Spectral decomposition of x-ray absorption spectroscopy datasets: Methods and applications

X-ray absorption spectroscopy (XAS) today represents a widespread and powerful technique, able to monitor complex systems under in situ and operando conditions, while external variables, such us sampling time, sample temperature or even beam position over the analysed sample, are varied. X-ray absorption spectroscopy is an element-selective but bulk-averaging technique. Each measured XAS spectrum can be seen as an average signal arising from all the absorber-containing species/configurations present in the sample under study. The acquired XAS data are thus represented by a spectroscopic mixture composed of superimposed spectral profiles associated to well-defined components, characterised by concentration values evolving in the course of the experiment. The decomposition of an experimental XAS dataset in a set of pure spectral and concentration values is a typical example of an inverse problem and it goes, usually, under the name of multivariate curve resolution (MCR). In the present work, we present an overview on the major techniques developed to realize the MCR decomposition together with a selection of related results, with an emphasis on applications in catalysis. Therein, we will highlight the great potential of these methods which are imposing as an essential tool for quantitative analysis of large XAS datasets as well as the directions for further development in synergy with the continuous instrumental progresses at synchrotron sources.