Is Structural Relaxation During Vitrification the Inverse of the Glass Transition?

Abstract

We have recently applied the second law to an isolated system, consisting of a system Σ such as a glass surrounded by an extremely large medium Σ, to show that the instantaneous temperature T(t), thermodynamic entropy S(T_0,t) and enthalpy H(T_0,t) of Σ decrease in any isothermal relaxation towards their respective equilibrium values under isobaric cooling. The Gibbs statistical entropy also conforms to the above relaxation behavior in a glass, which however is contrary to the conjecture by Gupta, Mauro and coworkers that the glass transition and the structural relaxation are inverse processes. However, they do not establish that the entropy loss during the glass transition is accompanied by a concomitant entropy gain of the medium to maintain the second law. They use a novel statistical formulation of entropy based on several conjectures such as it being zero for a microstate, but do not compare its behavior with the thermodynamic entropy. The formulation is designed to show the entropy loss. Its subsequent rise not only contradicts our result but also implies that the glass during relaxation must have a negative absolute temperature. To understand these puzzling results and the above conjecture, we have carried out a critical evaluation of their unconventional approach. We find that the inverse conjecture is neither supported by their approach nor by the second law. The zero-entropy microstate conjecture is only consistent with S(T_0,t)\equiv 0 at all temperatures, not just at absolute zero and is found to have no scientific merit. We show that the maximum entropy gain of the medium during the glass transition invalidates the entropy loss conjecture. After pointing out other misleading, confusing and highly exaggerated statements in their work, we finally conclude that their unconventional statistical approach and computational scheme are not appropriate for glasses.