Acta Biomaterialia 6 (2010) 756762

Contents lists available at ScienceDirect

Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actabiomat

Premature degradation of poly(a-hydroxyesters) during thermal processing of Bioglass-containing composites

Jonny J. Blaker a, Alexander Bismarck a, Aldo R. Boccaccini b, Anne M. Young c, Showan N. Nazhat d,*
a

Department of Chemical Engineering, Polymer and Composite Engineering (PaCE) Group, Imperial College London, London SW7 2AZ, UK Department of Materials, Imperial College London, London SW7 2BP, UK c Division of Biomaterials and Tissue Engineering, Eastman Dental Institute, University College London, 256 Grays Inn Road, London WC1X 8LD, UK d Department of Mining and Materials Engineering, McGill University, MH Wong Building, 3610 University Street, Montreal, Que., Canada H3A 2B2
b

a r t i c l e

i n f o

a b s t r a c t
Bioactive, biodegradable composites are increasingly being explored as bone replacement materials and as scaffolds for tissue engineering. Their properties are not only dependent on the properties of the ller and matrix, but are also determined by their interaction. This study investigated the effect on poly(D,L-lactide) (PDLLA) matrix when processed at high-temperatures in the presence of Bioglass particulate ller. Composites with different ller contents were compounded at elevated temperatures by co-extrusion followed by compression moulding and compared with composites of similar composition prepared by thermally induced phase separation (TIPS), a low-temperature processing route. It was found that the inclusion of Bioglass in PDLLA under elevated temperatures resulted in the degradation of the matrix, leading to a reduction in the mechanical properties of the composites and in the molecular weight of the matrix. Attenuated total reection Fourier transform infrared (ATR-FTIR) spectroscopy showed the presence of a peak at 1600 cm1 in the composite material, particularly when processed at elevated temperatures, whereas no peak at this wavelength was discernible for the pure PDLLA. Furthermore, timebased ATR-FTIR spectra taken at elevated temperatures on the TIPS-processed composites showed an increase in the intensity of the peak at 1600 cm1 and a concomitant reduction of the C@O stretch peak at 1745 cm1 with time. This suggested the formation of a carboxylate salt by-products as a consequence of a reaction at the interface between the Bioglass ller and the PDLLA matrix. Therefore, the results conrmed that this degradation was not solely due to shear effects during the extrusion process. This work thereby supports the assertion that, in the presence of Bioglass ller particles, poly(a-hydroxyester)-based composites should not be processed at elevated temperatures. 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Article history: Received 18 May 2009 Received in revised form 6 August 2009 Accepted 11 August 2009 Available online 14 August 2009 Keywords: Bioactive glass Silica Bioglass Scaffolds Biodegradable composite

1. Introduction It is desirable to reinforce biodegradable polymer matrices with stiff, bioactive llers to produce composites of tailored mechanical, surface and chemical properties for potential biomedical applications, such as in the augmentation or regeneration of bone [1]. These reinforcements may be in non-porous bulk polymers for use as xation plates and screws where high stiffness and torsional strength are required [2], or in porous polymer scaffolds for tissue engineering where polymers alone may be too weak to meet the demands of load bearing sites [3]. The inclusion of inorganic phases of tricalcium phosphate [4], hydroxyapatite (HA) [57] or bioactive glasses [810] into degradable poly(a-hydroxyester)-based matrices, e.g. poly(lactic acid), poly(glycolic acid) and their copolymers, presents a convenient way to augment mechanical properties and

* Corresponding author. Tel.: +1 514 398 5524; fax: +1 514 398 4492. E-mail address: showan.nazhat@mcgill.ca (S.N. Nazhat).

enhance bioactivity. Moreover, it can potentially provide a means for controlling the degradation kinetics (in vitro or in vivo) of these polymers by neutralizing acidic degradation products [11,12]. In vitro degradation has been retarded by incorporating granules of coral, due to the buffering effect of CaCO3 neutralizing the carboxyl end groups [13]. However, basic compounds have also been known to catalyse ester linkage scission and increase degradation rates of poly(a-hydroxyesters) [14]. A balance between neutralizing the acidic degradation products during degradation and causing base catalysis of such polymers is required. Filler properties, volume fraction, particle size, morphology, dispersion and agglomeration all have a marked effect on composite properties. The processing conditions themselves may inuence interfacial adhesion and induce voids affecting composite properties. Conversely, in some instances a poor interface may be desirable, e.g. in dense composites of poly(a-hydroxyesters), to facilitate the ingress of water and diffusion of polymer degradation products, thereby preventing autocatalysis mediated degradation commonly associated with these polymers [9].

