Fabrication of calcium phosphate fibres through electrospinning and sintering of hydroxyapatite nanoparticles

Materials Letters 106 (2013) 145–150

Contents lists available at SciVerse ScienceDirect

Materials Letters

0167-57http://d

$ThisCommomits nothe orig

n CorrE-m

journal homepage: www.elsevier.com/locate/matlet

Fabrication of calcium phosphate fibres through electrospinningand sintering of hydroxyapatite nanoparticles$

Pierre-Alexis Mouthuy a, Alison Crossley b, Hua Ye a,n

a Institute of Biomedical Engineering, Department of Engineering Science, University of Oxford, UKb BegbrokeNano—OMCS, Department of Materials, Oxford University Begbroke Science Park, Sandy Lane, Yarnton, Oxford OX5 1PF, UK

a r t i c l e i n f o

Article history:Received 10 November 2012Accepted 20 April 2013Available online 10 May 2013

Keywords:ElectrospinningSinteringHydroxyapatite nanoparticlesPLGAComposites

7X/$ - see front matter & 2013 The Authors. Px.doi.org/10.1016/j.matlet.2013.04.110

is an open-access article distributed undens Attribution-NonCommercial-No Derivativen-commercial use, distribution, and reproductinal author and source are credited.esponding author.ail address: [email protected] (H. Ye).

a b s t r a c t

Calcium phosphate (CaP) materials such as synthetic hydroxyapatite (HA, Ca10(PO4)6 � (OH)2) andβ-tricalcium phosphate (β-TCP, Ca3(PO4)2) are well-known for their potential in bone tissue engineeringand drug delivery applications. Processing such materials into submicrofibres might contribute toimprove their biocompatibility. This paper presents a new method for creating CaP submicrofibresthrough the electrospinning route. A thermal treatment at sintering temperature (1100 1C) was appliedto electrospun polymer fibres filled with hydroxyapatite nanoparticles to cause aggregation of thenanoparticles and vaporise the polymer matrix. The images taken by electron microscopy revealed thatthe treated samples maintained their submicrofibrous morphology. Moreover, Fourier transform infraredspectroscopy, X-ray diffraction and X-ray photoelectron spectroscopy analysis confirmed that theresulting fibres are made of hydroxypatite and tricalcium phosphates.

& 2013 The Authors. Published by Elsevier B.V. All rights reserved.

1. Introduction

Calcium phosphate (CaP) materials are known to support boneingrowth and to promote bone integration [1]. They are alsonon-toxic, biodegradable and can easily adsorb bioactive mole-cules at their surface. These characteristics make them usefulfor both tissue engineering (TE) and drug delivery applications[2,3]. Among CaP materials, synthetic hydroxyapatite (HA,Ca10PO46 � (OH)2) and β-tricalcium phosphate (β-TCP, Ca3(PO4)2)are most widely used. HA is the main form of calcium-phosphatecrystal found in human bone. Usually, the degradability of CaPmaterials varies according to the Ca/P ratio, with β-TCP being moresoluble than HA [4]. Therefore, mixing CaP materials may help tocontrol, to some extent, the degradation of the TE scaffold. Inparticular, HA/β-TCP ceramic scaffolds have been shown to bemore effective in bone repair than pure HA or pure β-TCP [5].

CaP solid bodies have been used for many applications but theyusually face the issue of being too brittle. The poor mechanicalproperties of CaP ceramic materials have severely obstructed theirclinical applications [6]. Fibres and other types of particle aregenerally more successful for filling bone defects and helping to

ublished by Elsevier B.V. All rights

r the terms of the CreativeWorks License, which per-

ion in any medium, provided

repair the damaged tissue. Furthermore, they can be dispersed inpolymer matrices for mechanical reinforcement purposes and forimproved biocompatibility [7].

