Dresden 2026 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
TT: Fachverband Tiefe Temperaturen
TT 54: Graphene: Electronic structure, excitations, etc. (joint session O/TT)
TT 54.4: Vortrag
Mittwoch, 11. März 2026, 15:45–16:00, HSZ/0204
Machine Learning for Bandstructure-Structure Relationships in Doped Graphene — •Benedict Saunders1, Lukas Hörmann1,2, Valdas Vitartas1, Chen Qian1, and Reinhard J Maurer1,2 — 1University of Warwick, Coventry, United Kingdom — 2University of Vienna, Vienna, Austria
The introduction of topological defects or dopants to graphene's honeycomb lattice is a common approach to tuning the electronic and transport properties of the material to suit a specific application. Nitrogen is a widely investigated dopant, which can be introduced into the lattice in various configurations, each with a distinct effect on the material's bandstructure. However, due to the combinatorial explosion of possible dopant configurations, conventional first-principles screening of the electronic bandstructure for all possible configurations is not feasible. Here, we address this limitation by employing the recently developed MACE-H machine learning model to predict the electronic Hamiltonian of nitrogen-doped graphene configurations directly for a series of previously generated defect configurations. This allows us to rapidly generate accurate bandstructures and densities-of-states for large numbers of configurations, enabling the identification of structure-property relationships as a function of temperature and composition.
Keywords: Graphene; Electronic structure; Machine learning; Neural network; Configurational explosion