Approche de la forme des cristallites par la diffraction des rayons X

Abstract

The work of this thesis concerns the determination of the morphology of the crystallites in a polycrystalline compound. Different shapes for the crystallites can be processed, it is the spherical shapes, cylindrical and parallelepiped. These morphologies were studied and translated as a program (DSIZE) by Langford. The objective of this paper was to achieve the treatment of prismatic hexagonal base, because we felt that this morphology often appears in polycrystalline compounds which possess hexagonal symmetry. Many questions are asked, then resolved during the work among which we mention the phase of the understanding and exploitation of various relationships used in this method of microstructural analysis. Thus, a lot of these relationships were demonstrated and several numeric values and curves related to these relationships were found, in order to go on knowledge devoid of any uncertainty. The principle of determining the morphology of the crystallites is that of least squares. After studying the method and the elements it uses, it was about building a hexagonal prism processing routine basis in accordance with the target set. This treatment has been integrated in Langford program which we used as a working basis, so that different symmetries can be exploited by the user complementarity in a single unit. The result of work has been verified by the realization of many theoretical and experimental cases tests and by the subsequent treatment of two morphologies, cylindrical and prismatic hexagonal base, whenever this was possible, allowing for comparisons and, therefore, confirm the results. Tests showed the proper functioning of the treatment used. At the end of the work, the aspects related to the convergence of computing and also limiting the range of validity of the initial values corresponding to the shape parameters were considered: the convergence took place for each of the examples treated with the values far enough to the initial values.