The crystal structure of Sr4PtO6 and two related compounds

Abstract

hydrolyze readily on contact with water. This rapid hydrolysis suggested that the platinum metal coordination polyhedra did not share oxygens with each other. Since the X-ray powder patterns were simple in appearance, it seemed that a crystal structure determination would be feasible as well as informative. Experimental Powder photographs of Sr4PtO s and Sr4IrO s could be indexed fairly well on the basis of a primitive cubic unit cell with a-6.88 A. The measured density of the platinum containing compound (6.52 g.cm.-3) corresponded to two formula weights for such a unit cell. Assuming a non-defect structure, only nine primitive cubic space groups could accommodate the numbers of atoms involved. In none of these space groups was it more than a one parameter problem to place all the cations. All the possibilities were examined and shown to be incorrect. It was, therefore, concluded that the compounds were not cubic and that single crystals studies would be necessary. Attempts to prepare single crystals by starting with the Sr4PtO s and Sr4IrO s powders were not successful. However, by prolonged heating of a pelleted sample of Sr4RhO s in a ceramic boat, a small, triangular, platy crystal was obtained. Zero and fourth level Weissen-berg photographs were taken of this crystal with the rotation axis normal to the plate. The absences indicated a rhombohedral lattice. The hexagonal cell dimensions were determined as a = 9.74, c = 11.84 A. Before further data could be accumulated, the crystal was lost and another of the same shape could not be found. Rather large acicular crystals were obtained by again starting with the Sr4RhO 6 powder but this time heating with SrF~. as a flux in a platinum crucible. Some of these crystals were mounted with the needle axis as rotation axis, and Weissenberg photographs were taken of the zero through fourth levels using the equi-inclination method. Intensity photographs were taken using the multiple film technique and were read by comparison with a calibrated intensity strip. Lorentz and polarization corrections were made using Lu's (1943) charts. No correction was made for extinction or absorption. Approximately the same unit-cell size, a = 9.74, c = 11-90 ~, was found for the needle shaped crystals as ~or the triangular one. However, although the symmetry was still trigonal, more reflections were observed for the needles, and the rhombohedral lattice absence rule was not obeyed. In fact, the first of these crystals examined appeared to have C6~ symmetry on upper levels, although for other crystals the upper level symmetry was clearly lower than C6z. On further investigation it was found possible, for all of the acicular crystals photographed, to divide the reflections into two groups, one for which h-k+l = 3n, and the other for which-h+k+l = 3n. These two sets could be arranged in a one to one correspondence with the relative intensities within one set the same as those in the other. The intensity ratio between the two sets was not the same for different crystals. These observations could be explained by assuming the crystals to be rotation twins with, in general, one contribution larger than the other. Granting this assumption, the lattice would be rhombohedral. The probable space group is one of R3c or R3c. By trial and error and the use of structure factor plots it was found that reasonable agreement between observed and calculated structure factors (R < 0.25) could be obtained for all levels but the second using space group R3c by placing 6 Sr in 'a', 6 Rh in 'b', and 18 Sr in 'e', with x = 0.370. For the second level, however, R was about 0-5. Attempts to improve this poor fit for the second level by changing to space group R3c were not successful. At this point it was decided to substitute iridium for rhodium in the calculations in order to see how the structure factors so calculated would compare with the observed powder data for Sr4Ir06. The fit was acceptable. At the same time it was noted that F(hk2) values calculated for Sr4IrO ~ gave a more satisfactory fit with the observed F(hk2) values for the single crystals (presumed to be Sr4Rh06) than did the structure factors calculated for Sr4RhO 6. There was, however , no possibility of appreciable amounts of iridium in these crystals. The only explanation which seemed reasonable was that platinum from the platinum crucible had entered into the crystals. (Atomic numbers: Rh, 45; Ir, 77; Pt, 78). To test this hypothesis of platinum substitution, the formula for the single crystals examined was assumed to be 8r4Rhl_~_Ptx06 and F(hk2) values were calculated (neglecting oxygen) for x = 0.00; 0.25; 0.50; 0.75; 1.00, and compared with the observed values. In each case the scale factor was taken so as to make .~lFol = .~lFcl and R was computed. The values for R in the order of increasing x are 0.46, 0.33, 0.23, 0.20, and 0.16. Thus, even though the single crystals were prepared from Sr4RhO6, the X-ray evidence indicates that the crystals examined were really Sr4PtO 6. A chemical analysis of the whole batch of crystals prepared in the platinum crucible was made and showed that there actually was about a third as much Pt as Rh in the material, quite enough, certainly, to result in a formula of Sr4PtO 6 for single crystals. Since the platinum metals in the +4 oxidation state usually take on 6-fold coordination, it was assumed that along the 3-fold axis (Pt-Sr-Pt-Sr) there would be a triangle of oxygens located between each pair of cations. Assuming each of these oxygens to be 2.05 A from Pt and 2.53/~ from Sr, the z parameter for 36 'f' was fixed at 0.093. The orientation of this triangle was determined by calculating first level structure factors at various angles of rotation. Using tlfis method, the x and y parameters for oxygen were chosen as x-0.181 and y = 0.011. The overall R factor was improved from 0.154 to 0.122 by including 36 oxygens at these positions. In summary, there are