This invention relates to solid phase deformation processes; more particularly, this invention relates to the solid phase deformation of brittle composites having a phase made up of orientable, semi-crystalline, thermoplastic polymeric material; and to improved polymeric composites obtained thereby.

During the last decade, a very substantial amount of research has been affected into the development of composites made up of a flexible polymer matrix reinforced with a ceramic powder. Also during the last twenty years, a very substantial amount of research has been affected into the improvement of, principally, the mechanical properties of flexible polymers such as those making up the matrix phase of the composites, by orientation. It has long been recognised that manipulation of the molecular orientation of a ductile polymer leads to a significant enhancement in the stiffness and strength along the orientation direction, as well as providing a means of tailoring the material anisotrophy. There are several well established technologies for inducing molecular orientation in ductile polymeric material, namely continuous tensile drawing (1-4), die drawing (5) and hydrostatic extrusion (6-10). Notably, all of these techniques have long been practised on ductile polymers, ie having those polymers having a high yield strength and a relatively low fracture strength and thus being able to resist breakage.

An example of a composite typically used in a solid phase deformation process and also made up of the, aforedescribed, flexible polymer matrix reinforced with a ceramic powder, is isotropic polyethylene (PE) reinforced with synthetic hydroxyapatite (HA). This material is typically compression moulded in order to provide a suitable product. This technique is used

largely because the material has brittle, as opposed ductile, properties, ie the composite is prone to breakage because the fracture strength is relatively low compared to the yield strength.

Reference herein to brittle composites is intended to include those composites which are unable to withstand loads, such as those greater than a minor physiological nature, and as a result of this suffer damage. A load of a minor physiological nature would be one typically withstood by a cortical bone.

Reference herein to ductile composites is intended to include those composites which are able to withstand loads, such as those greater than a minor physiological nature, and as a result of this do not suffer damage. A load typically withstood by such a composite would be of a major physiological nature and so be equal to or greater than the load withstood by skeletal bone.

Nevertheless, composites of a brittle sort, and preferably those including approximately 40 vol % HA, can be compression moulded to provide a bone substitute with ceramic particles homogenously distributed in the polymeric matrix. Typically, the composite is produced using a hot compounding process. A composite with 40 vol % HA is considered to have a ductility within the bounds of cortical bone; stiffness and strength suitable for prothesis under minor physiological loads; fracture toughness superior to bone; and favourable bioactivity, such as to encourage bone apposition and the virtual disappearance of the boundary between the natural bone and the composite a few months after implantation (11-14). A composite as afore described has now been commercialised under the trade mark HAPEX (Smith and Nephew, Richards, USA) and is finding increasing use in a range of

minor load bearing applications such as orbital floor reconstruction (15) and otologic implants.

Despite this success, there remain important areas of orthopaedic surgery which require bone substitute materials with considerably higher stiffness and strength, comparable to the values associated with cortical bone.

It follows that there is therefore a need to either produce new composites having properties of stiffness and strength that are suitable for the manufacture of protheses subject to major physiological loads, or alternatively for a process to be developed which essentially converts brittle composites into ductile composites and in so doing alters the stiffness and strength of the composite such that the yield strength is relatively greater than the fracture strength. It is with this latter need that the current invention is concerned.

We have, contrary to conventional expectations, been able to subject a relatively brittle composite to a particular process and as a result of this process transform the properties of the brittle composite in order to produce a ductile composite having stiffness and strength properties which make the material suitable for use, particularly, but not exclusively, in a wide range of orthopaedic implants.

It is therefore an object of the invention to provide a process which alters the stiffness or strength or ductile properties of a brittle material, such as a brittle polymer, thus increasing any one or more of the aforementioned properties with a view to providing a ductile material which is particularly, though not exclusively suitable for use in orthopaedic implants.

According to a first aspect of the invention there is therefore provided a process for the deformation of a workpiece comprising an orientable, thermoplastic, brittle polymer by passage of the workpiece through a die having both an entry side and an exit side, which process comprises positioning the workpiece at the entry side of the die and causing the workpiece to deform in the solid phase through the die and collecting the deformed workpiece through the exit side of the die.

In a preferred process of the invention said process of deformation may be undertaken by either die drawing or hydrostatic extrusion.

