Triiron(III) oxonium tetradecahydrogendotriacontaoxooctaphosphate tetrahydrate, (HaO)Fe a(HPO4)2(H2PO4)6.4H20, M r = 1030.6, monoclinic, C2/c, a = 16.797 (8), b = 9.528 (8), c = 17.609 (9) ,~, t = 90.64 (6) °, V = 2818 (3) A 3, Z = 4, D x = 2"43 Mg m -3, Mo Ka radiation, 2 = 0.71069/~, g = 21.7 cm -~, F(000) = 2067.8, T = 290 K, final conventional R fac to r=0 .091 , w R = 0 . 0 9 9 for 986 unique reflections and 218 variables. This analysis and M6ssbauer spectroscopy led to the formula (H30)Fe a(HPO4)2(H2PO4)6.4H20. In the structure, iron(Ill) orthophosphate layers perpendicular to the c axis * Present address: Institute of Inorganic Chemistry, University of Mainz, D-6500 Mainz, Federal Republic of Germany. 0108-2701/86/050525-04501.50 alternate with water layers. The iron phosphate layers contain FeO 6 octahedra which share vertices with O2P(OH)2 and O3P(OH) tetrahedra. Within this network are hollows that are occupied by oxonium ions. The pseudo symmetry of the layer is well described by the layer symmetry group 3p21. Neighbouring layers are connected by hydrogen-bonded water molecules. Fe--O distances vary from 1.94 (2) to 2.06 (2) A. The present compound is isostructural with analogous thallium and aluminium orthophosphates. Introduction. Iron phosphates are of importance in many technological and ecological areas, such as corrosion inhibition in systems carrying water, the interaction of phosphates as adhesives on iron and steel, © 1986 International Union of Crystallography 526 (H30)Fe3(HPO4)2(H2PO4)6.4H20 and the interaction of soil components (e.g. goethite, ~t-FeOOH) with fertilizers containing phosphates. Despite this importance little is known about the structural chemistry of the iron phosphates. Even orthophosphoric acid is very complex and contains a large assortment of linear and cyclic oligomers based on P O P bridges (Van Wazer, 1958, 1961). As many of the iron phosphate compounds occurring in technological problems are almost insoluble and unstable at high temperatures, preparation of single crystals is difficult and most attempts at characterization have been attempted by methods like M6ssbauer spectroscopy (see e.g. Meisel, Guttmann & Giitlich, 1983). Thus when we found small single crystals of the present compound in the reaction products of iron with phosphoric acid, we decided to elucidate the structure by X-ray analysis. Expedmental. Compound prepared by reaction of 4.5 g iron powder (ct-Fe) with 50 ml 85% HaPO 4. Reaction vessel was allowed to stand for eleven weeks at room temperature without stirring. The pale pink-grey precipitate formed contained small crystals. Mrssbauer spectra indicated presence of only Fe lu in the sample (Meisel, Ten Dolle, Mintjes, Fuggle, Bosman & Beurskens, 1986). Irregular crystal 0.15 x 0.12 x 0.03 mm selected for X-ray analysis. MoKa radiation, graphite-crystal monochromator, Nonius CAD-4 single-crystal diffractometer. Unit-ceU dimensions determined from angular settings of 25 reflections with 5 < 0 < 18 o. Space group C2/c from systematic absences and structure determination. Intensity data of 9898 reflections (full sphere up to 0 = 25 °) measured, using co-20 scan technique, scan angle 1.90 ° and variable scan rate with max. scan time 20 s per reflection. Intensity of primary beam checked throughout data collection by monitoring three reference reflections every 30 min. Decline in intensities up to 29%, probably due to decomposition, occurred over course of data collection. Smooth curve based on reference reflections used to correct for this drift. On all reflections, profile analysis performed (Lehmann & Larsen, 1974; Grant & Gabe, 1978); empirical absorption correction applied using ~ scans (North, Phillips & Mathews, 1968). (Correction factors in the range 0.77 to 1.00; analytical absorption corrections could not be performed because of irregular shape of crystal.) Space-group-symmetry-equivalent reflections averaged, Rin t = 0.12, resulting in 2419 unique reflections of which 986 observed with I > 3o(/) and Rtnt=0-04. No extinction correction. Lorentz and polarization corrections applied and data reduced to I Fol values. The three iron atoms found using Patterson techniques (SHELX84; Sheldrick, 1984). P and O atoms found with DIRDIF (Beurskens et aL, 1982). Structure refined by full-matrix least squares on I FI values, using SHELX76 (Sheldrick, 1976). Scattering factors from International Tables for X-ray Crystallography (1974). Isotropic refinement converged to R =0.119. At this stage empirical absorption correction applied (Walker & Stuart, 1983), resulting in further decrease of R to 0.104 (correction factors in range 0.876-1.143). During final stages of refinement positional parameters and anisotropic thermal parameters of all atoms refined. Location of hydrogen atoms not possible, due to quality of data and probable disorder of positions. Final conventional agreement factors were R =0.091 and wR=0.099 for 986 observed reflections and 218 variables. High R factor due to quality of crystal. Function minimized was ~W(Fo--Fc) 2 with w = 1/[o2(Fo) + 0.0004F J] with O(Fo) from counting statistics. Max. shift over e.s.d. ratio in last full-matrix least-squares cycle less than 0.01. Final difference Fourier map showed no peaks higher than 1 .0eA -3. Plots made with PLUTO (Motherwell, 1976). Table 1. Atomic parameters (with e.s.d.'s)