Analyzing Dynamical Processes with Local Molecular Field Theory

Abstract

Local molecular field (LMF) theory provides a framework for describing the collective response of a system to long-range interactions in nonuniform liquids. Based on this theory, different roles played by the short and long-range components of the intermolecular interactions can be disentangled in determining relevant structural and thermodynamic properties in equilibrium. Furthermore, in dynamical processes, nonlocal long-range interactions are often associated with long relaxation times, and can contribute significantly to the stability of the system in different phases. In this thesis, LMF theory is utilized to quantify and analyze the dynamical effects arising from long-range Coulomb interactions in aqueous solutions, while elucidating how they are connected to strong local forces and fluctuations.

The first half of the work concerns ionic and dipolar solvation dynamics, which plays an essential role in many solution phase chemical reactions. The physical models of Gaussian-smoothed charge and dipole distributions are conceptualized from LMF theory to investigate the molecular origins of linear and nonlinear effects in solvation dynamics. The long-range component of the solute-solvent electrostatic interaction is shown to underlie the linear response behavior of the system, while the short-range interactions introduce additional nonlinear effects. The LMF-based solvation models further demonstrate their functionality in probing the intrinsic dielectric dispersion of solvent water. The second half of the work is focused on the nucleation processes in the aqueous environment. Simulating crystal nucleation from solutions requires efficient treatments for intermolecular interactions to drive the transitions on time scales affordable to molecular dynamics simulations. For this purpose, a LMF-based molecular model is employed to capture the renormalized long-range interactions, and well-tempered metadynamics is adopted to enhance the fluctuations arising from short-range interactions. By comparing to a short-range reference model, the necessity of long-range interactions in explaining metastability is revealed. Temporal fluctuations and direct evidence for the two-step nucleation mechanism are observed through the analysis using a deep learning-based approach. The results about these two types of dynamical processes contribute to a deeper understanding of the roles of short and long-ranges interactions in the aqueous systems.