Carbonate Hydroxyapatite and Silicon-Substituted Carbonate Hydroxyapatite: Synthesis, Mechanical Properties, and Solubility Evaluations

The present study investigates the chemical composition, solubility, and physical and mechanical properties of carbonate hydroxyapatite (CO3Ap) and silicon-substituted carbonate hydroxyapatite (Si-CO3Ap) which have been prepared by a simple precipitation method. X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), X-ray fluorescence (XRF) spectroscopy, and inductively coupled plasma (ICP) techniques were used to characterize the formation of CO3Ap and Si-CO3Ap. The results revealed that the silicate (SiO4 4−) and carbonate (CO3 2−) ions competed to occupy the phosphate (PO4 3−) site and also entered simultaneously into the hydroxyapatite structure. The Si-substituted CO3Ap reduced the powder crystallinity and promoted ion release which resulted in a better solubility compared to that of Si-free CO3Ap. The mean particle size of Si-CO3Ap was much finer than that of CO3Ap. At 750°C heat-treatment temperature, the diametral tensile strengths (DTS) of Si-CO3Ap and CO3Ap were about 10.8 ± 0.3 and 11.8 ± 0.4 MPa, respectively.


Introduction
The use of hydroxyapatite (HA) as bone substitute is well known for its bioactivity and osteoconductivity in vivo [1,2]. However, the natural bone which differs from pure HA contains about 4-8 wt% carbonate along with several multisubstituted ions (Na + , Mg 2+ , K + , F − , Cl − , etc.) in its structure [3][4][5]. Carbonate substituted into the HA structure (CO 3 Ap) is of special interest because the CO 3 2− ion has an impact on different pathologies of human tissues, such as dental caries [6]. CO 3 Ap was also reported to be more soluble in vivo than HA and to increase the local concentration of calcium and phosphate ions that are necessary for new bone formation [7]. Moreover, CO 3 Ap is resorbed faster by osteoclasts and replaced with the new bone at a higher rate compared to HA [8]. CO 3 2− ion can replace OH − or PO 4 3− ions giving A-or B-type CO 3 Ap, respectively. If these substitutions take place simultaneously, an AB-type substitution occurs, as in the case of the bone mineral [7,9].
It was reported that Si enhances and stimulates osteoblast-like cell activity [10] in vitro and induces a higher dissolution rate in vivo [11]. The solubility was observed to increase with a decrease in structural order due to the presence of the foreign ions (i.e., CO 3 2− , SiO 4 4− ) in the HA structure [12]; nonetheless, only few papers have investigated ion release in synthetic fluids [11,13]. Therefore, the development of synthetic HA powders with a fully completed ionic substitution in the HA lattice is of great importance in order to mimic that of the natural bone.
Numerous research works have focused on the synthesis of HA biomaterial substituted with single-or multi-ion substitution of CO 3 2− [14], Si 4+ [3,15], and so forth, whereas the substitution of CO 3 2− along with other cations in the apatite structure was restricted to the cosubstitution of HA with the ionic pair of Mg 2+ /CO 3  The Scientific World Journal present in the material substituted completely the PO 4 3− in the HA structure or whether the replacement was partial. It was reported [12] that both CO 3 2− and SiO 4 4− reduced HA crystallinity, and the structure could host only a limited amount of the two ions before collapsing. Additionally, the final product contained CO 3 2− and SiO 4 4− , but there was a lack of experimental evidence on the competitive substitution of CO 3 2− and SiO 4 4− ions for PO 4 3− ions [12]. Recently, an extensive study on the SiO 4 4− and CO 3 2− cosubstituted HA was reported [18]. However, the preparation methods were carried out under air atmosphere and used CO 2 from the atmosphere as the CO 3 2− source, and as such, there was no control of CO 3 2− substitution level. Thus, the CO 3 2− ion present could indeed be doped-HA, where the foreign ion is just adsorbed on the surface of the crystals [12]. Moreover, there were few research works that studied the mechanical properties of the ion-substituted HA after heat-treatment.
Therefore, the purpose of the present work is to investigate the simultaneous substitution of SiO 4 4− and CO 3 2− into the HA structure in order to obtain a product which is closer to the natural bone. The competition between CO 3 2− and SiO 4 4− for substituting the PO 4 3− ions in the HA structure was also investigated. The aim of the work was also to evaluate the mechanical properties and the solubility of the siliconsubstituted carbonate HA as compared to that of carbonate HA.

