Fabrication and characterization of sol–gel derived 45S5 Bioglass ceramic scaffolds

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Fabrication and characterization of sol–gel derived 45S5 Bioglass�–ceramicscaffolds

Qi-Zhi Chen a,b,⇑, George A. Thouas c

a Department of Materials Engineering, Monash University, Clayton, Victoria 3800, Australiab Division of Biological Engineering, Monash University, Clayton, Victoria 3800, Australiac Department of Zoology, The University of Melbourne, Parkville, Victoria 3010, Australia

a r t i c l e i n f o

Article history:Received 31 March 2011Received in revised form 2 June 2011Accepted 5 June 2011Available online 13 June 2011

Keywords:Bioglass�

Sol–gelScaffoldMechanical propertiesCell infiltration

a b s t r a c t

Although Bioglass� has existed for nearly half a century its ability to trigger bone formation and tuneabledegradability is vastly superior to other bioceramics, such as SiO2–CaO bioactive glasses. The sol–gel pro-cess of producing glass foams is well established for SiO2–CaO compositions, but not yet established for45S5 composites containing Na2O. In this work the sol–gel derived 45S5 Bioglass� has for the first timebeen foamed into highly porous three-dimensional scaffolds using a surfactant, combined with vigorousmechanical stirring and subsequent sintering at 1000 �C for 2 h. It was found that the mechanicalstrength of the sintered sol–gel derived Bioglass� scaffolds was significantly improved, attributable tothe small fraction of material on the pore walls. More importantly, the compressive strength of thethree-dimensional scaffolds produced by this surfactant foaming method could be predicted using Gib-son and Ashby’s closed cell model of porous networks. A comparative experiment revealed that ionrelease from the sol–gel derived Bioglass� foams was faster than that of counterparts produced by thereplication technique. In vitro evaluation using osteoblast-like cells demonstrated that the sol–gelderived 45S5 Bioglass foams supported the proliferation of viable cell populations on the surface ofthe scaffolds, although few cells were observed to migrate into the virtually closed pores within thefoams. Further work should be focused on modifications of the reaction conditions or alternative foamingtechniques to improve pore interconnection.

� 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Tissue engineering aims at the regeneration of damaged tissueto its natural state, with its primary methodology of using scaffoldspopulated with signalling molecules and cells [1,2]. In the case ofbone tissue engineering two essential features of a scaffold arehigh porosity (�90%) and appropriate mechanical rigidity [3–5].Firstly, when cells have attached to the material surfaces theremust be enough space and sufficient channels to allow for nutrientdelivery and waste removal. Secondly, the scaffold should be ableto temporarily replace the mechanical function of damaged boneuntil sufficient new bone tissue has formed.

Highly porous foams made of ceramics can be fabricated by avariety of processes [6–14]. Among them, foaming sol–gel derivedbioactive glasses made by addition of a surfactant combined withmechanical stirring produce scaffolds of high porosities. Currentlysol–gel derived bioactive glasses have more simple compositionsthan 45S5 Bioglass�, and exhibit high bioactivity due to themesoporous texture inherent in the sol–gel process [15]. In sol–gel

foams a hierarchical structure has been achieved with macro-scalepores of diameter >500 lm connected by pore windows [16]. Mostwork to date has been carried out on the 70S30C composition(70 mol.% SiO2 and 30 mol.% CaO). The mechanical strength of sol–gel derived 70S30C glass foams has been improved by increasingthe sintering temperature (the final stage of foam synthesis) [17].It has been shown that compressive strength increases from 0.36to 2.26 MPa when annealing temperatures are increased from 600to 800 �C, and that increasing the annealing temperature does nothave a great impact on the connectivity of the macropore network[17]. The improved mechanical strength of 70S30C foams has beenattributed to the formation of a crystalline phase of calcium silicateCaSiO3 [17].

One of the concerns associated with heat treating bioactiveglasses at their crystallization temperatures is the possibility ofaffecting the biodegradability of the scaffold, as a satisfactorydegradability is another essential feature of a tissue engineeringscaffold. It has been found in previous work [18] that a mechani-cally competent crystalline phase (Na2Ca2Si3O9) (formed in 45S5Bioglass� during sintering at 1000 �C) could transform to adegradable amorphous calcium phosphate when the scaffold isincubated in an aqueous solution similar to biological fluids. How-ever, no similar transformation has been reported to take place in

1742-7061/$ - see front matter � 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.actbio.2011.06.005

⇑ Corresponding author at: Department of Materials Engineering, MonashUniversity, Clayton, Victoria 3800, Australia. Tel.: +61 3 99053599; fax: +61 399054940.

E-mail address: [email protected] (Q.-Z. Chen).

Acta Biomaterialia 7 (2011) 3616–3626

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /ac tabiomat


the crystallized 70S30C composition system [19]. Hence, the pri-mary objective of this study was to produce scaffolds from sol–gel derived 45S5 Bioglass�–ceramics, based on previously reportedformulations [16], and carry out a careful investigation of their bio-activity, mechanical properties and cytocompatibility.

A further challenge in the design of bone tissue engineeringscaffolds is that a high porosity does not necessarily guarantee res-toration of the vasculature of the engineered tissue, which is one ofthe major obstacles to successfully realizing the clinical use ofin vitro engineered tissue and organ substitutes [20]. The idealscaffold, which will also promote vascularization in vivo, has notyet been determined. Indeed, the major disadvantage of most scaf-folds reported to date is that cells tend to adhere only to the outerlayer of the scaffolds [21]. This may partially explain why mostscaffolds fail to vascularize, independent of their material proper-ties [20]. Therefore, another objective of this work was to carryout an in vitro evaluation of cell infiltration of the sol–gel derived45S5 Bioglass�–ceramic scaffolds, which will be a useful indicationof in vivo tissue penetration into the scaffolds.

