Dresden 2020 – wissenschaftliches Programm
Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
TT: Fachverband Tiefe Temperaturen
TT 58: Graphene I: Growth, Structure and Substrate Interaction (joint session O/TT)
TT 58.6: Vortrag
Donnerstag, 19. März 2020, 11:45–12:00, GER 37
Simulating the scattering of a hydrogen atom from graphene using a high-dimensional neural network potential. — •Sebastian Wille1,2, Marvin Kammler2, Martín L. Paleico3, Jörg Behler3, Alec M. Wodtke1,2, and Alexander Kandratsenka2 — 1Institute for Physical Chemistry, Georg-August University Göttingen, Germany — 2Department of Dynamics at Surfaces, Max Planck Institute for Biophysical Chemistry, Göttingen, Germany — 3Theoretical Chemistry, Georg-August University Göttingen, Germany
Understanding the formation of covalent bonds due to atomic-scale motions and energy dissipation pathways involved is an ongoing challenge in the field of chemistry. Recent measurements of the translational energy loss distribution of hydrogen atoms scattered from graphene showed a bimodal pattern. The second generation reactive empirical bond order potential was fitted to ab initio electronic structure data obtained from embedded mean-field theory to generate a potential energy surface (PES). First-principles dynamics simulations using the provided PES were able to reproduce the bimodal feature of the energy loss spectrum and were in qualitative agreement with experimental results. But these investigations could not fully provide a detailed description of the scattering and sticking mechanisms. Therefore, we developed a full-dimensional neural network PES by fitting to the density functional theory data in order to further reduce the remaining errors by the fitting procedure of the PES underlying molecular dynamics simulations performed.