Aging and diffusion in low dimensional environments

TL;DR

This paper investigates aging and diffusion in one-dimensional disordered systems like the Sinai model, revealing three large-time regimes and proposing a scaling approach for predicting aging behaviors in complex landscapes.

Contribution

It introduces a comprehensive analysis of aging and diffusion regimes in low-dimensional disordered environments, including a new exactly solvable model and a scaling method for barrier landscapes.

Findings

Identification of three large-time regimes: quasi-equilibrium, diffusion, and aging.

Numerical evidence supporting the existence of these regimes in the Sinai model.

Introduction of a new exactly solvable model demonstrating diverse aging behaviors.

Abstract

We study out of equilibrium dynamics and aging for a particle diffusing in one dimensional environments, such as the random force Sinai model, as a toy model for low dimensional systems. We study fluctuations of two times $(t_w, t)$ quantities from the probability distribution $Q(z,t,t_w)$ of the relative displacement $z = x(t) - x(t_w)$ in the limit of large waiting time $t_w \to \infty$ using numerical and analytical techniques. We find three generic large time regimes: (i) a quasi-equilibrium regime (finite $\tau=t-t_w$) where $Q(z,\tau)$ satisfies a general FDT equation (ii) an asymptotic diffusion regime for large time separation where $Q(z) dz \sim \bar{Q}[L(t)/L(t_w)] dz/L(t)$ (iii) an intermediate ``aging'' regime for intermediate time separation ($h(t)/h(t_w)$ finite), with $Q(z,t,t') = f(z,h(t)/h(t')) $. In the unbiased Sinai model we find numerical evidence for regime (i) and (ii), and for (iii) with $\bar{Q(z,t,t')} = Q_0(z) f(h(t)/h(t'))$ and $h(t) \sim \ln t$. Since $h(t) \sim L(t)$ in Sinai's model there is a singularity in the diffusion regime to allow for regime (iii). A directed model, related to the biased Sinai model is solved and shows (ii) and (iii) with strong non self-averaging properties. Similarities and differences with mean field results are discussed. A general approach using scaling of next highest encountered barriers is proposed to predict aging properties, $h(t)$ and $f(x)$ in landscapes with fast growing barriers. We introduce a new exactly solvable model, with barriers and wells, which shows clearly diffusion and aging regimes with a rich variety of functions $h(t)$.