The analysis of mixtures : Application of principal component analysis to XAS spectra

Many samples which are subjected to XAS analysis contain the element of interest in more than one chemical form. The interpretation of the spectra from such samples is often not straightforward, particularly if appropriate model systems are not available. We have applied principal component analysis (PCA) to real and simulated systems which contain mixtures of several species for a given element PCA has been extensively used for the analysis of other types of spectra, including MS, IR, and UV-VIS. The application of PCA to XAS is illustrated by examining the speciation of iron within coals. PCA can determine how many different species that contain a particular element are present in a series of spectra. In tandem with model compounds, principal component analysis can suggest which of the models may contribute to the observed XAS spectra.