Dresden 2011 – wissenschaftliches Programm

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

O: Fachverband Oberflächenphysik

O 36: Poster Session II (Metals; Nanostructures at surfaces; Surface or interface magnetism; Spin-Orbit Interaction at Surfaces; Electron and spin dynamics; Surface dynamics; Methods; Theory and computation of electronic structure)

O 36.118: Poster

Dienstag, 15. März 2011, 18:30–22:00, P4

Neural Network Potential-Energy Surfaces for Organic Molecules: First Applications to Tartaric Acid — •Sinja Klees, Tobias Morawietz, and Jörg Behler — Lehrstuhl für Theoretische Chemie, Ruhr-Universität Bochum, D-44780 Bochum, Germany

First-principles calculations provide an accurate description of many systems. However, the computational costs severely limit the size of the systems that can be studied. Therefore, the development of efficient interatomic potentials is an important prerequisite for studying complex systems in molecular dynamics simulations. Neural Networks represent a class of flexible functions, which allows to construct accurate potential-energy surfaces (PESs) based on a set of reference points. To date, most Neural Network potentials have been reported for very small molecules or inorganic solids. Here, we investigate the applicability of this method to large organic molecules. Using tartaric acid as a benchmark system we show that a highly accurate PES can be constructed systematically.