RELIABILITY OF THE SIMPLIFIED INTEGRAL-BREADTH METHODS IN DIFFRACTIONLINE-BROADENING ANALYSIS

A comparison between different simplified integral-breadth methods, of ten used in the Rietveld-refinement programs to calculate coherent dom ain size and lattice strain, is carried out. It is shown that systemat ic differences exist for both domain size and strain, when they simult aneously broaden diffraction lines. Among different approximations, th e values of domain size exceptionally scatter and sometimes are comple tely false (negative or not real). A comparison to the alternative Fou rier method shows that all the simplified integral-breadth methods ove restimate domain size, but especially the Cauchy-Cauchy approximation. The root-mean-square strain and the upper limit of strain are related in the general case of the Voigt strain-broadened line profile. It is shown that they should not differ much as the profile changes between the Gauss and Cauchy extremes. The pure-Gauss size-broadened profile is incompatible with the definitions of surface-weighted domain size a nd column-length distribution function.