Dresden 2026 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
MM: Fachverband Metall- und Materialphysik
MM 8: Materials for the Storage and Conversion of Energy II
MM 8.2: Talk
Monday, March 9, 2026, 16:00–16:15, SCH/A216
Voltage Profile Predictions for Na-Ion Cathodes: A Hybrid MLIP / PBEsol+U+V approach — •Stefan Schären1,2, Valentina Sanella1,3, Manuel Dillenz4, Nicola Marzari2, and Iurii Timrov1 — 1PSI, Villigen, Switzerland — 2EPFL, Lausanne, Switzerland — 3ETHZ, Zürich, Switzerland — 4DTU, Copenhagen, Denmark
Sodium-ion batteries are emerging as a sustainable alternative to lithium-ion batteries, but are limited by the lack of a cathode that offers high voltage, energy density and long-term stability. Here, we present a computational approach that accelerates the prediction of voltage profiles during cathode discharge, using a combination of self-consistent Hubbard-corrected DFT (PBEsol+U+V), and foundational machine-learning interatomic potentials (MLIPs). Using NaxV2(PO4)3 (NVP) as a test case, we show that MLIP-based pre-screening reduces the number of required DFT calculations to a fraction of those needed for a traditional cluster expansion, while retaining comparable accuracy. Furthermore, the self-consistent Hubbard U and V corrections yield electronic occupations that closely match the chemical intuition for oxidation states, enabling a direct analysis of vanadium redox activity and associated volume changes during desodiation. Altogether, the approach delivers a useful and predictive toolkit for identifying high-performance Na-ion cathodes in the future.
Keywords: Sodium-ion Batteries; Hubbard-corrected DFT; NaSICON; MLIP; Voltage Profiles