English

Logarithmically Slow Relaxation in Quasi-Periodically Driven Random Spin Chains

Disordered Systems and Neural Networks 2018-02-26 v2 Quantum Gases Statistical Mechanics Strongly Correlated Electrons

Abstract

We simulate the dynamics of a disordered interacting spin chain subject to a quasi-periodic time-dependent drive, corresponding to a stroboscopic Fibonacci sequence of two distinct Hamiltonians. Exploiting the recursive drive structure, we can efficiently simulate exponentially long times. After an initial transient, the system exhibits a long-lived glassy regime characterized by a logarithmically slow growth of entanglement and decay of correlations analogous to the dynamics at the many-body delocalization transition. Ultimately, at long time-scales, which diverge exponentially for weak or rapid drives, the system thermalizes to infinite temperature. The slow relaxation enables metastable dynamical phases, exemplified by a "time quasi-crystal" in which spins exhibit persistent oscillations with a distinct quasi-periodic pattern from that of the drive. We show that in contrast with Floquet systems, a high-frequency expansion strictly breaks down above fourth order, and fails to produce an effective static Hamiltonian that would capture the pre-thermal glassy relaxation.

Keywords

Cite

@article{arxiv.1708.00865,
  title  = {Logarithmically Slow Relaxation in Quasi-Periodically Driven Random Spin Chains},
  author = {Philipp T. Dumitrescu and Romain Vasseur and Andrew C. Potter},
  journal= {arXiv preprint arXiv:1708.00865},
  year   = {2018}
}

Comments

6+3 pages, 4+4 figures; v2. minor improvements; as published

R2 v1 2026-06-22T21:04:59.526Z