Factors affecting data reproducibility on an area detector

The reproducibility of data collected on a Seimens/ Nicolet area detector has been tested with respect to crystal orientation, X-ray collimation and scaling procedure. In addition, a comparison between area detector data and diffractometer data was made in order to determine the accuracy of the area detector data. The reproducibility of the area detector data, although acceptable, is somewhat affected by crystal orientation (shape effect) and appears to be insensitive to X-ray collimation, contrary to what would be expected. The scaling procedures used in the tests were XENGEN [Howard, Gilliland, Finzel, Poulos, Ohlendorf & Salemme (1987). J. Appl. Cryst. 20, 383-387] (one, two and three parameters), FBSCALE [Weissman (1982). In Computational Crystallography, edited by D. Sayre, pp. 56-63. Oxford: Clarendon Press] and SIMPLAD [Takusagawa (1987). J. Appl. Cryst. 20, 243-245]. It was found that, although these procedures do an adequate job of scaling area detector data, none can totally correct for crystal-shape effects. From our experience, XENGEN multi-parameter or SIMPLAD scaling on data sets having low redundancy must be used with care. When compared to diffractometer data, area detector data scaled by any of the above procedures is in good agreement, with merging R-factors (on F) in the range 3.45 to 4.55%. A plot of <F_ad> versus <F_d> shows a straight line with slope of 1(3)%. No saturation problems were observed.