Correlation Analysis in X-ray Photoemission Spectroscopy

X-ray photoelectron spectroscopy (XPS) is a powerful technique for surface analysis, but such analysis can be hindered by uncertainty in modelling spectra. Often, many spectral models have a similar goodness of fit, and distinguishing between them can be impossible without additional information. A further challenge is found in interpreting spectra from samples consisting of multiple chemical compounds. We show here how correlation analysis can be used to interpret large XPS datasets. Correlations in atomic concentrations and binding energies of core lines can be interpreted within a framework of an underlying chemical model and this can yield additional information compared with analysis of each spectrum individually. We give examples of the usage of this analysis on some simple systems, and discuss the potential and limitations of the technique.