Multiscale Molecular Modeling and AI-Accelerated Discovery of Polymeric Materials for Sustainability

Advancing a sustainable, low-carbon future depends critically on the discovery of materials with exceptional performance, durability, and manufacturability. Polymeric materials—particularly polymer electrolytes and polyelectrolytes—are central to such technologies, enabling functions in fuel cells, electrolyzers, batteries, and selective membranes. Their chemical diversity and architectural tunability allow for precise control over ionic conductivity, selectivity, and stability. Yet these properties emerge from hierarchical structures spanning molecular to macroscopic scales, and small changes in chemistry or morphology can propagate unpredictably across these scales. The vast combinatorial design space, coupled with the complexity of structure–property relationships, renders traditional trial-and-error development inefficient and incomplete. This thesis advances a unified framework for polymer materials discovery that integrates multiscale molecular modeling with artificial intelligence (AI). First, molecular simulations at atomistic and coarse-grained resolutions, anchored to experimental observables such as scattering and transport measurements, are used to reveal the microscopic interactions and mesoscale morphologies that govern ion transport. These studies identify physically meaningful descriptors—such as hydration structure, medium-range correlations, and percolation connectivity—that directly link chemical design to macroscopic performance. Second, physics-guided AI models, built on hierarchical, functional-group-based representations, predict key properties and propose chemically plausible candidates with optimized, multi-objective performance. The resulting design strategy couples mechanistic insight with data-driven exploration, enabling targeted navigation of vast chemical spaces. By uniting fundamental insight with predictive and generative tools, this work transforms qualitative heuristics into quantitative rules for designing polymeric materials. The methodology is broadly applicable and provide a pathway to accelerate the discovery of materials for sustainable technologies of the future.