Many-body Theory vs Simulations for the pseudogap in the Hubbard model

TL;DR

This paper compares many-body theoretical approaches and Monte Carlo simulations to understand the pseudogap phenomenon in the half-filled two-dimensional Hubbard model, highlighting discrepancies and limitations of certain theories.

Contribution

It demonstrates that self-consistent Eliashberg-type theories fail to reproduce the pseudogap observed in Monte Carlo simulations, emphasizing the need for more accurate many-body techniques.

Findings

Monte Carlo simulations show a pseudogap in the spectral weight.

Eliashberg-type theories do not reproduce the pseudogap.

Differences between precursor and strong-coupling pseudogaps are discussed.

Abstract

The opening of a critical-fluctuation induced pseudogap (or precursor pseudogap) in the one-particle spectral weight of the half-filled two-dimensional Hubbard model is discussed. This pseudogap, appearing in our Monte Carlo simulations, may be obtained from many-body techniques that use Green functions and vertex corrections that are at the same level of approximation. Self-consistent theories of the Eliashberg type (such as the Fluctuation Exchange Approximation) use renormalized Green functions and bare vertices in a context where there is no Migdal theorem. They do not find the pseudogap, in quantitative and qualitative disagreement with simulations, suggesting these methods are inadequate for this problem. Differences between precursor pseudogaps and strong-coupling pseudogaps are also discussed.