Journal of
Materials Chemistry B PAPER
Cite this: J. Mater. Chem. B, 2014, 2, 536
X-ray absorption spectroscopy shining (synchrotron) light onto the insertion of Zn2+ in calcium phosphate ceramics and its influence on their behaviour under biological conditions† Sandrine Gomes,ab Amandeep Kaur,ab Jean-Marie Nedelecab and Guillaume Renaudin*ab The present study gives a fine description of the Zn2+ location in Zn-doped Biphasic Calcium Phosphate (BCP) samples heat treated between 500 C and 1100 C. Structural considerations were used to explain the sample interactions with biological fluid (DMEM). X-ray Absorption Spectroscopy (XAS) experiments were used to characterize the powdered samples. The main results (1) indicate the presence of Zn2+ complexes physisorbed at the HAp surface for a sintering temperature of 500 C, (2) confirm the insertion of Zn2+ into the b-TCP phase using a substitution mechanism for a sintering temperature around 700 C, and (3) fully describe the insertion of Zn2+ into the HAp phase by an interstitial mechanism for heat treatment above 900 C (composition Ca10Znx(PO4)6(OH)22xO2x; xmax 0.25). The formation of the linear O–Zn–O entity with dZn–O ¼ 1.72(2) A ˚ has been clearly evidenced by Fourier transform amplitude fitting in the R-space (Zn two-fold coordination unambiguously described for the first time). The mineralisation stimulatory effect of Zn2+ incorporated into BCP has been explained by two different mechanisms. For samples heat treated between 500 C and 800 C, the stimulatory effect is attributed to the presence of soluble Zn2+ species: Zn2+ physisorbed at the HAp surface for sintering
Received 8th October 2013 Accepted 14th November 2013
treatment around 500 C and Zn2+ incorporated into about 20 wt% (weight percent) of the soluble Zndoped b-TCP phase for sintering treatment around 800 C. A sintering temperature above 900 C led to
DOI: 10.1039/c3tb21397h
the formation of an extremely weakly soluble and well-crystallized Zn-doped HAp phase which acts by
www.rsc.org/MaterialsB
facilitating the nucleation of a calcium phosphate phase at its surface.
1. Introduction The mechanism of Zn2+ insertion in Biphasic Calcium Phosphate (BCP) bioceramics has been recently deeply investigated by our team thanks to the detailed crystallographic investigations.1 Long range order characterisation has demonstrated that this mechanism is temperature dependent.2 Up to 600 C, Zn-doped BCP is mainly composed of undoped hydroxyapatite (HAp, Ca10(PO4)6(OH)2) and a part of the introduced Zn2+ cations is not detectable by the X-ray diffraction technique (either physisorbed at the HAp surface as already mentioned3,4 or incorporated into an amorphous compound). For temperatures around 800 C, Zn-doped BCP is still composed of undoped HAp (slightly Zn-doped HAp more precisely) together with
a Clermont Universit´e, ENSCCF, Institut de Chimie de Clermont-Ferrand, BP 10448, F-63000 Clermont-Ferrand, France. E-mail:
[email protected]; Fax: +33 4 73 40 70 95; Tel: +33 4 73 40 73 36 b
CNRS, UMR 6296, ICCF, F-63171 Aubi`ere, France
† Electronic supplementary 10.1039/c3tb21397h
information
536 | J. Mater. Chem. B, 2014, 2, 536–545
(ESI)
available.
See
DOI:
Zn-doped b-tricalcium phosphate (b-TCP, b-Ca3(PO4)2). The Zn doping mechanism for b-TCP is realized by substitution on the Ca4 and Ca5 crystallographic sites of the b-TCP structure, i.e. the low-density column.5 Finally, above 900 C Zn-doped BCP is mainly composed of the Zn-doped HAp phase with composition Ca10Znx(PO4)6(OH)22xO2x (with xmax 0.25). Starting from 1000 C, the Zn2+ cations are inserted into the HAp structure, thanks to an interstitial mechanism on the 2b Wyckoff site of the HAp structure.6 In spite of the extremely similar ionic size, the mechanism of incorporation of Zn2+ cations into the HAp structure is fundamentally different from the one of Mg2+.7 To realize the 2-fold coordination with the formation of linear O– Zn–O entities2 (a Zn2+ cation located in the 2b Wyckoff site is surrounded by two close hydroxyl crystallographic sites along the hexagonal channel), 3d atomic orbitals of the transition metal cation are certainly needed. One of the objectives of the present study is to conrm the formation of this O–Zn–O entity described for the rst time in Zn-doped HAp.1 X-ray Absorption Spectroscopy (XAS) at the Zn K-edge was used to investigate the Zn2+ local environment (coordination number and interatomic distances) in Zn-doped BCP samples containing a low amount
This journal is © The Royal Society of Chemistry 2014
Paper
Journal of Materials Chemistry B
of doping element. An amount of about 1 weight percent (wt%) of Zn in the whole sample is inevitably a limiting factor for the previously realized Rietveld analyses. Although joint Rietveld renements have been performed on synchrotron diffraction data and neutron diffraction data, the exact characterization of the local environment of Zn2+ is not allowed by long range order description with statistical disorder around the Zn atomic site. The statistical disorder between one O–Zn–O entity and two hydroxyl groups prevents the exact local structure description by long range order consideration (similar to the silicate/phosphate disorder in silicon containing HAp samples also characterized by X-ray and neutron joint Rietveld analyses).8,9 XANES/EXAFS studies have already been performed on Zn-doped apatitic materials in the last ve years with divergent conclusions: (1) Zn atoms are localised at the surface and not in the structure of pathological apatite,3,4 (2) Zn2+ cations substitute Ca2+ into the HAp structure either in the Ca1 site10 or in the Ca2 site,11 and (3) tetrahedral coordinated Zn2+ cations are located in the Ca2 site.12 It is now important to reinvestigate these systems thanks to the XAS technique in the light of the known temperature dependent mechanism of Zn2+ incorporation into BCP. Aerwards the obtained results (i.e. the exact description of the temperature dependent Zn2+ local environment) will be used to correctly interpret the data from the biological uid/Zn-doped BCP interaction study, which is the main issue of the present study. The important role of zinc in biological applications has been put forward in the literature.13–15 Uptake and release of Zn in the body are strongly mediated by the bone reservoir.15 When incorporated into bioceramics, it has been demonstrated that zinc has a stimulatory effect on bone formation and mineralization in vivo and in vitro,16,17 and that Zn incorporation into implants promotes bone formation around the material,18–20 improves biological properties,18,21 decreases the inammatory response22,23 and has an antibacterial effect.24 In order to explore in detail the effect of the Zn doping element in promoting bone formation, two series of samples with nominal composition Ca10Znx(PO4)6(OH)22xO2x (x ¼ 0.15 and 0.20) have been synthesized at sintering temperatures between 500 C and 1100 C (as well as an undoped series, with x ¼ 0.00, for comparison). The interactions with biological uid (DMEM) have been investigated. Finally, X-ray absorption spectroscopy (XANES and EXAFS; SAMBA beam line at SOLEIL synchrotron) has been used to correctly interpret the behaviour of the materials in DMEM.
