Investigation of the local environment of iodate in hydroxyapatite by combination of X-ray absorption spectroscopy and DFT modeling

Abstract

Iodate-substituted hydroxyapatites are novel phases of potential interest for the immobilisation of radioactive iodine-129. However, the local environment of the iodine in these phases is still unclear. Here, a combined experimental–computational strategy has been used to investigate the mode of incorporation of iodates (IO₃⁻) in Ca–hydroxyapatite and Sr–hydroxyapatite lattices. On one hand, I K- and L₃-edge XANES (X-ray Absorption Near Edge Structure) spectra are presented, showing that while the local structure around the iodate is similar in the two substituted hydroxyapatite lattices, it differs significantly from the one observed in a series of model compounds (NaIO₃, KIO₃, Ca(IO₃)₂·H₂O). I K-edge EXAFS (Extended X-ray Absorption Fine Structure) spectra were then analysed, revealing the lack of order around the iodate, and also the absence of local clustering of the iodates along the hydroxyl columns of the apatite. Further insight into the local environment of iodates in apatites was then obtained by DFT (Density Functional Theory) computational modeling of iodate-substituted Ca–hydroxyapatite lattices.

---

Introduction

Iodine-129 is a radioactive isotope which is produced by the nuclear fission of ²³⁵U and ²³⁹Pu. It is classified as “long-lived intermediate-level” waste (LL-ILW), its half-time being ∼1.6 × 10⁶ years and its specific activity ∼6.5 × 10⁶ Bq g⁻¹. Due to the radiotoxicity of iodine-129 for wildlife and humans, investigations have been carried out to determine ways for its capture and/or immobilisation in deep geological repositories.^1,2 In particular, different strategies have been proposed to try to trap iodine into robust ceramic matrices like apatites^3,4 and zeolites.⁵

Apatites are a family of compounds of general formula M₁₀(XO₄)₆Y₂, where M is generally a divalent metal cation (Ca²⁺, Pb²⁺,…), XO₄ an oxoanion (PO₄³⁻, VO₄³⁻,…), and Y a mono-charged anion (OH⁻, F⁻,…).⁶ Iodine has been incorporated in the apatite lattice under the form of iodide (I⁻) in phases like Pb₁₀(VO₄)_4.8(PO₄)_1.2I₂ and Ca₁₅(PO₄)₉IO,^3,7 and more recently as an iodate (IO₃⁻) in the Ca₁₀(PO₄)₆(OH)_1.6(IO₃)_0.4 phase.⁴ In the latter case, the mechanism of incorporation of iodates in the Ca–hydroxyapatite lattice (Ca₁₀(PO₄)₆(OH)₂, HA) was demonstrated using IR and ¹H solid state NMR spectroscopies, showing the partial substitution of hydroxyls by iodates along the channels that are running along the crystallographic c axis.⁴ Although this substitution scheme is consistent with the fact that (IO₃)⁻ and (OH)⁻ have the same charge, it seemed somewhat surprising because of the large difference in size and geometry of both anions. Thus, it appeared to us important to better characterise the local environment around the iodate in this structure, in order to (i) understand how the calcium hydroxyapatite structure accommodates to the presence of this anion, (ii) rationalise the confinement properties of this material, and thus (iii) be in position of proposing and developing new classes of apatite phases with improved confinement properties.

Substitution mechanisms of trace ions in apatite lattices are most commonly studied by X-ray and neutron diffraction, by performing Rietveld refinements on high-resolution powder patterns.⁸ Solid state NMR^9,10 and X-ray absorption spectroscopy⁹ are other experimental techniques which have proven their efficiency to provide information on the local environment of ionic substituents in apatites, especially when trace levels are incorporated, and when local structural disorder is present. Furthermore, computational modeling of substituted apatite lattices can provide complementary information on the structural changes occurring upon incorporation of various ions.^10–12

In the case of iodate-substituted HA, the X-ray diffraction powder patterns were found to be unsuitable for structural refinements by the Rietveld method.⁴ The purpose of this article is thus to show how insight into the local structure around iodates in phases like Ca₁₀(PO₄)₆(OH)_1.6(IO₃)_0.4 can be obtained by combining XANES (X-ray Absorption Near Edge Structure) and EXAFS (Extended X-ray Absorption Fine Structure) experiments, at both the K and L₃ edges of iodine, with computational modeling.

