Access Type

Open Access Dissertation

Date of Award

January 2018

Degree Type


Degree Name




First Advisor

H. Bernhard Schlegel


There are two topics in this dissertation: ground state and excited state modeling of a few series of transition metal complexes that facilitate solar energy conversion, and Born-Oppenheimer Molecular Dynamics (BOMD) simulations of molecular cations interacting with intense mid-infrared laser light.

In Chapter 2 and 3, a few series of transition metal complexes that facilitate solar energy conversion are studied computationally. Metal-to-ligand charge-transfer (MLCT) excited states of several (ruthenium) (monodentate aromatic ligand, MDA) chromophore complexes are modeled by using time-dependent density function theory (TD-DFT). The calculated MLCT states correlate closely with the heretofore unknown emission properties that were observed experimentally. The hydrogen evolution mechanisms of three new series of cobalt based water splitting catalysts are modeled by Density Functional Theory (DFT). The three series include: 1) a family of cobalt complexes with pentadentate pyridine-rich ligands, 2) a family of three heteroaxial cobalt oxime catalysts, namely [CoIII(prdioxH)(4tBupy)(Cl)]PF6, [CoIII(prdioxH)(4Pyrpy)(Cl)]PF6, and [CoIII(prdioxH)(4Bzpy)(Cl)]PF6, 3) a pentadentate oxime that has ligand incorporated water upon metal coordination and is water soluble. These calculations provide reasonable interpretations of the experimental observations.

In Chapter 4, 5 and 6, mode selective fragmentation of ClCHO+ and circular migration of hydrogen in protonated acetylene with intense mid-IR laser pulses are simulated by BOMD trajectory calculations. The ionization rate of ClCHO in the molecular plane has been calculated by time-dependent configuration interaction with a complex absorbing potential (TDCI-CAP), and is nearly twice as large as perpendicular to the plane, suggesting a degree of planar alignment can be obtained experimentally for ClCHO+, starting from neutral molecules. The BOMD simulations demonstrate circularly polarized light with the electric field in the plane of the molecule deposits more energy and yields larger branching ratios for higher energy fragmentation channels than linearly polarized light with the same maximum field strength. The trajectories with different pairs of the dual laser pulses give very different branching ratios. The difference in branching ratios is even more pronounced when one of the two pulses started one quarter of the total duration earlier than the other vs. the other way around for the same pulse pair. In protonated acetylene, hydrogen migration around the C2 core occurs by interchange between the Y shaped classical structure and the bridged, T-shaped non-classical structure of the cation, which is 4 kcal/mol lower in energy. The linearly and circularly polarized pulses transfer similar amounts of energy and total angular momentum to C2H3+. There is an appreciable amount of angular displacement of the three hydrogens relative to the C2 core for circularly polarized light, but only an insignificant amount for linearly polarized light. This suggests a propeller-like motion of the three hydrogens is induced only by the circularly polarized light.

In Chapter 7, an inherent problem in BOMD is explained and mostly circumvented by using ADMP method. Since BOMD is based on the Born-Oppenheimer approximation, the wavefunction of the system is converged at each time step to calculate the force for integrating the classical equations of motion. This resulted in an artifact manifested for a few trajectories as anomalously large charge oscillations on an H atom (H+/H/H–) when it was well-separated (beyond ca. 3 Å) from the rest of the molecule, thus absorbing an anomalously large amount of energy. ADMP method, an alternative to BOMD method, propagates the density matrix using extended Lagrangian dynamics. Our ADMP calculations in intense laser fields show that the accuracy is similar to BOMD while the charge oscillation problem is eliminated naturally because the electronic wavefunction is propagated rather than converged at each step.

In Chapter 8, in order to interpret the experimental results of two-electron angular streaking for methane molecule, TDCI calculations and BOMD trajectory calculations are carried out with the additional help from a logistic regression machine learning algorithm to analyze geometric changes. The ionization angular dependence of various stable and meta-stable structures of methane cation, as well as neutral methane, is calculated by our TDCI-CAP approach. The ionization is mostly along the C–H bond direction for the neutral methane and monocation in the tetrahedral geometry, while the directions of ionization for other geometries are less straightforward. The relaxation time needed for neutral methane geometry (tetrahedron shape) to collapse into D2d and C2v structures on the cation potential energy surface is estimated to be 3 fs (half of the initial population converted) by classifying geometries along BOMD trajectories with a machine learning algorithm.