Applications of the generalized gradient approximation to ferroelectric perovskites

TL;DR

This paper evaluates the Perdew-Burke-Ernzerhof GGA functional's sensitivity to the parameter , demonstrating the need for system-dependent tuning to accurately predict properties of ferroelectric perovskites.

Contribution

It investigates how varying the  parameter affects the accuracy of density functional theory predictions for ferroelectric perovskites, highlighting the necessity for a system-specific approach.

Findings

Original  value works for some solids but not all

Adjusting  improves agreement with experimental lattice parameters

Sensitivity of structural properties to  underscores need for a universal function

Abstract

The Perdew-Burke-Ernzerhof generalized gradient approximation to the density functional theory is tested with respect to sensitivity to the choice of the value of the parameter $\kappa$, which is associated to the degree of localization of the exchange-correlation hole. A study of structural and dynamical properties of four selected ferroelectric perovskites is presented. The originally proposed value of $\kappa$=0.804 %(best suited for atoms and molecules) works well for some solids, whereas for the ABO$_3$ perovskites it must be decreased in order to predict equilibrium lattice parameters in good agreement with experiments. The effects on the structural instabilities and zone center phonon modes are examined. The need of varying $\kappa$ from one system to another reflects the fact that the localization of the exchange-correlation hole is system dependent, and the sensitivity of the structural properties to its actual value illustrates the necessity of finding a universal function for $\kappa$.