arXiv:2610.05995
Free fermion models with an operator entanglement growing as $O(\sqrt{t})$
Federico Tonetto, Jérôme Dubail
Abstract
We study the time evolution of the Operator Space Entanglement Entropy (OSEE) of one-dimensional number-conserving free fermions subject to a random space- and time-dependent evolution. It is known that, in free-fermion systems without randomness, the OSEE of \emph{local} operators grows at most \emph{logarithmically} in time. Here we show that, in the presence of randomness, the OSEE of local spin operators involving a Jordan-Wigner string grows as $\mathcal O(\sqrt t)$, for all values of the Rényi index. We establish this in two models: a Brownian Hamiltonian and a brick-wall matchgate circuit. We also consider the OSEE of the evolution operator, and we show that it also grows as $\mathcal O(\sqrt{t})$, in contrast with the linear growth generically expected without randomness. We analytically compute the average entropies of the two operators by mapping the quantum models to the symmetric simple exclusion process, relying on an assumption of self-averaging which we support numerically. We also consider two other types of randomness: random evolution in time that keeps spatial translation invariance, and quenched disorder. In the former case, we find that the OSEE also grows as $\sqrt{t}$. In the latter case, we find that the entropy grows as $\mathcal O( \log \log t)$ for Hamiltonian evolution both for local operators with a Jordan-Wigner string and for the evolution operator, whereas it remains bounded at all times in the brick-wall matchgate circuit.
AI Summary
The paper asks whether logarithmic local-operator entanglement growth in clean free-fermion chains survives random evolution. It studies one-dimensional, number-conserving fermions under a Brownian nearest-neighbour hopping Hamiltonian (QSSEP) and Haar-random $U(2)$ number-conserving matchgate circuits, considering spatial, temporal, and spatiotemporal disorder. With independent randomness in space and time, the Rényi operator-space entanglement entropy (OSEE) grows as $S^\alpha_A(O)=2s_\alpha\sqrt{t}+o(\sqrt{t})$ for local spin operators containing a Jordan–Wigner string, and as $S^\alpha_A(U)=c_\alpha\sqrt{t}+o(\sqrt{t})$ for the evolution operator, for every fixed $\alpha>0$. The predicted von Neumann coefficients are $2s_1\simeq1.0475$ and $c_1\simeq0.8998$ in the chosen time units. Local operators without a string retain bounded OSEE. The calculation expresses both entropies through a common reduced Gaussian correlation matrix: string OSEE equals twice the entanglement of an evolving domain wall. Mapping averaged full counting statistics to the classical symmetric exclusion process, and assuming leading-order equality of annealed and quenched generating functions, yields correlation-matrix moments and integral formulas for the coefficients. Numerical simulations support these predictions. Purely temporal disorder with a two-site unit cell also gives numerical $\sqrt{t}$ growth, whereas one-site translation-invariant noise retains logarithmic string OSEE. Quenched disorder gives numerical $\log\log t$ growth for the chiral random-hopping Hamiltonian and saturation for the circuit. Limitations include unproved full-counting-statistics self-averaging, an unproved exclusion-process correspondence for the circuit, and no exact entropy treatment of the other disorder regimes. The models are free but not Yang–Baxter integrable.
Central claim may not hold: The matchgate circuit averages to a discrete-time exclusion process, but applying the continuous-time SSEP current large-deviation formula requires an asymptotic equivalence that the paper does not establish. The derivation for all positive Rényi indices lacks small-eigenvalue control needed to extend its integer-moment asymptotics to the entropy functions for 0<α≤1. Bounded variance of log χ does not imply the annealed–quenched equivalence claimed in the self-averaging subsection, because rare events can still change the exponential average at leading order.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.