Implementation of self-consistent MGGA functionals in augmented plane wave based methods

TL;DR

This paper presents a self-consistent implementation of meta-GGA functionals within augmented plane wave methods, demonstrating improved accuracy in electronic property calculations while maintaining computational efficiency.

Contribution

It introduces a novel self-consistent implementation of MGGA functionals in APW methods, enabling more accurate electronic structure simulations.

Findings

Accurate calculation of band gaps, lattice constants, and magnetic moments.

Validation of implementation through electric field gradient calculations.

Demonstrates the efficiency of MGGA in solid-state simulations.

Abstract

Functionals of the meta-generalized gradient approximation (MGGA) are nowadays widely used in chemistry and solid-state physics for the simulation of electronic systems like molecules, solids, or surfaces. Due to their dependency on the kinetic energy density, they are in principle more accurate than GGA functionals for various properties (geometry, binding energy, electronic structure, etc.), while being nearly as fast since they are still of the semilocal form. Thus, when an accuracy better than GGA is required, one may consider using a MGGA instead of the much more costly hybrid functionals or methods like the random-phase approximation or $GW$. In this work, the self-consistent implementation of MGGA functionals in APW based methods is presented. Technical aspects of the implementation are discussed, and calculations of band gaps, lattice constants, and magnetic moments are presented in order to validate our implementation. To test the changes of the electron density due to a MGGA, the electric field gradient on transition-metal atoms is calculated.