Machine Learning and Novel Solar Cell Concepts Group

The group researches novel concepts for solar cells, with a particular focus on perovskite/silicon tandem devices. The Simulation team develops a web-based PV simulation platform (XSolar-Hetero) to upload/simulate/compare measurements and to apply machine learning to the collected data, with a focus on the simulation of perovskite solar cells and perovskite/silicon tandems (in cooperation with the Emerging Solar Technologies Group). Novel structures investigated are local-area double-side contact passivated (“biPoly”) silicon bottom cells and various tunnel junctions to couple the silicon and perovskite sub-cells. The machine learning activities aim at extracting various device parameters from PL images of silicon solar cell precursors. The Processing team (until Apr 2021) developed silicon solar cells for use as bottom cell in 2-terminal perovskite/silicon tandems, with a focus on biPoly cells and on abcPoly cells (which have both contacts at the rear).