2. 2.1
Materials and methods Sol–gel elaboration of Zn-substituted BCP samples
The sol–gel route previously proposed by the authors was used to synthesize one undoped and two Zn-doped series of BCP samples.2 Briey, to produce 2 g of undoped BCP powder, 4.7 g of Ca(NO3)2$4H2O (Aldrich) and 0.84 g of P2O5 (Avocado Research chemicals) were dissolved in ethanol under stirring and reuxed at 85 C for 24 h. Then, this solution was kept at 55 C for 24 h to obtain a white consistent gel. The as-prepared gel was further oven-dried at 80 C for 10 h to obtain a white powder. The last step consists of the high temperature heat treatment: pellets of crushed white powder were heat treated for
This journal is © The Royal Society of Chemistry 2014
15 h at temperatures between 500 C and 1100 C (7 sintering temperatures – leading to seven samples per series of composition – were used: 500 C, 600 C, 700 C, 800 C, 900 C, 1000 C and 1100 C). To prepare Zn-substituted samples, the required amount of Zn(NO3)2$6H2O (Acros Organics) was added to the solution, simultaneously with Ca(NO3)2$4H2O (Aldrich). Nominal compositions have been calculated assuming the insertion of Zn2+ cations into an interstitial crystallographic site, i.e., Ca/P ¼ 1.67 (whatever the Zn added amount). In the following, samples are labelled ‘xZn-T’ with x ¼ 00, 15 and 20 for samples with the nominal compositions Ca10(PO4)6(OH)2, Ca10Zn0.15(PO4)6(OH)1.70O0.30 and Ca10Zn0.20(PO4)6(OH)1.60O0.40, respectively. T refers to the sintering temperature: T ¼ 500 C, 600 C, 700 C, 800 C, 900 C, 1000 C and 1100 C. Elemental analyses of the samples by ICP-AES have conrmed the targeted nominal compositions. 2.2
X-ray powder diffraction (XRPD) and Rietveld analyses
XRPD patterns were recorded on an X'Pert Pro Philips diffractometer, with q–q geometry, equipped with a solid detector X˚ XRPD Celerator and using Cu Ka radiation (l ¼ 1.54184 A). patterns were recorded at room temperature in the interval 3 < 2q < 120 , with a step size of D2q ¼ 0.0167 and a counting time of 200 s for each data value. A total counting time of about 200 min was used for each sample. An XRPD pattern was collected from the pure LaB6 NIST standard under the same experimental conditions in order to extract the instrumental resolution function to improve the peak prole tting during Rietveld renements. Rietveld renements of X-ray powder patterns were performed for each sample with the program FullProf.2k.25 The procedure used (both data-collection and renement strategy) corresponds to the general guidelines for structure renement using the Rietveld (whole-prole) method formulated by the International Union of Crystallography Commission on Powder Diffraction.26–28 The Rietveld strategy has been detailed in previous related studies.1,2 2.3
X-ray absorption spectroscopy (XAS)
Zn K-edge Extended X-ray Absorption Fine Structure (EXAFS) spectra, simultaneously with the X-ray Absorption Near Edge Structure (XANES) part of the spectra, were collected on the Zndoped samples and two reference compounds (ZnO with Zn2+ tetrahedral coordination, and a ZnAl–CO3 layered double hydroxide phase with Zn2+ octahedral coordination) at the SAMBA beam line at the SOLEIL synchrotron (Saint Aubin, France) in order to determine the electronic state as well as to describe accurately the coordination spheres of Zn atoms. The SOLEIL synchrotron was running at 2.75 GeV with an average current of 400 mA. The X-ray beam was obtained with a two crystal Si(111) monochromator which offers an energy resolution of DE/E ¼ 104 necessary to resolve the XANES structure. Harmonics were rejected from the beam thanks to the use of two palladium-coated mirrors tilted at 4.5 mrad. Calibrations of the experiments were made with Zn metallic reference foil (Kedge 9659 eV). Experiments were performed at the liquid nitrogen temperature (77 K) and atmospheric pressure. Spectra
J. Mater. Chem. B, 2014, 2, 536–545 | 537
Journal of Materials Chemistry B
were collected in the energy range between 9400 and 10 650 eV, with the energy step varying from 0.3 eV (XANES part) to 2.0 eV (end of the EXAFS part) and 3 s dwell time per point. Spectra were obtained three times running in direct absorption mode and six times running in uorescence mode using a SDD detector.29 The size of the beam was determined by a set of slits (200 mm 500 mm). Data treatments have been performed by using the programs Athena and Artemis from IFFEFIT package soware30 by merging the successively recorded absorption/uorescence spectra. Single scattering theory has been used here. Following Lengeler–Eisenberg normalization, EXAFS oscillations were Fourier Transformed (FT) using a Hanning window between 3.0 ˚ 1. The c(k) function was Fourier transformed using and 13.0 A 3 k weighting, and all shell-by-shell tting was done in R-space. Theoretical backscattering paths were calculated successively using ATOMS31 and FEFF6.32 2.4 Interaction of powdered samples with the biological uid (DMEM) The samples synthesized at T ¼ 500 C, 700 C, 900 C and 1100 C were immersed at 37 C for 1, 2, 6, 10 and 20 days in standard Dulbecco's Modied Eagle's Medium (DMEM; VWR International SAS, France); pH 7.3 (DMEM without D-glucose and without L-glutamine, Biochrome AG Cat. no. F0405). DMEM contained the following mineral ingredients: NaCl (6400 mg L1), KCl (400 mg L1), CaCl2 (200 mg L1), MgSO4$7H2O (200 mg L1), Fe(NO3)3$9H2O (0.1 mg L1), MgSO4$7H2O (200 mg L1) and NaHCO3 (3700 mg L1). DMEM has been considered rather than simulated body uid because it matches more closely the biological conditions in particular due to the presence of amino acids. Aer determining the Specic Surface Area (SSA) of the samples, thanks to nitrogen sorption isotherms (on a Quantachrom Autosorb-1b instrument), the volume of DMEM has been calculated for 10 mg of powder and a constant liquid/ SSA ratio according to the following equation:
