Abstract
The studies presented in this thesis employ electronic structure calculations and wavefunction analysis to identify plasmonic excited states in various systems. Additionally, the tools for analyzing time-derivative non-adiabatic couplings (NACs) were developed, and the calculation of NACs using semiempirical methods was implemented. Although many methods for classifying plasmonic excited states exist, classification methods must adhere to the definition closely. If only a single property is measured during classification, it may lead to classifying an excited state without accounting for all the characteristics of a plasmon. Therefore, properly classifying plasmons requires accounting for each characteristic, as they are a collective and coherent oscillation of the conduction band electrons.
The studies on silver nanowires and nanorods agree with previously established trends. Furthermore, the plasmon identification criteria developed in this thesis demonstrate how the size and shape of silver nanowires and nanorods control the character of the prominent absorbing longitudinal and transverse peaks. The excited states making up the peak areclassified as plasmon-like, collective, interband, or single-particle states. Similar approaches were used to study graphene-based systems, and the results agree with previous studies in the literature. Longitudinal and transverse plasmon-like modes overlap for armchair edges much more than for zig-zag edge-terminated graphene-based systems. The longitudinal and transverse modes for zig-zag edge-terminated systems have almost no overlap in the absorption spectra.
Moreover, progress was made toward studying hot carrier relaxation using plane-wave DFT calculations by systematically varying the kinetic energy cutoff when calculating molecular trajectories and comparing the calculated NACs. This study provides insight into the effect of the kinetic energy cutoff on NACs and the acceptable ranges for qualitative and quantitatively agreeable results with reference plane-wave DFT calculations. This allows practitioners to make an informed decision when choosing parameters for plane-wave DFT calculations to calculate NACs for non-adiabatic molecular dynamics (NAMD) simulations.
Finally, semiempirical methods were utilized for speed and fast calculation of configuration interaction (CI) states for NAMD simulations. If studying a large molecule or many excited states is desired, using time-dependent density functional theory (TD-DFT) for CI states becomes cost-prohibited. From data available in the literature, results from the INDO NAMD simulations are comparable with those from time-dependent density functional theory tight-binding (TD-DFTTB).
In conclusion, this work contributes significantly to studying and further understanding plasmons in small systems. The strides made toward accurately studying plasmon-like excited state dynamics using plane-wave DFT and the excellent agreement between INDO and DFT NACs demonstrate excellent progress in this field. Nonetheless, there is still room forinnovation in extracting plasmon-like characteristics from NAMD simulations and quantifying them.