journal of the mechanical behavior of biomedical materials 50 (2015) 277–289
Available online at www.sciencedirect.com
www.elsevier.com/locate/jmbbm
Research Paper
A comparative study of the effects of different bioactive fillers in PLGA matrix composites and their suitability as bone substitute materials: A thermo-mechanical and in vitro investigation R.L. Simpsona, S.N. Nazhatb, J.J. Blakerc,n, A. Bismarckd, R. Hille, A.R. Boccaccinif,g,n, U.N. Hansena, A.A. Amisa,h a
Department of Mechanical Engineering, Imperial College London, London SW7 2AZ, UK Department of Mining and Materials Engineering, McGill University, 3610 University Street, Montreal, Quebec H3A 0C5, Canada c Materials Science Centre, School of Materials, University of Manchester, Grovesnor Street, Manchester M1 7HS, UK d Department of Chemical Engineering, Polymer and Composite Engineering (PaCE) Group, Imperial College London, London SW7 2AZ, UK e Institute of Dentistry, Dental Physical Science Unit, Queen Mary, University of London, London E1 4NS, UK f Department of Materials, Imperial College London, London SW7 2BP, UK g Institute of Biomaterials, Department of Materials Science and Engineering, University of Erlangen-Nuremburg, 91058 Erlangen, Germany h Musculoskeletal Surgery Group, Imperial College School of Medicine, London W6 8RF, UK b
art i cle i nfo
ab st rac t
Article history:
Bone substitute composite materials with poly(L-lactide-co-glycolide) (PLGA) matrices and
Received 20 February 2015
four different bioactive fillers: CaCO3, hydroxyapatite (HA), 45S5 Bioglasss (45S5 BG), and
Received in revised form
ICIE4 bioactive glass (a lower sodium glass than 45S5 BG) were produced via melt blending,
5 June 2015
extrusion and moulding. The viscoelastic, mechanical and thermal properties, and the
Accepted 5 June 2015
molecular weight of the matrix were measured. Thermogravimetric analysis evaluated the
Available online 16 June 2015
effect of filler composition on the thermal degradation of the matrix. Bioactive glasses
Keywords:
caused premature degradation of the matrix during processing, whereas CaCO3 or HA did
Bioactive glass
not. All composites, except those with 45S5 BG, had similar mechanical strength and were
Hydroxyapatite
stiffer than PLGA alone in compression, whilst all had a lower tensile strength. Dynamic
Poly(α-hydroxyester) Bone substitute Composite
mechanical analysis demonstrated an increased storage modulus (E0 ) in the composites (other than the 45S5 BG filled PLGA). The effect of water uptake and early degradation was investigated by short-term in vitro aging in simulated body fluid, which indicated enhanced
n Corresponding authors at: School of Materials, University of Manchester (J.J. Blaker) and Institute of Biomaterials, University of Erlangen-Nuremberg (A.R. Boccaccini). E-mail addresses: [email protected] (J.J. Blaker), [email protected] (A.R. Boccaccini).
http://dx.doi.org/10.1016/j.jmbbm.2015.06.008 1751-6161/& 2015 Elsevier Ltd. All rights reserved.
278
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
water uptake over the neat polymer; bioactive glass had the greatest water uptake, causing matrix plasticization. These results enable a direct comparison between bioactive filler type in poly(α-hydroxyester) composites, and have implications when selecting a composite material for eventual application in bone substitution. & 2015 Elsevier Ltd. All rights reserved.
1.
Introduction
Resorbable polymers and ceramics have been developed for orthopaedic procedures, which may encourage bone growth, guiding and supporting regenerating bone for sufficient time such that it is capable of self-support (Cao and Hench, 1996). Ultimately these implanted materials degrade harmlessly in the body (Middleton and Tipton, 2000; Domb et al., 1997; Hench, 1991). However, neither polymers or ceramics, taken separately, have proven entirely suitable for use in bone replacement: polymers may have insufficient mechanical properties (Engelberg and Kohn, 1991), while ceramics often have an unpredictable, low tensile strength so they are unsuitable for use in locations of significant torsion or bending (Bronzino, 1995). Composites have been studied in efforts to obtain a superior bone substitute material, usually polymer matrices with particulate ceramic/glass fillers (Wang, 2003; Rezwan et al., 2006; Boccaccini and Blaker, 2005). These should result in increased mechanical properties over the polymer and impart bioactivity (Boccaccini et al.,
2002). Basic ion releasing ceramics/glasses have a pH buffering effect in the early stages of degradation (Kikuchi et al., 2002; Tsunoda, 2003; Schiller and Epple, 2003), which could reduce the degradation rate of hydrolytically degradable polymers, such as poly(lactide-co-glycolide) (PLGA) and polylactide (PLA) (Blaker et al., 2011). Whilst numerous composites have been investigated as bone analogues, the optimal composite has yet to be developed due to the complex nature of the intended application (Rezwan et al., 2006). In order to compare different composites accurately, it is necessary to use the same production method and testing regime because differences in these will result in variations in material and mechanical properties. Few examples of this type of comparison have been found (Bleach et al., 2001; Ignatius et al., 2001), making direct correlation difficult. The aim of this study was to compare four different particulate fillers in a PLGA matrix using similar processing routes and testing methods to enable direct comparison of the effects of filler selection on composite properties in the as-made state and during aging in simulated body fluid (SBF).
Fig. 1 – SEM images of (a) CaCO3 (rhombohedral particles), (b) HA with irregular shaped particles, (c) 45S5 BG showing fewer small particles, (d) ICIE4 BG with larger particles dispersed amongst smaller particles.
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
279
Table 1 – Particle size distribution of materials. D10, D50 and D90 ¼10, 50, 90 % (size by volume, obtained using a CILAS 1180 Particle size analyzer, Orléans, France), respectively, of particles were this size or less. (Average PLGA n ¼4, CaCO3 and HA n ¼ 5, 45S5 BG and ICIE4 BG n ¼3). Material
Average D10 particle size [μm]
Average D50 particle size [μm]
Average D90 particle size [μm]
PLGA (powder) CaCO3 HA 45S5 BG ICIE4 BG
73.5 3.6 0.7 7.5 0.7
198.4 16.4 3.8 35.3 5.2
346.2 24.9 14.6 62.1 41.1
Fig. 2 – Processing routes used to compare the effects of different fillers on PLGA.
2.
Materials and methods
2.1.
Materials
Granules of 95:5 poly(L-lactide-co-glycolide) (PLGA) semicrystalline poly(α-hydroxyester) co-polymer with a manufacturer specified inherent viscosity of 2.38 dl/g and melting range 160–170 1C, were used (PURAC, Gorinchem, The Netherlands). To enable subsequent processing, the granules were cryogenically milled to a powder in liquid nitrogen using an Alpine 100UPZII mill (Hosokawa Micron Ltd., Cheshire, UK), then dried in a vacuum oven. The powder was then sealed in moisture-proof sachets and stored in a
desiccator with indicating silica gel and phosphorous pentoxide until use. The particulate fillers were CaCO3 (Sigma-Aldrich Co. Ltd., Dorset, UK), sintered hydroxylapatite CAPTALs S (HA) (Plasma Biotal Ltd., Derbyshire, UK), 45S5 Bioglasss (45S5 BG) (of composition in mol%: SiO2 46.13, P2O5 2.61, CaO 26.91, Na2O 24.35); (MoSci Corporation, Montana, USA), and ICIE4 bioactive glass (ICIE4 BG) (of composition in mol%: SiO2 49.46, P2O5 1.07, CaO 42.87, Na2O 6.6) which was fabricated in-house (Elgayar et al., 2005). All fillers were dried at 250 1C in a fan oven for 4 h and stored as described previously for the PLGA powder. Particulate morphology and size can be compared from the scanning electron microscopy images in Fig. 1 and are summarized in Table 1.
280
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
2.2.
Extrusion of neat and composite PLGA material
The generic processing routes are presented schematically in Fig. 2. PLGA was combined with 25 vol% of bioactive filler. This loading was selected as it was expected to retain bioactivity and optimal mechanical properties of the composites while allowing the material to be easily processed. The mixtures were transferred to vials and agitated using a roller mixer (SRT1, Bibby Sterilin Ltd., Staffordshire, UK) for 2.5 h (based on previous trials and observations) to ensure homogeneous particle mixing. A twin-screw 5 cm3 micro-extruder (DSM research products, Maastricht, The Netherlands) blended the composite material at 30 rpm. Extrusion parameters were initially optimized (Table 2). using PLGA filled with CaCO3 (referred to as PLGC hereafter); these parameters were adjusted for the other material combinations: PLGA filled with HA (PLGH), 45S5 BG (PLG4B) or ICIE4 BG (PLGIB) as required to reduce degradation indicated by observable discoloration. The extruded materials were cryogenically ground under liquid nitrogen, to yield a powder of particle size o1 mm. Powders were dried in an oven for 5.5 h at 100 1C and stored as detailed previously, at 4 1C. For the extruded samples, matrix–filler bonding was not assessed as extrusion was used purely to obtain a more homogenous mix.
2.3.
Composite forming
Two methods were used to fabricate test specimens: rudimentary injection moulding and compression moulding (Fig. 2). The injection mould was filled with material and heated to the extrusion temperature given in Table 2, this mould was then transferred to a press (Moores Presses Ltd, Birmingham, UK) and compressed to 9.5 kN, which was maintained until the mould cooled to 30 1C. Compression moulding of flat sheets was conducted using an aluminium mould with the same conditions.
positive controls were PLGA alone and bioactive fillers, respectively. The homogeneity of the bioactive fillers in the composites was determined by comparing residual filler weights (obtained from TGA) with the theoretical weights (given in Table 2). The onset of thermal degradation was defined here as a loss of 40.5% relative to starting mass.