Experimental procedures

Sample origin and preparation

NaIO₃ (Alfa Aesar, Germany, 99.0% purity), KIO₃ (Alfa Aesar, Germany, 98.0% purity) and Ca(IO₃)₂·H₂O (Acros Organics, Belgium, 98.0% purity) were purchased from commercial sources and used as received. These three metal iodates were used as standards in the EXAFS and XANES characterisations.

The iodate-substituted Ca–hydroxyapatite (CaI–HA, Ca₁₀(PO₄)₆(OH)_1.6(IO₃)_0.4) was prepared according to our previously published procedure.⁴

A new iodate-substituted Sr–hydroxyapatite phase (SrI–HA) was also synthesised for the purpose of the XAS analyses. SrI–HA was synthesised using Sr(NO₃)₂ (Alfa Aesar, Germany, 99.0% purity), (NH₄)₂HPO₄ (VWR, France, 99.1% purity) and NaIO₃ (Alfa Aesar, Germany, 99.0% purity) as reagents. All solutions were prepared with ultrapure water, and ethylenediamine (Merck) was used to adjust their pH. The following reaction was performed at room temperature and under an N₂ atmosphere, in order to avoid as much as possible the incorporation of carbonates in the apatite lattice. First, an aqueous solution of strontium nitrate was prepared ([Sr²⁺] = 0.0556 mol L⁻¹; V = 100 mL; pH was adjusted to ∼11.0 using a 1 mol L⁻¹ solution of C₂H₄(NH₂)₂). Then, an aqueous solution of (NH₄)₂HPO₄ and NaIO₃ was prepared ([“P”] = 0.0334 mol L⁻¹, and [“I”] = 0.011 mol L⁻¹; V = 100 mL; pH adjusted to ∼11.0 by addition of concentrated C₂H₄(NH₂)₂), and added dropwise to the Sr solution under magnetic stirring. This led to the progressive precipitation of a white solid. The suspension was stirred at room temperature for 5 hours, during which small amounts of concentrated C₂H₄(NH₂)₂ were regularly added, in order to keep the pH close to its initial value. The precipitate was recovered by filtration, before being dried at 120 °C overnight. This solid was then washed to remove soluble by-products and physisorbed species like nitrate ions. The washing procedure consisted in a magnetic stirring of a 5 g L⁻¹ suspension of the powder in ultrapure water for one hour (typically, 1 g of the product collected after drying was suspended with 200 mL of ultrapure water). The suspension was filtered and the solid ultimately dried at 90 °C for ∼16 hours. After the complete treatment, ∼1 g of solid was obtained, which corresponds to a yield of ∼77% assuming the formation of Sr₁₀(PO₄)₆(OH)_1.7(IO₃)_0.3. This formula was established on the basis of elemental analyses and of the measurement of iodine concentration using an iodide selective electrode (Thermo Scientific). XRD and IR characterisations of this sample can be found in the ESI (see Fig. S1 and S2 in ESI†).

XAS acquisitions, data reduction and analysis

Iodine X-ray absorption spectroscopy measurements were performed at the Soleil Synchrotron (Saint-Aubin, France). The Soleil ring energy was 2.75 GeV and the current was 400 mA. Samples were ground to a fine powder, diluted in cellulose, pressed into pellets, and run at ∼77 K in transmission mode, except for the XANES spectrum of SrI–HA at the L₃-edge, which was recorded in fluorescence mode.

Analyses at the L₃-edge were performed on the LUCIA beamline.¹³ The X-ray incident energy on the sample was defined using a double-crystal Si(111) monochromator, whose calibration was performed by setting at 4966 eV the first inflexion point of a titanium foil XANES spectrum. The transmission-mode spectra were collected scanning the pre-edge (4500–4552 eV), edge (4552.2–4632 eV) and post-edge (4633–4820 eV) regions in 5.0, 0.2 and 1.0 eV steps respectively, with dwell times per point of 2.0 to 4.0 s. Typically, 2 to 3 XANES data sets were collected for each sample and averaged. The XANES spectrum recorded in fluorescence mode was scanned using a dwell time per point of 6.0 s; a total of 8 XANES spectra were collected and averaged.

