Publication date: 22nd July 2026
Hybrid lead halide perovskites (HLPs) have emerged as one of the most important classes of materials in the field of photovoltaics due to their exceptional optoelectronic properties and compositional flexibility. Yet, the thermal and chemical material instability, along with the toxicity of lead cations, remain open problems that require a precise control and understanding of the crystal growth and microstructure at the atomic scale. In this talk, we will discuss recent progress in physics-based models [1,2] as well as advanced machine learning approaches [3] for the large-scale molecular dynamics simulations of bulk, surfaces and interfaces. Specifically, we will report on the development of a new MYP2[4] model that, by the introduction of many-body interatomic terms, enables the simulation of crystal growth. This provides information on kinetics, activation energies, morphology and defects formation[5,6] while also allowing the study of complex 2D/3D interfaces and heterostructures in solution. We will conclude with a perspective on the evolving role of molecular simulation in the era of machine learning.
Funding through project CNR DIMENSIONING, computation support from CINECA through ISCRA Initiative (IscraB CCG-PVOX, IscraC - NEUROMYP)
