Skip to main content

GaP-Si solar cells achieve significant efficiency at lower cost

By Doug Main March 12, 2018
GaP-Si solar cells
Schematic showing the various layers of indium-tin oxide (ITO), silver (Ag), silicon (n-Si), gallium phosphide (GaP), molybdenum oxide (MoOx), and other components of two versions (a and b) of solar cells. The panels containing molybdenum were the most efficient. Credit: Journal of Materials Research

Silicon-based solar cells max out at an efficiency of 33.5%, which is governed by the Shockley-Queisser limit. For decades, researchers have been trying to find ways around this maximum efficiency. One promising method is to integrate silicon with III-V semiconductors. III-V materials are made of two components: an element in group III of the periodic table (such as Al, Ga, and In), and a group V element (such as N, P, As, and Sb). Because of their intermediate polarity and other traits, they are very efficient, and ideal for optoelectronics and high-speed electronics.

The Shockley-Queisser limit exists for silicon solar cells because only photons with a certain range of energy can be absorbed due to silicon’s bandgap. When III-V materials are integrated with silicon, they absorb lower energy photons in addition, increasing the overall efficiency.

Experimental devices containing III-V materials and using concentrated sunlight have achieved efficiencies of over 45%. But such devices are difficult and expensive to make. 

Scientists have now found a new way to incorporate III-V compounds with silicon to create a functioning solar cell, taking a step toward photovoltaics with high efficiencies. As described in an article published in a recent issue of the Journal of Materials Research, a team of scientists including Chaomin Zhang and Christiana Honsberg at Arizona State University integrated III-V materials into silicon photovoltaics as a series of thin layers.

The trouble with integrating semiconductors and silicon is that they have different electrical properties and different lattice constants. But gallium phosphide (GaP), the III-V semiconductor used in this study, has a slightly more similar molecular arrangement to silicon, Honsberg says.

The team found a way to tune the interface properties between the two substances such that only one charge carrier can cross the boundary between the materials, enabling completion of the circuit. They did this by putting a hole-selective contact on the front of the solar cell, and an electron-selective contact on the back.

The researchers created several different solar cells, using slightly different combinations of materials. The most efficient version has a layer of molybdenum oxide (that acts as a hole-selective contact) on its top, while a GaP layer toward the bottom acts as the electron-selective contact. This means that the GaP layer produces conduction electrons, while the molybdenum layers act to receive these carriers. The advantage of this design is that it does not rely on a typical p-n junction, which in the case of silicon is subject to the Shockley-Queisser limit. The design bypasses this limitation and opens the door for substantially higher efficiency solar cells, according to Honsberg.

Research on so-called carrier-selective contacts has become an important and emerging area in photovoltaics, according to Seth Hubbard, director of the NanoPower Research Laboratory and professor at Rochester Institute of Technology, who was not involved in the study.

To create photovoltaic panels, the research team started with wafers of silicon four inches in diameter. The researchers evaporated molybdenum oxide on top, allowing a 9 nanometer-thick layer to form. They added GaP to the underside using migration-enhanced epitaxy which deposits single crystals of GaP onto the silicon in a vacuum. The researchers also added a layer of indium-tin-oxide and silver to both sides of the wafer using RF sputtering. These layers allow electrons to reach silver electrodes implanted in the panels, which connect the GaP layer with the hole-selective contact (molybdenum oxide), completing a circuit.

The molybdenum panels reached an efficiency of 14.1%. That is still less than one third that achieved for other experimental III-V photovoltaics, but these are expensive and unlikely to be commercialized, whereas the currently described silicon chips could be commercialized perhaps in the next five years, Honsberg says.

This is the first demonstration of the device, and the efficiency is likely to greatly increase in the future with further modifications to and optimization of the design, Honsberg adds. “Hopefully we can match the efficiencies of present III-V devices but retain the cost advantages of silicon devices.”

Read the abstract in the Journal of Materials Research.