1742-7061/$ - see front matter 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2009.08.020

J.J. Blaker et al. / Acta Biomaterialia 6 (2010) 756762

757

Bioactive composites are commonly produced in the bulk state by mixing the component phases, compounding (at elevated temperatures), followed by elevated temperature injection or compression moulding and forging for the fabrication of plates, rods and screws [2,15]. Twin-screw extrusion is widely used as it induces a uniform particle distribution by exerting high shear forces. Porous composite scaffolds, on the other hand, are produced using a variety of high-temperature techniques, such as moulding in the presence of leachable particulates [16], the sintering of composite microspheres [17,18] and high-temperature supercritical CO2 processing or gas foaming [4]. They can also be produced through lowtemperature methods such as thermally induced phase separation (TIPS) [19], combined solvent casting and porogen leaching [5], solid freeform methods [20] and certain gas foaming methods [21,22]. The processing and fabrication of such scaffolds has been reviewed elsewhere [1,2]. Bioglass 45S5 is a proven bioactive material consisting of a surface reactive silicate which forms a layer of carbonated HA when in contact with biological uids. It is osteoconductive and osteoinductive [23]. Therefore, the use of Bioglass particles as reinforcing agents within polymer matrices is attractive as they are also expected to increase the mechanical properties of the composite. Wang et al. [15] reported an increase in stiffness and microhardness with ller volume fraction in Bioglass particulate-reinforced high-density polyethylene composites, which were produced through melt blending, compounding, powdering and compression moulding. However, this positive reinforcing effect of the Bioglass llers was not replicated in composites of poly(a-hydroxyesters) [8]. In one study, Niemel et al. [8] produced bioactive glass (with 53 wt.% SiO2) poly(D,L-lactide) composites by melt extrusion and demonstrated that the addition of the ller reduced the modulus and strength of the composite. In a further investigation, the mechanical properties of wet specimens of the same materials were assessed in vitro and in vivo [24]. The ndings were again consistent with increasing Bioglass particle inclusion reducing mechanical competency. The incorporation of small particles of a bioactive glass (53 wt.% SiO2) into poly(e-caprolactone-co-DL-lactide 96/4) through melt blending and compression moulding has also been shown to accelerate degradation [25]; the presence of the glass or processing could be responsible for the degradation. In contrast, other studies have shown that Bioglass addition to poly(a-hydroxyesters) using low-temperature processing may circumvent these problems [10,26]. Given the ndings highlighted above, it is hypothesized that the inclusion of Bioglass into poly(a-hydroxyesters) at elevated temperature leads to a reaction at the llermatrix interface that may degrade the polymer, thus resulting in the reduction of the mechanical properties of the composites. Therefore, this work investigated the effect of both the processing route and Bioglass content on the properties of poly(D,L-lactide) (PDLLA) matrix composites. BioglassPDLLA composites were produced using blending, high-temperature extrusion and compression moulding. These were assessed in terms of their structural, morphological and mechanical properties, and compared to the properties of composite foams of the same composition produced at low-temperature by TIPS to assess degradation due to processing.

Bioglass particles of chemical composition (in wt.%) SiO2, 45; CaO, 24.5; P2O5, 6; NaO2, 24.5 [27] were incorporated as the ller phase in the composites. The Bioglass particles were of particle size <20 lm, with a mean particle size of 4 lm. The glass density was 2.825 g cm3 and the surface area of the powder used was 2.7 m2 g1. The morphology of the 45S5 Bioglass particles was non-spherical, irregular and angular [28]. 2.2. High- and low-temperature processing of PDLLA and composite samples PDLLA and composite samples (lled with 2, 15 and 23 vol.% of Bioglass) of dimensions 3 5 6 mm3 were produced by coextrusion and subsequent compression moulding. Details of the composites and the equivalent Bioglass particulate weight percentages are presented in Table 1. Prior to extrusion, PDLLA and composite mixtures were cryogenically milled and blended using liquid nitrogen to produce a ne powder. A twin-screw DSM 5 cm3 microextruder (DSM Research Products, The Netherlands) was lled (to give a sample volume of 4.5 cm3) over a 5 min period at 150 C using a screw speed of 40 rpm. Samples were extruded under a nitrogen blanket. Previously, thermogravimetric analysis (TGA) data revealed the onset of thermal degradation of PDLLA in air to be 320 C [26]. Compression moulding was conducted at 150 C and 120 MPa, where the pressure was maintained until the polymer cooled to room temperature (taking approximately 30 min). Five samples per shot were obtained from the mould, and were carefully pressed out of the mould using an aluminium die of the same dimensions. Neat PDLLA and composite foams, lled with 2 and 15 vol.% Bioglass, were prepared by thermally induced phase separation (TIPS) and subsequent solvent sublimation, which has been described in detail elsewhere [19]. Briey, PDLLA was dissolved in dimethyl carbonate at a polymer weight to solvent volume ratio of 5% (w/v). The mixture was stirred overnight to obtain a homogeneous polymer solution. Given quantities of Bioglass particles were added to the polymer solution, which was then transferred to a lyophilization ask and sonicated for 15 min to achieve a homogeneous distribution of Bioglass particles in the solution and to break up agglomerates. The ask was then immersed in liquid nitrogen and maintained at 196 C for 2 h. The frozen mixture was placed under vacuum (102 Torr) and transferred to an ethylene glycol bath, which was maintained at 10 C. The solvent was sublimed at 10 C for 48 h and then at 0 C for a further 48 h. Finally, the resulting foam monolith (of 90 mm diameter and 10 mm thickness) was completely dried at room temperature in a vacuum oven until reaching constant weight. 2.3. Scanning electron microscopy (SEM) Samples were freeze-fractured using liquid nitrogen and a sharp blow with a razor blade to enable examination of interior cross-sections and fracture surfaces. The exterior and interior of

