We are happy to report that Dr. Jianjun Hu, along with collaborators Qi Wang (PI), Sophya Garashchuk, Linda Shimizu, and Chuanbing Tang, have received a research award from the Department of energy for their project titled "Data-science enabled investigation of the mechanisms for multiscale ion transport in functional electrolytes and for the radical generation in crystalline assemblies".
By integrating data science, machine learning and multiscale modeling/simulation with innovative experiments, we will investigate ion transport mechanisms and morphology/property relationships copolymer-based functional electrolyte membrane of great potential to become the next generation energy conversion and storage devices. We expand the adaptive materials design loop based on materials informatics to one incorporating multiscale modeling/simulation aided by data analytic tools and deep learning Technologies. Data science and multiscale modeling are essential for attaining fundamental understanding for the rational design of the complex molecular aggregates with desired properties. The research is relevant to ‘chemistry across multiple scales in complex environments important in catalysis, biochemistry or electrochemistry’ and to ‘far from equilibrium phenomena where dynamics is fast, such as in transport and separation in complex systems’ in basic energy science directions.
We will perform ab initio dynamics with quantum corrections for selected nuclei (OH- and H+) for realistic molecular models of polyelectrolytes. The computational bottleneck, ES evaluations (on the order of 107) will be alleviated by replacing many calculations with estimates from machine learning (ML) protocols. These simulations aim to elucidate the system-specific mechanisms of ion transport, and to yield parameters (mobility/conductivity) for a phase field model describing ion transport in phase-separated heterogeneous morphology. This model, targeting the ion transport in the heterogeneous material system, will be derived from the Onsager principle, in which the coefficients will be learned from simulation data generated from atomic level simulations as well as experimental data. Optimization of the morphology for targeted ion transport properties will be carried out as well to guide rational design of electrode materials.