2.4.2.
2.4.3.
Material characterization
2.4.1.
Thermogravimetric analysis (TGA)
Extruded samples were analyzed using TGA (Stanton Redcroft Simultaneous Thermal Analyzer (STA-780), London, UK). Triplicate samples of 15–20 mg were heated in air with a heating regime of ambient temperature to 500 1C to burn off the polymer, with a heating rate of 10 1C min 1 (the maximum temperature had to be increased to 700 1C for the PLGA–45S5 BG (PLG4B) composites as the glass sintered at approximately 450 1C, trapping the polymer). Negative and
Gel permeation chromatography (GPC)
GPC was used to assess the effect of filler type and processing on the molar mass of the polymer. GPC was carried out in duplicate for each sample by RAPRA Technology, Shropshire, UK. Chloroform solvent (10 ml) was added to 20 mg polymer or 30–40 mg composite samples, which were left for a minimum of 4 h to dissolve. Solutions were then filtered through a 0.2 μm polyamide membrane into sample vials, and put into an autosampler. Columns suited to medium/high molecular weight samples were used (Plgel guard plus 2 mixed bed-B, 30 cm, 5–10 μm). The flow rate used was 1.0 ml min 1 (nominal) at 30 1C. A refractive index detector was used. Viscotek Trisec 2000 and Trisec 3.0 software was used to collect and analyze the data, respectively. The equipment was calibrated using narrow distribution polystyrene calibrants (Polymer Laboratories Ltd), to give results as ‘polystyrene equivalent’ molecular weights.
2.4.4. 2.4.
Differential scanning calorimetry (DSC)
The effect of the various processing methods and filler type on the thermal properties of the composites was assessed using DSC (PYRIS Diamond DSC with Pyris series software, PerkinElmer Instruments, MA, USA). DSC was conducted on samples in duplicate, using a heating regime of 0–180 1C at a rate of 10 1C min 1 in a nitrogen atmosphere. Results are reported in terms of glass transition temperature (Tg) calculated from the ½ Cp (specific heat capacity), peak crystallization temperature (Tc) and melt peak Tm.
Mechanical properties
Tensile dog-bone shaped specimens of dimensions 2 2 10 mm3 (width thickness gauge length) were produced by injection moulding and tested with an Instron 5584 materials test machine (Instron Ltd, High Wycombe, UK). Compression test specimens 5 3 6 mm3 (width thickness height) were also produced by injection moulding and tested using an EZ-50 materials test machine (Lloyd Instruments, Hampshire, UK). Tests were conducted at 1 mm min 1 and a minimum of 10 repeat specimens were assessed to establish the Young's modulus, compressive and tensile strengths. Machine compliance was accounted for during data analysis
Table 2 – Target material composition and extrusion processing details. Material a
PLGA PLGC PLGH PLG4B PLGIB a
Vol% filler
Equivalent wt% filler
Processing temperature [1C]
Residence time in extruder [s]
0 25 25 25 25
0 44.1 45.5 42.1 43.6
220 220 220 180 190
0 420 420 51 120
Neat PLGA was not extruded. Residual weights were found to vary by o 5 wt% of the theoretical loadings, (TGA results shown in Fig. 3).
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
281
Fig. 3 – TGA of the PLGA composites, an assessment of the degree of their homogeneity. Td of PLGA¼281 1C. (a) PLGC, (b) PLGH, (c) PLG4B, (d) PLGIB; n ¼ 3 for each composite. Note: PLG4B was subjected to a higher temperature (see text in Section 2).
by using a video extensometer (tension) and correcting for machine stiffness (compression). Two-tailed t-tests were used to assess whether the composites produced results significantly different to PLGA alone.
2.4.5.
Dynamic mechanical analysis (DMA)
Three-point bend test specimens were cut from the centre of the compression moulded sheet for DMA testing and specimen edges were polished with 600, 1200 and 4000 grit SiC paper to remove cutting marks. Rectangular samples 24 3.5 1.7 mm3 were cut, giving a test span of 20 mm. A DMA 7e dynamic mechanical analyzer (PerkinElmer Instruments, MA, USA, with Pyris Series software) was used to perform temperature scan measurements over 25–65 1C at a heating rate of 4 1C min 1, with a 0.15% dynamic strain and 120% static to dynamic stress ratio, at 1 Hz. PLG4B was tested at 0.02% dynamic strain due to specimen failure at the higher strain. Nitrogen was used as a purge.
2.4.6.
Aging in simulated body fluid (SBF)
The effects of short-term aging and fluid absorption in SBF (Kokubo et al., 1990) were assessed for up to 21 days in terms of dynamic mechanical properties, weight and dimensional change. Samples were individually immersed in 40 ml of SBF in capped glass vials maintained at 37 1C under tangential agitation of 175 rpm, using an orbital shaker (C24 Incubator Shaker, New Brunswick Scientific, Edison, NJ, USA). Samples were removed from the SBF, tested, and then replaced to allow material properties to be tracked across consistent samples. Samples were removed at 1, 3, 7, 14, and 21 day(s); and on extraction were blotted dry with paper towel and weighed, their dimensions measured using Vernier callipers (taking 3 measurements per dimension), then subjected to DMA in three-point bending mode (sample orientation was maintained for consistency). All measurements were performed in quintuplet. DMA was conducted isothermally at 25 1C, using a regime of 0.02% dynamic strain, 120% static to dynamic stress ratio, at 1 Hz for 3 min.
282
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
3.
Results
3.1.
Thermal analyses of extruded and moulded materials
TGA of the extruded materials (Fig. 3) confirmed the actual weight percentage of filler in the composites to be equal to theoretical prediction, within 2.2–4.6%. The PLGA matrix degradation temperature (Td) of 281 1C did not alter upon melt extrusion in the presence of fillers CaCO3 or HA (Fig. 3a and b); whereas composites containing bioactive glasses 45S5 BG and ICIE4 BG exhibited a significantly reduced Td of 202 and 238 1C, respectively (Fig. 3c and d). Typical DSC traces are presented in Fig. 4 and summarized in Table 3. Fig. 4a shows that processing affected the neat PLGA, with a moderate reduction in Tm and Tg: it is evident that the as-received PLGA pellets possessed a degree of crystallinity associated with their thermal history, which may have acted to elevate the Tg. After extrusion all of the materials exhibited similar DSC profiles (Fig. 4b). However, DSC revealed significant changes in the materials with further processing (injection or compression moulding) (Fig. 4c and d). Injection moulding reduced Tgs, especially for the 45S5 BG-filled samples, with a Tg of 47.1 1C, in comparison to a Tg of 60 1C for PLGA. Neat PLGA differed to the composites in that it displayed a single Tm peak (with a flattened shoulder region), indicating melting of the crystallites either already present or formed during the DSC measurement, followed by some re-crystallization and melting of the remaining crystallites (Pijpers et al., 2002). The reductions
in Tg found for the 45S5 BG containing material, PLG4B (Fig. 4c and d), suggested substantial reduction in molecular weight.
3.2. Changes in molar mass due to processing and composition Processing reduced the molar mass of the PLGA matrix, as given by the GPC results (Fig. 5a and b). As-received PLGA pellets had a weight average molar mass of approximately 310 kg mol 1. The GPC elution profiles were similar for all materials assessed, with molar mass distributions remaining mono-modal. Milling and drying had negligible effect on molar mass. Of the processing steps, extrusion caused the greatest loss in molar mass, with reductions in PLGA Mw between 38% and 60%. Both injection and compression moulding caused further polymer degradation. Both types of bioactive glass were associated with massive loss of molar mass in comparison with HA or CaCO3 fillers. The materials degraded in the order PLGAoPLGCoPLGH⪡PLGIB⪡PLG4B.
3.3.
Mechanical and dynamic mechanical properties
With the exception of PLG4B, all the fillers significantly (Po0.0001) increased compressive and tensile Young's modulus, and the storage modulus of the composites (Table 4). The fillers did not influence compressive strength significantly, however, again, there was a reduction for the PLG4B composites. In tension, PLG4B was too brittle to be tested as it shattered during tightening of the tensile grips. The other
Fig. 4 – Representative examples of DSC curves: (a) effect of processing on PLGA; (b) effect of extrusion; (c) effect of injection moulding (IM); (d) effect of compression moulding (CM).
283
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
Table 3 – Summary of PLGA and composite thermal properties. Material
PLGA PLGC PLGH PLG4B PLGIB a b
TGAa
DSCb
DMAb
Td
Tg
Tc
Tm
Tg (from E″ peak)
281 282 280 202 238
59.8 58.3 59.7 57.1 58.6
116.3 109.6 107.8 106.5 107.6
161.0 158.2/165.2 159.2/165.0 152.7/160.2 156.0/164.0
54.2 54.7 53.7 46.1 52.7
After extrusion. After compression moulding (CM).
Fig. 5 – Change in weight average molecular weight during processing (n ¼ 2). (a) Compression moulding (CM); (b) injection moulding (IM).
