Synthesis, mechanical and biological characterization of ionic doped carbonated hydroxyapatite/ β-tricalcium phosphate mixtures

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

Author's personal copy

Synthesis, mechanical and biological characterization of ionic dopedcarbonated hydroxyapatite/b-tricalcium phosphate mixtures

S. Kannan a, S.I. Vieira b, S.M. Olhero c, P.M.C. Torres a, S. Pina a, O.A.B. da Cruz e Silva d, J.M.F. Ferreira a,⇑a Department of Ceramics and Glass Engineering, University of Aveiro, CICECO, 3810-193 Aveiro, Portugalb Department of Biology, Centre for Cellular Biology, University of Aveiro, 3810-193 Aveiro, Portugalc Department of Mechanical Engineering and Industrial Management, FEUP, University of Porto, Porto, Portugald Department of Health Sciences, Centre of Cellular Biology, University of Aveiro, 3810-193 Aveiro, Portugal

a r t i c l e i n f o

Article history:Received 25 June 2010Received in revised form 2 December 2010Accepted 6 December 2010Available online 10 December 2010

Keywords:Hydroxyapatiteb-Tricalcium phosphateCarbonatesIonic substitutionsMechanical features

a b s t r a c t

The influence of ionic substituents in calcium phosphates intended for bone and tooth replacement bio-medical applications is an important research topic, owing to the essential roles played by trace elementsin biological processes. The present study investigates the mechanical and biological evaluation of ionicdoped hydroxyapatite/b-tricalcium phosphate mixtures which have been prepared by a simple aqueousprecipitation method. Heat treating the resultant calcium phosphates in a carbonated atmosphere led tothe formation of ionic doped carbonated hydroxyapatite/b-tricalcium phosphate mixtures containing theessential ions of biological apatite. The structural analysis determined by Rietveld refinement confirmedthe presence of hydroxyapatite as the main phase, together with a considerable amount of b-tricalciumphosphate. Such phase assemblage is essentially due to the influence of substituted ions during synthesis.The results from mechanical tests proved that carbonate substitutions are detrimental for the mechanicalproperties of apatite-based ceramics. In vitro proliferation assays of osteoblastic-like cells (MC3T3-E1 cellline) to powders revealed that carbonate incorporation can either delay or accelerate MC3T3 prolifera-tion, although reaching the same proliferation levels as control cells after 2 weeks in culture. Further,the powders enable pre-osteoblastic differentiation in a similar manner to control cells, as indirectlymeasured by ALP activity and Type-I collagen medium secretion.

� 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Hydroxyapatite (Ca10(PO4)6(OH)2, HAP) and b-tricalcium phos-phate (b-Ca3(PO4)2, b-TCP) as bone substitutes are well knownfor their bioactivity and osteoconductive properties in vivo [1,2].The apatite phase that makes up the inorganic component of hardtissue can be described as a calcium-deficient apatite (Ca/P molarratio <1.67) along with several multi-substituted ions (Na+, Mg2+,K+, F�, Cl�, CO3

2�) in its structural lattice. Table 1 [3–5] reportsthe comparative composition and lattice parameters of inorganicapatites of adult human calcified tissues. The type and amount ofionic substitutions in the apatite phase of both bone and toothmay vary depending on the individual. Although these levels ofsubstitution are small, the different ions play a major role in thebiochemistry of the hard tissue [6–11]. A previous study reportedby the authors dealt with the incorporation of essential ions (bar-ring carbonates) in the structural lattice of HAP and b-TCP mixtures[12]. In that particular study, it was concluded that cations (Na+,Mg2+ and K+) were incorporated at the structural lattice of b-TCP,

whereas anions (Cl� and F�) tend to stabilize the structure ofHAP. However, the carbonates that constitute a major part in thebiological apatite next to calcium and phosphorous were not re-ported in that study.

The substitution of carbonates in HAP structure has been re-ported before [13,14], whereas the available studies concerningthe incorporation of carbonates along with cations in the apatitestructure are restricted to co-substituted hydroxyapatites withthe ionic pairs of Na+/CO3

2� and Mg2+/CO32� [15–18]. During the

cationic (Na+, Mg2+) and anionic (CO32�) co-substitutions, the for-

mation of pure HAP phase was confirmed with the charge imbal-ance between Na+ and Ca2+ ions being restored by CO3

2� ions inthe apatite structure [16]. Mg2+ ions prefer to stabilize the b-TCPphase during heat treatment, but tend to restore single-phaseHAP in the presence of CO3

2� ions even at 1300 �C [17]. A detailedstudy on Na+/CO3

2� co-substituted hydroxyapatite through Riet-veld refinement was also reported recently by Feki et al. [19].The major concern with the incorporation of carbonates in the apa-tite structure is its preservation during high-temperature heattreatment. The present study is an attempt to incorporate carbon-ates into the apatite structure along with the essential ions (Na+,Mg2+, K+, F�, Cl�) that are present in the biological apatite. A simple

1742-7061/$ - see front matter � 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.actbio.2010.12.009

⇑ Corresponding author. Tel.: +351 234 370242; fax: +351 234 370204.E-mail address: [email protected] (J.M.F. Ferreira).

Acta Biomaterialia 7 (2011) 1835–1843

Contents lists available at ScienceDirect

Acta Biomaterialia

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

Author's personal copy

aqueous precipitation technique followed by heat treatmentunder CO2 atmosphere was employed in the present study. Theresultant materials were subjected to structural characterizationand mechanical evaluation.

From a biological perspective, it has been shown that some bio-materials are able to modify directly the osteoblastic proliferationrate and its differentiation, such as the synthesis of alkaline phos-phatase (ALP) [20–23]. Therefore, in vitro proliferation and differ-entiation responses of osteoblast-like cells (MC3T3-E1 cell line)to the powders were studied. Various biological responses to thebiomaterials were assayed, including cell viability determinationsby the resazurin assay, photometric evaluation of ALP activity insecreted medium, and Type-I collagen secretion at the differentia-tion period.

