Disordered insulator in an optical lattice

TL;DR

This study experimentally investigates the disordered Bose-Hubbard model using optical speckle, revealing a disorder-induced superfluid-to-insulator transition and providing new insights into the interplay of disorder and interactions in quantum systems.

Contribution

First experimental realization of the disordered Bose-Hubbard model with controlled disorder, observing a disorder-induced insulator transition and constraining theoretical models.

Findings

Disorder induces a superfluid-to-insulator transition at high disorder strengths.

No evidence found for an insulator-to-superfluid re-entrant transition.

Increased disorder correlates with greater dissipation, excluding re-entrant superfluid predictions.

Abstract

Disorder can profoundly affect the transport properties of a wide range of quantum materials. Presently, there is significant disagreement regarding the effect of disorder on transport in the disordered Bose-Hubbard (DBH) model, which is the paradigm used to theoretically study disorder in strongly correlated bosonic systems. We experimentally realize the DBH model by using optical speckle to introduce precisely known, controllable, and fine-grained disorder to an optical lattice5. Here, by measuring the dissipation strength for transport, we discover a disorder-induced SF-to-insulator (IN) transition in this system, but we find no evidence for an IN-to-SF transition. Emergence of the IN at disorder strengths several hundred times the tunnelling energy agrees with a predicted SF--Bose glass (BG) transition from recent quantum Monte Carlo (QMC) work. Both the SF--IN transition and correlated changes in the atomic quasimomentum distribution--which verify a simple model for the interplay of disorder and interactions in this system--are phenomena new to the unit filling regime explored in this work, compared with the high filling limit probed previously. We find that increasing disorder strength generically leads to greater dissipation in the regime of mixed SF and Mott-insulator (MI) phases, excluding predictions of a disorder-induced, or "re-entrant," SF (RSF). While the absence of an RSF may be explained by the effect of finite temperature, we strongly constrain theories by measuring bounds on the entropy per particle in the disordered lattice.