Published on Feb 21, 2020
Three new developments have now occurred making economical TPV systems possible. The first development is the diffused junction GaSb cell that responds out to 1.8 microns producing over 1 W/cm2 electric given an IR emitter temperature of 1200 C. This high power density along with a simple diffused junction cell makes an array cost of $0.5 per Watt possible.
The second development is new IR emitters and filters that put 75% of the radiant energy in the cell convertible band. The third development is a set of commercially available ceramic radiant tube burners that operate at up to 1250 C. Herein, we describe a 1.5 kW TPV generator / furnace incorporating these new features. This TPV generator / furnace is designed to replace the residential furnace for combined heat and power (CHP) for the home.
burning a fuel. If very low cost solar cells can be made, a homeowner can generate his own electricity at rates below the electric utility rate. However, while this can be true during summer months, there is a problem in winter months when the sun doesn’t shine. Low Bandgap photovoltaic or “Solar” cells can solve this problem. A homeowner can put solar cells on his roof for electric power in the summer months and “solar” cells in his furnace for electric power during the winter months. The idea of “solar” cells in a heating furnace is called ThermoPhotoVoltaics or ThermoPV or TPV. The idea is that a ceramic element is placed in the flame in the furnace and this element then glows like the coals in a fireplace. “Solar” cells near by then convert the glow into electricity. Using TPV, the homeowner generates electricity whenever he needs heat. Therefore, it is not necessary to burn additional fuel.
While this TPV concept is simple, the problem has been that the two types of solar cells (or more accurately photovoltaic cells) are not the same. While the solar cells on the roof convert visible light into electricity, the TPV cells in the furnace need to convert infrared radiation into electricity. JX Crystals has invented and developed the required GaSb TPV cells.
After inventing the required IR sensitive cells, JX Crystals began to develop complete TPV systems. This effort then required us to invent and integrate several key components into a practical generator that can be manufactured economically. While the TPV idea was first proposed in 1960, three new developments have taken place in the last 10 years that now make economical TPV systems possible. The first new development is the diffused junction GaSb cell [1,2,3,4] that responds out to 1.8 microns producing over 1 W/cm2 electric given an IR emitter temperature of 1500 K (1225 C). Two TPV circuits incorporating these new cells are shown in figure 1a along with a power curve in figure 1b showing 2 Watts per cell. The power density is approximately 100 times more than the traditional planar
solar cell making cost of $0.5 per Watt possible.
TPV cells and solar cells are natural allies. They both produce electricity without the need for additional fuel. The bar graph shown in figure 6 is for the electricity use in a typical residence in NY. It illustrates how TPV and solar can work together to produce home electric power. This bar graph is typical for Mid Atlantic & New England states as well as states around the Great Lakes, Alaska and Canada. The economics for photovoltaics is now very exciting as the following calculation shows. Given investment for manufacturing and marketing, we expect our Residential TPV Furnaces (figure 3) to be selling for $1500 on top of the standard heating equipment cost. At what price will a Residential TPV Furnace begin to be cost effective? Referring to the bar graph in figure 5, the annual savings in a home-owner’s electric bill will be $376 per year at 10 cents per kWh. At this savings rate, the payback time will be 4 years
The generic TPV system
A generic TPV system consists of a:
(i) source of heat (e.g., a flame, radiative isotope, the sun),
(iii) semiconductor converter,
(iv) means of recirculating the sub-bandgap photons to conserve system energy, and
(v) power conditioning system.
The various options for items (ii), (iii), and (iv), as well as their performance are discussed in this paper. It must be appreciated that each of these processes has an efficiency of less than unity, thus implying that the overall system efficiency must necessarily be relatively modest. In the short-term, a system efficiency of 20% would be a very ambitious goal. Nearly all work undertaken to date has concerned the optimization of individual components, with relatively little work being devoted to the system as a whole. This rather unbalanced approach has been addressed, to some extent, in recent years but more work on complete systems—and especially the demonstration of systems—must be performed to demonstrate the full potential of TPV.
The semiconductor converter
Early work on TPV converters focused mainly on silicon and germanium converters (references). However, the quality of these elemental semiconductors was poor. It is the impressive progress in III-V devices that has led to the high performance of modern TPV devices and the consequent resurgence of interest in this field. Potential reductions in cost may be achieved with development of thin-film converters. Most, but not all, recent device work has concerned the semiconductors GaInAs and GaInAsSb grown by organo-metallic vapor-phase epitaxy (OMVPE) or by molecular-beam epitaxy (MBE). GaInAs is lattice-matched to an InP substrate when the atomic proportions of Ga:In are in the ratio 47:53.
However, this composition corresponds to a bandgap of 0.73 eV, which is too large for optimum efficiency. Hence, much effort has been devoted to fabricating InGaAs devices with bandgaps of as little as 0.5 eV. This material, of course, is lattice-mismatched to the InP substrate, and success in growing it relatively free of damaging defects (caused by strain relaxation) is of major importance to the technology. In addition, considerable effort has been devoted to the characterization of the alloys.
The quaternary alloy GaInAsSb can be grown lattice-matched to either GaSb or InAs substrates, although the former is mainly used. The optimum bandgap can be selected (in the range 0.38-0.7 eV) without compromising lattice-matching, and great progress on GaInAsSb devices has been made. At present, at least up to October 1998, there seems to be little to choose between these two rival alloy systems. Despite this progress in two appropriate multi-component alloys, the only commercially manufactured TPV system is the Midnight Sun, marketed by JX Crystals, Inc. This uses the binary semiconductor GaSb, with a bandgap of about 0.7 eV. Although this is nonideal, GaSb can be grown in large volume and can easily be processed into a device that performs well. The devices are fabricated from wafers of p-GaSb, with the junctions being formed by diffusion. Arrays of GaSb cells are connected in series and parallel combinations, and system outputs of as high as 600 watts have been recorded.
The radiator may be either broad- or narrowband. If the former is used, then the spectral emittance must be as close to unity as possible, and the material most commonly used has been silicon carbide. This has an emittance of about 0.9, and it will withstand temperatures up to about 1900 K. Optimization of the bandgap depends, as expected, on the radiator temperature, and there is inevitably a large proportion of the incident photons that are of too low an energy to be absorbed. For the system to have an acceptable efficiency, these must be re-absorbed to the radiator. This aspect will be discussed in the next section. Narrow-band radiators considered to date include ytterbia, erbia, and holmia. These radiate at energies of about 1.1, 0.7, and 0.5 eV, respectively. The selective radiation property arises from the electronic structure; valence electrons are screened from mutual interaction thereby preventing energy bands from forming. Great progress has been made with selective radiators in recent years, and earlier objections of less-than-ideal performance have largely been overcome. Again, at this stage, there seems to be little to choose between the two approaches.
More Seminar Topics:
Microbial Enhanced Oil Recovery,
Nanobiomaterials for Biosensing,
Protein Adsorption On Metal Oxides,
RF Absorption Involving Biological Macromolecules,
Soil Carbon Sequestration,
Solar Pond Technology,
Coal Bed Methane