2. Experimental procedure

2.1. Preparation of powders

The synthesis of pure HAP and of different types of ionic dopedcalcium phosphates was carried out according to procedure de-scribed by the authors elsewhere [12,24,25]. Calcium nitrate tetra-hydrate (Ca(NO3)2�4H2O; Vaz-Pereria-Portugal), diammoniumhydrogen phosphate ((NH4)2HPO4; Vaz-Pereria-Portugal), sodiumnitrate (NaNO3; Vaz-Pereria-Portugal), magnesium nitrate hexahy-drate (Mg(NO3)2�6H2O; Merck), potassium nitrate (KNO3, Merck),ammonium chloride (NH4Cl; Merck) and ammonium fluoride(NH4F; Merck) were used as starting chemical precursors forCa2+, PO4

3�, Na+, Mg2+, K+, Cl� and F� ions, respectively, for the

synthesis. Two different types of compositions were attempted toform ionic doped calcium phosphates, which were then heattreated in different atmospheres: air (BAP-1 and BAP-2) and CO2

(BAC-1 and BAC-2) (Table 2). For this purpose, a solution mixturecontaining anions (PO4

3�, F� and Cl�) was added slowly at a rateof 50 mL min�1 to the solution mixture containing nitrates ofcations (Ca2+, Na+, Mg2+ and K+) stirred at a rate of 1000 rpm (solu-tion concentrations of precursors are detailed in Table 2). After theaddition of all precursors, the pH of the mixed solution was �3,being then increased to pH 10 by adding concentrated ammoniumhydroxide (NH4OH; Sigma–Aldrich) solution. After the completionof addition, the reaction was performed at 90 �C for 2 h under aconstant stirring rate of 1000 rpm. Pure stoichiometric HAP with-out any additives was prepared under the same conditions to com-pare the results. The precipitated suspension was poured out fromthe reactor and allowed to settle down for 24 h for the maturationof precipitate. After 24 h, the precipitates were separated throughvacuum filtration and dried at 80 �C overnight. The precipitateswere washed and filtered repeatedly with de-ionized water. Thedried cakes were ground to fine powders and subjected to heattreatments in both controlled air and carbonated atmospheres atwide temperatures. Heat treatments in air were carried out in aThermolab furnace (Pt30%Rh/Pt6%Rh-thermocouple) with a heat-ing rate of 5 �C min�1 to achieve a predetermined temperaturerange and a dwelling time of 2 h and again cooled to room temper-ature at the rate of 5 �C min�1. Heat treatments in carbonatedatmosphere were carried out in the same furnace with a CO2 gasflow under 400 kPa and a heating rate of 5 �C min�1 to achieve apredetermined temperature range and a dwelling time of 2 h and

Table 1Composition of human biological apatite [3–5].

Composition (wt.%) Enamel Dentin Bone HA

Calcium (Ca) 36.5 35.1 34.8 39.6Phosphorus (P) 17.7 16.9 15.2 18.5Ca/P (molar ratio) 1.63 1.61 1.71 1.67Sodium (Na) 0.5 0.6 0.9 –Magnesium (Mg) 0.44 1.23 0.72 –Potassium (K) 0.08 0.05 0.03 –Carbonate (as CO3

2�) 3.5 5.6 7.4 –Fluoride (F) 0.01 0.06 0.03 –Chloride (Cl) 0.30 0.01 0.13 –

Crystallographic properties: Lattice parameters (±0.003 Å)a-axis (Å) 9.441 9.421 9.41 9.430c-axis (Å) 6.880 6.887 6.89 6.891Ignition products (800 �C) b-TCP + HA b-TCP + HA HA + CaO HATypical crystal sizes (nm) 100 lm � 50 � 50 35 � 25 � 4 50 � 25 � 4 200–600

Table 2Molar concentrations of the precursors used in the synthesis; the planned elemental compositions for samples BAP-1 and BAP-2 were exactly the same as for samples BAC-1 andBAC-2, only differing in the atmospheres (air or CO2, respectively), of heat treatments; in a similar manner, the stoichiometric hydroxyapatite is designated as HAP when heattreated in air and as BAC when heat treated in CO2 atmosphere.

Elements Sample code: BAC-1 Sample code: BAC-2

Planned concentration ofthe elements duringsynthesis (wt.%)

Elemental analysis afterheat treatment in CO2

atmosphere at 900 �C (wt.%)

Planned concentration ofthe elements duringsynthesis (wt.%)

Elemental analysis afterheat treatment in CO2

atmosphere at 900 �C (wt.%)

Ca 39.89 38.98 (±0.02) 39.89 39.31 (±0.02)P 18.50 18.15 (±0.02) 18.50 18.32 (±0.02)Na 0.572 0.49 (±0.02) 1.144 1.09 (±0.02)Mg 0.484 0.51 (±0.02) 0.968 0.93 (±0.02)K 0.584 0.46 (±0.02) 1.168 1.07 (±0.02)Cl 0.050 0.04 (±0.02) 0.100 0.09 (±0.02)F 0.050 0.032 (±0.01) 0.100 0.061 (±0.01)CO3 – 4.74 (±0.03) – 4.67 (±0.03)Ca/P ratio 1.67 1.66 (±0.01) 1.67 1.66 (±0.01)(Ca + Na + Mg + K)/

P ratio1.76 1.75 (±0.01) 1.86 1.85 (±0.01)

1836 S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843

Author's personal copy

again cooled to room temperature at the rate of 5 �C min�1. Theheat-treated powders were ground well once again and crushedto particle size <25 lm and used for characterization studies.

2.2. Powder characterization

The qualitative and quantitative analysis of crystalline phases inthe resultant calcium phosphates were determined by X-raydiffraction (XRD) analysis using a conventional Bragg–Brentanodiffractometer (Philips PW 3710, Eindhoven, The Netherlands)with Ni-filtered Cu Ka radiation. XRD data were recorded in the2h range of 5–140� (step size 0.02� and 6 s of counting time foreach step). The phase fractions were extracted by Rietveld-RIRrefinements, using GSAS software [26] and EXPGUI [27]. The back-ground was successfully fitted with a Chebyshev function with avariable number of coefficients, depending on its complexity. Thepeak profiles were modelled using a pseudo-Voigt function withone Gaussian and one Lorentzian coefficient. Lattice constants,phase fractions and coefficients corresponding to sample displace-ment and asymmetry were also refined.