2. Experimental procedures

2.1. Foam synthesis

Foams were prepared using a sol–gel derived 45S5 Bioglass�

composition (Fig. 1) as described elsewhere [16,22]. Sol preparationinvolved mixing of the reagents in the following order: deionizedwater, 2 N nitric acid, tetraethyl orthosilicate (TEOS), triethyl phos-phate (TEP), sodium nitrate, and calcium nitrate (all from Sigma–Al-drich). The molar ratio of water to the rest of the chemicals (R ratio)was 10:1. A volume of 50 ml of sol was foamed by vigorous agita-tion with the addition of 0.5 ml of Teepol� (Thames Mead Ltd.)and 1.5 ml of 5 vol.% HF (a gelation catalyst). Teepol� is a detergentcontaining a low concentration mixture of anionic (15%) and non-ionic surfactants (5%). As the gelling point was approached thefoamed solution was cast into cylindrical polymethylpropylenemoulds. The samples were aged, dried, thermally stabilized at600 �C and furnace cooled. The above procedures are summarizedin Fig. 1. The above foams were then sintered at 1000 �C for 2 h.

2.2. Characterization

The density qfoam of the scaffolds was determined from themass and dimensions of the sintered bodies. The porosity p wasthen calculated as:

p ¼ 1� ðqfoam=qsolidÞ ¼ 1� qrelative ð1Þ

where qsolid = 2.7 g cm�3 is the theoretical density of sintered 45S5Bioglass� [23].

The microstructure of the foams before and after immersion inDulbecco’s modified Eagles’s medium (DMEM) (tissue culture solu-tion used for biocompatibility studies, outlined below) was charac-terized by field emission gun (FEG) scanning electron microscopy(SEM) (JEOL 7001). Samples were gold coated and observed at anaccelerating voltage of 15–20 keV. Energy dispersive X-ray (EDX)spectra (Ka line) were collected at 20 keV, then processed usingan INCA (Oxford Instruments) program, using standard referencespectra.

Selected foams were also characterized using X-ray diffraction(XRD) analysis with the aim of assessing the crystallinity after sin-tering and the formation of hydroxyapatite (HA) crystals on strutsurfaces after different times of immersion in DMEM. The foamswere first ground to a powder. Then 0.1 g of the powder was col-lected for XRD analysis. A Philips PW 1700 Series automated pow-der diffractometer was used, employing Cu Ka radiation (at 40 kVand 40 mA) with a secondary crystal monochromator. Data werecollected over the range 2h = 10–60� using a step size of 0.04�and a counting time of 25 s per step.

Attenuated total reflectance Fourier transform infrared spec-troscopy (ATR-FTIR) analysis was conducted for the Bioglass�–ceramic powders, using a Nicolet 6700 spectrometer and a SmartOrbit single bounce diamond ATR accessory to analyse the chemi-cal bonds in the Bioglass�–ceramics. The data were collected in therange 500–2000 cm�1 wave number, at the standard resolution of0.09 cm�1. The data are presented as is, without ‘‘ATR correction’’,and so no account was made for the wavelength-dependent depthof penetration of the FTIR beam. Due to the intimate contact of thematerial with the diamond crystal surface the FTIR spectra did notrequire the use of an internal standard to allow quantitativecomparison.

2.3. Mechanical testing

The compressive strength of the foams was measured using anInstron 5848 mechanical tester at a cross-head speed of0.5 mm min�1 and with a 1 kN load cell. The samples were rectan-gular in shape, with normal dimensions: 20 mm in height and10 � 10 mm in cross-section. During compression testing the loadwas applied until densification of the porous samples started tooccur.

2.4. Assessment of bioactivity in DMEM

The size of all samples for these tests was 10 � 10 � 10 mm or10 � 10 � 20 mm (for compression strength testing). Three sam-ples were extracted from the DMEM solution after 7, 14, 21 and28 days. The DMEM was replaced twice a week because the cationconcentration decreased during the course of the experiments, as aresult of changes in the chemistry of the samples. Once removedfrom the incubation fluid the samples were rinsed gently, first inpure ethanol and then using deionized water, and left to dry atambient temperature in a desiccator.

2.5. Measurement of pH values and ion concentrations

Samples (1 g) of the materials to be tested were soaked in 10 mlof tissue culture medium in 15 ml conical tubes under standardcell culture conditions within a culture incubator (37 �C in humid-ified air containing 5% CO2). Acidity was measured at 4, 12, 24, 36and 48 h. This was followed by replacing the medium with freshDMEM, which was subsequently incubated for another 48 h.

Alkoxides: TEOS and TEP

1. Prepare a sol from the alkoxides and Ca(NO3)2

and NaNO3 in deionised water solventAdd Catalyst (HNO3) to

accelerate hydrolysis

Dry, burn and sinter the green body

2. Foam the sol by vigorous agitation

3. When the gelation of the foamed sol is nearly completed, cast the gel in moulds

Glass-ceramic foam

Surfactant for foaming, Catalyst (HF) for gelation

Add

4. Age, dry and sinter the gel at 600°C for 5 hrs and sinter at 1000°C for 2 hrs

Fig. 1. Flowchart of the production of sol–gel derived Bioglass� foams.

Q.-Z. Chen, G.A. Thouas / Acta Biomaterialia 7 (2011) 3616–3626 3617


During the second 48 h incubation the pH value of the mediumwas again measured at 4, 12, 24, 36 and 48 h. The acidity measure-ments were done by insertion of a pH glass electrode (HannaInstruments HI 1230B) attached to a pH meter (Hanna InstrumentsHI 98185) into the solutions via an incubator access port, allowingthe electrode to stabilize under incubation conditions for eachreading.

The solutions used to analyse ion concentrations were preparedby soaking 1 g of the materials in 10 ml of deionized water. Solu-tions were collected and transferred to new tubes at differentintervals up to 2 days. All solutions were analysed using ion chro-matography. An ICS-1000 (Dionex) ion chromatograph with at-tached autosampler was used for the analysis of sodium (Na+)and calcium (Ca2+) ions. An ICS-2500 (Dionex) was used for theanalysis of phosphate (PO3�

4 ) and silicon (SiO2�4 ) anions.

