A GaInP/Ga(In)As/GaNAsSb/Ge 4J solar cell grown using the combined MOVPE + MBE method is presented. This structure is used as a test bench to assess the effects caused by the integration of subcells and tunnel junctions into the full 4J structure. A significant degradation of the Ge bottom subcell emitter is observed during the growth of the GaNAsSb subcell, with a drop in the carrier collection efficiency at the high energy photon range that causes a~15% lower J sc and a V oc drop of~50 mV at 1-sun. The V oc of the GaNAsSb subcell is shown to drop by as much as~140 mV at 1-sun. No degradation in performance is observed in the tunnel junctions, and no further degradation is neither observed for the Ge subcell during the growth of the GaInP/Ga(In)As subcells. The hindered efficiency potential in this lattice-matched 4J architecture due to the degradation of the Ge and GaNAsSb subcells is discussed. At this stage of development, it is pertinent to have a closer look at the integration of components forming the 4J structure. For example, the insertion of the dilute nitride subcell in the 3J structure brings about added thermal loads to the Ge bottom subcell, an effect accentuated by the fact that dilute nitrides usually require an annealing step to improve their electronic properties. The dilute-nitride subcell itself suffers annealing during the growth of the upper subcells, which could be expected to have an impact on its performance. The focus has to be put in identifying and quantifying these effects, and redesigning the growth process to minimize their impact.In this work, we present a detailed characterization of our 4J solar cell, based on a prototype structure achieved by a combination of MOVPE + MBE growth methods, and using a GaNAsSb junction.This solar cell succeeds in integrating 4 component junctions into a monolithic, lattice matched structure, but it is still far from being a high efficiency device, mainly due to sub-optimum characteristics of the dilute nitride subcell. However, the device provides valuable insight into the integration of the dilute nitride subcell into a 4J solar cell. We focus mainly on the effect of thermal load on the performance of the Ge and GaNAsSb bottom subcells, and the tunnel junctions. It is found that significant losses are at stake, mainly in the Ge and GaNAsSb subcells, which can limit the potential of this 4J solar cell structure to compete in efficiency with other architectures.

No abstract

Virtual substrates based on thin Ge layers on Si by direct deposition have achieved high quality recently. Their application to high efficiency III-V solar cells is analyzed in this work. Replacing traditional Ge substrates with Ge/Si virtual substrates in standard lattice-matched and upright metamorphic GaInP/Ga(In)As/Ge solar cells is feasible according to our calculations using realistic parameters of state-of-the-art Ge solar cells but with thin bases (< 5µm). The first experimental steps are tackled by implementing Ge single-junction and full GaInP/Ga(In)As/Ge triple-junction solar cells on medium quality Ge/Si virtual substrates with 5µm thick Ge layers. The resultsshow that the photocurrent in the Ge bottom cell is barely enough to achieve current matching with the upper subcells, but the overall performance is poor due to low voltages in the junctions. Moreover, observed cracks in the triple-junction structure point to the need to reduce the thickness of the Ge + III-V structure or using other advanced approaches to mitigate the thermal expansion coefficient mismatch effects, such as using embedded porous silicon. Next experimental work will pursue this objective and use more advanced Ge/Si virtual substrates available with lower threading dislocation densities and different Ge thicknesses.Index Terms -III-V multijunction solar cell, germanium buffer, low cost substrate, virtual substrate. This is an accepted proceedings paper presented at IEEE PVSC 2019, in press.

This paper summarizes the state-of-the-art of the lattice matched GaInP/Ga(In)As/Ge triple-junction solar cell developed at the Solar Energy Institute of UPM (IES-UPM). Different research topics tackled over the last years about this structure are described. As result of this work, an efficiency of-40% at ~-415>i is presented.

No abstract