In principle, both die drawing and hydrostatic extrusion offer the potential of producing oriented polymers with sufficiently large dimensions for the fabrication of products such as orthopaedic implants. Both die drawing and hydrostatic extrusion are processes involving a billet which is made to pass through a convergent die by the application of either a tensile force (die drawing) or a back pressure (hydrostatic extrusion).

In yet a further preferred process of the invention said die is of a convergent nature.

In yet a further preferred process of the invention the workpiece is heated to a sufficiently high temperature to allow flow, but at a temperature below the workpiece melting point. Heating the workpiece in this way enables a high degree of molecular orientation to take place.

We prefer to use hydrostatic extrusion in order to effect the invention because the deformation occurs in a totally compressive stress field. Moreover, when

we use hydrostatic extrusion we prefer to provide a back pressure, in order to effect the extrusion, by using a technique which involves surrounding the workpiece with a suitable fluid and using a piston to produce a hydrostatic pressure in the fluid. Indeed, the name hydrostatic extrusion is usually restricted to this particular technique which eliminates friction between the workpiece and a container wall and lowers the die friction because of the presence of the pressurising fluid. An additional advantage of this technique is that heat is transmitted to the workpiece by the fluid, thus reducing the risk of uneven temperature gradients.

In yet a further preferred process of the invention said workpiece comprises an orientable, thermoplastic brittle polymer reinforced with particles of ceramic materials. More preferably still said workpiece comprises a composite of isotropic polyethylene and hydroxyapatite. More preferably still said composite includes, typically, synthetic hydroxyapatite in an amount of 40 vol %.

Preferably, the process effects a reduction in the bulk cross - section area of the workpiece (by "bulk cross-sectional area" is meant the area of the bulk of the workpiece normal to the machine direction). Preferably, the die is a reducing die.

Not all polymers are capable of existing with an extended chain crystalline morphology; however, the process of the present invention is applicable to ceramic reinforced linear polyethylene. From a commercial standpoint, the process of the present invention is of particular importance in relation to linear polyethylene, preferably having a weight average molecular weight (Mw) from 50,000 to 10,000,000, especially from 100,000 to 8.000,000

reinforced with hydroxyapatite with particle size 2 μm, specially from 3 μm and 10 μm and hydroxyapatite fraction from 10 volume per cent to 70 volume per cent, specially from 20 volume per cent to 60 volume per cent.

The term "workpiece" as used herein includes bars, strips, rods, tubes and other cross -sections of solid or hollow stock. The term includes both billets and other forms of stock of greater length, indeed, continuous stock, which may be formed as the process is performed, maybe utilised.

The orientable, semi-crystalline, thermoplastic polymer is reinforced with ceramic particles or fillers. Examples of useful fibrous fillers include glass, asbestos, metal, carbon and ceramic whiskers, such as those formed with silicon carbide. Examples of useful laminar fillers include mica, talc and graphite flakes. Chalk and ply may also be included. The amount of fillers, other than the ceramic particles, which may advantageously be included depends on the nature of the filler, but up to 50 volume per cent, preferably not less than 30 volume per cent, specially not less than 20 volume per cent may be incorporated.

In accordance with a preferred embodiment of this invention to the workpiece is caused to deform through the reducing die in the solid phase by hydrostatically extruding it there through. Draw-assisted hydrostatic extrusion, as described in UK Patent No. 1480479, may be utilised with advantage. The net hydrostatic pressure ( that is, the difference between the applied extrusion pressure and the atmospheric pressure) for extrusion to occur will, at a given extrusion temperature will have a value from 0.2 to 3.0 kbar, preferably from 0.5 to 2.0 kbar, for example 1 kbar, is found to be suitable.

It is also preferred that the reducing die temperature is above the Tg of the polymer matrix of the composite but below the melting point of said polymer matrix.

The invention provides a composite of an oriented, semi-crystalline thermoplastic polymer reinforced with ceramic particles, preferred by the process of the present invention.

This invention further provides composites of ceramic particles reinforcing an oriented linear polyethylene which has deformed to a deformation ratio no greater than 20; for example 10, and which has an axial modulus of at least 10 Gpa; for example, of at least 6 Gpa.

