- Main
First-Principles Discovery of Reaction Mechanisms and High-Performance Computational Tools for Minimum Energy Pathway Tracing
- Park, Heejune
- Advisor(s): Wang, Lee-Ping
Abstract
The exploration of potential energy surfaces (PESs) is essential for understanding chemical reaction mechanisms, yet the high dimensionality and complexity of PESs make locating minimum-energy pathways (MEPs) challenging. This thesis develops computational tools to explore PESs and discover reaction pathways. Chapter 1 outlines the theoretical foundations of electronic structure methods, from Hartree–Fock and post-Hartree–Fock approaches to density functional theory, and how the energies from these methods are used to evaluate interatomic forces to run ab initio molecular dynamics (AIMD) simulations; it also describes how AIMD trajectories are refined to characterize reaction mechanisms. Chapter 2 uses a retrosynthetic AIMD simulation nanoreactor approach to identify two previously unknown low-barrier pathways for pyrrole formation and one barrierless pathway for pyridine formation, offering new insights into the prebiotic synthesis of nitrogen-containing heterocycles. Chapter 3 introduces a novel interpolation method, TRICS (translation–rotation internal coordinates with splicing), to generate high-quality MEP approximations by employing multiple internal coordinate systems and iterative splicing; the method performs comparably or better than existing approaches across 679 reactions from three benchmark datasets. Chapter 4 develops a high-throughput nudged elastic band (NEB) framework on cloud infrastructure, enabling efficient parallel calculations and systematic evaluation of protocol parameters. This implementation achieves high convergence rates across diverse reactions and successfully reproduces results for transition-metal-catalyzed systems. Together, these tools advance the automated exploration and characterization of chemical reaction pathways on complex PESs.