Phase composition analysis by Rietveld analysis: when are we going to get serious about absolute determinations?

Full text: Most of the phase composition analysis results reported in the literature use so-called 'standardless phase composition analysis' which is a misnomer for 'relative phase composition analysis'. In this method, the ZMV formalism is used to generate relative weight concentrations ci and cj for phases 'i' and 'j' according to ci/cj = [si/sj].[(ZMV)i/(ZMV)j]...(1) where symbol 's' represents Rietveld scale factors, 'Z' is the number of phase formula units of weight 'M' per unit cell volume 'V'. Commonly, the relative concentrations are converted to 'standardless' absolute concentrations with the following formula ..Cj={si(ZMV)i / ΣSj(ZMV)j]} x 100 %...(2) whereby it is implicitly assumed that (i) systematic errors in the diffraction data and refinement model are negligible, and (ii) that amorphous material and undetected phases do not contribute significantly to the total composition. Use of formula (2) in effect 'forces' the sum of the reported concentrations to 100% and, in so doing, tends to mask systematic errors which bias the scale factors, eg due to microabsorption, preferred orientation and 2θ-dependent intensity errors. The development of procedures for absolute phase composition assessments comparable in accuracy with that typically achieved in elemental analysis (< 0.1% relative) requires the use of calibration standards, diffraction data of adequate quality and high quality refinements. When an internal standard is blended with the material, equation, ci [si/ss].[(ZMV)j/(ZMV)s]].cs... (3a) is used; subscript 's' signifying the internal standard. When an internal standard cannot be added to the material (eg with a solid ceramic or metal specimen), then equation ci [si/ss].[(ZMV)j/(ZMV)s].[μs/μi]. cs...(3b) is employed, where symbol '1' represents the phase mass attenuation coefficient. It is hoped that the Rietveld analysis practitioners will devise methods involving a suite of external standards and the use of equation (3b) to achieve the reliability of analyses attainable in XRF spectrometry. Copyright (1999) Australian X-ray Analytical Association Inc