Analyses at the K-edge were performed on the SAMBA beamline.¹⁴ The X-ray incident energy on the sample was defined using a double-crystal Si(220) monochromator. The monochromator calibration was performed by collecting the XANES spectrum of a tin foil and setting the first inflexion point at 29 200 eV. The XANES region was scanned with 5 eV, 1 s in the pre-edge (32 800–33 100), 1.0 eV, 1 s in the edge (33 100–33 210) and 1.5 eV, 1 s in the post-edge (33 210–33 300). The EXAFS region was scanned on a 20 Å⁻¹ range in the wave vector k space (Å⁻¹), setting E₀ at 33 169 eV, with steps of 0.05 Å⁻¹ and dwell time increasing from 2 to 4 s. Depending on the data quality, 2 to 10 XAS data sets were collected per sample and averaged.

XANES and EXAFS data were reduced using standard normalisation procedures using the ATHENA package.¹³ Although EXAFS studies were performed at both the I L₃ and K-edges, only the results of analyses at the K-edge are presented here, as they were found to be of better quality. EXAFS spectra were extracted using the Autobk algorithm. The EXAFS χ(k) functions were weighted by k³ to enhance oscillations at high k values. The k³-weighted EXAFS spectra were Fourier transformed over the k-range 3.0–18.0 Å⁻¹ using a Kaiser–Bessel apodisation window (window parameter = 3).

EXAFS data analyses were performed using the ARTEMIS software package.¹⁵ Theoretical EXAFS backscattering paths were calculated from the atomic positions resulting from the computational modeling. The paths lists were then obtained by running the subroutines ATOMS and FEFF6 included in the ARTEMIS program. Only simple scattering paths were considered for fitting. Paths selection was based on their relevance and relative contribution to the EXAFS signal (calculated by FEFF). The fitting analysis was restrained to the [0.85–4.2] Å domain of the radial distribution function obtained by the Fourier transform of the EXAFS oscillations. The amplitude factor S₀² was set at 1.1 for all shells of all samples.

Computational modeling

All calculations were performed using the ab initio plane-wave pseudopotential approach as implemented in the VASP code.¹⁶ The Perdew–Burke–Ernzerhof (PBE)¹⁷ generalised gradient approximation functional was chosen to perform the periodic DFT calculations. The valence electrons were treated explicitly and their interactions with the ionic cores are described by the Projector Augmented-Wave method (PAW), which allows a low energy cut-off equal to 400 eV to be used for the plane-wave basis.

A double cell (in the c direction) obtained from the hydroxyapatite monoclinic structure (ICSD 97849)¹⁸ was used as a starting point.‡ Four substitution models were obtained by replacing one of the columnar OH by an IO₃⁻ group, varying the initial orientation of the iodate.

All atomic positions of these models were then relaxed at 0 K without geometrical constraints, keeping the cell parameters fixed (a = 9.419 Å, b = 18.848 Å, c = 13.768 Å, γ = 119.98°). The integral over the first Brillouin zone was performed using a Monkhorst–Pack 1 × 1 × 1 k-point grid. The optimised structure was obtained after the total energy differences between the loops were less than 10⁻⁴ eV.

Results and discussion

Both iodine K- and L₃-edge X-ray Absorption Spectroscopy (XAS) experiments have been reported in the literature.^19–21 EXAFS experiments at the K-edge have been shown to be much more sensitive,²¹ due to the longer k-range of available oscillations. However, the XANES spectra are generally broad with few features, due to the substantial core–hole lifetime at the K-edge.¹⁹ In contrast, I L₃-edge XANES spectra have much more resolved features, but the shorter k-range (which is due to the proximity with the L₂ edge) hampers the determination of information on long-range distances around the iodine using the EXAFS data. Furthermore, oscillations coming from the calcium K-edge may still remain at the iodine L₃ edge,²¹ thereby hampering an accurate extraction of the iodine EXAFS oscillations (Fig. S3†). Thus, overall, there is a good complementarity between both edges, and results of XAS analyses performed at both edges are presented below.

XANES

The I L₃ and K-edge XANES spectra of the iodate substituted apatites CaI–HA and SrI–HA are shown in Fig. 1, together with those of a series of crystalline model compounds of known structure (NaIO₃, KIO₃ and Ca(IO₃)₂·H₂O). At both edges, the overall profile (edge position and features) is consistent with the presence of iodine under the form of an iodate.^20,21

Fig. 1 I L₃ and K-edge XANES spectra of the model compounds (Na(IO₃), K(IO₃), Ca(IO₃)₂·H₂O) and iodate-substituted HA (CaI–HA and SrI–HA). Details on the measurement conditions used to acquire the different XANES spectra can be found in the Experimental section.

