Detecting Many-Body Scars from Fisher Zeros

  • Yuchen Meng
  • , Songtai Lv
  • , Yang Liu
  • , Zefan Tan
  • , Erhai Zhao*
  • , Haiyuan Zou*
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

The far-from-equilibrium dynamics of certain interacting quantum systems still defy precise understanding. One example is the so-called quantum many-body scars (QMBSs), where a set of energy eigenstates evade thermalization to give rise to long-lived oscillations. Despite the success of viewing scars from the perspectives of symmetry, commutant algebra, and quasiparticles, it remains a challenge to elucidate the mechanism underlying all QMBS and to distinguish them from other forms of ergodicity breaking. In this work, we introduce an alternative route to detect and diagnose QMBS based on Fisher zeros, i.e., the patterns of zeros of the analytically continued partition function Z on the complex β (inverse temperature) plane. For systems with scars, a continuous line of Fisher zeros will appear off the imaginary β axis and extend upward, separating the β plane into regions with distinctive thermalization behaviors. This conjecture is motivated from interpreting the complex Z as the return amplitude of the thermofield double state, and it is validated by analyzing two models with QMBS, the P¯XP¯ model and the Ising chain in external fields. These models also illustrate the key difference between QMBS and strong ergodicity breaking including their distinctive renormalization group flows on the complex β plane. This "statistical mechanics"approach places QMBS within the same framework of thermal and dynamical phase transitions. It has the advantage of spotting scars without exhaustively examining each individual quantum state.

Original languageEnglish
Article number070402
JournalPhysical Review Letters
Volume135
Issue number7
DOIs
StatePublished - 15 Aug 2025

Fingerprint

Dive into the research topics of 'Detecting Many-Body Scars from Fisher Zeros'. Together they form a unique fingerprint.

Cite this