Regensburg 2022 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 12: Computational Materials Modelling: Physics of Ensembles 1

MM 12.1: Talk

Tuesday, September 6, 2022, 10:15–10:30, H44

Pinpointing Hubbard corrections to tackle inhomogenous nl subshells: The DFT+U(m) method — •Eric Macke¹, Iurii Timrov², Lucio Colombi Ciacchi¹, and Nicola Marzari² — ¹Hybrid Materials Interfaces Group and Bremen Center for Computational Materials Science, MAPEX, University of Bremen, Am Fallturm 1, 28359 Bremen, Germany — ²Theory and Simulation of Materials (THEOS) and National Centre for Computational Design and Discovery of Novel Materials (MARVEL), Ecole Polytechnique Federale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland

DFT+U remains a key tool in computational material science as it mitigates the DFT self-interaction error. While this simplistic approach often provides electronic and magnetic properties in fair agreement with literature, recent investigations revealed that DFT+U massively over-stabilizes high spin configurations of transition metal elements surrounded by strong ligand fields. In such compounds, the projection of occupied Kohn-Sham states onto the atomic basis frequently yields occupation numbers n_m^σ≈ 0.5 for certain magnetic quantum numbers m, so that the penalizing Hubbard term E_U = ∑_m,σ U/2 [n_m^σ(1−n_m^σ)] is maximized. We argue that if the occupation of a magnetic (spin-) orbital is dominated by the effect of hybridization with a ligand, Hubbard corrections should not be applied in the same way as for unaffected orbitals. To account for such inhomogeneous bonding regimes, we propose a more flexible scheme that enables the use of distinct Hubbard parameters U(m) for the same species, computable by means of ab initio methods.