Can correlations drive a band insulator metallic?

TL;DR

This paper investigates how on-site Coulomb repulsion affects a band insulator, revealing a transition to a metallic phase before becoming a Mott insulator, challenging traditional Hartree-Fock predictions.

Contribution

It demonstrates, using DMFT, that Coulomb interactions can induce a metallic state in a band insulator, a phenomenon not captured by Hartree-Fock theory.

Findings

Gap closes at critical Uc1, leading to a metallic phase

Transition from metal to Mott insulator at Uc2

Results differ from Hartree-Fock predictions of monotonic gap decrease

Abstract

We analyze the effects of the on-site Coulomb repulsion U on a band insulator using dynamical mean field theory (DMFT). We find the surprising result that the gap is suppressed to zero at a critical Uc1 and remains zero within a metallic phase. At a larger Uc2 there is a second transition from the metal to a Mott insulator, in which the gap increases with increasing U. These results are qualitatively different from Hartree-Fock theory which gives a monotonically decreasing but non-zero insulating gap for all finite U.