Regensburg 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
O: Fachverband Oberflächenphysik
O 85: New Methods: Theory
O 85.6: Talk
Thursday, March 20, 2025, 11:45–12:00, H25
Machine Learning for Polaronic Materials: TiO2(110) at the nanoscale — •Firat Yalcin1, Simon Trivisonne1, Viktor Birschitzky1, Carla Verdi2, and Michele Reticcioli1,3 — 1University of Vienna, Austria — 2University of Queensland, Australia — 3CNR-SPIN L’Aquila, Italy
The combination of machine learning (ML) with density functional theory accelerates material simulations, expanding both spatial and temporal scales. However, current ML methods struggle to address polaron trapping. We present a novel machine learning force field (MLFF) approach that incorporates polaron trapping descriptors, enabling large-scale studies of polaronic materials. Using TiO2(110) as a case study, we reveal how Nb dopants and oxygen vacancies affect polaron configurations and drive catalytic CO adsorption. Additionally, our method captures the dynamic evolution of polarons with unprecedented statistical robustness. This work advances fundamental understanding of defect-polaron interactions while offering a fully-automated and efficient computational suite for the study of polaronic materials.
Keywords: polaron; machine learning; molecular dynamics; DFT; defects