The Interplay of Dynamics and Order in Mixed Perovskites

Perovskites with the general composition ABX₃ are among the most versatile functional materials known, finding applications including photovoltaics, solid-state electrolytes, and ferroelectrics. A unifying feature across these contexts is the presence of soft lattice dynamics, i.e., low-energy structural distortions that mediate phase transitions and couple strongly to electronic, optical, and transport properties. Alloying on the A, B, or X site provides a powerful handle for tuning these dynamics and the associated functional responses, but it simultaneously introduces chemical disorder and competing structural motifs whose interplay is often poorly understood.
Understanding the phase behavior of perovskites is critical for optimizing their structural stability and optoelectronic performance. In this context, I will present an analysis of the phase diagram of MA_1-xFA_xPbI₃ obtained using a machine-learned interatomic potential (MLIP) based on the neuroevolution potential framework. The results reveal a morphotropic phase boundary (MPB) at approximately 27% FA content, delineating the transition between out-of-phase and in-phase octahedral tilt patterns. Phonon mode projections show that this transition coincides with a mode crossover composition, where the free energy landscapes of the underlying phonon modes become nearly degenerate. The results provide a systematic and consistent description of this important system, complementing earlier partial and sometimes conflicting experimental assessments. Furthermore, density functional theory calculations show that band edge fluctuations peak near the MPB, indicating an enhancement of electron–phonon coupling and dynamic disorder effects. These findings establish a direct link between phonon dynamics, phase behavior, and electronic structure, providing a further composition-driven pathway for tailoring the optoelectronic properties of perovskite materials. By demonstrating that phonon overdamping serves as a hallmark of the MPB, this study offers insights into the design principles for stable, high-performance perovskite solar cells.
Extending this perspective to the chalcogenide perovskite BaZrS_3xSe_3-3x, I will discuss results that show anion alloying gives rise to an unusual trans-type ordering of S and Se atoms, most pronounced at 33% S, which persists near room temperature and is confirmed by scanning transmission electron microscopy. Free-energy calculations yield the full temperature–composition phase diagram and demonstrate that composition, crystal structure, and anion ordering jointly control the optical band gap, with the order–disorder transition shifting the Tauc gap by approximately 0.16–0.19 eV.
Finally, to enable predictive simulations of perovskites where long-range electrostatics and polarization play a central role — including ferroelectrics, ionic conductors, and reactive interfaces — I will also present qNEP, a charge-aware extension of the neuroevolution potential framework. By incorporating explicit, environment-dependent partial charges treated via a particle–particle particle–mesh scheme, qNEP provides direct access to Born effective charges, dielectric response, and infrared spectra at a computational cost only modestly above that of a standard short-range potential, enabling simulations of million-atom systems on nanosecond time scales.