| Understanding electronic behavior in molecular and nano-scale systems is fundamental to the development and design of novel technologies and materials for application in a variety of scientific contexts from fundamental research to energy conversion. This dissertation aims to provide insights into this goal by developing novel methods and applications of first-principle electronic structure theory. Specifically, we will present new methods and applications of excited state multi-electron dynamics based on the real-time (RT) time-dependent Hartree-Fock (TDHF) and time-dependent density functional theory (TDDFT) formalism, and new development of the multi-configuration self-consist field theory (MCSCF) for modeling ground-state electronic structure. The RT-TDHF/TDDFT based developments and applications can be categorized into three broad and coherently integrated research areas: (1) modeling of the interaction between moleculars and external electromagnetic perturbations. In this part we will first prove both analytically and numerically the gauge invariance of the TDHF/TDDFT formalisms, then we will present a novel, efficient method for calculating molecular nonlinear optical properties, and last we will study quantum coherent plasmon in metal namowires using RT-TDDFT; (2) modeling of excited-state charge transfer in molecules. In this part, we will investigate the mechanisms of bridge-mediated electron transfer, and then we will introduce a newly developed non-equilibrium quantum/continuum embedding method for studying charge transfer dynamics in solution; (3) developments of first-principles spin-dependent many-electron dynamics. In this part, we will present an ab initio non-relativistic spin dynamics method based on the two-component generalized Hartree-Fock approach, and then we will generalized it to the two-component TDDFT framework and combine it with the Ehrenfest molecular dynamics approach for modeling the interaction between electron spins and nuclear motion. All these developments and applications will open up new computational and theoretical tools to be applied to the development and understanding of chemical reactions, nonlinear optics, electromagnetism, and spintronics. Lastly, we present a new algorithm for large-scale MCSCF calculations that can utilize massively parallel machines while still maintaining optimal performance for each single processor. This will great improve the efficiency in the MCSCF calculations for studying chemical dissociation and high-accuracy quantum-mechanical simulations. |