CaP fibres have been produced through several routes includingspinning, extrusion, precipitation, sol–gel and pyrolysis, and elec-trospinning [8–13]. Electrospinning is a simple technology thatallows the production of nano- and micro-fibres by applying ahigh voltage to a polymer solution. It has been used to producedifferent types of ceramic fibres through the use of chemicalprecursors mixed to the polymer solution and by performing athermal treatment on the resulting fibres [14,15]. However, theproduction of CaP fibres by electrospinning remains poorly inves-tigated. Wu et al. have created HA microfibres by electrospinning apolymer mixture containing Ca(NO3)2 �4H2O, (C2H5O)3PO (precur-sors) and by performing a thermal treatment on the collectedfibres at 600 1C for 1 h [13]. The generation of HA nanofibres hasbeen investigated by Kim and Kim [16]. They have based theirstrategy on the use of sol–gel precursors of the apatites and byadjusting the concentration of the sols, they could vary thediameter of the fibres in a range of a few micrometres to hundredsof nanometres. Calcination was performed at 700 1C for 2 h. Also,Tadjiev et al. have used the same precursors as Wu and colleaguesand heat treatments between 500 and 800 1C to produce β-TCPceramic nanofibres [17].

Sintering of pure HA particles is usually reported to occur above1000 1C. The choice of the sintering temperature is important as ithas an effect on the properties of the resulting sample. Mostinvestigators agree that pure HA (ratio CaP¼1.67) is stable in an

reserved.

P.-A. Mouthuy et al. / Materials Letters 106 (2013) 145–150146

air and argon atmosphere at temperatures upto 1200 1C [18–20].However, decomposition of HA at temperatures as low as 800 1Chas been observed for calcium deficient HA samples [19].

This paper investigates a new approach to obtain CaP sub-microfibres through the electrospinning route. In a previousresearch article [7], we have demonstrated that HA nanoparticlescan be incorporated into Poly(lactic-co-glycolic acid) (PLGA) fibresat high concentration (upto 50% in weight to volume ratio withinthe polymer solution). Here, we report that applying aheat treatment at sintering temperatures to such PLGA–HA com-posite fibres forces the aggregations of the nanoparticles andvaporises the polymer matrix, leading to the creation of CaPsubmicrofibres.

2. Materials and methods

Preparation of electrospinning solutions: Polymer solutions wereprepared by dissolving Poly(lactic-co-glycolic acid) (PLGA, ratio75:25, Mw: 66–107 kDa, Sigma-Aldrich Chemical Company Ltd.,Dorset, UK) into 1,1,1,3,3,3-hexafluoroisopropanol (HFIP, Apollo Scien-tific Limited, Cheshire, UK) at a concentration of 15% (weight tovolume ratio). Deficient hydroxyapatite synthetic nanoparticles withan average particle diameter of 93 nm (HA nanopowder, Sigma-Aldrich) was then added to the polymer solution at concentrations of50% (weight to volume ratio). Solutions were agitated at roomtemperature on a roller for at least 24 h to allow for completedissolution of the polymers. Each solution was then homogenised onice using a vibracell ultrasonicator (130W, 20 kHz; Sonics MaterialsInc., Newtown, USA) prior to electrospinning.

Electrospinning: The polymer solution was transferred to a 1 mlor 5 ml syringe and positioned in the pump for the electrospinningprocess. The solution was pushed towards the electrospinningnozzle at a rate of 1 ml/h and the voltage was set at 12.2 kV. Theelectrospun fibres were collected in the form of a randommesh, asdescribed previously in [7].

Preparation of calcium phosphate fibres: The PLGA–HA electro-spun samples were folded to fit into an alumina ceramic crucibleand the container was placed in the middle of a furnace (Uni-versity of Oxford, Materials Department, Parks Road). The heattreatment was applied in a 50 kPa Argon atmosphere at 1100 1Cfor 1 h. The furnace was brought to a temperature with a heatingrate of 0.27 1C/s and was cooled at a rate of 0.1 1C/s. The treatedsamples were collected next day and were removed forcharacterisation.

Scanning electron microscopy (SEM): Samples (PLGA–HA compo-sites before and after treatment) were mounted on an aluminiumstub using a carbon adhesive disk. The materials were then coatedwith a 2.5 nm layer of platinum using a Cressington 208 HRSputter coater (Vortex Control Systems Inc., Texas, USA) and highresolution images of the scaffolds were taken using a JSM 840Fscanning electron microscope (JEOL, Tokyo, Japan). On eachsample, three different areas were observed randomly at differentmagnifications.

