Next-Generation Multi-Junction Photovoltaic Design Paradigms and Adaptive Optics Techniques for Telecommunications Applications and the Global Energy Transition

Abstract

Multi-junction photovoltaic devices offer the best performance in both solar and laser-based photovoltaic applications but have greater complexity and higher cost than their single-junction counterparts. In this thesis I demonstrate several pathways to improve the applicability and competitiveness of multi-junction photovoltaic devices. First, I investigate multi-junction solar cells and their application to the global energy transition and niche high performance domains. I start by studying silicon-germanium-tin as a potential material to extend the indium gallium phosphide/indium gallium arsenide/germanium lattice-matched multi-junction solar cell material system. That project aims to leverage the maturity of the three-junction design while boosting device efficiency by improving current-matching through introduction of the silicon-germanium-tin subcell. Next, I examine subcell segmentation, a nascent multi-junction solar cell design paradigm. I elucidate how subcell segmentation eases the simultaneous lattice and current matching requirements for multi-junction solar cells, greatly increases the number of bandgap combinations which can achieve high performance, reduces junction growth constraints, and unlocks the potential for operation at the most extreme solar concentration factors. I then shift focus to multi-junction photonic power converters and laser-based applications. I introduce machine learning enhanced design of ten-junction photonic power converters operating at a telecommunications wavelength of 1550 nm. I show that the novel machine learning enhanced approach greatly increases the number and variety of optimal designs, reduces computational expense, and simultaneously offers a new pathway to discover and understand fundamental physics in the design space. Finally, I detail development of a cost-effective, infrared-capable adaptive optics technique. I demonstrate how a single-pixel camera with compressive sensing could be used as input to adaptive optics to overcome fast signal fading in free-space telecommunications and power-by-light systems.