Degree

Doctor of Philosophy (PhD)

Department

Chemistry

Document Type

Dissertation

Abstract

Polymer electrolyte membranes, an intriguing category of soft materials, are characterized by polymer backbones tethered with negatively or positively charged functional groups, finding applications from biological macromolecules to electrochemical separations in fuel cells. This dissertation focuses on computational investigations of polymer electrolytes in aqueous solutions using polypeptoids and ion exchange membranes. Polypeptoids, highly tailorable peptidomimetic polymers lacking backbone hydrogen bonding or chirality, were used as model systems. Micelles formed from these polypeptoid polymers exhibited electrostatic repulsion, which competes with charge-sodium interactions, adding complexity to predicting micelle shapes. This research refined the design strategy, enhancing micelle shape and structure through precise adjustments in the charged components along the polypeptoid backbone.

Chemical separations significantly contribute to global energy consumption. There is significant interest in using material-based separations and electrochemical processes powered by renewables to reduce energy consumption. Polymer electrolytes are used as ion-exchange membranes and binding agents for porous resin wafer conductors in electrically driven ionic separation processes. A novel machine learning framework was developed to predict ion activity coefficients by integrating molecular dynamics simulation data with the material properties of model compounds in ionic separations. Integrated solvation descriptors from atomistic simulations were incorporated as an additional dimension into an existing ML framework. This framework employs a limited dataset of experimental activity coefficients, polymer structure information, and molecular descriptors characterizing ion and polymer solvation to predict ion activity coefficients within Ion Exchange Membranes. The machine learning models achieved high accuracy, with an average mean absolute error of under 10%. This study demonstrates the effectiveness of the integrated approach in predicting ion exchange membrane activity coefficients, even with limited experimental data. This simplified approach aligns with industrial needs for efficient electrochemical separations—promising advancements in the field.

Date

10-26-2023

Committee Chair

Kumar, Revati

Download

Share

COinS