The structures of micro-crystalline compounds can sometimes be solved directly from their synchrotron powder diffraction patterns. However, the odds decrease as peaks broaden, as unit cell dimensions and the size of the asymmetric unit increase, and as the structural symmetry decreases. Many of the compounds which have a great deal of interest to DuPont seem to fall into one or more of these categories. Unfortunately, experience has shown that solving structures using techniques derived from single-crystal methods simply do not work very well with powder data taken on samples with the above "complications". For the powder technique to be generally useful as a routine method of crystal structure analysis, new techniques for solving structures need to be found.

Attendence at various crystallographic meetings has suggested that a variety of new techniques are being explored. Unfortunately, many of them appear to have serious limitations. In an effort to determine which of these techniques might be truly useful for DuPont's problems, the Corporate Center for Analytical Science (CCAS) has set aside a US$ 1000 prize to the first person/group who can solve the crystal structure of HAlF4 [Chem. of Materials, 7, 75 (1995)]. The preparation of this compound involves moderate heating of aluminum fluoride salts which contain organic cations. Typical of compounds made by thermal decomposition, the peaks in the x-ray powder diffraction are somewhat broad. Attempts to anneal the compound at higher temperatures results only in further decomposition to AlF3, so there is very little chance of improving the crystallinity of the material. Two high-resolution, synchrotron x-ray powder diffraction patterns on this material (one with and one without preferred orientation); neither has produced a structure using the traditional techniques available at DuPont. The challenge is now open to the crystallographic community.

Details: HAlF4; very thin, flat-plate crystallites; monoclinic with approximate cell parameters of a = 8.28, b = 6.20, c = 10.58 A, beta = 103.22 deg (a and b have been confirmed by electron diffraction from a pattern which also shows mirror symmetry; you're welcome to reindex the pattern); peaks are consistent with space groups C2, Cm, and C2/m; NMR indicates that all Al are octahedral; from known structures, Al octahedra may be corner-sharing (most common) or edge-sharing [far less common, but known [Amer. Chem. Soc., 115, 3028 (1993); Inorg. Chem., 32, 2985 (1993)]. Two sets of data were collected and have been posted ( as files of x-y pairs where x = two-theta and y = the normallized intensity:

(1) 1.0 mm capillary; wavelength = 0.69955 A; data is somewhat noisy but has little (no?) preferred orientation; X3b beamline at NSLS; two-theta range, 3 - 50.17 deg; filename = half4_x3b.xy.

(2) flat-plate sample, wavelength = 0.84878 A; data has preferred orientation effects - 00l reflections are emphasized; X7a beamline at NSLS; two-theta range, 4 - 60 deg; filename = half4_x7a.xy.

The correctness of any proposed structural model will be tested against a neutron diffraction pattern (HFBR - HRNPD) which is not being made available to the contestants. The first model which makes chemical sense (coordination numbers and bond values in agreement with previously reported results and pass I. D. Brown's bond-valence sum test) and which matches both the x-ray and neutron diffraction patterns [with R(Bragg) below 0.08] will be declared the winner and the authors of the model will be entitled to the US$ 1000. Proposed structural solutions should be sent to Richard L. Harlow, CRD, E228/316d, The DuPont Company, Wilmington, DE, 19880-0228, USA. Good Luck.

Two files of powder patterns are available for download by clicking below.



Return to the ACA Service Crystallography Homepage!!