9

Polymers for biodegradable medical devices VIII, Hydroxybutyrate-hydroxyvalerate copolymers: physical and degradative properties of blends with polycaprolactone

M. Yasin and B.J. Tighe Speciality Polymers Group, Department of Chemical Engineering and Applied Chemistry, Aston University, Aston Triangle, Birmingham 84 7ET, UK

The physical and degradative properties of polyhydroxybutyrate-hydroxyvalerate copolymer blends with polycaprolactone were investigated. Blends containing low levels of polycaprolactone (< 20%) were found to possess a considerable degree of compatibility, whilst those with higher levels of polycaprolactone were incompatible and showed phase separation behaviour. This incompatibility was most marked in blends containing approximately 50% of each component. In blends containing low levels of polycaprolactone, processing conditions governed the ease of crystallization of polycaprolactone in the polyhydroxybutyrate-hydroxyvalerate matrix and thus the mechanical property of the blend. The degradation rate of these blends was found to be influenced by a complex set of factors, including temperature, pH and polycaprolactone content of the blend. Although crystallinity affected the mechanical properties of the blends, its influence on the hydrolytic degradation rate was masked by the large difference in the molecular weight of the polyhydroxybutyrate-hydroxyvalerate copolymers (M, - 300 000) and polycaprolactone

(M, - 50 000). The polyhydroxybutyrate-hydroxyvalerate/polycaprolactone blends were found to be much more stable to hydrolytic degradation than polyhydroxybutyrate-hydroxyvalerate/ polysaccharide blends previously studied. Here the combined techniques of goniophotometry and surface energy measurements proved extremely valuable in monitoring the early stages of degradation, during which surface, rather than bulk degradation, processes predominate.

Keywords: Hydroxybutyrate-hydroxyvalerate copolymer, polycaprolactone blends, hydrolytic degradation, goniophotometry

Received 2 October 1990; revised 17 October 1990; accepted 21 November 1990

The basic mode of operation for surgical fixation devices such as clips and staples requires the device to have a certain degree of strength and the ability to undergo a degree of flexing without fracture. These property requirements are difficult to correlate with tensile parameters obtained from simple stress-strain curves, but some comparison can be made on the basis of initial tensile moduli which reflect the stiffness of different candidate materials. Although physical properties of polyhydroxybutyrate (PHB) can be varied by hydroxy- valerate (PHV) units, the extent to which these may be modified is limited. The use of blending techniques to increase the versatility of these materials is potentially valuable. We have previously describedIe the incor- poration of polysaccharides, primarily as a means of influencing the degradation profiles. Such blends are heterogeneous, with the polysaccharide acting as an

Correspondence to Dr M. Yasin.

incompatible filler in the PHB-PHV matrix. It would be both interesting and useful, however, to produce a range of degradable materials with a wide spectrum of physical properties by the investigation of compatible blends.

Although a common method of increasing the process- ability and lowering of the initial tensile modulus (i.e. a decrease in stiffness) of a polymer is to use plasticizers during processing operations, for biomedical applications use of plasticizers presents difficulties, primarily because of potential toxicity of substances normally used in this way. Polymers can in principle be used as polymeric plasticizers to change the physical properties. One potentially interesting polymer is the polyester poly- caprolactone, (PCL), which in recent years has been widely exploited for its unique ability to blend with many polymers over wide composition ranges5v6.

PCL is a crystalline, tough and flexible polymer5 and because of its low glass transition temperature ( Tg) - 60 to -70°C and an ability to increase molecular mobility of

0 1992 Butterworth-Heil 92/010009/08

rlemann Ltd Biomaterials 1992, Vol. 13 No. 1

10 Polymers for biodegradable medical devices: M. Yasin and 6.J. Tighe

the polymer chains, PCL resin has been used as a polymeric plasticizer7. PCL is difficult to melt-process by itself, because of its low melting temperature (T,) of 63’C. However, because of its ease of blending and added advantage of being biodegradable’-“, PCL blends with the PHB-PHV copolymers are potentially interesting, particularly for medical applications. Although various copolymers of c-caprolactone with DL-lactide and glycolide have been described*1-*7, physical blends of PCL and the PHB-PHV copolymer series have not been reported.

In this paper, the use of blending techniques to modify the properties of PHB-PHV copolymers by incorporation of synthetic biodegradable polymers, typified by PCL, is described. The main theme of the paper is, however, the development of novel biomaterials, rather than the investigation of degradation mechanism pathways.

EXPERIMENTAL

Materials

PHB-PHV copolymers of 12 and 20% hydroxyvalerate content and weight average molecular weights (M,) of 350 000 and 300 000, respectively, were obtained from Marlborough Biopolymers, ICI, Teesside, UK, in hydroxyapatite (1% w/w) nucleated ‘All-Bran’ form. PCL was supplied by Aldrich Chemical Co., Inc., USA, in pellet form, (A&., of 54 500 and number average molecular weight (M,) of 20 500).