In yet a further preferred process of the invention the said workpiece is extruded through the said die at an extrusion ratio of at least 5 : 1 and more preferably between 5: 1 and upto and including 11 : 1, preferably 8: 1.

According to one further aspect of this invention there is provided a process for the deformation of a workpiece comprising an orientable, thermoplastic polymer reinforced with particles of ceramic by passage of the solid phase through a die having both an entry side and an exit side, which process comprises providing the workpiece comprising the orientable, thermoplastic polymer reinforced with particles of ceramic, the composite being initially at the entry side of the die, causing the workpiece to deform in the solid phase through the die, and collecting the deformed workpiece through the exit side of the die.

According to a yet further aspect of the invention there is provided the

product produced by any of the aforementioned processes.

The invention will now be described by way of example only with reference to the following information wherein:-

Table 1 shows a schedule of temperature-pressure to produce plates and billets of 100% PE;

Table 2 shows the results of mechanical testing of extruded products and more specifically flexural test parameters;

Table 3 shows simple beam theory formula for three point ending;

Table 4 shows the maximum standard deviation, expressed as a per cent, for both die swell and flexural tests;

Table 5 shows extrusion characteristics and flexural properties of HA/PE composites;

Figure 1 is a longitudinal, cross-sectional, diagrammatic representation of a hydrostatic extrusion rig;

Figure 2 shows DSC melting endotherm of 100% Rigidex HM 4560 XP. a) As supplied, b) process through a twin screw extruder, c) compression moulded, d) and e) hydrostatically extruded to ER= 5: 1 and ER= 8: 1, respectively;

Figure 3 shows DSC melting endotherm of HA/PE composite (40 vol %

HA), a) Processed through the twin screw extruder and compression moulded, b), c) and d) hydrostatically extruded to ER= 5: 1, ER= 8:1 and ER= 11: 1, respectively.

Materials

HA is a calcium phosphate ceramic [Ca _]0 (PO) ₆ (OH,)]. Grade P88, manufactured by Biotal Ltd, UK, a synthetic HA, was used, with 4.14 μm average size (20) and a density of 3.156 g cm ^'3 The polymer was Rigidex HM 4560XP (BP Chemicals Ltd., UK), as used in HAPEX. This is an ethylene hexene co-polymer with less than 1.5 butyl branches per 1000

carbon atoms, ^M _™ = 225,000, « = 24,000 and 0.945 g cm ^'3 density.

Preparation of Composites

Details of the preparation of isotropic HA/PE composites have been given by Wang et al (14). Briefly, the HA and PE powders were pre-mixed in a domestic blender and compounded in a twin screw-extruder at temperatures between 210 and 250 ^g C. The extruded rods were pelletized and powderized with a mesh size of 1 mm.

It has been reported(16) that the compounding process does not completely break down the HA agglomerations for composites with HA content greater than 40 vol %. In an attempt to improve the homogeneity of a composite with 50 vol % HA, this was further powderized using sieves with 500 μm, 200 μm and 80 μm perforations. The added stages were carried out in a Fritsch Pulverisette 14 Rotor-Speed Mill (Fritsch Gmbh, Idai-Oberstein, Germany) fitted with a 12 knife stainless steel rotor at a speed of 16,000 rpm.

The material to be fed into the Rotor - Speed Mill was kept in a beaker which was immersed in liquid nitrogen. However, no liquid nitrogen was poured into the machine.

Preparation of Samples (rectangular plates and cylindrical billets)

The isotropic composite samples for flexural testing were 70 x 40 mm rectangular plates 2 mm thick, compression moulded in a stainless steel mould using a hydraulic hot press. The powder in the mould was sandwiched between two sheets of aluminium foil to facilitate the release of the sample. The temperature during compaction was monitored with a probe connected to an electronic thermometer. 5 mm of the sensing end of the probe could be inserted with a tight fit into holes bored in the mould. The composite was maintained at 180 ² C for 20 minutes under low pressure to ensure good thermal contact between the mould and the two hot plates of the press. At the end of this period the compressive load was increased rapidly to 9 tonne (30 MPa pressure), followed by switching off the heating and water cooling of the hot press. The mould was maintained at constant pressure until it reached a temperature of 50 ^Q C in about 30 minutes, when it was removed from the press to cool down to room temperature before retrieving the sample.