Elemental analyses for the presence of all elements except fluo-rine and carbonates were made using X-ray fluorescence spectros-copy (Philips PW2400 X-ray fluorescence spectrometer). Thevacuum of the chamber was lower than 2 Pa. The error associatedwith each chemical element could be determined as ±1 of the lastdigit of the measured values. Fluorine was determined by theselective ion electrode method. Carbonates were determined byelemental analysis (CE110 coupled with Thermal Finnigan DeltaPlus mass spectrometer). The investigated samples, each with5 lg, were heated up to 1020 �C in He atmosphere over a chromatecatalyst. IR spectra for the powders were obtained by IR Fourierspectrometry (FTIR; model Mattson Galaxy S-7000, USA). Forthis purpose each powder was mixed with KBr in the proportionof 1/150 (by weight) for 15 min and pressed into a pellet using ahand press. Each IR spectrum was the average of 32 scans collectedat 2 cm�1 resolution at room temperature.

2.3. Preparation of samples by slip casting and mechanical testing

The apatite powders were first subjected to calcination at900 �C in both air and carbonated atmospheric conditions in orderto remove residual nitrates and adsorbed water. Aqueous suspen-sions containing 50 vol.% solids loading were then prepared by dis-persing the apatite powders in de-ionized water with the help of0.4 wt.% Targon 1128 (BK Ladenburg, Germany), based on the drymass of solids. The suspensions were de-agglomerated by ball mill-ing for 24 h in polypropylene bottles using alumina balls (12 mmdiameter) by maintaining a 1:3 weight ratio between powderand balls. After removing the balls, the suspensions were degassedfor 5 min by vacuum pumping. Afterwards, the suspensions werecast into plaster moulds to consolidate samples with dimensions50 � 5 � 5 mm3. The partial dehydration of the green samples un-der ambient conditions was followed by two drying steps of 12 heach at 40 �C and then at 90 �C. The dried slip cast bodies were sin-tered for 2 h at different temperatures (900, 1100, 1200 and1300 �C) in air/CO2 atmospheres. The evaluated mechanical prop-erties of the sintered samples were hardness (H) and flexuralstrength. The data on hardness were collected using a hardnesstester (Leitz Wetzler, Germany) by holding the indenter tip (with137�) under 5 kg load for 15 s on the surface of the sample polishedto mirror finish. The results presented are the average of at least 15indentations per specimen in a total of five specimens per group.The flexural strength was measured using a 3-point bend test(JIS-R1601) with a load cell of 10 kN and a support span of30 mm [28]. About 15–20 samples were tested per case in the3-point bend test analysis, and the results were averaged out for

consideration. The bulk density (BD), apparent porosity and waterabsorption (WA) capacity of the sintered samples were measuredaccording to Archimedes’ principle (ASTM C372) using a Mettlerbalance (AG 245, Mettler Toledo, Nänikon, Switzerland). On aver-age, three density measurements were performed for each samplein this study (±0.01 error).

2.4. MC3T3-E1 cell culture and resazurin cytotoxicity assay

The powders were crushed and ground to particle size <25 lmfollowed sterilization by c-radiation prior to being tested in cellcultures. The osteoblastic cell line MC3T3-E1 (ATCC CRL-2593)was used for the biological characterization. MC3T3-E1 cells weremaintained at 37 �C in a humidified atmosphere of 5% CO2 in air, in2 mM glutamine-containing minimum essential a-medium in Ea-gle’s balanced salt solution supplemented with 10% (v/v) fetal bo-vine serum, 1% (v/v) of a 100 U mL�1 penicillin and 100 mg mL�1

streptomycin solution (Gibco BRL, Invitrogen) and 3.7 g L�1 NaH-CO3. Sub-confluent cultures (80–90% confluence) were split 1:5using a 0.25% trypsin/EDTA (Gibco BRL, Invitrogen) solution at 5%CO2, 37 �C. In order to determine the powders’ cytotoxicity, thenon-toxic and reversible resazurin metabolism assay was used[29].

The range of linearity between resazurin reduction to resorufinand the density of MC3T3-E1 viable cells was first determined, beingobserved to occur at a cellular density between 1 and7 � 104 cells cm�2 [30]. To determine the dose-dependent cell via-bility and proliferation when exposed to the powders, 1 � 105 cellswere seeded in 35 mm wells (density 1 � 104 cells cm�2), and cul-tured in cell media supplemented with 0 or 1 mg mL�1 of powders.This powder concentration chosen as 1 mg mL�1 already resultedin a saturated solution. Powder formulation was chosen, to the det-riment of bulk samples, mainly due to the higher surface area of con-tact between the biomaterials and cells and conditioned mediumwhen using powder formulation. Following 1.5, 3, 7, 15, 23 and30 days of incubation at 37 �C in a humidified 5% CO2 and 95% airatmosphere, growth medium was removed by aspiration and re-placed with fresh medium containing 10% of a resazurin (Sigma Al-drich) solution (0.1 mg mL�1 resazurin in phosphate buffer saline(Pierce, Perbio)). After 4 h of incubation, 1 mL of the medium wascollected, and cell viability was determined by spectrophotometri-cally (Cary 50 BIO, Varian) measuring resazurin reduction at 570and 600 nm. For each day, a final resazurin value (ODf) was calcu-lated as the ratio OD 570/OD 600 nm minus the OD 570/OD600 nm ratio of a negative control (resazurin media incubated for4 h in the absence of cells). For each time point, the ODf levels of cellsincubated with 0 mg mL�1 powder were taken as 100%, and cell via-bility was calculated as a percentage of these control values. Allexperiments were carried out in triplicate and expressed as themean ± standard error of the mean (SEM). Cells morphology wasevaluated by phase contrast (PhC) microscopy in an inverted Olym-pus IX81 epifluorescence microscope upon 1 day in culture, andmicrophotographs were taken at 18 �C with a Digital CCD mono-chrome camera F-View II (Soft Imaging System).

2.5. Determination of ALP activity

MC3T3-E1 cells were seeded at 1 � 104 cells cm�2 in 35 mmwells and cultured on 0 (control) and 1 mg mL�1 of powders. ALPactivity of exposed cells was evaluated at 1.5, 3, 7, 15, 23 and30 days of incubation at 37 �C by a modified BCIP-NBT assay [31–34]. This assay is based on a chromagenic reaction initiated bythe cleavage of the phosphate group of BCIP by ALP present inthe cells media. This reaction produces a proton, which reducesNBT to an insoluble purple precipitate. Briefly, 200 lL of eachwell-conditioned medium was incubated with 100 lL of a BCIP/

S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843 1837

Author's personal copy

NBT (Promega) 1:2 solution in universal ALP buffer (100 mM citricacid, 100 mM KH2PO4, 100 mM sodium tetraborate decahydrated,100 mM Tris, 100 mM KCl; pH 11), following the manufacturer’sinstructions. After 2 h of incubation at 37 �C, the purple precipi-tates were solubilized with 600 lL of a 10% SDS solution. ALPactivity was determined spectrophotometrically at 560 nm and ex-pressed as OD, upon subtraction of the blank OD (200 lL of freshcell media plus BCIP/NBT substrate, dissolved in 10% SDS).

