Many-body tunneling in a double-well potential

TL;DR

This paper introduces a new approach to evaluate Wannier functions emphasizing their tails, and demonstrates that nonstandard Hubbard terms significantly influence many-body tunneling dynamics in double-well potentials, especially at higher interactions.

Contribution

The paper develops an alternative method for Wannier functions focusing on tails and incorporates nonstandard Hubbard terms, enhancing the understanding of many-body tunneling beyond the standard Hubbard model.

Findings

Nonstandard Hubbard terms significantly alter tunneling dynamics.

Density-induced tunneling modifies the single-particle tunneling parameter.

Pair tunneling enables coherent propagation not captured by the standard model.

Abstract

We present an approach for evaluating Wannier functions, offering an alternative perspective on their role in many-body systems. Unlike traditional methods, such as the maximally localized Wannier functions approach, which focuses on minimizing the function tails, our approach emphasizes these tails. Using perturbative analytical approximations and extensive numerical simulations on an exactly solvable model, we address nonstandard Hubbard terms and demonstrate their critical influence on many-body dynamics. Specifically, we study tunneling dynamics in arbitrary double-well potentials, moving beyond the standard Hubbard model to include nonstandard terms such as density-induced tunneling and pair tunneling. Our results reveal that these terms significantly modify the dynamics predicted by the standard Hubbard model: density-induced tunneling modifies the single-particle tunneling parameter $\Omega_0$, while pair tunneling enables coherent propagation not captured by the standard model. We show that the discrepancies between the standard and nonstandard Hubbard models grow with increasing interaction strength, potentially leading to novel transport behaviors. However, at lower interaction strengths, both models converge, as nonstandard terms become negligible. These findings have important implications for phenomena such as superconductivity in twisted bilayer graphene and metal-insulator transitions. Our model aligns well with numerical simulations of lowest-band parameters and is strongly supported by experimental observations of second-order atom tunneling in optical double-well potentials. This strong agreement with experimental data highlights the accuracy and potential of our approach in providing a more comprehensive framework for describing complex many-body systems than the standard Hubbard model.