Preparation of samples

Solvent cast films Films of PHB-PHV/PCL blends were prepared by conventional solvent-casting techniques, from 10% (w/v) solutions of the polymers, using both methylene chloride and a mixture of methylene chloride:tetrahydrofuran. Glass Petri dishes were used as the casting substrate and a range of blends containing O-100% PCL were prepared at room temperature.

Injection-moulded samples The PHB-PHV copolymers were melt blended with PCL using a heated two-roll mill (Bridge Micromill, David Bridge Co., Ltd, Castleton, Rochdale, UK], for 5 min at 140°C and 126°C for the 12% PHV and 20% PHV PHB- PHV copolymers, respectively. Loadings of O-100% w/w PCL blends were prepared in a batch size of 40 g. The injection moulding of these PHB-PHV/PCL blends was performed on a electrically heated bench-top injection moulder (SP 1, Small Power Machines Co., Ltd, Chippenham, UK). The PHB-PHV/PCL blends were injection moulded at 150°C for the 12% PHV blends and 140°C for the 20% PHV blends using a preheated mould [- 5O’C). Pure PCL was injection moulded at 150°C, using a warm mould [- 25°C). Samples were moulded into a dumb-bell-shaped test piece with cross-sectional dimensions 4 X 1 mm.

Hydrolytic degradation studies

Degradation of the PHB-PHV/PCL blends was carried out in three aqueous buffer solutions’. Samples were

placed in small screw-capped bottles containing 50 ml of buffer and maintained at 37 f O.l’C or 50 + O.l’C in water baths. The samples were removed periodically, washed with distilled water and placed between filter paper to remove the surface water. The samples were weighed wet and again after drying in a vacuum oven at 5o°C for 2 d.

Degradation monitoring techniques

Degradation of the PHB-PHV/PCL blends was followed by an array of techniques, including gravimetry [water uptake and dry weight loss), goniophotometric analysis’, scanning electron microscopy (SEM) and surface energy measurements’. Polymer crystallinity was monitored by X-ray diffraction’ and density using a density gradient column prepared from carbon tetrachloride and xylene in ratios ranging from 68:32 (v/v) to 16:84 (v/v). Molecular weight was determined by gel permeation chroma- tography (GPC), using chloroform as solvent at the SERC Polymer Supply and Characterisation Centre, RAPRA, Shawbury, UK. Tensile properties of the moulded samples were monitored, after removal from degradation media, washing with distilled water and drying with filter paper at a cross-head speed of 5 mm/min and gauge length of 10 mm using an Instron tensometer [Model TM- M, Instron Ltd, High Wycombe, UK), at 20°C.

RESULTS AND DISCUSSION

Initial compatibility studies

The films of the PHB-PHV/PCL blends, cast from methylene chloride and methylene chloride:tetrahydro- furan mixture, dried to a range of optical clarities. Those with up to 10% PCL in the blend were clear and with increasing PCL content, the films became more opaque. Although it appears that PHB-PHV/PCL blends con- taining low levels of PCL are compatible, optical clarity itself is not a true test for polymer-polymer compatibility. The haziness of the films containing higher levels of PCL was possibly due to the crystallization of the PCL component in the blend. Indeed, quenching experiments involving the warming up of films containing > 10% PCL to 126°C and rapid quenching in liquid nitrogen resulted in previously opaque films becoming completely transparent.

SEM studies showed no phase separation between the PHB-PHV and PCL components in blends with low PCL content, suggesting that some degree of compatibility exists in these blends. Optical microscopy showed that the blends low in PCL contained minute polymer spherulites evenly distributed throughout the film and increasing in size with increased PCL content. In addition, more conventional phase separation occurred in blends containing > 10% PCL. In a typical case, a film of the PHB-PHV/PCL blend containing 20% PCL, under observation in a polarizing microscope, showed a large number of very small spherulites, evenly distributed. On heating the film on a hot stage to 130’C, the crystalline structures formed a melt. On cooling, spherulites recrystallized from the melt and began to grow radially out from the centre. The spherulitic growth was stopped only by the interaction of neighbouring spherulites.

Biomaterials 1992, Vol. 13 No. 1

Polymers for biodegradable medical devices: M. Yasin and B.J. Tighe 11

In these observations, the PHB-PHV/PCL blends containing 10% PCL seemed to show many of the characteristics of a compatible system. Other blends in the range investigated showed inferior mechanical properties to either of the component materials and were obviously incompatible. Melt-processing studies were undertaken on a range of compositions, but work was concentrated on the PHB-PHV/PCL (9O:lO) together with the PHB-PHV/PCL (10:90) blend for purposes of comparison.