Fourier transform infrared spectroscopy (FTIR): The surfacechemistry of the treated samples was analysed by using a FTIRspectrometer (Bruker Optics Limited, Ettlingen, Germany). PurePLGA fibres, PLGA–HA composite fibres and pure HA nanopowderswere also analysed as controls. Each spectrum was acquired andaveraged with a resolution of 4 cm−1 by accumulation of 128 scansand the signal was measured in a range between 400 and4000 cm−1.

X-ray diffraction (XRD): Samples were prepared by grinding theelectrospun materials into powder and spreading the powder ontoa crystal of silicon covered with a layer of silicone grease. Excesspowder was removed and the crystal and powder were exposed to

an X-ray beam. X-ray diffraction measurements were carried outusing a fully automated Siemens D5000 powder diffractometer,employing a Cu Kα radiation (λ¼0.15406 nm) and a secondarymonochromator. The samples were continuously spun during datacollection and scanned using a step size of 0.051 2θ over the rangeof 51–751 2θ, with a count time of 12 s/step. For comparisonpurposes, the X-ray diffraction of the HA nanopowder and of thePLGA–HA fibres were also recorded.

X-ray photoelectron spectroscopy (XPS): Samples were mountedon a stub using a double sided adhesive tab before being placedinto the ultrahigh vacuum (2�10−7 mPa) analysis chamber of thespectrometer. XPS analysis was carried out using a VG Clam 4 MCDanalyser system with X-ray radiation from the Mg Kα band(hν¼1253.6 eV). Recording was performed with an energy of100 eV and a take-off angle of 901. The spectra were analysedusing Microcal Origin 6.0 software and the assignment of thepeaks was performed with reference to the UKSAF and NISTdatabases. As controls, HA nanopowder and PLGA–HA fibres werealso analysed.

3. Results

Sample morphology: The morphologies of PLGA–HA electrospunfibres before (left side) and after the treatment at 1100 1C (rightside) are shown in Fig. 1. Interestingly, the treated samples(referred as “CaP fibres” in this paper) retain the submicrofibrousmorphology of PLGA–HA composites. Fig. 1(e) and (f) also revealsthat the nanoparticles observed on PLGA–HA fibres have disap-peared to leave a smooth and segmented surface in CaP fibres.

FTIR analysis: The FTIR spectra obtained from the differentsamples are indicated in Fig. 2. The spectra obtained from purePLGA fibres and from HA nanopowder serve as controls. In thePLGA spectrum, the C–O characteristic bands in the region 1065–1280 cm−1 can be seen. On the other hand, the spectrum of HAnanopowder reveals the characteristic peak assigned to PO4

−3: ν1vibration mode at about 964 cm−1, ν3 vibration mode at 1031 cm−1

and 1091 cm−1 (asymmetric). The spectrum of untreated PLGA–HAsample displays both characteristics from PLGA and HA, asdescribed in [7]. On the other hand, the spectrum of CaP fibresseems to have less characteristics of PLGA (CQO and C–O bands).However, the aspect of the curve is slightly different from the onerecorded for the HA nanopowder. Indeed, weak absorption peakscan be detected at about 993, 1043 and 1067 cm−1 near the ν3vibration mode of PO4

−3 (1031 cm−1 and 1091 cm−1). Anothersmall shoulder could be observed at about 950 cm−1 near the ν1vibration mode characteristic (964 cm−1).

XPS analysis: The experimental binding energies of the O 1s, C1s, P 2p, and Ca 2p electrons and proportions of the correspondingatomes are indicated in Table 1. In PLGA–HA fibres, C 1s was foundat 284.4 eV, O 1s at 530.7 eV, Ca 2p at 345.5 eV, and P 2p at130.2 eV. Carbon was the dominant element (51%) while only2.37% of Ca and 2.95% of P were detected in the sample (ratio Ca/P¼1.24). In HA nanopowder, O was found to be the dominantelement with 50.87%, while Ca and P reached values as high as13.94% and 10.13% respectively (ratio Ca/P¼1.37). Ca 2p and P 2pwere found at slightly higher binding energies: 347.5 eV and133.3 eV, respectively. CaP fibres showed very similar observationwhen compared to the pure nanopowder in terms of bindingenergies and percentages of elements. Although the ratio O/C isslightly lower, 1.49 against 2.03 for the HA nanopowder, O was stillthe major element at 46.41%. Ca and P reached values of 12.78%and 9.6%, respectively (ratio Ca/P¼1.33).

