Defect structures of sodium and chloride co-substituted hydroxyapatite and its osseointegration capacity

A defect structure and osseointegration capacity of sodium and chloride co-substituted hydroxyapatite (NaClAp) were newly studied. The NaClAp was prepared by reacting H3PO4 and Ca(OH)2 with NaNO3 and NH4Cl followed by sintering; pure hydroxyapatite (HAp) was synthesized as a control. After sintering, the co-substitution of Ca and OH with Na and Cl, respectively, produced charged point defects at Ca and PO4 sites. Also, OH molecules partially adopted a head-on structure. The calculated total system energy of NaClAp was higher, whereas the binding energies between each constituent elements and system were lower than those of HAp. These results suggest that NaClAp was less stable than HAp, due to the formation of various defects by co-substitution of Na and Cl. Indeed, NaClAp exhibited higher dissolution behavior in simulated body fluid (SBF) compared with HAp. Accordingly, this increased the capability to produce low crystalline hydroxyl carbonate apatite, likely due to the increasing degree of apatite supersaturation in SBF. Besides, the NaClAp granules showed noticeable improvements in osseointegration capacity four weeks after in vivo test compared with HAp. Collectively, these results imply that the defects made by multiple ion substitutions are useful to increase osseointegration capacity of hydroxyapatite.

Many attempts have been made to explain the defect structures of the hydroxyapatite substituted with cationic ions having different electronic valences. Several studies tried to show the entire defect structures and the process of defect formation [8,10,19]. However, the suggested defect structures were incomplete because they did not exactly follow three rules for analyzing defect chemical reactions such as site, charge, and mass balances [8,10,19]. The Kröger-Vink notation [20] coupled with three rules for analyzing defect chemical reactions is necessary to solve this problem.
The substitution of single foreign element, which had different electronic valences, electronegativity, and sizes, with one of the constituent elements of hydroxyapatite increased the total system energy and decreased the binding energies between constituent elements of apatite and the system [19,21]. The different valences, electronegativity, and sizes of foreign elements are expected to make the hydroxyapatite unstable energetically after sintering by causing lattice distortions and point defects. This will cause the dissolution of constituent elements in vivo and consequently enhance the osseointegration capacity by the formation of low crystalline hydroxyl carbonate apatite (HCA).
Herein, we attempted to maximize the total system energy by simultaneously replacing two constituent elements of hydroxyapatite, Ca and OH, with Na and Cl, respectively. There have been no studies about the co-substitution of different elements with different charges. We describe in detail the synthesis of Na and Cl co-substituted hydroxyapatite (NaClAp) and the formation mechanism of defect structures during the process using Krö ger-Vink notation [20]. The NaClAp was then evaluated, relative to pure HAp, in regard to HCA producing capability in simulated body fluid (SBF) and osseointegration capacity. The ab initio method was used to analyze the improved osseointegration capacity of NaClAp compared to HAp.

Materials and methods
Syntheses of Na and Cl co-substituted hydroxyapatite and pure hydroxyapatite The Ca(OH) 2 , H 3 PO 4 , NaNO 3 , and NH 4 Cl were used to produce Na and Cl co-substituted hydroxyapatite. The molar ratio of Ca(OH) 2 /H 3 PO 4 /NaNO 3 /NH 4 Cl was 9.7:6:0.3:1.0.
The Ca(OH) 2 powder was produced by calcining CaCO 3 (Wako Pure Chemicals) at 1050°C for 3 h, and then slaking the resultant CaO with water. In order to synthesize Na and Cl co-substituted hydroxyapatite, the NaNO 3 (Aldrich) and NH 4 Cl (Aldrich) were dissolved in a 1 M Ca(OH) 2 suspension. Next, 1 M H 3 PO 4 (Wako Pure Chemicals) was reacted with the Ca(OH) 2 mixture at room temperature with fast stirring. This process was repeated only apatite was identified after sintering the precipitate at 1100°C for 1 h. The suspension was matured for 1 d at ambient condition and subsequently dried completely in a forced air circulation oven at 70°C to retain water-soluble NaNO 3 and NH 4 Cl. The dried block was subsequently crushed and sieved to separate powders under 75 lm. From now, the apatite substituted by Na and Cl together is abbreviated as NaClAp. Single hydroxyapatite powder was made as a control conforming to the identical processes except for adding NaNO 3 and NH 4 Cl. After this, single hydroxyapatite is abbreviated as HAp.
The HAp and NaClAp disks (/ 10 9 t2 mm) were made under a 30 MPa to evaluate their capabilities of producing HCA and dissolution in the SBF. After being embedded the disks into the pre-sintered and ground NaClAp granules to suppress the volatilization of Na and Cl from the sample, the sintering was conducted at 1100°C for 3 h at a ramping rate of 5°C/min. For osseointegration capacity tests, porous apatite granules were prepared using previously reported method [19]. In brief, polyethylene glycol (PEG; Mw 10000, Aldrich) powder was used as a porogen by generating two different size distributions (75-212 and 212-425 lm) by sieving. The two sizes of PEG were mechanically mixed with the apatite powder. The mixing ratio among apatite, small-sized PEG (75-212 lm), and large-sized PEG (212-425 lm) was 60: 20: 20 (wt %). The resulting mixtures were compacted into disks with a diameter of 50 mm and a height of 30 mm under a pressure of 20 MPa, followed by sintering at 1100°C for 3 h at a heating rate of 5°C/min. The sintered disks were fractured and sieved to generate granules with a size range of 425-600 lm.

