抄録
We have developed a new deconvolution method for eliminating instrumental aberrations from powder X-ray diffraction data. The method is based on the analytical model of the instrumental aberrations and fast Fourier transform calculation. It is much more advantageous than conventional methods, because it does not require measurement of a reference sample nor complicated analytical processes. The method is applied to the diffraction data of fine SiC powder sample with broad crystallite size distribution. The deconvoluted data clearly showed symmetric diffraction peak profiles with sharp peak top and long tails, which are characterized by a ′super-Lorentzian′ line shape. Assuming spherical crystallite shape with log-normal size distribution, the logarithmic (relative) distribution width has been estimated at 0.93(3), which is close to the value 0.97, evaluated by a laser diffraction method. The frequency of the deformation-type stacking fault along the 111-direction is also estimated at 0.005 from systematic shifts of the diffraction peaks.