2.6. Cytocompatibility in vitro (ISO 10993)

The cytocompatibility study was performed according to thestandard cytotoxicity assessment set by the International Stan-dardization Organization (ISO 10993). Samples of 0.2 g of bioactiveglass were preconditioned by soaking in tissue culture medium for48 h. The extractant medium was prepared by placing the precon-ditioned bioactive glass in 2 ml samples of cell culture medium(DMEM supplemented with 10% foetal calf serum (FCS), 1% L-gluta-mine and 0.5% penicillin/streptomycin) for 24 h at 37 �C under 5%CO2 in a culture incubator. HA was used as the material control,and 2 ml of cell culture medium alone was the negative control.Prior to exposure to these extracts osteoblast-like cells (MG63)were seeded in standard medium at a density of approximately2000 cells well�1 in 96-well tissue culture treated plates (Falcon�,BD Bioscience, North Ryde, Australia), under standard incubationconditions (37 �C and 5% CO2), with the medium being changedevery second day. When the cell monolayers had reached 80% con-fluence (at day 4) the medium in each well was entirely replacedwith 0.2 ml of extract medium samples (medium pre-exposed tomaterial) or control medium (material control, medium pre-ex-posed to PleuraSeal�; negative control, medium only). All cultureswere then allowed to proceed for 2 days.

At the end of the incubation period the culture medium was col-lected and the degree of cell death was determined by measure-ment of lactate dehydrogenase (LDH) levels [24], as released intothe culture medium (‘‘released LDH’’), using a commercial kit (Sig-ma–Aldrich TOX-7) as described previously [25]. Finally, each wellcontaining living cells was filled with 0.2 ml of fresh cell culturemedium and cells lysed using TOX-7 solution. These lysates werethen used to determine the cellular LDH content, which equatesto the number of living cells per well (‘‘total LDH’’). The overallLDH level was determined by measuring the absorbance of thesupernatant from the centrifuged medium at 490 nm (after sub-traction of the background absorbance at 690 nm) using a multi-well plate format UV–vis spectrophotometer (Thermo Scientific).The absorbance results for LDH were converted into the numberof cells using a linear standard curve (not shown). Hence, cytotox-icity can be expressed as:

Dead cellsð%Þ ¼ ½released LDH=ðreleased LDHþ total LDHÞ� � 100 ð2Þ

2.7. AlamarBlue™ cell proliferation test

Cell proliferation was assessed using a commercial Alamar-Blue™ assay kit (Life Technologies). AlamarBlue™ is non-toxic tocells. The assay does not interrupt cell culture, allowing continuousmeasurement of cell proliferation kinetics. Hence, the Alamar-Blue™ assay is appropriate to evaluate the long-term cytotoxicityof biomaterials due to biodegradation under physiological

conditions. MG63 cells were seeded at 5000 cells ml�1 in each wellof a 48-well plate and cultured in the medium with extracts. Incu-bated medium in wells with neither cells nor testing material ex-tracts were used as negative controls. After culture for 48 h100 ll of AlamarBlue™ indicator was added to each well (exceptfor the background controls) and incubated under culture condi-tions for 6 h. The medium were then transferred to a new plate, fol-lowed by absorbance determination at wavelengths of 570 and600 nm in a spectrophotometer (Thermo Scientific, Pathtech,Australia).

The above procedures were repeated every 48 h until conflu-ence was reached (typically day 6). Cell proliferation was quanti-fied as the percentage reduction in AlamarBlue, i.e.:

reduced AlamarBlueð%Þ ¼ ð½eoxðk2ÞAðk1Þ� eoxðk1ÞAðk2Þ�=½eredðk1ÞA� ðk2Þ� eredðk2ÞA� ðk1Þ�Þ � 100 ð3Þ

where A(k1) and A(k2) are the values of absorbance of test wellsmeasured at wavelengths k1 and k2, and A� (k1) and A� (k2) arethe values of absorbance of negative control wells containing onlymedium and AlamarBlue™ without cells. All values were blankedwith the readings of the background controls. The other parametersin Eq. (3) are as follows: eox(k1) = 80,586, eox(k2) = 117,216,ered(k1) = 155,677 and ered(k2) = 14,652.

2.8. Cell seeding on the sol–gel derived scaffolds

The sol–gel derived, sintered scaffolds were cut to a size of�15 mm in diameter and �5 mm in thickness. All scaffolds weresterilized in dry heat at 180 �C for 2 h and preconditioned in cul-ture medium for 48 h prior to cell culture. MG63 cells were seededon the sterilized scaffolds at a variety of concentrations: 1000,2000, 5000, 10,000 and 20,000 cells per sample. The samples werekept at 37 �C in an atmosphere of 5% CO2 for incubation periods of1 and 6 days.

2.9. Preparation of SEM samples after tissue culture

The above cell-cultured 45S5 Bioglass�–ceramic scaffold speci-mens were fixed in 2.5% glutaraldehyde in 0.14 M sodium cacodyl-ate buffer (pH 7.3) (both Sigma–Aldrich) at 48 �C overnight, thendehydrated in a graded series of alcohols (50%, 70%, 90%, and twochanges of 100% ethanol), washed with hexamethyldisilazane for1–2 min, and placed in a desiccator overnight. After 24 h the sam-ples containing fixed cells were mounted onto stubs using Araldite(Devcon, Wellingborough, UK) and Liquid Dag (Neubauer Chemik-alen, Munster, Germany) and left to air dry overnight. Specimenswere then gold coated and observed by FEG SEM using a JEOL7001 at 15 keV.

2.10. Statistical analysis

All experiments were run with five samples per experimentalgroup, and data are shown as means ± standard errors of the mean.One-way analysis of variance (ANOVA) with Tukey’s post hoc testwas performed to analyse the statistical difference, and signifi-cance levels were set at P < 0.05.

3. Results and discussion

3.1. Macroporous structure of the foams

Two typical macroporous structures of the sintered sol–gel de-rived Bioglass�–ceramic foams are shown in Fig. 2. At high foam

3618 Q.-Z. Chen, G.A. Thouas / Acta Biomaterialia 7 (2011) 3616–3626


porosity (>90%) (Fig. 2a) the pore size was mainly in the range 0.2–1 mm. Most medium sized (�0.5 mm in diameter) and larger(�1 mm in diameter) pores had small windows (0.03–0.3 mm indiameter) in their walls. However, the macroporous structure ofthe foams was dominated by completely closed pores at porositiesof below 85% (Fig. 2b).

During sintering at 1000 �C for 2 h the foams shrank consider-ably, with a reduction in porosity of on average 5%. In the case ofhigh porosities the small windows in the cell walls persisted inthe macroporous structure (Fig. 2a), indicating that the connectivestructure of the foams was maintained during sintering at hightemperatures.