Table 1 Composition of BioglassPDLLA composites prepared by co-extrusion and compression moulding. Volume of glass (%) Equivalent weight of glass (%) Bioglass mass (g) Polymer mass (g) Theoretical density (g cm3) 0 0 0 5.76 1.29 2.18a 4.76 0.28 5.55 1.50 15.14b 28.57 1.92 4.81 1.62 22.92c 40 2.91 4.37 1.89

2. Experimental 2.1. Materials Racemic PDLLA with an inherent viscosity of 1.62 dl g1 and density of 1.26 g cm3 was obtained from Purac biochem (Gorinchem, The Netherlands) and used as received. Melt-derived 45S5

Theoretical density was calculated according to the rule of mixtures. Note: a, b and c are nominally referred in the text as 2, 15 and 23 vol.%.

758

J.J. Blaker et al. / Acta Biomaterialia 6 (2010) 756762

extruded and compression moulded samples were examined to assess the inuence of Bioglass particle inclusion on morphology and microstructure. Samples were gold coated for 120 s under a current of 20 mA before SEM examination using an accelerating voltage of 20 kV in a JEOL 5610LV SEM (JEOL, USA). 2.4. Mechanical analysis of compression moulded samples Compression testing was performed to BS EN ISO 640 using a Zwick/Roell Z010. Specimens were strained at a rate of 1 mm min1 to a maximum of 30% strain. The compressive modulus (Ec) was determined between strain values of 0.05 and 0.25%. The compressive strength (rm) was recorded at the maximum stress sustained by the sample and the nominal compressive strain (ecm) was determined at this value. The compression strength at yield (ry) was determined at the rst stress at which an increase in strain occurs without an increase in stress. Nominal compressive yield strain (ecy) is dened here as the strain corresponding to the compressive yield stress. Samples were tested in quintuplet and statistical analysis was performed to conrm signicant differences in the mechanical properties when compared to the pure PDLLA control using the Students t-test to determine p-values at a signicance level of 0.05. 2.5. Size-exclusion chromatography (SEC) The weight-average molecular weight (Mw) of the as-received PDLLA, and processed PDLLA/Bioglass-lled composites produced by extrusion, compression moulding and thermally induced phase separation, was determined by SEC using a HewlettPackard HP1090 SEC apparatus equipped with three Ultrastyragel columns (102105 ). Tetrahydrofuran was used as an eluent (ow rate: 1 ml min1), Bioglass particles were ltered prior to analysis, and calibration was performed using polystyrene standards (Polymer Laboratories Ltd., Shropshire, UK). 2.6. Fourier transform infrared spectroscopy (FTIR) Attenuated total reectance (ATR)-FTIR spectra of PDLLA, 45S5 Bioglass and composites as extruded, compression moulded and TIPS foams were obtained at room temperature using a Golden Gate Single Reection Diamond ATR attachment (Graseby Specac, UK) in a series 2000 FTIR spectrometer (PerkinElmer Instruments). The resolution was set to 4 cm1, the wavenumber ranged between 4000 and 500 cm1 and four scans were performed. FTIR studies were additionally performed on PDLLA and PDLLA with 15 vol.% Bioglass TIPS foam as they were subsequently heated to 170 C and held at this temperature to assess any chemical changes that may occur, without the inuence of shear, as would be exerted in the twin screw extruder. Spectrum Timebase v2.0 (Perkin Elmer) was used to resolve changes in spectra occurring as a function of time for samples placed onto the diamond ATR at 170 C for up to 9 h (spectra were recorded every 8 s), using a wavenumber range between 4000 and 700 cm1. One advantage of performing kinetic experiments directly on the diamond of the ATR is that, once the sample has reached the experiment temperature, sample contact should remain constant and changes in absorbance at any wavenumber can then be taken as an indication of the level of chemical reaction. If, upon subtracting one spectrum from that at a later time, the same difference spectrum is always observed, an identical chemical process is likely to be occurring at all times without the formation of reactive intermediates [29,30]. If ut and uf are the fractions of PDLLA reacted at time t and nally, respectively, then, assuming a Beer Lambert-type law:

A0 At A0 A1 A0 Af /f A0 A1 /t

1 2

where At, Ao, Af and A1 are the absorbance at time t, initially, nally and with complete reaction, respectively. Unfortunately these terms cannot be calculated using such expressions unless A1 is known. Dividing Eq. (1) by Eq. (2), however, gives the reaction extent (n), which can be determined as:

/t A0 At /f A0 Af

This expression is useful because the reaction extent is the same irrespective of whichever wavenumber is used. If it is known that only the reactive component and none of the products or other components in the composites absorbs at a given wavenumber then A1 is zero and from Eqs. (1) and (3), it is possible to write:

/t n/f

A0 At A0

This expression also assumes that background scattering on the FTIR spectrum is negligible or can be estimated. Both Eqs. (3) and (4) are used in the following to quantify reaction rates at 170 C.

