Experimental Validation of Quantitative XANES Analysis for Phosphorus Speciation

The quantitative approach used in x-ray absorption spectroscopy (XAS) experiments is oftentimes based on statistical goodness-of-fit criteria, which do not explain the accuracy of the components obtained from the fittings. This study was performed to validate the linear combination (LC) approach used in quantitative XAS analysis by estimating the accuracy of this procedure. Near-edge Kα1 fluorescence XAS spectra were acquired for known binary mixtures of Ca, Al, and Fe phosphates in varying proportions and for the individual compounds. All combinations of the spectra of model compounds were fitted to the spectra of the known mixtures to obtain their relative abundance. The binary combinations produced the best fit with chi2 values ranging from 0.02 to 0.25. The relative error associated with the fitting ranged from as low as 0.8 to 17% for thoroughly mixed samples. The relative error was small when the proportion of Ca phosphate in the mixture was high but the error was large at low abundance of this component in the mixture. Because the interpretation of the XANES result largely depends on the relative proportion of species in the sample obtained by LC, we therefore recommend acquiring a spectrum for a mixture of certified reference compounds that mimics the composition of the sample being investigated at the beamline to estimate the accuracy of the proportions obtained from quantitative x-ray absorption near-edge structure (XANES) analysis.