Reinvestigation of the crystal structure of Ca2Ce8(SiO4)6O2 apatite by Rietveld refinement

The lattice parameters of apatite-type Ca2Ca8(SiO4)6O2 determined from X-ray powder diffraction data are inconsistent with those found in the literature for the same type of material. Hence the structure was redetermined and compared with previously published data.


Chemical context
Contaminated metallic wastes produced by the nuclear industry need to be managed. This is often achieved by melting them with an oxide slag to make the waste packages more dense and to concentrate the residual plutonium or uranium oxide contamination of the metal. Laboratory work on actinides is facilitated by the use of surrogates that mimic their properties of interest. Cerium can thus be used to simulate the presence of plutonium (Ramsey et al., 1995). The reactivity of cerium(IV) oxide added to melted stainless steel and an SiO 2 -CaO-Al 2 O 3 glass slag under neutral conditions and at high temperature was studied. Powdered stainless steel from Alfa Aesar and an SiO 2 -CaO-Al 2 O 3 lab-made glass frit were fused at 1723 K under argon for 6 h in a graphite crucible. Cerium(IV) oxide was introduced to simulate PuO 2 . The Ellingham diagram predicts that Ce IV is reduced to Ce III under these conditions, which is the predominant cerium form at 1723 K. The same behavior is expected for Pu (Ellingham, 1944). After melting, SEM observations of the sample showed that cerium was concentrated in the glass, as expected, but in two different forms. Cerium is present in the homogeneous part of the glass, at a typical content of 5 wt.% but is also found inside large crystals of hundreds of micrometers across. The X-ray diffractogram of bulk material shows an amorphous bump, attributed to the glass, and a typical pattern for an apatite structure in space group type P6 3 /m. The approximate composition of this phase was determined as Ca 2.4AE0.3 Ce 7.6AE0.3 (SiO 4 ) 6 O 2 and a cell volume of roughly 560 Å 3 . However, the diffraction data is inconsistent with the PDF card for Ca 2 Ce 8 (SiO 4 ) 6 O 2 (#00-055-0835; ICDD, 2015) with a cell volume of 530.96 Å 3 (Skakle et al., 2000). The difference between the two volumes, 5.2%, cannot be explained by the difference in composition between the two phases. For instance, the cell volumes of the apatites Ca 2 Nd 8 (SiO 4 ) 6 O 2 (#00-028-0228) and Ca 2.8 Nd 7.8 (SiO 4 ) 6 O 2 (#04-007-5969) are 552.20 and 551.76 Å 3 , respectively, a difference of just 0.08%. Moreover, the difference between the cell volumes of Ca 2 La 8 (SiO 4 ) 6 O 2 (#00-029-0337) and Ca 4 La 6 (SiO 4 ) 6 O 2 (#04-007-9090) is 0.3%. In other words, the 5.2% difference between the cell volume of Ca 2 Ce 8 (SiO 4 ) 6 O 2 (Skakle et al., 2000) and Ca 2.4AE0.3 Ce 7.6AE0.3 (SiO 4 ) 6 O 2 (this work) cannot be explained by their differing calcium contents. These are the reasons why the structure of Ca 2 Ce 8 (SiO 4 ) 6 O 2 was reinvestigated in the present work.