For composites with 40 vol % HA or lower, the above procedure resulted in excessive flash and no control on the thickness of the plate. In these cases the pre-determined thickness (2 mm) was achieved by placing suitable spacers in the mould and using a moderate excess of composite powder.

For 100 % PE, the technique described above produced plates with a large

number of voids. The problem was due to inhomogeneous cooling from the melt and the associated volume shrinkage. In highly crystalline polymers, such as polyethylene, the main contribution to shrinkage arises from the recrystallization process taking place during solidification, while thermal contraction makes only a small contribution. Cawood and Smith (17) studied the problem and the following procedure, based on their work, produced excellent void-free polyethylene plates.

The mould, with sufficient amount of powder polymer but no spacers, was placed in the press under the smallest possible pressure, just enough to ensure contact with the hot plates. The assembly was brought to 175 ² C, relieving the pressure as this tended to rise. After 20 minutes at 175 ² C the heating was switched off and the temperature lowered at a rate of 2 to 4 -C per minute using the water cooling system of the press, while the pressure was adjusted according to the schedule shown in Table 1. Below 100 ⁹ C the pressure remained constant and the mould was left in the press until it reached a temperature of 50 ^Q C, when it was removed from the press to cool down to room temperature before retrieving the sample.

The billets for hydrostatic extrusion were cylindrical rods 60 mm long and 12 mm diameter. These were produced in a stainless steel mould and the temperature-loading schedules were similar to those used for the rectangular plates of the corresponding materials, namely HA/PE composite or 100 % polyethylene. No spacers were used.

Some billets of HA/PE composites were produced by injection moulding.

The equipment used was an SL2 Air Operated Machine, made by J.B. Engineering, Chippenham, UK. The moulding temperature was 180 ³ C and

the rod dimensions were 90 mm long and 10 mm diameter.

Hydrostatic Extrusion

Details of the experimental procedure are described by Gibson and Ward (7) and only a brief summary will be presented here.

In Figure 1, the apparatus comprises a generally cylindrical hydrostatic extrusion vessel 1 containing an axially aligned chamber 2, having a diameter of 40mm and an effective length of 170mm, which is closed at an upstream end. The downstream and of the chamber is formed with two, internal, axially symmetric shoulders 4 and 5 which are each threaded to receive a threaded conical reducing die 6 in which a workpiece 7 is seated. The extrusion vessel has a port 12 which communicate with a pressure varying system (not shown); and is provided with a circumferential band heater 15 enabling the tooling temperature to be maintained within 2 ^S C.

The hydrostatic extrusion apparatus shown in Fig. 1. Die 6 which has a cone semi-angle of 15 ² and bore diameters of 1.8 mm, 2.5 mm, 3.0 mm or 3.5 mm, according to the extrusion ratio (ER = 5: 1, 8: 1 or 11: 1) and the diameter of the cylindrical billets. These were machined with a 15 ^s nose to create an initial pressure seal. At the end of the nose a constant diameter stub was also machined, which protruded a few millimetres through the die. A cable attached to the stub was used to drive a rotary potentiometer to provide a displacement signal which was recorded from the beginning of the extrusion. A haul-off load of 100 g was attached to the free end of the cable to ensure a firm drive of the rotary potentiometer. The back 3 mm of the billet was machined to a larger diameter to act as a plug and prevent the violent release

of pressure and hot fluid at the end of a run.

The pressurising fluid was castor oil (J.L. Seaton, Hull, U.K.). The billets were coated with two layers of Evostick (Evode Ltd., UK) to avoid direct contact between the polymer matrix and the pressurising fluid, which involves a risk of stress cracking (7). Each applied layer of glue was allowed to dry for several hours. It was found that the Evostick coating peeled-off during extrusion and did not go through the die.

The extrusion pressure was a function of the material and the extrusion ratio, varying between 19 MPa and 207 MPa. There was little control of the extrusion rate, which was about 1.5 mm min ^"1 (vide infra).

The use of the apparatus is described in the following example which illustrates the invention.

The billet was prepared as aforedescribed and a nose was then machined on each billet so that it would accurately mate with the reducing die as shown in the accompanying drawing. Two reducing dies having bore diameters of

2.5 and 3.5mm were used, the conical semi-angle was 15 ² in each case.

