Ageing and Relaxation in Glass Forming Systems

TL;DR

This paper presents a unified view of glass formation, linking local structural heterogeneity with slow dynamics and aging, and challenges the notion of a finite-temperature divergence in relaxation times.

Contribution

It introduces a generic class of glass formers with competing states, explaining aging and relaxation through local shear modulus variations and structural patches.

Findings

Relaxation events occur where local shear modulus is low or negative.

Aging involves localized transitions between local minima in the potential surface.

Relaxation times increase super-exponentially with decreasing temperature, with no finite-temperature divergence.

Abstract

We propose that there exists a generic class of glass forming systems that have competing states (of crystalline order or not) which are locally close in energy to the ground state (which is typically unique). Upon cooling, such systems exhibit patches (or clusters) of these competing states which become locally stable in the sense of having a relatively high local shear modulus. It is in between these clusters where ageing, relaxation and plasticity under strain can take place. We demonstrate explicitly that relaxation events that lead to ageing occur where the local shear modulus is low (even negative), and result in an increase in the size of local patches of relative order. We examine the ageing events closely from two points of view. On the one hand we show that they are very localized in real space, taking place outside the patches of relative order, and from the other point of view we show that they represent transitions from one local minimum in the potential surface to another. This picture offers a direct relation between structure and dynamics, ascribing the slowing down in glass forming systems to the reduction in relative volume of the amorphous material which is liquid-like. While we agree with the well known Adam-Gibbs proposition that the slowing down is due to an entropic squeeze (a dramatic decrease in the number of available configurations), we do not agree with the Adam-Gibbs (or the Volger-Fulcher) formulae that predict an infinite relaxation time at a finite temperature. Rather, we propose that generically there should be no singular crisis at any finite temperature: the relaxation time and the associated correlation length (average cluster size) increase at most super-exponentially when the temperature is lowered.