XRD analysis: The XRD patterns of the samples are shown inFig. 3. The chemical identification was performed by comparingthe experimental X-ray patterns to standards compiled by the

Fig. 1. SEM images showing the morphology of fibres electrospun from a PLGA solution containing 50% (w/v) of HA nanoparticles (left side) and the morphology of the samefibres after the treatment at 1100 1C (right side). Images are shown at different magnifications to reveal surface structures.

P.-A. Mouthuy et al. / Materials Letters 106 (2013) 145–150 147

International Centre for Diffraction Data (ICDD). The analysis showthat the diffractogram obtained for the PLGA–HA sample and HAnanopowder identified synthetic hydroxyapatite (Ca5(PO4)3OH)with a probability above 99% for both. The pattern observed forCaP fibres identified hydroxyapatite with a probability of 81% withsmall amounts of whitlockite (Ca9(Mg,Fe++)(PO4)6 � (PO3OH)) andβ-TCP (Ca3(PO4)2) with probabilities of 14% and 11% respectively.

4. Discussions

In previous research, CaP fibres have been successfully producedby electrospinning through the combinative use of chemical pre-cursors and heat treatments at temperature below 800 1C [13–17].The strategy presented in this paper differs in the way that the CaPmaterial was directly incorporated into the electrospun compositefibres in the form of nanoparticles. Furthermore, the electrospin-ning process was followed by the heat treatment at sinteringtemperatures, i.e. above 1000 1C. This simple way of creating CaPfibres avoids the use of chemical precursors, often toxic andexpensive, and potentially allows the creation of nano- andmicro-fibres from a wide variety of ceramic material. Here, we

successfully demonstrated the approach with PLGA as a sacrificialpolymer. However, cheaper materials such as polyvinylalcohol andpolyvinylpyrrolidone could be used since these polymers havealready been shown to be suitable for the preparation of ceramicfibres using a high temperature treatment [15]. Polymers havinglow melting points, such as polycaprolactone, were not able tomaintain the fibrous structure during the thermal treatment (datanot shown). To enable the sintering process, particles must bedensely packed. Therefore fibres prepared from a solution contain-ing highly concentrated HA nanoparticles (50% w/v) were selectedfor this work. Moreover, the large amount of nanoparticles in thepolymer matrix might have played a role in maintaining the fibrousmorphology while the polymer was subsequently molten andvaporised.

The SEM images have revealed that the resulting CaP fibresacquire a segmented aspect, while the nanoparticles (seen at thesurface of PLGA–HA composites) are not visible anymore (Fig. 1,right side). These observations suggest that sintering has occurred.In sintering processes, small individual structures merge togetherby diffusion to form bigger structures. To confirm the chemicalnature of the heat-treated samples, further characterisationsinvolving FTIR, XPS and XRD were performed.

Table 1Experimental binding energies of the O 1s, C 1s, P 2p, and Ca 2p electrons andproportions of the corresponding elements measured in PLGA–HA fibres, HAnanopowder and CaP fibres by XPS analysis.

Element Binding energy (eV) Proportion (%)

PLGA–HA fibres O 1s 530.704 41.78C 1s 284.404 51.28P 2p 131.217 2.37Ca 2p 345.478 2.95

HA nanopowder O 1s 531.744 50.87C 1s 284.442 25.07P 2p 133.258 10.13Ca 2p 347.519 13.94

CaP fibres O 1s 531.485 31.21C 1s 285.185 46.41P 2p 133 9.60Ca 2p 347.261 12.78

Fig. 2. FTIR spectra of (a) PLGA fibres, (b) HA nanopowder, (c) PLGA–HA composite fibres, and (d) sintered PLGA–HA fibres. The pure PLGA spectrum shows the C–Ocharacteristic bands in the region 1065–1280 cm−1. The spectrum of HA nanopowder reveals the characteristic peak assigned to PO43−: ν1 vibration mode at about 964 cm−1,ν3 vibration mode at 1031 cm−1 and 1091 cm−1 (asymmetric) respectively. The spectrum of untreated PLGA–HA 50% displays both characteristics of PLGA and HA. Howeverthe spectrum of CaP fibres shows the loss of the characteristics of PLGA (C–O bands), suggesting the successful thermal degradation of the polymer, while the characteristicsof HA (PO43−) remain present.

