Dissertation Information

Title: From Solvation to Storage: Computational and Machine Learning Insights into Organic Molecules and Battery Materials

Program: Computing Ph.D. - Data Science

Advisor: Dr. Oliviero Andreussi

Committee Members: Dr. Edoardo Serra, Dr. Konrad Meister

Abstract

This proposal outlines a two-part research effort that integrates machine learning and computational chemistry to address challenges in two interconnected domains: the modeling of solvation phenomena in organic molecules and the discovery of advanced materials for energy storage.

In the first part, we investigate the thermodynamic and structural properties of solvated organic compounds, with a focus on how molecular features influence solvation free energy, conformational behavior, and solvent interactions. Through a combination of classical and quantum chemical calculations, alongside machine learning models trained on curated datasets, we aim to identify key molecular descriptors that govern solvation behavior across diverse solvents. A substantial portion of this work has been completed, including dataset generation, model development, and preliminary validation.

The second part addresses the discovery of promising anode materials for battery applications, with an emphasis on bulk metallic glasses (BMGs). This research centers on the Co–P system, which exhibits favorable structural and electrochemical characteristics. Using density functional theory, molecular dynamics simulations, and machine-learned interatomic potentials, we have investigated the composition–structure relationships governing glass formation. This project is nearing completion, and results are being prepared for peer-reviewed publication.
Together, these studies demonstrate how data-driven and physics-based approaches can be combined to accelerate the design of molecules and materials, with broad implications for chemical modeling, solvent selection, and sustainable energy storage technologies.