The determination of amorphous by quantitative Rietveld analysis is in principal a very straightforward procedure: A certain amount of a fully crystalline standard is added and thoroughly mixed with the sample, X-ray powder data are collected and a quantitative Rietveld analysis is performed. The deviation of found and actual weight percentages of the standard can be used to compute the amorphous content of the sample. Occasionally, even that last step is already included in the Rietveld software.
However, recent investigations have shown that this method can give misleading results due to several factors which can distort the results. For one thing fully crystalline standards are much harder to obtain than assumed previously. This is best illustrated by the case of the NIST standard SRM676 which was originally assumed to be a fully crystalline material and therefore an ideal internal standard for quantitative Rietveld analysis. As of 2005 NIST admits an amorphous content of approximately 8 % which originally passed unnoticed. More seriously, microabsorption effects have been found to heavily distort the results of quantitative Rietveld analyses. This effect can under ideal circumstances be corrected with the Brindley correction. Usually, however, the premises of the Brindley correction such as spherical particle size and homogeneous particle distribution are not met and employing the Brindley correction may yield even less reliable results than overlooking the microabsorption problem altogether.
We propose to avoid these problems by developing a new set of internal standards based on ultrahard carbides. These materials are well crystalline and show a clear cleavage. Therefore amorphization on grinding is unlikely. By using carbides with a wide range of chemical composition and therefore linear absorption coefficients we aspire to minimize the absorption contrast. On this occasion we will present the first results of this approach to tackle the problems associated with the internal standard method. |