Melt processing and physical properties

Although it was possible to injection mould PHB-PHV/ PCL blends, without using nucleating agents, con- siderable time was needed to allow the moulding to crystallize before the mouldings could be removed from the mould. To alleviate these problems and to produce consistent injection mouldings, PHB-PHV copolymers containing 1% w/w hydroxyapatite were used to prepare the PHB-PHV/PCL blends.

Tensile tests were carried out on mini-dumb-bell injection-moulded test pieces of the PHB-PHV/PCL blends prepared with loadings of 10 and 90% w/w PCL in the blend, using 12% PHV and 20% PHV PHB-PHV copolymers (nucleated with 1% w/w hydroxyapatite). The resultant tensile parameters are given in Table I. The effect on mechanical properties produced by the intro- duction of a small amount of PCL into the matrix of PHB-PHV copolymers is very similar for both the 12 and 20% PHV PHB-PHV copolymers. The addition of 10% PCL reduces both the yield strength and the initial tensile modulus, with a concurrent small increase in the percentage elongation at break, in comparison to un- blended PHB-PHV copolymers.

With increasing PCL content in the PHB-PHV/PCL blends, the properties of the blend would be expected to be more influenced by the PCL component rather than the PHB-PHV component. The properties of the blends containing 90% PCL might therefore be expected to be similar to those of PCL itself. Although this is reflected in comparatively similar values for the yield strength and initial modulus of PCL and blends high in PCL, the ultimate tensile strength and elongation at break of PCL are higher than the corresponding values for the blends containing 90% PCL. This can be explained in terms of polymer crystallinity. PCL is injection moulded at nearly 100°C above its T,, and the resultant cooling under pressure to room temperature ensures that the material is

very crystalline, as evidenced from X-ray dioffractograms. The incorporation of 10% PHB-PHV copolymer into the PCL matrix decreases the crystallinity marginally. Because of the vast differences in the Tg and T,,, of the two components, the PHB-PHV copolymers crystallize out of the melt first, at temperatures that are above the T,,, of PCL. SEMs of the 90% PCL-loaded blends show discrete phases, so it is probable that the blends high in PCL are incompatible, in both technological and thermo- dynamic terms. X-ray diffractograms of PCL and various blends with the PHB-PHV copolymers show that PCL, together with the PHB-PHV/PCL blends high in PCL content have a higher crystallintiy than those with low PCL content.

The role played by PCL in the low PCL-content blends is more complex than one of a simple polymeric plasticizer. In the case of the 10% PCL blend with 20% PHV PHB-PHV copolymer, a dramatic effect was observed which was consistent and reproducible between different batches. In essence, factors which promote the ready crystallization of PCL reduce its effectiveness as a plasticizer. Amongst the factors that influence this are the initial injection temperature and the temperature of the mould. Because the PHV content of the PHB-PHV copolymers alters its melting point, this is reflected in the injection temperature used to mould the copolymer. In particular, a 10°C reduction in injection temperature was involved in moving from a 12% PHV to a 20% PHV copolymer. Additionally, it was difficult to maintain the mould temperature at a precise level for each batch of mouldings, especially when differences in barrel temperature were involved. Since the mould temperature (- 50°C) was in the region of the melting point of PCL, it was found possible to promote or inhibit crystallization of PCL with the combination of moulding conditions used. The dramatic difference between conditions under which PCL crystallizes readily and with difficulty are illustrated in the tensile results for the 10% PCL blends with the 12 and 20% PHV PHB-PHV copolymers, respectively (Table I). It should be noted however, as subsequent results show, annealing these samples (whether in aqueous media or in an air oven), produces crystallization of the PCL.

Accelerated hydrolytic degradation

Owing to the low melting point of PCL, 50°C was chosen as the upper limit for accelerated degradation of these blends. The weight loss profiles at 50’C for these blends

Table 1 Tensile parameters of injection-moulded samples of 12% and 20% PHV PHB-PHV copolymers and their blends with PCL.

Blend YS (MPa) %EY ITM (MPa) UTS (MPa) %EB

12% PHV PHB-PHV copolymer 28.7 k 1.8 12 k 2 500 + 30 27.1 f 1.8 15k2 20% PHV PHB-PHV copolymer PCL 19.1 IO.2 13 i 2 310 zk 10 17.3 + 0.9 22 + 2 PCL 16.3 f 1.2 18 k 6 150 + 20 20.1 * 1.2 400 + 60 12% PHV, PHB-PHV/PCL (9O:lO) 19.9 f 0.1 13* 2 300 + 30 19.9 Ik 0.1 13+2 12% PHV, PHB-PHV/PCL (10:90) 15.0 + 0.8 43 * 5 150 * 10 15.0 + 0.8 43 i- 5 20% PHV, PHB-PHV/PCL (9O:lO) 18.0 + 1.2 18 f 4 300 ?I 40 19.2 f 0.7 400 + 50 20% PHV, PHB-PHV/PCL (10:90) 15.3 + 0.3 31 f2 150 + 10 15.3 z!I 0.3 31 +2

YS = yield strength (MPa). %EY = % elongation at yield. ITM = initial tensile modulus (MPa).