3. Results and discussion 3.1. Observations of extruded and compression moulded samples During the extrusion of the neat polymer and composite samples, some bubble formation was observed (Fig. 1a and b). The extruded materials became more amber in colour with increasing Bioglass content. This suggested chemical changes were occurring despite the operating temperature of 150 C being signicantly below the degradation temperature of PDLLA (320 C by TGA) [26]. A sweet smelling odour was also noted during processing of the composites, which was not detected with the neat polymer. Fig. 1c shows the 15 vol.% BioglassPDLLA composite at high magnication, demonstrating that, although there was a homogeneous distribution of the Bioglass particles within PDLLA, there was a poor interface and voids were present between the ller and the matrix. Compression moulded composites were observed to embrittle progressively with increasing Bioglass content. SEM examination of freeze-fractured compression moulded samples conrmed the homogeneous distribution of Bioglass particles within the PDLLA and the presence of voids at the interface region between the particles and the matrix (Fig. 2a). The presence of interior and exterior porosity was also reported by Niemel et al. [8] in their melt-extruded bioactive glasspolylactide composites. SEM analysis of freeze-fractured samples revealed brittle fracture surfaces, e.g. surfaces characterized by smooth planes and no induction of plastic deformation, which was particularly evident for the composites at the higher Bioglass volume fractions (Fig. 2b). 3.2. Effect of Bioglass on the mechanical properties of the composites The compressive mechanical properties of the neat and composite samples are summarized in Table 2. All parameters demonstrated a reduction in values relative to pure PDLLA. Statistical analysis conrmed that only the 15 vol.% Bioglass composite demonstrated statistically signicant differences in the compressive modulus when compared to that of the pure PDLLA. All other

J.J. Blaker et al. / Acta Biomaterialia 6 (2010) 756762

759

Fig. 2. SEM of compression moulded composites: (a) the as-processed compression moulded surface of 2 vol.% BioglassPDLLA at high magnication demonstrating a poor interface and void formation between the Bioglass particles and the PDLLA matrix; (b) freeze-fractured sample of 23 vol.% BioglassPDLLA at high magnication. The sample fractured in a brittle manner and little evidence of bubble formation was observed for this sample.

Fig. 1. SEM of freeze-fractured surfaces of high-temperature extrusion processed composites: (a) 2 vol.% BioglassPDLLA; (b) 23 vol.% BioglassPDLLA at low and high magnication (inset) (note the presence of bubbles throughout the material); (c) 15 vol.% BioglassPDLLA at high magnication. Note the apparent poor interface and voids between the Bioglass (angular and irregular) particles and the PDLLA matrix.

The nding that the mechanical properties of compression moulded composites decrease with increasing Bioglass content in PDLLA aligns with similar ndings reported for other bioactive glass/poly(a-hydroxyesters) composites [8,24,25]. This result is in contrast to other melt processed bioactive lled poly(a-hydroxyester) matrix composites, e.g. HA [4,6], tricalcium phosphates [4] and phosphate-based glass particles [31]. It has been previously reported that the lack of improvement in mechanical properties of Bioglass-lled composites was due to a poor glass/polymer interface, the presence of porosity or the fact that Bioglass has a lower modulus than other llers, e.g. 35 vs. 80 GPa for HA [32].

3.3. Effect of Bioglass on the molecular weight of PDLLA compressive mechanical properties of the composites with the various Bioglass inclusions were signicantly reduced when compared to those of neat dense PDLLA. Size-exclusion chromatography was carried out to investigate the effect of Bioglass addition on the molecular weight of the PDLLA matrix. Fig. 3 shows the effects of high-temperature extru-

Table 2 Compressive mechanical properties of dense PDLLA and composites lled with Bioglass fabricated by extrusion and subsequent compression moulding. Bioglass ller (vol.%) 0 2 15 23 Compressive modulus Ec (GPa) 1.97 0.1 1.86 0.2 (p = 0.3033) 1.60 0.1 **(p = 0.0004) 1.65 0.4 (p = 0.1209) Maximum compressive strength rm (MPa) 86.4 6 72.6 9 *(p = 0.0214) 46.8 3 **(p = 0.0001) 46.3 13 **(p = 0.0002) Nominal compressive strain ecm (%) at rm 5.20 0.4 4.56 0.2 *(p = 0.0126) 4.17 0.6 *(p = 0.0127) 3.13 0.2 **(p = 0.0001) Compression stress at yield ry (MPa) 65.2 5 54.0 6 *(p = 0.0125) 33.0 3 **(p = 0.0001) 35.3 9 **(p = 0.0002) Nominal compressive strain

ecy (%) at ry
3.40 0.3 2.90 0.1 *(p = 0.0077) 1.90 0.4 **(p = 0.0002) 2.10 0.1 **(p = 0.0001)

The values are presented in terms of means standard deviation (n = 5). Statistically signicant differences in the mechanical properties of the different composites when compared to those of the dense PDLLA are indicated by , as determined by the p-values. ** means very signicant.

