Collisionless drag for a one-dimensional two-component Bose-Hubbard model

TL;DR

This paper theoretically explores the collisionless drag phenomenon in a one-dimensional two-component Bose-Hubbard model, revealing conditions under which the drag becomes significant and its impact on correlation functions and spin dynamics.

Contribution

It provides the first detailed tensor network analysis of collisionless drag in a 1D Bose-Hubbard model, linking drag to the pair-superfluid phase and long-range correlations.

Findings

Drag can become comparable to superfluid density near the pair-superfluid phase.

Large drag influences the long-range behavior of correlation functions.

Mesoscopic systems favor large drag, avoiding BKT transition effects.

Abstract

We theoretically investigate the elusive Andreev-Bashkin collisionless drag for a two-component onedimensional Bose-Hubbard model on a ring. By means of tensor network algorithms, we calculate the superfluid stiffness matrix as a function of intra- and interspecies interactions and of the lattice filling. We then focus on the most promising region close to the so-called pair-superfluid phase, where we observe that the drag can become comparable with the total superfluid density. We elucidate the importance of the drag in determining the long-range behavior of the correlation functions and the spin speed of sound. In this way, we are able to provide an expression for the spin Luttinger parameter $K_S$ in terms of drag and the spin susceptibility. Our results are promising in view of implementing the system by using ultracold Bose mixtures trapped in deep optical lattices, where the size of the sample is of the same order of the number of particles we simulate. Importantly, the mesoscopicity of the system, far from being detrimental, appears to favor a large drag, avoiding the Berezinskii-Kosterlitz-Thouless jump at the transition to the pair-superfluid phase which would reduce the region where a large drag can be observed.