(16h) Transferable Bond Valence Interatomic Potentials for the Accurate and Reliable Nanoscale Simulation of Ferroelectric Solid Solutions

Ferroelectrics and their solid solutions, e.g., Ba_xSr_1-xTiO₃ (BST), find application in a number of important technologies ranging from heterogeneous catalysis (e.g., direct NO_x decomposition and the selective partial oxidation of methane) to electroacoustic transduction (e.g., microphones and SONAR) and random-access memory. The design of next-generation materials for these applications, however, relies on the development of simulation methods that are quantum-mechanically accurate, reliable (i.e., reproduce experiments), and efficient for the nanoscale modeling of solid solutions. Here, we develop a transferable interatomic potential for simulating BST (i.e., works for any x from BaTiO₃ to SrTiO₃) with up to millions of atoms and over several microseconds. Moreover, this tool builds upon the bond valence molecular dynamics method, which has been used successfully in the past to study structural phase transitions, the nucleation of ferroelectric domains, and the dynamics of the walls separating these domains. A key feature of this tool is that it accurately models the temperature-composition phase diagram of BST and reveals the role of Sr-induced dipole scatter in the reduction of global electric polarization and the enhancement of the order-disorder character of the ferroelectric-paraelectric phase transition. Due to its success, this tool lays the foundation for investigating and designing state-of-the-art alloys for ferroelectric catalysis, transduction, and computer memory.[1]

[1] R. B. Wexler, Y. Qi, and A. M. Rappe, Phys. Rev. B 100, 174109 (2019) DOI: https://doi.org/10.1103/PhysRevB.100.174109