Multi-junction solar cells based on gallium and germanium have achieved solar to electricity conversion efficiencies of 46% in highly concentrated sunlight and an efficiency of 34.5% in ambient sunlight. However, these types of solar cells remain extremely expensive, and even the using highly concentrated sunlight which reduces the amount solar cell area required by a large factor, these cells have been unable to complete with the falling costs of conventional silicon solar cells.
A collaboration between researchers at The MASDAR institute of the Unite Arab Emirates and MIT has developed a dual junction solar cell which consists of a gallium arsenide phosphide cell grown on a silicon germanium substrate which is then bonded to a silicon substrate which acts as the bottom cell in the tandem cell design. The silicon germanium underlayer gives good performance properties to the gallium based top cell but blocks light from the bottom silicon cell. In the step cell design a patterning process is used to etch away part of the top cell exposing the silicon beneath directly to incoming solar radiation.
A news articlehas been published by the MASDAR institute describing some aspect of this dual junction cell design. An abstract of a Journal of Applied Physics paper about the theoretical performance limits of this cell design (38.7% max efficiency) is also available. The article and the abstract leave many puzzles in my mind about the physics of this cell and about the sources of cost reduction relative to the more usual designs of multi-junction gallium based cells.
Nevertheless the researchers are sufficiently enthused about the economic potential of this cell design that they are planning to create a startup company to try to commercialize it. Presumably the market will be in the currently moribund area of concentrated photovoltaics (CPV), which uses concentrating optics and dual axis trackers to produce electricity from high efficiency PV designs.