SKM 2023 – wissenschaftliches Programm

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

MM: Fachverband Metall- und Materialphysik

MM 16: Energy Conversion

MM 16.6: Vortrag

Dienstag, 28. März 2023, 11:45–12:00, SCH A 215

Modelling Hard Carbon with High-Dimensional Neural Network Potentials — •Alexander L. M. Knoll^{1, 2}, Jafar Azizi³, Holger Euchner⁴, Axel Groß³, and Jörg Behler^{1, 2} — ¹Lehrstuhl für Theoretische Chemie II, Ruhr-Universität Bochum, 44780 Bochum, Germany — ²Research Center Chemical Sciences and Sustainability, Research Alliance Ruhr, 44780 Bochum, Germany — ³Institut für Theoretische Chemie, Universität Ulm, 89081 Ulm, Germany — ⁴Institut für Physikalische und Theoretische Chemie, Eberhard Karls Universität Tübingen, 72076 Tübingen, Germany

Hard carbon (HC) is widely regarded as a promising anode material for lithium and sodium ion batteries. However, this material is notoriously difficult to characterise experimentally, as its properties strongly depend on the carbon precursor and the conditions of preparation. Ab initio investigations could help to illuminate its structure, but are severely hampered by its complexity involving large graphitic domains as well as amorphous motifs. To bypass these challenges, we present a high-dimensional neural network potential (HDNNP) for HC that accurately reproduces short-range and dispersion interactions in the system. We show that elaborate structural models can be constructed with this potential using various simulation protocols and demonstrate their validity through comparison to experimental data.