Tandem solar cell structures are the only strategy that has been shown to exceed the detailed balancing efficiency limit of high-quality single-junction solar cells. To continue to improve the efficiency of cost-effective terrestrial solar power generation, many people around the world are considering hybrid tandem connections of different subcells, especially designs with silicon solar cells as bottom subcells. In this project, we investigated the possibilities of multiple tandem designs, including those with three-terminal (3T) and four-terminal (4T) configurations. The use of 3T and 4T designs can be used for efficient and economical hybrid tandem designs, utilizing the best subcellular materials, such as emerging perovskite materials. In particular, the three-terminal configuration has not been fully studied before. We have laid the groundwork for understanding the operation of 3T tandem in this project: developing a taxonomy for naming, a method for measuring and interconnecting, and a model that simply describes 3T tandem. The coupling between electrical and optical subcells is also measured and modeled.
An important part of this work is the fabrication of new example tandem structures, including 4T GaAs/Si, 3T GaInP/Si, 3T GaAs/Si, and 3T GaInP/GaAs devices. Using these high-quality tandem cells, we have been able to clearly demonstrate efficient achievability, as well as subtle physical effects such as photon cycling and luminescence coupling. We have developed and demonstrated basic building block tools such as transparent conductive adhesives (TCAs) and 3T silicon bottom cells with crossed back contacts (IBCs), which can also be used in many other tandem designs. We have tested the reliability of these tools and equipment under standardized tests and outdoor measurements. We found that 4T GaAs/Si tandems are relatively easy to manufacture and robust under real outdoor conditions. Although we have demonstrated working mixing 3T III-V/TCA/SI IBC tandem, we have experienced low yields even with our best processes. Further work is needed to increase the processing yield of these devices.
Therefore, we also created tandem cells using the full III-V3T tandem process, which is very powerful and yielding, allowing the use of 8 nearly identical 3T tandems to create voltage-matched strings in many different configurations. Using these robust 3T tandem examples, we are able to measure and precisely characterize 3T tandem behavior to predict their operation over varying spectra and temperatures. The optoelectronic equivalent circuit model is very versatile, suitable for hybrid series connections, and includes operating 3T Si IBC units. This generic model has been released to the public as PVcird, an open-source software based on python. We calculate the impact of these new tandem device designs on real-world energy production and show that the relative performance of different tandem configurations is contextual and can be designed using the tools developed here.