760

J.J. Blaker et al. / Acta Biomaterialia 6 (2010) 756762

100

a
0.4

80

polymer Mw [% of intial]

Absorbance [arbitrary]

TIPS foams post-extrusion post-compression moulded

Compression moulded 15 vol.% TIPS 15 vol.% PDLLA (PURAC) 45S5 Bioglass

0.3
1600 cm
-1

60

0.2

40

0.1

20
0.0

0 0 5 10 15 20

2000

1750

1500

1250

1000
-1

750

500

Bioglass filler content [vol. %]

Fig. 3. Change in the Mw of the PDLLA and BioglassPDLLA composites produced at high-temperature extrusion (D) and compression moulding () and compared to TIPS-produced composites (s). Data were normalized as a percentage of the original polymer Mw.
-1

Wavenumber [cm ]

b
peak wavenumbers ratio 1600:1745 cm

0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00
Extruded Compression moulded TIPS

sion and subsequent compression moulding on the Mw of PDLLA as a function of Bioglass content and in comparison with composites produced using TIPS. The as-received PDLLA had an Mw of 260 kg mol1 and a polydispersity of 2.2. The combination of high-temperature extrusion and an increasing Bioglass content resulted in a dramatic reduction in the Mw of PDLLA. Whereas only a slight reduction in Mw was observed for the TIPS-produced composites (255 kg mol1 for the pure PDLLA TIPS foam and 244 kg mol1 for the 15 vol.% Bioglass-lled TIPS foam), this reduction was not signicant in contrast to the high-temperature processing route. Studies on melt-extruded bioactive glass-poly(e-caprolactoneco-DL-lactide 96/4) composites have also shown a reduction in Mw of the matrix with the inclusion of bioactive glass particles of <45 lm [25]. It was reported that during hydrolytic degradation the Mw decrease was more rapid in the presence of bioactive glass than in the neat polymer matrix. In contrast, such a reduction in the polymer Mw was not demonstrated when Bioglass was added to high-density polyethylene, also produced under elevated temperature extrusion and compression moulding [15]. 3.4. Effect of Bioglass on PDLLA at elevated temperatures The FTIR spectra of as-received PDLLA, 45S5 Bioglass and composites lled with 15 vol.% Bioglass produced by high-temperature extrusion and subsequent compression moulding, or TIPS are shown in Fig. 4a. Above 2000 cm1 no signicant peak variations with time or composition were observed (data not shown). The composite samples did, however, display a weak peak at 1600 cm1 that was absent in the polymer and Bioglass spectra. This peak was most apparent for composites processed at hightemperatures and higher Bioglass content. The ratio of the height of this peak divided by that at 1745 cm1 increased systematically with increasing Bioglass content (Fig. 4b). More marked FTIR spectral changes were observed for the BioglassPDLLA composites that were subjected to increased exposure time at elevated temperatures. As an example, the FTIR spectra of TIPS-produced 15 vol.% BioglassPDLLA foam held at 170 C for up to 7 h is provided in Fig. 5a. These spectra exhibit a signicant reduction in the C@O stretch peak at 1745 cm1 and an increase in intensity of the 1600 cm1 peak with time. The

10

15

20

Bioglass as filler in PDLLA [vol. %]

Fig. 4. (a) FTIR spectra of as-received raw materials of PDLLA (Purac) and 45S5 Bioglass, as well as compression moulded, and TIPS composites lled with 15 vol.% Bioglass for comparison of low and high-temperature processing techniques, respectively. (b) Effect of processing technique on the ratio of peak absorbance at 1600 and 1745 cm1 as a function of vol.% Bioglass-ller.

CAH peaks at 1447 and 1370 cm1 additionally underwent a slight reduction and were partially replaced by peaks at lower wavenumbers. Furthermore, there was a shift in the CAO/CAC stretch peaks at 1179 and 1075 cm1 to lower wavenumbers. Difference spectra obtained by subtracting earlier from later time spectra had several peaks and troughs (Fig. 5b). Their positions and relative intensities were time independent, indicating the same chemical change at all times. In comparison, no change in the spectra was observed when the neat polymer was exposed to 170 C for 2 h. The reaction extent calculated using the absorbance at 1600 cm1 and Eq. (3) is given in Fig. 5c. From the gradient of this plot it can be seen that the reaction rate begins to slow substantially after 15,000 s. At the end of the experiment (25,000 s), however, the peak at 1745 cm1 (C@O) had only declined by 25%. From Fig. 5c the fractional absorbance change at 1745 cm1 would be expected to decrease to almost zero if all the PDLLA is decomposed. The fraction of unreacted PDLLA, calculated using the 1745 cm1 peak and Eq. (4), follows the same trend as the reaction extent curve as would be expected if the product has no absorbance at 1745 cm1 (compare Eqs. (3) and (4)). It is possible that by 25,000 s the reaction may have reached equilibrium or that only the PDLLA in close proximity to the ller particles may be able to react.

J.J. Blaker et al. / Acta Biomaterialia 6 (2010) 756762

761

a
0.5
-1 1745 cm

Time at 170 C
1059s 5273s 10000s 15292s 25032s
-1 1595 cm

0.4

0.3

0.2

0.1

0.0 2000 1750 1500 1250 1000 750

b
0.10
PDLLA TIPS 1 to 7k s 15 vol.% TIPS 1 to 7k s 15 vol.% TIPS 1 to 25k s

0.05

0.00

-0.05

the effects of shear experienced during melt extrusion the onset of thermal degradation decreased with increasing Bioglass content. The spectral changes detected in the composites are quite different from those previously observed for degradation of PLA in phosphate-buffered solutions, biotic media or bovine serum albumin solutions [33,34], none of which exhibited an intense peak at 1600 cm1. In this study, the peak at 1600 cm1 is at a too low wavenumber or intensity for most C@O and C@C stretch peaks, respectively unless the electrons in the former can be delocalized in compounds such as carboxylate salts or b diketones. Silica-based glasses such as Bioglass consist of SiAOASi bonds, which at the surface terminate in hydroxyl groups. These, in the presence of moisture, may be dissociated to SiAO and associated instead with Na+ or Ca2+ counterions, which may catalyse polymer chain scission. The reduction in intensity of the C@O peak (at 1745 cm1) and other peaks between 1050 and 1200 cm1, due to the CAO group of the ester, provides evidence for the loss of ester groups. Generally, poly(a-hydroxyesters) degrade by random chain scission by ester hydrolysis in a process that is autocatalysed through the generation of carboxylic acid end groups. The lack of change in the spectrum around 3300 cm1 (data not shown) suggests that alcohols and carboxylic acids, which would form upon this conventional hydrolysis of an ester, are not present at high levels in the nal material. The appearance of peaks at 1600 and 1451 cm1 in the difference spectra, however, are consistent with asymmetric and symmetric stretches respectively of a CO2 (carboxylate) chemical group, which associates with Na+ or Ca2+ counterions present as network modiers in the glass. The following reaction is proposed:

Absorbance

2000

1800

1600

1400

1200
-1

1000

800

RSiO R0 CO2 R00 ! RSiOR R0 CO 2

The R00 group on the glass surface may then degrade further by a similar reaction with other SiO groups, leading to small oligomeric fragments. The sweet odour that was apparent during compression moulding suggests the loss of ester fragments via evaporation. It is also possible that the RSiOR00 group may be partially degraded by surrounding water, leading to RSiOH and small alcoholic R00 OH molecules that evaporate at the processing temperature. Bioglass has been shown to be of a high surface energy [35] and will adsorb water to give SiAOH on the surface. Its subsequent inclusion will hydrolyse the PDLLA causing chain scission of the polymer, with an increased reaction rate at elevated temperatures and increased pH (base catalysis of the ester). The extent of PDLLA degradation was reduced using a lower fraction of Bioglass due to fewer SiAOH groups being present at the surface, thus leading to a lower degree of chain scission. The extrusion process involves high shear stresses and permits interaction between the glass particles and the polymer chains in the melt, thereby increasing the likelihood of surface reactions. This is in contrast to TIPS-processing, where the polymer and Bioglass are mixed in a suitable solvent for the polymer, prior to rapid freezing, to induce the phase separation. In this case, there is less (or no) potential for the premature degradation of the polymer. The FTIR investigation carried out on the TIPS-produced materials at 170 C conrmed that interaction between Bioglass and PDLLA can increase thermal degradation even without the shear forces exerted by the extrusion process. 4. Conclusions This study has demonstrated that, in the presence of Bioglass, the high-temperature processing of PDLLA can cause signicant reduction in its molecular weight. Conversely, the processing of the Bioglass-lled poly(a-hydroxyester) foams by low-tempera-