3.2. Chemical characteristics of the foams

FTIR analysis (Fig. 3) revealed that decomposition of the two ni-trates NaNO3 and Ca(NO3)2 was incomplete after ageing at 600 �Cfor 5 h, with a peak of NO3 remaining at �820 cm�1. However, aftersintering at 1000 �C for 2 h the characteristic peaks of the two ni-trates NaNO3 and Ca(NO3)2 disappeared, leaving typical Bio-glass�–ceramic peaks. Meanwhile, two broad silicate absorptionbands at about 1085 and 910 cm�1 were observed in the spectraof the sol–gel derived powders, aged at either 600 �C or at 600 �Cfollowed by sintering at 1000 �C. The peaks at 729, 696, 627 and615 cm�1 can be attributed to the crystalline phase, which is con-firmed by the XRD results (Fig. 4). The peaks at 627 and 615 cm�1

are ascribed to the deformation modes of P–O bonds in crystallinephosphate, and the peaks at 729 and 696 cm�1 are assigned to thesymmetric Si–O–Si stretching vibration in crystalline silicate SiO4�

4 ,

as previously described [26,27]. Furthermore, the FTIR spectrumprofile of the present sol–gel derived, sintered powders are verysimilar to those of sintered NovaBone� Bioglass�, as reported pre-viously [19]. Hence, the sol–gel process described in Chen et al.[19] can be applied to produce a 45S5 Bioglass�–ceramic reproduc-ibly and reliably.

The crystalline phase Na2Ca2Si3O9 could be identified in the sol–gel derived 45S5 Bioglass�–ceramic material by XRD analysis(Fig. 4). Even prior to soaking in DMEM apatite peaks were evidentin the XRD spectra of the four powders shown. However, this maynot be considered surprising, because the aqueous environment ofthe sol–gel process has been found to favour the precipitation ofHA from the amorphous structure of silica-based bioactive glasses[28,29].

3.3. Bioactivity of the sol–gel derived 45S5 Bioglass�–ceramic foams

3.3.1. Morphology observation (SEM)The surface morphology of the walls of pores in the sol–gel de-

rived, sintered 45S5 Bioglass�–ceramic foams before and afterincubation in DMEM for up to 4 weeks can be seen in Fig. 5. After

Fig. 2. Macroporous structures of sol–gel derived 45S5 Bioglass�–ceramic foams.(a) The porosity of the foam is 91.1%; the size of most pores is in the range 0.25–1 mm with small windows (<0.3 mm in diameter) in their walls. (b) The porosity ofthe foam is 84.9%; most cells are completely closed.

Fig. 3. FTIR spectra of a sol–gel derived 45S5 material aged at 600 �C or sintered at1000 �C for 2 h, and a sintered commercial Bioglass�. The spectrum of the sol–gelderived Bioglass�–ceramics sintered at 1000 �C has the same profile as that of thecommercial NovaBone 45S5 Bioglass� sintered at 1000 �C for 2 h.

Fig. 4. X-ray diffraction patterns of sol–gel derived 45S5 Bioglass�–ceramics agedat 600 �C for 5 h and sintered at 1000 �C for 2 h. All spectra were obtained using0.05 g powder. The major peaks of combeite (Na2Ca2Si3O9) and hydroxyapatite aremarked by (5) and (.), respectively.



soaking in DMEM for 7 days these pore wall surfaces were popu-lated with newly formed fine particles (�1 lm in size) (Fig. 5b),which have been identified as bone-like apatite precipitates (Sec-tion 3.3.2). The population of particles increased with increasingimmersion time in DMEM (compare Fig. 5c and d). The foam sur-faces were nearly entirely covered by a layer of fine precipitatesafter incubation in DMEM for 2 weeks (Fig. 5c). Further evolutionwas the formation of precipitate agglomerates on the existing pre-cipitate layer (Fig. 5d).

3.3.2. X-ray diffraction and EDX analysisXRD spectra of the sol–gel derived, sintered 45S5 Bioglass�–

ceramic material before and after soaking in DMEM are shown inFig. 6. XRD spectra were generated from the whole material of usedpowder, rather than from a surface layer of powder particles.Hence, the spectra obtained are representative of the structurethroughout the material. The diffraction peaks of the Na2Ca2Si3O9

phase became shorter with increasing incubation time in DMEM,eventually almost disappearing after incubation for 4 weeks, leav-ing a broad halo pattern (indicating an amorphous structure). EDXanalysis revealed that after soaking in DMEM for 4 weeks the com-position of the amorphous powder was identical to that of bone-like apatite (i.e. HA) (Fig. 7). The formation of amorphous apatitethroughout the material indicates that it is likely to have suitablebone bonding, as well as suitable degradability akin to the previ-ously reported properties of amorphous HA and related amorphouscalcium phosphates [30].

3.3.3. pH change during incubation in DMEMWhen soaking in an aqueous environment ion release from

45S5 Bioglass�–ceramics was expected to cause changes in thepH value and ion concentration of the surrounding fluid. Previousstudies on this subject have been conducted with deionized water,phosphate-buttered saline (PBS), simulated body fluid (SBF) andTris-buffered solutions [31–35]. We focused on the results of soak-ing in DMEM tissue culture medium, which is relevant to the eval-uation of biocompatibility of the materials under standard cell

Fig. 5. Surface morphology of the pore walls in sol–gel derived 45S5 Bioglass� foams. (a) As-sintered at 1000 �C for 2 h. (b–d) Sintered followed by immersion in tissueculture medium (DMEM) for 1, 2 and 4 weeks, respectively.

Fig. 6. X-ray diffraction patterns of sol–gel derived 45S5 Bioglass�–ceramic foamsbefore and after immersion in tissue culture medium (DMEM) for up to 4 weeks. Allspectra were obtained using 0.05 g powder. The major peaks of combeite(Na2Ca2Si3O9) are marked by (5).

Fig. 7. Energy dispersive X-ray (EDX) results on the sol–gel derived 45S5 Bioglass�–ceramic foams after immersion in tissue culture medium (DMEM) for 4 weeks. Thecomposition of 45S5 and the stoichiometric compositions of two calcium phos-phates, hydroxyapatite (HA) and tricalcium phosphate (CaP), are also shown forcomparison.



culture conditions in vitro. For comparison, 45S5 Bioglass–ceramicscaffolds produced by the replication technique [14] and sinteredunder the same conditions were assessed as well.

