Modelling charge self-trapping in wide-gap dielectrics: Localization problem in local density functionals

TL;DR

This paper investigates the challenges of modeling charge self-trapping in wide-gap dielectrics using density functional theory, highlighting the dependence on exchange-correlation functionals and the limitations of Kohn-Sham approaches.

Contribution

It demonstrates how the stability of self-trapped states varies with the amount of exact exchange in hybrid functionals and discusses the limitations of standard DFT in capturing self-trapping phenomena.

Findings

Self-trapped exciton not observed in plane-wave DFT calculations.

Stability of self-trapped hole depends on the exact exchange percentage.

Kohn-Sham methods inadequately represent key contributions to self-trapping energy.

Abstract

We discuss the adiabatic self-trapping of small polarons within the density functional theory (DFT). In particular, we carried out plane-wave pseudo-potential calculations of the triplet exciton in NaCl and found no energy minimum corresponding to the self-trapped exciton (STE) contrary to the experimental evidence and previous calculations. To explore the origin of this problem we modelled the self-trapped hole in NaCl using hybrid density functionals and an embedded cluster method. Calculations show that the stability of the self-trapped state of the hole drastically depends on the amount of the exact exchange in the density functional: at less than 30% of the Hartree-Fock exchange, only delocalized hole is stable, at 50% - both delocalized and self-trapped states are stable, while further increase of exact exchange results in only the self-trapped state being stable. We argue that the main contributions to the self-trapping energy such as the kinetic energy of the localizing charge, the chemical bond formation of the di-halogen quasi molecule, and the lattice polarization, are represented incorrectly within the Kohn-Sham (KS) based approaches.