Measuring correlated electron dynamics with time-resolved photoemission spectroscopy

TL;DR

This paper investigates the capabilities and limitations of time-resolved photoemission spectroscopy in capturing the dynamics of correlated electrons, highlighting the challenges in reconstructing Green functions and identifying observable phenomena like Mott insulator oscillations.

Contribution

It provides a theoretical analysis of how time-resolved photoemission spectra relate to electronic Green functions in correlated systems, revealing what can and cannot be inferred from experimental data.

Findings

Real-time Green functions during metallic state buildup cannot be directly extracted from photoemission signals.

Collapse-and-revival oscillations in Mott insulators can be observed as spectral weight oscillations.

The study uses the Falicov-Kimball model within dynamical mean-field theory to simulate these phenomena.

Abstract

Time-resolved photoemission experiments can reveal fascinating quantum dynamics of correlated electrons. However, the thermalization of the electronic system is typically so fast that very short probe pulses are necessary to resolve the time evolution of the quantum state, and this leads to poor energy resolution due to the energy-time uncertainty relation. Although the photoemission intensity can be calculated from the nonequilibrium electronic Green functions, the converse procedure is therefore difficult. We analyze a hypothetical time-resolved photoemission experiment on a correlated electronic system, described by the Falicov-Kimball model in dynamical mean-field theory, which relaxes between metallic and insulating phases. We find that the real-time Green function which describes the transient behavior during the buildup of the metallic state cannot be determined directly from the photoemission signal. On the other hand, the characteristic collapse-and-revival oscillations of an excited Mott insulator can be observed as oscillating weight in the center of the Mott gap in the time-dependent photoemission spectrum.