Towards an automated quality control of XAS data

The quality of fluorescence X-ray absorption spectroscopy (XAS) data strongly depends on the identification and elimination of contributions suffering from artificial deviations. To enhance detection of deviations, XAS data are converted here to difference spectra and cumulative difference spectra. A variety of statistical criteria and procedures are examined for their application in the quality control of such data. The criterion best suited in this case is determined and a strategy for the automatic elimination of artefacts is developed: deviation-affected spectra are iteratively removed from the data pool. A threshold is defined to avoid unnecessary reduction of the experimental data pool. Exemplarily the procedure is applied for the quality control of BioXAS data.