Rigorous wave function embedding with dynamical fluctuations

TL;DR

This paper introduces a wave function-based method to efficiently capture dynamical fluctuations in strongly correlated systems, offering a computationally cheaper and systematically improvable alternative to traditional dynamical mean-field theory.

Contribution

The authors develop a frequency-independent, zero-temperature wave function approach that models environmental fluctuations in correlated systems, improving computational efficiency and convergence.

Findings

Successfully applied to Bethe lattice Hubbard model

Demonstrated rapid convergence in 1D Hubbard chain

Offers benefits over traditional DMFT in computational cost

Abstract

The dynamical fluctuations in approaches such as dynamical mean-field theory (DMFT) allow for the self-consistent optimization of a local fragment, hybridized with a true correlated environment. We show that these correlated environmental fluctuations can instead be efficiently captured in a wave function perspective in a computationally cheap, frequency-independent, zero-temperature approach. This allows for a systematically improvable, short-time wave function analogue to DMFT, which entails a number of computational and numerical benefits. We demonstrate this approach to solve the correlated dynamics of the paradigmatic Bethe lattice Hubbard model, as well as detailing cluster extensions in the one-dimensional Hubbard chain where we clearly show the benefits of this rapidly convergent description of correlated environmental fluctuations.