Roberto Car is the Ralph W. Dornte Professor in Chemistry at Princeton University, where he is also a faculty member in the Princeton Institute for the Science and Technology of Materials. He conducts research on the simulation of molecular dynamics phenomena.. Professor Car studied physics and attained a doctorate in 1971 in nuclear technology at the Politecnico di Milano. He was a postdoc at the University of Milan and an assistant at the École Polytechnique Fédérale de Lausanne (EPFL). He was employed at the Thomas J. Watson Research Center of IBM, and he was associate professor for physics at SISSA in Trieste, and professor for physics at the University of Geneva (and director of the IRRMA of EPFL). He is a professor in the Theory Department, of the Fritz Haber Institute of the Max Planck Society.
Abstract:
Studies of the ferroelectric phase transition with the Deep Potential model
In the Deep Potential (DP) framework, the dependence on the atomic coordinates of the potential energy and of the dielectric polarization is represented by deep neural networks trained on density functional theory (DFT) data. The model retains quantum mechanical accuracy at nearly the cost of empirical force fields, opening the way to large scale molecular dynamics (MD) simulations with ab-initio predictive power. Here, I report recent studies of the ferroelectric phase transition made possible by DPMD. Two prototypical perovskite materials are considered: lead titanate (PTO) and lead magnesium niobate (PMN). In crystalline PTO the ferroelectric phase transition exhibits order-disorder and soft mode features. PMN is a chemically disordered material whose dielectric response exhibits glassy behavior, called relaxor, that bears similarity with magnetic spin glasses and fragile structural glasses. In both cases DPMD simulations predict finite temperature properties in close agreement with experiment, providing fresh insight into the microscopic mechanisms behind the structural and dielectric behavior. In PTO, the simulations reveal that the spectral behavior attributed to soft mode derives from relaxational disorder of the cell dipoles. In PMN, the inherent potential landscape gives clues on why relaxors behave like magnetic spin glasses and fragile structural glasses.
Supported by the DOE Award DE-SC0019394
