Many-body effects in TiSe2: Can GW describe an Excitonic Insulator?

TL;DR

This study combines ab initio GW calculations with experimental ARPES and STS measurements to investigate whether GW can accurately describe the excitonic insulator phase in TiSe2, revealing a band gap and renormalized band shape consistent with excitonic effects.

Contribution

The paper demonstrates that GW calculations can reproduce key features of TiSe2's excitonic insulator phase, bridging theoretical predictions with experimental observations.

Findings

GW predicts a ~0.2 eV band gap in TiSe2.

Band shape is strongly renormalized, showing a mexican hat feature.

Theoretical results agree well with experimental spectra.

Abstract

We present both theoretical ab initio GW and experimental angle-resolved photoemission (ARPES) and scanning tunneling (STS) spectroscopy results on TiSe2. With respect to the density-functional Kohn-Sham metallic picture, the many-body GW self-energy leads to a ~ 0.2 eV band gap insulator consistent with our STS spectra at 5 K. The band shape is strongly renormalized, with the top-of-valence moved towards a circle of points away from \Gamma, arising in a mexican hat feature typical of an excitonic insulator. Our calculations are in good agreement with experiment.