Effect of Strong Correlations on the Disorder-Induced Zero Bias Anomaly in the Two-Site Anderson-Hubbard Model
Hong-Yi Chen, W. A. Atkinson

TL;DR
This paper investigates how strong electron correlations influence the formation of a disorder-induced zero bias anomaly in a simplified two-site Anderson-Hubbard model, revealing different underlying mechanisms in weakly and strongly correlated regimes.
Contribution
It provides an analytical study of the ZBA formation in a two-site model, highlighting the role of many-body effects in the strongly correlated limit, which were previously not well understood.
Findings
In weak correlations, ZBA arises from level repulsion between molecular orbitals.
In strong correlations, ZBA is mainly due to suppression of triplet excitations.
Many-body effects are crucial for understanding ZBA in strongly correlated systems.
Abstract
Several recent exact diagonalization calculations have established that the Anderson-Hubbard model has a disorder-induced zero bias anomaly (ZBA) (also called a disorder-induced pseudogap) in the density of states. In order to understand the physics of the ZBA, we study a simplified problem---an ensemble of two-site molecules with random site energies---for which analytical results are possible. For this ensemble, we examine how the ZBA forms in both the weakly correlated (mean field) and strongly correlated limits. In the weakly correlated case, the ZBA can be understood as the result of level repulsion between bonding and antibonding molecular orbitals. A similar level repulsion occurs in the strongly correlated case too, but a larger contribution to the ZBA comes from the suppression of a triplet excitation mode. This inherently many-body mechanism does not have a counterpart in…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
