Eyring theory for plasticity in amorphous polymers violates Curie's principle

In 1936, Eyring introduced a model for plastic flow which still forms the bedrock of practically all studies on plasticity of glassy polymers. Though the concept of activated process he introduced is fundamentally relevant, we argue here that under no circumstances can the Eyring model be correct as it violates Curie's principle, which is a basic physical requirement of statistical mechanics. An alternative model was proposed by [Long et al., Phys. Rev. Mater., 2018, 2, 105601] to describe the acceleration of the dynamics by an applied stress, in which the elastic energy stored at the length scale of dynamical heterogeneities ξ ≈ 3 - 5 nm reduces the free energy barriers for relaxation. While this model still considers α-relaxation as an activated process, as did Eyring, it fully complies with Curie's principle. It is based on a Landau expansion of the free energy barriers as a function of the applied stress. We argue that, due to the large length scale involved in the α-relaxation, only the leading quadratic order term should be retained, as higher order terms are negligible. We discuss a few recent experiments which confirm these features. This model opens the way to set glassy polymers plasticity into the realm of out-of-equilibrium statistical physics and condensed matter physics, which we argue is the appropriate framework for considering the physics of glass transition and mechanical properties of glassy polymers.