2.6. Type-I collagen medium secretion upon powder exposition

MC3T3-E1 cells were seeded at 1 � 104 cells cm�2 in 35 mmwells and cultured on 0 and 1 mg mL�1 of powders for 23 days at37 �C. Cell conditioned media were collected into 1% SDS solutions,and cells were harvested with a 1% SDS boiling solution. Sampleswere boiled for 10 min and sonicated, and the total mass contentwas determined using the BCA kit (Pierce). Mass-normalized cellmedia samples were subjected to reducing 7.5% SDS–PAGE inTris–Glycine buffer, and electrophoretically transferred onto nitro-cellulose membranes. Precision Plus Dual Color (Bio-Rad) wereused as protein standards. Immunoblotting of the transferred pro-teins was performed by incubating membranes O/N with the anti-

Collagen Type-I primary antibody (Novus Biologicals, Germany),after blocking non-specific binding sites with non-fat dried milkin TBS-T (10 mM Tris–HCl at pH 8.0, 150 mM NaCl, 0.5% Tween).Detection was achieved using a horseradish peroxidase-linked sec-ondary antibody and an ECL kit (GE Healthcare).

2.7. Statistical analysis

Statistical significance analysis was conducted using SPSS 16.0software by one-way analysis of variance, followed by theTukey–Kramer post hoc test, with the level of statistical signifi-cance set at p < 0.05.

3. Results and discussion

3.1. Phase evolution

The XRD patterns and FTIR spectra of the as-dried powders at80 �C revealed the presence of poorly crystalline apatite phase rep-resented by broad diffraction patterns. No significant details couldbe evidenced from the as-dried powders owing to the presence ofexcess adsorbed water, residual nitrates and ammonium ions. Heat

Fig. 1. Observed (crosses), calculated (continuous line) and difference curve from the Rietveld refinement of (a) BAC-1 and (b) BAC-2 samples heat treated at 1300 �C for 2 h inCO2 atmosphere. Markers representing the phase reflections correspond to HAP, b-TCP from bottom to top. (Refined parameters for BAC-1 composition: Rwp = 0.042,Rp = 0.021, v = 3.15; refined parameters for BAC-2 composition: Rwp = 0.039, Rp = 0.027, v = 3.11).

1838 S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843

Author's personal copy

treatment of the powders under a carbonated atmosphere at awide range of temperatures led to the formation of crystalline apa-tite phase. The elemental analysis data presented in Table 2 con-firm the presence of all the intended elements in the powders.No carbonates were detected in the powders that were heat trea-ted in an air atmosphere (HAP, BAP-1 and BAP-2), whereas thepresence of carbonates was confirmed in the powders heat treatedin a CO2 atmosphere (BAC, BAC-1 and BAC-2). The carbonate con-tent in the BAC composition was estimated as 4.62 (±0.03) wt.%from the analysis. Similar carbonate contents were measured forthe BAC-1 and BAC-2, as reported in Table 2. The refined XRD pat-terns presented in Fig. 1 confirmed the expected evolution of crys-talline phases. HAP and b-TCP phases, commonly known asbiphasic calcium phosphate mixtures, were detected in the refinedXRD patterns. The BAC-1 and BAC-2 compositions were confirmedto possess Ca/P molar ratio of 1.66 (±0.01) after elemental analysis,in good consistency with the planned values. The BAC-1 powdershowed a phase composition of 93.7 wt.% HAP and 6.3 wt.%b-TCP, while the BAC-2 consisted of 86.4 wt.% HAP and 13.6 wt.%b-TCP. In spite of the similar Ca/P molar ratios of both BAC-1 andBAC-2 powders, the observed variations in the weight ratio ofHAP/b-TCP phases between the powders are mainly attributed totheir different levels of ionic (Na+, Mg2+, K+, F�, Cl� and CO3

2�) sub-stitutions. The BAP-1 powder showed a phase composition of75.4 wt.% HAP and 24.6 wt.% b-TCP, while the BAP-2 indicated66.3 wt.% HAP and 33.7 wt.% b-TCP. This significant variation inthe phase composition observed between BAP-1 and BAP-2 pow-ders is mainly due to the different substitution levels of ions(Na+, Mg2+, K+, F�, Cl�) in the composition, as confirmed by elemen-tal analysis. It should be noted that single-phase hydroxyapatitewas detected for HAP and BAC compositions after refinement, thusconfirming the influence of ionic (Na+, Mg2+, K+, F�, Cl�) substitu-ents on the formation of biphasic mixtures (HAP + b-TCP) for thecompositions BAP-1, BAP-2, BAC-1 and BAC-2.

From the above results, it can be concluded that the inclusion ofcarbonates showed a significant influence on the resultant HAP/b-TCP phase ratios of the biphasic mixtures, especially in thepresence of other ionic (Na+, Mg2+, K+, F�, Cl�) substitution. Noadditional phases other than HAP and b-TCP were detected inany of the investigated compositions. The refined lattice data forHAP phase in both BAC-1 and BAC-2 compositions confirm its typ-ical hexagonal structure (space group P63/m) and the values are inagreement with the crystal data of HAP reported by Kay et al. [35]and also for the synthetic stoichiometric HAP obtained through thesame process in the present study. In case of b-TCP phase in theinvestigated compositions, the refined lattice data confirm itsrhombohedral structure (space group R3c), and the data are inagreement with the values reported by Dickens et al. [36] for sin-gle-phase crystalline b-TCP (a = b = 10.439 Å, c = 37.375 Å, a =b = 90�, c = 120�).

