Printer Friendly

Solar photovoltaics: out of the lab and onto the production line.

Solar Photovoltaics: Out of the Lab and onto the Production Line

Laboratory solar cell conversion efficiencies have risen markedly in recent years. During the same time, the lag in translating these laboratory successes to the market has also grown. Researchers in both industry and government have placed new emphasis on developing innovative methods to manufacture cost-effective photovoltaics.

Dreams of clean abundant low-cost solar electric power have persisted for good reasons: no fuel to burn, few parts to fix, minimal supervision required. While solar photovoltaic (PV) power has long been one of the brightest prospects on the energy-supply horizon, the prospect of large-scale PV utility power generation and the widespread use of PVs in homes and offices has seemed to dim as time passed.

Though it still provides only a tiny fraction of our nation's huge energy needs, the photovoltaic industry has in fact made substantial progress in recent years. Some 46.6 megawatts of photovoltaic modules were shipped in 1990, continuing the 25 to 30 percent annual market growth rate of recent years. Worldwide PV industry revenues for 1991 could hit $800 million, according to industry monitors. Though the PV industry has seen business failures, major consolidations, and new strategic partnerships in recent years, the number of organizations engaged in PV research and development or manufacturing has grown to more than 200 worldwide.

Today, solar power is at work primarily in remote locations far from utility power grids as midsize stand-alone systems (a few watts to a few thousand watts) and in consumer products such as outdoor lighting systems, personal electronics, battery chargers, and automotive sunroofs for cooling and heating (a few milliwatts to a few watts).

Improvements in processing and product, along with a greater effort in marketing, have reduced PV costs to about $6 per peak watt (midday power output). That price translates into a cost of electricity production of $0.25 to $0.50 per kilowatt-hour. Though still not competitive with today's utility-grid power costs, that price is a great improvement. U.S. Department of Energy (DOE) goals for the 1995-2000 time period is about $0.12 to $0.20 per kilowatt-hour.

If the continued high cost of solar electric power has kept utilities out of PV technology, the dearth of demand for giant utility PV systems has kept prices high by not allowing the economies of scale derived from large-volume production, industry observers said. To combat these circumstances, the DOE has recently announced a new strategy designed to accelerate the commercial use of solar electric technology, especially by utilities. The ambitious Solar 2000 plan calls for the deployment of 1400 megawatts of American-made PV systems worldwide (with 900 megawatts in the United States) by the year 2000, 10 times the current base. These goals will require a substantial investment.

Funding Shift

It is estimated that the U.S. government has placed nearly $1 billion into PV research since the early 1970s, while private industry has invested more than twice that amount. "In the 1988 to 1990 period, the PV budget bottomed out at about $35 million per year," said Tom Surek, manager of PV programs at the National Renewable Energy Laboratory (NREL, formerly the Solar Energy Research Institute) in Golden, Colo. "As increased emphasis has been placed on PVs, especially on their commercialization, the budgets have risen to $46 million in 1991 and $60 million in 1992."

Years of DOE funding support for basic PV research in government, industrial, and university laboratories have surely borne fruit, yielding a range of solar cell technologies that regularly set new records for energy conversion efficiency as engineers apply insights derived from improved understanding of PV materials and cell design.

"The record-setting cells and other recent successes came about mainly because of federal funding directed to basic PV research through the 1980s," said Richard J. King, program manager at the DOE PV Technology division (Washington, D.C.). "At the same time, however, the lag in performance between the lab devices and the products on the market was widening. When people realized this was happening, they started looking at the manufacturing aspects of the business."

Manufacturing Technology

The result has been a refocusing of research support on advancing manufacturing technology to drive the commercialization of cost-effective PVs. A prime example is the Photovoltaic Manufacturing Technology (PVMaT) project, a five-year DOE initiative aimed at improving the understanding of PV production technology, said NREL-based project leader, Ed Witt. PVMaT is a government/industry cost-shared partnership that is designed to advance PV manufacturing technology, reduce PV module manufacturing costs, and expand production capacity, according to Witt. He said the DOE will provide about $55 million over five years to study both generic and company-specific manufacturing difficulties. The PV industry will provide matching funds for each PVMaT subcontract.

In Phase 1 of the PVMaT project firms were asked to identify their current capabilities in manufacturing and process development, specific manufacturing routes that may lead to increased production and reduced manufacturing costs, and problems that are impeding progress. Twenty-two PV manufacturers were awarded subcontracts of up to $50,000 each (see page 55). Phase 2 will consist of multiple competitive procurements during the program period to solve the process-specific problems PV makers identified in Phase 1. Later, companies may be teamed to undertake generic research that is of interest to several companies.

Another notable effort is a small Congressionally mandated project aimed at studying the integration of PVs into the roofs, walls of office buildings, and homes, said Surek. "This could be a high-value application. Integral PV systems could offset up to 70 percent of a building's electrical load."

Some PV researchers said that subcontracts for basic research have been squeezed by the new emphasis on manufacturing. Though total PV funding has risen, the same projects that so successfully boosted laboratory conversion efficiencies will soon be strapped for cash, they said.

Industry is not sitting around waiting for federal money, however. Besides demonstrating performance improvements, the industry has begun operating new automated factory lines for producing current-technology PV devices and modules in larger volumes as researchers seek to improve manufacturing processes.

In the field, solar power arrays are proving out their long-term application. One of the test programs, Photovoltaics for Utility Scale Applications (PVUSA), based in Davis, Calif., is a national cooperative research and development project that is assessing and demonstrating the viability of utility-type PV electric-generating systems. PVUSA, an outgrowth of a Pacific Gas & Electric Co. venture, provides a test bed for installation, operation, and evaluation of emerging PV module technologies and balance-of-system configurations (all the equipment in the PV system other than the solar modules) in a utility setting. The project is fielding a dozen different PV technologies from a range of PV companies.

PV Strategies

In general, the problem with high-performance solar cells such as the single-crystal gallium arsenide PVs used on spacecraft is that they cost too much, experts said. One school of thought argues that if cells are so expensive, then the incoming sunlight should be gathered and focused with cheap concentrator lenses to increase the output of each cell. But although laboratory concentrator cells have exceeded efficiencies of 30 percent, the balance-of-system costs for these systems are higher. In addition, experts said that concentrators make sense only in regions with high solar irradiance.

The principal way in which researchers have tried to make PVs cheaper is to increase their energy conversion efficiencies. If one doubles the efficiency, the system cost can be more than halved. For the same energy output, high-efficiency systems require fewer modules, and therefore fewer support structures, as well as less wiring, land, and installation labor.

Crystalline and Semicrystalline Silicon

High-performance high-priced single-crystal silicon has been the workhorse for PV power applications since the industry's inception. It currently accounts for 35 percent of the total market. These durable and reliable devices are made from many silicon wafers sliced individually from cylindrical crystalline silicon ingots by saws. The wafers are then soldered together and encapsulated into a module. The cost of the wafers and the labor-intensive production process make crystalline silicon PVs relatively expensive.

One of today's leading single-crystal silicon devices, a 17.5 percent efficiency cell, is now being produced by England's BP Solar Ltd. (Andover, Mass.) at a new plant in Spain. The production line makes use of the new laser-grooved buried-grid process developed by Martin Green and his colleagues at the University of New South Wales (Sydney, Australia). Now licensed to BP Solar, this technique minimizes the blockage of incident light normally caused by front-mounted electrodes by burying them inside the cell. Grooves cut in the silicon are later filled with conducting material to form electrodes.

A radically different approach to single-crystal silicon PVs is being pursued by Texas Instruments Inc. (Dallas) and Southern California Edison Co. (Rosemead, Calif.), which in mid-1991 announced the joint development of spheral solar technology--an innovative method to produce solar cells from thousands of silicon microspheres embedded in aluminum foil (see page 53).

In 1990, about one-third of PVs shipped were made from semicrystalline silicon. Semicrystalline silicon devices, in which the active portion of the device is composed of several relatively large crystal grains (about a square centimeter in area), are generally not as efficient as single-crystal silicon cells, but they cost less. The most popular commercial method to produce these PV cells involves a process in which molten silicon is directly cast in a square mold and then allowed to solidify into an ingot that is subsequently sliced into wafers. Demand for this type of PV module is growing. For example, Kyocera Corp. (Sakura, Japan), the biggest manufacturer of PV power modules in Japan, reportedly plans to double its production capacity of semicrystalline silicon devices from 6 megawatts per year to 12 megawatts by the end of 1992. By 1995, Kyocera plans to boost capacity to 30 megawatts.

Thin-Film Amorphous Silicon

Another route to cost-effective solar cells is to make inherently inexpensive PV materials that work relatively well. This is the approach of thin-film amorphous (noncrystalline) silicon PVs. Since its introduction in the late 1970s, shipments of amorphous silicon PVs have grown rapidly to 15 megawatts in 1990, more than 30 percent of the market. It is used predominantly for consumer electronics and small power applications.

"Amorphous silicon has been successful because you can lay down thin coatings onto inexpensive substrates relatively easily and cheaply," said David Carlson, vice president and general manager of the Solarex Thin-Film division (Newtown, Pa.). "Unfortunately, low-efficiency (5 percent for modules) holds it back." Improvements in material quality, module design, and large-volume production facilities are necessary to reduce overhead costs, he said.

"Amorphous silicon's big problem is that its efficiency degrades by from 10 to 50 percent when it's exposed to light," said Subhendu Guha, vice president for research and technology at United Solar Systems Corp. (USSC) in Troy, Mich. This phenonmenon is known as the Staebler-Wronski effect, named for those who reported it in 1977. Amorphous silicon has no long-range order. The absence of a crystal lattice means that many atoms do not bond strongly with their nearest neighbors as they would in a crystal. According to the conventional wisdom, those weak bonds break when they are exposed to sunlight, forming defects called dangling bonds that can serve as sites where electron-hole pairs recombine. "Light creates new defects that gobble up electron-hole pairs," Guha explained. The charge carriers never make it to the electrodes, which cuts electrical output. Annealing of the silicon by the sun's heat removes new dangling bonds and the opposing processes eventually reach an equilibrium state, but at a lower-than-optimal efficiency.

The Staebler-Wronski effect can be minimized by suitably adding hydrogen to "tie up" or passivate the dangling bonds and by making the active films so thin that the electron-hole pairs are never far from the electrodes, so that recombination is harder. Nevertheless, the phenomenon remains a prime research topic.

A chemical vapor deposition apparatus is generally used to make amorphous silicon cells. Using large vacuum chambers filled with precisely controlled mixtures of silicon, germanium, and hydrogen gases, as well as small quantities of various doping agents such as phosphorus or boron, technicians deposit micron-thick films on glass or stainless-steel substrates. Radio-frequency electrodes inside the tanks generate electronic fields that ionize the gases. The ionized atoms in the resulting hot plasma plate out onto the substrate surface, which is heated to several hundred degrees C. Layers with specific opto-electronic properties are produced by adding different dopants to the gas mixtures during process deposition cycles. Lasers scribe narrow gaps that are filled in by conductive materials to connect cells into high-voltage large-area modules. Other deposition stations along the production line lay down antireflective coatings and films that act as electrodes and encapsulants.

Makers of amorphous silicon PVs are cutting labor costs and increasing fabrication rates by installing large production lines with automated equipment such as robot materials handlers, conveyor belts, scribing lasers, and computerized precision process controls. Solarex Thin-Film division, for example, has constructed a totally automated pilot line that produces 12-inch by 13-inch 5-watt modules with a stabilized efficiency of 5 percent, Carlson said. This line is capable of producing about 1 megawatt of PVs a year with a 90 percent yield.

Another producer of amorphous silicon PVs is Advanced Photovoltaic Systems (Fairfield, Calif.), the successor to Chronar Corp. APS has begun construction of a new 10-megawatt-per-year capacity manufacturing facility. The company makes 2.5-foot by 5-foot amorphous silicon panels rated at 50 watts. The company reportedly plans to install a 400-kilowatt array consisting of 9600 modules at the PVUSA site in Davis, Calif.

A leader in the amorphous silicon industry is United Solar Systems, which is a recently formed joint venture between Energy Conversion Devices (Troy, Mich.) and Canon Inc. (Lake Success, N.Y.). "ECD needed a partner to take its technology to large-scale production," Guha said.

"Inside glow-discharge chambers, we deposit six layers of amorphous silicon (two identical n-i-p cells) onto 2000-foot rolls of 14-inch-wide stainless-steel sheet in an automated process," Guha explained. "We stopped production for a year starting in July 1990 to make the line more efficient and cost-effective by installing more automation and better reliability controls. The 2-megawatt-capacity line fabricates 4-square-foot 6.2 percent efficiency (stabilized) panels composed of same band-gap tandem cells connected in series: six layers (n-i-p) of amorphous silicon along with a textured aluminum silicon dielectric layer and an aluminum/zinc oxide back reflector.

The main technique used to push higher the efficiencies of amorphous silicon is to stack different band-gap-energy cells on top of each other like a layer cake. Each cell absorbs a different portion of the solar spectrum, yielding increased utilization of the incoming rays. In a stacked amorphous silicon cell, the silicon absorbs the visible light while the silicon germanium collects the infrared light, increasing total light collection. "The beauty of this approach is that all you have to do to make these stacked cells is to add a few more deposition chambers to your line," he said.

Researchers at Solarex recently obtained 10 percent efficiency in a 12-inch by 13-inch module with triple-junction devices using a top cell of an amorphous silicon carbon alloy (1.9-electron-volt (eV) band gap), an amorphous silicon cell in the middle, and amorphous silicon-germanium alloy (1.4 to 1.45 eV) for the bottom junction. The front contact was made of textured tin oxide to trap incident light inside, and the reflective back contact layer is composed of an indium tin oxide over silver layer.

Meanwhile, USSC has demonstrated a triple-stacked amorphous silicon laboratory cell that has a 13.7 percent efficiency. The top device is amorphous silicon (1.8-eV band gap), the middle junction is amorphous silicon with a little lower band gap (1.7 eV), and the bottom is an amorphous silicon-germanium alloy (1.4-1.5 eV). The low-band-gap amorphous silicon-germanium alloy allows the cell to capture the longer-wavelength photons (red and infrared) not absorbed in the top layers.

USSC is planning to construct a large-capacity facility for the production of the triple cell that is to be running by 1994. At the same time, ECD engineers are developing a similar line that will manufacture 3 megawatts of triple cells a year for an ECD joint venture in Moscow, Russia, called Sovlux. The Russian plant is to be operational in early 1992. The production triple-cell modules, he said, are expected to have from 8 to 9 percent (stabilized) efficiency.

Thin-Film Polycrystalline PVs

Astropower Inc. (Newark, Del.) is taking a substantially different approach toward making efficient cost-effective silicon-based PVs. Its silicon-film device is a thin-film polycrystalline silicon solar cell that is produced by depositing a photoactive layer onto a low-cost ceramic substrate. The 100-square-centimeter cells are 10.9 percent efficient. A smaller-area version with a 15.7 percent efficiency shows that improvement is possible, said Allen Barnett, president of Astropower. He also noted that Astropower has produced a silicon-film PV panel composed of 36 cells (0.45 by 0.96 m) that produces 34.2 watts.

Commercial feasibility has been demonstrated by Astropower's pilot line housed in a new 40,000-square-foot building. The silicon-film line is producing tens of square meters of cells per day, Barnett said, some to fill a recent order for 20 kilowatts of PVs from Virginia Power Co. The line has the capacity to produce about 0.5 megawatt of PVs annually, he said. By the end of next year, its output should reach 3 megawatts.

The idea behind Astropower's silicon-film approach is to reduce manufacturing costs by making thin-film polycrystalline silicon cells in a continuous fashion. In the process, 100-micron-thick layers of polycrystalline silicon are grown on a special low-cost thermal-expansion-matched conductive ceramic substrate 10 centimeters on a side.

Astropower's proprietary process involves first depositing a reflective metal barrier on the ceramic panels and then a thin solution of silicon and tin. The substrate provides both mechanical support and back surface conduction while matching the thermal expansion properties of silicon. A conveyor belt brings the package through a furnace that causes the silicon to precipitate out onto the ceramic surface and to grow into planar polycrystalline silicon grains. The finished polycrystalline cells are then screen-printed with electrodes and soldered together into modules like crystalline silicon cells.

Barnett calls the first cell his company produced Product I. Future versions will be improved by the adoption of device-modeling optimization techniques. Product II, a 14 percent efficiency cell that is under development, will feature a metallurgical barrier/reflecting layer between a thinner (30-micron) silicon layer and the ceramic to minimize charge-carrier recombination as well as light-trapping textured surfaces to increase total internal reflection of light to enhance the chances of absorption. Barnett believes that Product III will reach 17 or 18 percent by employing a monolithically integrated high-voltage structure in which embedded interconnects make module-making less costly, and an insulating substrate, which electrically isolates segments of the active layer.

"In general, the process is a major materials-handling problem," Barnett noted. "The manufacturing difficulties we faced were bigger than we anticipated." That's one reason Astropower recently entered into a technology-assistance agreement with Dow Chemical Co. (Midland, Mich). "Dow engineers have materials and process expertise that we can use, and we can help introduce Dow to the PV industry," he said.

Other nonsilicon polycrystalline thin-film PV technologies have also made substantial progress. From cell efficiencies of about 8 to 9 percent in the early 1980s, these thin-film PVs have attained efficiencies of about 14 percent, while demonstrating excellent long-term stability. These technologies are particularly attractive because they hold potential for high-volume production of large-area PV modules. In fact, some experts see in the continued impressive progress of nonsilicon thin-film PV materials the start of a shift in which these products would replace amorphous silicon in certain applications and compete with crystalline silicon. Ken Zweibel, thin-film PV program leader at NREL, predicted that commercial PV products using these thin-film semiconductor materials will reach 10 percent efficiency by 1995.


One of these high achievers is a polycrystalline material called copper-indium-diselenide (CIS). Major participants in this technology are Siemens Solar Industries (Camarillo, Calif.), which purchased the former Arco Solar a few years ago, International Solar Electric Technology (Inglewood, Calif.), Boeing Aerospace & Electronics (Seattle), and Martin Marietta Corp. (Denver).

Siemens has demonstrated a CIS cell with a conversion efficiency of 14.1 percent, and a 4-square-foot 37.7-watt CIS module with a 9.7 percent efficiency as well. These devices have a micron-thick layer of CIS coated with a 0.03-micron layer of cadmium sulfide. Theoretical efficiencies for copper-indium-diselenide PVs have been estimated at as high as 23.5 percent.

PV experts said that CIS offers many intrinsic benefits. With a wide band-gap window, it absorbs much of the solar spectrum. It also has high optical-absorption properties, which means a thin layer can absorb a lot of light, thus lowering the amount of material needed. And, unlike amorphous silicon, it doesn't degrade in sunlight. In addition, CIS is more forgiving of impurities and compositional and thickness variations than other thin films, said Charles Gay, president of Siemens Solar Industries.

Various deposition processes are used in producing CIS PVs, including coevaporation, electrodeposition/selenization, spraying, and screen printing. Siemens holds patents on a procedure in which the copper and indium are sputtered onto a surface, and the deposited material is then selenized. Gay said that CIS cell manufacture generally involves three deposition steps and three laser patterning steps. After deposition of a layer, lasers cut lines in the material and then subsequent depositions fill in the groove to create the interconnect pattern. CIS modules, he said, are similar to others, with a 1-foot by 4-foot sheet of tempered glass on top of the active layers. The package is encapsulated in ethylene vinyl acetate.

Prototype CIS modules are now undergoing environmental testing, Gay said. Siemens has a CIS pilot line running, which is to produce CIS modules for demonstration projects including PVUSA and others in Japan and Germany.

"We're focused on large-area reproducibility of the processing," Gay said. "For the lab cells, we worked with high-purity CIS. We're now evaluating feasibility of using low-cost commercial-grade feed stock and mapping the sensitivities of processing changes on cell performance. The real need is to understand the things that control performance on the manufacturing line, such as yield and uniformity. Over the years, we've found that processes do not necessarily scale up as you'd wish."

Another participant in the CIS area is International Solar Electric Technologies (ISET). The company currently makes 12 percent cells in the laboratory. In ISET's low-cost two-stage production process, copper and indium are deposited by electron beam onto molybdenum-coated glass substrates. The copper-indium layer is then exposed to a selenium-bearing gas that reacts to form high-quality CIS. ISET engineers are said to have overcome adhesion problems with the cells' CIS/molybdenum interface by inserting a thin layer of tellurium between the two materials. The tellurium layer is said to improve the cell's overall performance. Company president, Vijay K. Kapur, believes his company can eventually make 12 to 15 percent efficient CIS modules, which means it could produce power at about $1 per watt.

Cadmium Telluride

Another promising thin-film polycrystalline PV material, cadmium telluride, is receiving substantial attention from companies such as Photon Energy Inc. (El Paso, Tex.), Solar Cells Inc. (Toledo), Japan's Matsushita Battery, and England's BP Solar. In fact, a prototype cadmium telluride cell recently demonstrated the highest efficiency ever seen in a thin-film cell, 14.6 percent. Developed by T.L. Chu at the University of South Florida (Tampa), the record-setting cell bodes well for the future of this technology, according to NREL's Zweibel. "It provides a good deal of confidence that the future of cadmium telluride is bright." Theoretical efficiencies for cadmium telluride cells are calculated to be as high as 27.5 percent.

Zweibel noted that the material's light-absorbing characteristics are well matched to the solar spectrum (1.45-eV band gap) and can be manufactured inexpensively through electrodeposition and spraying. There are problems, however. It is difficult to make electrical connections with the material, and cells are sensitive to water vapor.

Chu's 14.6 percent cell uses several innovative, but simple, deposition processes. He first coats glass with a transparent conductive tin-oxide layer and follows that by depositing cadmium sulfide with a solution-growth (precipitation) process. The active cadmium telluride layer is subsequently produced by a closed-space sublimation (evaporation) process. Zweibel suspects that the cell's high performance arises from the good electrical junction produced by the solution-grown cadmium sulfide layer.

A year-old start-up company, Solar Cells, employs a similar closed-space sublimation process to build its cadmium telluride PVs. In this procedure, cadmium telluride powder is heated in a container to about 700 [degrees] C, whereupon it sublimates (evaporates) onto a slightly cooler substrate surface. This procedure provides for very high rate deposition, Zweibel noted, adding that Solar Cells is designing equipment that will make the cadmium telluride layer for a demonstration 4-foot by 6-foot panel in under a minute.

Photon Energy has built a pilot production line with a capacity of about 1 megawatt. It uses an inexpensive spray deposition process to fabricate 12.7 percent efficient cells and 8.1 percent efficient square-foot modules rated at 6.8 watts. The company has also produced a 4-square-foot prototype module with a conversion efficiency of 6.5 percent. Photon Energy is reportedly negotiating with a Fortune-500 company about a corporate link to provide capital for enhanced production.

