Regensburg 2025 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 5: Poster

AKPIK 5.4: Poster

Donnerstag, 20. März 2025, 15:00–16:30, P2

Machine learning to resolve the structure: Perovskites and related materials — •Ekaterina Kneschaurek, Vladimir Starostin, Valentin Munteanu, Constantin Völter, Damian Balaz, Mikhail Romodin, Maik Hylinski, Dmitry Lapkin, Ivan Zaluzhnyy, Alexander Hinderhofer, Alexander Gerlach, and Frank Schreiber — University of Tübingen, Tübingen, Germany

Recent advances in synthesis of novel materials used in solar cells, such as perovskites, are supported by structural analysis, using X-rays. To unravel the complexity of such materials, we utilize both in situ (studying crystallization kinetics) and ex situ (accessing phase composition) X-ray diffraction. The high-resolution scattering patterns are recorded by large 2D detectors, resulting in enormous amounts of data, which are difficult to analyze manually [1]. In some cases, there are no crystallographic information files (.cif) for the newly synthesized materials. An initial guess, based on the composition of the studied material, can estimate ranges of expected scattering signals from distinct phases. The deep learning (DL) model can detect Bragg peaks within the diffraction pattern, while the application of clustering and 2D Gaussian fitting enables the processing of complex data to be conducted more effectively. Preliminary data analysis of 2D patterns in real time can increase the efficiency of beamtimes and provide a feedback loop to optimize the parameters of the experiment.

[1] Hinderhofer, A., Greco, A., Starostin, V., Munteanu, V., Pithan, L., Gerlach, A., Schreiber, F. (2023). J. Appl. Cryst. 56, 3-11.

Keywords: machine learning; perovskites; X-ray diffraction; 2D detectors; structural analysis