Table 4 – Mechanical properties (s.d.): Ec ¼ compressive Young's modulus (n Z10), σc ¼ compressive strength (n Z10); Et ¼ tensile Young's modulus (n Z 10); σt ¼ tensile strength (n Z10); E0 ¼ storage modulus (isothermal at 25 1C) (n¼ 5). N.B. PLG4B could not be tested in tension as it shattered when gripped in the test machine. Material
Ec [GPa]
σc [GPa]
Et [GPa]
σt [MPa]
E0 [GPa]
PLGA PLGC PLGH PLG4B PLGIB
3.4 5.5 5.9 3.5 5.9
92.3 91.0 93.1 69.0 93.1
3.7 7.2 8.8 – 7.5
65.3 45.6 51.7 – 35.8
3.3 6.2 6.1 3.1 5.9
n
(0.1) (0.3)n (0.2)n (0.6) (0.3)n
Significantly different to PLGA (Po0.0001).
(5.0) (2.5) (2.5) (7.7)n (4.5)
(0.3) (0.7)n (2.0)n (1.4)n
(2.2) (7.2)n (9.7)n (7.3)n
(0.1) (0.3)n (0.2)n (0.2) (0.1)n
284
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
fillers significantly (Po0.05) reduced the tensile strength. DMA is useful as a corroborative technique to DSC and is often more sensitive in determining thermal transitions. After compression moulding, the Tg defined by the peak in loss modulus (E″) of the 45S5 BG containing composite was 46.1 1C and significantly reduced from that of the neat PLGA, of 54.2 1C. DMA data are summarized in Fig. 6a, and the ratios of E0 and tan δ of the composites to neat PLGA shown as functions of temperature, in Fig. 6b and c, respectively (Nazhat et al., 2000, 2001; Bleach et al., 2002). Fillers initially caused an increase in storage modulus, which was much more pronounced for PLGC, PLGH and PLGIB. The damping behaviour indicated by tan δ was similar for all materials other than PLG4B, which had a higher initial tan δ.
3.4.
Aging in SBF
The effective uptake of water as a function of time in SBF is indicated by weight and dimension increases (swelling) (Fig. 7a and b). The samples containing bioactive glass exhibited greatest changes: PLG4B samples increased in weight by 51%, and PLGIB by 24% after 7 days in SBF; there were no further increases in dimensions or weight. Moderate swelling was exhibited by PLGC and by PLGH to a lesser extent, whereas PLGA did not swell significantly. Immersion in SBF affected the viscoelastic properties in terms of E0 and tan δ (Fig. 8a and b). Some specimens broke in
testing or warped while immersed; therefore not all were tested in triplicate (see Fig. 8 for details). The composites containing bioactive glass experienced a rapid loss in E0 and an increase in tan δ. In contrast, the HA and CaCO3 containing composites retained similar properties to the PLGA specimens.
4.
Discussion
In this study of composites with bioactive particulate fillers in a PLGA matrix, the most significant finding was that those filled with bioactive glasses exhibited greater detrimental effects to the PLGA matrix than either HA or CaCO3. Bioglass 45S5s is highly active when implanted into bone defects, yet was the filler most detrimental to the PLGA matrix when processed under elevated temperature conditions, including melt extrusion and moulding. The thermal degradation temperature of PLGA was affected significantly by the inclusion of the bioactive glass formulations, this was not so when filled with either HA or CaCO3. Molar mass measurements showed that the highest overall degradation was experienced by the PLG4B composites, followed by the PLGIB composites. PLGC and PLGH showed comparable degradation during all process steps while PLG4B and PLGIB underwent similar degradation following extrusion, yet PLG4B showed pronounced degradation after compression moulding and injection moulding, probably due to the exposure
Fig. 6 – (a) Storage modulus (thick lines) and tan δ (thin lines) for PLGA and composites. Mean (for each 0.1 1C temperature increment recorded) 71s.d. (calculated at 25, 35, 45, 55 and 37 1C). n ¼ 3, (b) storage modulus and (c) tan δ, ratio of composite value to PLGA value. Mean71s.d.
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
285
Fig. 7 – (a) Percent weight increase over time in SBF, n ¼ 5, except PLG4B (n ¼ 4 at 3 and 7 days, and 2 thereafter) and PLGIB (n ¼3 at 14 and 21 days); (b) percent increase in specimen width over time in SBF. n¼ 5 (an average of three measurements along the width per specimen), except PLG4B (n ¼ 4 at 1 day, 3 at 3 days and 2 thereafter) and PLGIB (n ¼ 3 at 14 and 21 days). Mean71s.d.
times and high temperatures required by the viscosity of the mixture. Such degradation was avoided in the extrusion process for bioactive glass containing samples as the residence time in the extruder was less. The reductions in polymer matrix molar mass for the bioactive glass composites were not encountered in a previous study with relatively inert high density polyethylene matrices (Wang et al., 1998); this suggests that the reduction in molar mass observed when processing PLG4B was because PLGA is susceptible to degradation via hydrolysis. Similar reductions in molar mass were reported for composites of melt extruded bioactive glass and poly(εcaprolactone-co-DL-lactide 96/4) (Rich et al., 2002). The reduced strength for PLG4B may relate to the lower molar mass of the polymer (Fennel and Hill, 2001; Shikinami and Okuno, 1999). Stiffness is normally increased when stiffer particulate filler is added, especially in compression. However, polymer stiffness can be reduced if its molecular weight has greatly reduced, which is likely for PLG4B. Verheyen et al. (1992) found that increases in particle size increased strength and compressive stiffness. They suggested that this was because the material response was dominated by poor
interfacial adhesion and therefore the particles acted as defects in the material, thus the composite with smaller particles had more defects and reduced mechanical properties. The morphologies of the particles differ significantly, as shown in Fig. 1, the Bioglass 45S5s particles are shard-like, with acute angles likely to act as stress raisers, in comparison to the CaCO3 particles which are of a more regular, rhombohedral shape. HA and ICIE4 BG particles exhibit far less acute angles compared to Bioglass 45S5s. The presence of undercuts on these particles will enable enhanced mechanical interlocking with the polymer matrix (provided sufficient wetting). However, the degradation of the particle/matrix interface (and matrix) due to processing will negate this enhancement. During in vitro or in vivo degradation gaps at the interface will serve to increase fluid penetration and further detriment mechanical properties. In our work scanning electron microscopy of the fracture surface (Fig. 9) showed some voids at the matrix–filler interface, suggesting poor bonding, which can explain the reduction in tensile strength for all composites. As with many particulate filled polymer matrix composites, the effects of defects such as
286
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
Fig. 8 – (a) Storage modulus after aging in SBF; (b) tan δ after aging in SBF. n ¼ 5 (except: PLG4B; n¼ 4 at 1 day, 3 at 3 days, 2 thereafter. PLGIB: n ¼3 at 7 days and thereafter). Mean71s.d.
voids, poor wetting at the particle–matrix interface, and cracks initiating from sharp particulates are exacerbated in the tensile mode, contributing to the reduced strengths observed in comparison to the compressive mode. Surface functionalization to improve bonding may improve mechanical properties in the future. The DMA data suggests that at physiological temperature PLG4B had inferior mechanical properties to PLGA alone. Water uptake increased significantly in the presence of all fillers, particularly with the bioactive glasses. Ceramic fillers in PLA increase water absorption (Bleach et al., 2001), this also occurs when HA is combined with hydrophobic polymers (Padilla et al., 2002). Some of the findings from the aging study can be explained due to the enhanced ingress of water facilitated by the presence of filler particles, which are generally high surface energy materials, and poor interfaces which may have acted as capillaries. Osmotic effects may also be present, due to dissolution of the fillers, particularly the more soluble bioactive glasses. Although the fillers served to increase the stiffness of the PLGA initially, after approximately 7 days in SBF this effect was lost (o1 day in the case of
45S5 BG filled composites), due to fluid ingress and plasticization of the matrix. The composite materials with more soluble fillers and lower molecular weight polymer matrices underwent a more rapid loss in stiffness. The effects of heat and shear during manufacture of the polymer specimens as the primary cause of degradation may be ruled out since it was only the bioactive glass fillers that caused significant degradation. This suggests that there is a chemical interaction between certain silica-containing or silica doped phosphate based glass formulations with hydrolysable polyesters (Blaker et al., 2005a; Shah Mohammadi et al., 2010). Similar findings have been found for Bioglass 45S5s reinforced poly(DLlactide) (PDLLA) composites (Blaker et al., 2005a, 2010). In previous work on PDLLA (Blaker et al., 2010) the inclusion of Bioglass 45S5s under elevated temperatures resulted in degradation of the matrix, leading to reduction in mechanical properties and matrix molecular weight. Bioglass 45S5s has a high-energy surface (Blaker et al., 2005b) and will adsorb water to give SiOH- at its surface. Its inclusion into poly(α-hydroxyesters) will result in polymer hydrolysis, with an increased reaction rate at elevated temperatures and increased pH (base catalysis of the ester). Silica-doped phosphate-based glasses
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
287
Fig. 9 – SEM micrographs of the failure surfaces of tensile PLG and composite specimens, tested at ambient temperature. (a, b) PLG, (c–e) PLGC, (f–h) PLGH, (i–k) PLG4B, (l–n) PLGIB. Scale bar: left images 900 μm, centre images 150 μm, right images 30 μm.
288
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
have also demonstrated a degradative effect on compression moulded polycaprolactone matrices (Shah Mohammadi et al., 2010). In the development of bone analogous materials, a composite of bioactive particles in a polymer matrix should have improved mechanical properties, provided that there is good adhesion between the filler and matrix. This can be achieved by mechanical interlocking, which can be dependent on the size and surface roughness of the particles, or by a chemical bond as achieved through surface treatment. A roughened bioactive glass surface may improve interfacial interaction: sol–gel produced glasses have porous surfaces, unlike the smooth surface of melt-derived glasses as tested here (Sepulveda et al., 2002).