Evaluation of HCA producing capability in SBF
The HCA producing capability of HAp and NaClAp was evaluated in SBF [22]. The SBF was made by dissolving CaCl 2 , MgCl 2 Á6H 2 O, KCl, NaCl, NaHCO 3 , Na 2 SO 4 , and K 2 HPO 4 Á3H 2 O in deionized water. The tris(hydroxymethyl) aminomethane and 1 M HCl solution were used to buffer the SBF at pH 7.25. It was subsequently filtered (0.22 lm, Millipore). Before the test, a UV lamp was used for 30 min to sterilize the samples. Subsequently, they were immersed into 30 ml of SBF and then kept for different time intervals at 36.5°C. Next, the samples were mildly rinsed three times with deionized water to prevent the precipitation of various salts after drying and then air-dried under ambient conditions.

Analysis
A field emission scanning electron microscopy (FE-SEM; S-4700, Hitachi) and X-ray diffractometer (XRD; D8 Advance, Bruker) were used to evaluate the microstructures and crystal phases, respectively. Microstructural characteristics were determined with image analysis software (ImageJ, NIH). The shortest diagonal of each grain was determined, and it was used to represent the grain size. The average grain size was determined by measuring at least 500 crystals. Relative densities of the sintered samples were measured by Archimedes' principle. The lattice constants of the sintered HAp and NaClAp were assessed with Rietveld refinement (TOPAS v.2.0, Bruker) [23]. The least squares approach was repeated to refine a theoretical line profile until it matches the measured profile.
An electron probe X-ray microanalyzer (EPMA; JXA-8500F, JEOL) was used to characterize the existence of Ca, P, Na, and Cl on the apatites. The 102 9 102 lm area was examined with a resolution of 0.4 lm.
The Fourier-transform infrared spectroscopy (FT-IR; Spectrum 100, PerkinElmer) was used to assess the functional groups of the samples at a resolution of 4 cm -1 . The sintered samples were ground and then mixed 350-fold with potassium bromide after drying for 1 d in a convection oven at 70°C. The potassium bromide pellets with the samples were compacted into disks under a compression of 1 MPa. The background noise was corrected using the pure potassium bromide disk.
The chemical composition of sintered NaClAp was determined by an inductively coupled plasma atomic emission spectroscopy (ICP-AES; Optima-8300, PerkinElmer) and ion chromatography (IC; ICS-2000, Dionex) after dissolving it in 1 N HCl solution. The concentrations of Ca, P, Na, and Cl dissolved from the samples into SBF after various soaking times were also measured by ICP-AES and IC. For measuring the pH of the SBF, a pH meter (DK-20, Horiba) was used. For each soaking time, five samples were used and the values were averaged and showed as the mean ± standard deviation.
First principle calculation of total system energy The first principle calculations were performed to optimize structures of the apatites. The density functional theory was used with the exchange-correlation energy functional treated by Perdew-Burke-Ernzerhof (PBE) from a generalized gradient approximation [24] with a projector augmented wave [25]. In the hydroxyapatite unit cell, 5% of Ca and 25% of OH were replaced by Na and Cl, respectively, to optimize ab initio calculation procedure. The selfconsistent electronic density function and total energy were calculated with a plane-wave basis set that had a cutoff energy of 400.0 eV (29.40 Ry) using the Vienna ab initio simulation package [26]. The selfconsistent field calculation was repeated until the total energy difference of the systems between the adjacent iterating steps was less than 10 -6 eV. All constituent elements in the apatite system were fully relaxed for structural optimization until the maximum Hellmann-Feynman forces [27] reached a range of ± 2.0 meV/Å with an ionic relaxation scheme based on the conjugate gradient method [28]. The non-shifted and C-point centered 3 9 3 9 3 and 3 9 3 9 1 k-point grids with the Monkhorst-Pack scheme were used to sample Brillouin zone [29]. For Brillouin zone integration, a linear tetrahedron method with Blöchl correction was implemented for the smearing method [30].

