Randomised measurements of a disorder-induced entanglement transition in a neutral atom quantum processor

TL;DR

This paper demonstrates a method to measure entanglement entropy in a neutral atom quantum simulator using randomised measurements, revealing a transition from chaos to localisation due to programmable disorder.

Contribution

It introduces a new protocol leveraging local energy tuning and global fields to measure entanglement without local gate control in an analogue quantum simulator.

Findings

Disorder induces a transition from chaotic to localised entanglement dynamics.

The method resolves disorder-specific entanglement spreading within decoherence times.

The approach extends the capabilities of programmable analogue quantum simulators.

Abstract

The development and spread of entanglement in complex quantum systems is central to exploring many-body phenomena out of equilibrium. Measuring entanglement dynamics can shed light on information scrambling and thermalisation, namely on transitions from many-body quantum chaos to localisation in disordered, interacting systems. In quantum computing systems, entanglement entropy and other nonlinear functions of the density matrix have been recently measured, in particular by using the randomised measurement toolbox. However, it is difficult to implement the required arbitrary unitary rotations on specific subsystems without universal local control. Here we devise and demonstrate the measurement of entanglement entropy in a programmable analogue quantum simulator using a randomised measurement protocol that leverages local energy tuning together with a global field to bypass the need for local gate control. We implement this on a commercially available neutral-atom quantum simulator, QuEra's Aquila, and use it to show how programmable disorder in the local Hamiltonian parameters leads to a transition from chaotic to localised entanglement dynamics. Given current decoherence times, we clearly resolve disorder-specific, time-dependent entanglement spreading in small systems. Our work extends the utility of programmable analogue quantum simulators, and opens further opportunities for wider randomised measurement toolboxes in a range of other analogue systems.