Each billet was, in turn, urged into position in the reducing die, the tooling was assembled as shown in the accompanying drawing and the chamber filled with a silicone oil (DC 550 ex Dow Corning) and suitably bled. The band heater was energised, the silicon oil being raised to an equilibrium temperature of 115 ⁹ C and the pressure raised to 1.4 kbar (when using the reducing die with 3.5 mm bore diameter). The workpiece (or billet) was extruded through the die at an extrusion speed of up to 20mm min-1, for

example 1 mm min-1.

At the termination of the extrusion, the pressure was, in each case released and the temperature allowed to fall below 100 ² C before the product was extracted.

Differential Scanning Calorimetry (DSC)

The effect of the hydrostatic extrusion on the morphology of the PE matrix was qualitatively assessed by studying the melting behaviour of the material. For this purpose, a Perkin Elmer Differential Scanning Calorimeter DSC7 (Perkin Elmer Corp., Norwalk, Connecticut, USA) was used with a scanning rate of 10 ² C min ^"! and 2 to 10 mg of material for each run.

Mechanical Testing

Samples were tested in three point bending. Two types of geometry were used, i) bars with rectangular cross sections and, ii) cylindrical rods (hydrostatically extruded samples). Tests of rectangular bars followed ASTM 790 recommendations (18), while rods were tested with their original extruded diameters. It has been shown that the results obtained with both geometries are fully comparable (19).

Three main measurements were made in flexure; modulus (FM), strength (FS) and ductility (FD). The formulae used to calculate these properties are given by the simple beam theory (22). For convenience, the flexural test parameters and the formulae are shown in Tables 2 and 3, respectively. Note that the flexural modulus was measured with two different gauge lengths,

namely 130 mm (FM , ₃₀ ) and 50 mm (FM ₅₀ ). The terms "strain" and "ductility" refer to the maximum strain (in %) in the sample at a given deflection. For all samples the following condition applied.

Gaugelength -> 15

Thickness^ diameter) as required by ASTM 790 (26).

Some rods did not break but, instead, the load-deflection graph exhibited a peak stress. In these cases the flexural strength (FS) and ductility (FD _m ) were measured at maximum stress. A further value of the maximum strain in the rod (FD. ₂₅ ) was calculated when the load decreased by 25 % of the peak stress. However, FD. ₂₅ is only quoted for samples which showed no crack formation, as evidenced by a step-like decrease in the load-deformation graph.

All the mechanical tests were carried out at room temperature (22.0 = 1.5 ² C) using an Instron machine TT-CM (Instron Ltd, High Wycombe, UK). A minimum of four nominally identical samples were used for each material, but this number was often increased through the availability of samples made for other experiments (for example, the modulus, a non-destructive test, was measured with up to 16 nominally identical samples). The maximum standard deviations (SD, in per cent of the property value) are shown in Table 4.

RESULTS

Previous work (7) showed that linear polyethylene should be extruded at 100 ² C for optimum balance between the increasing deformability of the material and loss of product properties due to annealing effects at higher

temperatures. Preliminary experiments indicated that the composite systems studied in this research required a higher temperature for successful hydrostatic extrusion and 115 ² C was adopted throughout.

The preliminary experiments also showed that the extrusion characteristics and the mechanical properties of the extrudates are not affected by the type of moulding of the billets (compression or injection moulding). Therefore, this parameter is omitted from the presentation and discussion of the results.

Table 5 shows the flexural properties of the systems tested. The table also includes a range of values associated with cortical bone. (20) It is seen that hydrostatic extrusion of HA/PE composites can produce improvements in the stiffness and strength of over 100 %, accompanied by increases in ductility of at least 400 % . Thus, the hydrostatically extruded systems achieve levels of stiffness and strength within the range associated with cortical bone, whilst exhibiting a vastly superior ductility.