Immersion of sol–gel derived 45S5 Bioglass�–ceramic foams in-creased the pH value of DMEM more rapidly than those fabricatedby the replication technique (Fig. 8). With the former a saturatedstate (pH �9.0) was reached after immersion for 12 h, whereaswith the latter a similar saturated state (pH �8.7) was reachedafter immersion for 48 h (Fig. 8a). The saturated pH values ofDMEM caused by incubation of the two types of scaffolds, however,were not significantly different. During the second 48 h incubationthe increase in pH value of the medium was not as great as duringthe first 48 h, the saturated values being 8.4 and 7.9 on average forthe sol–gel derived and replicated scaffolds, respectively (Fig. 8b).

3.3.4. Ion release during incubation in deionized waterThe faster incremental change in pH value caused by the sol–gel

derived 45S5 scaffolds can be directly attributed to the faster re-lease of Na+ ions from this type of foam, as shown in Fig. 9. A com-parison of Fig. 9a and b reveals that ion release was considerablyfaster during the first 24 h in the group of sol–gel derived scaffolds,compared with those fabricated by the replication method. It mustbe mentioned that the porosity of the scaffolds used in this testwere all around 0.90. We believe that the rapid release of ions fromthe sol–gel derived scaffolds is attributable to the large surfacearea in foams of this type, compared with those fabricated by thereplication method.

3.4. Compressive strength of the foams

The compressive strength values of the sol–gel derived, sinteredfoams (without immersion in DMEM) are given in Fig. 10. There

Fig. 8. pH as a function of scaffold immersion time in the DMEM tissue culturemedium during the (a) first and (b) second 48 h immersion.

Fig. 9. Analysis of Na, Ca, P and Si in deionized water in which the scaffolds wereimmersed for 48 h. The scaffolds were produced by (a) the replication technique or(b) the sol–gel foaming method.

Fig. 10. Compression strength values of sol–gel derived Bioglass� foams sintered at1000 �C for 2 h (blue dots), as well as theoretical calculated compressive strengthvalues of closed cell and open cell foams (curves). Theoretical values werecalculated from Eqs. (4) and (5), respectively.



was an abrupt increase in strength values when the foam porositydropped to �85%. This phenomenon is very likely related to thechange in macroporous structure: from partially open cells(Fig. 2a) to a nearly completely closed cell structure (Fig. 2b).

The compressive strength values of the sol–gel derived foams(Fig. 10) are collectively higher than those of scaffolds producedby the replication technique, which is close but lower than the the-oretical strength values of open cell foams [14]. The primary reasonfor this is the openness of the pores in the macroporous network.The scaffolds produced by the replication methods are fully open[18,36], whereas the sol–gel derived foams contain some partiallyclosed pores (Fig. 2). To engineer bone tissue a partially or fullyopen porous network is more desirable, for vascular infiltrationof the artificial scaffolds [20]. Hence, the improvement in themechanical properties of the sol–gel derived scaffolds wasachieved at the cost of macroporous structure.

Fig. 11 shows the compressive strength values of the sol–gel de-rived sintered foams after incubation in DMEM for 4 weeks, whichdecreased dramatically. The decrease in strength of the Bioglass�–ceramic foams can be attributed to the formation of an amorphousbone-like apatite, as discussed above (Figs. 5–7).

3.5. Comparison of theoretical compressive strength values withexperimental data

The modelling of the mechanical behaviour of highly porousmaterials has been presented by Gibson and Ashby [37]. The theo-retical compressive collapse stress (rtheo) of a foam with open orclosed cells can be expressed as a function of the relative density(qfoam/qsolid) of a cellular structure by Eqs. (4) and (5), respectively:

rtheo

rfs¼ 0:2

qfoam

qsolid

� �32

; ðopen cellsÞ ð4Þ

rtheo

rfs¼ 0:2 /

qfoam

qsolid

� �32

þ ð1� /Þ qfoam

qsolid

� �; ðclosed cellsÞ ð5Þ

where / is the fraction of material contained in the cell edges (thus1 � / is the fraction of material distributed in the cell faces) and rfs

is the modulus of rupture of the strut. Theoretical calculations showthat the modulus of rupture is typically about 1.1 times larger thanthe tensile strength in brittle materials [37].

These two equations are derived assuming that the cell walls(including cell edges and faces) are fully dense. Hence, Eqs. (4)and (5) are not directly applicable to the unsintered sol–gel derivedfoams in the present experimental work, because their cell wallsare made of a mesoporous aerogel, as described above. However,the two equations were thought applicable to the sintered foamsbecause sintering at 1000 �C effectively densified the mesoporousstructure (Fig. 5a). The tensile strength of dense glass–ceramicscomposed of 45S5 Bioglass� is 42 MPa [23].

For open cell foams the compressive strength (Eq. (4)) is com-pletely contributed by the cell edges. The calculated compressivestrength for open cell foams is given in Fig. 10. Foams with closedcells present a more complicated behaviour. Their strength is thesum of two contributions. The first is the contribution of cell edge

bending 0:2 / qfoamqsolid

� �32

� �. The second contribution is from the mem-

brane stresses in the cell faces ð1� /Þ qfoamqsolid

� �� [37]. When a closed

cell foam is loaded the bending of the cell edges causes the cellfaces to stretch. It is the membrane stresses of cell faces that confersignificantly higher strength to closed cell foams compared withtheir open cell counterparts. Measurements of / for rigid polyure-thane foams, which have a similar macroporous structure to thatshown in Fig. 2a, show that this value is typically 80% [38]. In thiswork / = 0.7, 0.8 and 0.9 were tried in the calculation of the

compressive strength of closed cell foams, and the results areshown in Fig. 10.

It should be noted that the investigated foams in this work donot exhibit a completely closed cell structure. Hence, it is not sur-prising to find that the experimental data are collectively lowerthan the calculated values for completely closed cell foams (with/ = 0.80). The small windows in the cell faces indicate that the /value (the fraction of solid material contained in the cell edges)of the present foams is likely to be greater than 80%. Fig. 10 showsthat most experimental data are located around the curve for/ = 0.85.

