Effect of quasiperiodic and random noise on many-body dynamical decoupling protocols

TL;DR

This paper investigates how quasiperiodic and random noise influence many-body dynamical decoupling protocols, revealing distinct relaxation regimes and spectral features that affect symmetry preservation.

Contribution

It extends previous work by analyzing the effects of different noise types on dynamical decoupling, identifying relaxation regimes and their spectral origins.

Findings

Initial relaxation to a noise-independent prethermal plateau

Linear relaxation with a noise-dependent rate

Logarithmic relaxation regime for quasiperiodic noise

Abstract

Symmetries (and their spontaneous rupturing) can be used to protect and engender novel quantum phases and lead to interesting collective phenomena. In Ref. 1, the authors described a general dynamical decoupling (polyfractal) protocol that can be used to engineer multiple discrete symmetries in many-body systems. The present work expands on the former by studying the effect of quasiperiodic and random noise on such a dynamical scheme. We find generally that relaxation of engineered symmetry generators proceeds by i) an initial relaxation on microscopic timescales to a prethermal plateau whose height is independent of noise, ii) a linear relaxation regime with a noise-dependent rate, followed by iii) a slow logarithmic relaxation regime that is only present for quasiperiodic noise. We glean the essential features of these regimes via scaling collapses and show that they can be generally explained by the spectral properties of the various noise waveforms considered. In particular, the quasiperiodic noise is characterised by highly time dependent spectrum with a noise floor that mimics white noise, and peaks that grow sharper with time. We argue that both the noise floor and peaks contribute to the initial linear-in-time relaxation while the logarithmic regime is initiated when the peaks become sufficiently well resolved and cease to contribute to further relaxation. We provide numerical evidence to justify these findings.