3.2. Influence of carbonates

Figs. 2 and 3 present the effect of temperature on the latticeparameters of hydroxyapatite phase for the different powder com-positions investigated. Over the investigated range of tempera-tures, all the powders show a relatively uniform trend in thevariation in a-axis parameter values that fit well within the errorlimits, whereas for the c-axis parameter values an irregular trendis observed for all the compositions. To be precise, all the carbon-ate-containing apatite compositions show higher a-axis parametervalues in comparison with apatites that contain no carbonates.This is especially true for the BAC sample, which could be termedcarbonated HAP without any other ionic substitutions. The consid-erable increases in the a-axis parameter for carbonated apatitescoincides well with the findings previously reported [37].

However, the comparatively lower a-axis parameter values forBAC-1 and BAC-2 towards BAC account for the presence of anionicsubstitutions (F� and Cl�) in the hydroxyapatite lattice.

FTIR spectra presented in Fig. 4 provide useful informationabout the characteristic nature of apatite. All the investigatedcompositions show characteristic bands due to PO4

3� ions (m1

963 cm�1, m3 1036 and 1095 cm�1, m4 568 and 600 cm�1) for

9.400

9.420

9.440

9.460

9.480

9.500

9.520

800 1000 1200 1400Temperature (ºC)

a -a

xis

(Å)

BAC

BAC-1

BAC-2

BAP-1

BAP-2

HAP

Fig. 2. Effect of heat treatment temperature on the a-axis lattice parameter ofhydroxyapatite phase for different powder samples.

6.870

6.872

6.874

6.876

6.878

6.880

6.882

6.884

6.886

6.888

6.890

800 900 1000 1100 1200 1300 1400Temperature (ºC)

c -a

xis

(Å)

BACBAC-1BAC-2BAP-1BAP-2HAP

Fig. 3. Effect of heat treatment temperature on the c-axis lattice parameter ofhydroxyapatite phase for different powder samples.

400 1000 1600 2200 2800 3400 4000Wave Number (cm-1)

Tra

nsm

itta

nce

(a.u

)

OH

OH

CO3

CO3

F

F

HAP

BAP-1

BAP-2

BAC

BAC-1

BAC-2

Fig. 4. FTIR spectra for the different apatite compositions after heat treatment at1100 �C.

S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843 1839

Author's personal copy

hydroxyapatite phase. The presence of hydroxyl group for the stan-dard HAP could be confirmed from the stretching vibration at3570 cm�1 and libration mode at 635 cm�1. The corresponding de-crease in the signal intensity of the OH group for both BAC-1 andBAC-2 compositions in comparison with the BAC compositioncould be noticed from the FTIR spectra. In fact, a similar decreasein the intensity of OH signals is also observed for the BAP-1 andBAP-2 samples towards HAP sample. The observed weak signalfor OH group is due to the incorporation of substituted fluorineand chlorine at the OH lattice of hydroxyapatite. In addition, thebands observed at �730 cm�1 for BAC-1, BAC-2, BAP-1 and BAP-2could be attributed to the presence of fluorine [38,39] in the com-position, and this is also found to be consistent with the elementalanalysis. However, no bands ascribed to chlorine could be observedin the FTIR spectra, as its presence in the apatite lattice could beconfirmed from its capability to mask the OH signal [40]. The infor-mation about carbonate incorporation in the apatite structurecould also be obtained from the FTIR spectra of BAC, BAC-1 andBAC-2 powders, as the characteristic bands of CO3

2� could beviewed at 1470 cm�1 (m3), 875 cm�1 (m2) and 1555 (m3) cm�1. Theband assignments at 1470 cm�1 and 875 cm�1 could be attributedto B-type carbonated apatite, whereas the bands located at1555 cm�1 could be attributed to A-type carbonated apatite[41,37]. It has already been reported that the calcium phosphateapatite constituent of bone mineral consists of a mixed AB-typesubstitution [42,43]. The most complex substitution type is there-fore also the most interesting one concerning developing a syn-thetic carbonate-substituted HAP as a bone-substitute material.The observed results from the present study confirm the formationof AB-type carbonated apatite, along with the presence of all essen-tial trace elements in the structure.

3.3. Mechanical properties

The influence of composition on sintering behaviour and on themechanical properties of different apatite-based materials over awide range of temperatures was investigated. Fig. 5 representsthe density and WA capacity, while Fig. 6 reports the flexuralstrength and hardness of the samples. An increase in density withincreasing sintering temperature is evident from Fig. 5, with theHAP sample registering the maximum value. A reverse phenome-non is observed in the case of the WA capacity of all the samples,where a decreasing trend with increasing temperature could befound. This decreasing trend could be attributed to the reductionin porosity of samples sintered at higher temperatures. Thedecreasing porosity of samples with increasing sintering tempera-ture is also responsible for the general improvement in flexural

strength and hardness observed for all the samples. It is also ex-pected that grain size developed during the sintering process willplay a role on both density and mechanical properties, but its spe-cific influence is out of the scope of the present study. HAP samplesshow better mechanical properties in comparison with other spec-imens. It should be noted that the mechanical properties of purecarbonate substituted hydroxyapatite (BAC) are comparably infe-rior to the stoichiometric HAP sample. The incorporation of all ionsis accompanied by a further decrease in mechanical properties.Even though, the flexural strength obtained for BAC, BAC-1 andBAC-2 specimens from the present investigation are significantlyhigher than the values reported by Merry et al. [44] for carbonatedhydroxyapatite.

3.4. Biological properties

The cell viability of MC3T3-E1 osteoblastic-like cells cultured inthe presence of the powders for 1.5, 3, 7, 15, 23 and 30 days wasevaluated by the non-destructive and non-toxic resazurin metabo-lism assay (Fig. 7). In the resazurin reduction test, viable cells re-duce this non-toxic blue-coloured probe dye to pink resorufin,being an assay used to indirectly measure cell viability and prolif-eration. During the proliferative period (1–15 days in culture), thelevels of resazurin reduction by cells exposed to the powders werelower than control conditions (0 mg mL�1 powders), but all sam-ples reached a similar plateau level at the 15th day in culture(Fig. 7a). Indeed, an initial delay of cell proliferation could be de-tected, although from days 3 to 15 the proliferation rates of cellsexposed to powders were higher than for control cells. The initiallower metabolic levels of cells exposed to the powders appear tocorrelate with both general and specific characteristics of thepowders. Generally, these delays may correlate with two factors:presence of apatite [45], and a slightly more basic pH for the pow-der-containing media (7.8, against 7.5 for powder-free media) atthe first day in culture. This slightly basic pH was further bufferedto physiological pH (�7.4) with further days in culture (day 3).Nonetheless, it may initially result in lower cellular enzymatic effi-ciency and hence in slower processes (e.g., cell division and metab-olism) and lower resazurin values than control cells. Further, thedifferent relative metabolic levels at day 3, and different prolifera-tive rates further on, appear to correlate with specific characteris-tics of the powders. At this period, carbonate incorporation eitherdelayed or accelerated MC3T3 osteoblasts proliferation. It couldbe identified that BAC-1 and BAC had delayed proliferative profileswhen compared with BAP-1 and HAP, which had no carbonates intheir composition, whereas BAC-2 that possesses carbonates in its