Polymeric materials usually swell on leaving the die. The effect is known as "die swell" and Gibson and Ward (7) define it as

^{e d} -xlOO ^d « (1)

where d _c and d _d are the extrudate and die bore diameters, respectively. Expression (1) will be referred to as diametrical die swell. However, when dealing with composite materials it is more useful to define the die swell in terms of the cross-sectional areas of the extrudate (A _e ) and die bore (A^) (vide infra), namely

which will be called X-section die swell. The values of the diametrical and X-section die swells are given in Table 5, while Table 4 shows the maximum standard deviations applying to the die swell measurements, in per cent of their average value. Owing to the die swell effect, Table 5 gives two extrusion ratios, namely the nominal (calculated from the die bore diameter) and the actual (calculated from the extrudate diameter).

Table 5 shows that the incorporation of HA into the polyethylene is accompanied by a significant reduction of the die swell effect.

DISCUSSION

Flexural Measurements

Table 5 indicates that the modulus and strength of HA/PE composites increased with increasing hydroxyapatite content and extrusion ratio, although some levelling-off appears to have taken place for the highest values of these variables used in this work. However, the mechanical properties of the composites have already attained the values for cortical bone. The ductility of the composites decreased sharply with increasing hydroxyapatite fraction, but the application of hydrostatic extrusion was associated with substantial improvements of this property, to the extent that all the hydrostatically extruded systems displayed a ductility above the maximum value for cortical bone.

The trends shown in Table 5 for the flexural properties of unextruded HA/PE

composites (increase of modulus and strength and decrease of ductility for increasing hydroxyapatite content) are in agreement with the results of Abra et al (16) and Wang et al (14) for the tensile properties of similar systems.

The results presented in Table 5 suggest that an HA/PE composite with optimum mechanical properties can be obtained with 40 vol % HA and ER - 8: 1. Higher HA content could improve the biological properties of the material, but these systems present considerable production difficulties (16) while not providing generally superior mechanical properties. Extrusion ratios above 8: 1 are possible, notwithstanding the increase of the extrusion pressure, but there appears to be no significant gains in the mechanical properties.

Characteristics of the Hydrostatic Extrusion of HA PE Composites

Gibson and Ward (7) made a thorough empirical study of the hydrostatic extrusion of polyethylenes with different molecular weight. Their work provides a suitable basis to gain some understanding of various features of the hydrostatic extrusion of the composites. Broadly, three main aspects will be discussed; a) die swell, b) factors affecting the control of the process and c) extrudate faults.

a) Die swell. It may be safely assumed that die swell of HA/PE composites is associated with the polyethylene phase only, with no contribution from the HA phase. A further assumption will be made, namely that the geometrical changes take place at constant volume. Under these conditions, the cross-section die swell of the HA/PE systems should be inversely proportional to the HA fraction.

Examination of Table 5 reveals that the inverse proportionality regime overestimates the cross-section die swell of the composites.

This discrepancy may be qualitatively understood on the grounds that, at a given nominal extrusion ratio, the deformation of the polyethylene phase increases with increasing HA content. Gibson and Ward (7) have shown that for polyethylene with Λ/ _w ~ 100,000, die swell decreases rapidly to zero as the nominal extrusion ratio increases towards 10: 1. A similar trend should apply to Rigidex HM 4560XP. Therefore, the incorporation of HA decreases the die swell through two mechanisms, i) a decrease of the polymeric phase fraction, ii) a further reduction owing to the increased deformation of this phase.

b) Factors affecting the control of hydrostatic extrusion. In the "Experimental" section above it was mentioned that there was little control of the extrusion rate of HA/PE composites. This characteristic may be compared with results reported by Gibson and Ward, (7) namely, for low nominal extrusion ratios the slope of extrusion pressure vs. extrusion rate is small and the latter can be chosen over wide limits. At higher extrusion ratios the extrusion rate reaches a limiting value whereupon large increases in pressure have very little effect on the extrusion rate. For example, for a polyethylene of Λ/ _w ~ 100,000 and ER~10: 1, extrusion rates of between 0.1 mm min ^"1 and 10 mm min ^"1 can be achieved with pressures varying between 35 MPa and 55 MPa, respectively. On the other hand, the same polymer with ER=20: 1 has a limiting extrusion rate of about 1.5 mm min ^"1 , that is, similar to the HA/PE composites. Although the extrusion ratios applied to these systems are 11.1 at most, the behaviour of the material may reflect the fact that the deformation is concentrated in the polymeric phase only.

