From first- to second-principles
modelling of ferroelectric perovskites:
a historical perspective

Professor Philippe Ghosez
Theoretical Materials Physics, Q-MAT, CESAM, University of Liège, Belgium
12th November 2020 16:00 GMT+2

Please register: https://us02web.zoom.us/meeting/register/tZModOysrT8uGNN_X2c-Kd6pLyqJtwsyoUNS

Abstract: Since the early 1990s, many efforts have been devoted to the first-principles modelling of functional oxides from density functional theory. This concerned, at first, bulk compounds and later on included thin films, superlattices and various kinds of heterostructures. Initially restricted to reproduce experimental findings, such calculations progressively became predictive. They were useful to identify and discuss various properties (ferroelectricity, piezoelectricity, multiferroism), and to rationalize their tuning by external constraints. However, such studies were mostly restricted to zero Kelvin and small systems. During the 90’s, Zhong, Vanderbilt and Rabe pioneered the development of so-called “effective Hamiltonians” directly fitted on first-principles data and providing access to the finite temperature properties of ferroelectric perovskites. This method is still very popular nowadays and relies on a projection within a restricted subspace including only the most relevant lattice degrees of freedom. Later, Wojdel et al. generalized the approach in order to include all lattice degrees of freedom while still determining all parameters from first-principles data, in a so-called second-principles framework as recently implemented in MULTIBINIT. Going further, such lattice models also remain compatible with the explicit treatment of additional relevant electronic degrees of freedom. Here, I will first review how the theory and modelling of ferroelectrics has evolved since the early stages, one century ago. Then, through selected examples, I will discuss the new perspectives offered by recent second-principles approaches.