UTM = ultimate tensile modulus (MPa) %EB = % elongation at break.

Biomaterials 1992. Vol. 13 No. 1

12 Polymers for biodegradable medical devices: M. Yasin and B.J. Tighe

Table 2 Time for 10% (t,,J weight loss of injection-moulded samples of PHB-PHV/PCL blends maintained in pH 10.6, 7.4 and 2.3 buffer at 50°C.

Blend pH 2.3

12% PHV PHB-PHV/PCL blends

pH 7.4 pH 10.6

12% PHV copolymer 250 PHB-PHV/PCL (9O:lO) 250 PHB-PHV/PCL (50:50) 220 PHB-PHV/PCL (33:67) 230 PHB-PHV/PCL (10:90) 240 PCL 210

20% PHV PHB-PHV/PCL blends

240 200 240 160 190 100 210 140 220 160 180 150

20% PHV copolymer 230 220 150 PHB-PHV/PCL (9O:lO) 230 210 140 PHB-PHV/PCL (50:50) 210 180 90

PHB-PHVIPCL (io:90) 250 590 Tlo PCL 210 180 150

are characterized by a initial slow rate, followed by a secondary enhanced phase. The t,, values (time taken for 10% dry weight loss), for the PCL blends with the PHB- PHV copolymers (Table Z), illustrate that PCL degrades faster than either of the PHB-PHV copolymers, probably because of its lower molecular weight. The degradation of the PHB-PHV/PCL blends at 50°C can be generalized to be fastest under alkaline conditions and slowest under acidic ones. Degradation under alkaline conditions is about 2-3 times faster than under acidic conditions.

A closer examination of the weight loss profiles of the PHB-PHV/PCL blends, typified in Figure 1, shows the presence of a compositonal dependence reflecting some differences in compatibility and crystallization behaviour. The trends for the 12% PHV and 20% PHV PHB-PHV copolymers are broadly similar. In both cases the blends with approximately equal proportions of PHB-PHV and PCL components show lowest stability. The detailed behaviour of other blends in this series shows some variability, but is consistent with the view that high levels of incompatibility and independent crystallization of the two components leads to reduced hydrolytic stability, largely due to the heterogeneous behaviour of these blends. It has already been noted that small differences in processing conditions can have noticeable effects on the crystallization phenomena. To this extent, some inherent slight variability in the behaviour of these

Table 3 Surface energy parameters of undegraded injection-moulded samples of unfilled 20% PHV PHB-PHV copolymer, PCL and their blends.

samples is to be expected. In general, the behaviour of samples moulded under slightly different conditions becomes more uniform with progressive exposure to the aqueous environment (at 50’C) due to the annealing of the samples.

A brief comparative study of the hydrolytic degradation of the solvent-cast films of the PHB-PHV/PCL blends at 50°C, in pH 7.4 buffer was carried out. The rate of degradation of these solvent-cast films was found to be faster than the injection-moulded samples over a similar timescale. This again is due to the differences in the crystallinity and polymer matrix compaction, induced by the two methods of fabrication. This difference parallels that observed with the unblended PHB-PHV copolymers* and is reflected in the measured differences in density of these samples.

Surface energy values by contact angle measurements (Table 3) of the injection-moulded samples of the PHB- PHV/PCL blends indicate a decrease in the polar com- ponent of the surface energy with increasing PCL content. This suggests that the surfaces of the PHB- PHV/PCL blends with high PCL content will be the most hydrophobic. This increased hydrophobicity will reduce the wettability and ease of hydrolytic degradation of the surface, if other factors remain constant. However, it is apparent from the weight loss data for these blends in the

100

3 90

2 9 : 80

m ‘Z ._ c - 70

60 I I I I ! L

0 50 100 150 200 250

Time (days)

Figure 1 Progressive weight loss of injection-moulded 12% PHV PHB-PHV/PCL blends maintained in an aqueous environ- ment at 50°C and pH 10.6. (0) 12% PHV PHB-PHV copolymer, (A) PHB-PHV/PCL (9O:lO) blend, (+) PHB-PHV/PCL (50:50) blend, (0) PHB-PHVIPCL (33:67) blend, (X) PHB-PHV/PCL (10:90) blend, (m) PCL.

Parameter/blend 20% PHV 20% PHV 20% PHV 20% PHV PHB-PHV PHB-PHV/PCL PHB-PHV/PCL PHB-PHV/PCL Copolymer (9O:lO) blend (50:50) blend (10:90) blend

PCL

CA water (degrees) 61.0 64.0 65.0 70.0 71.0 CA Mel (degrees) 35.0 31.0 27.0 26.0 25.0 (vp) (mN/m) 13.4 10.9 9.7 6.9 6.4 (ud) (mN/m) 35.7 38.0 40.0 41.3 41.9 (u’) (mN/m) 49.1 48.9 49.7 48.2 48.3

CA = average contact angle. (up) = polar component of surface energy. (ud) = dispersive component of surface energy. (yf) = total surface energy.

Biomaterials 1992, Vol. 13 No. 1

Polymers for biodegradable medical devices: M. Yasin and B.J. Tighe 13 _.....~

various buffers that the contribution of the surface energy to the degradation rate is overridden by the ove~helming combined effects of the lower molecular weight of the PCL and the contributions of crystallinity. Thus, although there is a difference in the polar component in the series of PHB-PHV/PCL blends, the actual difference in the interfacial tension between water and the two polymers (i.e. PHB-PHV and PCL), is too small to have any real effect on the overall wetting of the surfaces.

Physiological hydrolytic degradation

The surface and bulk properties of the PHB-PHV/PCL blends in a pH 7.4 buffer at 37°C were investigated using combined techniques of tensile strength, weight loss, goniophotometry, surface energy and molecular weight. The gravimetric degradation profiles for both the 12 and 20% PHV PHB-PHV/PCL blends illustrate the rather stable nature of these blends to hydrolytic degradation, in comparison to the PHB-PHV/polysaccharide blends

under similar conditions . 3*4 The relatively small change in the dry weight for these blends makes it difficult to predict compositional dependence of the hydrol~ic degradation rate under these conditions accurately. Furthermore, the period of study undertaken may be regarded as the initial phase in the hydrolytic degradation process, during which surface, rather than bulk degrada- tion, processes are predominant.

Although during the initial stages of degradation there is relatively little weight loss, dramatic changes occur in the tensile properties of these blends (Table 4). There is a steady increase in both the yield strength and the initial tensile modulus on immersion in pH 7.4 buffer at 37”C, for PCL and for both the 12 and 20% PHV PHB-PHVI PCL blends. The mode of fracture of the blends changes from an initial ductile mode to one of brittle fracture during the first few days of immersion, due to an increase in the crystallinity of the samples, brought about by the annealing process. This change in crystallinity was monitored by X-ray diffraction studies and is illustrated inFigure 2, which shows the changes in the diffractogram

Table 4 Tensile parameters of injection-mot&fed samples of 12% and 20% PHV PHB-PHV copolymers and their blends with PCL maintained in an aqueous pH 7.4 buffer at 37°C.

Time (days) YS (MPaf

PCL 0 16.3 rt 1.2 7 18.5 + 1.1

22 15.4 L 0.8 62 16.5 It: 0.9

112 15.4 t 0.7 180 4.3 * 0.7

12% PHV PHB-PHVIPCL (9O:iO) blend 0 19.9 + 0.1 7 19.5 f 0.5

22 15.4 f 0.8 62 15.5 f 0.5

112 15.2 f 0.1 180 12.8 rt 0.2 250 10.4 * 0.1

12% PHV PHB-PHVIPCL (10:90) blend 0 15.0 I 0.8 7 14.4 f 0.3

22 13.8 + 0.2 62 10.0 -f 0.1

112 9.4 i 0.2 180 3.7 rfr 0.3

20% PHV PHB-PHVIPCL (9O:lO) blend 0 18.0 z!z 1.2 7 21.1 + 1.8

22 17.9 Ifr 3.1 62 21.0 i 1.0

112 20.1 + 1.0 180 17.5 I 0.8 250 15.6 -t 0.5

20% PHV PHB-PHV/PCL (10:90) blend 0 15.3 f 0.3 7 15.6 & 0.2

22 14.2 f 0.4 62 12.3 t 0.3

112 7.3 i 0.1 180 5.9 + 0.2

%E at Y

18 + 6 32+ 10 12+2 10+2 11,2

5+2

13+2 11+2 13f2

7+2 6+2 6+2 6+2

43 + 5 21+-2 21 +2

5t2 6+2 4+2

18?4 16+2 25 f 6 11+4 13 t 2 11+2 17t4

31+2 14 f 2 23 + 2 12 + 2

5*2 652

ITM (MPa)

150 If- 20 180 f 10 190 f 10 210 f 10 200 f 20 110 i 30

300 2 30 500 f 10 270 + 50 500 f 10 380 f 50 430 f 10 310 + 50

150 ?I 10 170 + 20 140 f 20 210 f 10 220 * 40

80t 10

300 * 40 250 t 20 220 * 10 300 f 10 270 + 10 300 t 30 230 + 10

150 + 10 190 * 20 160 f 10 220 + 10 200 i: 10 130 t 30

UTS [MPa) %E at break

20.1 -t 1.2 400 + 60 18.5 rt 1.1 32f 10 15.4 ?1 0.8 12+2 16.5 It 0.9 lo+ 2 15.4 * 0.7 11+2 4.3 + 0.7 5f2

19.9 rtr 0.1 13 + 2 19.5 ?I 0.5 11+2 15.4 f 0.8 13+2 15.5 f 0.5 7+2 15.2 ?z 0.1 6+2 12.8 t: 0.2 6+2 10.4 * 0.1 622

15.0 k 0.8 43 f 5 14.4 z!I 0.3 21 +2 13.8 rt 0.2 21 * 2 10.0 I: 0.1 522

9.4 f 0.2 6+2 3.7 k 0.3 4f2

19.2 + 0.7 400 f 50 15.9 rt 0.8 36 I? 6 15.9 I!I 1.0 66+ 15 19.5 t 0.9 48 + 8 17.3 * 1.1 24 f 4 16.3 i 0.8 18k 3 15.2 ir 0.2 21 f5

15.3 rt 0.3 31 ?I 2 15.6 rt 0.2 14 f 2 14.2 i: 0.4 23 2 2 12.3 i 0.3 12 +_ 2

7.3 _+ 0.1 522 5.9 + 0.2 6f2

YS = yield strength (MPa). %EY = % elongation at yield. ITM = initial tensile modulus (MPa).

UTM = ultimate tensile modulus (MPa) %EB = % elongation at break.

Riomaterials 1992, Vol. 13 No. 1

14 Polymers for biodegradable medical devices: M. &sin and B.J. Tighe

7 days

0 day

SO 40 30

ZB(degrees.1

20 10

Figure 2 X-ray diffraction traces of 20% PHV PHB-PHV/PCL (9O:lO) blend after various periods of immersion in pH 7.4 buffer at 37°C.

traces of the 20% PHV PHB-PHVIPCL (9020) blend from 0 to 22 d degradation, With increasing incubation time, the initial tensile modulus increases, with a concurrent drop in the ultimate tensile strength and percentage elongation at both yield and at break. Pitt et af.% 12*13 suggest that the increase in c~stallinity with incubation for PCL may be associated with chain cleavage in the amorphous phase, and crystallization of the resulting unstrained tie segments, facilitated by the low T8 (-60 to -7OOC) of PCL. A similar explanation may be used to account for the changes in the crystallinity of the PHB-PHV/PCL blends. During the later stages of degradation, as the molecular weight decreases, there is progressive deterioration in the tensile properties.

During the initial stages of degradation, where surface rather than bulk degradation occurs, gravimetric tech- niques are of limited use. We have previously described the valuable and convenient techniques of goniophoto- metric analysis and surface energy measurements that may be used to monitor surface change& 3* *. AIthough SEM provides a very valuable and detailed morphological investigation of degraded surfaces, it has the inherent disadvantage of sample destruction. However, gonio- photometry and surface energy measurements (via contact angle measurements), are non-destructive and complimentary techniques which can be used to provide valuable information on the progress of surface erosion of a sample over a period of time.

Figure 3 shows the comparative changes in the gonio- photometric and surface energy parameters for PCL over 180 d.With < 2% weight loss, dramatic changes in the specular angle (I,) and gloss factor occur. The I, and gloss factor have decreased, with a concurrent increase in the diffuse reflectance (Id) value over the time period of O-180 d. Similarly, there is a considerable increase in the polar component and concurrent decrease in the dispersive component of surface energy over the same time period.

Figure 4 shows the changes in the goniophotometric

ti x-x-x X x- 0

I I

0 50 100 150 200

Time (days)

Figure 3 Combined surface energy and goniophotometric data for inj~tion-moulded PCL samples after immersion in an aqueous pH 7.4 buffer at 37°C over 200 d. (A) polar com- ponent of surface free energy, (El) dispersive component of surface free energy, ( n ) specular intensity (I,), (X) diffuse reflectance (Id), (+) Gloss factor.

X-X 1 x , x 0 I

X-

0 50 100 150 200

Time (days)

Figure 4 Combined surface energy and goniophotometric data for injection-moulded 20% PHV PHB-PHV/PCL (9O:lO) blend after immersion in an aqueouspH 7.4 buffer at 37°C over 200 d. (A) polar component of surface free energy, (Cl) dispersive component of surface free energy, (W) specular intensity (I,), (X) diffuse reflectance (fd), (+) Gloss factor.

and surface energy parameters for the 20% PHV PHB- PHV/PCL (9O:lO) blend over 180 d. Even after 22 d immersion, although weight loss is negligible, dramatic changes in the I, and polar component of surface energy are apparent. Figure 5 illustrates the goniophotometric and surface energy parameters of the corresponding high PCL (go%,) blend over the same time period. The increase in the surface energy is due to an increase in the hydroxyl and carboxyl groups on the surface, brought about by surface hydrolysis. Comparison of Figure 4 with Figure 5 indicates that the blend high in PCL content degrades faster than the blend containing lower levels (1070) of PCL.

CONCLUSION

From the gravimetry studies, the changes to the wet weight of these blends illustrate that, unlike PHB-PHV

Biomaterials 1992, Vol. 13 No. 1

Polymers for biodegradable medical devices: M. Yasin and f3.J. Tighe 15

x---x-x X X-

0 ! 1 I c 0 50 100 150 200

Time [days)

Figure 5 Combined surface energy and goniophotometric data for injection-moulded 20% PHV PHB-PHV/PCL (10:90) blend after immersion in an aqueous pH 7.4 buffer at 37°C over 200 d. (A) polar component of surface free energy, (El) dispersive component of surface free energy, (a) specular intensity (I,), (X) diffuse reflectance (Id), (+) Gloss factor.

copolymer blends with polysaccharides3’4, they take up much less aqueous buffer. This is expected because PCL can be considered to be a more hydrophobic component, in comparison to the polysaccharides and hence less likely to absorb water (buffer). The consequences of this on the hydrolytic degradation rate of these become apparent during the early stages of degradation, where little weight loss occurs,

Although X-ray diffraction studies indicate that the 20% PHV PHB-PHV/PCL (9O:lO) blend has a lower crystallinity than the corresponding blend containing 9070 PCL (68 k 5% and 86 t 5%, respectively), the former blend is hydrolytically more stable than the latter. This is possibly because the effect of a small difference in the crystallinity of the two blends is outweighed by the contributions from the high molecular weight of the PHB-PHV copolymer component in the blend. Although crystallinity plays an important role in the physical properties of these blends, its contributions to the hydrolytic degradation rate are overshadowed by the vast differences in the molecular weight of the two component polymers. The molecular weight of the PHB- PHV copolymers is some four times greater than that of the PCL, and as a result although the two blends mentioned here have differing degrees of crystallinities, the hydrolytic degradation rate is governed by the level of the two component polymers in the blend, which becomes apparent during later stages of degradation. The IZ% PHV PHB-PHV/PCL blends behaved similarly.

Surface features are good indicators for the onset of hydrolytic polymer degradation, where gravimetric methods are of limited use. The vastly superior changes in experimental values of surface parameter values (gloss factor from goniophotometry and the polar component of the surface energy from contact angle measurements) compared to simple weight loss, have been shown to be more reliable in detecting changes during the early stages of degradation for the PHB-PHV copolymers and their blends with PCL. During these early stages, surface rather than bulk degradation processes predominate.

Less rapid changes occur in the bulk properties of

PHB-PHV matrices such as crystallinity and tensile strength when compared to the surface measurements alone. However, some significant changes do occur, particularly at later stages of experimental monitoring. Here, molecular weight does fall as expected and usefully supplements results obtained by gravimetry. Although tensile strength is relatively unimportant in the deter- mination of the hydrolytic degradation process, it does show that PCL-filled PHB-PHV blends have inherently tess mechanical strengths than the unfilled copolymers and illustrates transitions in blend behaviour related to quenching or annealing the PCL component.

The least stable blends again seem to be those containing around 50% of each component. A possible explanation as to the faster degradation rate of the 50150 composition blends, in comparison to the other com- positions, is that at this particular loading the PHB-PHV and the PCL components show maximum incompatibility. This incompatibility and increased porosity, combined with possibly a low level of crystallinity, increases the degradation rate of the blend. This again indicates the complexity of the factors affecting the hydrolytic degradation of this series of blends.

These blends have a rather lower initial tensile modulus than the corresponding PHB-PHV/poly- saccharide system4. Surgical fixation devices (e.g. clips, staples), which provide support in the early stages of wound healing for tissues, need a certain degree of flexibility to facilitate proper mechanical function of the device. Combination of PCL and polysaccharide com- ponents, at low loadings (- 10%) with the high valerate PHB-PHV copolymers (20% PHV), should produce blends possessing excellent physical properties with reasonable hydrolytic degradation rates which may find application in these areas. PHB-PHV/PCL blends could also be used for the fabrication of medical devices, where neither the physical or degradative properties are important, but rather the matrix porosity is important, such as drug delivery systems.

In summary, a wide range of melt-p~cessable, biodegradable materials have been produced, the physical and degradative properties of which can be altered by varying the amounts of the component polymers used, and which have a complex compositional dependence. Certainly, these blends are a promising group of materials, the degradative and physical properties of which need to be fully studied. Our current activities involve the use of high molecular weight PCL with a narrower molecular weight distribution. The properties of this interesting group of materials will be described in further publica- tions.

ACKNOWLEDGEMENT

We would like to acknowledge Suresh Nair for the work on the initial compatibility of the PHB-PHV/PCL blends, which he did as part of his Final Year project.

REFERENCES

1 Holland, S.J., Jolly, A.M., Yasin, M. and Tighe, B.J., Polymers for biodegradable medical devices. II. Hydroxybutyrate-hydrox~alerate copolymers: hydro-

Biomaterials 1992. Vol. 13 No. 1

16 Polymers for biodegradable medical devices: hf. Yasin and B.J. Tjghe

lytic degradation studies, Biomaterials 1987, 8, 289-295 Yasin, M., Holland, S.J. and Tighe, B.J., Polymers for biodegradable medical devices. V. Hydroxybutyrate- hydroxyvalerate copolymers: effects of processing on hydrolytic degradation rates, Homaterials 1990, II, 451-454 Yasin, M., Holland, S.J., Jolly, A.M. and Tighe, B.J., Polymers for biodegradable medical devices. VI. Hydroxybutyrate-hydroxyvalerate copolymers: accelerated degradation of blends with polysaccharides, Biomaterials 1989, 10, 400-412 Holland, S.J., Yasin, M., and Tighe, B.J., Polymers for biodegradable medical devices. VII. Hydroxybutyrate- hydroxyvalerate copolymers: degradation of copolymers and their blends with polysaccharides under in vitro physiological conditions, Biomaterials 1990, 11, 206- 215 Brode, G.L. and Koleske, J.V., Lactone polymerisation and polymer properties, I. Macromol. Sci. Chem. 1972, A6(6), 1109-1144 Koleske, J.V., Blends containing poly(c-caprolactone) and related polymers, in Polymer Blends (Eds D.R. Paul and S. Newman), Vol. 2, Academic Press, New York, USA, 1978, pp 369-389 Hammer, CF., Polymeric plasticisers, in PolymerBlends (Eds D.R. Paul and S. Newman], Vol. 2, Academic Press, New York, IJSA, 1978, pp 219-241 Pitt, C.G., Chasalow, F.I., Hibionada, Y.M., Klimas, D.M. and Schindler, A., Aliphatic polyesters. 1. The degradation of poly(e-caprolactone) in viva,]. Appl. Polym. Sci. 1981, 26, 3779-3767 Woodward, SC., Brewer, P.S., Moatamed, F., Schindler, A. and Pitt, C.G., The intracellular degradation of poly(s-

10

11

12

13

14

15

16

17

caprolactone), J, Biomed. Mater. Res. 1985,19,437-444 Fields, R.D., Rodriquez, F. and Finn, R.K., Microbial degradation of polyesters: Polycaprolactone degraded by P. Pullans, J. Appl. Polym. Sci. 1974, 18, 3571-3579 Wise, D.I. and Gregory, J.B., A new composite material for fixation of orthopedic surgical implants, in Bio- polymeric Controlled Release Systems [Ed. D.L. Wise), Vol. 2, CRC Press, Boca Raton, Florida, USA, 1984, pp 231-238 Pitt, C.G., Gratzl, M.M., Kimmel, G.L., Surles, J. and Schindler, A., Aliphatic polyesters. II. The degradation of poly@L-Lactide), poly(e-caprolactone) and their co- polymers in viva, Biomaterials 1981, 2, 215-220 Schindler, A., Jeffcoat, R., Pitt, C.G., Wall, M.E. and Zweidinger, R., Biodegradable polymers for sustained drug delivery, Contemp. Top. Polym. Sci. 1987, 2, 251-289 Gilbert, R.D., Stannett, V., Pitt, D.G. and Schindler, A., The design of biodegradable polymers: Two approaches, in Developments in Polymer Degradation [Ed. N. Grassie), Vol. 4, Applied Science Publishers, London, UK, 1982, pp 259-293 Kricheldorf, H.R., Mang, T. and Jonte, J.M., Poly- lactones. 1. copolymerisation of glycolide and &-caprolactone, Macromolecules 1984, 17, 2173-2181 Shalaby, SW. and Jamiolkowski, D.D., Synthesis and intrinsic properties of crystalline copolymers of c-caprolactone and glycolide, PolymerPreprints 1985,26 (21, 90 Shalaby, SW. and Jamiolkowski, D.D., Absorbable fibres of c-caprolactone/glycolide copolymers and their biological properties, Polymer Preprints 1965, 26 (21, 200

Biomaterials 1992. Vol. 13 No. 1