Charge density analysis of two polymorphs of antimony(III) oxide Andrew E. Whitten, a Birger Dittrich, b Mark A. Spackman,* a Peter Turner c and Trevor C. Brown a a Chemistry, University of New England, Armidale, NSW, 2351, Australia. E-mail: mspackma@une.edu.au b Institut für Kristallographie, Freie Universität Berlin, Takustr. 6, D-14195 Berlin, Germany c Crystal Structure Analysis Facility, University of Sydney, NSW, 2006, Australia Received 8th October 2003, Accepted 10th November 2003 First published as an Advance Article on the web 24th November 2003 High-resolution X-ray diffraction data have been collected on the cubic polymorph of antimony() oxide (senarmontite) to determine the charge distribution in the crystal. The results are in quantitative agreement with crystal Hartree–Fock calculations for this polymorph, and have been compared with theoretical calculations on the orthorhombic polymorph (valentinite). Information about the nature of bonding and relative bond strengths in the two polymorphs has been extracted in a straightforward manner via topological analysis of the electron density. All the close contacts in both polymorphs are found to be similar in nature based on the value of the Laplacian, the magnitude of the electron density and the local energy density at the bond critical points, and these characterise the observed interactions as substantially polar covalent, similar to molecular calculation results on Si–O and Ge–O. Electrostatic potential isosurfaces reveal the octopolar nature of this function for senarmontite, and shed light on the observed packing arrangement of Sb 4 O 6 molecules in the crystal. Introduction Antimony compounds combined with halides have long been known to retard the propagation of flames. Historically senar- montite, the cubic polymorph of antimony trioxide, has been used as an additive in various products such as plastics, while the other polymorph, valentinite, is of little commercial value. Changing the reaction conditions can produce varying proportions of either polymorph, and we anticipate that information on bonding in each of the polymorphs may assist in understanding the reasons for the preferential form- ation of a given polymorph in certain conditions. This paper reports the results of charge density investigations of the two polymorphs. Although charge density analysis is now an established sub- field of crystallography, the number of studies carried out on compounds containing relatively heavy atoms remains rather small. Challenging problems often faced in the data analysis of such systems include large absorption, extinction and anomal- ous dispersion effects, and the possibilities of anharmonic thermal motion and anisotropic extinction. In addition, the low ratio of valence to core electrons in these compounds makes it difficult to study bonding features in the crystal, as the contri- bution to structure factors from core electrons tends to swamp the signal from valence electrons, except at low values of sin θ/λ. However, it is in precisely this region of reciprocal space that the attenuating effect of extinction on the observed intensities is the greatest. Corrections can be made for the problems mentioned here, but they nevertheless have the potential to compromise the charge density analysis, so care has to be taken and these limitations recognised when analysing the data. It is possible to minimise some of these effects by the use of very high-energy synchrotron radiation, as demonstrated by recent studies on stishovite (SiO 2 ), cuprite (Cu 2 O) and YBa 2 Cu 3 O 6.98 , 1 although this is not yet a routine solution. Problems associated with thermal motion can also be minim- ised by conducting the experiments at ultra-low temperatures, but in minerals such as those being studied presently, thermal motion is generally sufficiently reduced at moderately low temperatures. An increasingly common method of analysing experimental and theoretical electron densities is via the topology of the electron density, derived from the “atoms in molecules” theory of Bader. 2 As remarked by one of us in a recent review of the literature, 3 this type of analysis is becoming the de facto standard in the field, especially applied to high-quality X-ray data. The topological analysis differs from the traditional deformation density, which is a representation of the way in which the electron density is distorted from that of a super- position of spherical atoms due to the effects of bonding. The latter method does not provide quantitative details about bond- ing, whereas topological analysis of the electron density allows straightforward extraction of this type of information. This type of analysis is important in the present study as quantitative information regarding the bonding in these two polymorphs is expected to provide an indication of how the interactions in the crystal might relate to the different reactivities of the two polymorphs. Experimental Crystallography and X-ray data collection Suitable crystals of senarmontite were prepared by sublimation of senarmontite powder at 600 °C under a nitrogen atmosphere in a Pyrex sublimation vessel. 4 Yields from the sublimation method were low and crystals of suitable size were only col- lected after repeating the process many times, each for periods of 5–7 days. Preliminary diffraction experiments showed that most crystals were affected by twinning and possessed high mosaicity, but extensive searching eventually yielded a suitable specimen on which the data was later collected. 5 Data collection was undertaken at the University of Sydney on a Bruker SMART 1000 CCD X-ray diffractometer. The crystal was attached with Exxon Paratone N to a short length of fibre supported on a thin piece of wire inserted in a steel mount- ing pin. The crystals were then quenched in a cold nitrogen gas stream from an Oxford Cryosystems Cryostream, while X-rays were produced from graphite monochromated, Mo-Kα radi- ation, generated from a sealed tube. Data collection was under- taken in three spheres with the camera at 30, 65 and 102° in 2θ and 4.0 cm from the sample. Each sphere was collected using ω scan increments of 0.2°, and with the  axis at 0, 120°, for the first sphere, 30, 150, 270° for the second and 90, 210 and 330° for the third. Exposure times were 10, 15 and 20 s, respectively, for each of the three camera positions. The first 50 frames of DOI: 10.1039/ b312550e 23 This journal is © The Royal Society of Chemistry 2004 Dalton Trans. , 2004, 23–29 Published on 24 November 2003. Downloaded by University of Queensland on 12/10/2015 03:29:32. View Article Online / Journal Homepage / Table of Contents for this issue