Animal experiment
For analyzing the osseointegration capacity of HAp and NaClAp, eight New Zealand white male rabbits weighing 2-2.5 kg were used. The surgery was carried out under both general (ketamine; 7.5 mg/kg, Yuhan and Rompun; 3.5 mg/kg, Bayer Korea) and local (2% lidocaine with 1:100000 epinephrine) anesthesia. Iodine and 5% chlorhexidine digluconate were used to disinfect the surgical site. A midline opening was conducted from the nasofrontal area to the external occipital protuberance. Then, a skin periosteal flap was raised to expose the calvarial bone surface. A trephine bur (8 mm in diameter, 3i-Implant Innovations) was then used to outline the boundary of the bone defect. Subsequently, the bicortical calvarial bone was detached with a round bur. Copious saline irrigation was used during the processes. Two defects were made for each rabbit. Afterward, the respective defects were filled with one of each of the two samples. Next, the closing of periosteum and skin in layers was conducted with 4-0 chromic gut sutures (Ethicon). After three days of surgery, antibiotics were injected intramuscularly.
Four weeks after implantation, all eight rabbits were sacrificed. The 10% neutral buffered formalin was used to fix the bone collected en bloc. A super low-viscosity embedding media (Technovit 7200, Exakt Apparatebau) was used after washing with water and dehydrating to fix the samples without decalcification. Coronally section of the samples was made at a thickness of 80 lm using a diamond disk saw (Secotom-15, Struers). Then, grinding and polishing to 40 lm were carried out using a polisher (Tegrapol-35, Struers). The staining was conducted with Sanderson's Rapid Bone Stain TM and a van Gieson counterstain (Surgipath Medical Industries). The histology was observed using an optical microscope (DM5000B, Leica). Areal percentages of new bone, grafted samples, and connective tissues from whole defect area were measured with image analysis software (ImageJ, NIH).

Animal research ethics
Approval was obtained for all animal experiments, which were performed according to the guidelines of the Institutional Animal Care and Use Committee of Seoul National University (SNU-120222-1).

Statistical analysis
The SPSS Ò Statistics (Ver. 25, SPSS, IBM Ò ) was used for statistical analyses. Data were examined with a paired Student's t test and represented as the mean ± standard deviation with significant level of P \ 0.01.