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

800 900 1000 1100 1200 1300 1400Temperature (ºC)

Den

sity

(g/

cm3)

0

5

10

15

20

25

30

WA

Cap

acit

y (%

)

HAP

BAC

BAC-1

BAC-2

HAP

BAC

BAC-1

BAC-2

Fig. 5. Evolution of BD (thick lines and full symbols) and WA capacity (tiny linesand open symbols) of different apatites consolidated by slip casting from anaqueous suspension containing 50 vol.% solids and sintered at differenttemperatures.

0

100

200

300

400

500

600

800 900 1000 1100 1200 1300 1400Temperature (ºC)

Har

dnes

s (H

v)

0

10

20

30

40

50

60

70

80

Fle

xura

l Str

engt

h (M

Pa)

HAP

BAC

BAC-1

BAC-2

HAP

BAC

BAC-1

BAC-2

Fig. 6. Evolution of hardness (thick lines and full symbols) and flexural strength(tiny lines and open symbols) of different apatites consolidated by slip casting froman aqueous suspension containing 50 vol.% solids and sintered at differenttemperatures.

1840 S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843

Author's personal copy

structure increased its proliferative rate relative to BAP-2. Impor-tantly, lower proliferative rates correlate with specific alterationsin cell morphology for a more differentiated-like state: decreasedcytoplasmic volume and cell elongation, already observed at thefirst day in culture (Fig. 7b). For example, BAC-2 has both higherproliferation levels and fewer alterations in cell morphology thanBAP-2; cells exposed to BAP-1 have morphology and proliferationlevels similar to control cells. Significantly, more drastic alterationsin cell morphology were noticed for cells in direct contact with thepowders, which may be due to the roughness/porosity of thepowder surface or to the release of specific ions from the powders.These results are in accordance with other authors [45] whohave been using calcium phosphates in contact with osteoblastsand observed that these materials are more favourable for cell dif-ferentiation, resulting in initial slow growth. These authors alsoobserved that, in the presence of the powders, cell morphologyaltered to a more elongated shape, while being less numerous thannon-exposed control cells [45].

Phosphorylation of the matrix is important in bone calcificationand, when mature, osteoblastic cells are able to secrete ALP. Hence,ALP activity assays were performed in the medium of MC3T3-E1cells exposed to the powders at increasing days in culture. Resultsindicated that MC3T3-E1 cells maintain their capability to expressactive ALP enzymes on powders (Fig. 8a). Further, the absolute ALPactivity profiles of cells exposed to powders follow that of non-ex-posed cells, being lower in the proliferative period (1–15 days), andexhibiting a rapid increase during the differentiation period (15–30 days). In the latter period, the ALP activity levels of control cellswere slightly higher than those of cells exposed to powders(p < 0.05). While carbonate incorporation into BAP-1 had a nega-tive effect on ALP activity (higher levels for BAP-1 and lower forBAC-1), BAP-2 and BAC-2 presented similar absolute ALP activities.Further, carbonate substitution into HAP greatly increased its abil-ity to promote absolute ALP activity (compare BAC with HAP val-ues). Similar time–course graphs for the maturation period were

(a)

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0 5 10 15 20 25 30Days in culture

Res

azur

in A

ssay

(O

.D.F

)

Control cellsBAP-1BAP-2HAPBAC-1BAC-2BAC

(b) C BAP-1 BAP-2 HAP

BAC-1 BAC-2 BAC

Fig. 7. Cell viability assays of MC3T3-E1 osteoblastic-like cells exposed to the powders: (a) cell viability of cells cultured with the powders; results presented are themean ± SEM of at least three independent experiments; (b) phase contrast microphotographs of MC3T3-E1 cells exposed to the powders for 1 day in culture. Bar 100 lm.

0.0

0.2

0.4

0.6

0.8

1.0

0 5 10 15 20 25 30

Days in culture

AL

P A

ctiv

ity

Control cellsBAP-1BAP-2BAC-1BAC-2HAPBAC

0.0

0.1

0.2

0.3

0.4

0.5

0 5 10 15Days in culture

AL

P a

ctiv

ity

/ Res

azu

rin

rat

io

Control cells BAP-1BAP-2 HAPBAC-1 BAC-2BAC

a

b

Fig. 8. (a) ALP activity (OD) of conditioned medium of MC3T3-E1 cells cultured on1 mg mL�1 powders for several days in culture (1.5, 3, 7, 15, 23, 30). Results arepresented as the mean of three independent experiments ±SEM. (b) In order toannul dependence of ALP levels on cell number, the ALP data were normalized forthe samples’ relative resazurin data, taking resazurin control values as 1.0.

S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843 1841

Author's personal copy

obtained when the ALP data were indirectly normalized to cellnumber (data not shown). This similarity was expected, since cellnumber was observed to be similar in the maturation period(Fig. 7a). Nonetheless, in the proliferative period, the ALP valuesare notably much higher for cells exposed to BAP-2, in comparisonwith the other samples (Fig. 8b; data obtained upon division ofcrude ALP levels by the samples’ relative OD resazurin values, tak-ing control resazurin OD levels as 1). Noticeably, this induction of amarker of maturation at an early period correlates well with theobserved BAP-2 low capacity to induce MC3T3 proliferation at thisperiod (Fig. 7a) and its higher induction of alterations on cell mor-phology (Fig. 7b). The related BAC-2, more able to induce cell pro-liferation and to maintain normal cell morphology, lost thiscapacity to induce early ALP activity. Other authors observed thatcarbonate incorporation into apatite had negative effects on osteo-blast differentiation [46] but, as demonstrated by the ALP resultshere described, the effects of carbonated apatites in osteoblastALP activity cannot be generalized. It seems reasonable to admitthat the effect of carbonate might be influenced by the presenceof other ionic substitutions in the lattice. While the presence ofcarbonates does not seem to have any correlation with ALP activ-ity, the same cannot be said concerning the cell morphology. Asa matter of fact, the most significant morphological alterationswere observed in the cells exposed to BAP-2 (Fig. 7b), changes thatappear better correlated with higher initial ALP activity. Hence,alterations in cell morphology and ALP activity may be more re-lated to some specific carbonate incorporation-induced alterationsin the lattice structure and/or its dissolution behaviour in the phys-iological fluid, therefore exerting different effects in the cellsgrown when in contact with the powders.

