Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
HL: Fachverband Halbleiterphysik
HL 13: Poster I
HL 13.39: Poster
Monday, March 18, 2024, 15:00–18:00, Poster E
Exploring anharmonicity in metal halide double perovskites using machine-learned ACE potentials — •Mattis Gossler and Bernd Meyer — Interdisciplinary Center for Molecular Materials and Computer Chemistry Center, FAU Erlangen-Nürnberg, Germany
Metal halide perovskites (MHPs) have gained much attention for their exceptional photo-electrical properties, making them ideal for photovoltaic applications. Seeking eco-friendly alternatives to traditionally lead-based MHPs, double-cation perovskites A2IMIMIIIX6 offer fine-tuned photo-electrical properties through their variable composition.
MPHs are softer compared to other inorganic semiconductors, displaying significant structural fluctuations at room temperature. Understanding their electronic and structural properties involves anharmonic vibrational modes, best explored through ab initio molecular dynamics (AIMD). To overcome the severe simulation limitations of AIMD, we employ an on-the-fly active-learning workflow to train an atomic cluster expansion (ACE) machine-learned interatomic potential with minimal human intervention. The capabilities of the generated ACE potential are then tested to investigate structural and dynamic properties of the double perovskite Cs2AgBiCl6 from the CANBIC family at elevated temperatures, which are still ambiguous from experimental data.
Keywords: machine learning; atomic cluster expansion; metal-halide double perovskites