Results
The phase analyses were carried out to confirm the formations of HAp and NaClAp using XRD (Fig. 1). As a result, only an apatite phase with low crystallinity was identified for both specimens. The functional groups of the HAp and NaClAp powders were evaluated by FT-IR (Fig. 2). The band found at Figure 1 XRD patterns of as-synthesized hydroxyapatite (HAp) and sodium and chloride co-substituted hydroxyapatite (NaClAp) powders. D: hydroxyapatite. carbonate [31,32], while the bands at 1453 and 1421 cm -1 were indexed to the bending (t 3 ) and stretching (t 1 ) modes of carbonate [33], respectively. The band found at about 1383 cm -1 was indexed to symmetric stretching mode (t 3 ) of nitrate [34], which came from initial NaNO 3 . All bands except carbonate and nitrate bands were from hydroxyapatite.
The microstructures of the HAp and NaClAp after sintering were observed by FE-SEM (Fig. 3). The grain shapes and sizes were similar for both materials. Specifically, the average grain sizes of HAp and NaClAp were 0.32 ± 0.14 and 0.39 ± 0.16 lm, respectively. The sintered density of HAp was about 98%, while that of NaClAp was about 96%.
The crystal phase analysis was carried out for HAp and NaClAp disks ( Fig. 4) with XRD after sintering. The crystal phase of HAp was pure hydroxyapatite (JCPDS 09-0432), whereas that of NaClAp was chlorine-bearing hydroxyapatite (JCPDS 70-0794). In fact, the empirical formula of sintered NaClAp determined by ICP-AES and IC was Ca 9.70 Na 0.10 (PO 4 ) 5.32 O 0.68 (OH) 1.51 Cl 0.49 .
To assess the changes in hydroxyapatite structure due to sintering, the functional groups of HAp and NaClAp were characterized by FT-IR (Fig. 5). In addition to the hydroxyapatite bands, two new sets of bands were determined in the NaClAp. The first set of bands was observed at 3545, 3495, and 3460 cm -1 which came from OH-Cl [35]. The second new band was observed at 670 cm -1 , which came from -HO:OH-a 'head-on' configuration [36]. All bands except OH-Cl and -HO:OH-were from hydroxyapatite.
The microstructures of the HAp and NaClAp disks after sintering and corresponding elemental distributions of Ca, P, Na, and Cl measured by EPMA are illustrated in Fig. 6. Only Ca and P were found in the grains of the HAp disk (Fig. 6a). In contrast, Na and Cl, in addition to Ca and P, were homogeneously distributed in every grains of NaClAp disk (Fig. 6b).
Microstructures and detailed microstructures of HAp disks (Fig. 7) and NaClAp disks (Fig. 8) after immersing in SBF for different time periods were examined by FE-SEM. No noticeable changes of microstructures were observed for HAp within 1 w. On the contrary, small acicular apatite crystals covered the whole NaClAp disk surface after only 3 h (Fig. 8a). Moreover, the thickness of newly formed apatite layer progressively increased with time.
Changes in the elemental concentrations of Ca, P, Na, and Cl as well as changes in pH and ionic activity product (IAP) of apatite upon soaking in SBF are shown in Fig. 9. No distinct changes were observed after immersing HAp in SBF. Specifically, immersing of HAp in SBF produced only a slight lessening in the concentrations of Ca and P while no change in the  pH. On the contrary, the concentration of Ca and the pH of the SBF were increased up to 1 d by immersing NaClAp, but these measures decreased fast thereafter. Also, the concentration of P in the SBF a little dropped in the first 6 h and then decreased rapidly thereafter. Meanwhile, the changes of Na and Cl concentrations of the SBF in which NaClAp had been immersed were negligible. The IAP of apatite in SBF was determined using the dissolution data of Ca and P as well as changes in pH over immersing period (Fig. 9f). For the NaClAp, IAPs showed the highest value after approximately 6 h of immersing, after which they dropped fast. On the contrary, the decrease of IAP for HAp was negligible within 1 w.