Experimental Procedure
A precipitation method was adopted to prepare CO 3 Ap using Ca(OH) 2 (96% purity, FLUKA, 21181) and H 3 PO 4 (15M, MERCK, 100573, Germany) with CO 2 gas as the carbonate source [14]. The Ca/P molar ratio of the precursors was designed to be similar to Ca/P molar ratio of biological bone, which is 1.67 [2]. Initially, a solution of 300 mL of H 3 PO 4 1 M was gradually added to 500 mL of Ca(OH) 2 1 M under vigorous stirring at 400 rpm, whilst CO 2 gas was passed through the reaction flask during the reaction. According to Landi et al. [14], to obtain the highest carbonation degree and favor B-type CO 3 Ap precipitation with respect to A-type, the CO 2 flow was set at 0.5 bubble/s as the outlet flux. Similar to CO 3 Ap, the Si-CO 3 Ap was prepared using silicon tetraacetate [Si(COOCH 3 ) 4 ] (98% purity, SIGMA-ALDRICH) as the Si precursor. Based on the chemical formula proposed by Gibson et al. [20] for silicon-substituted HA (Si-HA), the amount of reagents was calculated by assuming that one SiO 4 4− ion would substitute for one PO 4 3− ion based on a stoichiometric HA; Ca/(P+Si) molar ratio = 1.67. Si(COOCH 3 ) 4 was dissolved in the Ca(OH) 2 solution under continuous stirring for 2 hours before adding the H 3 PO 4 solution. In this research work, the Si content was chosen to be 1.6 wt% which had been shown to be the optimum amount for the enhancement of the mechanical properties of Si-HA reported in our previous study [21], where the Ca/(P+Si) ratio = 1.84.
The reactions took place in a reaction flask which was placed in a heating mantle to control the reaction temperature at 40 ∘ C ± 1. The pH of the solution was monitored using a pH meter. NH 4 OH 29% (J.T.Baker, USA) was added to maintain the pH of the solution at 9.4 ± 0.1. After the reaction was completed, the slurry was continuously stirred for 2 h without CO 2 gas. It was then allowed to mature at room temperature for 24 h. Subsequently, it was filtered and washed with deionized water to remove any residue before being dried in an oven at 70 ∘ C for 24 h. The dried CO 3 Ap and Si-CO 3 Ap powders were then ground with an agate pestle and mortar. For the DTS test, the CO 3 Ap and Si-CO 3 Ap powders were compacted by uniaxial hydraulic pressing equipment using a die with 8 mm diameter at a pressure of 10 MPa. The thickness of samples was about 2.91-3.25 cm. Alcohol 70% was used to clean the mold. The compacted samples were then heat-treated at different temperatures of 650, 700, and 750 ∘ C with a heating rate of 3 ∘ C/min and soaked for 2 h in CO 2 atmosphere (80 mL/min) which was passed through 150 mL distilled water. The syntheses of CO 3 Ap and Si-CO 3 Ap were repeated three times to confirm the reproducibility of the materials.
The as-synthesized and heat-treated powders were characterized using an X-ray diffractometer (XRD; D5000 Siemens) for phase identifications. Peak (002) was chosen for determining the crystallite size since it is one of the strongest peaks without any overlapping in the CO 3 Ap and Si-CO 3 Ap patterns. The lattice parameters ( and ) of the as-prepared CO 3 Ap and Si-CO 3 Ap samples were determined through the (hkl) peaks position of the apatite from XRD patterns according to (1) as follows [22,23]: Fourier transform infrared spectroscopy (FTIR; Perkin-Elmer FT-IR 2000, FTIR spectrometer) was used to study the silicon and carbonate substitutions of the different functional groups, such as OH − , PO 4 3− , CO 3 2− , and SiO 4 4− in the CO 3 Ap and Si-CO 3 Ap samples. The carbonate content of powders was analyzed using an elemental analyzer (CHN test; Perkin Elmer series 2, 2400 CHNS/O). The chemical composition (Si and Ca) was determined by inductive coupled plasma (ICP) spectrometer (ICP/AES, ARL-3410). X-ray fluorescence spectrometer (XRF; Rigaku RIX-300 wavelength dispersive) was used to study the Ca/P ratio of the asprepared powders. The particle size of the powder (with ultrasonic dispersion) was measured using a Malvern Mastersizer X (Malvern Instruments, Malvern, UK). The powder before being characterized had been passed through a 75 m sieve.