Structural commentary
Apatites are mineral phases whose general formula is A 10 (XO 4 ) 6 Z 2 , where A = Ca, Sr, Ba, or many rare earth elements, X = B, Si, P, V, As and Z = OH, Cl, F, O, etc. (Byrappa &Yoshimura, 2001). Generally, apatites crystallize in the hexagonal crystal system in space group P6 3 /m. There are two types of A cations in these structures: Type I (Wyckoff position 4f) A cations are aligned along the threefold rotation axes. Theses cations are separated on each of these axes by one-half the value of the c parameter. Type I cations are sometimes called columnar cations. They are ninefold coordinated by oxygen atoms and these columns of AO 9 polyhedra are linked together by XO 4 tetrahedra, with three of the oxygen atoms belonging to one column and the fourth to an adjacent column (Elliott, 1994). This results in a skeleton of XO 4 tetrahedra (point group symmetry m..) alongside the columnar A cations. This skeleton defines channels that are collinear to the c axis and which correspond to the sixfold screw axes. The Z anions and the remaining A cations, also called type II cations and located on mirror planes (6h), are located inside these channels with the Z anions positioned in ellipsoidal cavities along the c axis. Type II cations are sevenfold coordinated, and these sites are smaller than those centered on type I cations. For the present structure, type I cations are statistically occupied by Ca and Ce whereas type II cations are solely occupied by Ce.
The refined title structure is displayed in Fig. 1. In the case of the synthesized Ca 2 Ce 8 (SiO 4 ) 6 O 2 apatite, the possible substitution of trivalent Ce by tetravalent Ce would require one positive charge, which should be balanced by replacing calcium by a monovalent cation. No such element was detected by energy-dispersive X-ray fluorescence (EDS). Such a substitution of Ce 3+ by Ce 4+ in a britholite was investigated by Terra (2005). However, the characterization of the substitution gave unclear results, revealing that the substitution was not successful. This indicates that the presence of tetravalent Ce in the synthesized apatite is also unlikely. It should be noted that there is no obvious explanation for the differences between our structure model (in particular in terms of lattice parameters) and the structure model with the same composition reported by Skakle et al. (2000;#00-055-0835). These authors stated that the difference might result from a different synthesis route, i.e. hydrothermal for their Ce-apatite versus solid-state routes for other Ln-apatites. However, in our opinion, differences in the synthesis route can result only in slight differences between the resulting structures for a given composition. On the other hand, differences occur mainly because of slight variations in the composition, especially for wet synthesis routes for which the presence of carbonate or hydrogenphosphate in the structure can be difficult to avoid. Another known source of variations in the lattice volume is the presence of radiation defects, but this does not seem relevant here. Table 1 reports some bond lengths compared with the already published Ce-apatite #00-055-0835 (Skakle et al., 2000) and La-apatite #00-029-0337 ( Table 1 Selected bond lengths (Å ) and angles ( ) in the redetermined structure and the already published Ce apatite structure (Skakle et al., 2000) and Ca 2 La 8 (SiO 4 ) 6 O 2 .

Figure 1
Polyhedral representation of the Ca 2 Ce 8 (SiO 4 ) 6 O 2 structure. SiO 4 tetrahedra are in blue, the mixed Ca/Ce sites are shown as pink/yellow and the Ce sites as yellow spheres, respectively. 1977). As already noticed by Skakle et al. (2000), the Si1-O1 bond length of their structure model was rather short (1.42 Å ), with the corresponding SiO 4 tetrahedron highly distorted, as shown by the distorsion index calculated by VESTA (Momma & Izumi, 2011). The Si-O bond lengths of the redetermined structure are in good agreement with expected values and those of other La-apatites (Table 1). Likewise, the distorsion index of the SiO 4 tetrahedron is smaller with an order of magnitude for the redetermined structure and other Laapatites.
The linear correlation indicates that the ionic radius of the lanthanide controls the cell volume. For the Ce-apatite (#00-055-0835), the cell volume in the published structure is 530.96 Å 3 , which does not follow the general trend (Skakle et al., 2000). However, the structure reported in the current article fits with the linear correlation between ionic radius and cell volume. Table 2 compiles the corresponding lattice parameters and cell volumes.

