SKM 2023 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

TT: Fachverband Tiefe Temperaturen

TT 68: Focus Session: Making Experimental Data F.A.I.R. – New Concepts for Research Data Management II (joint session O/TT)

TT 68.9: Talk

Friday, March 31, 2023, 12:00–12:15, WIL A317

Deep learning surface scattering data analysis for processing large synchrotron datasets — •Vladimir Starostin, Valentin Munteanu, Linus Pithan, Alexander Gerlach, Alexander Hinderhofer, and Frank Schreiber — Institute of Applied Physics, University of Tübingen, Germany

In situ real-time surface scattering experiments such as grazing-incidence wide-angle X-ray scattering (GIWAXS) produce large amounts of data, frequently exceeding the capabilities of traditional data processing methods. Here we demonstrate an automated pipeline for the analysis of GIWAXS images, based on a machine learning architecture for object detection, designed to conform to the specifics of the scattering data [1]. Our pipeline enables real-time GIWAXS analysis and is designed to be employed at synchrotron facilities. We also present FAIR data strategies and traceable data resources from the raw data to the corresponding scientific publication and vice versa [2] including intermediate processing steps.

We demonstrate our method on real-time tracking of lead halide perovskite structure crystallization processes, which are relevant for hybrid solar cell applications. However, our approach is equally suitable for other crystalline thin-film materials by design. In general, the solution substantially accelerates the analysis process of GIWAXS images, potentially boosting the speed of scientific discoveries in material science.

[1] V. Starostin et al. npj Comput Mater 8, 101 (2022)

[2] V. Starostin et al. Synch Rad News 13, 31–37 (2022)