The densities of the heat-treated CO 3 Ap and Si-CO 3 Ap compacts were measured using Archimedes' principle. The diametral tensile strengths (DTS) of the heat-treated CO 3 Ap and Si-CO 3 Ap compacts were tested at a strain rate of 0.5 mm/min. The DTS test involves compressing a sample diametrically, inducing a stress that causes the sample to yield in tension. In this test, a disk sample was placed between two platens and then vertically compressed until it broke [24]. During loading, the applied force was recorded and the tensile stress was calculated using (2) where max is maximum load at failure (N) and ℎ and are the thickness and diameter of the compacts (mm), respectively.
The Scientific World Journal 3 The solubility evaluation was performed in triplicate on the as-synthesized CO 3 Ap and Si-CO 3 Ap compacts (8 mm diameter die, 10 MPa) by immersing the compacts in a simulated body luid (SBF) solution at 36.5 ∘ C. The SBF solution was prepared according to the procedure described by Kokubo and Takadama [25]. The tests were carried out within 1 and 7 days. After the predetermined soaking time, the samples were removed and the liquid mediums were analyzed by ICP.
The released ion was estimated by subtracting the initial ion concentration of the SBF solution from the ion concentration of the SBF solution after immersion. Statistical analysis was performed to evaluate the statistical differences between the sample sets by employing one factor analysis of variance (ANOVA) when comparing more than two sample populations. Significant differences were considered at the 95% level ( < 0.05). Table 1 shows the physical and chemical properties of the assynthesized CO 3 Ap and Si-CO 3 Ap samples. The mean particle size of the as-synthesized Si-CO 3 Ap sample is significantly smaller than that of the as-synthesized CO 3 Ap sample. This can be attributed to the substitution of Si in the HA structure, as reported in previous research works [21,26].

Physical and Chemical Composition Analyses.
In the same table, the Ca/P molar ratios of the assynthesized CO 3 Ap and Si-CO 3 Ap samples show much higher values than those of the predetermined ratios. This indicated that the substitution of CO 3 2− and SiO 4 4− ions for the PO 4 3− groups in the HA had taken place. These substitutions reduce the amount of PO 4 3− group, thus leading to an increase in the Ca/P ratio [14,20]. However, the Ca/P ratio in this study was in the range of the Ca/P molar ratio of CO 3 Ap reported previously, which was of 1.7-2.6 [27].
The Si contents are also included in Table 1. Si measured in the as-synthesized Si-CO 3 Ap sample is about 0.85 wt%, and this is much lower than the starting value (1.6 wt%). The rest of the Si unaccounted for will be explained in the FTIR analysis. It was suggested that an amount of only 1 wt% Si substituted into HA was sufficient to elicit important bioactive improvements [12], and, hence, the Si-substituted CO 3 Ap in this research work could be considered to approach this enhancement.
After heat-treatment at a temperature range of 650-750 ∘ C, the carbonate amount slightly decreases compared to the as-prepared samples ( Table 2). This is due to the fact that carbonate absorbed had desorbed upon heat-treatment. The amount of carbonate is close to the typical amount of carbonate in human bone [28].   Figure 1 shows the XRD patterns of the assynthesized CO 3 Ap and Si-CO 3 Ap powders. The broad peaks indicate the formation of HA phase with low crystallinity, and no secondary crystalline phases were observed. The poor crystallinity was due to the low synthesis temperature and the substitution of SiO 4 4− and CO 3 2− ions limited the crystallization of the HA phase [18,21].

XRD Analysis.
The crystallite size determined using Scherrer's equation and the lattice parameters are given in Table 3. The CO 3 2− and SiO 4 4− substitutions in HA structure led to changes in the crystal lattice parameters [4,18]. Previous studies had shown that the -axis decreased and the -axis increased with increasing CO 3 2− or SiO 4 4− in the HA structure [3,6]. The values presented in Table 3 for the as-prepared powders in this present research work also show a similar trend with previous works. The SiO 4 4− groups are larger and have a more negative charge than either PO 4 3− or CO 3 2− ions [15,18]. Additionally, the substitution of SiO 4 4− and CO 3 2− for PO 4 3− contributes to reducing the crystallite size, as has been observed previously in other studies [12,18,21].
Numerous studies showed that both -and -axis dimensions increased with the silicon content [18,29,30]. Considering the substitution of SiO 4 4− in the CO 3 Ap, it is possible that, -and -axis dimensions are higher than those of CO 3 Ap (   After heat-treatment at 650 ∘ C to 700 ∘ C, pure CO 3 Ap and Si-CO 3 Ap are still observed and no secondary phases are detected (Figure 2(a), (b), (d), and (e)). However, a new phase, CaCO 3 , is clearly observed in Si-CO 3 Ap samples heattreated at 750 ∘ C due to the decomposition of the Si-CO 3 Ap samples.
Sintering of CO 3 Ap at high temperatures (≥900 ∘ C) [15,31] produces hydroxyapatite (HA) and CaO. In the CO 2 -rich atmosphere, the CaCO 3 obtained was due to the reaction of CaO and CO 2 . Therefore, a mixture of CO 3 Ap and CaCO 3 is observed after heat-treatment in CO 2 atmosphere. The decomposition temperature decreased with an increase of the carbonate [31] and/or silicon content [15,32]. Since the heat-treatment process was carried out at low temperatures, such decomposition did not occur in the CO 3 Ap sample but did occur in Si-CO 3 Ap sample at 750 ∘ C. The simultaneous substitution of SiO 4 4− and CO 3 2− ions for the PO 4 3− ions of the HA structure increased the defects in HA structure and produced more OH − vacancies [13] compared to CO 3 Ap. The formation of OH vacancies has been proven to accelerate the decomposition process [23]. Thus, the formation of CaCO 3 in the Si-CO 3 Ap could be explained by a similar mechanism as the decomposition of CO 3 Ap.  [4,18,33], whereas the bands located at 1505 cm −1 could be attributed to A-type CO 3 Ap [28]. The characteristic OH − bands of HA at 630 cm −1 are not clearly visible in all FITR spectra. In fact, a similar decrease in the intensity of OH − signals was also observed due to the substitution of CO 3 2− at the OH − lattice of HA [33]. In this case, the substitution of CO 3 2− and SiO 4 4− ions for PO 4 3− would create an OH − loss needed to compensate the charge balance, thus resulting in the weak of OH − signal. Additional bands are also observed in the Si-CO 3 Ap sample at about 800 cm −1 and 480 cm −1 which do not appear in CO 3 Ap sample. The band at 480 cm −1 is assigned to the SiO 4 4− in the apatite structure [15]. However, the band at about 800 cm −1 might be assigned to either the silicate group [30] or to the O-Si-O bending in the SiO 2 amorphous phase [22,34]. As detected by ICP, the amount of Si in Si-CO 3 Ap sample is much lower than the starting value (Table 1); the silicate species which could not totally be incorporated in the apatite structure exist on the surface of the materials as an amorphous phase [22,35] and/or remain in mother liquors after precipitation [36]. The remaining Si suggests that the competition arising between the SiO 4 4− and CO 3

2−
ions occupies the PO 4 3− sites. The polymerization of the silicate species at the surface was reported elsewhere [37]. In another research work [38], the amorphous SiO 2 phase in -TCP containing Si-substitution showed a significantly higher MC3T3-E1 osteoblast-like cell number compared to pure -TCP. Therefore, the presence of SiO 2 would not cause toxicity to the cells and would not affect cell differentiation. change the symmetry and stability of an apatite structure [39]. As a result of these substitutions, shifts and splitting of the PO 4 vibration bands at about 500-700 cm −1 occur in the apatite IR spectra (Figure 3). It has already been reported that the calcium phosphate apatite constituent of bone mineral consists of a mixed AB-type substitution [40]. The results from the present study confirm the formation of AB-type carbonated apatite along with the presence of Si in the structure. Thus, this complex substitution type is also of utmost importance when the development of a synthetic bone-substitute material is sought.

