Self-organized error correction in random unitary circuits with measurement

Random measurements have been shown to induce a phase transition in an extended quantum system evolving under chaotic unitary dynamics, when the strength of measurements exceeds a threshold value. Below this threshold, a steady state with a subthermal volume law entanglement emerges, which is resistant to the disentangling action of measurements, suggesting a connection to quantum error-correcting codes. We quantify these notions by identifying a power-law decay of the mutual information I ({x} : (A) over bar) alpha x(-3/2) in the volume-law-entangled phase, between a qudit located a distance x from the boundary of a region A, and the complement (A) over bar, which implies that a measurement of this qudit will retrieve very little information about (A) over bar. We also find a universal logarithmic contribution to the volume law entanglement entropy S-(2)(A) = kappa L-A + 3/2 log L-A which is intimately related to the first observation. We obtain these results by relating the entanglement dynamics to the imaginary time evolution of an Ising model, to which we apply field-theoretic and matrix-product-state techniques. Finally, exploiting the error-correction viewpoint, we assume that the volume-law state is an encoding of a Page state in a quantum error-correcting code to obtain a bound on the critical measurement strength pc as a function of the qudit dimension d: p(c) log[(d(2) - 1)(p(c)(-1)-1)] <= log[(1 - p(c))d]. The bound is saturated at p(c) (d ->infinity) = 1/2 and provides a reasonable estimate for the qubit transition: p(c) (d = 2) <= 0.1893.

Automated summary

Random measurements induce a phase transition in an extended quantum system under chaotic dynamics, resistant to disentangling action when strength exceeds a threshold. The study quantifies notions by power-law decay of mutual information and logarithmic contribution to entanglement entropy. Utilizing an error-correction viewpoint, a bound on critical measurement strength is obtained as a function of qudit dimension, relevant for qubit transition estimates.