At the L₃-edge, the XANES spectra show many features (especially a pre-edge at 4.562 keV), corresponding to transitions mainly towards vacant 5d orbitals.²⁰ A composite oscillation is then observed between 4.582 and 4.596 keV, associated to multiple scattering events suffered by the photoelectron embedded in a long or medium range order. Differences are observed for all model samples, which is consistent with the differences in local environment of the iodate ion in these phases. The XANES spectra of CaI–HA and SrI–HA are similar, and also resemble the spectrum of NaIO₃.

At the K-edge, much broader transitions are observed,¹⁹ the edge corresponding to transitions towards vacant hybridised 5p levels, and the region between ∼33.184 and 33.200 keV to multiple scattering. The analogy of the XANES spectra of CaI–HA and SrI–HA is still apparent. However, their comparison with NaIO₃ reveals differences, notably in the “multiple scattering” region above 3.18 keV. Thus, in contrast with our previous hypothesis (which was based on I L₃-edge XANES spectra only),⁴ analogies between the local environment of the iodate in CaI–HA and NaIO₃ do not hold when the XANES spectra recorded at the I K-edge are compared. This underscores the importance of performing XANES analyses at both edges, before drawing conclusions on iodate local environments.

All in all, based on the I L₃ and K-edge XANES analyses, it can be concluded that the local structure around the iodates in CaI–HA and SrI–HA are similar, but that they are different to those encountered in the model compounds NaIO₃, KIO₃ and Ca(IO₃)₂·H₂O. The I L₃-edge XANES spectra also differ from those of a simple iodate anion in solution,²¹ which shows that neighbouring atoms beyond the first oxygen coordination shell contribute to the multiple scattering region in both iodate-substituted HA phases. Thus, in view of gaining insight into the arrangement of the more distant cations (Ca or Sr) and anions (phosphates) around the iodate, EXAFS experiments were carried out.

EXAFS

The EXAFS oscillations and their Fourier transforms are shown in Fig. 2 and S4.† For all compounds, the main oscillation arises from the 3 O atoms covalently bound to the iodine, which leads to the strong peak centered at ∼1.4 Å after Fourier transformation (distance not corrected from back-scattering phase shift).§ As the purpose of the EXAFS analyses was to gain insight into the arrangement of atoms beyond these 3 oxygen atoms, a detailed study on model crystalline phases was first performed (NaIO₃, KIO₃ and Ca(IO₃)₂·H₂O), in order to determine what type of information could be extracted from their EXAFS spectra.

Fig. 2 Experimental and fitted I K-edged EXAFS data of K(IO₃), Ca(IO₃)₂·H₂O and CaI–HA, showing (a) the EXAFS oscillations, and (b) the Fourier transform of the EXAFS data. Details on the measurement conditions used to acquire the different EXAFS spectra can be found in the Experimental section, as well as the fitting procedure.

By looking at the Fourier transforms of the EXAFS data of the 3 models, it appears that only K(IO₃) and Ca(IO₃)₂·H₂O have significant contributions beyond 2 Å. Simulations of their EXAFS data were carried out by choosing different paths, based on the iodine local environment in each of the crystal structures (Fig. S5†).²² The best fits are shown in Fig. 2, and the parameters used for the fits are reported in Table 1.¶ In both cases, discrepancies in the distances and exact number of back-scattering atoms can be observed, in comparison with the experimental structures. Nevertheless, these simulations show that for crystalline compounds, contributions from sufficiently “heavy” neighbouring cations like K⁺ can be seen between 2 and 4 Å, but also and mainly from nearby iodine atoms (as for the Ca(IO₃)₂·H₂O structure). To our knowledge, this is the first time that the I K-edge EXAFS data of K(IO₃) and Ca(IO₃)₂·H₂O is analysed in such depth.