Nevertheless, the pressure was occasionally increased in an attempt to speed up the process, but the extrusion rate entered an exponential regime leading to a blow up. The short section of the extrudate coming out of the die during the exponential regime showed faults similar to the stick-slip effect reported by Gibson and Ward (7).

The pressure required to maintain steady-state extrusion was often lower than the initial breakthrough pressure. This characteristic was also observed by Gibson and Ward (7) for the hydrostatic extrusion of polyethylene. These authors advanced an explanation in terms of the high static friction coefficient at the billet-die interface, according to Pugh's discussion. (21) When the billet begins to move, fluid is drawn into this interface, causing the friction coefficient to decrease. Gibson and Ward (7) do not state what happens if the pressure is not decreased after breakthrough, but for HA/PE composites it was found that the extrusion rate could enter the exponential regime referred to above, leading to a blow-up. No satisfactory explanation of this behaviour can be given, but it should be noted that the friction coefficient at the billet-die interface should be higher for the composites than for unfilled polyethylene, as suggested by the observation that the filled material produce significant wear of the stainless-steel die. Also, the flow of d e pressurizing fluid into the billet-die interface should be affected by the different surface properties of the composite or polyethylene billets.

DSC Studies

DSC is a useful experimental technique to study the effect of the various processing stages on the morphology of HA/PE composites and their qualitative correlation with the resultant mechanical properties. Two main

cases are considered, i) DSC of 100 % Rigidex HM 4560XP (Fig. 2) and, ii) DSC of composites with 40 vol % HA (Fig. 3).

i) DSC of 100% Rigidex HM 4560XP. Figure 2a shows the scan of the polyethylene powder as supplied and (Fig. 2b) after processing through the twin screw extruder under similar conditions to those used to incorporate HA into the polyethylene. The melting endotherms are similar, indicating that the compounding process did not result in any significant morphological changes in the polymer. The melting range seen in Figs. 2a and 2b is somewhat lower than expected for polyethylene, owing to Rigidex HM 4560XP being a copolymer (see Experimental section).

Compression moulding of the polyethylene produces a moderate increase of the melting point (Fig. 2c). A further increase is seen after hydrostatic extrusion to both ER = 5: 1 (Fig 2d) and ER = 8:1 (Fig. 2e), indicating an oriented morphology which resulted in a 100 % improvement in stiffness and strength when compared with the unextruded polymer (Table 5). Figures 2d and 2e show that the polymer with extrusion ratios of 5: 1 and 8: 1 display broadly similar melting behaviour. Consequently, the corresponding stiffness and strength were also similar.

The failure of higher nominal extrusion ratios to produce a stiff er and stronger polymer may be partly due to die swell, giving rise to relatively similar actual extrusion ratios (Table 5). Moreover, Gibson and Ward (7) have shown that the effect of higher extrusion ratios on the stiffness and strength of the extrudate decreases with increasing molecular weight. Rigidex HM 4560XP lies in the middle-to-high range of molecular weight for commercial polyethylenes and its mechanical properties may not be very

sensitive to the extrusion ratio.

ii) DSC of HA/PE composites. Figures 3a and 2b show the melting behaviour of the polyethylene after processing in the twin screw extruder with and without HA respectively. It may be seen that the incorporation of the filler leads to a moderate decrease of the melting peak of the polymer, probably reflecting the higher energy moφhology associated with the creation of interfaces and/or the hydroxyapatite particles providing low energy nucleation sites, resulting in smaller crystals. Compression moulding of the composite increased the endotherm peak temperature (Fig. 3b), as it was the case for the unfilled polymer. Further substantial increases were induced by hydrostatic extrusion with ER = 5: 1 and ER = 8: 1 (Figs. 3c and 3d, respectively). However, Fig. 3e shows that ER = 11: 1 has not resulted in an additional increase of the endotherm peak temperature. The stiffness and strength follow similar trends, (Table 5) namely a major increase with ER =

5: 1, followed by a moderate increase up to ER = 8: 1 with little change thereafter up to ER = 11 : 1. Both the melting behaviour and the progression of the stiffness and strength may be qualitatively understood on the grounds of an increasingly oriented matrix as the extrusion ratio increased. The orientation was enhanced by the incorporation of HA, as discussed in the die swell section. For extrusion ratios above 8:1 any additional deformation has taken place through moφhological changes which has little effect on the melting behaviour of the matrix and on the stiffness and strength of the composite, although the ductility decreased (Table 5).