P.-A. Mouthuy et al. / Materials Letters 106 (2013) 145–150148

FTIR analysis confirmed that the polymer was burnt out andevaporated during the thermal treatment since no characteristic ofPLGA was observed in the spectrum of CaP fibres. Moreover, noband corresponding to carbonates was observed in the range1400–1550 cm−1, which suggests that carbonate ions were onlypresent in trace quantities. This may be due to the volatile natureof the carbonate group. Also, the emerged peaks from the splittingof the ν1 and ν3 vibration modes are characteristic of tricalciumphosphate [21]. This suggests that the decomposition of HAstarted during the heat treatment. Moreover, consultation of theliterature indicates that the spectra observed for CaP fibres istypical for a mixture of HA and β-tricalcium phosphate [22–24].This last observation is in agreement with the outcomes of theXRD analysis: HA was identified as the main phase of the CaP

fibres, but β-tricalcium phosphates (β-TCP) and whitlockite (a Mg-containing β-TCP) were also identified.

The XPS data clearly indicates that the HA nanoparticles used inthis experiment are deficient in calcium (Ca/P ratio¼1.37). Thiscould explain why the HA decomposition occurs before 1200 1Csince deficient HA start their decomposition at temperatures lowerthan pure HA (ratio¼1.67) [19]. In the literature, this decompositionis described as partial although the reason remains unknown. Thecommonly accepted decomposition reaction for deficient HA is

Ca10(PO4)6(OH)2-3Ca3(PO4)2+CaO+H2O

The XPS analysis also reveals low calcium phosphate levels inthe PLGA–HA samples, despite the fact that the fibres are filledwith a large amount of nanoparticles. Since XPS is a surfacechemical analysis technique (with analysis depth less than10 nm), this suggests that the nanoparticles lying at the surfaceof the fibres were embedded in a thin layer of polymer. Thepresence of carbon observed by XPS at the surface of the HAnanoparticles might result from atmospheric contamination byCO2 and volatile fatty acids. At the surface of the CaP fibres, anadditional reason for the presence of carbon might be theexistence of residual atoms from the degradation–evaporation ofthe PLGA polymer. This would explain the higher proportion ofcarbon when compared to the native HA nanopowder and why theratio of oxygen to carbon is lower. The slightly lower percentage ofoxygen in CaP fibres, compared to the HA nanopowder, may resultfrom the loss of OH groups (dehydroxylation) that occur at hightemperature. For pure HA, this reaction is known to occur atsintering temperatures below 1200 1C with a conversion degree of0.4 to 0.5 [25]. A simple way to present the reaction formula fordehydroxylation would be

Ca10(PO4)6(OH)2-Ca10(PO4)6O+H2O

Fig. 3. XRD patterns obtained for (a) PLGA–HA fibres, (b) HA nanopowder, and (c) CaP fibres. In (a) and (b), Hydroxyapatite (Ca5(PO4)3OH was identified at significant levels(p¼99.84% and 99.88%, respectively). In (c), Hydroxyapatite (Ca5(PO4)3OH, p¼81.16%), whitlockite (Ca9(Mg,Fe++)(PO4)6 (PO3OH), p¼13.72%) and β-TCP (Ca3(PO4)2,p¼10.92%) were identified at significant levels.

P.-A. Mouthuy et al. / Materials Letters 106 (2013) 145–150 149

The resulting oxyhydroxyapatite cannot be detected by XRDdue to the high similarity of its crystalline structures with hydro-xyapatite. It could not be seen by FTIR either, since the bandcharacteristic of –OH was observed below 400 cm−1 which isbelow the range explored in this study.

