Module integration of III-V compound semiconductor/silicon two-junction solar cells with three and four terminals
Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering
Final Report Abstract
During my reseach fellowship at the National Renewable Energy Laboratory (NREL), Golden, Colorado, USA, I investigated the interconnection of III-V compound semiconductor/silicon two-junction solar cells. Photovoltaic modules with multijunction solar cells convert solar irradiation more efficiently to electrical energy than present modules. The multijunction cells consist of two or more subcells, which are stacked on top of each other. The annually generated energy yield depends on the one hand on the properties of the subcells and on the other hand on their interconnection. It is possible to increase the energy yield of tandem devices by up to 20% if the cells are not connected in series, but contacted separately. The latter can be achieved by using tandem cells with three (3T) or four terminals (4T). However, there is a lack of adequate interconnection concepts especially for 3T devices, which can reach similar high efficiencies as 4T devices. To close that gap is the aim of this research project during the fellowship. With an analytical model, I analyzed the performance of interconnected solar cells as function of the top cell’s band gap of the top cell assuming a Si bottom cell. The model predicted an optimal band gap for 4T strings of 1.81 eV with a maximum efficiency ηmax of 37.8% and infinite string of 3T devices ηmax of 37.8% at 1.80 eV. This shows that the results that 3T cells perform as well as 4T devices is valid also on module level for voltage matched cells. The maximum efficiency for 2T devices is 37.4% at 1.67 eV. 3T devices show the same robustness against spectral changes as 4T devices. To investigate the increase in annual energy yield of 3T devices compared to 2T devices, an energy yield model was developed. It includes spectral as well as temperature effects. The temperature is determined using outdoor data of existing modules, which is then rescaled by the efficiency of the module I aim to predict. In case of a GaInP on Si device the energy yield is increased by 5.7% comparing 3T and 2T devices. 4T devices produce 8.5% more energy than 2T devices. In conceptual works, interconnection schemes were developed and analyzed with network simulation. Losses occurring at the string ends of 3T devices can be omitted using an alternating arrangement of two different tandem devices using the same material system. Further, a 3D interconnection scheme was proposed. A promising approach for 3T devices are tandem devices built with interdigitated back contact cells (IBC) as bottom cells, which feature the contacts for both polarities on the rear side. For 3T devices the IBC cells have an additional front contact. Compared to a middle contact, they have the advantage that no grooves through one junction or no intermediate contact grid are required. To the best of my knowledge, no experimental 3T tandem devices with a 3T bottom cell was ever published, even though it has been proposed and simulated in various papers. A plain process sequence for these 3T tandem devices was developed. Instead of fabricating the III-V top cell on a separate handle, it is processed directly on the Si bottom cell. This enables to reduce optical and electrical losses and the material consumption. A first working proof-of-concept device reaching 20.2% efficiency was realized.
Publications
- “III-V/Si tandem cell to module interconnection – comparison between different operation modes”, Proceeding of the 45th IEEE Photovoltaic Specialists Conference, Washington D.C., USA, 2017
Henning Schulte-Huxel, Emily L. Warren, Manuel Schnabel, Paul Stradins, Daniel Friedman, Adele C. Tamboli
(See online at https://doi.org/10.1109/PVSC.2017.8366558)
