Combining semi-local exchange with dynamical mean-field theory: electronic structure and optical response of rare-earth sesquioxides

TL;DR

This paper introduces a combined ab initio dynamical mean-field theory and semi-local exchange approach to accurately model the electronic structure and optical properties of rare-earth sesquioxides, capturing complex 4f state behaviors.

Contribution

The study develops a novel hybrid computational method that effectively describes both localized 4f electrons and extended states in rare-earth sesquioxides, improving upon previous models.

Findings

Successfully reproduces the evolution of the optical gap across the RE series.

Identifies the role of 2p-4f hybridization in shifting 4f states within the gap.

Links optical conductivity features to 4f states within the p-d gap.

Abstract

In rare-earth semiconductors, wide ligand $p$ and rare-earth 5$d$ bands coexist with localized, partially filled 4$f$ shells. A simultaneous description for both extended and localized states represents a significant challenge for first-principles theories. Here, we combine an {\it ab initio} dynamical mean-field theory approach to strong local correlations with a perturbative application of the semi-local modified Becke-Johnson exchange potential to correct the semiconducting gap. We apply this method to calculate the electronic structure and optical response of the light rare-earth sesquioxides RE$_2$O$_3$ (RE= La, Ce, Pr and Nd). Our calculations correctly capture a non-trivial evolution of the optical gap in RE$_2$O$_3$ due to a progressive lowering of the 4$f$ states along the series and their multiplet structure. 2$p$ $-$ 4$f$ hybridization is found to induce a substantial upward shift for the occupied 4$f$ states occurring within the $p-d$ gap, thus reducing the magnitude of the optical gap. We show that a characteristic plateau observed in the optical conductivity in the Pr and Nd sequioxides right above their absorption edge is a fingerprint of 4$f$ states located within the $p-d$ gap.