Thin film solar cells made using Cu(InGa)(SeS)2 absorber layers have achieved module efficiencies over 17% and are in the early stages of commercialization using a metal precursor/H2Se/H2S process. We have designed and are implementing a high-throughput, rapid thermal processing (RTP) system to produce the Cu(InGa)(SeS)2 films and have characterized absorber layers and solar cells. Our method, which is a variation of the precursor reaction method most commonly used in industry, greatly reduces residence time and eliminates toxic H2Se from the manufacturing process.
The precursor reaction method consist of a two-step process: (1) deposition of a metallic Cu-In-Ga precursor on a substrate, usually Mo-coated glass, and (2) reaction of the precursor with selenium and sulfur sources, usually H2Se and H2S at 400—550 °C, to form a Cu(InGa)(SeS)2 film. Precursor reaction methods are desirable since commercially available sputtering systems can be used to deposit the metal precursor uniformly over areas >1m2 which is required for flat panel solar modules. However, conventional precursor reaction methods are limited by two challenges: (1) H2Se is extremely toxic with a threshold limit for human exposure of 50 ppb; (2) the reaction kinetics are very slow requiring residence times of about 1 hour to completely react the precursor. An alternative approach is to deposit a layer of selenium on the metallic precursor and react the “selenium-capped” precursor in an H2S environment. In addition to the improved safety afforded by such an H2Se-free process, this approach also yields films with much better adhesion than obtained with H2Se/H2S-reacted precursors. Because of the improved adhesion, the films can withstand reactions at higher temperature, which allows residence times of less than 10 minutes.
In this presentation, we will discuss results from our investigation of Cu(InGa)(SeS)2 films produced at temperatures greater than 550 °C from Se-capped precursor films. A factorial experimental design was used to determine which process variables—temperature ramp rate, dwell temperature, Se cap thickness, and H2S concentration—affect material properties, specifically material composition and crystallinity as characterized by energy dispersive x-ray spectroscopy and x-ray diffraction, respectively. Dwell temperature and H2S concentration were determined to be significant factors affecting material composition. Devices were fabricated and current-voltage behavior measured under standard test conditions and the highest efficiency we have obtained to date is 12.9%.