Systems containing both HA and TCP phases have beenintensively investigated and have shown potential for bone tissueengineering [26,27]. β-TCP has exceptionally good tissue compat-ibility, allowing direct bonding between the regenerated and thenative bone without intermediate connective tissue, and encoura-ging faster bone regeneration and resorbability than HA. HA/TCPsystems have been developed in order to regulate the resorptionkinetics through varying the ratio HA/TCP. For similar reasons, thechemistry of the CaP fibres is likely to induce an excellent responsefrom bone tissue.

Measurements of mechanical robustness of the CaP fibresremain to be conducted, but initial inspection has indicatedprofound scaffold fragility compared to the PLGA–HA composites.Although the polymer-based fibres are more flexible and lessfragile, the absence of polymer matrix may be preferred forapplications such as bone tissue engineering and drug delivery.For PLGA in particular, the acidity of the polymer degradationproduct can cause tissue inflammation or deactivation of the drug.If bone tissue engineering applications are considered, whererobustness is required, CaP fibres will necessitate further improve-ment to their mechanical properties. Alternatively, in their current

condition, fibres may be more appropriate as coating or fillingmaterial for ceramic or metal implants, in order to enhanceosteoconduction and facilitate the incorporation of the implantinto the native bone tissue.

5. Conclusions

To conclude, the results presented in this paper have demon-strated that CaP fibres can be prepared by electrospinning a PLGA–HA composite solution and subsequently applying a thermaltreatment at sintering temperature to the electrospun fibres. Thepolymer phase was evaporated at the high temperatures while thenanoparticles were sintered, allowing the samples to retain theirsubmicrofibrous morphology. Hydroxyapatite was identified to bethe main phase in the treated samples. Tricalcium phosphateswere also found, as a result of the dehydroxylation and decom-position of deficient HA at high temperature. The chemistry of thenew fibres holds great potential for bone tissue engineering.However, more investigations are to be carried out, in particularto improve their mechanical properties, as these are currentlyinsufficient for the fibres to be used as bone scaffold. Nevertheless,this simple approach is promising for the creation of submicro-fibres made of a wide variety of ceramic materials without the useof toxic and expensive chemical precursors.

P.-A. Mouthuy et al. / Materials Letters 106 (2013) 145–150150

Acknowledgements

The authors would like to thank EPSRC for supporting PAMouthuy's Ph.D. study with a studentship. Also we thank Mr IanLloyd for his assistance in the sample preparation.

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.matlet.2013.04.110.

References

[1] Kokubo T, Takadama H. How useful is SBF in predicting in vivo bonebioactivity? Biomaterials 2006;27:2907–15.

[2] Bose S, Tarafder S. Calcium phosphate ceramic systems in growth factor anddrug delivery for bone tissue engineering—a review. Acta Biomater2012;8:1401–21.

[3] Habraken WJEM, Wolke JGC, Jansen JA. Ceramic composites as matrices andscaffolds for drug delivery in tissue engineering. Adv Drug Delivery Rev2007;59:234–48.

[4] Klein CPAT, de Blieck-Hogervorst JMA, Wolke JGC, de Groot K. Studies of thesolubility of different calcium phosphate ceramic particles in vitro. Biomater-ials 1990;11:509–12.

[5] Nery EB, LeGeros RZ, Lynch KL, Lee K. Tissue response to biphasic calciumphosphate ceramic with different ratios of HA/βTCP in periodontal osseousdefects. J Periodontol 1992;63:729–35.

[6] Ducheyne P. Bioceramics: material characteristics versus in vivo behavior.J Biomed Mater Res 1987;21:219–36.

[7] Mouthuy PA, Ye H, Triffitt J, Oommen G, Cui Z. Physico-chemical characterisa-tion of functional electrospun scaffolds for bone and cartilage tissue engineer-ing. Proc Inst Mech Eng Part H J Eng Med 2010;224:1401–14.

[8] Ramanan SR, Venkatesh R. A study of hydroxyapatite fibers prepared via sol–gel route. Mater Lett 2004;58:3320–3.