Evaluation of Mechanical Properties and Microstructure.
The mechanical and physical properties were evaluated in terms of diametral tensile strength (DTS) and bulk density. In Figure 4, the density of CO 3 Ap sample is higher than that of Si-CO 3 Ap sample at any heat-treatment temperatures. This can be explained by the higher lattice parameters of both CO 3 2− and SiO 4 4− cosubstitution compared to the single CO 3 2− substitution (Table 3). It can also be seen that the density of the CO 3 Ap samples significantly increases with increasing heat-treatment temperatures, whilst there is only a slight change in the density of the Si-CO 3 Ap samples. The substitution of Si reduced the density of the materials compared to HA as reported previously [15,21] due to the change of unit cell parameters in the silicon-substituted materials. Therefore, the effect of silicon became significant which slowed down the densification process upon heat-treatment. In the present research work, the densities of CO 3 Ap and Si-CO 3 Ap are significantly lower compared to that of a fully dense HA (3.16 g/cm 3 ) due to the low heat-treatment temperatures. Si-HA [21] Pure HA [21] Si-HA [19] Heat-treatment temperature ( ∘ C) Figure 6: DTS versus heat-treatment temperatures for various carbonate hydroxyapatites. Figure 5 shows that the DTS of both CO 3 Ap and Si-CO 3 Ap samples significantly increase with increasing temperatures. The increase in DTS value of CO 3 Ap with the increasing heat-treatment temperatures can be explained by the increase in density as shown in Figure 4. However, although a slightly higher density was obtained for the Si-CO 3 Ap, the DTS of Si-CO 3 Ap increases significantly with increasing heat-treatment temperatures. This is due to the cosubstitution of CO 3 2− and SiO 4 4− . This cosubstitution induced the smaller particle size (Table 1). In addition, Si substitution was reported to impede grain growth during heat-treatment [41] and so increased the DTS value. By comparison, the DTS of CO 3 Ap samples appear to be slightly higher compared to those of Si-CO 3 Ap samples. This difference in strength was evaluated to be > 0.05, and as such, this difference in DTS value of Si-CO 3 Ap is insignificant compared to CO 3 Ap. However, its density is significantly lower ( < 0.05) indicating the positive effect of SiO 4 4− and CO 3 2− cosubstitutions on this matter. As reported, the effect of silicon on the increase of mechanical strength was evidenced at higher heat-treatment temperatures, that is, 1200 ∘ C and above, as compared to Si-free samples [21]. Conversely, at lower temperatures, this effect was not so apparent where the strengths of Si-samples and Si-free samples were comparable based on previous studies [21,41] and even lower [19] due to the lower density. Hence, due to the low heat-treatment temperatures employed in this research work, the difference in strength between CO 3 Ap and Si-CO 3 Ap samples is not that significant. Figure 6 presents a comparison of the DTS values of the present materials at 750 ∘ C with those of samples in previous research works [19,21]. Interestingly, at the same Si content (about 0.8 wt%), the DTS values of CO 3 Ap and Si-CO 3 Ap at 750 ∘ C are about 10.8 ± 0.3-11.8 ± 0.4 MPa, and these are higher than those of Si-substituted HA samples at 1250 ∘ C [21] and much higher than that of Si-HA sample at 1300 ∘ C [19]. This demonstrates that, at these low heat-treatment temperatures, the cosubstitution of carbonate and Si in the HA structure would increase the strength of the final product.
In Figure 6, the DTS of Si-substituted HA sample [21] is higher than that of pure HA because the SiO 4 4− substitution impeded grain growth at high temperatures and, therefore, The Scientific World Journal 7 increased the strength of the materials [41]. The DTS of CO 3 Ap and Si-CO 3 Ap in the present work are also higher than that of pure HA. It was explained [42] that the CO 3 2− and SiO 4 4− substitutions also reduced the grain size of the final product and resulted in an increase of the strength of the samples.

Solubility Evaluation.
