The structure of platinic acid

Abstract

H2Pt(OH) 6, monoclinic, C2/c, a = 8.459 (5), b = 7.184 (5), c = 7.429 (4) A,/~ = 93.71 (2) ° at 297 K, Z -4, D,n = 4.17, D c = 4.408 Mg m -3. The structure is a slight distortion of the K2PtCI 6 type, with nearly regular Pt(OH) 6 octahedra forming an approximately face-centred cubic array. Short O . . . O distances between octahedra suggest strong hydrogen bonding. Introduction. This study was undertaken to establish the structural formula of platinic acid, which has been written variously as PtO2.4H20, Pt(OH)4.2H20 and, most commonly, H2Pt(OH)6 (Mellor, 1937; Hartley, 1973). While the crystal structures of its Li, Na and K salts have been determined (Tr6mel & Lupprich, 1975; Bjfrling, 1941), previous X-ray data on the parent acid appear to be limited to an incomplete description of its unindexed powder pattern (Keenan, Giesemann & Smith, 1954). The compound was prepared as a pale-yellow powder from Na2PtC16 by Bellucci's method (Sidgwick, 1950), except that a more crystalline product was obtained by performing the precipitation with acetic acid from a hot solution. The X-ray powder pattern was indexed, and lattice parameters were determined by least squares from Guinier photographs taken with Cu Ka~ radiation, with ThO 2 (a = 5.5972 A) as an internal standard. The large e.s.d.'s in the parameters reflect differences between samples; for a single sample the accuracy is about five times better. The source of the variation is not known, but impurities, including water, or disorder are possibilities. When part of the sample used for the structure determination was decomposed at I073 K to yield Pt metal, the weight loss was within 1% of the theoretical value, which suggests reasonable purity. All reflexions with h + k :/: 2n, and hOl reflexions with l :/: 2n were absent, so that the space group is Cc or C2/c. However, all reflexions with k + l :/: 2n were very weak; indeed only 7 of the 72 accessible reflexions of this class were actually observed, which suggests an F-centred array of Pt atoms and implies that the space group is C2/c. Intensities were obtained from Guinier photographs with a recording micro0567-7408/79/123014-02501.00 densitometer and measurement of the peak areas with a planimeter. The region of reciprocal space scanned, 20 < 86 o, contains 95 unique reflexions with k + l = 2n, but overlap reduces the number of distinct peaks to 70, of which one was not observed. The much weaker reflexions with k + l v~ 2n were not included in the refinement. During the refinement it became clear that there was a dependence of intensity on Bragg angle which was not well accounted for by the Guinier Lorentzpolarization factor with absorption corrections (Sas & de Wolff, 1966) unless unrealistic values for the specimen absorption and temperature factor were assumed. Diffractometer intensities were therefore collected with Ni-filtered Cu Ka radiation from a pure sample, and also from samples diluted with flour and with PVA emulsion: agreement between the three sets was good, indicating the absence of preferred orientation. The final data set was a composite of the diffractometer and Guinier data, with the Guinier data adjusted for the difference in Lp factors for the two techniques. A set of relative scattering amplitudes was derived from the Guinier intensity data, assuming all the reflexions in an unresolved group to have the same value of F, and used to construct a Patterson map which indicated that the O atoms formed an almost regular octahedron about Pt. O atom coordinates estimated from the Patterson map, together with scale and an overall isotropic temperature factor, were refined with POWDER (Rossell & Scott, 1975) and the refinement converged rapidly. Scattering factors for Table 1. Fractional atomic coordinates E.s.d.'s of the last digit are in parentheses. Overall isotropic temperature factor B = 1.65 (11) A 2.