Crystal structure of tetrawickmanite, Mn2+Sn4+(OH)6

The crystal structure of tetrawickmanite, a tetragonal hydroxide-perovskite mineral, has been determined for the first time by means of single-crystal X-ray diffraction. It is characterized by alternating corner-linked [Mn2+(OH)6] and [Sn4+(OH)6] octahedra whose sense of rotation varies along c, in contrast to its dimorph, the cubic wickmanite.

The crystal structure of tetrawickmanite, ideally Mn 2+ Sn 4+ (OH) 6 [manganese(II) tin(IV) hexahydroxide], has been determined based on single-crystal X-ray diffraction data collected from a natural sample from Lå ngban, Sweden. Tetrawickmanite belongs to the octahedral-framework group of hydroxideperovskite minerals, described by the general formula BB'(OH) 6 with a perovskite derivative structure. The structure differs from that of an ABO 3 perovskite in that the A site is empty while each O atom is bonded to an H atom. The perovskite B-type cations split into ordered B and B 0 sites, which are occupied by Mn 2+ and Sn 4+ , respectively. Tetrawickmanite exhibits tetragonal symmetry and is topologically similar to its cubic polymorph, wickmanite. The tetrawickmanite structure is characterized by a framework of alternating cornerlinked [Mn 2+ (OH) 6 ] and [Sn 4+ (OH) 6 ] octahedra, both with point-group symmetry 1. Four of the five distinct H atoms in the structure are statistically disordered. The vacant A site is in a cavity in the centre of a distorted cube formed by eight octahedra at the corners. However, the hydrogen-atom positions and their hydrogen bonds are not equivalent in every cavity, resulting in two distinct environments. One of the cavities contains a ring of four hydrogen bonds, similar to that found in wickmanite, while the other cavity is more distorted and forms crankshaft-type chains of hydrogen bonds, as previously proposed for tetragonal stottite, Fe 2+ Ge 4+ (OH) 6 .