Table 1 Parameters used in the fits of the I K-edge EXAFS data of K(IO₃), Ca(IO₃)₂·H₂O and CaI–HA (data shown in Fig. 2)^a

Compound	S0^2	ΔE (eV)	R^b	Paths	N^c	R^d (Å)	σ^2^d (Å^2)
a The S0^2 values were extracted from the FEFF fits of the EXAFS data of the model compounds, and then applied to the fitting of iodate-substituted apatite phase.b The equation defining R values can be found in ref. 15 and 23 (and in the related documents).c The number of neighbours was set to a fixed value during the fit. For KIO3 and Ca(IO3)2·H2O, it corresponds to an average over the different crystallographically distinct sites in the structure.d Standard deviations were calculated from the fit.e A comparison of the fits with and without inclusion of this backscattering shell of 5 O atoms was carried out, showing that the fit is of better quality when these O atoms are included (χ^2 decreases from 465 to 297, and R from 0.009 to 0.006), and that their contribution appears between 2.0 and 2.6 Å.
KIO3	1.1	9.94	0.007	O	3	1.81 ± 0.01	0.002 ± 0.001
				O	3	2.76 ± 0.01	0.012 ± 0.002
				K	4	3.80 ± 0.02	0.014 ± 0.003
				K	4	3.98 ± 0.03	0.015 ± 0.004
Ca(IO3)2·H2O	1.1	9.61	0.006	O	3	1.81 ± 0.01	0.002 ± 0.001
				O	5^e	2.93 ± 0.03	0.023 ± 0.005
				Ca	1.5	3.61 ± 0.02	0.011 ± 0.002
				I	4	3.92 ± 0.01	0.010 ± 0.001
CaI–HA	1.1	9.15	0.010	O	3	1.82 ± 0.01	0.002 ± 0.001

---

In the case of CaI–HA, no clear contribution beyond 2 Å could be observed (Fig. 2). The data was thus fitted considering only the presence of the 3 O atoms of the iodate. The lack of back-scattering signals can be due to the presence of distributions in local environments around the iodate in the crystal structure, resulting in the absence of any privileged back-scattering pathway. However, it also indicates the absence of significant clustering of the iodates along the OH channels. Indeed, if two successive hydroxyls had been substituted by iodates in the Ca–HA lattice, I⋯I distances below 4 Å would have been expected, and should have been apparent on the spectra, given the previous observations on Ca(IO₃)₂·H₂O.

Because Sr²⁺ cations are much better back-scatterers than Ca²⁺, it was expected that more information on the positioning of the cations around the iodate would be reached by analyzing the EXAFS spectra of SrI–HA. As shown in Fig. S4,† the spectra representing the Fourier transform of the EXAFS data seem to show very weak signals beyond 2 Å. However, despite long acquisition times, the level of noise on the spectra was high (due to the strong absorption of the Sr), meaning that no unambiguous simulation of this EXAFS spectrum was possible.

Overall, the EXAFS analyses carried out at the I K-edge show that additional information on the local environment of the iodates in HA was reached: (i) there is a large range of I⋯X distances around the iodate, due to the positioning of the iodate in the HA crystal structure, and/or to the presence of different orientations of the iodates in the lattice; and (ii) iodates are well dispersed along the hydroxyl columns and do not tend to cluster. The former point shows that the difficulties in exploiting the XRD powder patterns of these phases for structure resolution by Rietveld refinements originate not only in the crystallite size and shape (small acicular-shaped crystallites), but also in the lack of long range order around the iodate.

DFT modeling of iodate-substituted hydroxyapatites

In order to gain more direct insight into how the Ca–HA lattice can adapt to the incorporation of iodates along the OH columns, computational models were developed. More specifically, geometry optimisations of substituted-HA lattices were carried out, the initial orientation of the iodate differing from one model to the other. It should be noted that in order to reduce the computational cost of these DFT calculations, the iodate content chosen in the calculations (Ca₁₀(PO₄)₆(OH)_1.75(IO₃)_0.25) was lower than the one obtained experimentally (Ca₁₀(PO₄)₆(OH)_1.6(IO₃)_0.4). With such a composition, no clustering of iodates along the OH columns can occur, in line with experimental observations.

The optimised structures of four different iodate-substituted HA are shown in Fig. S6.† The relative energy of these four models is different after optimisation (Table S1†), suggesting that the configuration proposed in “model 1” is the most probable. As only four different models of iodate-substituted HA were tested here, other more stable configurations (with a different iodate orientation) may have been overlooked. Nevertheless, from this preliminary study, it is clear that the HA lattice can adapt to different iodate orientations, making very likely the possibility of having different iodate orientations in the CaI–HA lattice.

A comparison of the optimised structure of non-substituted and iodate-substituted HA lattices for the most stable model is shown in Fig. 3. It reveals that the incorporation of the iodate along the OH columns leads to distortions in the HA lattice, notably by impairing the alignment of the Ca(II)-type cations along the c direction. This results in a larger distribution in Ca⋯O distances in the substituted hydroxyapatite, in agreement with the Ca K-edge XANES data (Fig. S7†). The orientation of the OH groups is affected as well, by loss of co-linearity with the c axis. These structural changes are also observed for the less stable models (Fig. S6†). It is worth noting that in all models, the analysis of the I⋯X distances (X = O, Ca, P) around the iodate clearly shows that they progressively increase beyond 2 Å, and that there is no specific distance at which several “heavy” back-scatterers are present.||

Fig. 3 Representation of the optimised geometries of Ca–HA (left) and iodate-substituted CaI–HA (right). Ca, P, O, H, and I atoms are in green, dark brown, red, white, and blue, respectively.

All in all, this computational study confirms the hypotheses made concerning the absence of back-scattering beyond 2 Å in the EXAFS spectra of CaI–HA: (i) the HA lattice can adapt to several orientations of the iodate along the column; and (ii) whatever the iodate orientation, no predominant back-scattering pathway could be made evident. As a matter of fact, attempts to simulate the EXAFS data of CaI–HA using the geometrical parameters around the iodate in model 1 were unsuccessful, further underscoring the absence of a unique iodate orientation along the columns.

Given the evidence of multiple scattering on the I L₃-edge XANES spectrum of CaI–HA, another way to test the validity of the computational models would be to perform ab initio calculations of the I L₃-edge XANES spectra, based on the full multiple scattering (MS) theory. However, the preliminary studies we carried out in this direction on the model compounds clearly showed that such calculations are not straightforward for iodine compounds, and that they will require a complete separate study in itself.

Conclusions

In this manuscript, we have shown that insight into the local environment of the iodate in substituted hydroxyapatite phases can be obtained by combining X-ray absorption spectroscopy experiments (at both the I K and L₃ edges) and DFT computational modeling. Both experiments and simulations show the absence of local ordering around the iodate in the structures.

Although this study provides complementary information about the structure of iodate-substituted HA phases to what had already been established,⁴ a higher level of description of the iodate local environment may be accessible, by (i) improving the computational models (by starting from larger supercells, to reach iodate contents more similar to those actually obtained experimentally, and also to model the disorder actually present along the OH columns), (ii) performing ab initio MS calculations of the I L₃-edge XANES spectra of these structural models, to compare them to those observed experimentally, and (iii) studying iodate-substituted HA phases by “NMR-crystallography”,²⁴ to further validate the models. Such directions will be looked into in the future.

Acknowledgements

This work was supported by CEA, CNRS and the MATINEX National Research Group. We acknowledge the French synchrotron SOLEIL for provision of synchrotron radiation facilities (projects 20110099 and 20110704). DFT calculations were performed on the IDRIS supercomputer centre of the CNRS (Project 091461).

