Researchers at Chalmers University of Technology and the University of Birmingham have used advanced machine learning-driven molecular dynamics simulations to elucidate the low-temperature phase of formamidinium lead iodide (FAPbI3).
Their study identifies the detailed crystal structure and octahedral tilt pattern characteristic of this phase and reveals that the rotational dynamics of formamidinium (FA) cations become arrested in a metastable configuration upon cooling. This cation freezing explains the persistent experimental challenges in accessing the thermodynamic ground state of FAPbI3.
The work leverages machine-learned interatomic potentials to perform large-scale molecular dynamics simulations encompassing millions of atoms and extending to simulation timescales thousands of times longer than conventional methods, thus providing an unprecedented atomic-level description.
Experimental validation through nuclear magnetic resonance (NMR) and inelastic neutron scattering (INS) at cryogenic temperatures (~-200°C) confirms the simulation predictions, establishing strong agreement between theory and experiment.
These insights into the low-temperature structural and dynamical behavior of FAPbI3 are vital for rational design and stabilization of halide perovskite materials, including mixed-halide compositions, to improve phase stability and device performance. This fundamental understanding advances the engineering of efficient, cost-effective, and durable perovskite solar cells, addressing the critical need for scalable renewable energy technologies.
