New Insights into the Glass Transition from Computational Prediction and Evolutionary Design

In most polymers, the glass transition is one of the most important phenomena determining performance properties including mechanical response, processability, and transport behavior. For this reason, understanding and controlling the glass transition is a longstanding goal of polymer science and soft condensed matter physics. However, the vast range of timescales associated with glass formation, coupled with a lack of an agreed-upon theoretical description of the problem, have posed major challenges to achieving this goal. Here I describe a new approach to this problem, combining efficient molecular dynamics simulations, physics-based heuristics, machine learning, and evolutionary algorithms to predict, understand, and design the glass transition.