CAREER: Uncovering Mechanical Structure-Property Relations in Highly Entangled Polymers via Multiscale Modeling and Interpretable Machine Learning

About This Grant

This Faculty Early Career Development Program (CAREER) award will support research that advances predictive design of polymer materials by uncovering how molecular entanglements govern mechanical performance. Polymer entanglements form naturally in long-chain materials and are known to significantly enhance stiffness, strength, and toughness. Recent studies have demonstrated that replacing chemical crosslinks with entanglements in certain hydrogels can increase toughness by several orders of magnitude, highlighting their potential to improve materials used in applications ranging from industrial elastomers and adhesives to biomedical implants and soft robotic systems. Despite their importance, the role of entanglements in determining macroscopic mechanical behavior remains poorly understood due to the difficulty of experimentally observing these nanoscale features and the challenge of connecting molecular physics to bulk response. This project addresses these challenges by integrating multiscale mechanics, machine learning, and experimental validation to enable faster and more reliable prediction of polymer properties based on synthesis choices. In parallel, this CAREER award will support new educational initiatives that train students at the intersection of mechanics, materials science, and machine learning, with an emphasis on model interpretability and hands-on demonstrations to improve student learning and workforce preparedness. The CAREER project will support research that seeks to establish a multiscale modeling framework linking molecular-scale polymer physics to emergent mechanical properties in highly entangled networks. The central goal is to develop physics-informed machine learning models that capture the mechanical role of entanglements and enable predictive optimization of polymer gels. To achieve this, coarse-grained molecular dynamics simulations will be used to model entangled polymer networks, which will then be distilled into graph-based representations suitable for graph neural networks (GNNs). These GNNs will be trained to reproduce the micromechanical behavior of the molecular simulations while enabling mechanical property predictions at substantially larger spatiotemporal scales than existing discrete approaches. Model predictions will be validated through experimental mechanical testing. The project will focus on polyacrylamide hydrogels, a well-characterized system in which entanglements dominate mechanical behavior due to their high density relative to chemical crosslinks. While this material system serves as a model platform, the framework will provide a general roadmap for physics-informed, machine-learning-driven design of entangled elastomeric polymers. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.