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

O: Fachverband Oberflächenphysik

O 34: Poster: Surface Dynamics & Electron-Driven Processes

O 34.1: Poster

Dienstag, 19. März 2024, 18:00–20:00, Poster C

Targeted growth of metastable interface polymorphs — •Simon B. Hollweger, Anna Werkovits, and Oliver T. Hofmann — Institute of Solid State Physics, TU Graz, Austria

Growing a specific interface polymorph of organic molecules on inorganic substrates is far from trivial. The target structure is usually not the one that is thermodynamically most favorable, but rather an energetically higher-lying, metastable structure. It is primarily unclear how the growth conditions need to be tuned to maximize the yield of a higher-energy polymorph. These conditions are usually explored experimentally by a tedious trial-and-error routine. To avoid this obstacle we theoretically investigate the growth behavior of such systems. We discuss a class of systems where we can start from a lower-energy polymorph and through systematic changes of pressure and temperature induce a structural change to the desired target structure. For that two requirements need to be fulfilled, (a) the metastable target structure has a similar packing density as the initial structure and (b) an 'elevator' polymorph exists that has a higher packing density and is thermodynamically accessible at lower temperatures. To explore in which range of substrate-molecule and molecule-molecule interactions the necessary conditions are fulfilled we employ a combination of kinetic Monte Carlo simulations and ab-initio thermodynamics. This allows us to formulate structure-to-property relationships. For actually determining 'growth recipes' for a specific target structure we utilize Deep Reinforcement Learning combined with kinetic Monte Carlo simulations to obtain optimal temperature and pressure curves.

Keywords: growth; kinetic Monte Carlo; Deep Reinforcement Learning; polymorphism