Automated optimisation of scanning parameters for SAXS/WAXS using machine learning for radiation-sensitive materials

This project aims at developing a machine learning (ML)-driven approach to automatically optimize X-ray scanning parameters regarding radiation sensitive samples for synchrotron-based scanning scattering (SAXS and WAXS) experiments. By adjusting and monitoring critical parameters, such as scan speed, orientation (meshing in y and x, spiral), distance between points, scan size and beam energy and flux, we aim to minimize radiation damage while ensuring the data quality during data collection. This methodology will be tested on radiation-sensitive model samples such as polysaccharides (i.e. cellulose, starch, and chitin), using the advanced capabilities of beamline ID13. Ultimately, the project will contribute to improving the accuracy of synchrotron experiments, particularly for radiation-sensitive samples.