In the case of crystalline HA, the degree of micro-and macroporosities, defect structure, and amount and type of other phases present have a significant influence on the dissolution rate [43]. In this study, the immersion of the CO 3 Ap and Si-CO 3 Ap compacts (surface area = 150.8 mm 2 ) into SBF solution produced noticeable changes in the ion concentrations of the solution. Figures  7(a), 7(b), and 7(c) show the ion concentration of Ca, Si and changes in pH value of the medium after a certain period of immersion time, respectively. According to Boanini et al. [12], crystallinity and crystal dimensions significantly affected the solubility and, as a consequence, ion release. Thus, a decrease in structural order due to the presence of foreign ions might be responsible for the observed increase in solubility.
In Figure 7, the Ca 2+ and Si 4+ ion concentrations as well as the pH of the SBF solution increase with soaking duration which indicates the dissolution of Ca 2+ and Si 4+ ions. It had been reported that the initial dissolution of implant materials plays an important role in enhancing their bonding to the bone [32]. With an increase in the soaking duration, Ca 2+ concentrations and pH value continuously increase due to the ionic exchange between H + within the SBF solution and Ca 2+ in the CO 3 Ap and Si-CO 3 Ap compacts [44,45]. The increase of solution pH generally facilitates the nucleation of apatite [46].
The release of Si 4+ ions was also observed continuously over the whole investigation period. It was reasoned out that the amorphous layer surrounding the apatite grains dissolved within the first period of immersion in SBF leaving a more stable and less soluble core [13]. As solubility is highly sensitive to the structural and chemical compositions of the apatite samples, the crystallite size is a key factor for in vitro behavior of synthetic apatite [47]. In this manner, the resorbability of CO 3 Ap and Si-CO 3 Ap could be promoted by a smaller crystallite size when CO 3 2− and SiO 4 4− were cosubstituted; the amorphous shell can be thicker and yield a more intense and prolonged ion release [13]. In addition, the Ca 2+ release in Si-CO 3 Ap compacts is slightly higher compared to CO 3 Ap, which suggests a better solubility (Figure 7(a)) that leads to a faster super-saturation with respect to HA, a faster nucleation, and growth of apatite on the surface of the compacts [36].
By comparison, the Ca 2+ release for CO 3 Ap and Si-CO 3 Ap samples in this study is much higher than that of Mgsubstituted fluorapatite [48] and HA [44,49] under the same conditions. It was reported that the solubility of materials increases with increasing ionic substitutions into the HA lattice and decreasing crystallinity which is represented by the higher ion release in the SBF solution [13,16,49]. Therefore, the CO 3 Ap and Si-CO 3 Ap obtained in this work are of higher solubility compared to the above-mentioned materials.
Based on the solubility evaluations using SBF, the solubility of CO 3 Ap and Si-CO 3 Ap is such that it is predicted that ions would continuously exist in actual physiological conditions. This is further reinforced by a previous work [13]. These materials could supply elements which are essential for osteoblast activity and new bone tissue formation [13]. The simultaneous presence of such elements can further enhance the cell response.

Conclusions
Carbonate hydroxyapatite and silicon-substituted carbonate hydroxyapatite powders were successfully synthesized by a simple and high-yield process. The crystallite and mean particle size of Si-CO 3 Ap sample was significantly smaller than that of CO 3 Ap sample due to the cosubstitution of SiO 4 4− and CO 3 2− in the HA structure. No secondary phases were detected in CO 3 Ap and Si-CO 3 Ap samples after heattreatment in the temperature range of 650 ∘ C to 700 ∘ C. CaCO 3 was observed in Si-CO 3 Ap sample after heat-treatment at 750 ∘ C, whilst the purity of CO 3 Ap was retained. The SiO 4 4− and CO 3 2− cosubstituted HA structure led to a significant decrease in density compared to a single CO 3 2− substituted HA structure, whilst the DTS of both samples showed insignificant differences.
The competition between SiO 4 4− and CO 3 2− ions had taken place to occupy the PO 4 3− site. Si-CO 3 Ap existed in the form of AB-type carbonated apatite, and the presence of SiO 4 4− in the structure is of utmost interest in developing a synthetic bone-substitute material. The total amount of carbonate and silicon and the crystal size of the powder obtained mimic those of biological apatites. The silicon substitution improved the solubility of Si-CO 3 Ap which prolongs the ion release compared to that of Si-free CO 3 Ap. The present materials possess low crystallinity and the CO 3 2− content is close to that found in natural bone, and, in combination with the high strength, these materials could be ideal for bone substitutes.