Measurement-Driven Phase Transition within a Volume-Law Entangled Phase

TL;DR

This paper investigates a phase transition between two volume-law entangled phases in quantum systems with measurements, identifying a separability transition characterized by changes in entanglement and computational complexity.

Contribution

The study introduces an analytically solvable model revealing a new separability phase transition, and proposes the entangling power as an order parameter for this transition.

Findings

Identified a phase transition between fully-entangled and separable volume-law phases.

Established a connection between the transition and a mean-field percolation model.

Proposed a benchmarking task for quantum computers based on distribution deviations.

Abstract

We identify a phase transition between two kinds of volume-law entangled phases in non-local but few-body unitary dynamics with local projective measurements. In one phase, a finite fraction of the system belongs to a fully-entangled state, one for which no subsystem is in a pure state, while in the second phase, the steady-state is a product state over extensively many, finite subsystems. We study this "separability" transition in a family of solvable models in which we analytically determine the transition point, the evolution of certain entanglement properties of interest, and relate this to a mean-field percolation transition. Since the entanglement entropy density does not distinguish these phases, we introduce the entangling power - which measures whether local measurements outside of two finite subsystems can boost their mutual information - as an order parameter, after considering its behavior in tensor network states, and numerically studying its behavior in a model of Clifford dynamics with measurements. We argue that in our models, the separability transition coincides with a transition in the computational "hardness" of classically determining the output probability distribution for the steady-state in a certain basis of product states. A prediction for this distribution, which is accurate in the separable phase, and should deviate from the true distribution in the fully-entangled phase, provides a possible benchmarking task for quantum computers.