Table 5 shows that the incoφoration of HA had a strong reinforcing effect on the mechanical properties of unextruded polyethylene. However, it is not yet possible to assess the reinforcement effect of the filler in the extruded

material because the presence of HA may have affected the mechanical properties of the composite by increasing the deformation of the polymeric matrix, as well as acting as a straight reinforcement.

Hydrostatic extrusion has been successfully applied to highly filled particulate hydroxyapatite reinforced polyethylene composites. The resultant mechanical properties of the extruded product exhibit significantly increased ductility (and hence toughness). Therefore, HA/PE composites processed by this route can be used in major load bearing skeletal prostheses.

REFERENCES

1) G. Capaccio and I.M. Ward, Nature Phys. Sci., 243, 143 (1973).

2) P. Smith and P.J. Lemstra, J. Mater. Sci., 15, 505 (1980).

3) G. Capaccio and I.M, Ward, Polymers, 15, 233 (1974). 4) G. Capaccio, T.A. Crompton and I.M. Ward, J. Polym. Sci. - Polym. Phys. Ed., 14, 641 (1976).

5) P.D. Coates and I.M. Ward, Polymer, 20, 1553 (1979).

6) A.G. Gibson, I.M. Ward, B.N. Cole and B. Parsons, J. Mater. Sci. Lett., 9, 1193 (1974). 7) A.G. Gibson and I.M. Ward, J. Polym. Sci. - Polym. Phys. Ed., 19,

2015 (1978).

8) P.S. Hope, A.G. Gibson and I.M. Ward, J. Polym. Sci. - Polym. Phys. Ed., 18, 1213 (1980).

9) P.S. Hope and I.M. Ward, J. Mater. Sci., 16, 1511 (1981). 10) A.G. Gibson: PhD Thesis, The University of Leeds (1977).

11) W.Bonfield, C.Doyle and K.E. Tanner in "Biological and biomedical performance of biomateriais" edited by P.Christel, A.Meunier, A.J.C Lee (Elsevier Science Publishers, Amsterdam 1986), pp 153-158.

12) W.Bonfield, J.Biomed. Eng., 10 (1988) 552-526. 13) W.Bonfield in "Advanced services in ceramics - vol 1. An Introduction to Biosciences", edited by L.Littench and J.Wilston (World Scientific

Publishing Co., Singapore 1993) pp 299-303.

14) M.Wang, D.Porter and W.Bonfield, Brit. Ceram. Trans., 93 (1994) 91-

95. 15) K.E.Tanner, R.N.Downes and W.Bonfield, Brit. Ceram. Trans., 93

(1994) 104-107.

16) J. Abram, J. Bowman, J.C. Behiri and W. Bonfield, Plast. Rubber

Process. Applic, 3, 261 (1984).

17) M.J. Cawood and G.A.H. Smith, Polymer Testing, 1, 3 (1980).

18) ASTM D 790M-93, American Society for Testing and Materials, Easton, Maryland (1993). 19) N.H. Ladizesky, I.M. Ward and W. Bonfield, J. Appl. Polm. Sci. (to be submitted).

20) W. Bonfield, J. Biomed. Eng., 10, 522 (1988).

21) H.Ll.D. Pugh, Mechanical Behaviour of Materials Under Pressure, Elsevier Science Publishers, Amsterdam, (1970). 22) J.G. Williams, Stress Analysis of Polymers, Ellis Horwood, Chichester, (1980).

TABLE 1

Schedule of Temperature - Pressure to Produce Plates and Billets of 100% PE

ω a. ca

< u H

3 X a

TABLE 3

Simple Beam Theory Formulae for Three Point Bending

Were FM = flexural modulus; FS = flexural strength; W = load; δ = deflection; L = gauge length; d = sample thickness; b = sample width;

D = sample diameter.

TABLE 4

Maximum SD's for Die Swell and Flexural Tests (in per cent)

TABI E 5 Extrusion Characteristics and fiexural Properties of HA/PE Composite

CO

C. CD CΛ.

CO

IE m

DO cz r~ m

Λ35