J. Appl. Cryst. (2000). 33, 953±957 H. Ehrenberg et al.  Resonant X-ray diffraction 953 research papers Journal of Applied Crystallography ISSN 0021-8898 Received 26 July 1999 Accepted 16 March 2000 # 2000 International Union of Crystallography Printed in Great Britain ± all rights reserved Resonant X-ray diffraction using high-resolution image-plate data H. Ehrenberg, a * M. Knapp, a T. Hartmann, a H. Fuess a and T. Wroblewski b a Institute for Materials Science, Darmstadt University of Technology, Petersenstr. 23, D-64287 Darmstadt, Germany, and b Hamburger Synchrotron Laboratory, Notkestr. 85, D-22607 Hamburg, Germany. Correspondence e-mail: helmut@tu-darmstadt.de The experimental setup for the collection of synchrotron X-ray powder diffraction data from samples with high absorption (R > 10) is described. It consists of a combination of a vacuum chamber with an image-plate system. A numerical absorption correction for the applied geometry has been derived and the data were corrected accordingly. Values for f 0 (Er) and f 00 (Er) were re®ned from eight measurements on Er 5 Re 2 O 12 above and below the Er L III absorption edge. Successful re®nement of the crystallographic data has veri®ed the high quality of the collected intensities. 1. Introduction The pronounced energy dependence of the X-ray scattering factor near absorption edges, f = f 0 + f 0 + if 00 , allows an element and oxidation-state speci®c variation of the real and imaginary parts of f in experiments at X-ray sources with variable wavelength. A summary of the theoretical back- ground and applications of resonant anomalous X-ray scat- tering can be found in work by Materlik et al. (1994). However, experiments on polycrystalline samples are often limited by severe absorption, especially at long wavelengths. Cox & Wilkinson (1994) recommend that the radius R of the capillary in a Debye±Scherrer experiment be chosen such that R does not exceed 2, a condition which cannot always be met, especially in the case of resonant experiments where the wavelength is ®xed by the absorption edge and only the radius R can be varied or the sample diluted. For much larger values of R, synchrotron radiation of high intensity and a detection method that allows the simultaneous collection of data at different diffraction angles need to be employed. At the same time, good resolution is required in order to minimize the correlation between re®ned parameters, especially in the range of pronounced overlap at high diffraction angles. At beamline B2 at HASYLAB in Hamburg, Germany, we have combined several experimental conditions to apply resonant X-ray scattering in Debye±Scherrer mode up to wavelengths of approximately 2.4 A Ê . This setup is described in this contribution. The handling and analysis of the raw data is outlined for Er 5 Re 2 O 12 , on which experiments near the Er L III edge at 8.3575 keV, i.e. 1.4835 A Ê , have been performed. 2. Experimental The powder diffractometer B2 is situated at a bending magnet of the storage ring DORIS-III. The vertical divergence of 0.4 mrad can be reduced by a collimating mirror, half way between the storage ring and a Ge(111) double-crystal monochromator (Ehrenberg et al., 1998). In addition to providing better resolution at high diffraction angles, a platinum-coated mirror and the Ge crystals suppress higher order harmonics very effectively and therefore no such contribution was detected. Diffraction patterns are collected in Debye±Scherrer mode with a rotating capillary in a vacuum chamber with pressure below 1 mbar (10 2 Pa) to avoid air scattering along the beam path (Gierlotka et al., 1997). This reduces the background and allows experiments up to a maximum wavelength of approximately 2.4 A Ê , limited by the incident spectrum. Three slit systems can be used for beam conditioning: slit 1 in front of the mirror, slit 2 behind the monochromator and slit 3 near the sample. The primary beam is cut by slit 1 to the size of the monochromating crystals, which reduces the background level. The beam size at the sample position is adjusted by slit 2, and the optimum setting is obtained when the beam covers all of the capillary. Slit 3 cuts off the radiation from the scattering at slit 2. An image-plate system collects data over a range of 2 = 135  : two image- plate stripes (5  40 cm) are situated in a cylindrical holder of 300 mm radius, covering the 2 ranges up to 61  , and from 63 to 135  , respectively. The blind angle between 61 and 63  is caused by a metal bridge utilized for mechanical stability against atmospheric pressure. However, the bridge position can be shifted by 1.5  by rotating the vacuum chamber. After exposure each stripe has to be carefully aligned on an image- plate scanner (Molecular Dynamics, Storm 820) to guarantee that the scanning directions are strictly parallel and perpen- dicular to the scattering plane. The spatial resolution of the scanner is 50 mm; therefore intensities are obtained in steps of approximately 2 = 0.00955  . Er 5 Re 2 O 12 was prepared by subsolidus reaction, as previously described for Tm 5 Re 2 O 12 (Ehrenberg et al., 1999). As an impurity, a very small amount (0.5 mol%) of Er 2 O 3 was identi®ed in the reaction product. The sample was ground and