It is also worth mentioning that the Gibson and Ashby theory[37] is built on a geometrically uniform model, i.e. the size of cells,as well as the struts and cell walls, is identical throughout the por-ous network. In reality there is no such perfect network, and thestruts of a real porous scaffold vary in thickness and always containcracks and voids. Hence, a real porous network will always prema-turely rupture at these weak sites where the size of the struts andcell walls are thinner than the mean values. Hence, the experimen-tal values for the compressive strength of scaffolds will always bebelow the curve of theoretical values for the same porosity and /parameter. Actually, a theoretical curve calculated from the Gibsonand Ashby model (Eqs. (4) or (5)) sets an upper boundary of com-pressive strength that can be achieved by a real scaffold. Anyexperimental value significantly higher than the upper limit shouldbe questionable.

In summary, the compressive strength values of the sol–gel de-rived sintered foams, which were close to the theoretical valuespredicted by Eq. (5) with / = 0.85, indicate that the foams primarilycontained closed cells throughout the network, in agreement withthe SEM observations (Fig. 2). Although only �15% of the materialis contained in the cell faces, this small quantity of material plays akey role in affecting the mechanical performance of the foams.

3.6. Cytocompatibility of the sol–gel derived Bioglass� scaffolds

The biocompatibility of melt derived 45S5 and sol–gel derivedS70C30 has previously been extensively evaluated, both in vitroand in vivo [39–41]. Given that the current sol–gel derived 45S5was produced by a completely new process which involved acidiccatalysts, HNO3 and HF, it was essential to evaluate the biocompat-ibility of the materials, using the cytocompatibility standard de-scribed in ISO 10993 and required by the FDA for medicallyapplied biomaterials. HA materials were used as controls in thiswork.

Fig. 11. Comparison of the compressive strength of sol–gel derived Bioglass� foamsbefore and after immersion in a DMEM tissue culture medium for 4 weeks. Foamswere sintered at 1000 �C for 2 h.



The medium containing the extracts of sintered powders werefound to support proliferation of MG63 cells and visual microscopicobservation showed no gross differences in cell proliferation be-tween the negative control, HA and sol–gel derived Bioglass�–ceramics groups (Fig. 12). This was quantitatively confirmed usingthe LDH (Fig. 13) and AlamarBlue™ (Fig. 14) techniques, revealingthat the growth kinetics of MG63 cells were not statistically differ-ent between the three groups of medium tested. The sol–gel de-rived, sintered glass–ceramics were found to have very similarcytocompatibility to HA, as indicated in Fig. 13. The proportion ofdead cells observed in MG63 cultures exposed to the sol–gel

derived Bioglass�–ceramics was on average slightly higher thanthat for HA, but the difference was insignificant.

In this study the extractant medium was collected after soakingthe preconditioned sol–gel Bioglass� for 24 h, with a pH value of�8.2, shown in Fig. 8b, which is considerably higher than the pHlevel of human body fluids (typically in the range 7.2–7.4).Although most cell types are sensitive to extracellular pH, and sur-vive optimally in a physiological medium with a pH value in thisrange, osteoblasts and osteoblast-like cells (MG63) can withstandmoderately alkaline conditions [42–45], with some studies show-ing increased metabolic activity of primary osteoblasts withincreasing pH from 7.0 to 7.8 [43,46]. In another study the pH valueof the culture medium without cells reached 8.1 in the presence ofa bioactive glass after 2 days incubation, and human osteoblasticcells cultured in contact with this bioactive glass retained theirphysiological, morphological and biochemical properties [45].The pH value of the same culture medium with cells reached a pla-teau at a lower value (7.8) [45], indicating that osteoblast cellshave the ability to buffer their surrounding medium, which is likelydue to their acidification by lactic acid secretion with increasing

Fig. 12. Typical morphology of MG63 cells cultured for 2 days in (a) the standardculture medium, (b) the extractant medium for hydroxyapatite, (c) the extractantmedium for sol–gel derived Bioglass�–ceramics 1000 �C for 2 h, respectively. Themagnification of the images is the same.

Fig. 13. Cytotoxicity of the test materials using MG63 cells, detected by measuringthe release of LDH into the medium containing the extracted substances during2 days culture. Medium with hydroxyapatite extracts served as the positive control.The sol–gel derived Bioglass�–ceramics were sintered at 1000 �C for 2 h. Nosignificant differences were found in the percentage of dead cells between any twoof the three groups (P > 0.05) n = 5.

Fig. 14. MG63 cell proliferation kinetics measured by the AlamarBlue™ technique.The initial plating density was 5000 cells ml�1 in each well of a 48-well plate (n = 5).Medium with the hydroxyapatite extract was the positive control. Sol–gel derivedBioglass�–ceramics were sintered at 1000 �C for 2 h. The differences between anytwo groups were not significant (P > 0.05) throughout the culture duration. Forclarity, error bars are not shown in the graph.



cell density. Cells exposed to the extractant of the sol–gel Bioglass�

may have undergone an initial shock when exposed to alkalinemedium, however, this appeared to have been alleviated over time.As a result, no significant cytotoxicity was revealed, compared withthe HA group.

While we cannot conclude that the absence of cytotoxicity wasnot simply due to compensation by the cells as a stress response toa non-physiological pH range, we have shown previously withelastomeric Bioglass composites that we can more closely controlthe acidification of the medium by altering the manufacturing con-ditions of the biomaterial [47–49]. This might be a future consider-ation for the present foam, depending on the specific tissue typethat it would be engineered for.

3.7. Cell infiltration of the sol–gel derived foams

MG63 cells were directly seeded onto sol–gel derived, sinteredBioglass�–ceramic scaffolds in 48-well tissue culture plates. Thesample size was �8 � 8 � 5 mm, and seeding densities were 1–20 � 103 cells per well. Regardless of what seeding density wasused, MG63 cells were primarily located on the surface of thefoams (Fig. 15a), with very few cells immigrating into the poresthrough small windows (Fig. 15b). Nonetheless, SEM observationsrevealed a normal morphological phenotype of this cell type on thesurface matrix of the sol–gel derived Bioglass�–ceramic scaffoldsafter 6 days culture (Fig. 16), with some evidence of active attach-ment by cell membrane processes (Fig. 16b). Hence, we envisagethat the porous structure in the present sol–gel derived

Bioglass�–ceramic foams could impede the in-growth of vesselsinto the scaffolds, as well as the regeneration of bone tissue, eventhough attachment of cells to the scaffold surfaces may be unaf-fected. To achieve a better interconnection between pores the for-mation of homogenously distributed bubbles and their coalescencemust be carefully controlled by using alternative surfactant typesand matched drying, ageing and sintering protocols. Alternatively,

Fig. 15. (a) Scanning electron micrographs of cultured MG63 osteoblast-like cellson the surface of sol–gel derived Bioglass�–ceramic scaffolds, cultured for 1 daywhen cells were dividing. (b) Section of a cell-seeded scaffolds, with almost no cellson the pore wall.