Mineralogical and crystal-chemical context
Tetrawickmanite, ideally Mn 2+ Sn 4+ (OH) 6 , belongs to the octahedral-framework group of hydroxide-perovskites, described by the general formula BB'(OH) 6 with a perovskite derivative structure. The structure of hydroxide-perovskites differs from that of an ABO 3 perovskite in that the A site is empty while each O atom is bonded to a hydrogen atom. The lack of A-site cations makes them more compressible than perovskite structures (Kleppe et al., 2012) and elicits an industrial interest for their potential use in hydrogen storage at high pressures (Welch & Wunder, 2012).
Tetrawickmanite was initially described by White & Nelen (1973) from a pegmatite at the Foote Mineral Company's spodumene mine, Kings Mountain, North Carolina. From the X-ray diffraction pattern and the crystal morphology, they determined that tetrawickmanite exhibits tetragonal symmetry and is topologically similar to its polymorph, the cubic wickmanite. A second occurrence of tetrawickmanite at Lå ngban, Sweden, was reported by Dunn (1978) and described as tungsten-rich tetrawickmanite with tungsten substituting for tin in the structure.
In the course of identifying minerals for the RRUFF Project (http://rruff.info), we were able to isolate a single crystal of tetrawickmanite from Lå ngban with composition (Mn 2+ 0.94 Mg 0.05 Fe 2+ 0.01 ) AE=1 (Sn 4+ 0.92 W 6+ 0.05 ) AE=0.97 (OH) 6 . Thereby, this study presents the first crystal structure determination of tetrawickmanite by means of single-crystal X-ray diffraction.
Hydroxide-perovskites have the vacant A site in a cavity in the centre of a distorted cube formed by eight octahedra at the corners. According to the Glazer notation for octahedral-tilt systems in perovskites (Glazer, 1972), wickmanite, the cubic polymorph of tetrawickmanite, is an a + a + a + -type perovskite, with three equal rotations ( Fig. 1a) while tetrawickmanite is of a + a + c À type and it changes the senses of rotation in alternate layers along the c-axis direction (Fig. 1b). This difference in octahedral-tilt systems is similar to that observed during compressibility studies of cubic burtite [CaSn 4+ (OH) 6 ; Welch & Crichton, 2002] and tetragonal stottite [Fe 2+ Ge 4+ (OH) 6 ; Ross et al., 2002]. As the authors pointed out, the variance in the octahedral-tilt systems leads to distinct hydrogen-bonding topologies between burtite and stottite, similar to those observed between wickmanite and tetrawickmanite.
Wickmanite has a single type of cavity with the H atom disordered over two positions, forming a ring of four hydrogen-bonds with two other hydrogen-bonds at the top Framework of alternating corner-linked [Mn 2+ (OH) 6 ] and [Sn 4+ (OH) 6 ] octahedra in (a) wickmanite (Basciano et al., 1998) and (b) tetrawickmanite, with change in senses of rotation in alternate layers along the c-axis direction. Yellow and grey octahedra represent Mn and Sn sites, respectively. Blue spheres represent H atoms.
As stated earlier, the compressibilities of cubic burtite and tetragonal stottite, with unit-cell volumes 535.8 and 426 Å 3 , respectively, have been studied and their hydrogen bonding has been compared (Welch & Crichton, 2002;Ross et al., 2002). By analogy, a study of the compressibility of the polymorphs wickmanite and tetrawickmanite, with much closer unit-cell volume values (488.26 and 482.17 Å 3 , respectively), might also help in understanding the connection between hydrogen-bonding topologies and compression mechanisms in hydroxide-perovskites. Kleppe et al. (2012) studied pressure-induced phase transitions in hydroxide-perovskites based on Raman spectroscopy measurements of stottite [Fe 2+ Ge 4+ (OH) 6 ] up to 21 GPa. In their work, they proposed the monoclinic space group P2/n for stottite at ambient conditions derived from the presence of six OH-stretching bands in the Raman spectra in the range 3064-3352 cm À1 . We refined the structure of tetrawickmanite in space group P2/n (R 1 = 0.0215) and performed the Hamilton reliability test (Hamilton, 1965). The test indicated that the better structural model for tetrawickmanite is based on the tetragonal space group P4 2 /n at the 92% confidence level. Moreover, analysis of the anisotropic displacement parameters showed that the tetragonal model displays ideal rigidbody motion of the strong polyhedral groups (Downs, 2000), thus corroborating a tetragonal structure for tetrawickmanite.
The Raman spectrum of tetrawickmanite in the OHstretching region (2800-3900 cm À1 ) is displayed in Fig. 4. The minimum number of peaks needed to fit the spectrum in this region (using pseudo-Voigt line profiles) is seven, which is in agreement with the number of hydrogen bonds derived from the structure (Table 1). According to the correlation of O-H stretching frequencies and O-HÁ Á ÁO hydrogen-bond lengths in minerals by Libowitzky (1999), the most intense peaks (3062, 3145, 3253 and 3374 cm À1 ) are within the range of calculated wavenumbers for the HÁ Á ÁO distances between 2.75 and 2.86 Å and they correspond to the strongest hydrogen bonds in the structure.

Experimental
The tetrawickmanite specimen used in this study was from Lå ngban, Sweden, and is in the collection of the RRUFF project (deposition R100003: http://rruff.info/R100003). Its chemical composition was determined with a CAMECA SX100 electron microprobe at the conditions of 20 kV, 20 nA and a beam size of 5 mm.   Raman spectrum of tetrawickmanite in the OH-stretching region (2800-3900 cm À1 ). At the top right, the spectral deconvolution obtained with seven fitting peaks using pseudo-Voigt line profiles.  6 . The Raman spectrum of tetrawickmanite was collected from a randomly oriented crystal on a Thermo-Almega microRaman system, using a 532 nm solid-state laser with a thermoelectric cooled CCD detector. The laser was partially polarized with 4 cm À1 resolution and a spot size of 1 mm.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 2. Electron microprobe analysis revealed that the tetrawickmanite sample studied here contains small amounts of W, Mg and Fe. However, the structure refinements with and without a minor contribution of these elements in the octahedral sites did not produce any significant differences in terms of reliability factors or displacement parameters. Hence, the ideal chemical formula Mn 2+ Sn 4+ (OH) 6 was assumed during the refinement, and all non-hydrogen atoms were refined with anisotropic displace-ment parameters. All H atoms were located from difference Fourier syntheses. The hydrogen atoms H1-H4 were modelled as statistically disordered around the parent O atom. H atom positions were refined freely; a fixed isotropic displacement parameter (U iso = 0.03 Å ) was used for all H atoms.

Special details
Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes. Refinement. Refinement of F 2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F 2 , conventional R-factors R are based on F, with F set to zero for negative F 2 . The threshold expression of F 2 > σ(F 2 ) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F 2 are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger.