95

The intermittent nature of sunlight poses for all earth-based solar energy schemes the problem of stable output with discontinuous input. In the gasoline engine, this problem is solved with the flywheel. In the car's electrical system, it is solved by the storage battery. However, storage of large quantities of electricity is a difficult engineering problem. Where topography permits, some commercial power companies rely on the potential energy in water impounded behind hydroelectric dams, and even find advantage in pumping water back behind such dams with off peak power so that it will be available to meet peak loads. (The Storm King Project was for this purpose.) However, pumpedback storage assumes that base load energy will be continuously available. For intermittent solar energy it has been suggested that chemical storage of electric power might be a feasible solution. Vast quantities of electricity are tied up in some of the more common materials. For example, the quantity of electricity represented by a ton of aluminum is 50,000-60,000 kilowatt-hours. (Unfortunately, no way is yet available for converting aluminum back into electricity.)

5

However, if water is dissociated by electrolysis (passing a direct current through it), quantities of oxygen and hydrogen are liberated and might be collected. These might then be recombined in a fuel cell, to generate electricity again, without prohibitive loss of efficiency. Or the hydrogen might be used elsewhere as fuel to run a gas turbine to generate electricity. Or it could be supplied to the natural gas industry as a substitute for natural gas (although hydrogen provides only about one-fourth the thermal energy of natural gas per unit of volume). Another possibility is that.it could be reacted with carbon or hydrocarbon materials to produce methane. Hydrogen might be used as a substitute for gasoline in motor vehicle engines. It could be combined with calcium or uranium to form the hydride; in this solid, inert form it could be transported conveniently and safely, and when heated moderately, these hydrides would release the hydrogen at a rate that could be readily controlled. The metals could be reused indefinitely.

Conceptually, such a solar-electric system could be operated primarily to produce hydrogen, which could then be used as the universal source of packaged thermal energy to generate electric power, mechanical energy, and heat or cold. The electrolysis process is readily interruptible, without time lag in starting up again. Moreover, the product of combustion of hydrogen is water-presenting no noxious emissions of sulfur or carbon monoxide. Still another advantage that has been claimed for hydrogen as an energy form is that it can perhaps be piped underground more cheaply than electricity can be conveyed by overhead transmission lines. The obvious disadvantage of hydrogen, of course, is its explosive combustibility—a hazard that it

The fact that aluminum metal requires so much electrical energy to produce it has significance here. Pot-lines to reduce alumina to metal take a heavy toll of electricity supply. Stockniled aluminum represents a stockpile of both bauxite and energy. Importation of bauxite commits large amounts of electric energy domestically to converting It to aluminum metal. The importation of metallic aluminum is the same essentially as the importation of both bauxite and electrical energy. It is also striking that aluminum is recycled to a much smaller extent (20%) than is lead (50%) or copper (40%), although to remelt aluminum takes vastly less electrical energy (1300 to 2000 kwh. per ton) than to produce primary aluminum (50,000 to 60.000 per ton)-about the same as to remelt steel (1240 kwh per ton). If one is to examine the "economics" of energy rather than monetary units, one must give attention to anomalies like this. The concept of "storing" interruptible electrical supply from solar energy in the form energy-intensive materials needed by industry seems to be valid regardless of whether they are used as such or are converted back into electrical energy as in the case discussed in the text of the use of hydrogen in fuel cells.

96

shares with gasoline. Whether technology can learn to handle hydrogen as safely as it has learned to handle gasoline remains to be determined.

The attractiveness of the hydrogen system technically, and its versatility and compatibility with both biological and thermal solar radiation systems suggests that the system itself may warrant some research emphasis. It is also to be observed that hydrogen can be produced (along with methane and carbon monoxide) by the high temperature dissociation of steam in the presence of burning coal. (This was, in fact, the traditional method of producing illuminating gas the so-called "water gas"-in municipal gas plants, before the advent of natural gas.) In the event that natural gas reserves become exhausted more quickly than is now foreseen, it might be feasible to resort to use of coal gas as an interim measure, with subsequent phasing in hydrogen or methane generated from a solar complex.

II. THE PROSPECT FOR SOLAR CELL TECHNOLOGY

Two principal concepts have been advanced to use solar cells in a large scale power system (i.e., in the hundreds or thousands of megawatts). One is a land-based array of solar cells, with solar energy somewhat concentrated by a system of flat mirrors. Success in this approach depends on a radical reduction in the cost of the semiconductor material used, and some resolution of the problem of intermittent solar radiation.

An alternative concept would place the solar cell array in a synchronous orbiting satellite, continuously exposed to solar radiation; the electricity produced from the cell array would then be converted to microwave energy, beamed to earth, collected and converted into electricity for transmission and use. Success in this approach, according to the Cornell Workshop analysis, would require "tenfold reductions in (1) weight/kilowatt of the space array, (2) cost/unit weight to orbit, and (3) component cost **** Questions were also raised concerning the environmental hazard of a high microwave radiation flux density (100 watts per square meter)."

Cost Considerations of Solar Cell Materials

Cost of the semiconductor materials to be used in the large-scale generation of power really comes down to a question of the possibility of improving the cost/effectiveness of high purity silicon. Some research in other semiconductor materials might be warranted, and even the attractive if remote possibility of organic (plastic) semiconductors has not been completely abandoned. But in the opinion of an ad hoc panel of the National Academy of Sciences on solar cell efficiency: “In considering the various alternative devices, it is clear that the silicon solar cell offers the best opportunity for increased efficiency at lowest cost and shortest payoff time."

་་ ་

In its report, the ad hoc committee offered the judgment that the efficiency of the silicon cell might be increased from its present 11 percent to 20-22 percent through improvement in the quality of the silicon. With respect to "Cost of Solar Cells the committee in an appendix to its report stated as follows:

Op. cit., p. 118.

National Research Council. Ad Hoc Panel On Solar Cell Eciency. Solar Cells: Outlook for Improved Efficiency. Washington, National Academy of Sciences, 1972, p. 18.

Appendix: Cost of Solar Cells

97

"Contrary to the experience with many other semiconductor devices, the cost of solar cells (about $100,000 per kW generating capability) has remained almost unchanged during the past decade. There are a number of factors responsible for this. First, there has been no significant increase in the market for solar cells, whose use is confined almost exclusively to power supplies on satellites. The small market has discouraged introduction of mass production techniques. Second, this particular application sets very stringent requirements on cell reliability. This in turn requires intensive, expensive testing of every cell. Third, limits of area on space vehicles encourage cutting cells to rectangular shapes. This results in additional costs because of the loss of otherwise usable silicon crystal material. The as-grown crystals are circular in shape, and if cells of circular shape were sufficient, savings could be effected. However, power per unit area or weight would be increased.

"According to E. L. Ralph, simple change in shape, etc., would make it possible to produce 11 percent efficient cells at a cost of $15,000 per kW generating capability, a factor of about 6 decrease from percent levels. Ralph assumes that the cost of the silicon single crystals will remain unchanged. This cost factor provides a serious ultimate limitation on the cost of the cell. Yet, the ultimate cost is set to a considerable extent by the size of the market. Current U.S. production of single-crystal silicon amounts to about 100 tons per year. The amount of silicon that would be needed to generate enough power to match current U.S. terrestrial electric power needs would be about 2 million tons. (These cells would cover a total area of about 5 × 103 sq miles. This area is comparable with the area of the United States currently covered by structures and roads.) Silicon is an extremely abundant element so that the cost is not based on limited supplies of the raw material. It is reasonable to expect that if the demand for silicon were to increase from the current level to the level required by a large-scale terrestrial market, the price should drop substantially. This has certainly been the case in the history of other materials like aluminum and, more recently, titanium.

1

"Ralph points out the use of a simple conical egg carton' concentrator has increased the power produced per cell by a factor of 2.5. The cost of power generated with such a system including the $15,000 per kW solar cell, would further reduce costs to about $6000 per kW. Higher sunlight concentration would reduce costs still further. Ralph estimates that the price could ultimately be reduced to the range of $1000 per kW. He cites the fact that the Odeillo, France, 1000-kW solar furnace was built at a cost of $2000 per kW. A simpler design would be sufficient for large-scale electric power generation. "To date, there has been no attempt to explore the costs of largescale solar-energy conversion systems; therefore the costs cited here are conjectural."

1E. L Ralph. "A Plan to Utilize Solar Energy as an Electric-Power Source," Proc. of the Eighth Photovoltaic Specialists Conference, Seattle, Washington, Aug. 1970, IEEE, Catalogue No. 70C 32 ED.

F. Tromke. "Le Four Solair de 1000 kw d'Odeillo-Font-Romen," Rev. Hautes Temp. Refract. 1,5-14 (1964).

98

Magnitude of the Requirements; Abundance of Silicon

With respect to the sheer quantitative availability of silicon, there is no problem: it is the most abundant material in the earth's crust. On the basis of the principle that materials should be used quantitatively in some correspondence to their relative availabilities, silicon is a decidedly under-used material. However, the problem of its use in solar cells is that of processing. On this subject, the ad hoc committee observed that improvement in technology had leveled out over the last five years. Only two of the five companies originally engaged in producing high purity silicon for commercial solar cells remained in the business. Production, largely for the space program, was trending toward the more expensive batch process. Efforts to increase the efficiency of the silicon cells had "not been encouraged or supported" and "present-day solar-cell technology does not take advantage of the most sophisticated silicon technology in materials, contacts, or surface treatments, nor do production methods make use of automation, which would probably result in significant cost savings."

Nevertheless, as a practical matter, it is hard to envision meeting a substantial fraction of U.S. energy needs with solar cells. To meet total U.S. energy needs by this system would entail deployment of cells and reflectors in an earth-based system to intercept some 35,000 square miles surely an ambitious undertaking.10

However, use of solar cells as one of a variety of solar energy systems, or perhaps as a part of a hybrid system, might prove of ultimate importance.

Silicon Cell R&D Projects Supported by NSF

Three research projects sponsored by the National Science Foundation under its program of "Research Applied to National Needs" (RANN) are aimed at materials improvement in the field of cell semiconductors. These are described as follows:

Work statement summaries for these three projects, supplied by NSF, are as follows:

(1)

This project is concerned with exploratory research to examine advanced methods for fabricating thin films of polycrystalline silicon photovoltaic cells. The longer term goal is to develop automated continuous processes for producing these cells and to reduce cell costs per watt substantially below those of present production methods. The advanced methods of fabrication to be studied will include (1) vacuum

• Ibid., p. 10.

10 "Chance for Solar Energy Conversion." Chemical and Engineering News, Dec. 20. 1971, p. 39. The article makes reference also to a "new proposal" for an "electromagnetic ware energy converter" for which an estimated conversion emciency of 50 percent was claimed to be possibly attainable.

99

evaporation of silicon onto flexible substrates at high temperatures, to try to obtain satisfactory thin (perhaps 10 microns) polycrystalline films suitable for large-area photovoltaic cells; (2) sputtering of silicon to form thin polycrystalline films; (3) electron-beam ion plating of silicon to form suitable films; and, (4) chemical vapor deposition of silicon containing gases onto heated substrates to obtain suitable polycrystalline films. Present production inethods for silicon solar cells require high-purity single crystal silicon material that is sawed into slices about 0.01 inch thick; and then the slices are polished, diffused, activated, and mounted with electrical contacts to provide reliable operating cells at a cost of several hundred dollars, or more, per watt of electrical power capability. Cells produced by present techniques have played an important role in many solar space power systems; but, present production methods do not appear to be capable of producing the required volume of solar cells or of reducing cell costs by a factor of more than 100 to make terrestrial solar power electrical systems economically practical. The immediate objectives of this research are to learn to grow suitable polycrystalline films and to learn to form p-n junctions on them to obtain reliable, efficient, large-area photovoltaic cell structures at very low cost per watt.

(2)

This research project is directed to further understanding and development of CdS/Cu,S solar cells to obtain longerlife, higher-performance, more economical cells for applications in large and small-scale solar energy conversion systems. The principal objectives are (1) improved understanding of the basic properties and conversion mechanisms of CdS/Cu,S cells, (2) improved cell lifetimes and methods for accelerated lifetests (goal in excess of 20 year lifetimes), (3) improved performance and conversion efficiency at elevated temperatures, (4) improved production techniques to increase reliability of cells and decrease production costs, (5) economic analyses of production costs of CdS solar cells, (6) limited production of cells and testing of a demonstration solar cell unit, (7) systems analysis to identify key parameters and trade-offs in applications; and, (8) improved understanding of possible public hazards arising from accidental local distribution of cadmium from damaged cells. The project will include an evaluation of the feasibility of using systems of CdS cells for local solar energy harvesting to produce heat and electrical power for individual commercial and residential buildings in the United States and to produce larger blocks of power for local industries requiring direct current power.

The objective of this project is to develop a more efficient and cheaper photovoltaic device using Schottky Barrier Diode (SBD) principles. This project includes calculations to determine the optical properties and to select thicknesses of various metal coatings on semiconductor substrates for pro