Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
CPP: Fachverband Chemische Physik und Polymerphysik
CPP 30: Emerging Topics in Chemical and Polymer Physics, New Instruments and Methods I
CPP 30.2: Vortrag
Mittwoch, 19. März 2025, 16:45–17:00, H34
Deep Learning-Driven GISAXS Data Processing for Nanostructure Characterization — •Yufeng Zhai1, Shachar Dan1, Julian Heger2, Peter Müller-Buschbaum2, and Stephan Roth1,3 — 1Deutsches Elektronen-Synchrotron (DESY), Hamburg, Notkestr. 85, Germany — 2Technical University of Munich, TUM School of Natural Sciences, Department of Physics, Chair for Functional Materials, Garching 5, Germany — 3Royal Institute of Technology (KTH), Stockholm, Sweden
Nanostructured materials, are at the forefront of advanced applications in various fields, owing to their unique physical and chemical properties. Grazing incidence small-angle X-ray scattering (GISAXS) has emerged as a powerful technique for probing the morphology of these nanostructures, offering valuable insights into electron density distributions both at the surface and within thin films. In our approach, we first simulate GISAXS pattern using the Distorted Wave Born Approximation (DWBA) model to generate high-quality training datasets. We then apply deep learning techniques, specifically convolutional neural networks (CNNs), to predict size distributions from GISAXS data. Our results demonstrate that CNNs are highly robust under varying noise conditions and present a promising, time-efficient approach for overcoming the challenges of conventional scattering analysis. This study highlights the potential of integrating advanced computational methods and new analytical tools to enhance the characterization of nanostructures.
Keywords: GISAXS; Deep Learning