Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz
AKPIK 6: AI Methods for Materials Science
AKPIK 6.1: Vortrag
Donnerstag, 20. März 2025, 16:30–16:45, H5
Bayesian Optimization for High-Resolution Transmission Electron Microscopy — •Xiankang Tang1, Lei Jin2, Yixuan Zhang1, Rafal Dunin-Borkowski2, and Hongbin Zhang1 — 1Institute of Materials Science, TU Darmstadt, 64287 Darmstadt, Germany — 2Ernst Ruska-Centrum für Mikroskopie und Spectroskopie mit Elektronen, FZ Jülich, 52428 Jülich, Germany
Advances in machine learning technologies make it possible to automatize material characterizations, indispensable for the near-future implementation of autonomous experimentation for solid-state materials. High-resolution transmission electron microscopy (HRTEM) allows the study of the atomic structure of solid materials with sub-Angstrom resolution. By matching experimental and simulated images, unknown experimental parameters and crystal structures can be determined. However, this process entails strong domain expertise and can be time-consuming. In this work, we implement a Bayesian optimization-based approach to automatize the image analysis processes. Combined with some prior information, such as the experimentally measured aberration, the 3D crystal structure of the specimen can be reconstructed from a single HRTEM image. Specifically, by defining the mean square error(MSE) of pixel intensity values between experimental and simulated images, our method effectively captures both global and local image features. This approach not only achieves an exact match in absolute image contrast but also identifies unknown experimental parameters, optimizes atomic positions, and reveals surface morphology at atomic resolution.
Keywords: High-resolution transmission electron microscopy; Bayesian Optimization; 3D crystal structure reconstruction