Static vs. dynamical mean field theory of Mott antiferromagnets

TL;DR

This paper compares static and dynamical mean field theories in describing Mott antiferromagnets, revealing significant differences in spectral properties and bandwidth, with experimental validation through photoemission data.

Contribution

It demonstrates that dynamical mean field theory captures important spectral features and renormalizations missed by static mean field approaches in Mott antiferromagnets.

Findings

Dynamical mean field theory shows a strongly renormalized Slater band.

Spectral weight shifts to spin-polaron side bands.

Experimental spectra of Cr-doped V₂O₃ align with dynamical mean field theory predictions.

Abstract

Studying the antiferromagnetic phase of the Hubbard model by dynamical mean field theory, we observe striking differences with static (Hartree-Fock) mean field: The Slater band is strongly renormalized and spectral weight is transferred to spin-polaron side bands. Already for intermediate values of the interaction $U$ the overall bandwidth is larger than in Hartree-Fock, and the gap is considerably smaller. Such differences survive any renormalization of $U$. Our photoemission experiments for Cr-doped V$_2$O$_3$ show spectra qualitatively well described by dynamical mean field theory.