5.
Conclusions
This study has enabled the direct comparison between a poly(αhydroxyester), 95/5 poly(L-lactide-co-glycolide) (PLGA), and PLGA matrix composites filled with four different particulate fillers produced with a consistent processing and testing regime. Composites processed with calcium carbonate or hydroxylapatite underwent limited polymer degradation, yet significantly increased the modulus of the polymer; whereas bioactive glass fillers resulted in significant polymer degradation, with reduced thermo-mechanical properties due to the lowered molar mass of the PLGA matrix and poor matrix–filler interface bond. Shortterm aging studies also suggest that the stiffening effect of these composites is reduced markedly after a short period in SBF due to fluid ingress and plasticization of the matrix. The bioactive glass filled composites exhibited the poorest properties due to the premature degradation of the PLGA matrix during processing; these composites should be processed via lower temperature routes to reduce such degradation. The evidence suggests that the BG fillers assessed herein, manufactured using these methods, have limited utility and are not suitable for strengthening the polymers for the use of bone replacement. PLGC or PLGH would be the preferable fillers based on this data; further assessment of bioactivity (e.g. cell culture) would allow us to discern the most appropriate material of the systems investigated here for use in bone replacement applications.
Acknowledgements RLS was supported by an EPSRC doctoral training grant, the Rosetrees Trust, the University of London Central Research Fund, the Institution of Mechanical Engineers, and the Society of Orthopaedic Medicine. The authors are grateful to Steven Lamoriniere for making the moulds used in the forming of the composites. The authors have no conflicts of interest.
Appendix A.
Supporting information
Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jmbbm. 2015.06.008.
r e f e r e n c e s
Blaker, J.J., Maquet, V., Je´roˆme, R., Boccaccini, A.R., Nazhat, S.N., 2005a. Mechanical properties of highly porous PDLLA/ Bioglasss composite foams as scaffolds for bone tissue engineering. Acta Biomater. 1 (6), 643–652. Blaker, J.J., Maquet, V., Boccaccini, A.R., Jerome, R., Bismarck, A., 2005b. Wetting of bioactive glass surfaces by poly(alphahydroxyacid) melts: interaction between Bioglasss and biodegradable polymers. E-Polymers 5 (1), 248–260. Blaker, J.J., Bismark, A., Boccaccini, A.R., Young, A.M., Nazhat, S. N., 2010. Premature degradation of poly(a-hydroxyesters) during thermal processing of Bioglasss-containing composites. Acta Biomater. 6 (3), 756–762. Blaker, J.J., Nazhat, S.N., Maquet, V., Boccaccini, A.R., 2011. Longterm in vitro degradation of PDLLA/Bioglasss bone scaffolds in acellular simulated body fluid. Acta Biomater. 7 (2), 829–840. Bleach, N.C., Tanner, K.E., Kelloma¨ki, M., To¨rma¨la¨, P., 2001. Effect of filler type on the mechanical properties of self-reinforced polylactide–calcium phosphate composites. J. Mater. Sci. Mater. Med. 12 (10–12), 911–915. Bronzino, J.D., 1995. In: The Biomedical Engineering Handbook. CRC Press, USA. Bleach, N.C., Nazhat, S.N., Tanner, K.E., Kelloma¨ki, M., To¨rma¨la¨, P., 2002. Effect of filler content on mechanical and dynamic mechanical properties of particulate biphasic calcium phosphate–polylactide composites. Biomaterials 23 (7), 1579–1585. Boccaccini, A.R., Blaker, J.J., 2005. Bioactive composite materials for tissue engineering scaffolds. Expert Rev. Med. Devices 2 (3), 303–317. Boccaccini, A.R., Roether, J.A., Hench, L.L., Maquet, V., Je´roˆme, R., 2002. A composites approach to tissue engineering. In: Lin, H. T., Singh, M., (Eds), Ceramic Engineering & Science Proceedings of the 26th Annual Conference on Composites, Advanced Ceramics, Materials and Structures: B, Florida, 23 (4), pp. 805–816. Cao, W., Hench, L.L., 1996. Bioactive materials. Ceram. Int. 22 (6), 493–507. Domb, A.J., Kost, J., Wiseman, D.M., 1997. Handbook of biodegradable polymers. In: Florence, AT., Gregoriadis, G. (Eds.), Drug Targeting and Delivery. Harwood Academic Publishers, The Netherlands. Elgayar, I., Aliev, A.E., Boccaccini, A.R., Hill, RG., 2005. Structural analysis of bioactive glasses. J. Non-Cryst. Solids 351 (2), 173–183. Engelberg, I., Kohn, J., 1991. Physico-mechanical properties of degradable polymers in medical applications: a comparative study. Biomaterials 12 (3), 292–304. Fennel, B., Hill, RG., 2001. The influence of poly(acrylic acid) molar mass and concentration on the properties of polyalkenoate cements. J. Mater. Sci. 36 (21), 5193–5202. Hench, L.L., 1991. Bioceramics: from concept to clinic. J. Am. Ceram. Soc. 74 (7), 1487–1510. Ignatius, A.A., Wolf, S., Augat, P., Claes, L.E., 2001. Composites made of rapidly resorbable ceramics and poly(lactide) show adequate mechanical properties for use as bone substitute materials. J. Biomed. Mater. Res. 57 (1), 126–131. Kikuchi, M., Koyama, Y., Takakuda, K., Miyairi, H., Shirahama, N., Tanaka, J., 2002. In vitro change in strength of β-tricalcium phosphate/copolymersised poly-L-lactide composites and their application for guided bone regeneration. J. Biomed. Mater. Res. 62 (2), 265–272. Kokubo, T., Kushitani, H., Sakka, S., Kitsugi, T., Yamamuro, T., 1990. Solutions able to reproduce in vivo surface-structure changes in bioactive glass–ceramic A-W. J. Biomed. Mater. Res. 24 (6), 721–734.
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
Middleton, J.C., Tipton, A.J., 2000. Synthetic biodegradable polymers as orthopedic devices. Biomaterials 21 (23), 2335–2346. Nazhat, S.N., Joseph, R., Wang, M., Smith, R., Tanner, K.E., Bonfield, W., 2000. Dynamic mechanical characterization of hydroxyapatite reinforced polyethylene: effect of particle size. J. Mater. Sci. Mater. Med. 11 (10), 621–628. Nazhat, S.N., Kelloma¨ki, M., To¨rma¨la¨, P., Tanner, K.E., Bonfield, W., 2001. Dynamic mechanical characterization of biodegradable composites of hydroxyapatite and polylactides. J. Biomed. Mater. Res. 58 (4), 335–343. Padilla, S., del Real, R.P., Vallet-Regı´, M., 2002. In vitro release of gentamicin from OHAp/PEMA/PMMA samples. J. Control. Release 83 (3), 343–352. Pijpers, T.F.J., Mathot, V.B.F., Goderis, B., Scherrenberg, RL., van der Vegte, E.W., 2002. High-speed calorimetry for the study of the kinetics of (de)vitrification, crystallization, and melting of macromolecules. Macromolecules 35 (9), 3601–3613. Rezwan, K., Chen, Q.Z., Blaker, J.J., Boccaccini, A.R., 2006. Biodegradable and bioactive porous polymer/inorganic composite scaffolds for bone tissue engineering. Biomaterials 27 (18), 3413–3431. Rich, J., Jaakkola, T., Tirri, T., Na¨rhi, T., Yli-Urpo, A., Seppa¨la¨, J., 2002. In vitro evaluation of poly(ε-caprolactone-co-DL-lactide)/ bioactive glass composites. Biomaterials 23 (10), 2143–2150. Schiller, C., Epple, M., 2003. Carbonated calcium phosphates are suitable pH-stabilising fillers for degradable polyesters. Biomaterials 24 (12), 2037–2043.
289
Sepulveda, P., Jones, J.R., Hench, L.L., 2002. In vitro dissolution of melt-derived 45S5 and sol–gel derived 58S bioactive glasses. J. Biomed. Mater. Res. 61 (2), 301–311. Shah Mohammadi, M., Ahmed, I., Marelli, B., Rudd, C., Bureau, M. N., Nazhat, SN., 2010. Modulation of polycaprolactone composite properties through incorporation of mixed phosphate glass formulations. Acta Biomater. 6 (8), 3157–3168. Shikinami, Y., Okuno, M., 1999. Bioresorbable devices made of forged composites of hydroxyapatite (HA) particles and polyL-lactide (PLLA): Part 1. Basic characteristics. Biomaterials 20 (9), 859–877. Tsunoda, M., 2003. Degradation of poly(DL-lactic acid-co-glycolic acid) containing calcium carbonate and hydroxyapatite fillers – effect of size and shape of the fillers. Dent. Mater. J. 22 (3), 371–382. Verheyen, C.C.P.M., de Wijn, J.R., van Blitterswijk, C.A., de Groot, K., 1992. Evaluation of hydroxylapatite/poly(L-lactide) composites: mechanical behaviour. J. Biomed. Mater. Res. 26 (10), 1277–1296. Wang, M., 2003. Developing bioactive composite materials for tissue replacement. Biomaterials 24 (13), 2133–2151. Wang, M., Hench, L.L., Bonfield, W., 1998. Bioglasss/high density polyethylene composite for soft tissue applications: preparation and evaluation. J. Biomed. Mater. Res. 42 (4), 577–586.
