Regensburg 2019 – wissenschaftliches Programm

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

O: Fachverband Oberflächenphysik

O 68: Poster Wednesday: Nanostructures

O 68.1: Poster

Mittwoch, 3. April 2019, 17:45–20:00, Poster B2

Simulation of Brass Nanoparticles using a Neural Network Potential — •Jan Weinreich, Martín Leandro Paleico, and Jörg Behler — Universität Göttingen, Institut für Physikalische Chemie, Theoretische Chemie, Tammannstr. 6, 37077 Göttingen, Germany

High-dimensional neural network potentials (HDNNPs) are a powerful method to study complex systems, because they offer the accuracy of density functional theory (DFT) calculations at the computational costs of simple empirical potentials thus enabling large-scale Monte Carlo and molecular dynamics simulations. In this work we construct a HDNNP for brass nanoparticles, which are important industrial catalysts for methanol synthesis. The potential is applied to determine the most stable structures of large brass clusters of varying composition and size. The analysis of the surface structures emerging in Monte Carlo simulations reveals that depending on the system size the copper to zinc ratio in the surface layers can strongly differ from the interior of the particles.