The atomic structures of HAp [Ca 10 (PO 4 ) 6 (OH) 2 ] and 1 9 1 9 2 NaClAp super-cells [Ca 19 Na(PO 4 ) 11-O(OH) 3 Cl] were optimized by cell volume, cell shape, and atom position. The part of Ca and OH sites was set to the Na and Cl and the CO 3 site was set to single oxygen after releasing CO 2 gas during the sintering. The substitution amount of Na and Cl was set to integral for easy calculations. The OH chain of NaClAp was set to a 'head-on' configuration of HO:OH [36] along the [0 0 1] axis based on FT-IR result (Fig. 5). The atomic structures of the HAp unit cell (Fig. 10a) and 1 9 1 9 2 Na and Cl co-substituted super-cells (Fig. 10b) in the crystal plane (1 0 0) with full ionic relaxation are shown in Fig. 10. Table 1 shows the lattice constancies of apatites compared with experimental data (ab initio and Rietveld) and previous reports [37,38]. The lattice constancies of HAp and NaClAp determined by PBE were in good agreement with results attained by the  Rietveld method and those reported earlier [37,38] within approximately 1% error.
The total energy differences between HAp and NaClAp were determined with DE defined as: where E tot (NaClAp) and E tot (HAp) are the total energies of the corresponding systems; l Ca , l PO 4 , l OH , l H , l Na , and l Cl are the chemical potentials of the corresponding systems; and n is the number of corresponding systems. The determined total energy difference between HAp and NaClAp was ? 15.437 eV. This suggests that the structure of NaClAp was energetically more unstable than HAp.  Table 2 shows the binding energies between apatite structures and the constituent elements of apatite. The binding energy was defined as where E tot (structure), E tot (structureelements), and l elements are the total energies and chemical potentials of the corresponding systems, respectively. The differences of binding energies with the atomic positions in the crystal structures were negligible. Therefore, the average binding energies were used in this experiment ( Table 2). The binding energies of Ca, PO 4 , and OH were notably decreased by Na and Cl substitutions. This means that the bond between the NaClAp and its constituent elements came to be energetically unstable. The binding energies of Na and Cl were much smaller than other constituent elements. The higher total system energy and less stable binding energies of constituent elements of NaClAp compared to HAp explains the higher releasing behaviors of NaClAp in SBF (Fig. 9).
The osseointegration capacities of HAp and NaClAp granules were evaluated using a calvarial defect model in New Zealand white rabbits. Figure 11 displays the FE-SEM image of a NaClAp granule that was used for the osseointegration capacity test. The structure of the HAp granules was the same as that of NaClAp granules. Figure 12 displays the histological images of the calvarial defects after 4 w of surgery for each group. In case of HAp granules, they were mostly enclosed by connective tissues (pale blue) and the amount of newly formed bone (pink) was small (Fig. 12a, c). On the contrary, the newly formed bone (pink) around the NaClAp granules (Fig. 12b, d) was dense and sound. The average occupied % of HAp granules was 48.3 ± 4.5%, while that of NaClAp granules was 45.7 ± 3.3% (p [ 0.01). The areal % of newly formed bone were 5.8 ± 1.8% for the HAp-treated defects while 52.4 ± 6.1% for NaClAp-treated defects (p \ 0.01). Meanwhile, the areal % of soft tissue were 45.9 ± 4.9% for HAp-treated defects and 2.9% ± 1.5% for NaClAp-treated defects, respectively (p \ 0.01).

