Insights into the physics of glass formation from simulations of thin film dynamics

A major motivation for 35 years of research on nanoconfined glass formation has been the hope that confinement and broken symmetry near interfaces might reveal the details of dynamical correlations in glass forming liquids and thus resolve the nature of glass formation itself. A central challenge has been lack of a settled, cohesive picture of the dependences of altered dynamics on position, domain size, and temperature. Here, I discuss results establishing a cohesive phenomenology of interface and confinement effects on glass formation at the longest timescales accessible to simulation. Key findings include a temperature-invariant fractional reduction in activation barriers, an exponential gradient of activation barriers and glass transition temperatures in the first 5-10 nm near the free surface, saturation of the range of this gradient upon cooling, and the emergence of non-interface finite-size effects in thin films. I report on new results establishing the long-ranged fate of this gradient beyond the near-surface exponential. We find that the phenomenology of altered thin films dynamics is profoundly at odds with predictions of a number of theories of glass formation. On the other hand, the phenomenology is predicted near-quantitatively at a zero-parameter level by the Elastically Collective Nonlinear Langevin Equation (ECNLE) Theory of Schweizer and coworkers. This theory predicts altered dynamics in films as a result of alteration and truncation of local caging constraints and long-ranged elastic activation barriers. Together with complementary evidence for the success of the ECNLE theory in predicting bulk glass formation, these findings may have major implications for our understanding of the underlying mechanism of glass formation.