Concentrating the Sun

Jerry R. Kukulka, Dr. Richard R. King and Dr. Nasser H. Karam

Researchers worldwide are looking for the combination that will unlock the door to more efficient solar energy. A variety of approaches are being tested that, hopefully, will raise efficiencies to a level that will bring solar to parity with conventional power generating systems.

Figure 1. Concentrator receivers from Solar Systems Pty. Ltd. of Hawthorn, Australia, are designed for utility application at 400× to 500×. Courtesy of Solar Systems Pty. Ltd., Australia.

One such technique uses concentrated photovoltaic collector technology (CPV) that tracks the sun and either reflects or refracts its light, concentrating it 200 to 1000 times onto small solar cells. Although this has the disadvantage of requiring large areas of real estate, it nevertheless has shown promise for high-efficiency, low-cost power generation.

Figure 2. A concentrator receiver from Amonix Inc. of Torrance, Calif., is designed for utility application at 250 to 400×. Courtesy of Amonix Inc.

At the heart of the technology is the multijunction solar cell, a tiny device that can convert more than 40 percent of the sun’s energy into electrical energy. This works because modern multijunction solar cells can exploit the solar spectrum more efficiently than other technologies (Figure 3).

Figure 3. A multijunction solar cell composed of multiple layers of various semiconductor materials can convert more than 40 percent of incoming sunlight into usable electricity.

Splitting up the spectrum

Multijunction concentrator solar cells are based on III-V semiconductors – the same (GaAs) material used to make LEDs. The cells achieve their high efficiency by combining several solar cells, or p-n junctions, into a multijunction cell. Each cell in the multijunction structure, called a subcell, is composed of a different semiconductor material (absorber), each with a different energy bandgap that absorbs light in a different portion of the solar spectrum.

The subcells are electrically connected in series and are positioned in optical series such that the highest bandgap is in the top subcell, which absorbs the highest energy photons in the solar spectrum but does not absorb lower energy photons (lower than its bandgap).

Photons with energy lower than the bandgap of the top cell are transmitted to the subcell beneath, which has the second highest bandgap. Each subcell is connected to the next in the stack using a “tunnel junction,” a thin insulating barrier between two conducting electrodes that is optically transparent and electrically conducting. In this way, the multijunction cell divides the broad spectrum into wavelength bands, each of which can be used more efficiently by the individual subcells than in the single junction case.

By optimizing both the number of wavelength bands (subcells) and the size of the bands, the maximum current generation in each band is achieved and, by minimizing the optical and electrical losses, the conversion efficiency of the CPV cells can be increased to greater than 50 percent.

While it may seem paradoxical that a relatively complex system can deliver lower electricity costs, the answer is straightforward. CPV systems use relatively small quantities of cells inside a larger optical system that tracks the sun. Volume production of multijunction cells can be expected to deliver cells of nearly 40 percent efficiency at a cost of $5/cm², resulting in a receiver that represents 20 to 30 percent of the entire cost for most CPV systems.

This would allow power costs to fall as low as 12 cents per kilowatt hour. A decrease in Balance of Systems (BOS) costs for the larger systems (such as central receiver stations) will reduce the cost of energy even further. However, the solar cell efficiency is much more a factor for leveraging costs than the BOS for reducing the price of generating electricity and, hence, drives significant development activities.

In the past eight years, the output power of terrestrial III-V multijunction cells used in solar concentrators has doubled compared with single-junction GaAs cells. The lattice-matched three-junction GaInP/GaInAs/Ge cell used in CPV is the result of innovation and refinement over the past 10 years.

This cell structure is created by growing as many as 20 thin, single-crystal layers of III-V materials onto a single-crystal Ge wafer that is 100 mm in diameter. Metallorganic vapor phase epitaxy is employed in growing the layers. The requirements for crystal growth quality, thickness and doping concentration uniformity are extraordinarily high. Growing the structure of a multijunction solar cell is equivalent to precisely and uniformly covering a football field with 20 layers of snow, ranging in thickness from ? to less than 1/1000 of an inch.

The efficiency of this type of cell can be further improved with metamorphic multijunction solar cell design, a process in which crystal dislocations form in a metamorphic buffer – a region with a graded semiconductor composition, grown first on the substrate. The crystal structure relaxes so that a new, larger lattice constant is reached. This structure then can be used as a virtual substrate for subsequent semiconductor growth at the larger lattice constant. These layers can take better advantage of their various bandgaps.

Recent advances in III-V metamorphic multijunction solar cell design for terrestrial concentrator use have resulted in even greater photovoltaic conversion efficiency. Metamorphic three-junction concentrator cells developed by Spectrolab Inc. of Sylmar, Calif., have reached a record efficiency of 40.7 percent under the terrestrial solar spectrum, the first photovoltaic cell of any type to achieve the 40 percent-plus milestone.

Global research continues on advanced multijunction cell architectures. Incorporation of metamorphic semiconductor materials and increased numbers of junctions – among other device structure advances – can be expected to increase practical concentrator cell efficiencies to 45 or even 50 percent. By using efficiency as a powerful lever to bring down the cost, it will become possible to deploy solar cell technology on a larger scale and to change the way we generate electricity.

Meet the authors

Jerry R. Kukulka is chief engineer, Richard R. King is principal scientist, and Nasser H. Karam is vice president of advanced technology products, all at Spectrolab Inc. in Sylmar, Calif.; e-mail: jkukulka@spectrolab.com; rking@spectrolab.com; nkaram@spectrolab.com.