Dresden 2026 – scientific programme

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

MM: Fachverband Metall- und Materialphysik

MM 5: Topical Session: Physics-driven Artificial Intelligence for Materials I

MM 5.2: Talk

Monday, March 9, 2026, 10:45–11:00, SCH/A251

Screening of high-entropy oxides as oxygen conductors for fuel cells — •Jesper R. Pedersen, Ciku Parida, Benjamin H. Sjølin, and Ivano E. Castelli — Department of Energy Conversion and Storage, Technical University of Denmark, Kgs. Lyngby 2800, Denmark

High-entropy materials are at the cutting edge of materials design and this work investigates their suitability for use in fuel cells. However, screening high-entropy materials is computationally demanding, especially for properties such as oxygen migration barriers which are critical for components used in fuel cell cathodes. We aim to address this challenge by fine-tuning the MACE foundation model to a generated database containing more than 400 nudged elastic band (NEB) calculations spanning the chemical space of double perovskites. We show this new model achieves accuracy similar to the DFT calculations not only on structural relaxations, but also for the more complex task of saddle-point discovery. This enables the continued rapid exploration of the chemical space beyond the initial screening, leading to new insights into factors governing oxygen migration in fuel cell cathodes.

Keywords: Fuel cell; High-entropy oxides; Data-driven discovery; High-throughput screening; Machine learning interatomic potential