The polarization factor for coherent scattering from a crystalline or non-crystalline sample can be calculated easily for any number of pre-and post-sample scatterers of any orientation. No complicated visualization is necessary. Either an analytical or a numerical method can be used. Simple repeated matrix multiplication is used to determine the polarization factor. The technique is illustrated for a Bonse-Hart instrument having 13 scattering processes.
IntroductionWhen only one or two scattering processes are considered, the ideal polarization factor often can be derived easily considering the orientation and components of the electric vector before and after each scattering process. The ideal polarization factor is the usual one used in kinematical diffraction theory. For more complicated cases the geometry can become very difficult to visualize, and analytical equations for the ideal polarization factor P can become very complicated. Vincent (1982) derived equations for calculating the ideal polarization factor for any number of pre-sample scatterers, with the restriction that all scattering planes are either normal or parallel to all other scattering planes for all of the pre-sample scatterers. He indicates that equations can be derived for other orientations of planes, but the resulting equations become complicated. Post-sample scatterers were not considered.A general method for determining the ideal polarization factor has existed for many years. The origin of it is over a century old (Stokes, 1852). The Stokes parameters describe the state of polarization of radiation. Since this early work, a large number of persons have developed methods for calculating ideal polarization factors based on matrix operations on the Stokes parameters. McMaster (1961) 0108-7673/83/050773-05501.50 subject and gave several references. The matrix method has seldom been used in X-ray scattering studies, although it was used to derive the intensity of secondary scattering in non-crystalline materials (Dwiggins & Park, 1971;Dwiggins, 1972). An advantage of using such a matrix method is that either numerical or analytical expressions can be determined using only simple matrix multiplication.In this paper the matrix method is adapted for X-ray scattering studies used for investigations of structure. Only coherent radiation is considered at first, but incoherent radiation will be discussed. Although X-ray tubes produce unpolarized characteristic radiation, completely polarized radiation is allowed because synchrotron radiation that is nearly completely polarized now is used for some studies (Vincent, 1982). Circular polarization is not allowed, although it is possible to include it, because the resulting equations would be much more complicated and because highintensity X-ray sources with circular polarization are not yet available.When strongly diffracting crystals are used as monochromators, as often is the case in single-crystal structure studies, dynamical rather than kinematical theory applies and the degree of perfec...