Synthesis and crystallization
The usual synthesis protocol for Ca 2 Ln 8 (SiO 4 ) 6 O 2 apatites is a calcination of CaO, Ln 2 O 3 and SiO 2 in appropriate amounts under air above 1673 K (Nicoleau et al., 2016). A trivalent Ln precursor is used, with the same oxidation state as found in the final apatite. However, cerium oxide is usually only available as CeO 2 , in which cerium is tetravalent. The synthesis of Ce 2 O 3 is known to be quite difficult and this phase is not fully stable under air (Bä rnighausen & Schiller, 1985;Strydom & van Vuuren, 1987;Perrichon et al., 1994;Hamm et al., 2014). Hence, a particular synthesis protocol was adopted to successfully prepare the Ca 2 Ce 8 (SiO 4 ) 6 O 2 phase. Metallic silicon was added to a mixture of CeO 2 , SiO 2 and CaO to

Figure 2
Plot of the volume of Ca 2 Ln III 8 (SiO 4 ) 6 O 2 Ln-apatites versus ionic radii of Ln 3+ ions in a VIII-coordinated environment. ensure a double reaction during the thermal treatment: (i) in situ reduction of CeO 2 to Ce 2 O 3 with oxidation of Si to SiO 2 and (ii) synthesis of the apatite by calcination of Ce 2 O 3 , SiO 2 and CaO. Silicon was chosen because it reduces CeO 2 to Ce 2 O 3 and is inert to CaO, SiO 2 and the alumina crucible used for the reaction (Ellingham, 1944). Moreover, the product of the reaction relates to the final composition of the apatite without by-products. The following amounts of precursors were used: 688.5 mg of CeO 2 , 28.1 mg of Si, 120.2 mg of SiO 2 and 56.1 mg of CaO. The mixture was manually milled three times in an agate mortar and gently pressed in an alumina crucible. The sample was then heat treated under argon at 1873 K for 1 h with a heating ramp of 10 K min À1 , then cooled at a controlled rate of 30 K min À1 . A radially and axially shrunk pellet was obtained with a homogeneous brownish color, which was crushed in an agate mortar before X-ray diffraction measurements. The powdered sample was analyzed in a Panalytical XPert MPD Pro diffractometer in Bragg Brentano geometry for 5 h with 2 varying between 15 and 130 , using copper radiation. The powder was polished for SEM observation and EDS measurements. The average composition measured from six points was Ca 2.1 Ce 7.9 -(SiO 4 ) 6 O 2 , which is close to the fixed composition of Ca 2 Ce 8 (SiO 4 ) 6 O 2 chosen for the refinement.

Refinement
Details of the data collection and structure refinement are summarized in Table 3 and Fig. 3. The occupancies of the Si and O atoms were fixed to unity in agreement with the general observation that there are no vacancies on these sites for apatites. The total occupancies of the 6h and 4f sites were constrained to unity and the Ca 2 Ce 8 (SiO 4 ) 6 O 2 composition was kept fixed. The 6h site was considered as fully occupied by cerium since the refined calcium content was very low (0.9%). It is the same case as in the Pr-apatite structure (#00-029-0362) but not for the La (#00-029-0337) and Nd (#00-028-0228) apatites where calcium contents on the 6h site were deter-mined to be 1.4% and 4%, respectively. The 4f site was modelled as half-occupied by Ca and Ce ions. Isotropic displacement parameters (U iso ) were constrained to 0.008 Å 2 for the 4f site atoms, 0.006 Å 2 for the 6h site atoms, 0.005 Å 2 for Si and 0.003 Å 2 for oxygen sites. The residual electron density is 5.75 e Å À3 at a distance of 0.97Å from site Ce_b. Experimental and calculated X-ray powder data profiles of Ca 2 Ce 8 (-SiO 4 ) 6 O 2 , with difference plot.  (Panalytical, 2011); cell refinement: JANA2006 (Petříček et al., 2014); data reduction: JANA2006 (Petříček et al., 2014); program(s) used to solve structure: JANA2006 (Petříček et al., 2014); program(s) used to refine structure: JANA2006 (Petříček et al., 2014); molecular graphics: VESTA (Momma & Izumi, 2011); software used to prepare material for publication: publCIF (Westrip, 2010).