A synchrotron study of Na2.27Ho7.73(SiO4)6O0.72

A well crystallized powder sample of sodium holmium orthosilicate oxyapatite, Na2.27Ho7.73(SiO4)6O0.72, was obtained after mechanical milling and thermal treatment at 1123 K. Crystal structure analysis was performed from the results of Rietveld refinement of the synchrotron diffraction data. As in other rare-earth orthosilicate apatites, sodium cations appear located sharing with holmium the 4f Wyckoff position at the center of a tricapped trigonal prism. In its turn, holmium almost fully occupies the 6h position at the center of a seven-coordinated pentagonal bipyramid. A small quantity of Na atoms was found at this site. No vacancies are present in the two independent crystallographic sites available for Ho and Na atoms.

A well crystallized powder sample of sodium holmium orthosilicate oxyapatite, Na 2.27 Ho 7.73 (SiO 4 ) 6 O 0.72 , was obtained after mechanical milling and thermal treatment at 1123 K. Crystal structure analysis was performed from the results of Rietveld refinement of the synchrotron diffraction data. As in other rare-earth orthosilicate apatites, sodium cations appear located sharing with holmium the 4f Wyckoff position at the center of a tricapped trigonal prism. In its turn, holmium almost fully occupies the 6h position at the center of a seven-coordinated pentagonal bipyramid. A small quantity of Na atoms was found at this site. No vacancies are present in the two independent crystallographic sites available for Ho and Na atoms.
Portions of this research were carried out at the Stanford Synchrotron Radiation Lightsource (http://smb.slac.stanford. edu/powder/), a national user facility operated by Stanford University on behalf of the US Department of Energy, Office of Basic Energy Sciences. The authors thank Manuel Aguilar for the conventional X-ray diffraction measurements at the Instituto de Física, Universidad Nacional Autó noma de Mé xico, and Carlos Linares for the WDS measurements at the Laboratorio Universitario de Petrología, Universidad Nacional Autó noma de Mé xico, and Angel Osornio for technical support. IR acknowledges a fellowship from the Consejo Nacional de Ciencia y Tecnología (CONACyT) and projects LiY 9 (SiO 4 ) 6 O 2 (Redhammer & Roth, 2003). The structural data used for the title OAp phase was established considering the isostructural compound NaY 9 (SiO 4 ) 6 O 2 reported by Gunawardane et al. (1982). For this compound, the chemical formula Na x Y 10 -x (SiO 4 ) 6 O 2 represents a model in which the sodium ion is introduced in the apatite structure substituting for yttrium. In the present work the nominal formula used for holmium oxyapatite and the possible incorporation of sodium and fluorine is represented as Na x Ho 10 -x (SiO 4 ) 6 O 2 -y F y . The crystal structure of OAp has an arrangement similar to some other alkaline rare-earth oxyapatites already reported such as NaY 9 (SiO 4 ) 6 O 2 and LiY 9 (SiO 4 ) 6 O 2 (Redhammer & Roth, 2003).

Experimental
The mechanical milling at 612 rpm (FRITSCH Pulverisette mill model 06.102) of sodium holmium orthosilicate oxyapatite (OAp) was carried out considering the nominal composition Na 2 Ho 8 (SiO 4 ) 6 F 2 from stoichiometric mixtures of Ho 2 O 3 (Aldrich.99.9%), SiO 2 (Aldrich 99.6%) and NaF (Analit.Analytic grade) following the method of Rodríguez-Reyna et al. (2006) and Fuentes et al. (2006). Once the mechanochemical processing finished, the sample was heated in a tube furnace for 15 h at 1073 K in air atmosphere. Conventional X-ray powder diffraction data showed reflections that match with the isostructural oxyapatite NaY 9 (SiO 4 ) 6 O 2 (PDF file 35-404). At this step of synthesis, the sample exhibited very poor crystallinity. After an additional thermal treatment at 1123 K, the crystallization of the sample was reached with the presence of quartz (PDF file 46-1045) as a secondary phase. From wavelength dispersive spectrometry (WDS),chemical analysis was performed by means of a Jeol JXA-8900R, EPMA spectrometer.
The average atomic content with standard uncertainty for each element in the sample were 12.4(1.9) for sodium, 42.2(4.6) for holmium and 45.2(6.3) for silicon, and give a Na:Ho ratio of 0.29. The expected values, according to the initial mixture of reactives were 12.5 for sodium, 50.0 for holmium and 37.5 for silicon. These quantities changed as consequence of the process of mechanical milling. Fluorine could not be observed so the phase originally labeled as OAp could be considered as pure oxyapatite phase. The atomic content for each element in the unit cell for OAp was calculated dividing by 5.46 the averaged WDS values giving 2.27 for sodium, 7.23 for holmium and 8.28 for silicon. The excess of silicon was interpreted taking into account the presence of quartz measured as a secundary phase by x-ray diffraction. Therefore, the content of silicon had to be split in such way that 6 silicon atoms were corresponded to OAp and 2.28 to quartz for each unit cell of oxyapatite.

Refinement
The incident beam was calibrated using LaB 6 (NIST SRM 660) as reference standard. The starting parameters for performing the Rietveld refinement were, for the OAp phase, the data from the isostructural NaY 9 (SiO 4 ) 6 O 2 (ICSD 27191, Gunawardane et al., 1982); and for quartz the data from ICSD 90145 (Gualtieri, 2000). Atomic coordinates were refined for the OAp phase considering the holmium and sodium atoms distributed in each one of the two available sites according to the Na:Ho ratio of 0.29 (obtained by WDS). In the following step, the occupation factors for Na and Ho were constrained in such a way that the total amount of both Na and Ho be constant keeping the Na:Ho ratio at 0.29. This ratio was allowed to change for each one of the crystallographic sites. At the end of the refinement, the thermal isotropic parameters were refined independently for each site. These parameters were set to the same refined value for all the O atoms in the structure.
Bond valence calculations were made using the recommended bond-valence parameters for oxides published by Brese & O'Keeffe (1991). Bond valence sum around Ho1 and Ho2 gave the values of 3.01 and 2.75 respectively, being the last lower than the expecting valence of 3. One way to explain the low bond valence sum is considering the Na substituting for Ho. In such case, bond valence sum for Na2 in the Ho2 site is found to be 1.51, and the weighted bond valence sum results equal to the weighted atomic valence when the fraction of Na2 is 0.33 in the Ho2 site. This quantity is reasonably close to 0.38 obtained by Rietveld refinement. For the Ho1 site, the bond valence sum around Ho is practically the expecting valence of 3 which agrees with the very low proportion of Na (0.12) obtained in the refinement. For the Ho1-O4 bonds there are three Ho1 atoms coordinated to O4 oxygen with a bondlengths of 2.197 Å. Each Ho1-O4 bond contributes with a bond valence of 0.62 giving a bond valence sum of 1.87. This value agrees with the fact that this site is partialy occupied by oxygen O4 by an occupation factor of 0.36. With all this results, the chemical formula obtained was found to be Na 2.27 Ho 7.73 (SiO 4 ) 6 O 0.72 , which is charge balanced. The composition agrees with the WDS results and sodium content. The final Rietveld refinement of the synchrotron diffraction pattern is shown in Fig. 1 and the crystal structure is represented in Fig. 2. Fig. 1. Rietveld refinement for OAp employing synchrotron radiation data. Observed (crosses), calculated (solid line) and difference (bottom trace) plots are represented. Vertical marks correspond to the allowed Bragg reflections for OAp (first row) and SiO 2 (second row).