III-Sb-based solar cells and their integration on Si

Overview & Synopsis

III-Sb materials have demonstrated their potential for multiple opto-electronic devices, with applications stretching from communications to environment. However, they remain an almost unexplored segment for classical photovoltaic systems. In this research, we intend to demonstrate that III-Sb-based devices are promising candidates for high-efficiency, low-cost solar cells. Their benefits are two-fold: not only do they offer a wide range of lattice-matched alloys and low-resistivity tunnel junctions, but they also enable direct growth on Si substrates. We thus investigate the building blocks of a GaSb-based multi-junction solar cell integrated onto Si. First, we develop the photovoltaic growth and processing by fabricating homo-epitaxial GaSb cells. Intensity-voltage (J-V) measurements approach the state of the art with 1-sun efficiency of 5.9%. Then, we integrate a GaSb single-junction cell on a Si substrate by molecular beam epitaxy (MBE). X-ray diffraction (XRD) and atomic force microscopy (AFM) analysis show structural and morphological properties close to the best reported in the literature for similar metamorphic buffers. We further adapt the cell configuration to circumvent the high defect density at the GaSb/Si interface. The heteroepitaxial cell results in a reduced efficiency of 0.6%. Nevertheless, this performance is close the most recent advancements on GaSb heteroepitaxial cells on GaAs, despite a much larger mismatch. Last, we investigate the epitaxy of AlInAsSb. This alloy could in theory reach the widest range of bandgap energies while being lattice-matched to GaSb. However, it presents a large miscibility gap, making it vulnerable to phase segregation. AlInAsSb only counts few experimental reports in the literature, all referring to unoptimized growth conditions and abnormally low bandgap energies. We successfully grow good-quality layers with Al composition x_{Al} ranging from 0.25 to 0.75, showing no macroscopic sign of decomposition. Yet, transmission electron microscopy (TEM) observations point to nanometric fluctuations of the quaternary composition. Photoluminescence (PL) data is studied to determine the alloy's electronic properties. We eventually propose and fabricate a tandem cell structure, resulting in 5.2% efficiency. Quantum Efficiency (QE) measurements reveal that the top subcell is limiting the tandem performance. Numerical fits to both J-V and QE data indicate improvement paths for each building block.