The quantitative determination of the crystalline and amorphous content by the Rietveld method: application to glass ceramics with different absorption coefficients

The Rietveld method provides an accurate determination of the crystalline and the amorphous fractions in polyphase mixtures. To determine the content of the amorphous phase, the sample is diluted with an internal standard which is considered as a component itself and refined with the other phases. Although the method is accurate for systems containing crystalline and amorphous phases with an absorption coefficient comparable to that of the standard, it was not thoroughly tested for systems containing a weakly or highly absorbing amorphous phase. Commonly, this is the case for glass ceramics which may contain elements with fairly different atomic number.X-ray powder diffraction data refined with the Rietveld method were utilized for the determination of the amorphous content in a number of systems containing glasses with different absorption coefficients (Li-, B-, Al-, P-, Zn-, Zr-, Ba-, Pb-glass) to check the accuracy of the method. The mixtures were prepared diluting the glass phase with corundum NIST 676 and hydroxyapatite NIST 2910 in fixed proportions. The refinements were carried out with GSAS and the new code QUANTO, especially developed for automatic quantitative phase analysis (QPA) of polycrystalline mixtures eventually including an amorphous component.The glass weights estimated by the two programs are both in good agreement with the nominal value (45 wt%) although a minor dependency of the calculated amorphous fraction on the linear absorption coefficient of the glass was observed and attributed to microabsorption effects. The Brindley (1945) correction applied here according to the formulation of Taylor and Matulis (1991) revealed not to be fully satisfactory for mixture containing an amorphous component with a very different absorption coefficient, especially when highly accurate determinations are required. This is in concert with the general concern that the existing models for the correction of microabsorption effects should be revised as they are not adequate even for systems containing only crystalline phases (Madsen et al., 2001).The method was successfully applied to a commercial glass ceramic indicating that it can be applied with a good accuracy also to glass ceramics with fairly different absorption coefficients.