BACKGROUND CORRECTION OF NEUTRON AND X-RAY FIBER DIFFRACTION DATA. Ivanova, M. and Makowski, L., Chemistry Department, Florida State University, Tallahassee, FL 32306-3006, USA

Background estimation is the limiting factor in processing of fiber diffraction data. Background in fiber diffraction pattern is not readily measurable except at resolution low enough that adjacent layer lines do not overlap. Estimation at larger diffraction angles is problematic since there is no way to reliably identify a point on the pattern where the only contribution is the background. Where this positions can be identified, it is possible to estimate the background by fitting the observed background points to a set of cylinder functions, A parameterization of the background can also be integrated to estimation of the layer line intensities bay angular deconvolution. This works for cases where the background can be well represented by a limited number of cylinder functions. To address the more general problem we used the observation that the spatial frequencies making up the background are much lower than those that constitute the data, and used iterative low pass filtering to separate the contributions from the layer line data and from the background. This process has been integrated into angular deconvolution to provide accurate estimates for data from x-ray and neutron diffraction patterns that could not be successfully processed by previous methods. We have applied the method successfully to x-ray diffraction patterns from mutant filamentous bacteriophage fddx and neutron diffraction patterns from specifically deuterated M13 phage particles.