arXiv:2610.04197
How to distinguish chaos from integrability using OTOC
Alexey Milekhin, Anders Wallberg
Abstract
The decay of out-of-time-ordered correlators (OTOCs) to a small residual value signals the scrambling of quantum information. This behavior is generic: it occurs in both chaotic and interacting integrable many-body systems, and therefore cannot by itself distinguish between them. In this paper we analytically study the late-time saturation value of a closely related observable, operator purity, and show that its scaling with system size robustly distinguishes chaotic from integrable systems. Our approach is based on the Mazur bound, which provides concrete scaling estimates and makes explicit why the locality of integrable charges matters. We support our findings with numerical simulations of one-dimensional spin chains.
AI Summary
The paper asks whether late-time scrambling observables distinguish interacting integrable dynamics from chaos, despite OTOC decay occurring in both. It studies the infinite-time average of operator purity $P_A(t)=\operatorname{Tr}\rho_{A_1A_2}(t)^2$, an averaged infinite-temperature OTOC associated with the evolution operator’s Choi state. Writing purity as a SWAP autocorrelation in two copies allows application of the Mazur bound. For orthogonal traceless conserved charges, the bound contains a baseline $d_A^{-2}$ plus nonnegative contributions determined by partial traces of each charge onto $A$. For a family of quasi-local charges with uniformly bounded weights and increasing support lengths, it gives $d_A^2\overline{P_A}\gtrsim L_A^3/(3L^2)$, hence at least linear growth in $L_A$ at fixed $r=L_A/L$. By contrast, energy conservation and its powers yield the thermodynamic-limit lower bound $d_A^2\overline{P_A}\geq(1-r^2)^{-1}$. The paper argues that chaotic systems have a size-independent residual at fixed $r<1/2$, using non-resonance and completeness arguments for conserved operators in the doubled system. Exact diagonalization of integrable and perturbed XXX, XYZ, and Ising chains supports this distinction. Infinite-time averages reach $L=12$; sampling times between $L^2$ and $4L^2$ reaches $L=14$, with $L_A=L/2$. Limitations include the uniform-weight requirement for the integrable bound, stronger unverified assumptions behind an exponential enhancement, and the absence of a general proof of chaotic saturation or polynomial saturation times. Sampling independence is assumed; finite-temperature extensions and individual-OTOC diagnostics remain open.
Central claim may not hold: The integrable linear-growth theorem requires uniformly positive localization weights for an orthogonal charge family, a condition the paper does not establish for its interacting numerical examples. The claimed bounded chaotic saturation is not established because the Hamiltonian-power calculation controls fixed-degree contributions to a lower bound, without controlling the growing-degree contributions needed to invoke completeness.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 13+7 pages, 2 figures