Matsushita Battery has been making screen-printed cadmium telluride cells since the late 1970s. These small cells power handheld calculators. Screen-printing is a low-cost process, but the company's module efficiencies are limited to under 10 percent by a thick cadmium sulfide layer.

BP Solar also has a research effort in cadmium telluride. It has demonstrated a 14 percent small-area cell produced with layers of electrodeposited cadmium telluride and solution-grown cadmium sulfide.

Texas Instruments' Spheral Solar Technology

An innovative PV approach called spheral solar technology is being developed jointly by Texas Instruments Inc. (Dallas) and Southern California Edison Co. (Rosemead, Calif.). Spheral solar technology is a patented process that produces many tiny single-crystal silicon devices from low-cost lower-purity silicon feedstock. Seventeen thousand of these silicon microspheres--each acting as an independent photovoltaic (PV) cell--are embedded in a thin aluminum foil substrate to produce a 4-inch-square half-volt cell, said Jules D. Levine, associate project manager for Texas Instruments. The best spheral solar cells have exhibited a conversion efficiency of 10.2 percent, though Levine believes 15 percent is possible with process improvements. The technology is the result of six years of joint research and a combined investment of about $10 million, he explained. The idea was to use abundant lower-cost 99 percent pure metallurgical-grade silicon ($0.50 to $1 per pound) instead of the 99.5 percent pure semiconductor-grade silicon that is currently used (up to $5 to $7 per pound).

The two firms plan to develop a pilot line to produce prototype spheral solar modules and hope to perfect the manufacturing process by the end of 1992, Levine said. Initial applications are to be demonstrations for utilities and remote power installations.

According to Levine, the new process begins when ground particles of metallurgical silicon are melted in a furnace. Surface tension causes them to form into tiny globules of molten silicon covered with an impurity-rich slag that has sweated to the surface. As the silicon solidifies, it forms millimeter-sized single-crystal spheres. The impurities are then ground off the surface in an abrasive racetrack apparatus. This melt/grind procedure is repeated three to five times until the microspheres have attained the desired purity levels, Levine explained.

Remaining impurities in the silicon are treated to tie them up, or passivate them, so they do not cause internal charge carriers to recombine. Residual boron in the original silicon melt dopes the sphere material so that it functions as a p-type semiconductor. The outer surface of each microsphere is formed into an n-type semiconductor layer by diffusing a phosphorous dopant into it.

Smooth aluminum foil is embossed in a hexagonal pattern and etched with potassium hydroxide to produce small apertures (0.53-mm) spaced 0.81 millimeter apart. The spheres are self-loaded into the holes by a vacuum chuck, and then thermomechanically bonded to the foil in a press. A backside etching step of the n-type outer layer then exposes the p-type core of each sphere in the rear.

A thin layer of temperature-resistant polyimide plastic is then applied to the back side and cured. Selective abrasion of the rear sphere tip opens a pathway through the film. The front side is then etched to remove much of the n-type layer, which helps to increase output current. The back bond is applied by thermomechanically attaching a second aluminum foil to the rear. Finally, an antireflection coating of tin oxide and a plastic encapsulant layer are added. Output yield of the fabrication process is about 90 percent, he noted.

The aluminum foil, said Levine, serves four functions: as a flexible matrix to hold the spheres in place, an electrical bus to collect the generated current, a mask for the backside etching step, and a back reflector of light.

Selected DOE Photovoltaic Manufacturing Technology Phase 1 Subcontractors

Flat-Plate Crystalline Silicon Modules

Crystal Systems Inc. (Salem, Mass.): optimize the fixed abrasive slicing technique for a reduction in crystalline silicon ingot wafer slicing cost. Mobil Solar Energy Corp. (Billerica, Mass.): reduce material usage and improve edge-defined film-fed growth process for production of polycrystalline silicon PVs. Siemens Solar Industries (Camarillo, Calif.): improve single-crystall silicon growth process, wafer sawing, cell processing, and module fabrication in manufacture of crystalline silicon modules. Solarex Corp. (Rockville, Md.): improve control of crystalline silicon growth, better process yields, increase automation use, and reduce module material costs. Westinghouse Electric Corp. (Pittsburgh): reduce material and process costs for the dendritic web process for polycrystalline silicon PVs.

