Degree
Doctor of Philosophy (PhD)
Department
Chemistry
Document Type
Dissertation
Abstract
Interfaces where two or more condensed phases meet govern phenomena central to catalysis, electrochemical energy conversion, separations, and environmental remediation. Their structural heterogeneity and dynamic nature make them challenging to probe experimentally, but computer simulations provide molecular-level insights into processes that remain inaccessible to experiments. This dissertation employs ab initio molecular dynamics (AIMD), classical molecular dynamics, and machine-learned interatomic potentials to investigate the structure, dynamics, and proton transport at graphene oxide (GO)–liquid interfaces.
The first project explores aqueous GO interfaces across a spectrum of oxidation states that extends from highly oxidized sheets to fully reduced graphene. Deep neural network force fields trained on AIMD data enable nanosecond-scale simulations while retaining first-principles fidelity. These simulations reveal how interfacial water structure, hydrogen-bond networks, and reactivity respond to surface functionalization. Oxidized GO stabilizes long-lived hydronium species and creates acidic environments, whereas reduced surfaces suppress proton activity and approach the inert behavior of pristine graphene.
The second project examines oxidation-controlled proton transfer at GO–water interfaces. Deep potential molecular dynamics with explicit hydrated protons highlights two contrasting mechanisms. Oxidized GO promotes spontaneous interfacial proton release, while reduced GO stabilizes proton adsorption at the surface. These results show how surface chemistry governs proton localization and mobility, offering design principles for GO as either a proton exchange membrane material or a solid acid catalyst.
The third project extends these studies to GO–methanol interfaces, where AIMD simulations identify the formation of protonated methanol and interfacial methoxide species. Development of a message passing neural network potential (MACE) enables nanosecond-scale tracking of protonated methanol, demonstrating how machine-learned models expand the reach of interfacial simulations and open new directions for predictive studies of complex reactive environments.
Altogether, this dissertation demonstrates that molecular simulations provide a unified framework for understanding interfacial chemistry at GO surfaces in both aqueous and non-aqueous systems. More broadly, the integration of AIMD with machine learning establishes a transferable paradigm for studying reactive interfaces and lays the foundation for designing next-generation materials for energy, catalysis, and environmental technologies.
Date
10-29-2025
Recommended Citation
Azom, Golam, "Computational Nanoscale Investigations of Structure, Dynamics, and Reactivity at Solid–Liquid Interfaces" (2025). LSU Doctoral Dissertations. 6915.
https://repository.lsu.edu/gradschool_dissertations/6915
Committee Chair
Kumar, Revati