To confirm that osteoblastic cells exposed to the powders in-deed maintain the ability to differentiate at similar levels to controlcells, the profile of Type-I collagen cellular secretion was also ana-lysed (Fig. 9). Results indicate that, at the differentiation period,following 23 days of exposure to the powders, non-exposed andpowder-exposed cells secrete similar collagen amounts into theextracellular milieu. These results, together with the proliferationand ALP data, confirm that the powders are biocompatible and thatthe effects of carbonate substitutions in the biological activity ofapatites cannot be generalized.

4. Conclusions

In conclusion, it was shown that a simple aqueous precipitationtechnique followed by heat treatment under CO2 atmosphere isappropriate to prepare an apatite model which contains the essen-tial ions found in natural bony apatite. The significance of this ap-proach is that the concentration levels of all the essential elementscould be varied by simply adjusting the initial concentration ofprecursors during synthesis. Most significantly, carbonates couldbe incorporated into the apatite structure, and average contentsof 4.68 ± 0.03 wt.% could be preserved at high temperatures up to1300 �C. The formation of AB-type carbonated apatite from thepresent investigation simulating natural bone could be an addedadvantage. However, the incorporation of carbonate in ionic doped

apatites was shown to be detrimental for the final mechanicalproperties in comparison with stoichiometric HAP. The biologicaltests demonstrated that carbonated powders were biocompatible,and MC3T3-E1 osteoblastic-like cells exposed to the carbonatedpowders showed equal proliferation capacities at the end of theproliferative period, similar ALP activities and equal Type-I colla-gen secretion. Further work will be required to evaluate the effectsof incorporating different levels of carbonate into the apatite struc-ture on its overall performance.

Acknowledgements

The authors thank CICECO, University of Aveiro, for financialsupport, and also FCT for the fellowship grants to S.I.V. (SFRH/BPD/19515/2004), S.M.O. (SFRH/BPD/27013/2006), P.M.C.T. (SFRH/BD/62021/2009) and S.P. (SFRH/BD/BPD/64119/2009).

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 1 and 8 are diffi-cult to interpret in black and white. The full colour images can befound in the on-line version, at doi:10.1016/j.actbio.2010.12.009.

References

[1] Ducheyne P, Qiu Q. Bioactive ceramics: the effect of surface reactivity on boneformation and bone cell function. Biomaterials 1999;20:2287–303.

[2] Engin NO, Tas AC. Preparation of porous Ca10(PO4)6(OH)2 and Ca3(PO4)2

bioceramics. J Am Ceram Soc 2000;83:1581–4.[3] Legeros RZ. Calcium phosphates in oral biology and medicine. Basel: Karger;

1991.[4] Daculsi G, Bouler J-M, Legeros RZ. Adaptive crystal formation in normal and

pathological calcifications in synthetic calcium phosphate and relatedbiomaterials. Int Rev Cytol 1997;172:129–91.

[5] Dorozhkin SV. Calcium orthophosphates. J Mater Sci 2007;42:1061–95.[6] Ginty F, Flynm A, Cashman KD. The effect of dietary sodium intake on

biochemical markers of bone metabolism in young women. Br J Nutr1998;79:343–50.

[7] Ergun C, Webster TJ, Bizios R, Doremus RH. Hydroxylapatite with substitutedmagnesium, zinc, cadmium, and yttrium. I. Structure and microstructure. JBiomed Mater Res 2002;59:305–11.

[8] Wiesmann HP, Plate U, Zierold K, Hohling HJ. Potassium is involved in apatitebiomineralization. J Dent Res 1998;77:1654–7.

[9] Kim HW, Noh YJ, Koh YH, Kim HE. Pressureless sintering and mechanical andbiological properties of fluor-hydroxyapatite composites with zirconia. J AmCeram Soc 2003;86:2019–26.

[10] Farley JR, Wergedal JE, Baylink DJ. Fluoride directly stimulates proliferationand alkaline phosphatase activity of bone-forming cells. Science 1983;222:330–2.

[11] Neuman WF, Neuman MW. The chemical dynamics of bone mineral. Chicago:University of Chicago press; 1958.

[12] Kannan S, Goetz-Neunhoeffer F, Neubauer J, Ferreira JMF. Ionic substitutions inbiphasic hydroxyapatite and b-tricalcium phosphate mixtures: structuralanalysis by Rietveld refinement. J Am Ceram Soc 2008;91:1–12.

[13] Morgan H, Wilson RM, Elliott JC, Dowker SEP, Anderson P. Preparation andcharacterization of monoclinic hydroxyapatite and its precipitated carbonateapatite intermediate. Biomaterials 2000;21:617–27.

[14] Suchanek WL, Shuk P, Byrappa K, Riman RE, Ten Huisen KS, Janas VF.Mechanochemical hydrothermal synthesis of carbonated apatite powders atroom temperature. Biomaterials 2002;23:699–710.

[15] De Maeyer EAP, Verbeeck RMH, Naessens DE. Effect of heating on theconstitution of Na+ and CO3

2� containing apatites obtained by hydrolysis ofmonetite. Inorg Chem 1994;33:5999–6006.

[16] De Maeyer EAP, Verbeeck RMH, Naessens DE. Stoichiometry of sodium andcarbonate-containing apatites obtained by hydrolysis of monetite. Inorg Chem1993;32:5709–14.

[17] Gibson IR, Bonfield W. Preparation and characterization of magnesium/carbonate co-substituted hydroxyapatites. J Mater Sci Mater Med 2002;13:685–93.

[18] Fleet ME, Liu X. Coupled substitution of type A and B carbonate in sodium-bearing apatite. Biomaterials 2007;28:916–26.

[19] Feki HEL, Savariault JM, Salah AB, Jemal M. Sodium and carbonate distributionin substituted calcium hydroxyapatite. Solid State Sci 2000;2:577–86.

