Machine Learning-Enhanced Multiscale Simulations for Soft Material Modeling
Research
Molecular Design of Biosensors
Big Idea:
Understand the structure-performance relationships of near-infrared (NIR) fluorescent polymer dots for deep-tissue imaging by using ML-enhanced multiscale simulations & data science approaches.Potential Impact:
Develop high quantum yield NIR polymer dots for tumor-targeted labeling.
Charge Transport of Flexible Electronics
Big Idea:
Integrate molecular dynamics (MD) simulations, quantum mechanism approaches, and ML techniques into a unified computing platform to efficiently explore the relationship between charge transport properties and molecular structures of organic semiconductors (OSCs).
Potential Impact:
Understand the charge transport mechanism and advance the molecular design of OSCs
Machine Learning for Multiscale Structure Characterizations
Big Idea:
Soft materials exhibit a multi-length-scale nature that plays a critical role in their applications and processing. However, characterizing these disordered systems has remained a longstanding challenge. To address this, we are developing data-driven approaches to uncover and identify their structures across multiple length scales.
Potential Impact:
Efficiently exploring the multiscale structures of soft materials, enabling improved structural and morphological control.
Degradation for Sustainable Polymers
Big Idea:
Develop ML-based diffusion models to simulate heterogeneous catalytic upcycling, explore polymer diffusion within and around catalyst matrices, elucidate chain release mechanisms, and model reaction kinetics using electronic coarse-grained approaches.
Potential Impact:
Facilitate upcycling techniques to effectively reduce plastic waste.