Flat-Plate Thin-Film Silicon Modules

AstroPower Inc. (Newark, Del.): lower cost and improve thin-film polycrystalline silicon on ceramic substrate PV production process. Chronar Corp. (Lawrenceville, N.J.): automate production of amorphous silicon PVs. Glasstech Solar Inc. (Golden, Colo.): improve proprietary glass-in/panel-out consent for in-line processing of amorphous silicon PVs. Energy Conversion Devices (Troy, Mich.): incorporate narrow-band-gap material into amorphous silicon cells produced in continuous process and add proprietary microwave chemical vapor deposition process and automation onto line. Iowa Thin Films Technology Inc. (Ames, Iowa): improve and lower cost of automated processing of monolithic amorphous silicon modules deposited on continuous polymer substrate. Silicon Energy Corp. (Chatsworth, Calif.): optimize and automate manufacturing of multiple-junctin amorphous silicon PVs.

Flat-Plate Polycrystalline Thin-Film Modules

Boeing Aerospace & Electronics (Seattle): develop and scale-up copper-indium-gallium-selenium module production process. Photon Energy Inc. (El Paso, Tex.): improve high-throughput spray-processing of cadmium telluride/cadmium sulfide cells with increased automation and optimized design. Siemens Solar Industries (Camarillo, Calif.): better materials usage and increase automation of copper-indium-diselenide PV line. Solar Cells Inc. (Toledo): improve close-spaced sublimation process for high-throughput production of cadmium telluride PVs.

Concentrator Modules

Alpha Solarco Inc. (Cincinnati); Entech Inc. (Dallas); Kopin Corp. (Taunton, Mass.); Solar Engineering Applications (San Jose, Calif.); Solar Kinetics Inc. (Dallas); and Spectrolab Inc. (Sylmar, Calif.).

Photovoltaic Production Equipment

Global Photovoltaic Specialists Inc. (Canoga Park, Calif.) and Spire Corp. (Bedford, Mass.).

PHOTO : Field trials. Advanced thin-film copper-indium-diselenide (CIS) photovaltaic modules produced by Siemens Solar Industries undergo outdoor environmental testing. CIS units like these have demonstrated long-term stability at nearly 10 percent efficiency.

PHOTO : Rapid film deposition. Next-generation large-area cadmium telluride photovoltaic devices are produced by a promising closed-space sublimation process at Solar Cells Inc. Heated cadmium telluride powder evaporates in thin layers onto enclosed substrates.

PHOTO : Amorphous silicon module. United Solar Systems Corp. produces its 22-watt Uni-Solar battery charger modules by depositing thin films of noncrystalline silicon onto 2000-foot-long rolls of stainless-steel sheet. These commercial photovoltaic modules are about 5 percent efficient.

PHOTO : Efficient progress. Amorphous silicon (a-Si), cadmium telluride (CdTe), and copper-indium-diselenide (CIS) thin-film photovoltaic modules have shown consistent improvement in energy conversion efficiency since the early 1980s. The Department of Energy's efficiency goal is 15 percent. [Note: Arco Solar is now Siemens Solar.]

PHOTO : Single-crystal microspheres. Texas Instruments Inc. and Southern California Edison Co. have developed an innovative and potentially cost-effective solar cell technology based on thousands of tiny silicon devices embedded in aluminum foil. The spheral solar process starts with low-cost low-purity silicon and refines it in a unique manner.
COPYRIGHT 1992 American Society of Mechanical Engineers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:includes related articles
Author:Ashley, Steven
Publication:Mechanical Engineering-CIME
Date:Jan 1, 1992
Previous Article:Tapping landfills for energy.
Next Article:Safety at a glance: upgrading the displays in a nuclear plant control room.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters