The determination of chromium speciation in solid samples is critical for environmental and industrial purposes. Several analytical methods exist to perform such a determination either directly in solid state or liquid state after an extraction step, each of them having some limitations. In this study, the use of a high-resolution wavelength-dispersive X-ray fluorescence spectrometer to determine and quantify chromium species is investigated by looking at the differences in the Kβ transition profiles between Cr(0), Cr(III) and Cr(VI) compounds. Three different approaches were tested and compared to determine the Cr(VI) fraction of known mixtures: relative height and peak fitting using calibration mixtures, partial least square regression (PLS) of pure compounds, and principal component regression (PCR) of pure compounds. The accuracy of these methods was found to be about the same with an average relative error in the range of 15%. However, PLS and PCR can be easily implemented in an automated way contrary to peak fitting which can be sometimes perceived as analyst-dependant. Another advantage of using PLS and PCR is that information concerning the other oxidation states present in the sample can be retrieved. Finally, PLS and the peak height approach can be used up to 0.5% total chromium which make the XRF an alternative technique to X-ray induced photoelectron spectroscopy (XPS) for chromium speciation in solid state.
Published by Elsevier B.V.