<b>ABSTRACT</b><br/>Machine learning (ML) methods have gained increasing traction the interest of scientist for many applications in materials science from discovery and synthesis of new materials to prediction of materialrtial properties and their device level performance. Block/random co-polymer electrolytes such as poly (styrene-block-vinyl pyridine) are used as ionic exchange and binders for porous resin wafer conductors in electrically driven ionic separations. Macromolecular architecture of these polymers has huge impact on transport properties such as perm selectivity, conductivity, etc. in ion exchange membranes. This study focuses on developing a data driven machine learning model for electrically driven/ electrochemical ion separations. This model will bridge molecular scale attributes from molecular dynamics (MD) simulations to material properties of model materials and translational materials used in ionic separations. Both existing and novel data will be used to train this model. The classical molecular dynamics (MD) simulations with conventional force fields (OPLSAA, etc.) are used to elaborate the structural and dynamical properties of novel systems including diffusion constants, rotational dynamics, largest water cluster size, conductivity, hopping rates, average number of waters in first hydration shell, etc. Machine learning methods such as Support Vector Regression (SVR) and other machine learning tools will be used to connect molecular level attributes from molecular dynamics to material properties polymer electrolytes. Our preliminary SVR model includes the concentration dependences of the activity coefficients of polymer electrolytes using molecular dynamics finger prints, based on available molecular dynamics and experimental data on co-polymer electrolytes. Furthermore, other properties such as conductivity and, ionic selectivity will be used to further expand the existing model, thereby enabling the guided design of nthese can be used to guide the search for new materials suitable for ionic separations. The results of this work will be presented and discussed.

Bridging the length scales in ionic separation via data-driven machine learning

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.