Available online at www.sciencedirect.com
www.elsevier.com/locate/jmbbm
Research Paper
A comparative study of the effects of different bioactive fillers in PLGA matrix composites and their suitability as bone substitute materials: A thermo-mechanical and in vitro investigation R.L. Simpsona, S.N. Nazhatb, J.J. Blakerc,n, A. Bismarckd, R. Hille, A.R. Boccaccinif,g,n, U.N. Hansena, A.A. Amisa,h a
Department of Mechanical Engineering, Imperial College London, London SW7 2AZ, UK Department of Mining and Materials Engineering, McGill University, 3610 University Street, Montreal, Quebec H3A 0C5, Canada c Materials Science Centre, School of Materials, University of Manchester, Grovesnor Street, Manchester M1 7HS, UK d Department of Chemical Engineering, Polymer and Composite Engineering (PaCE) Group, Imperial College London, London SW7 2AZ, UK e Institute of Dentistry, Dental Physical Science Unit, Queen Mary, University of London, London E1 4NS, UK f Department of Materials, Imperial College London, London SW7 2BP, UK g Institute of Biomaterials, Department of Materials Science and Engineering, University of Erlangen-Nuremburg, 91058 Erlangen, Germany h Musculoskeletal Surgery Group, Imperial College School of Medicine, London W6 8RF, UK b
art i cle i nfo
ab st rac t
Article history:
Bone substitute composite materials with poly(L-lactide-co-glycolide) (PLGA) matrices and
Received 20 February 2015
four different bioactive fillers: CaCO3, hydroxyapatite (HA), 45S5 Bioglasss (45S5 BG), and
Received in revised form
ICIE4 bioactive glass (a lower sodium glass than 45S5 BG) were produced via melt blending,
5 June 2015
extrusion and moulding. The viscoelastic, mechanical and thermal properties, and the
Accepted 5 June 2015
molecular weight of the matrix were measured. Thermogravimetric analysis evaluated the
Available online 16 June 2015
effect of filler composition on the thermal degradation of the matrix. Bioactive glasses
Keywords:
caused premature degradation of the matrix during processing, whereas CaCO3 or HA did
Bioactive glass
not. All composites, except those with 45S5 BG, had similar mechanical strength and were
Hydroxyapatite
stiffer than PLGA alone in compression, whilst all had a lower tensile strength. Dynamic
Poly(α-hydroxyester) Bone substitute Composite
mechanical analysis demonstrated an increased storage modulus (E0 ) in the composites (other than the 45S5 BG filled PLGA). The effect of water uptake and early degradation was investigated by short-term in vitro aging in simulated body fluid, which indicated enhanced
n Corresponding authors at: School of Materials, University of Manchester (J.J. Blaker) and Institute of Biomaterials, University of Erlangen-Nuremberg (A.R. Boccaccini). E-mail addresses: [email protected] (J.J. Blaker), [email protected] (A.R. Boccaccini).
http://dx.doi.org/10.1016/j.jmbbm.2015.06.008 1751-6161/& 2015 Elsevier Ltd. All rights reserved.
278
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
water uptake over the neat polymer; bioactive glass had the greatest water uptake, causing matrix plasticization. These results enable a direct comparison between bioactive filler type in poly(α-hydroxyester) composites, and have implications when selecting a composite material for eventual application in bone substitution. & 2015 Elsevier Ltd. All rights reserved.
1.
Introduction
Resorbable polymers and ceramics have been developed for orthopaedic procedures, which may encourage bone growth, guiding and supporting regenerating bone for sufficient time such that it is capable of self-support (Cao and Hench, 1996). Ultimately these implanted materials degrade harmlessly in the body (Middleton and Tipton, 2000; Domb et al., 1997; Hench, 1991). However, neither polymers or ceramics, taken separately, have proven entirely suitable for use in bone replacement: polymers may have insufficient mechanical properties (Engelberg and Kohn, 1991), while ceramics often have an unpredictable, low tensile strength so they are unsuitable for use in locations of significant torsion or bending (Bronzino, 1995). Composites have been studied in efforts to obtain a superior bone substitute material, usually polymer matrices with particulate ceramic/glass fillers (Wang, 2003; Rezwan et al., 2006; Boccaccini and Blaker, 2005). These should result in increased mechanical properties over the polymer and impart bioactivity (Boccaccini et al.,
2002). Basic ion releasing ceramics/glasses have a pH buffering effect in the early stages of degradation (Kikuchi et al., 2002; Tsunoda, 2003; Schiller and Epple, 2003), which could reduce the degradation rate of hydrolytically degradable polymers, such as poly(lactide-co-glycolide) (PLGA) and polylactide (PLA) (Blaker et al., 2011). Whilst numerous composites have been investigated as bone analogues, the optimal composite has yet to be developed due to the complex nature of the intended application (Rezwan et al., 2006). In order to compare different composites accurately, it is necessary to use the same production method and testing regime because differences in these will result in variations in material and mechanical properties. Few examples of this type of comparison have been found (Bleach et al., 2001; Ignatius et al., 2001), making direct correlation difficult. The aim of this study was to compare four different particulate fillers in a PLGA matrix using similar processing routes and testing methods to enable direct comparison of the effects of filler selection on composite properties in the as-made state and during aging in simulated body fluid (SBF).
Fig. 1 – SEM images of (a) CaCO3 (rhombohedral particles), (b) HA with irregular shaped particles, (c) 45S5 BG showing fewer small particles, (d) ICIE4 BG with larger particles dispersed amongst smaller particles.
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
279
Table 1 – Particle size distribution of materials. D10, D50 and D90 ¼10, 50, 90 % (size by volume, obtained using a CILAS 1180 Particle size analyzer, Orléans, France), respectively, of particles were this size or less. (Average PLGA n ¼4, CaCO3 and HA n ¼ 5, 45S5 BG and ICIE4 BG n ¼3). Material
Average D10 particle size [μm]
Average D50 particle size [μm]
Average D90 particle size [μm]
PLGA (powder) CaCO3 HA 45S5 BG ICIE4 BG
73.5 3.6 0.7 7.5 0.7
198.4 16.4 3.8 35.3 5.2
346.2 24.9 14.6 62.1 41.1
Fig. 2 – Processing routes used to compare the effects of different fillers on PLGA.
2.
Materials and methods
2.1.
Materials
Granules of 95:5 poly(L-lactide-co-glycolide) (PLGA) semicrystalline poly(α-hydroxyester) co-polymer with a manufacturer specified inherent viscosity of 2.38 dl/g and melting range 160–170 1C, were used (PURAC, Gorinchem, The Netherlands). To enable subsequent processing, the granules were cryogenically milled to a powder in liquid nitrogen using an Alpine 100UPZII mill (Hosokawa Micron Ltd., Cheshire, UK), then dried in a vacuum oven. The powder was then sealed in moisture-proof sachets and stored in a
desiccator with indicating silica gel and phosphorous pentoxide until use. The particulate fillers were CaCO3 (Sigma-Aldrich Co. Ltd., Dorset, UK), sintered hydroxylapatite CAPTALs S (HA) (Plasma Biotal Ltd., Derbyshire, UK), 45S5 Bioglasss (45S5 BG) (of composition in mol%: SiO2 46.13, P2O5 2.61, CaO 26.91, Na2O 24.35); (MoSci Corporation, Montana, USA), and ICIE4 bioactive glass (ICIE4 BG) (of composition in mol%: SiO2 49.46, P2O5 1.07, CaO 42.87, Na2O 6.6) which was fabricated in-house (Elgayar et al., 2005). All fillers were dried at 250 1C in a fan oven for 4 h and stored as described previously for the PLGA powder. Particulate morphology and size can be compared from the scanning electron microscopy images in Fig. 1 and are summarized in Table 1.
280
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
2.2.
Extrusion of neat and composite PLGA material
The generic processing routes are presented schematically in Fig. 2. PLGA was combined with 25 vol% of bioactive filler. This loading was selected as it was expected to retain bioactivity and optimal mechanical properties of the composites while allowing the material to be easily processed. The mixtures were transferred to vials and agitated using a roller mixer (SRT1, Bibby Sterilin Ltd., Staffordshire, UK) for 2.5 h (based on previous trials and observations) to ensure homogeneous particle mixing. A twin-screw 5 cm3 micro-extruder (DSM research products, Maastricht, The Netherlands) blended the composite material at 30 rpm. Extrusion parameters were initially optimized (Table 2). using PLGA filled with CaCO3 (referred to as PLGC hereafter); these parameters were adjusted for the other material combinations: PLGA filled with HA (PLGH), 45S5 BG (PLG4B) or ICIE4 BG (PLGIB) as required to reduce degradation indicated by observable discoloration. The extruded materials were cryogenically ground under liquid nitrogen, to yield a powder of particle size o1 mm. Powders were dried in an oven for 5.5 h at 100 1C and stored as detailed previously, at 4 1C. For the extruded samples, matrix–filler bonding was not assessed as extrusion was used purely to obtain a more homogenous mix.
2.3.
Composite forming
Two methods were used to fabricate test specimens: rudimentary injection moulding and compression moulding (Fig. 2). The injection mould was filled with material and heated to the extrusion temperature given in Table 2, this mould was then transferred to a press (Moores Presses Ltd, Birmingham, UK) and compressed to 9.5 kN, which was maintained until the mould cooled to 30 1C. Compression moulding of flat sheets was conducted using an aluminium mould with the same conditions.
positive controls were PLGA alone and bioactive fillers, respectively. The homogeneity of the bioactive fillers in the composites was determined by comparing residual filler weights (obtained from TGA) with the theoretical weights (given in Table 2). The onset of thermal degradation was defined here as a loss of 40.5% relative to starting mass.
