High-Efficiency Spectrophotovoltaic System for Orbital Power Generation

Abstract

An analysis was performed to determine the economic feasibility of a concentrating Spectrophotovoltaic orbital electrical power generation system, capable of output power levels on the order of 100 kW. In this system dichroic beamsplitting mirrors are used to divide the solar spectrum into several wavebands. These mirrors consist of transparent substrates coated on the frontside with multilayer stacks of transparent dielectric materials. With this technique very high reflectivities can be achieved over a well-defined spectral region. Wideband antireflection coatings are applied to the back to insure good transmission of the nonreflected part of the spectrum. Absorption of the reflected and transmitted wavebands by solar cells with matched energy bandgaps increases the cell efficiency while decreasing the amount of heat which must be rejected. The optical concentration is performed in two stages. The first concentration stage employs a Cassegrain-type telescope, resulting in a short system length. The output from this stage, after being spectrally selected by dichroic beamsplitters, is directed to compound parabolic concentrators which comprise the second stage of concentration. In order to determine the optimum energy bands, ideal efficiencies for one-, two-, three-, and four-cell systems were calculated under 1000 sun, AMO conditions. The efficiencies ranged from 32.2% for a one-cell system to 52.8% for a four-cell system. Various combinations of Ge, Si, GaAs, and GaP were chosen to calculate more reasonable efficiencies. Account was taken of losses due to grid shadowing, cell reflection, dichroic mirrors and nonunity spectral response. Oneand four-cell systems were found to be less efficient than most of the twoand three-cell systems which had efficiencies on the order of 2697o-33%. A cost tradeoff analysis was performed to determine an optimum system configuration. The primary mirror was found to be responsible for 70%-85% of the system cost. Neglecting solar cell development cost, the twoand three-cell systems were found to be the most cost effective. The optimum system consisted of an jj/3.5 optical system, a beamsplitter to divide the spectrum at 0.9 /im, and two solar cell arrays, GaAs and Si. For a 100-kW power level, with solar cells maintained at a temperature of 300 K, the primary mirror diameter was calculated to be 20.43 m and the projected cost per peak watt of this system is $2.52/Wp.