25, 747 (1985).
21. A. E. Zachariades and T. Kanamoto, Polym. Eng. Sci., 28, 658 (1986).
22. A. E. Zachariades and J. A. Logan, J. Polym. Sci.: Polym. Phys. Ed., 21, 821 (1983).
23. R. W. Truss, K. S. Hah, J. F. Wallace, and P. H. Geil, Polym. Eng. Sci., 20, 747 (1980).
24. P. Gao, M, K. Cheung, and T. Y. Leung, Polymer, 37, 3265 (1996).
25. A. E. Zachariades, M. P. C. Watts, T. Kanamoto, and R. S. Porter, J. Polym. Sci.: Polym. Lett. Ed., 17, 485 (1979).
26. T. Y. Leung, PhD, thesis, The Hong Kong University of Science and Technology, Hong Kong (1995).
27. P. Smith, P. J. Lemstra, and H. C. Booij, J. Polym., Sci.: Polym. Phys. Ed., 19, 877 (1981).
28. A. M. Kotliar and R. A. Black, J. Polym. Sci.: Part B: Polym. Phys., 28, 1033 (1990).
29. C. Whitehouse, M. L. Liu and P. Gao, to be published in Polymer.
30. C. Whitehouse, M. L. Liu, and P. Gao, in preparation.
31. S. Mazur, in M. Narkis, and N. Rosenzweig, eds., Polymer Powder Technology, Wiley, Chichester, U. K. (1995).
32. B. Wunderlich and C. Cormier, J. Polym. Sci.: Polym. Phys. Ed., 5, 987 (1967).
33. S. Ottani and R. S. Porter, J. Polym. Sci.: Part B: Polym. Phys., 2 9, 1179 (1991).
34. H. H. Chuah, R. E. DeMicheli, and R. S. Porter, J. Polym. Sci., Polym. Lett Ed., 21, 791 (1983).
|Printer friendly Cite/link Email Feedback|
|Author:||Whitehouse, C.; Liu, M.L.; Gao, P.|
|Publication:||Polymer Engineering and Science|
|Date:||May 1, 1999|
|Previous Article:||A new method for estimating the cellular structure of plastic foams based on dielectric anisotropy.|
|Next Article:||Polypropylene reactor blends: composition evaluation by analytical TREF.|