2.4.2.
2.4.3.
Material characterization
2.4.1.
Thermogravimetric analysis (TGA)
Extruded samples were analyzed using TGA (Stanton Redcroft Simultaneous Thermal Analyzer (STA-780), London, UK). Triplicate samples of 15–20 mg were heated in air with a heating regime of ambient temperature to 500 1C to burn off the polymer, with a heating rate of 10 1C min 1 (the maximum temperature had to be increased to 700 1C for the PLGA–45S5 BG (PLG4B) composites as the glass sintered at approximately 450 1C, trapping the polymer). Negative and
Gel permeation chromatography (GPC)
GPC was used to assess the effect of filler type and processing on the molar mass of the polymer. GPC was carried out in duplicate for each sample by RAPRA Technology, Shropshire, UK. Chloroform solvent (10 ml) was added to 20 mg polymer or 30–40 mg composite samples, which were left for a minimum of 4 h to dissolve. Solutions were then filtered through a 0.2 μm polyamide membrane into sample vials, and put into an autosampler. Columns suited to medium/high molecular weight samples were used (Plgel guard plus 2 mixed bed-B, 30 cm, 5–10 μm). The flow rate used was 1.0 ml min 1 (nominal) at 30 1C. A refractive index detector was used. Viscotek Trisec 2000 and Trisec 3.0 software was used to collect and analyze the data, respectively. The equipment was calibrated using narrow distribution polystyrene calibrants (Polymer Laboratories Ltd), to give results as ‘polystyrene equivalent’ molecular weights.
2.4.4. 2.4.
Differential scanning calorimetry (DSC)
The effect of the various processing methods and filler type on the thermal properties of the composites was assessed using DSC (PYRIS Diamond DSC with Pyris series software, PerkinElmer Instruments, MA, USA). DSC was conducted on samples in duplicate, using a heating regime of 0–180 1C at a rate of 10 1C min 1 in a nitrogen atmosphere. Results are reported in terms of glass transition temperature (Tg) calculated from the ½ Cp (specific heat capacity), peak crystallization temperature (Tc) and melt peak Tm.
Mechanical properties
Tensile dog-bone shaped specimens of dimensions 2 2 10 mm3 (width thickness gauge length) were produced by injection moulding and tested with an Instron 5584 materials test machine (Instron Ltd, High Wycombe, UK). Compression test specimens 5 3 6 mm3 (width thickness height) were also produced by injection moulding and tested using an EZ-50 materials test machine (Lloyd Instruments, Hampshire, UK). Tests were conducted at 1 mm min 1 and a minimum of 10 repeat specimens were assessed to establish the Young's modulus, compressive and tensile strengths. Machine compliance was accounted for during data analysis
Table 2 – Target material composition and extrusion processing details. Material a
PLGA PLGC PLGH PLG4B PLGIB a
Vol% filler
Equivalent wt% filler
Processing temperature [1C]
Residence time in extruder [s]
0 25 25 25 25
0 44.1 45.5 42.1 43.6
220 220 220 180 190
0 420 420 51 120
Neat PLGA was not extruded. Residual weights were found to vary by o 5 wt% of the theoretical loadings, (TGA results shown in Fig. 3).
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
281
Fig. 3 – TGA of the PLGA composites, an assessment of the degree of their homogeneity. Td of PLGA¼281 1C. (a) PLGC, (b) PLGH, (c) PLG4B, (d) PLGIB; n ¼ 3 for each composite. Note: PLG4B was subjected to a higher temperature (see text in Section 2).
by using a video extensometer (tension) and correcting for machine stiffness (compression). Two-tailed t-tests were used to assess whether the composites produced results significantly different to PLGA alone.
2.4.5.
Dynamic mechanical analysis (DMA)
Three-point bend test specimens were cut from the centre of the compression moulded sheet for DMA testing and specimen edges were polished with 600, 1200 and 4000 grit SiC paper to remove cutting marks. Rectangular samples 24 3.5 1.7 mm3 were cut, giving a test span of 20 mm. A DMA 7e dynamic mechanical analyzer (PerkinElmer Instruments, MA, USA, with Pyris Series software) was used to perform temperature scan measurements over 25–65 1C at a heating rate of 4 1C min 1, with a 0.15% dynamic strain and 120% static to dynamic stress ratio, at 1 Hz. PLG4B was tested at 0.02% dynamic strain due to specimen failure at the higher strain. Nitrogen was used as a purge.
2.4.6.
Aging in simulated body fluid (SBF)
The effects of short-term aging and fluid absorption in SBF (Kokubo et al., 1990) were assessed for up to 21 days in terms of dynamic mechanical properties, weight and dimensional change. Samples were individually immersed in 40 ml of SBF in capped glass vials maintained at 37 1C under tangential agitation of 175 rpm, using an orbital shaker (C24 Incubator Shaker, New Brunswick Scientific, Edison, NJ, USA). Samples were removed from the SBF, tested, and then replaced to allow material properties to be tracked across consistent samples. Samples were removed at 1, 3, 7, 14, and 21 day(s); and on extraction were blotted dry with paper towel and weighed, their dimensions measured using Vernier callipers (taking 3 measurements per dimension), then subjected to DMA in three-point bending mode (sample orientation was maintained for consistency). All measurements were performed in quintuplet. DMA was conducted isothermally at 25 1C, using a regime of 0.02% dynamic strain, 120% static to dynamic stress ratio, at 1 Hz for 3 min.
282
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
3.
Results
3.1.
Thermal analyses of extruded and moulded materials
TGA of the extruded materials (Fig. 3) confirmed the actual weight percentage of filler in the composites to be equal to theoretical prediction, within 2.2–4.6%. The PLGA matrix degradation temperature (Td) of 281 1C did not alter upon melt extrusion in the presence of fillers CaCO3 or HA (Fig. 3a and b); whereas composites containing bioactive glasses 45S5 BG and ICIE4 BG exhibited a significantly reduced Td of 202 and 238 1C, respectively (Fig. 3c and d). Typical DSC traces are presented in Fig. 4 and summarized in Table 3. Fig. 4a shows that processing affected the neat PLGA, with a moderate reduction in Tm and Tg: it is evident that the as-received PLGA pellets possessed a degree of crystallinity associated with their thermal history, which may have acted to elevate the Tg. After extrusion all of the materials exhibited similar DSC profiles (Fig. 4b). However, DSC revealed significant changes in the materials with further processing (injection or compression moulding) (Fig. 4c and d). Injection moulding reduced Tgs, especially for the 45S5 BG-filled samples, with a Tg of 47.1 1C, in comparison to a Tg of 60 1C for PLGA. Neat PLGA differed to the composites in that it displayed a single Tm peak (with a flattened shoulder region), indicating melting of the crystallites either already present or formed during the DSC measurement, followed by some re-crystallization and melting of the remaining crystallites (Pijpers et al., 2002). The reductions
in Tg found for the 45S5 BG containing material, PLG4B (Fig. 4c and d), suggested substantial reduction in molecular weight.
3.2. Changes in molar mass due to processing and composition Processing reduced the molar mass of the PLGA matrix, as given by the GPC results (Fig. 5a and b). As-received PLGA pellets had a weight average molar mass of approximately 310 kg mol 1. The GPC elution profiles were similar for all materials assessed, with molar mass distributions remaining mono-modal. Milling and drying had negligible effect on molar mass. Of the processing steps, extrusion caused the greatest loss in molar mass, with reductions in PLGA Mw between 38% and 60%. Both injection and compression moulding caused further polymer degradation. Both types of bioactive glass were associated with massive loss of molar mass in comparison with HA or CaCO3 fillers. The materials degraded in the order PLGAoPLGCoPLGH⪡PLGIB⪡PLG4B.
3.3.
Mechanical and dynamic mechanical properties
With the exception of PLG4B, all the fillers significantly (Po0.0001) increased compressive and tensile Young's modulus, and the storage modulus of the composites (Table 4). The fillers did not influence compressive strength significantly, however, again, there was a reduction for the PLG4B composites. In tension, PLG4B was too brittle to be tested as it shattered during tightening of the tensile grips. The other
Fig. 4 – Representative examples of DSC curves: (a) effect of processing on PLGA; (b) effect of extrusion; (c) effect of injection moulding (IM); (d) effect of compression moulding (CM).
283
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
Table 3 – Summary of PLGA and composite thermal properties. Material
PLGA PLGC PLGH PLG4B PLGIB a b
TGAa
DSCb
DMAb
Td
Tg
Tc
Tm
Tg (from E″ peak)
281 282 280 202 238
59.8 58.3 59.7 57.1 58.6
116.3 109.6 107.8 106.5 107.6
161.0 158.2/165.2 159.2/165.0 152.7/160.2 156.0/164.0
54.2 54.7 53.7 46.1 52.7
After extrusion. After compression moulding (CM).
Fig. 5 – Change in weight average molecular weight during processing (n ¼ 2). (a) Compression moulding (CM); (b) injection moulding (IM).