00

Wavenumber [cm ]

c
Fractional absorbance change

0.25

1.0

0.20

0.8

0.15

0.6

0.10

0.4

0.05

0.2

0.00

5000

10000

15000

20000

0.0 25000

time [s]
Fig. 5. ATR-FTIR analyses of TIPS-produced foams of PDLLA and 15 vol.% Bioglass PDLLA composites at 170 C as a function of time. (a) FTIR spectra of 15 vol.% BioglassPDLLA TIPS foam composite up to 25,000 s. There is a marked reduction in the peak at 1745 cm1 and a concomitant increase at 1600 cm1 as a function of time. (b) Change in FTIR spectra of PDLLA TIPS foam between 1 and 7000 s, and of 15 vol.% BioglassPDLLA TIPS foam between 1 and 7000 s and 1 and 25,000 s. (c) Reaction extent (Eq. (3)) calculated using the absorbance at 1600 cm1 and fractional absorbance change at 1745 cm1 (Eq. (4)) of 15 vol.% BioglassPDLLA TIPS foam composite. Both exhibit similar trends as expected from Eq. (4) and indicate a nal reaction level of 25%.

The ndings here align with TGA results reported in a previous study [26], which showed increasing amounts of Bioglass present in PDLLA foams (produced via TIPS) reduce the onset of thermal degradation for the PDLLA from 300 to 220 or 190 C, with 2 and 15 vol.% Bioglass content. It is worth noting that without

Reaction extent

762

J.J. Blaker et al. / Acta Biomaterialia 6 (2010) 756762 [14] Makino K, Ohshima H, Kondo T. Mechanism of hydrolytic degradation of poly(L-lactide) microcapsules: effects of pH, ionic strength and buffer concentration. J Microencapsul 1986;3:20312. [15] Wang M, Hench LL, Boneld W. Bioglass/high density polyethylene composite for soft tissue applications: preparation and evaluation. J Biomed Mater Res 1998;42:57786. [16] Aho AJ, Tirri T, Kukkonen J, Strandberg N, Rich J, Seppl J, et al. Injectable bioactive glass/biodegradable polymer composite for bone and cartilage reconstruction: concept and experimental outcome with thermoplastic composites of poly(e-caprolactone-co-D,L-lactide) and bioactive glass S53P4. J Mater Sci Mater Med 2004;15:116573. [17] Lu HH, El-Amin SF, Scott KD, Laurencin CT. Three-dimensional, bioactive, biodegradable, polymer-bioactive glass composite scaffolds with improved mechanical properties support collagen synthesis and mineralization of human osteoblast-like cells in vitro. J Biomed Mater Res 2003;64:46574. [18] Yao J, Radin S, Leboy PS, Ducheyne P. The effect of bioactive glass content on synthesis and bioactivity of composite poly (lactic-co-glycolic acid)/bioactive glass substrate for tissue engineering. Biomaterials 2005;26:193543. [19] Maquet V, Boccaccini AR, Pravata L, Notingher I, Jrme R. Porous poly(ahydroxyacid)/Bioglass composite scaffolds for bone tissue engineering. I. Preparation and in vitro characterisation. Biomaterials 2004;25:418594. [20] Hutmacher DW, Sittinger M, Risbud MV. Scaffold-based tissue engineering: rationale for computer-aided design and scaffold free-form fabrication systems. Trends Biotechnol 2004;22:35462. [21] Orava E, Korventausta J, Rosenberg M, Jokinen M, Rosling A. In vitro degradation of porous poly(DL-lactide-co-glycolide) (PLGA)/bioactive glass composite foams with a polar structure. Polym Degrad Stab 2007;92:1423. [22] Howdle SM, Watson MS, Whitaker MJ, Popov VK, Davies MC, Mandel FS, Wang JD, Shakesheff KM. Supercritical uid mixing: preparation of thermally sensitive polymer composites containing bioactive materials. Chemical Communications 2001;(1):10910. [23] Hench LL. Bioceramics. J Am Ceram Soc 1998;81:170528. [24] Niiranen H, Pyhlt T, Rokkanen P, Kellomki M, Trml P. In vitro and in vivo behaviour of self-reinforced bioabsorbable polymer and self-reinforced bioabsorbable polymer/bioactive glass composites. J Biomed Mater Res 2004;69:699708. [25] Rich J, Jaakkola T, Tirri T, Nrhi T, Yli-Urpo A, Seppl J. In vitro evaluation of poly(e-caprolactone-co-DL-lactide)/bioactive glass composites. Biomaterials 2002;23:214350. [26] Blaker JJ, Maquet V, Jrme R, Boccaccini AR, Nazhat SN. Mechanical properties of highly porous PDLLA/Bioglass composite foams as scaffolds for bone tissue engineering. Acta Biomater 2005;1:64352. [27] Hench LL. Bioceramics: from concept to clinic. J Am Ceramic Soc 1991;74:1487510. [28] Sepulveda P, Jones JR, Hench LL. Characterization of melt-derived 45S5 and solgel-derived 58S bioactive glasses. J Biomed Mater Res 2001;58(6):73440. [29] Hoffman MP, Young AM, Gbureck U, Nazhat SN, Barralet JE. FTIR-monitoring of a fast setting brushite bone cement: effect of intermediate phases. J Mater Chem 2006;16:3199206. [30] Young AM, Rafeeka SA, Howlett JA. FTIR investigation of monomer polymerisation and polyacid neutralisation kinetics and mechanisms in various aesthetic dental restorative materials. Biomaterials 2004;25:82333. [31] Georgiou G, Mathieu L, Pioletti DP, Bourban P-E, Mnson J-AE, Knowles JC, et al. Polylactic acidphosphate glass composite foams as scaffolds for bone tissue engineering. J Biomed Mater Res B (Appl Biomater) 2007;80:3223. [32] Hench LL, Wilson J, editors. An introduction to bioceramics. 2nd ed.. London: World Scientic; 1999. [33] aiker E, Gmsderelioglu M, Gner A. Degradation of PLA, PLGA homo- and copolymers in the presence of serum albumin: a spectroscopic investigation. Polym Int 2000;49:72834. [34] Pamula E, Dobrzynski P, Bero M, Paluszkiewicz C. Hydrolytic degradation of porous scaffolds for tissue engineering from terpolymer of L-lactide, epsiloncaprolactone and glycolide. J Mol Struct 2005;744:55762. [35] Blaker JJ, Maquet V, Boccaccini AR, Jerome R, Bismarck A. Wetting of bioactive glass surfaces by poly(alpha-hydroxyacid) melts: interaction between Bioglass biodegradable polymers. E-polymers 2005;art. no. 023.