Notes and references

1. D. F. Sava, M. A. Rodriguez, K. W. Chapman, P. J. Chupas, J. A. Greathouse, P. S. Crozier and T. M. Nenoff, J. Am. Chem. Soc., 2011, 133, 12398; K. W. Chapman, D. F. Sava, G. J. Halder, P. J. Chupas and T. M. Nenoff, J. Am. Chem. Soc., 2011, 133, 18583; J. T. Hughes, D. F. Sava, T. M. Nenoff and A. Navrotsky, J. Am. Chem. Soc., 135, 16256; C. Falaise, C. Volkringer, J. Facqueur, T. Bousquet, L. Gasnot and T. Loiseau, Chem. Commun., 2013, 49, 10320; D. F. Sava, K. W. Chapman, M. A. Rodriguez, J. A. Greathouse, P. S. Crozier, H. Zhao, P. J. Chupas and T. M. Nenoff, Chem. Mater., 2013, 25, 2591.
2. J. L. Krumhansl and T. M. Nenoff, Appl. Geochem., 2011, 26, 57; J. D. Pless, J. B. Chwirka and J. L. Krumhansl, Environ. Chem. Lett., 2007, 5, 85; T. J. Garino, T. M. Nenoff, J. L. Krumhansl and D. X. Rademacher, J. Am. Ceram. Soc., 2011, 94, 2412; D. F. Sava, T. J. Garino and T. M. Nenoff, Ind. Eng. Chem. Res., 2012, 51, 614.
3. M. Uno, M. Shinohara, K. Kurosaki and S. Yamanaka, J. Nucl. Mater., 2001, 294, 119; C. Guy, F. Audubert and J. E. Lartigue, C. R. Phys., 2002, 3, 827; L. Campayo, S. Le Gallet, Y. Grin, E. Courtois, F. Bernard and F. Bart, J. Eur. Ceram. Soc., 2009, 29, 1477; J.-M. Montel, C. R. Geosci., 2011, 343, 230; M. C. Stennett, I. J. Pinnock and N. C. Hyatt, J. Nucl. Mater., 2011, 414, 352; F. Lu, Z. Dong, J. Zhang, T. White, R. C. Ewing and J. Lian, RSC Adv., 2013, 3, 15178.
4. L. Campayo, A. Grandjean, A. Coulon, R. Delorme, D. Vantelon and D. Laurencin, J. Mater. Chem., 2011, 21, 17609; D. Laurencin, A. Grandjean, and L. Campayo, Fr. Demande, 2011, FR 2957913 A1 20110930 10/52216 (patent deposited on 03/26/2010).
5. Y. Watanabe, T. Ikoma, H. Yamada, Y. Suetsugu, Y. Komatsu, G. W. Stevens, Y. Moriyoshi and J. Tanaka, ACS Appl. Mater. Interfaces, 2009, 1, 1579; K. W. Chapman, P. J. Chupas and T. M. Nenoff, J. Am. Chem. Soc., 2010, 132, 8897; T. J. Garino, T. M. Nenoff, J. L. Krumhansl and D. X. Rademacher, Ceram. Trans., 2010, 217, 35.
6. J. C. Elliott, Structure and chemistry of the apatites and other calcium orthophosphates, in Studies in inorganic chemistry, Elsevier, Amsterdam, 1994, vol. 18; Y. Pan and M. E. Fleet, Rev. Mineral. Geochem., 2002, 48, 13.
7. P. A. Henning, S. Lindin and V. Petricek, Acta Crystallogr., Sect. B: Struct. Sci., 1999, 55, 165.
8. S. Gomes, J.-M. Nedelec and G. Renaudin, Acta Biomater., 2012, 8, 1180; S. Gomes, J.-M. Nedelec, E. Jallot, D. Sheptyakov and G. Renaudin, Cryst. Growth Des., 2011, 11, 4017; M. D. O'Donnell, R. G. Hill and S. K. Fong, Mater. Lett., 2009, 63, 1347; A. Kyriacou, Th. Leventouri, B. C. Chakoumakos, V. O. Garlea, C. B. dela Cruz, A. J. Rondinone and K. D. Sorge, J. Mater. Sci., 2013, 48, 3535.
9. D. Laurencin, A. Wong, W. Chrzanowski, J. C. Knowles, D. Qiu, D. M. Pickup, R. J. Newport, Z. Gan, M. J. Duer and M. E. Smith, Phys. Chem. Chem. Phys., 2010, 12, 1081; S. Gomes, G. Renaudin, A. Mesbah, E. Jallot, C. Bonhomme, F. Babonneau and J.-M. Nedelec, Acta Biomater., 2010, 6, 3264.
10. D. Laurencin, N. Almora-Barrios, N. de Leeuw, C. Gervais, C. Bonhomme, F. Mauri, W. Chrzanowski, J. C. Knowles, R. J. Newport, A. Wong, Z. Gan and M. E. Smith, Biomaterials, 2011, 32, 1826; H. Yi, E. Balan, C. Gervais, L. Segalen, F. Fayon, D. Roche, A. Person, G. Morin, M. Guillaumet, M. Blanchard, M. Lazzeri and F. Babonneau, Am. Mineral., 2013, 98, 1066.
11. J. R. Tolchard, J. E. H. Sansom, M. S. Islam and P. R. Slater, Dalton Trans., 2005, 1273; J. R. Tolchard, P. R. Slater and M. S. Islam, Adv. Funct. Mater., 2007, 17, 2564; D. Bazin, X. Carpentier, I. Brocheriou, P. Dorfmuller, S. Aubert, C. Chappard, D. Thiaudie, S. Reguer, G. Waychunas, P. Jungers and M. Daudon, Biochimie, 2009, 91, 1294; J. Terra, E. Rodrigues Dourado, J.-G. Eon, D. E. Ellis, G. Gonzalez and A. M. Rossi, Phys. Chem. Chem. Phys., 2009, 11, 568.
12. S. Peroos, Z. Du and N. H. de Leeuw, Biomaterials, 2006, 27, 2150; N. H. De Leeuw, Phys. Chem. Chem. Phys., 2002, 4, 3865; Y. Tang, H. F. Chappell, M. T. Dove, R. J. Reeder and Y. J. Lee, Biomaterials, 2009, 30, 2864; D. E. Ellis, J. Terra, O. Warschkow, M. Jiang, G. B. Gonzalez, J. S. Okasinski, M. J. Bedzyk, A. M. Rossi and J. G. Eon, Phys. Chem. Chem. Phys., 2006, 8, 967; A. Al-Yasari, A. Jones, A. Orera, D. C. Apperley, D. Driscoll, M. S. Islam and P. R. Slater, J. Mater. Chem., 2009, 19, 5003.
13. A. M. Flank, G. Cauchon, P. Lagarde, S. Bac, M. Janousch, R. Wetter, J. M. Dubuisson, M. Idir, F. Langlois, T. Moreno and D. Vantelon, Nucl. Instrum. Methods Phys. Res., Sect. B, 2006, 246, 269.
14. V. Briois, E. Fonda, S. Belin, L. Barthe, C. La Fontaine, F. Langlois, M. Ribbens and F. Villain, SAMBA: The 4–40 keV X-ray absorption spectroscopy beamline at SOLEIL, UVX 2010 (10^e Colloque sur les Sources Cohérentes et Incohérentes UV, VUV et X; Applications et Développements Récents), EDP Sciences, 2011, 41.
15. B. Ravel and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT, J. Synchrotron Radiat., 2005, 12, 537.
16. G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169; G. Kresse and J. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758.
17. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865; J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1997, 78, 1396.
18. Y. Suetsugu and J. Tanaka, J. Mater. Sci.: Mater. Med., 2002, 13, 767.
19. I. Bonhoure, A. M. Scheidegger, E. Wieland and R. Dähn, Radiochim. Acta, 2002, 90, 647.
20. M. L. Schlegel, P. Reiller, F. Mercier-Bion, N. Barré and V. Moulin, Geochim. Cosmochim. Acta, 2006, 70, 5536.
21. Y. S. Shimamoto and Y. Takahashi, Anal. Sci., 2008, 24, 405.
22. B. W. Lucas, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1984, 40, 1989; E. Alici, T. Schmidt and H. Lutz, Z. Anorg. Allg. Chem., 1992, 608, 135; N. I. Sorokina, V. V. Ilyukhin, V. I. Pakhomov and N. V. Belov, Dokl. Akad. Nauk SSSR, 1979, 249, 613.
23. M. Newville, J. Synchrotron Radiat., 2001, 8, 322.
24. NMR Crystallography, ed. R. K. Harris, R. E. Wasylishen and M. J. Duer, Wiley, 2009.

---

Footnotes

† Electronic supplementary information (ESI) available: Characterisations of SrI–HA by XRD and IR, I K-edge XAS of CaI–HA, I-K-edge EXAFS data of Na(IO₃) and SrI–HA, local environment of the iodine in the model compounds, all DFT models of iodate-substituted Ca–HA phases, and their relative energies. See DOI: 10.1039/c3ra47691j

‡ Although the CaI–HA phase synthesised experimentally is hexagonal (with a disorder in the orientation of the OH groups along the columns), a monoclinic structure was used as a starting point for the simulations. Indeed, the purpose of the simulations was solely to study the effect of the incorporation of the iodate along the OH columns, and not to look at the exact orientation of the OH groups from one column to the other. The advantage of the larger monoclinic cell is that by substitution of 1 OH only, iodate contents fairly similar to those obtained experimentally can be obtained, without having to use supercells.

§ Due to the phase shift, the distances after Fourier transform of the EXAFS data appear slightly shorter than the actual distances: for example, I–O distances in iodates are ∼1.8 Å, but they appear at ∼1.4 Å in Fig. 2 and S4.†

¶ Many different back-scattering pathways were considered for each model sample, starting from the experimental distribution of atoms around the iodine (Fig. S5†). The fitted data reported in Table 1 corresponds to the most reasonable set of parameters derived.

|| It is worth noting that we had verified on the Sr(IO₃)₂·H₂O, Ca(IO₃)₂·H₂O and Ca(IO₃)₂ phases that the computational methodology used in our calculations is capable of properly reproducing the local geometry around the iodate (see ESI,† Table S2).

---