Optoelectronic Design and Prototyping of Spectrum-Splitting Photovoltaics

Citation

Lloyd, John Vickery (2018) Optoelectronic Design and Prototyping of Spectrum-Splitting Photovoltaics. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/G1CG-E962. https://resolver.caltech.edu/CaltechTHESIS:05212018-185007265

Abstract

Global energy production is dominated by the combustion of fossil fuels but in order to avoid the projected consequences of anthropogenic climate change it is necessary that humankind reduce the carbon intensity of its energy supply. Fortunately the sun supplies a ubiquitous flow of energy of with excellent thermodynamic quality to earth. Massive investment and manufacturing scale has driven the costs of photovoltaic systems to levels competitive with fossil fuel generation, and yet commercial photovoltaic systems convert power from the sun into electricity with less than 20% efficiency. In this thesis we consider the thermodynamic and practical limits to the power conversion efficiency of photovoltaic systems and seek to design systems that address the greatest sources of loss, namely the lack of sub-bandgap absorption and the thermalization of excited carriers. We present several designs of spectrum-splitting systems that utilize optical structures to allocate incident broadband solar radiation into narrower spectral bands which can be converted by multiple distinct photovoltaic cells at greater efficiency. Furthermore, we report on the design and fabrication of thin film III-V single-junction cells at bandgaps spanning the solar spectrum for incorporation within spectrum-splitting systems. These devices were fabricated by utilizing epitaxial lift-off processes from both GaAs and InP wafers as proof of scalability. We additionally report on the fabrication and characterization of series of a spectrum-splitting prototypes. This design featured seven distinct spectral bands with single-junction photovoltaic cells designed to convert them with highest possible efficiency, and the ultimate prototype exhibited an 84.5% spectrum splitting efficiency and 30.2% power conversion efficiency under a standard AM1.5D solar spectrum. We also report a technical pathway to raise the prototype efficiency to a record breaking 45.2%. Finally, we present an optical design of a spectrum-splitting module that is informed by a technoeconomic analysis which drastically reduces the complexity and cost relative to the fabricated prototype.

Thesis Files