Aging in the trap model as a relaxation further away from equilibrium

TL;DR

This paper investigates the aging process in the trap model below the glass transition, revealing that the 'distance to equilibrium' can increase during aging for infinite traps and exhibits unique scaling behaviors for finite traps.

Contribution

It introduces a novel analysis of the 'distance to equilibrium' in the trap model, showing unexpected increase during aging and unique scaling laws for finite trap numbers.

Findings

elta S(t) increases during aging with infinite traps.

or finite traps, elta S(t) reaches a maximum before decreasing.

haracteristic times scale non-linearly with trap number N.

Abstract

The aging regime of the trap model, observed for a temperature T below the glass transition temperature T_g, is a prototypical example of non-stationary out-of-equilibrium state. We characterize this state by evaluating its "distance to equilibrium", defined as the Shannon entropy difference \Delta S (in absolute value) between the non-equilibrium state and the equilibrium state with the same energy. We consider the time evolution of \Delta S and show that, rather unexpectedly, \Delta S(t) continuously increases in the aging regime, if the number of traps is infinite, meaning that the "distance to equilibrium" increases instead of decreasing in the relaxation process. For a finite number N of traps, \Delta S(t) exhibits a maximum value before eventually converging to zero when equilibrium is reached. The time t* at which the maximum is reached however scales in a non-standard way as t* ~ N^(T_g/2T), while the equilibration time scales as N^(T_g/T). In addition, the curves \Delta S(t) for different N are found to rescale as ln t/ln t*, instead of the more familiar scaling t/t*.