Investigation of spectrum splitting systems for concentrating and one-sun photovoltaics

Spectrum splitting techniques have the potential to improve the conversion efficiency of multijunction cells. The spectral separation of the light redirects the part of the spectrum that is usually wasted in the TJ-cell to the Si-cell. Moreover, concentrating the solar irradiance reduces the solar cell area while improving the conversion efficiency. This concept was developed into two spectrum splitting prototypes. The concentrator prototype was built as a "proof of technology" for a concentrating photovoltaic (CPV) power tower system. The minimodule prototype was designed to use the light-trapping effect of prism concentrators to maximise the conversion efficiency. This thesis uses ray-tracing simulations to evaluate the benefits and challenges of spectrum splitting architectures in commercial systems under real operating conditions. For the spectrum splitting power tower, the simulations studied the effects on the system efficiency for different spectral irradiances, concentration levels, flux distribution, and angle of incidence of the rays. The annual average efficiency and energy yield calculations showed that the spectrum splitting CPV receiver should be used in small to medium size systems. Additionally, the minimodule simulations investigated the angular tolerance and conversion efficiency for different prism geometries. The total internal reflection (TIR) effect is the dominant factor on the minimodule system efficiency. Simulations showed that the TIR effect on the minimodule faces could disrupt the flux distribution among the cells, leading to efficiency losses. The different incidence angles magnify this effect as the sun moves along the sky. The findings of this work clearly showed that spectrum splitting systems should be optimised for conversion efficiency rather than power output. Moreover, the complex optical path (optical element - filter - Si cell - TJ cell) must be followed carefully to reap the benefits of the spectrum splitting techniques