Spectral-splitting concentrating photovoltaic-thermal (CPV-T) systems are solar energy harvesting systems capable of concurrently absorbing heat and generating electricity. This dissertation contributes to the development of spectral-splitting CPV-T systems by assessing their performance and environmental impacts, optimizing nanoparticle-based optical filters, and analyzing economic feasibility, the specifics of which are described in detail below. In Chapter 2, a simulation of a small-scale spectral-splitting CPV-T system utilizing a nanofluid-based optical filter was developed. In this simulation, the long-term energy generation by the photovoltaic system and thermal energy absorption by the nanofluid filter were obtained, and the latter was used in a household water heating system. Additionally, a parametric analysis was performed to assess the effects of different parameters on the performance of the system, and an environmental impact analysis provided an understanding of the carbon dioxide offset generated by the proposed system. The simulations indicate that the proposed PV-T system can offset a total of 1.317 tons of carbon dioxide per year per household. In Chapter 3, the essential task of finding the best nanofluid-based optical filters for spectral-splitting PV-T systems was accomplished. A multiparticle optimization routine was developed for selection of the best nanoparticle parameters, such as volume fraction and size, for three different base fluids (water, ethylene glycol, and Therminol VP-1) and solar cells (including Si, GaAs, and GaInP/GaAs). Efficiency values near 40% were obtained for Si and GaAs solar cells based on the filter efficiency metric, and a lowercost nanofluid solution for Si was proposed. Novel insights on the parameters affecting the plasmon resonance and damping of indium tin oxide nanocrystals were also detailed. In Chapter 4, an assessment of the cost of implementation of CPV-T power plants using the spectral-splitting technique was performed. Three types of solar power plant were analyzed, including a purely thermal parabolic trough, a hybrid Si-based, and a hybrid GaAs-based power plant. The levelized cost of energy (LCOE) algorithm included inputs obtained from random variables to obtain LCOE values and energy generation for each type of power plant as a probability distribution. The results indicated that a thermal-only solar power plant is more economical at lower solar multiple values, and hybrid systems have lower energy costs at higher solar multiples. In summary, the analyses and results described in this dissertation aim to provide a deeper understanding of the underlying physics of spectral-spliting CPV-T power conversion, and facilitate their real-world implementation.