[9] Tanahashi M, Kamiya K, Suzuki T, Nasu H. Fibrous hydroxyapatite, grown inthe gel system: effects of pH of the solution on the growth rate andmorphology. J Mater Sci: Mater Med 1992;3:48–53.

[10] Nakahira A, Sakamoto K, Yamaguchi S, Kaneno M, Takeda S, Okazaki M. Novelsynthesis method of hydroxyapatite whiskers by hydrolysis of alpha-

tricalcium phosphate in mixtures of water and organic solvent. J Am CeramSoc 1999;82:2029–32.

[11] Kolos EC, Ruys AJ, Roger GJ. A novel biomimetic approach for the manufactureof hydroxyapatite fibres. Key Eng Mater: Bioceram 2003;15:47–50.

[12] Ota Y, Iwashita T, Kasuga T, Abe Y. Novel preparation method of hydroxya-patite fibers. J Am Ceram Soc 1998;81:1665–8.

[13] Wu Y, Hench LL, Du J, Choy KL, Guo J. Preparation of hydroxyapatite fibers byelectrospinning technique. J Am Ceram Soc 2004;87:1988–91.

[14] Sigmund W, Yuh J, Park H, Maneeratana V, Pyrgiotakis G, Daga A, et al.Processing and structure relationships in electrospinning of ceramic fibersystems. J Am Ceram Soc 2006;89:395–407.

[15] Ramaseshan R, Sundarrajan S, Jose R, Ramakrishna S. Nanostructured ceramicsby electrospinning. J Appl Phys: Focussed Rev 2007;102:1–17.

[16] Kim. HW, Kim HE. Nanofiber generation of hydroxyapatite and fluor-hydroxyapatite bioceramics. J Colloid Interface Sci 2006;77:323–8.

[17] Tadjiev TR, Chun SS, Kim HM, Kang IK, Kim SY. Preparation and characteriza-tion of β-tricalcium phosphate nanofibers via electrospinning. Key Eng Mater:Bioceram 2007;342:817–20.

[18] Ye H, Liu XY, Hong H. Characterization of sintered titanium/hydroxyapatitebiocomposite using FTIR spectroscopy. J Mater Sci: Mater Med 2009;20:843–50.

[19] Zhou J, Zhang X, Chen J. High temperature characteristics of synthetichydroxyapatite. J Mater Sci: Mater Med 1993;4:83–5.

[20] Muralithran G, Ramesh S. The effects of sintering temperature on the proper-ties of hydroxyapatite. Ceram Int 2000;26:221–30.

[21] Jillavenkatesa A, Condrate Sr RA. The infrared and raman spectra of α-andβ-tricalcium phosphate (Ca3(PO4)2). Spectrosc Lett 1998;31:1561–70.

[22] LeGeros RZ, Lin S, Rohanizadeh R, Mijares D, LeGeros JP. Biphasic calciumphosphate bioceramics: preparation, properties and applications. J Mater Sci:Mater Med 2003;14:201–9.

[23] Slosarczyk A, Paluszkiewicz C, Gawlicki M, Paszkiewicz Z. The FTIR spectro-scopy and QXRD studies of calcium phosphate based materials produced fromthe powder precursors with different Ca/P ratios. Ceram Int 1997;23:297–304.

[24] Lin FH, Liao CJ, Chen KS, Sun JS. Preparation of high-temperature stabilizedbeta-tricalcium phosphate by heating deficient hydroxyapatite withNa4P2O7 � 10H2O addition. Biomaterials 1998;19:1101–7.

[25] Cihlar J, Buchal A, Trunec M. Kinetics of thermal decomposition of hydro-xyapatite bioceramics. J Mater Sci 1999;34:6121–31.

[26] Barinov SM. Calcium phosphate-based ceramic and composite materials formedicine. Russ Chem Rev 2010;79:13–29.

[27] Zhang L, Hanagata N, Maeda M, Minowa T, Ikoma T, Fan H, et al. Poroushydroxyapatite and biphasic calcium phosphate ceramics promote ectopicosteoblast differentiation from mesenchymal stem cells. Sci Technol AdvMater 2009;10:1–9.