ture methods, such as TIPS, are not detrimental to the polymer. FTIR investigations suggested that thermal processing may give rise to carboxylate salt by-products and oligomeric fragments. This was also conrmed through FTIR studies on the TIPS composites when exposed to 170 C for prolonged periods. Reduction in mechanical properties arising with molecular weight reduction further supported the assertion that, in the presence of Bioglass ller particles, PDLLA and other poly(a-hydroxyester)-based composites should not be processed at elevated temperatures. Therefore, low-temperature processing routes of BioglassPDLLA produce attractive composites for bone tissue engineering scaffolds and bone replacement materials.

Acknowledgements Funding from the Canadian Natural Sciences and Engineering Research Council and Canadian Foundation for Innovation, as well as the UK Engineering Physical Sciences Research Council, is gratefully acknowledged.

References
[1] Rezwan K, Chen QZ, Blaker JJ, Boccaccini AR. Biodegradable and bioactive porous polymer/inorganic composite scaffolds for bone tissue engineering. Biomaterials 2006;27:341331. [2] Boccaccini AR, Blaker JJ. Bioactive composite materials for tissue engineering scaffolds. Expert Rev Med Devices 2005;2:30317. [3] Shikinami Y, Okuno M. Bioresorbable devices made of forged composites of hydroxyapatite (HA) and poly-L-lactide (PLLA). Part I. Basic characteristics. Biomaterials 1999;20:85977. [4] Mathieu LM, Mueller TL, Bourban P, Pioletti DP, Mller R, Mnson JE. Architecture and properties of anisotropic polymer composite scaffolds for bone tissue engineering. Biomaterials 2006;27:90516. [5] Thomson RC, Yaszemski MJ, Powers JM, Mikos AG. Hydroxyapatite ber reinforced poly(a-hydroxy ester) foams for bone regeneration. Biomaterials 1998;19:193543. [6] Nazhat SN, Kellomki M, Tanner KE, Torml P, Boneld W. Dynamic mechanical characterisation of biodegradable composites of hydroxyapatite and polylactides. J Biomed Mater Res B (Appl Biomater) 2001;58:33543. [7] Kothapalli CR, Shaw MT, Wei M. Biodegradable HA-PLA 3-D porous scaffolds: effect of nano-sized ller content on scaffold properties. Acta Biomater 2005;1:65362. [8] Niemel T, Niiranen H, Kellomki M, Trml P. Self-reinforced composites of bioabsorbable polymer and bioactive glass with different bioactive glass contents. Part I. Initial mechanical properties and bioactivity. Acta Biomater 2005;l:23542. [9] Li S, McCarthy S. Further investigations on the hydrolytic degradation of poly(DL-lactide). Biomaterials 1999;20:3544. [10] Zhang K, Wang Y, Hillmyer MA, Francis LF. Processing and properties of porous poly(L-lactide)/bioactive glass composites. Biomaterials 2004;25:2489500. [11] Ara M, Watanabe M, Imai Y. Effect of blending calcium compounds on hydrolytic degradation of poly (DL-lactic acid-co-glycolic acid). Biomaterials 2002;23:247983. [12] Van Der Meer SAT, De Wijn JR, Wolke JGC. The inuence of basic ller materials on the degradation of amorphous D- and L-lactide copolymer. J Mater Sci Mater Med 1996;7:35961. [13] Li S, Vert M. Hydrolytic degradation of the coral/poly(DL-lactic acid) bioresorbable material. J Biomater Sci Polym Ed 1996;7(9):81727.