Thermal photovoltaics see first commercial light

Thermal photovoltaics see first commercial light
DENVER — Though a nascent and still largely defense-funded field, thermal photovoltaics (TPV) is beginning to spawn some practical applications.

The first commercial TPV-based product has already hit the market: an emergency 30-watt power source for charging a sailboat's navigation-equipment batteries from JX Crystals Inc. (Issaquah, Wash.), introduced this summer. The power generator is the first in a line of "Midnight Sun" products that marks the evolution of TPV away from a strict dependence on military funding and applications.

The technology is also being eyed by electric-vehicle researchers who view it as a compatible energy source for battery-powered vehicles. Gallium antimonide (GaSb) TPV cells have been built into the power system of a hybrid electric vehicle designed and built at Western Washington University (Bellingham, Wash.). Called the Viking 29, the electric prototype uses 600-watt TPV generators to charge 360 volts of NiCd (nickel-cadmium) batteries. Natural gas is the fuel for the TPV generator and the main power plant is a 75-kW (100-hp) electric motor which is more than 90 percent efficient throughout its operating range. Without the TPV system, the car has a range of 50 miles. With it, the range is increased to more than 200 miles.

Acceleration figures for the vehicle were not available, but Michael Seal, professor at Western Washington, said, "This isn't your typical electric vehicle. It is a very high-performance car."

TPV-based design
These developments, and other TPV-based designs, are being presented this week at the 4th Thermophotovoltaic Generation of Electricity Conference, sponsored by the National Renewable Energy Lab, which is seeing a record number of papers.

Unlike the more familiar solar-cell technology, TPV systems convert light at the infrared frequencies generated by hot objects. Because these IR radiators are heated by readily available fuels like natural gas or propane, TPV could be an important source of electric power for a variety of terrestrial systems.TPV systems generally consist of a ceramic radiator heated to a temperature somewhere between 1,200°C to 1,500°C to produce infrared energy, which is then directed toward photovoltaic arrays that are optimized to convert the radiation directly into electricity. The radiator might be broadband or selective in its optical characteristics, each approach having pros and cons. There might also be a thin-film array of micro antennae applied to the surface of the TPV cells to convert a broadband IR spectrum into a highly selective spectrum of very limited wavelengths, nearly perfectly tuned to the bandgap of the specific semiconductor material used in the arrays.

Not everyone is convinced that the complexity of such systems is worth it, however. "It's hard to imagine how quaternary [compound semiconductor] technology gets cheap enough to put in your home, although maybe there are applications for it in the military and other price-insensitive areas," said Jason Keyes, manager of business development for JX Crystals. "Our goal is to make TPV cheap enough so that if you can generate a few hundred watts, you can do it at $1 [a] watt. We're not there yet, but we know how to get there." His company will present a paper describing the next product in its Midnight Sun series, a propane- or natural-gas-fired stove that generates 25,000 Btu and 100 watts. This stove, like the company's sailboat battery charger, uses GaSb-based TPV cells.

It is designed for use in homes not connected to the electricity grid, or as a backup power and heat supply in case of power outages. The stove can be used either as a fireplace insert or as a vented heater located in the corner of a family room. The system is in the prototype stage and Keyes said the price target is $2,000.

The more familiar solar-cell technology has moved farther along the road to low-cost commercial applications because silicon has a bandgap (1.14 eV) well-matched to the spectrum of sunlight. Indeed, the great majority of solar cells are made from single-crystal or amorphous silicon, although various III-V and II-VI material combinations are used to make solar cells for space and other applications.

Yet TPV cells are not subject to the vagaries of sunlight, and even though their energy-conversion efficiencies may be lower at present, they have the potential to generate significantly more power for a given size than solar cells. The main reason is that TPV cells are positioned only 1 to 2 cm away from the radiative source, compared with the 150 million kilometers, give or take, between the sun and the earth. The TPV environment is photon-rich and so the power density is higher, even though the temperature of the radiator is far less than that of the sun.

TPV cells are generally built from semiconductors with bandgaps ranging from about 0.5 eV to 0.75 eV. Such low bandgaps match the lower-energy photons radiated by TPV generators. These materials are indium arsenide (InAs) alloyed with gallium arsenide (GaAs) to form the ternary (three-element) alloy InGaAs. The bandgap can be tuned by varying the value concentration of the various elements. Alternatively, antimony can be added to the mix to make a quaternary (four-element) alloy to bring potentially greater power output, but with additional cost and complexity. Silicon has also been considered for possible use in TPV cells but is no longer the main avenue of current TPV research.

Typical output

The output for a single-junction TPV cell is typically about 0.5 watts per square centimeter, whereas a typical flat-plate, single-junction silicon solar cell will produce power on the order of 15 milliwatts per square centimeter. Even the highest-efficiency research solar cells produce only about 30 milliwatts per square centimeter, according to Timothy Coutts, a prominent TPV and solar-cell researcher and Research Fellow at the National Renewable Energy Laboratory (NREL) in Golden, Colo., a unit of the Department of Energy.

"People have tried for years to make silicon-based TPV systems, and while the jury is still out on that, tremendous progress is being made with III-V material systems," Coutts said. "For example, the rare-earth oxide ytterbia has an emission band corresponding to silicon's bandgap, but you have to heat it to very high temperatures to get enough useful photons, which creates thermal-engineering issues, and even then not enough radiant energy is pushed out at the frequencies required." A TPV cell that combines an InP substrate with device layers that absorb at about 2 microns in the IR spectrum, say InGaAs, works well when used in conjunction with a broad-band source.

Coutts also said computer models show that multijunction TPV cells, with layers of materials designed to capture energy from multiple wavelengths in the infrared spectrum, theoretically could produce up to 5 watts per square centimeter, though these have not yet been demonstrated.

At the conference, of which Coutts is chairman, many papers will address the two areas of greatest debate among TPV researchers: whether it is better to design cells to work with broadband or selective-emission radiators, and whether the greater potential power output of TPV cells made from ternary and quaternary compounds outweighs the material's challenges and cost.

One highlight on the selective-emission side is a paper from a group at NASA-Lewis (Cleveland) concerning a promising erbium-doped YAG (yttrium-aluminum-garnet) thin-film selective emitter. The paper presents the results of a study of thermal gradient effects across the thickness of the film. It is the first detailed theoretical and computational study of temperature gradients and their effect on performance.

In terms of TPV-cell operating characteristics, a paper from Sarnoff Corporation (Princeton, N.J.) discusses InGaAsSb cells grown on GaSb by molecular-beam epitaxy, one of the first uses of the technique for TPV. The overriding concern the paper addresses is that while a GaSb substrate is commercially available, InGaAs can be lattice-matched to it only at one specific atomic ratio of indium and gallium, which does not happen to be optimum for TPV cells. By adding a fourth element, Sb, the researchers say it appears the lattice match can be maintained and a variety of bandgaps can be produced. They built TPV cells that featured internal quantum efficiencies as high as 95 percent.

One paper that may herald the entry of Japanese researchers into the field is from Ishikawajima-harima Heavy industries (Tokyo), discussing potential TPV applications for civilian and industrial use. Researchers from the company are presenting the results of a study arguing that with regard to system efficiency, TPV can compete with conventional small-scale generators.

The Viking-9 electric vehicle uses eight TPV cells to triple the car's range from 50 to 150 miles