Single-crystal structure refinements and crystal chemistry of synthetic trioctahedral micas

To study structural changes with changing chemical composition in the octahedral sheet of trioctahedral potassium-rich 1M micas, 3 natural and 12 synthetic micas were examined by singlecrystal X-ray diffraction. Samples with Ni, Mg, Co, Fe, and Al in the octahedral sheet and with Si, Al and one sample with Ga in the tetrahedral sheet were prepared at high temperatures and pressures, which yielded crystallites £200 mm in size. For samples with approximately AlSi3 composition of the tetrahedral sheet and with no octahedral Al, mean M1-O and M2-O octahedral bond lengths correlate very well with the mean ionic radius of the octahedral cations. For these samples, the M1 and M2 sites are found to be very similar or identical within experimental error in terms of mean M-O bond lengths, bond angles, and polyhedral distortion parameters. Octahedral distortion is negatively correlated with the size of the octahedral cation. Octahedra in Ni-mica show the largest deviation from ideal octahedral geometry, whereas those in annite are closest to ideal octahedral geometry, but they are still significantly flattened. Octahedral Al prefers the M2 site. This causes the mean M1-O bond length to decrease less with increasing Al content as compared to the mean M2-O distance. This Al preference for M2 also causes the M2 site to become smaller and more distorted than the M1 site. Refinement of the Mg-Fe ratios in the octahedral sites along the annite-phlogopite join shows that the two cations are statistically distributed over M1 and M2. Tetrahedra are regular and show only a small elongation along c*. Tetrahedral distortion parameters for the AlSi3 micas show no correlation with chemistry of the octahedral layer. However, the mean T-O bond lengths increase slightly with increasing size of the octahedral cation. With decreasing size of the lateral dimension of the octahedral sheet in the (001) plane, the tetrahedral sheet shows increasing ditrigonal distortion. Largest tetrahedral rotation angles are observed for synthetic near end-member siderophyllite with a = 11.5∞ and for tetra-gallium-phlogopite KMg3GaSi3O10(OH)2 with a = 10.8∞. Mineral names used here refer to synthetic products as well. REDHAMMER AND ROTH: STRUCTURE OF SYNTHETIC TRIOCTAHEDRAL MICAS 1465 investigated. The major goals of this study were (1) to correlate structural changes with mean octahedral ionic radii (Shannon 1970), (2) to correlate deviations from ideal geometry (distortion) of polyhedra with chemical composition and mean M-O bond lengths, and (3) to provide structural data of synthetic end-member (and natural near end-member) compositions for subsequent theoretical electric-field gradient calculations, particularly for annite, sideropyhllite, and phlogopite compositions. Recently, a larger number of structural investigations have been done by single-crystal methods on natural (especially Brigatti and Davoli 1990; Brigatti et al. 1991; Brigatti and Poppi 1993; Brigatti et al. 1996a, b; Brigatti et al. 1998; Brigatti et al. 2000a, 2000b; Brigatti et al. 2001) and synthetic (McCauley et al. 1973; Comodi et al. 1999) micas close to the annite-phlogopite join. Concerning cationic ordering of Fe, Mg on the two non-equivalent M1 and M2 octahedral sites of natural trioctahedral 1 M micas Brigatti and Davoli (1991) and Brigatti et al. (2000a) noted the possibility of a slight preference of Fe for the M1 site. Based on geometric and chemical features, Cruciani and Zanazzi (1994) noted a preferential partitioning of high-charge cations (Ti, Al, Cr, and Fe) in the M2 site. Very recently, Brigatti et al. (2001) also found preferences of Ti for the M2 site in Ti rich trioctahedral 1 M micas. For Al clear evidence for cationic ordering was found very recently. In Fe, Al-rich micas close to the annitesiderophyllite binary, Brigatti et al. (2000a) note a preference of octahedral Al for the M2 site. The same was found independently by Redhammer et al. (2000) from Rietveld refinements. Previously, Guggenheim and Eggleton (1987) suggested a crystal chemical reason why Al would prefer the M2 over the M1 site in the micas. EXPERIMENTAL METHODS Syntheses of mica single crystals (Table 1) were performed in a high-pressure/high-temperature piston-cylinder apparatus at temperatures of 1200–1250 ∞C and pressures of 2.5 GPa for at least 48 hours. About 10–20 mg of the starting material, together with excess of water, was placed in a small platinum capsule (2–4 mm in length, inner diameter 1.5 mm) that was welded tight. After completing the experiment, the integrity of the capsule was tested by weighting the capsule, opening it, drying at 80 ∞C, and weighting it again. Starting materials were prepared either by the gelling method as described by Hamilton and Henderson (1968) (mixtures A20, A40, A60, and A80, which were previously prepared by Redhammer et al. 1995) or by mixing K2CO3 and metal oxides (NiO, MgO, CoO, Fe2O3, SiO2, Al2O3) in the stoichiometry of the desired composition (cf. Redhammer 2001). Electron-microprobe analysis (JEOL JXA 8600, acceleration voltage 15 kV, beam current 30 nA, beam focused to about 3 mm spot size) was used to determine the chemical composition of the mica products. Minerals were used as standards. Because of the small amounts of sample product available, standard epoxy mounts were not made and H2O content and Mössbauer data could not be determined. Instead, several mica flakes were fixed with carbon tape onto a glass slide, and were coated with carbon and then analyzed. This procedure gives consistent results, although the weight percent sums are lower than expected. Calculated mineral formula in Table 2 are based on O10(OH)2. Single-crystal X-ray diffraction data sets were measured with an imaging-plate diffractometer system (Stoe-IPDS, MoKa radiation, pyrolytic graphite monochromator). Intensity data were collected to 56.5∞ 2q within a j-range of 0–229.5∞, and the j increment was 1.5∞/image. This procedure produces 153 images per measurement. Except for samples no. NiPhl, no. Phl, no. CoAn and no. Ann, all lattice parameters were determined from the single-crystal X-ray diffraction measurements (about 1500 measured reflections). For the former samples, sufficient amounts and purity were obtained for powder X-ray diffraction (Siemens D500, CuKa radiation, 10–90∞ 2q graphite monochromator, silicon used as an internal standard). For these samples, the lattice parameters of the powder TABLE 1. Experimental conditions and results of mica synthesis Sample* Starting material composition T † P † t† Products‡ NiPhl KNi3Si3AlO10(OH)2 1250 2.5 67 mi, (ol) Phl KMg3Si3AlO10(OH)2 1250 2.5 61 mi GaPhl KMg3Si3GaO10(OH)2 1250 2.5 61 mi CoAn KCo3Si3AlO10(OH)2 1250 2.5 72 mi, (ol) A20 K(Mg2.4Fe0.6)Si3AlO10(OH)2 1250 2.5 59 mi, (mt, sa), gl A40 K(Mg1.8Fe1.2)Si3AlO10(OH)2 1200 2.5 66 mi, (mt, sa), gl A60 K(Mg1.2Fe1.8)Si3AlO10(OH)2 1200 2.5 63 mi, mt, sa, gl A80 K(Mg0.6Fe2.4)Si3AlO10(OH)2 1200 2.5 60 mi, mt, sa, gl A100 KFe3Si3AlO10(OH)2 1300 2.5 51 gl 1200 2.5 47 gl 1100 2.5 63 gl 1000 2.6 39 gl 950 2.5 49 mi (powder) Mga1.2 K(Mg1.2Fe1.8)Si3AlO10(OH)2 1200 2.5 53 mi, mt, sa, gl Mga1.6 K(Mg1.6Fe1.4)Si3AlO10(OH)2 1200 2.5 70 mi, (mt, sa), gl CoNi1.8 K(Co1.8Ni1.2)Si3AlO10(OH)2 1250 2.5 53 mi, ol CoNi1.2 K(Co1.2Ni1.8)Si3AlO10(OH)2 1250 2.5 73 mi, (ol) Sd87 K(Fe2.13Al0.87)(Al1.87Si2.13)O10(OH)2 700 0.2 631 mi * In the following Tables, individual crystals coming out from a specific sample are denoted by ...#, e.g., crystal A20#2 is the 2nd crystal measured from experiment A20. † T = temperature (in DEGC), P = pressure (in GPa), t = experiment duration (in h). ‡ mi = mica, mt = magnetite, sa = sanidine, gl = glass, ol = olivine; mineral abbreviations in parentheses are below 5 vol% of experimental product (estimated from optical microscopy). REDHAMMER AND ROTH: STRUCTURE OF SYNTHETIC TRIOCTAHEDRAL MICAS 1466 X-ray diffraction experiments were used in the structure refinement. Only crystals of sufficient size (at least 0.05 ¥ 0.05 ¥ 0.01 mm), which gave sharp reflections and showed well-defined periodicity along c*, were used for intensity data collection. About 25 crystals were examined for each composition because many crystals showed significant numbers of stacking faults along c* direction. All crystals examined and subsequently investigated were 1M polytypes, therefore all refinements were performed in space group C2/m (12). The programs X-SHAPE (Stoe and Cie 1996) and SHELXL (Sheldrick 1997) were used for absorption correction and structure refinement, respectively. The atomic positions of Hazen and Burnham (1973) were used as initial values for the structure refinements. The position of the hydrogen atom was obtained from difference-Fourier maps and each was close to the position determined by Joswig (1972) for phlogopite, i.e., x = 0.098(2), y = /2, z = 0.3007(8).