Fig. 16. Scanning electron micrographs of cultured MG63 osteoblast-like cells onthe sol–gel derived Bioglass�–ceramic scaffolds cultured for: (a) 1 day, just dividedcells; (b) 4 days, widely spread cells (such that it is hard to recognize the boundaryof the cell in the photo); (c) 6 days, a spreading cell, showing membranousprocesses in contact with the scaffold surface.



a new approach that combines the sol–gel process (a material fab-rication technique) with other foaming methods, such as the repli-cation technique, could be used to construct a suitably strong foamwith a more open porous structure.

4. Conclusions

The novel sol–gel derived 45S5 Bioglass�–ceramic material hasbeen foamed into three-dimensional scaffolds using a surfactantcombined with vigorous mechanical stirring. The structure,mechanical properties, bioactivity, cytocompatibility and cell infil-tration of the porous materials were investigated by standard tech-niques. The major conclusions are listed below.

(1) Porous structure: the sol–gel foaming process using a surfac-tant and mechanical stirring produces scaffolds primarilycontaining closed pores throughout the three-dimensionalstructure.

(2) Compressive strength: the compressive strength of the sin-tered foams can be quantitatively described by the Gib-

son–Ashby equation rtheorfs¼ 0:2 /

qfoam

qsolid

� �32 þ ð1� phiÞ qfoam

qsolid

� �,

which is based on the model of a closed cell foam.(3) Bioactivity: the sol–gel derived, sintered 45S5 Bioglass�–

ceramic foams release alkaline ions more rapidly and thusmore rapidly increased the pH value of the surroundingmedium than counterparts produced by the replicationtechnique, due to the larger surface area of the porous struc-ture produced by the surfactant with mechanical stirringmethod.

(4) The sol–gel derived 45S5 Bioglass�–ceramic material sup-ports surface proliferation of viable cells. However, theclosed cell porous structure prevents cell migration intothe pores and thus into the three-dimensional network.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figures 3, 4, 6, 7, 8, 9,10, 11, 12, 13 and 14, are difficult to interpret in black and white.The full colour images can be found in the on-line version, atdoi:10.1016/j.actbio.2011.06.005.

References

[1] Burg KJL, Porter S, Kellam JF. Biomaterial developments for bone tissueengineering. Biomaterials 2000;21:2347–59.

[2] Vacanti JP, Vacanti CA. The history and scope of tissue engineering. In: LanzaRP, Langer R, Vacanti JP, editors. Principles of tissue engineering. San Diego,CA: Academic Press; 2000. p. 3–7.

[3] Temenoff JS, Lu L, Mikos AG. Bone tissue engineering using syntheticbiodegradable polymer scaffolds. In: Davies JE, editor. BoneEngineering. Toronto: EM Squared; 2000. p. 455–62.

[4] Bruder SP, Caplan AI. Bone regeneration through cellular engineering. In: LanzaRP, Langer R, Vacanti JP, editors. Principles of tissue engineering. San Diego,CA: Academic Press; 2000. p. 683–96.

[5] Hutmacher DW. Scaffolds in tissue engineering bone and cartilage.Biomaterials 2000;21:2529–43.

[6] Ishizaki K, Komarneni S, Nanko M. Porous materials: processing technologyand applications. Dordrecht: Kluwer; 1998. p. 12–66.

[7] Leong KF, Cheah CM, Chua CK. Solid freeform fabrication of three-dimensionalscaffolds for engineering replacement tissues and organs. Biomaterials2003;24:2363–78.

[8] Reed JS. Principles of ceramic synthesis. 2nd ed. Chichester: Wiley; 1988.[9] Komlev VS, Barinov SM. Porous hydroxyapatite ceramics of bi-modal pore size

distribution. J Mater Sci Mater Med 2002;13:295–9.[10] Fukasawa T, Deng ZY, Ando M, Ohji T, Goto Y. Pore structure of porous

ceramics synthesized from water-based slurry by freeze-dry process. J MaterSci 2001;36:2523–7.

[11] Lyckfeldt O, Ferreira JMF. Processing of porous ceramics by ‘‘starchconsolidation’’. J Eur Ceram Soc 1998;18:131–40.

[12] Sepulveda P, Binner JGP. Processing of cellular ceramics by foaming and in situpolymerisation of organic monomers. J Eur Ceram Soc 1999;19:2059–66.

[13] Jones JR, Ahir S, Hench LL. Large-scale production of 3D bioactive glassmacroporous scaffolds for tissue engineering. J Sol–Gel Sci Technol2004;29:179–88.

[14] Chen QZ, Thompson ID, Boccaccini AR. 45S5 Bioglass�-derived glass–ceramicscaffolds for bone tissue engineering. Biomaterials 2006;27:2414–25.

[15] Pereira MM, Clark AE, Hench LL. Effect of texture on the rate of hydroxyapatiteformation on gel–silica surface. J Am Ceram Soc 1995;78:2463–8.

[16] Jones JR, Hench LL. Factors affecting the structure and properties of bioactivefoam scaffolds for tissue engineering. J Biomed Mater Res B Appl Biomater2004;68B:36–44.

[17] Jones JR, Ehrenfried LM, Hench LL. Optimising bioactive glass scaffolds for bonetissue engineering. Biomaterials 2006;27:964–73.

[18] Chen QZ, Boccaccini AR. Coupling mechanical competence and bioresorbabilityin Bioglass�-derived tissue engineering scaffolds. Adv Eng Mater2006;8:285–9.

[19] Chen QZ, Li YA, Jin LY, Quinn JMW, Komesaroff PA. A new sol–gel process forproducing Na2O-containing bioactive glass ceramics. Acta Biomater2010;6:4143–53.

[20] Laschke MW, Harder Y, Amon M, Martin I, Farhadi J, Ring A, et al. Angiogenesisin tissue engineering: breathing life into constructed tissue substitutes. TissueEng 2006;12:2093–104.

[21] Kannan RY, Salacinski HJ, Sales K, Butler P, Seifalian AM. The roles of tissueengineering and vascularisation in the development of micro-vascularnetworks: a review. Biomaterials 2005;26:1857–75.

[22] Sepulveda P, Jones JR, Hench LL. Bioactive sol–gel foams for tissue repair. JBiomed Mater Res 2002;59:340–8.

[23] Hench LL, Wilson J. An introduction to bioceramics. 2nd ed. London: WordScientific; 1999.

[24] Schildhauer TA, Gekle CJE, Muhr G. Biomaterials in the skeletal system. Chirurg1999;70:888–96.

[25] Berchtold MW, Brinkmeier H, Muntener M. Calcium ion in skeletal muscle: itscrucial role for muscle function, plasticity, and disease. Physiol Rev2000;80:1215–65.

[26] Boccaccini AR, Chen Q, Lefebvre L, Gremillard L, Chevalier J. Sintering,crystallisation and biodegradation behaviour of Bioglass�-derived glass–ceramics. Faraday Discuss 2007;136:27–44.

[27] Lefebvre L, Chevalier J, Gremillard L, Zenati R, Thollet G, Bernache-Assolant D,et al. Structural transformations of bioactive glass 45S5 with thermaltreatments. Acta Mater 2007;55:3305–13.

[28] Rezwan K, Chen QZ, Blaker JJ, Boccaccini AR. Biodegradable and bioactiveporous polymer/inorganic composite scaffolds for bone tissue engineering.Biomaterials 2006;27:3413–31.

[29] Bellantone M, Hench LL. Bioactive behaviour of sol–gel derived antibacterialbioactive glass. Bioceramics 2000;192–1:617–20.

[30] Hench LL. Bioceramics. J Am Ceram Soc 1998;81:1705–28.[31] Peitl O, Zanotto ED, Hench LL. Highly bioactive P2O5–Na2O–CaO–SiO2 glass–

ceramics. J Non-Cryst Solids 2001;292:115–26.[32] Clupper DC, Mecholsky JJ, LaTorre GP, Greenspan DC. Bioactivity of tape cast

and sintered bioactive glass–ceramic in simulated body fluid. Biomaterials2002;23:2599–606.

[33] Saravanapavan P, Hench LL. Dissolution of bioactive gel–glass powders in theSiO2–CaO system. Bioceramics 2003;15:240–2. 213–216.

[34] Chen QZ, Ahmed I, Knowles JC, Nazhat SN, Boccaccini AR, Rezwan K. Collagenrelease kinetics of surface functionalized 4555 Bioglass�-based porousscaffolds. J Biomed Mater Res A 2008;86A:987–95.

[35] Saravanapavan P, Jones JR, Pryce RS, Hench LL. Bioactivity of gel–glass powdersin the CaO–SiO2 system: a comparison with ternary (CaO–P2O5–SiO2) andquaternary glasses (SiO2–CaO–P2O5–Na2O). J Biomed Mater Res A2003;66A:110–9.

[36] Chen QZ, Boccaccini AR, Zhang HB, Wang DZ, Edirisinghe MJ. Improvedmechanical reliability of bone tissue engineering (Zirconia) scaffolds byelectrospraying. J Am Ceram Soc 2006;89:1534–9.

[37] Gibson LJ, Ashby MF. Cellular solids: structure and properties. 2nd ed. Oxford:Pergamon Press; 1999. p. 429–52.

[38] Reitz DW, Schuetz MA, Glicksman LR. A basic study of aging of foam insulation.J Cell Plastics 1984;20:104–13.

[39] Saravanapavan P, Selvakumaran J, Hench LL. Indirect cytotoxicity evaluation ofsoluble silica, calcium, phosphate silver ions. Bioceramics 2004; 16:254–2,785–8.

[40] Saravanapavan P, Jones JR, Verrier S, Beilby R, Shirtliff VJ, Hench LL, et al.Binary CaO–SiO2 gel–glasses for biomedical applications. Bio-Med Mater Eng2004;14:467–86.

[41] Stanley HR, Hall MB, Clark AE, King CJ, Hench LL, Berte JJ. Using 45S5 bioglasscones as endosseous ridge maintenance implants to prevent alveolar ridgeresorption: a 5-year evaluation. Int J Oral Maxillofacial Implants1997;12:95–105.

[42] Kaysinger KK, Pierce WM, Nerland DE. Quantitative analysis of 2-oxoglutaratein biological samples using liquid-chromatography with electrochemicaldetection. Anal Biochem 1994;222:81–5.

[43] Kaysinger KK, Ramp WK. Extracellular pH modulates the activity of culturedhuman osteoblasts. J Cell Biochem 1998;68:83–9.

[44] Kirchhof K, Hristova K, Krasteva N, Altankov G, Groth T. Multilayer coatings onbiomaterials for control of MG-63 osteoblast adhesion and growth. J Mater SciMater Med 2009;20:897–907.



[45] Josset Y, Nasrallah F, Jallot E, Lorenzato M, Dufour-Mallet O, Balossier G, et al.Influence of physicochemical reactions of bioactive glass on the behavior andactivity of human osteoblasts in vitro. J Biomed Mater Res A2003;67A:1205–18.

[46] Ramp WK, Lenz LG, Kaysinger KK. Medium pH modulates matrix, mineral, andenergy-metabolism in cultured chick bones and osteoblast-like cells. BoneMiner 1994;24:59–73.

[47] Liang SL, Cook WD, Thouas GA, Chen QZ. The mechanical characteristics andin vitro biocompatibility of poly(glycerol sebacate)–Bioglass� elastomericcomposites. Biomaterials 2010;31:8516–29.

[48] Chen QZ, Quinn JMW, Thouas GA, Zhou XA, Komesaroff PA. Bone-likeelastomer-toughened scaffolds with degradability kinetics matching healingrates of injured bone. Adv Eng Mater 2010;12:B642–8.

[49] Chen QZ, Jin LY, Cook WD, Mohn D, Lagerqvist EL, Elliott DA, et al. Elastomericnanocomposites as cell delivery vehicles and cardiac support devices. SoftMatter 2010;6:4715–26.


Fabrication and characterization of sol–gel derived 45S5 Bioglass ceramic scaffolds

Documents

Transcript of Fabrication and characterization of sol–gel derived 45S5 Bioglass ceramic scaffolds