Avoiding surface and absorption effects in XRD quantitative phase analysis

The Rietveld method has modified the approach used for quantitative phase analysis. No external or internal standards are needed and the method suffers from instrumental and sample preparation errors to a lower extent. Moreover in the Rietveld method there are some problems affecting the modal analysis; in particular the most pervasive are the preferred orientations, the absorption contrast effects (sometimes called microabsorptions) and the sample artefacts (differences between surface and bulk). Analytical theories have been reported to correct all these cases [1-3], but a knowledge of the microstructure is necessary to avoid misleading analyses. These errors are more severe in X-ray diffraction than in the neutron case, due to the difference in sample representability. In XRD, by selecting the more appropriate theory and wavelength, most of the errors can be corrected or reduced. Examples of ceramic matrix composites toughened by a ductile metallic phase are reported. The large, highly absorbing, metal inclusions can spread over the surface during the cutting and polishing procedures of the composite specimens. The deformation process may develop also a small amount of preferred orientation. XRD analyses by the Rietveld method show how the combined use of MoKα and CuKα radiations with simple absorption and preferred orientation corrections drastically reduces the error in quantitative phase determination. It was found, for both radiations, that the choice of the correct absorption theory and microstructural parameters are determinant to the analysis results.