Mott metal-insulator transition from steady-state density functional theory

TL;DR

This paper introduces a computationally efficient i-DFT method to determine spectral functions and capture the Mott metal-insulator transition in bulk systems, using an idealized STM setup and exchange-correlation bias approximations.

Contribution

The paper develops a novel i-DFT approach to accurately model spectral functions and the Mott transition, incorporating exchange-correlation effects based on Fermi-liquid theory.

Findings

Successfully captures the Mott metal-insulator transition in different lattice structures.

Provides a practical method to extract spectral functions from differential conductance.

Demonstrates the effectiveness of exchange-correlation bias approximations in i-DFT.

Abstract

We present a computationally efficient method to obtain the spectral function of bulk systems in the framework of steady-state density functional theory (i-DFT) using an idealized Scanning Tunneling Microscope (STM) setup. We calculate the current through the STM tip and then extract the spectral function from the finite-bias differential conductance. The fictitious non-interacting system of i-DFT features an exchange-correlation (xc) contribution to the bias which guarantees the same current as in the true interacting system. Exact properties of the xc bias are established using Fermi-liquid theory and subsequently implemented to construct approximations for the Hubbard model. We show for two different lattice structures that the metal-insulator transition is captured by i-DFT.