Discussion
In this study, the defect structures and their effects on the osseointegration capacity of sodium and chloride co-substituted hydroxyapatite were newly studied.
When Ca(OH) 2 , H 3 PO 4 , NaNO 3 , and NH 4 Cl were reacted to produce NaClAp powder, Na-substituted low crystalline carbonate apatite powder with NH 4 Cl was formed (Figs. 1, 2; Scheme 1). As the amount of Na substituting for Ca (x in Scheme 1) increased, CO 3 substitution for PO 4 was stoichiometrically scaled to balance the charge and mass.
The substitution of Na for Ca was confirmed by a little increased amount of CO 3 ions after preparation compared with the pure apatite counterpart (Fig. 2) [8]. Our evidence suggests that the substitution of Cl for OH did not occur at this stage: The chlorine- bearing apatite (ClAp) was not detected in XRD analysis (Fig. 1) and OH-Cl peaks, which occur at 3545, 3495, and 3460 cm -1 [35], were not observed in the FT-IR spectra (Fig. 2), either. Indeed, the substitution of Cl for OH has been reported to start above 600°C [35] and this supports our results.
After sintering, Na and Cl co-substituted hydroxyapatite lacking carbonate ions (Scheme 2) was obtained.
The sintered NaClAp was the chlorine-bearing hydroxyapatite (Fig. 4) with a chlorine content of 0.59 according to JCPDS 70-0794 [39]. Similarly, the empirical formula of sintered NaClAp could be expressed as Ca 9.70 Na 0.10 (PO 4 ) 5.32 O 0.68 (OH) 1.51 Cl 0.49 Figure 9 Changes in concentrations of a calcium, b sodium, and c phosphorous as well as changes in d pH, e chlorine concentration, and f ionic activity products (IAP) of apatite in SBF after soaking hydroxyapatite (HAp) and sodium and chloride cosubstituted hydroxyapatite (NaClAp) disks for the indicated time intervals.
based on the reaction formula in Scheme 2. This means that about a quarter of the OH site was substituted with Cl. The large Cl ion prefers the halfway position between the Ca-triangles, whereas the OH ion places slightly above or below the plane of the three Ca [35]. With the increase in Cl concentration, the a-axis expands from 9.418 Å in the hydroxyapatite to 9.634 Å in the chlorapatite, while the c-axis contracts from 6.884 to 6.778 Å . This implies that the unit cell volume expands with the Cl substitution [35] and explains the peak shift from the peak position of pure hydroxyapatite to chlorine-bearing hydroxyapatite (Fig. 4b). Meanwhile, the substituted amount of Na and Cl in NaClAp was less than those initially added and this is likely due to their volatilizations during the sintering process. Figure 10 Crystal structures of a a hexagonal unit cell of hydroxyapatite (space group P6 3 /m) consisting of 10 calcium atoms, 6 phosphorous atoms, 26 oxygen atoms, and 2 hydrogen atoms, b 1 9 1 9 2 super-cells of sodium and chloride co-substituted hydroxyapatite replacing one calcium ion with one sodium ion, one hydroxide ion with one chloride ion, and one oxygen ion at carbonate ion vacancy compared to the hydroxyapatite structure. The substitution of Ca with Na produced negatively charged point defect (Na in the Ca site) as shown in Scheme 3. Meanwhile, the CO 3 ions in PO 4 site were thermally decomposed as CO 2 gas during sintering and produced positively charged point defect (O in the PO 4 site) as shown in Scheme 3. This phenomenon was described as the defect incorporation reaction following Krö ger-Vink notation in Scheme 3 as balancing the charge and mass [20]. Concurrently, a portion of the OH sites also accommodated Cl (Schemes 2, 3), thereby generating OH-Cl bonds and caused a partially 'head-on' structure of HO:OH in an hydroxyl configuration (Fig. 5) [36].
Elemental mappings of HAp and NaClAp disks by EPMA (Fig. 6) supported XRD and FT-IR findings   that Na and Cl were distributed homogeneously all over the apatite grain surfaces of the NaClAp disks ( Fig. 6b) but not the HAp disks (Fig. 6a). Collectively, these results obviously implied the co-substitution of Na and Cl for the Ca and OH sites, respectively, of hydroxyapatite after sintering.
ICP-AES evaluation of SBF while NaClAp disks were soaking (Fig. 9) suggests a mechanism by which low crystalline hydroxyl carbonate apatite (HCA) formed on the surface of the NaClAp. The Ca (Fig. 9a) and OH ions (Fig. 9d) were continuously released from NaClAp disks up to 1 d. The subsequent decrease in the concentrations of Ca, P, and OH ions implied that consumption of these ions to form HCA crystals was faster than their release rate. It is noteworthy that the ionic activity product (IAP) of apatite in SBF showed a maximum value (10 -95.9 ) after approximately 6 h and decreased rapidly thereafter (Fig. 9f). Actually, the high maximum IAP coupled with the low solubility of apatite in water (5.5 9 10 -118 at 37°C [40]) was sufficient to produce nucleation and growth of apatite crystals in SBF. However, the changes in IAPs were also negligible for HAp disks (Fig. 9f). This observation supports only a minor producing capability of HCA on HAp (Fig. 7). Collectively, these results imply that the fast dissolution of Ca and OH ions from NaClAp into SBF resulted in apatite supersaturation and consequently induced the formation of HCA on the NaClAp disks.
The total system and binding energies of apatite can explain the dissolution behaviors of constituent elements of apatite in SBF. Specifically, the total system energy of NaClAp was markedly elevated compared with that of HAp (? 15.437 eV) when Na was substituted for Ca and Cl for OH in HAp. Besides, the binding energies between apatite system and constituent elements of apatite (Ca, PO 4 , and OH) decreased with Na and Cl co-substitution. This explains that the higher solubility of NaClAp compared with that of HAp in SBF.
The NaClAp, with enhanced HCA producing capability in SBF, demonstrated considerably higher osseointegration capacity compared with HAp ( Fig. 12). Specifically, dense and sound bone was formed on the NaClAp granules without intervening connective tissues. On the contrary, the amount of newly formed bone on HAp granules was small. These results are consistent with previous reports, which showed high capability to form HCA [41][42][43][44][45][46][47][48].
Together, the results of this study suggest that NaClAp has improved osseointegration capacity over HAp. The enhanced HCA producing capability in SBF and osseointegration capacity of NaClAp were largely attributed to improved solubility, which could be explained by a higher total system energy achieved by the formation of various defects. This means that the co-substitution of multiple ions into hydroxyapatite structure is a useful way to improve the osseointegration capacity of hydroxyapatite.

Conclusions
The defect structures of Na and Cl co-substituted hydroxyapatite and its osseointegration capacity were newly studied. Substitutions of Na for Ca and Cl for OH in HAp produced negatively charged sodium in calcium site, positively charged oxygen in phosphate site, OH-Cl bonds, and a partially 'headon' structure of HO:OH in an hydroxyl configuration. These defects made NaClAp energetically unstable and consequently improved the HCA formation and osseointegration capacities. Taken together, our results indicate that the co-substitution of multiple ions into hydroxyapatite structure is a useful method to improve the osseointegration capacity of hydroxyapatite.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licen ses/by/4.0/.