Paper
composed of the HAp phase. The b-TCP phase is stabilized for the intermediate temperature with the maximum amount observed at 800 C. Fig. 1 illustrates these results for the two main phases HAp and b-TCP. Combined with the two expected HAp and b-TCP phases, a-CDP (composition Ca2P2O7), CaCO3, CaO and ZnO are temperature dependent impurities. a-CDP and CaCO3 (which transforms into CaO on heating) are present up to 700 C and ZnO is observed up to 900 C. As deeply described in a previous study,2 Zn2+ cations are inserted in the interstitial 2b Wyckoff site into the HAp structure above 900 C only. Structural parameters of the HAp phase (Table 1) evidence the late interstitial mechanism of incorporation of Zn, contrary to the substitution mechanism observed for b-TCP (Table S2, ESI†) at intermediate temperatures. The HAp lattice parameters show anisotropic variation between 900 C and 1100 C (contraction of the a lattice parameter and expansion along the c lattice parameter) combined with an increase of the Zn occupancy in the interstitial 2b Wyckoff site to reach the targeted compositions: Ca10Zn0.16(1)(PO4)6(OH)1.68(2)O0.32(2) and Ca10Zn0.23(1)(PO4)6(OH)1.54(2)O0.46(2) for the 15Zn-1100 and 20Zn-1100 samples, respectively. The variations of lattice parameters for b-TCP agree with a substitution mechanism (simultaneous decrease of both a and c lattice parameters due to the calcium substitution for a smaller Zn2+ cation) which takes place along the low density column of the structure. According to our previous studies1,2 three kinds of samples are distinguishable in terms of Zn location: - Samples heat treated at 500 C and 600 C with about one third of Zn2+ contained in ZnO and the other two thirds not detectable by XRPD (physisorbed at the HAp surface or included in an amorphous compound), - Samples heat treated between 700 C and 900 C with one part of Zn2+ contained in ZnO (still about one third) and the other main part incorporated into the b-TCP structure, - Samples heat treated at 1000 C and 1100 C with the whole 2+ Zn cations inserted into the HAp structure.
vol. of DMEM (mL) ¼ 5 SSA (m2 g1)
At each immersion time, the concentration of the three elements Ca, P and Zn was determined by ICP-AES in the biological uid aer elimination of the solid using a 0.2 mm lter. The SSA is temperature dependent as densication of the materials occurs on heating. The values are about 15 m2 g1, 10 m2 g1, 4 m2 g1 and 2 m2 g1 for sintering temperatures of 500 C, 700 C, 900 C and 1100 C, respectively.
3.
Results and discussion
3.1
Mineralogical analysis of the samples
To correctly interpret the behaviour of our samples, the mineral compositions have been extracted from the Rietveld treatment performed on laboratory X-ray diffraction patterns. Mineral compositions of the three series of samples are shown in Table S1 (ESI†). In agreement with previous results2 samples heat treated at 500 C and between 900 C and 1100 C are mainly
538 | J. Mater. Chem. B, 2014, 2, 536–545
Fig. 1 Mineral composition of the two main phases (HAp and b-TCP) as a function of the sintering temperature for the three series undoped (black squares and dotted lines), 15Zn-T (blue circles and solid lines) and 20Zn-T (red diamonds and solid lines). Lines are guides for eyes.
This journal is © The Royal Society of Chemistry 2014
Paper Table 1
Journal of Materials Chemistry B Structural parameters of the HAp phase obtained by Rietveld refinements HAp structural parameters
Samples
˚) a (A
˚) c (A
˚ 3) V (A
Zn Occa
Rened composition
00Zn-500 00Zn-600 00Zn-700 00Zn-800 00Zn-900 00Zn-1000 00Zn-1100
9.4196(1) 9.4197(1) 9.41874(9) 9.41954(5) 9.42007(3) 9.42025(2) 9.42037(2)
6.88419(10) 6.88410(9) 6.88356(8) 6.88241(4) 6.88100(3) 6.88127(2) 6.88176(2)
528.94(1) 528.97(1) 528.846(9) 528.848(5) 528.799(3) 528.840(2) 528.891(2)
— — — — — — —
Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2
15Zn-500 15Zn-600 15Zn-700 15Zn-800 15Zn-900 15Zn-1000 15Zn-1100
9.4228(1) 9.42044(9) 9.42057(8) 9.41927(5) 9.41859(3) 9.41766(2) 9.41740(2)
6.88155(9) 6.88233(8) 6.88365(7) 6.88273(5) 6.88622(2) 6.88989(2) 6.89068(2)
529.14(1) 528.943(9) 529.058(8) 528.842(6) 529.034(3) 529.211(2) 529.242(2)
— — 0.5(3) 0.4(3) 3.2(3) 5.4(3) 7.9(3)
Ca10(PO4)6(OH)2 Ca10(PO4)6(OH)2 Ca10Zn0.01(1)(PO4)6(OH)1.98(2)O0.02(2) Ca10Zn0.01(1)(PO4)6(OH)1.98(2)O0.02(2) Ca10Zn0.06(1)(PO4)6(OH)1.88(2)O0.12(2) Ca10Zn0.11(1)(PO4)6(OH)1.78(2)O0.22(2) Ca10Zn0.16(1)(PO4)6(OH)1.68(2)O0.32(2)
20Zn-500 20Zn-600 20Zn-700 20Zn-800 20Zn-900 20Zn-1000 20Zn-1100
9.4230(1) 9.4194(1) 9.41829(8) 9.41858(5) 9.41933(3) 9.41892(2) 9.41704(2)
6.8814(1) 6.88251(9) 6.88265(7) 6.88280(4) 6.88786(2) 6.89540(2) 6.89485(2)
529.16(1) 528.84(1) 528.726(9) 528.770(5) 529.242(3) 529.776(2) 529.522(2)
2.8(3) 2.2(3) 1.2(3) 2.0(3) 5.0(3) 10.4(3) 11.5(3)
Ca10Zn0.06(1)(PO4)6(OH)1.88(2)O0.12(2) Ca10Zn0.04(1)(PO4)6(OH)1.92(2)O0.08(2) Ca10Zn0.03(1)(PO4)6(OH)1.94(2)O0.06(2) Ca10Zn0.04(1)(PO4)6(OH)1.92(2)O0.08(2) Ca10Zn0.10(1)(PO4)6(OH)1.80(2)O0.20(2) Ca10Zn0.21(1)(PO4)6(OH)1.58(2)O0.42(2) Ca10Zn0.23(1)(PO4)6(OH)1.54(2)O0.46(2)
a
Zn occupancy in the 2b Wyckoff site.
During Rietveld renements the occupancy factors of all calcium and phosphorus crystallographic sites have been systematically tested. Whatever the temperature and the Zndoping level, no vacancies have been observed contrary to the proposed Ca-decient chemical composition for Zn-doped HAp.11 For both b-TCP and HAp phases the occupancies did not deviate signicantly from unity (except the half-lled Ca4 site in b-TCP). Aer testing, all occupancies were xed to unity (and Ca4 occupancy xed to 1/2 for b-TCP). As already mentioned in our previous studies1,2 no calcium substitution has been characterized in the HAp structure, contrary to the b-TCP case and contrary to the conclusion found in the literature.10–12 On the other hand, an XRPD invisible part of Zn was determined for doped samples heat treated between 500 C and 700 C, in agreement with indication of Zn2+ physisorbed at the HAp surface.3,4 During the last runs of Rietveld renement, Zn substitution in a-CDP (i.e. x renement in Ca2xZnxP2O7 solid solution) was considered. A maximal x value around 0.1 was observed (correlated with a weak decrease of the rened lattice parameters; decrease of 0.4% of the unit cell volume compared with published undoped a-CDP data from ICSD #22225). Such an amount of Zn2+ substituted in a-CDP did not allow us to explain the two thirds of Zn not detectable by long range order investigation. On the other hand, the XRPD invisible part of Zn was not correlated with the a-CDP weight amount and the Zn2P2O7 phase (not isostructural to Ca2P2O7) was not observed (the three a-, b-, and g-Zn2P2O7 polymorphs – corresponding to the ICSD #18315, ICSD #24153 and ICSD #51095 data, respectively – have been tested). This journal is © The Royal Society of Chemistry 2014
3.2
XAS analyses of the 20Zn-T series
The seven samples of the 20Zn-T series have been investigated thanks to XAS spectra recorded in transmission mode. This series has been chosen because of the single phase character of the two last 20Zn-1000 and 20Zn-1100 samples (Table S1† and Fig. 1). For the other 15Zn-T series, only the two 15Zn500 and 15Zn1100 samples have been measured. A set of two reference materials have been used: ZnO and ZnAl–LDH (a layered double hydroxide with composition Zn2Al(OH)6(CO3)1/2$2H2O). Zn2+ presents a tetrahedral coordination in zincite ZnO with four ˚ 33 whereas it presents an octahedral distances around 1.97 A, ˚ 34 coordination in ZnAl–LDH with six distances around 2.00 A. 3.2.1 Temperature dependence of the spectra. Fig. 2 and 3 show the XANES part of the spectra and the EXAFS modulations for the 20Zn-T series and for the two reference compounds, respectively. The temperature dependence of the spectra agrees with previous long range order characterizations.1,2 Modications appear mainly between 600 C and 700 C (the temperature at the beginning of the Zn substitution into the b-TCP structure), and around 900 C (the temperature at the beginning of the Zn insertion into the HAp structure). In agreement with PXRD analysis which indicates one third of Zn2+ contained in crystallized ZnO (Table S1†) with tetrahedral coordination, XANES spectra and EXAFS modulations from 20Zn-500 and 20Zn-600 samples are extremely similar to those from ZnO as shown by arrows ‘1’ in Fig. 2 and 3. The Zn2+ part not detectable by long range order investigation is then characterized in XANES spectra by the rst le shoulder at 9664.5 eV
J. Mater. Chem. B, 2014, 2, 536–545 | 539
Journal of Materials Chemistry B
Fig. 2 Normalized XANES spectra at the Zn K edge (E0 ¼ 9662 eV) for samples from the 20Zn-T series (solid lines) and the two reference compounds (ZnO and ZnAl–LDH, dashed lines). Arrows labelled ‘1’, ‘?’ and ‘3’ refer to signals from Zn in ZnO, not detectable by XRPD and inserted into HAp, respectively.
The k-weighted EXAFS signal, c(k), extracted from the absorption measurements from the 20Zn-T series (solid lines) and the two reference compounds (ZnO and ZnAl–LDH, dashed lines). Arrows labelled ‘1’, ‘?’ and ‘3’ refer to signals from Zn in ZnO, not detectable by XRPD and inserted into HAp, respectively.
Fig. 3
(arrow with label ‘?’ in Fig. 2), and in EXAFS modulation by shoulders (also identied by label ‘?’ in Fig. 3) corresponding to the modulations of octahedral Zn2+ in ZnAl–LDH. This indicates that the not detectable part of Zn2+ is certainly in octahedral coordination (physisorbed or in an amorphous compound). Spectra of the 20Zn-800 sample are relatively similar to those of the formers, nevertheless we can observe the formation of a new rst le shoulder at 9662.5 eV (arrow ‘3’ in Fig. 2). EXAFS modulations are quasi invariant up to a sintering temperature of 800 C. We can only observe that the shoulder identied by arrow ‘3’ in Fig. 3 is more pronounced. Spectra present important modications when Zn2+ is inserted into the HAp structure (i.e. above 900 C). XANES spectra for samples 20Zn-900, 20Zn-1000 and 20Zn-1100 present a rst sharp white line at 9662.5 eV (arrow ‘3’ in Fig. 2) followed by supplementary peaks at 9669 eV and 9677 eV. This remarkable sharp white line
540 | J. Mater. Chem. B, 2014, 2, 536–545
Paper
Fig. 4 The k-weighted amplitude of the Fourier transform uncorrected for the phase shift for samples from the 20Zn-T series (solid lines) and the two reference compounds (ZnO and ZnAl–LDH, dashed lines). Arrows labelled ‘1’, ‘2’ and ‘3’ refer to signals from Zn in ZnO, substituted in b-TCP and inserted into HAp, respectively.
traduces a new symmetry for Zn atoms without modication of the effective charge (as E0 is still close to 9662 eV, characteristic of Zn2+). EXAFS modulations are also highly modied for temperatures above 900 C, with apparition of the main ˚ 1 (arrow ‘3’ in Fig. 3). In Fig. 2 and 3 modulation at k ¼ 4.06 A arrows with labels ‘1’, ‘?’ and ‘3’ refer to signals from Zn atoms in ZnO, not detectable by XRPD, and inserted into HAp, respectively. In Fig. 4, the arrow with label ‘2’ refers to the Zn atom substituted in the b-TCP structure. Fourier transformed amplitudes (not corrected for phase shi) are represented in Fig. 4 in the R-space. Radial distribution from the 20Zn-500 sample is similar to that from ZnO with ˚ (arrow ‘1’) and the second peak at R ¼ the rst peak at R ¼ 1.52 A ˚ Nevertheless, the intensity ratio between these two rst 2.90 A. peaks for 20Zn-500 did not correspond to the one of ZnO. The second contribution is less intense in 20Zn-500, indicating that a part of Zn2+ in the sample has only one oxygen shell (without second neighbours). This observation indicates that the Zn2+ part not detectable by XRPD is more probably physisorbed at the HAp surface, and not incorporated into an amorphous compound. According to the previous observation on the XANES part and EXAFS modulations it seems that Zn2+ forms an octahedral complex physisorbed at the HAp surface. Already for sample 20Zn-600, the radial distribution presents important modications, namely the appearance of a distribution at R ¼ ˚ (arrow ‘3’ in Fig. 4). This short contribution becomes the 1.20 A only rst shell for samples heat treated at 1000 C and 1100 C ˚ i.e. the signal (and the second shell is observed at R ¼ 2.58 A), corresponding to Zn2+ inserted into the HAp structure with short Zn–O distances. This indicates clearly that Zn2+ incorporation into HAp is already weakly realized at 600 C in agree˚ (arrow ment with a previous study.2 The contribution at 1.52 A ‘1’ in Fig. 4) is shied toward higher interatomic distance for the intermediate sintering temperature (close to the value observed for octahedral Zn2+ in ZnAl–LDH), indicating an increase of the coordination number in agreement with the substitution in Ca4 (unusual 3 + 3 coordination) and Ca5
This journal is © The Royal Society of Chemistry 2014
Paper
(octahedral coordination) crystallographic sites in b-TCP around 800 C (arrow ‘2’ in Fig. 4). 3.2.2 Zn2+ insertion into HAp. Samples 20Zn-1000 and 20Zn-1100 are single phase, composed of Zn-doped HAp only. Radial distribution functions are ideal data to conrm the local environment of Zn2+ inserted into the interstitial 2b site of the HAp structure. The k3-weighted tted Zn K-edge EXAFS data of 20Zn-1100 are shown in Fig. 5 and t results are gathered in Table 2. Fit has been performed with the Artemis program in ˚ < R < 4 A, ˚ not corrected for phase shis. Eight the range 1 A direct paths have been considered in this R range. Six multiple paths (minor contributions to the Fourier transform amplitudes, not indicated in Table 2) have been added to slightly improve the t. The number of neighbours for each shell has been checked in the rst step, and they were xed to their crystallographic values (to minimise the number of rened parameters) as the deviations were small. The obtained results unambiguously conrm the insertion of Zn2+ cations into the 2b Wyckoff site of the HAp structure.1,2 The rst amplitude observed for a short R distance agrees with the O–Zn–O linear entity. A comparison with reference materials is available in ˚ for Zn-doped HAp (two-fold Fig. 4: see arrow ‘3’ at R ¼ 1.21 A ˚ for ZnO (four-fold coordicoordinated), arrow ‘1’ at R ¼ 1.52 A ˚ for ZnAl–LDH (six-fold nated) and arrow ‘2’ at R ¼ 1.66 A coordinated). ˚ (Table 2). This The rened Zn–O distance is 1.72(2) A distance is comparable to the interatomic distance dNi–O ¼ ˚ for the equivalent O–Ni–O linear entity in K2NiO2.35 1.68 A ˚ only Rietveld analyses indicate a Zn–O distance about 1.42 A ˚ (indicated by ‘shi' in Table 2) (Table 2). The difference of 0.30 A between XRPD long range order characterization and local structure EXAFS analysis is due to statistical disorder around the 2a Wyckoff site along the hexagonal channel in the Zndoped HAp structure (a unique crystallographic site, O4,
Fig. 5 (Left) k3-weighted amplitude of the Fourier transform uncorrected for phase shift (black line) and best fit in R-space (dotted red line) for samples 20Zn-1100 (independent points: 19.6, variables: 15 and R-factor: 0.022). (Right) Representation of the Zn-local environment; projection along the hexagonal c axis (top) and normal to the hexagonal c axis (bottom). Atom labels correspond to structural description reported by Rodriguez-Lorenzo et al.6 (black central spheres for Zn atoms, large grey spheres for neighbour O4 oxygen atoms belonging to the linear O–Zn–O entity, blue large spheres for Ca2 calcium atoms and pink tetrahedra for phosphate groups).
This journal is © The Royal Society of Chemistry 2014
Journal of Materials Chemistry B Table 2 Comparison of the Zn-local environment determined by XRPD Rietveld analysis and fit of the k3-weighted EXAFS raw data for 20Zn-1100. Atom labels correspond to structural description reported by Rodriguez-Lorenzo et al.6
Zn shells
EXAFS t
XRPD Rietveld
Shi
Zn–X
˚ Biso(X) (A ˚ 2) dZn–X (A) ˚ s 2 (A ˚ 2) CN dZn–X (A)
Zn–O4 Zn–O3(*) Zn–Ca2 Zn–P1(*) Zn–O3(*) Zn–O1(*) Zn–O3 Zn–O2(*)
2 6 6 6 6 6 6 6
1.42 2.92 2.92 4.01 4.13 4.38 4.87 4.90
¼ Biso(O1) ¼ Biso(O1) 0.83(2) 0.54(4) ¼ Biso(O1) 0.80(4) ¼ Biso(O1) ¼ Biso(O1)
1.72(2) 2.69(5) 2.97(2) 3.5(1) 3.3(2) 4.0(2) 4.79(6) 4.56(9)
0.001(2) ¼ s2 (P1) 0.007(3) 0.015(8) ¼ s2 (P1) ¼ s2 (P1) 0.006(9) ¼ s2 (P1)
˚ D(dZn–X) (A) +0.30 0.23(*) +0.05 0.5(*) 0.8(*) 0.4(*) 0.08 0.34(*)
(*)
: atoms belonging to the 6 neighbouring phosphate groups.
designs either the hydroxyl groups or oxygen atoms from O–Zn– O entities). EXAFS results allows correct characterization of this ˚ The tted distance of the six calcium distance dZn–O ¼ 1.72(2) A. cations (from Ca2 site) agrees with Rietveld renement (a shi ˚ is observed). On the other hand, tted EXAFS data of only 0.05 A clearly indicate that the six neighbouring phosphate groups approach the inserted Zn2+ cations. The tted distances of the ve atoms forming the PO43 anions are shied between 0.23 ˚ toward Zn2+, by comparison with long range order and 0.8 A, results from Rietveld analysis (see asterisks in Table 2). Fitted Debye–Waller factors (Table 2) are in good agreement with the mechanism of Zn insertion into HAp. The small s2 value for O4 corresponds to the formation of rigid O–Zn–O entities, whereas large s2 values for atoms belonging to phosphate groups are explained by Zn2+ attraction which led to local disorder. Neighbouring Ca2+ cations are not affected by the Zn2+ insertion: they are not shied from their position in the undoped HAp. 3.2.3 Zn2+ localisation at 500 C. Samples heat treated at 500 C (and also 600 C) present two kinds of Zn2+ cations. Rietveld results indicate that about one third of Zn2+ is contained in the ZnO phase. The two other thirds are not detectable by XRPD, either physisorbed at the HAp phase or incorporated into an amorphous phase. These two possibilities should have different EXAFS signatures. The formation of a Zn2+ complex physisorbed at the HAp surface will generate one rst shell of ˚ In the case oxygen atoms only (at a distance dZn–O of about 2.0 A). of Zn2+ contained in an amorphous compound, an equivalent rst oxygen shell should be observed, but almost a second shell (formed by cations Ca2+ or Zn2+) should also be present in the radial distribution (Fourier transform amplitude of EXAFS data). The superposition of Fourier transform amplitudes of the ZnO reference material, 15Zn-500 and 20Zn-500 samples (Fig. 6, top le), evidences that Zn-doped samples heat treated at 500 C ˚ and R present the two radial contributions from ZnO (R ¼ 1.55 A ˚ but relative intensities are not respected. The second ¼ 2.90 A) ˚ relative to the second Zn-shell, contribution close to R ¼ 2.90 A, is clearly less intense for Zn-doped samples (about half intensity). This indicates that about half Zn2+ cations are contained in the ZnO phase. The second half Zn2+ cations present the rst ˚ only, corresponding to the oxygen shell (with R close to 1.55 A)
J. Mater. Chem. B, 2014, 2, 536–545 | 541
Journal of Materials Chemistry B
Paper
(Top left) k3-weighted amplitude of the Fourier transform uncorrected for phase shift for the ZnO reference material (dashed line), 15Zn500 sample (bold line) and 20Zn-500 sample (black line). (Top right) Plot of the best fit in R-space (dotted red line; independent points: 17.0, variables: 9 and R-factor: 0.014) for the ZnO reference material (black line). (Bottom left) Plot of the best fit in R-space (dotted red line; independent points: 13.0, variables: 4 and R-factor: 0.021) for 15Zn-500 (black line). (Bottom right) Plot of the best fit in R-space (dotted red line; independent points: 13.0, variables: 3 and R-factor: 0.045) for 20Zn-500 (black line). Fig. 6
Results of the k3-weighted EXAFS raw data fit for ZnO, 15Zn-500 and 20Zn-500, x indicates the fitted proportion of Zn2+ from the physisorbed hydroxyl–aqua complex Table 3
Zn shells
ZnO
15Zn-500
20Zn-500
Neighbour Zn–X
CN
˚) dZn–X (A
˚ 2) s 2 (A
˚) dZn–X (A
˚ 2) s 2 (A
x
˚) dZn–X (A
˚ 2) s 2 (A
x
Zn–O Zn–Zn ‘Zn–O complex’
4 12 4 6
1.97(1) 3.23(1) — —
0.001(1) 0.005(1) — —
1.97(—) 3.23(—) 1.94(2) 1.94(2)
0.001(—) 0.005(—) 0.001(—) 0.001(—)
1x 1x 0.47(6) 0.37(6)
1.97(—) 3.23(—) 1.84(—) 1.84(—)
0.001(—) 0.005(—) 0.001(—) 0.001(—)
1x 1x 0.4(—) 0.4(—)
physisorbed Zn2+ complex. In the rst step, the Fourier transform amplitude of the ZnO reference material was tted in the R-space in order to extract tted parameters for the two rst shells of ZnO (Fig. 6, top right). The obtained tted parameters for ZnO were used, as xed values, for tting the Fourier transform amplitude of both Zn-doped samples heat treated at 500 C (Fig. 6, bottom). Physisorbed Zn2+ cations were also considered during the tting procedure by considering a unique rst shell. Only two parameters were then tted: the Zn– O interatomic distance for the physisorbed Zn2+ complex and the relative proportion of Zn2+ from ZnO and from the physisorbed Zn2+ complex. Fitting results indicate about 60% of Zn atoms contained in ZnO and 40% of Zn atoms physisorbed at the HAp surface in both samples heat treated at 500 C (Table 3). The large R values (see the Fig. 6 caption) illustrate that these renements brought rough, but highly interesting,
542 | J. Mater. Chem. B, 2014, 2, 536–545
quantitative estimation of the two Zn2+ populations. Fitting did not allow us to discriminate the four-fold or six-fold coordination for physisorbed Zn2+ cations. Nevertheless, previous observation on XANES spectra favours an octahedral coordination. The rst radial contribution for 20Zn-500 is shied toward low R values indicating certainly that a small amount of Zn2+ cations are already inserted into HAp. In agreement with Rietveld results,2 the Fourier transform amplitude of samples heat treated at 600 C already evidences the insertion of Zn2+ into ˚ (Fig. 4) for 20Zn-600. HAp: see minor R contribution at 1.21 A
3.3
DMEM interactions
Undoped HAp series as well as the two Zn-doped HAp series were used to follow the interaction with DMEM for 20 days. Samples heat treated at 500, 700, 900 and 1100 C were
This journal is © The Royal Society of Chemistry 2014
Paper
Journal of Materials Chemistry B
Fig. 7 Zn concentration in DMEM determined by ICP-AES as a function of interaction time for the 15Zn-T series (left) and the 20Zn-T series
(right).
considered to evidence Zn2+ leaching in DMEM from Zn-doped powder and calcium phosphate precipitation (decrease of the Ca and P concentrations in DMEM). 3.3.1 Zn-doped BCP interaction with DMEM. Before interpreting the role of the Zn2+ cation in the rate of calcium phosphate precipitation in DMEM, in the presence of Zn-doped BCP powder, we have to determine the Zn2+ leaching from powder. Fig. 7 shows the evolution of the Zn2+ concentration in DMEM according to the interaction time and heat treatment. Fig. 7 evidences that the different locations of Zn2+ in the powder led to radically different rates of Zn2+ leaching. For samples heat treated at 500 C, 700 C and 900 C, Zn2+ cations are present in DMEM since the rst days of the interaction. The Zn2+ concentration observed with 15Zn-500 certainly corresponds to the part of Zn atoms physisorbed at the HAp surface. This concentration increases for the 15Zn-700 sample in agreement with the large part of Zn2+ cations which substitutes the highly soluble b-TCP phase (Table S1†). From 900 C, the amount of Zn2+ in DMEM decreases considerably to reach a nearly zero value for 15Zn-1100. The insertion of Zn2+ into the interstitial site of the HAp structure makes its dissolution practically impossible. When considering the 20Zn-T series, the same behaviours are observed for the rst days of interaction with
Fig. 8 Ca (spheres) and P (squares) concentrations in DMEM determined by ICP-AES as a function of interaction time for the undoped series heat treated at 500 C (black), 700 C (green), 900 C (blue) and 1100 C (red).
This journal is © The Royal Society of Chemistry 2014
higher Zn2+ concentrations. Aer several days of interaction, the two series behave differently. Large amounts of Zn2+ in DMEM with the 20Zn-T series decrease, indicating that a sufficient concentration of Zn2+ is needed in biological uid to improve the precipitation of calcium phosphate (i.e. to improve mineralisation). The two samples heat treated at 900 C present a similar behaviour. The Zn2+ concentration in DMEM decreases from day 10 for 15Zn-900 and 20Zn-900. The presence of Zndoped HAp (already formed in large quantity at 900 C) seems to stimulate the mineralisation at the HAp surface by improving nucleation at the HAp/DMEM interface. 3.3.2 Sintering temperature effect on calcium phosphate precipitation. To correctly interpret the role of Zn2+, it is important to characterize before the effect of the sintering temperature in the absence of Zn2+. Fig. 8 shows the evolution of the Ca and P concentrations in DMEM as a function of sintering temperature and interaction time. Surprisingly, no precipitation is observed for the undoped sample heat treated at 500 C. The sample heat treated at 700 C, which contains the soluble b-TCP phase, quickly shows a decrease of both Ca and P elements (i.e. precipitation of calcium phosphate). This precipitation of calcium phosphate is delayed for samples mainly composed of crystalline HAp (i.e. heat treated at 900 C and 1100 C). The formation of a calcium phosphate layer at the surface of crystalline HAp occurs in a signicant manner only at day 20, aer several days of incubation necessary for nucleation. 3.3.3 Zn2+ inuence on calcium phosphate precipitation. Fig. 9 shows the role of Zn2+ in the formation of the calcium phosphate layer for the four investigated sintering temperatures. For samples heat treated at 500 C, a large amount of Zn2+ is needed to stimulate the calcium phosphate precipitation. Actually, no effect was observed with 15Zn-500, whereas a late precipitation of calcium phosphate is observed at day 20 with 20Zn-500. The simultaneous decrease of the Zn2+ concentration (Fig. 7) indicates that the formation of the solid is certainly due to the incorporation of Zn2+ cations. The kinetics of calcium phosphate precipitation observed for samples heat treated at 700 C (Fig. 8) is affected by the presence of Zn2+. Two explanations have to be considered: (1) either the Zn-doped b-TCP phase is less soluble than the undoped b-TCP phase or (2) the large amount of soluble Zn-doped b-TCP (about 20 wt%, Table
J. Mater. Chem. B, 2014, 2, 536–545 | 543
Journal of Materials Chemistry B
Paper
Fig. 9 Ca (spheres) and P (squares) concentrations in DMEM determined by ICP-AES as a function of interaction time for the undoped series (black), the 15Zn-T series (blue) and the 20Zn-T series (red) heat treated at 500 C (top left), 700 C (bottom left), 900 C (top right) and 1100 C (bottom right).
S1†) counterbalances partially the calcium phosphate precipitation. The large amount of Zn2+ in DMEM with samples heat treated at 700 C (Fig. 7) favours the latter explanation. Finally the presence of Zn2+ in samples heat treated at 900 C and 1100 C has no signicant effect. Nevertheless, Ca and P concentrations are slightly smaller in the presence of Zn2+ (namely at day 10 for the samples heat treated at 900 C, and already since the rst days for samples heat treated at 1100 C) indicating a stimulatory effect on mineralisation. Almost no Zn2+ cations are detectable in DMEM (Fig. 7) for samples heat treated at 1100 C, indicating that the stimulatory effect should be attributed to the improvement of the calcium phosphate nucleation sites at the Zn-doped HAp surface.
4. Conclusions The present XAS analyses of different Zn-doped BCP samples denitively conrm the mechanism of Zn2+ incorporation into BCP samples1,2 and highlight the interaction with biological uid (DMEM). XANES (Fig. 2) and EXAFS (Fig. 3 and 4) evidence the successive location of Zn2+ in the samples according to the temperature and mineral composition (Table S1† and Fig. 1). Samples heat treated at 500 C contain a small amount of ZnO phase (about 0.5 wt%) and a part of Zn2+ cations is physisorbed at the HAp surface. The corresponding quantities of physisorbed Zn2+ cations are indeed available to interact with biological uid if sufficient amount has been incorporated (Fig. 9,
544 | J. Mater. Chem. B, 2014, 2, 536–545
top le). From 600 C a small amount of Zn2+ cations is inserted into the interstitial sites of the HAp structure (arrow ‘3’ in Fig. 4), nevertheless the main part of Zn2+ substitutes calcium in the b-TCP phase forming a soluble Zn-doped b-TCP phase. Then, the samples heat treated at 700 C deliver a large amount of Zn2+ cations in biological uid (Fig. 7). The effect on mineralisation has been determined here due to the compensation effect between Zn-doped b-TCP dissolution and calcium phosphate precipitation when considering Ca and P concentrations in DMEM (Fig. 9, bottom le). Finally the t of the EXAFS data from samples heat treated at 1100 C has allowed us to provide a ne description of the local environment of Zn2+ inserted into the HAp structure. The formation of the linear O–Zn–O entity has been unambiguously conrmed, and the Zn–O interatomic ˚ Table 2). The six distance is now determined (dZn–O ¼ 1.72(2) A, phosphate anions, forming the second shell (together with six Ca2+ cations, Fig. 5), are shied toward the central Zn2+ cation ˚ compared to crystallographic long range order of about 0.5 A considerations. Zn2+ inserted into Zn-doped HAp is not available during the interaction with biological uid; nevertheless nucleation of a calcium phosphate layer at the HAp surface is facilitated during the rst days of contact. The positive mineralisation properties known for Zn-doped BCP powder can be explained by two different mechanisms according to the sintering temperature. For samples heat treated between 500 C and 800 C, the presence of the soluble Zn2+ cation (either physisorbed at the HAp surface for samples heated at 500 C, or
This journal is © The Royal Society of Chemistry 2014
Paper
substituted in the soluble Zn-doped b-TCP phase for samples heated around 700 C) allows the precipitation of calcium phosphate containing certainly zinc atoms. For samples heat treated above 900 C, the formation of a well crystalline Zndoped HAp phase facilitates the nucleation of calcium phosphate at the HAp surface.
Acknowledgements This work was supported by ANR under project NANOSHAP (ANR-09-BLAN-0120-03). We acknowledge SOLEIL for provision of synchrotron radiation facilities and we would like to thank Stephanie Belin for assistance in using beamline SAMBA (proposal 20120148). One of the authors, A. Kaur, wants to thank her originating University – School of Chemistry, VIT University, Vellore, India – to have allowed her master thesis in the French ICCF laboratory.
References 1 S. Gomes, J.-M. Nedelec, E. Jallot, D. Sheptyakov and G. Renaudin, Chem. Mater., 2011, 23, 3072–3085. 2 S. Gomes, J.-M. Nedelec and G. Renaudin, Acta Biomater., 2012, 8, 1180–1189. 3 D. Bazin, X. Carpentier, I. Brocheriou, P. Dorfmuller, S. Aubert, C. Chappard, D. Thiaudi` ere, S. Reguer, G. Waychunas, P. Jungers and M. Daudon, Biochimie, 2010, 91, 1294–1300. 4 A. Bigi, E. Foresti, M. Gandol, M. Gazzano and N. Roveri, J. Inorg. Biochem., 1995, 58(1), 49–58. 5 M. Yashima, A. Sakai, T. Kamiyama and A. Hoshikawa, J. Solid State Chem., 2003, 175(2), 272–277. 6 L. M. Rodriguez-Lorenzo, J. N. Hart and K. A. Gross, J. Phys. Chem. B, 2003, 107(33), 8316–8320. 7 S. Gomes, G. Renaudin, E. Jallot and J.-M. Nedelec, ACS Appl. Mater. Interfaces, 2009, 1(2), 505–513. 8 S. Gomes, G. Renaudin, A. Mesbah, E. Jallot, C. Bonhomme, F. Babonneau and J.-M. Nedelec, Acta Biomater., 2010, 6, 3264–3274. 9 S. Gomes, J.-M. Nedelec, E. Jallot, D. Sheptyakov and G. Renaudin, Cryst. Growth Des., 2011, 11, 4017–4026. 10 D. Bazin, X. Carpentier, O. Traxer, D. Thiaudiere, A. Somogyi, S. Reguer, G. Waychunas, P. Jungers and M. Daudon, J. Synchrotron Radiat., 2008, 15, 506–509. 11 K. Matsunaga, H. Murata, T. Mizoguchi and A. Nakahira, Acta Biomater., 2010, 6, 2289–2293. 12 Y. Tang, H. F. Chappell, M. T. Dove, R. J. Reeder and Y. J. Lee, Biomaterials, 2009, 30(15), 2864–2872. 13 E. J. Murray and H. H. Messer, J. Nutr., 1981, 111, 1641–1647.
This journal is © The Royal Society of Chemistry 2014
Journal of Materials Chemistry B
14 D. Bazin, P. Chevallier, G. Matzen, P. Jungers and M. Daudon, Urol. Res., 2007, 35, 179–184. 15 A. Ito, M. Otsuka, H. Kawamura, M. Ikeuchi, H. Ohgushi, Y. Sogo, et al., Curr. Appl. Phys., 2005, 5, 402–406. 16 M. Yamaguchi, H. Oishi and Y. Suketa, Biochem. Pharmacol., 1987, 36(22), 4007–4012. 17 M. Yamaguchi and R. Yamaguchi, Biochem. Pharmacol., 1986, 35(5), 773–777. 18 X. Wang, A. Ito, Y. Sogo, X. Li and A. Oyane, Acta Biomater., 2010, 6, 962–968. 19 A. Ito, H. Kawamura, M. Otsuka, M. Ikeuchi, H. Ohgushi and K. Ishikawa, Mater. Sci. Eng., C, 2002, 22, 21–25. 20 A. Ito, K. Ojima, H. Naito, N. Ichinose and T. Tateishi, J. Biomed. Mater. Res., 2000, 50, 178–183. 21 E. Jallot, J.-M. Nedelec, A. S. Grimault, E. Chassot, A. Grandjean-Laqueriere, P. Laquerriere and D. LaurentMaquin, Colloids Surf., B, 2005, 42, 205–210. 22 F. Velard, D. Laurent-Maquin, J. Braux, C. Guillaume, S. Bouthors, E. Jallot, et al., Biomaterials, 2010, 31, 2001– 2009. 23 A. Grandjean-Laquerriere, P. Laquerriere, E. Jallot, J.-M. Nedelec, M. Guenounou, D. Laurent-Maquin, et al., Biomaterials, 2006, 27, 3195–3200. 24 J. D. Li, Y. B. Li, Y. Zuo, G. Y. Lv, W. H. Yang and Zh. Y. Tian, Mater. Sci. Forum, 2006, 510–511, 890–893. 25 J. Rodriguez-Carvajal, PROGRAM FullProf.2k – version 3.20, Laboratoire L´ eon Brillouin (CEA-CNRS), Saclay, France, 2005, FullProf.2k manual http://www-llb.cea.fr/fullweb/fp2k/ fp2k_divers.htm; See also J. Rodriguez-Carvajal and T. Roisnel, EPDIC-8, 23–26 May 2002; Trans. Tech. Publication: Uppsala, Sweden; Mater. Sci. Forum 2004, 123, 443. 26 L. B. McCusker, R. B. Von Dreele, D. E. Cox, D. Lou¨ er and P. Scardi, J. Appl. Crystallogr., 1999, 32, 36–50. 27 R. J. Hill, J. Appl. Crystallogr., 1992, 25, 589–610. 28 N. V. Y. Scarlett, I. C. Madsen, L. M. D. Cranswick, T. Lwin, E. Groleau, G. Stephenson, et al., J. Appl. Crystallogr., 2002, 35, 383–400. 29 P. Lechner, A. Pahlke and H. Soltau, X-Ray Spectrom., 2004, 33, 256–261. 30 N. Newville, J. Synchrotron Radiat., 2001, 8, 322–324. 31 B. Ravel, J. Synchrotron Radiat., 2001, 8, 314–316. 32 J. J. Rehr, J. Mustre de Leon, S. I. Zabinsky and R. C. Albers, J. Am. Chem. Soc., 1991, 113, 5135. 33 S. C. Abrahams and J. L. Bernstein, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1969, 25, 1233–1236. 34 G. M. Lombardo, G. C. Pappalardo, F. Punzo, F. Costantino, U. Costantino and M. Sisani, Eur. J. Inorg. Chem., 2005, 2005, 5026–5034. 35 H. Rieck and R. Hoppe, Z. Anorg. Allg. Chem., 1973, 400, 311– 320.
J. Mater. Chem. B, 2014, 2, 536–545 | 545