[20] Naji A, Harmand M. Cytocompatibility of two coating materials, amorphousalumina and silicon carbide, using human differentiated cell cultures.Biomaterials 1991;12:690–4.

150 a(I)

C BAP-1 BAP-2 HAP BAC-1 BAC-2 BAC

ß(I)

Fig. 9. Immunoblot analysis of medium secreted Type-I collagen of MC3T3-E1 cellsexposed to the powders for 23 days in culture. a(I) and b(I), procollagen Type Imonomeric and dimeric forms, respectively. Migration of molecular weight markersis indicated to the right.

1842 S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843

Author's personal copy

[21] Malik M, Puleo D, Bizios R, Doremus R. Osteoblasts on hydroxyapatite, aluminaand bone surfaces in vitro: morphology during the first 2 hours of attachment.Biomaterials 1992;13:123–8.

[22] Puleo D, Preston K, Shaffer J, Bizios R. Examination of osteoblast–orthopaedicbiomaterial interactions using molecular techniques. Biomaterials 1993;14:111–20.

[23] Valerio P, Pereira MM, Goes AM, Leite MF. The effect of ionic products frombioactive glass dissolution on osteoblast proliferation and collagen production.Biomaterials 2004;25:2941–8.

[24] Kannan S, Ventura JM, Ferreira JMF. In situ formation and characterization offlourine-substituted biphasic calcium phosphate ceramics of varied F-HAP/b-TCP ratios. Chem Mater 2005;17:3065–8.

[25] Kannan S, Lemos AF, Ferreira JMF. Synthesis and mechanical performance ofbiological-like hydroxyapatites. Chem Mater 2006;18:2181–6.

[26] Larson AC, von Dreele RB. GSAS: general structure analysis system LANSCE,MS-H805; Los Alamos National Laboratory. NM: Los Alamos; 1998.

[27] Toby BH. EXPGUI, A graphical user interface for GSAS. J Appl Cryst2001;34:210–3.

[28] Testing Method for Flexural Strength (Modulus of Rupture) of HighPerformance Ceramics, Japanese Industrial Standard JIS R 1601, JapaneseStandards Association, Tokyo, Japan, 1981.

[29] Ueno A, Miwa Y, Miyoshi K, Horiguchi T, Inoue H, Ruspita I, et al. Noma TConstitutive expression of thrombospondin in MC3T3-E1 osteoblastic cellsinhibits mineralization. J Cell Phys 2006;209:322–32.

[30] Pina S, Vieira SI, Rego P, Torres PMC, e Silva OAB, da Silva EF, et al. Biologicalresponses of brushite-forming Zn- and ZnSr-substituted beta-tricalciumphosphate bone cements. Eur Cell Mater 2010;20:162–77.

[31] Kannan S, Goetz-Neunhoeffer F, Neubauer J, Rebelo AHS, Valerio P, FerreiraJMF. Rietveld structure and in vitro analysis on the influence of magnesium inbiphasic (hydroxyapatite and beta-tricalcium phosphate) mixtures. J BiomedMater Res B-Appl Biomater 2009;90:404–11.

[32] Horwitz JP, Chua J, Noel M, Donatti JT, Freisler J. Substrates for cytochemicaldemonstration of enzyme activity. 11. Some dihalo-3-indolyl phosphates andsulfates. J Med Chem 1966;9:447–8.

[33] Wolf P, Horwitz P, Vazquez J, Von Der Muehll E. A new histochemical stain forneutrophilic leukocyte alkaline phosphatase. Enzymologia 1968;35:154–6.

[34] Wolf P, Von der Muehll E, Praisler K. A test for bacterial alkaline phosphatase:use in rapid identification of serratia organisms. Clin Chem 1973;19:1248–9.

[35] Kay MI, Young RA, Posner AS. Crystal structure of hydroxyapatite. Nature1964;204:1050–2.

[36] Dickens B, Schroeder LW, Brown WE. Crystallographic studies of the role of Mgas a stabilizing impurity in b-Ca3(PO4)2. The crystal structure of pure b-Ca3(PO4)2. J Solid State Chem 1974;10:232–48.

[37] Gibson IR, Bonfield W. Novel synthesis and characterization of an AB-typecarbonate-substituted hydroxyapatite. J Biomed Mater Res 2002;59:697–708.

[38] Rodriguez-Lorenzo LM, Hart JN, Gross KA. Influence of fluorine in the synthesisof apatites. Synthesis of solid solutions of hydroxy fluorapatite. Biomaterials2003;24:3777–85.

[39] Kannan S, Ferreira JMF. Synthesis and thermal stability HAP and fl-TCPcomposites with co-substituted sodium, magnesium and fluorine. Chem Mater2006;18:198–203.

[40] Kannan S, Rebelo A, Ferreira JMF. Novel synthesis and structuralcharacterization of fluorine and chlorine co-substituted hydroxyapatites. JInorg Biochem 2006;100:1692–7.

[41] LeGeros RZ, Trautz OR, Klein E, LeGeros JP. Two types of carbonate substitutionin the apatite structure. Experimentia 1969;25:5–7.

[42] Emerson WH, Fischer ED. The infrared absorption spectra of carbonate incalcified tissue. Arch Oral Biol 1962;7:671–83.

[43] Rey C, Collins B, Goehl T, Dickson IR, Glimcher MJ. The carbonate environmentin bone mineral: a resolution-enhanced Fourier transform infraredspectroscopic study. Calcif Tissue Int 1989;45:157–64.

[44] Merry JC, Gibson IR, Best SM, Bonfield W. Synthesis and characterisation ofcarbonate hydroxyapatite. J Mater Sci Mater Med 1998;9:779–83.

[45] Richard D, Dumelie N, Benhayoune H, Bouthors S, Guillaume C, Lalun N, et al.Behavior of human osteoblast-like cells in contact with electrodepositedcalcium phosphate coatings. J Biomed Mater Res B Appl Biomater2006;79:108–15.

[46] Yang Liang, Perez-Amodio Soledad, Barre‘re-de Groot Florence YF, EvertsVincent, van Blitterswijk Clemens A, Habibovic Pamela. The effects ofinorganic additives to calcium phosphate on in vitro behaviour of osteoblastsand osteoclasts. Biomaterials 2010;31:2976–89.

S. Kannan et al. / Acta Biomaterialia 7 (2011) 1835–1843 1843