Table 4 – Mechanical properties (s.d.): Ec ¼ compressive Young's modulus (n Z10), σc ¼ compressive strength (n Z10); Et ¼ tensile Young's modulus (n Z 10); σt ¼ tensile strength (n Z10); E0 ¼ storage modulus (isothermal at 25 1C) (n¼ 5). N.B. PLG4B could not be tested in tension as it shattered when gripped in the test machine. Material
Ec [GPa]
σc [GPa]
Et [GPa]
σt [MPa]
E0 [GPa]
PLGA PLGC PLGH PLG4B PLGIB
3.4 5.5 5.9 3.5 5.9
92.3 91.0 93.1 69.0 93.1
3.7 7.2 8.8 – 7.5
65.3 45.6 51.7 – 35.8
3.3 6.2 6.1 3.1 5.9
n
(0.1) (0.3)n (0.2)n (0.6) (0.3)n
Significantly different to PLGA (Po0.0001).
(5.0) (2.5) (2.5) (7.7)n (4.5)
(0.3) (0.7)n (2.0)n (1.4)n
(2.2) (7.2)n (9.7)n (7.3)n
(0.1) (0.3)n (0.2)n (0.2) (0.1)n
284
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
fillers significantly (Po0.05) reduced the tensile strength. DMA is useful as a corroborative technique to DSC and is often more sensitive in determining thermal transitions. After compression moulding, the Tg defined by the peak in loss modulus (E″) of the 45S5 BG containing composite was 46.1 1C and significantly reduced from that of the neat PLGA, of 54.2 1C. DMA data are summarized in Fig. 6a, and the ratios of E0 and tan δ of the composites to neat PLGA shown as functions of temperature, in Fig. 6b and c, respectively (Nazhat et al., 2000, 2001; Bleach et al., 2002). Fillers initially caused an increase in storage modulus, which was much more pronounced for PLGC, PLGH and PLGIB. The damping behaviour indicated by tan δ was similar for all materials other than PLG4B, which had a higher initial tan δ.
3.4.
Aging in SBF
The effective uptake of water as a function of time in SBF is indicated by weight and dimension increases (swelling) (Fig. 7a and b). The samples containing bioactive glass exhibited greatest changes: PLG4B samples increased in weight by 51%, and PLGIB by 24% after 7 days in SBF; there were no further increases in dimensions or weight. Moderate swelling was exhibited by PLGC and by PLGH to a lesser extent, whereas PLGA did not swell significantly. Immersion in SBF affected the viscoelastic properties in terms of E0 and tan δ (Fig. 8a and b). Some specimens broke in
testing or warped while immersed; therefore not all were tested in triplicate (see Fig. 8 for details). The composites containing bioactive glass experienced a rapid loss in E0 and an increase in tan δ. In contrast, the HA and CaCO3 containing composites retained similar properties to the PLGA specimens.
4.
Discussion
In this study of composites with bioactive particulate fillers in a PLGA matrix, the most significant finding was that those filled with bioactive glasses exhibited greater detrimental effects to the PLGA matrix than either HA or CaCO3. Bioglass 45S5s is highly active when implanted into bone defects, yet was the filler most detrimental to the PLGA matrix when processed under elevated temperature conditions, including melt extrusion and moulding. The thermal degradation temperature of PLGA was affected significantly by the inclusion of the bioactive glass formulations, this was not so when filled with either HA or CaCO3. Molar mass measurements showed that the highest overall degradation was experienced by the PLG4B composites, followed by the PLGIB composites. PLGC and PLGH showed comparable degradation during all process steps while PLG4B and PLGIB underwent similar degradation following extrusion, yet PLG4B showed pronounced degradation after compression moulding and injection moulding, probably due to the exposure
Fig. 6 – (a) Storage modulus (thick lines) and tan δ (thin lines) for PLGA and composites. Mean (for each 0.1 1C temperature increment recorded) 71s.d. (calculated at 25, 35, 45, 55 and 37 1C). n ¼ 3, (b) storage modulus and (c) tan δ, ratio of composite value to PLGA value. Mean71s.d.
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
285
Fig. 7 – (a) Percent weight increase over time in SBF, n ¼ 5, except PLG4B (n ¼ 4 at 3 and 7 days, and 2 thereafter) and PLGIB (n ¼3 at 14 and 21 days); (b) percent increase in specimen width over time in SBF. n¼ 5 (an average of three measurements along the width per specimen), except PLG4B (n ¼ 4 at 1 day, 3 at 3 days and 2 thereafter) and PLGIB (n ¼ 3 at 14 and 21 days). Mean71s.d.
times and high temperatures required by the viscosity of the mixture. Such degradation was avoided in the extrusion process for bioactive glass containing samples as the residence time in the extruder was less. The reductions in polymer matrix molar mass for the bioactive glass composites were not encountered in a previous study with relatively inert high density polyethylene matrices (Wang et al., 1998); this suggests that the reduction in molar mass observed when processing PLG4B was because PLGA is susceptible to degradation via hydrolysis. Similar reductions in molar mass were reported for composites of melt extruded bioactive glass and poly(εcaprolactone-co-DL-lactide 96/4) (Rich et al., 2002). The reduced strength for PLG4B may relate to the lower molar mass of the polymer (Fennel and Hill, 2001; Shikinami and Okuno, 1999). Stiffness is normally increased when stiffer particulate filler is added, especially in compression. However, polymer stiffness can be reduced if its molecular weight has greatly reduced, which is likely for PLG4B. Verheyen et al. (1992) found that increases in particle size increased strength and compressive stiffness. They suggested that this was because the material response was dominated by poor
interfacial adhesion and therefore the particles acted as defects in the material, thus the composite with smaller particles had more defects and reduced mechanical properties. The morphologies of the particles differ significantly, as shown in Fig. 1, the Bioglass 45S5s particles are shard-like, with acute angles likely to act as stress raisers, in comparison to the CaCO3 particles which are of a more regular, rhombohedral shape. HA and ICIE4 BG particles exhibit far less acute angles compared to Bioglass 45S5s. The presence of undercuts on these particles will enable enhanced mechanical interlocking with the polymer matrix (provided sufficient wetting). However, the degradation of the particle/matrix interface (and matrix) due to processing will negate this enhancement. During in vitro or in vivo degradation gaps at the interface will serve to increase fluid penetration and further detriment mechanical properties. In our work scanning electron microscopy of the fracture surface (Fig. 9) showed some voids at the matrix–filler interface, suggesting poor bonding, which can explain the reduction in tensile strength for all composites. As with many particulate filled polymer matrix composites, the effects of defects such as
286
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
Fig. 8 – (a) Storage modulus after aging in SBF; (b) tan δ after aging in SBF. n ¼ 5 (except: PLG4B; n¼ 4 at 1 day, 3 at 3 days, 2 thereafter. PLGIB: n ¼3 at 7 days and thereafter). Mean71s.d.
voids, poor wetting at the particle–matrix interface, and cracks initiating from sharp particulates are exacerbated in the tensile mode, contributing to the reduced strengths observed in comparison to the compressive mode. Surface functionalization to improve bonding may improve mechanical properties in the future. The DMA data suggests that at physiological temperature PLG4B had inferior mechanical properties to PLGA alone. Water uptake increased significantly in the presence of all fillers, particularly with the bioactive glasses. Ceramic fillers in PLA increase water absorption (Bleach et al., 2001), this also occurs when HA is combined with hydrophobic polymers (Padilla et al., 2002). Some of the findings from the aging study can be explained due to the enhanced ingress of water facilitated by the presence of filler particles, which are generally high surface energy materials, and poor interfaces which may have acted as capillaries. Osmotic effects may also be present, due to dissolution of the fillers, particularly the more soluble bioactive glasses. Although the fillers served to increase the stiffness of the PLGA initially, after approximately 7 days in SBF this effect was lost (o1 day in the case of
45S5 BG filled composites), due to fluid ingress and plasticization of the matrix. The composite materials with more soluble fillers and lower molecular weight polymer matrices underwent a more rapid loss in stiffness. The effects of heat and shear during manufacture of the polymer specimens as the primary cause of degradation may be ruled out since it was only the bioactive glass fillers that caused significant degradation. This suggests that there is a chemical interaction between certain silica-containing or silica doped phosphate based glass formulations with hydrolysable polyesters (Blaker et al., 2005a; Shah Mohammadi et al., 2010). Similar findings have been found for Bioglass 45S5s reinforced poly(DLlactide) (PDLLA) composites (Blaker et al., 2005a, 2010). In previous work on PDLLA (Blaker et al., 2010) the inclusion of Bioglass 45S5s under elevated temperatures resulted in degradation of the matrix, leading to reduction in mechanical properties and matrix molecular weight. Bioglass 45S5s has a high-energy surface (Blaker et al., 2005b) and will adsorb water to give SiOH- at its surface. Its inclusion into poly(α-hydroxyesters) will result in polymer hydrolysis, with an increased reaction rate at elevated temperatures and increased pH (base catalysis of the ester). Silica-doped phosphate-based glasses
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
287
Fig. 9 – SEM micrographs of the failure surfaces of tensile PLG and composite specimens, tested at ambient temperature. (a, b) PLG, (c–e) PLGC, (f–h) PLGH, (i–k) PLG4B, (l–n) PLGIB. Scale bar: left images 900 μm, centre images 150 μm, right images 30 μm.
288
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
have also demonstrated a degradative effect on compression moulded polycaprolactone matrices (Shah Mohammadi et al., 2010). In the development of bone analogous materials, a composite of bioactive particles in a polymer matrix should have improved mechanical properties, provided that there is good adhesion between the filler and matrix. This can be achieved by mechanical interlocking, which can be dependent on the size and surface roughness of the particles, or by a chemical bond as achieved through surface treatment. A roughened bioactive glass surface may improve interfacial interaction: sol–gel produced glasses have porous surfaces, unlike the smooth surface of melt-derived glasses as tested here (Sepulveda et al., 2002).
5.
Conclusions
This study has enabled the direct comparison between a poly(αhydroxyester), 95/5 poly(L-lactide-co-glycolide) (PLGA), and PLGA matrix composites filled with four different particulate fillers produced with a consistent processing and testing regime. Composites processed with calcium carbonate or hydroxylapatite underwent limited polymer degradation, yet significantly increased the modulus of the polymer; whereas bioactive glass fillers resulted in significant polymer degradation, with reduced thermo-mechanical properties due to the lowered molar mass of the PLGA matrix and poor matrix–filler interface bond. Shortterm aging studies also suggest that the stiffening effect of these composites is reduced markedly after a short period in SBF due to fluid ingress and plasticization of the matrix. The bioactive glass filled composites exhibited the poorest properties due to the premature degradation of the PLGA matrix during processing; these composites should be processed via lower temperature routes to reduce such degradation. The evidence suggests that the BG fillers assessed herein, manufactured using these methods, have limited utility and are not suitable for strengthening the polymers for the use of bone replacement. PLGC or PLGH would be the preferable fillers based on this data; further assessment of bioactivity (e.g. cell culture) would allow us to discern the most appropriate material of the systems investigated here for use in bone replacement applications.
Acknowledgements RLS was supported by an EPSRC doctoral training grant, the Rosetrees Trust, the University of London Central Research Fund, the Institution of Mechanical Engineers, and the Society of Orthopaedic Medicine. The authors are grateful to Steven Lamoriniere for making the moulds used in the forming of the composites. The authors have no conflicts of interest.
Appendix A.
Supporting information
Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jmbbm. 2015.06.008.
r e f e r e n c e s
Blaker, J.J., Maquet, V., Je´roˆme, R., Boccaccini, A.R., Nazhat, S.N., 2005a. Mechanical properties of highly porous PDLLA/ Bioglasss composite foams as scaffolds for bone tissue engineering. Acta Biomater. 1 (6), 643–652. Blaker, J.J., Maquet, V., Boccaccini, A.R., Jerome, R., Bismarck, A., 2005b. Wetting of bioactive glass surfaces by poly(alphahydroxyacid) melts: interaction between Bioglasss and biodegradable polymers. E-Polymers 5 (1), 248–260. Blaker, J.J., Bismark, A., Boccaccini, A.R., Young, A.M., Nazhat, S. N., 2010. Premature degradation of poly(a-hydroxyesters) during thermal processing of Bioglasss-containing composites. Acta Biomater. 6 (3), 756–762. Blaker, J.J., Nazhat, S.N., Maquet, V., Boccaccini, A.R., 2011. Longterm in vitro degradation of PDLLA/Bioglasss bone scaffolds in acellular simulated body fluid. Acta Biomater. 7 (2), 829–840. Bleach, N.C., Tanner, K.E., Kelloma¨ki, M., To¨rma¨la¨, P., 2001. Effect of filler type on the mechanical properties of self-reinforced polylactide–calcium phosphate composites. J. Mater. Sci. Mater. Med. 12 (10–12), 911–915. Bronzino, J.D., 1995. In: The Biomedical Engineering Handbook. CRC Press, USA. Bleach, N.C., Nazhat, S.N., Tanner, K.E., Kelloma¨ki, M., To¨rma¨la¨, P., 2002. Effect of filler content on mechanical and dynamic mechanical properties of particulate biphasic calcium phosphate–polylactide composites. Biomaterials 23 (7), 1579–1585. Boccaccini, A.R., Blaker, J.J., 2005. Bioactive composite materials for tissue engineering scaffolds. Expert Rev. Med. Devices 2 (3), 303–317. Boccaccini, A.R., Roether, J.A., Hench, L.L., Maquet, V., Je´roˆme, R., 2002. A composites approach to tissue engineering. In: Lin, H. T., Singh, M., (Eds), Ceramic Engineering & Science Proceedings of the 26th Annual Conference on Composites, Advanced Ceramics, Materials and Structures: B, Florida, 23 (4), pp. 805–816. Cao, W., Hench, L.L., 1996. Bioactive materials. Ceram. Int. 22 (6), 493–507. Domb, A.J., Kost, J., Wiseman, D.M., 1997. Handbook of biodegradable polymers. In: Florence, AT., Gregoriadis, G. (Eds.), Drug Targeting and Delivery. Harwood Academic Publishers, The Netherlands. Elgayar, I., Aliev, A.E., Boccaccini, A.R., Hill, RG., 2005. Structural analysis of bioactive glasses. J. Non-Cryst. Solids 351 (2), 173–183. Engelberg, I., Kohn, J., 1991. Physico-mechanical properties of degradable polymers in medical applications: a comparative study. Biomaterials 12 (3), 292–304. Fennel, B., Hill, RG., 2001. The influence of poly(acrylic acid) molar mass and concentration on the properties of polyalkenoate cements. J. Mater. Sci. 36 (21), 5193–5202. Hench, L.L., 1991. Bioceramics: from concept to clinic. J. Am. Ceram. Soc. 74 (7), 1487–1510. Ignatius, A.A., Wolf, S., Augat, P., Claes, L.E., 2001. Composites made of rapidly resorbable ceramics and poly(lactide) show adequate mechanical properties for use as bone substitute materials. J. Biomed. Mater. Res. 57 (1), 126–131. Kikuchi, M., Koyama, Y., Takakuda, K., Miyairi, H., Shirahama, N., Tanaka, J., 2002. In vitro change in strength of β-tricalcium phosphate/copolymersised poly-L-lactide composites and their application for guided bone regeneration. J. Biomed. Mater. Res. 62 (2), 265–272. Kokubo, T., Kushitani, H., Sakka, S., Kitsugi, T., Yamamuro, T., 1990. Solutions able to reproduce in vivo surface-structure changes in bioactive glass–ceramic A-W. J. Biomed. Mater. Res. 24 (6), 721–734.
journal of the mechanical behavior of biomedical materials 50 (2015) 277 –289
Middleton, J.C., Tipton, A.J., 2000. Synthetic biodegradable polymers as orthopedic devices. Biomaterials 21 (23), 2335–2346. Nazhat, S.N., Joseph, R., Wang, M., Smith, R., Tanner, K.E., Bonfield, W., 2000. Dynamic mechanical characterization of hydroxyapatite reinforced polyethylene: effect of particle size. J. Mater. Sci. Mater. Med. 11 (10), 621–628. Nazhat, S.N., Kelloma¨ki, M., To¨rma¨la¨, P., Tanner, K.E., Bonfield, W., 2001. Dynamic mechanical characterization of biodegradable composites of hydroxyapatite and polylactides. J. Biomed. Mater. Res. 58 (4), 335–343. Padilla, S., del Real, R.P., Vallet-Regı´, M., 2002. In vitro release of gentamicin from OHAp/PEMA/PMMA samples. J. Control. Release 83 (3), 343–352. Pijpers, T.F.J., Mathot, V.B.F., Goderis, B., Scherrenberg, RL., van der Vegte, E.W., 2002. High-speed calorimetry for the study of the kinetics of (de)vitrification, crystallization, and melting of macromolecules. Macromolecules 35 (9), 3601–3613. Rezwan, K., Chen, Q.Z., Blaker, J.J., Boccaccini, A.R., 2006. Biodegradable and bioactive porous polymer/inorganic composite scaffolds for bone tissue engineering. Biomaterials 27 (18), 3413–3431. Rich, J., Jaakkola, T., Tirri, T., Na¨rhi, T., Yli-Urpo, A., Seppa¨la¨, J., 2002. In vitro evaluation of poly(ε-caprolactone-co-DL-lactide)/ bioactive glass composites. Biomaterials 23 (10), 2143–2150. Schiller, C., Epple, M., 2003. Carbonated calcium phosphates are suitable pH-stabilising fillers for degradable polyesters. Biomaterials 24 (12), 2037–2043.
289
Sepulveda, P., Jones, J.R., Hench, L.L., 2002. In vitro dissolution of melt-derived 45S5 and sol–gel derived 58S bioactive glasses. J. Biomed. Mater. Res. 61 (2), 301–311. Shah Mohammadi, M., Ahmed, I., Marelli, B., Rudd, C., Bureau, M. N., Nazhat, SN., 2010. Modulation of polycaprolactone composite properties through incorporation of mixed phosphate glass formulations. Acta Biomater. 6 (8), 3157–3168. Shikinami, Y., Okuno, M., 1999. Bioresorbable devices made of forged composites of hydroxyapatite (HA) particles and polyL-lactide (PLLA): Part 1. Basic characteristics. Biomaterials 20 (9), 859–877. Tsunoda, M., 2003. Degradation of poly(DL-lactic acid-co-glycolic acid) containing calcium carbonate and hydroxyapatite fillers – effect of size and shape of the fillers. Dent. Mater. J. 22 (3), 371–382. Verheyen, C.C.P.M., de Wijn, J.R., van Blitterswijk, C.A., de Groot, K., 1992. Evaluation of hydroxylapatite/poly(L-lactide) composites: mechanical behaviour. J. Biomed. Mater. Res. 26 (10), 1277–1296. Wang, M., 2003. Developing bioactive composite materials for tissue replacement. Biomaterials 24 (13), 2133–2151. Wang, M., Hench, L.L., Bonfield, W., 1998. Bioglasss/high density polyethylene composite for soft tissue applications: preparation and evaluation. J. Biomed. Mater. Res. 42 (4), 577–586.
