Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
O: Fachverband Oberflächenphysik
O 42: Focus Session: Innovation in Machine learning PRocEsses for Surface Science (IMPRESS)
O 42.6: Vortrag
Dienstag, 17. März 2020, 12:15–12:30, TRE Phy
(Re)interpreting TCNE adsorption on Cu(111) with machine learning — Alexander Egger1, Lukas Hörmann1, Andreas Jeindl1, Milica Todorovic2, Patrick Rinke2, and •Oliver T. Hofmann1 — 1TU Graz, Austria — 2Aalto University, Helsinki, Finland
Tetracyanoethylene (TCNE) layers on Cu(111) surfaces are a prototypical organic/inorganic interface. Measured vibrational spectra [1] of this interface evoke charge transfer between the substrate and molecules in the second layer and beyond. However, such “long-range” charge transfer defies our current understanding of organic/inorganic interfaces and is at variance with results from conventional density-functional theory (DFT) calculations.
In this work, we employ two new structure search methods, SAMPLE [2] and BOSS [3], that combine DFT with machine-learning algorithms, to computationally determine the geometric structure of TCNE mono- and bilayers on Cu 111. We then calculate vibrational spectra to compare to experiment. Our results show that the first TCNE layer re-orients with respect to the surface before molecules in the second layer are adsorbed. This reorientation changes the interaction with the surface qualitatively and brings the computed vibrational spectra into agreement with the measured ones. This structural phase change explains the observed spectral features without having to evoke long-range charge transfer. [1] Erley and Ibach, J. Phys. Chem. 91 2947 (1987) [2] Hörmann et al., CPC 244, 143 (2019) [3] Todorović et al., njp Comp. Mater. 5, 35 (2019)