Concentrating PV

Why the PV world is starting to focus on concentrating technologies Close up of a 35 kW concentrating solar unit AMONIX Concentrating light on a solar cell greatly increases its output. With improvements in technology now starting to offer efficiencies of 38%, the concentrating PV sector says that US$3/W systems are just a couple of years away. Jackie Jones reports. The last two to three years have seen an upsurge of interest in concentrating PV technology, as participants at the International Conference on Solar Electric Concentrators held in Arizona, US, in May, those at this year's European PV Solar Energy Conference and Exhibition in Barcelona, Spain, and at the upcoming PV Solar Energy Conference and Exhibition (Shanghai, China, October) will all testify. In 2004 just 1 MW of new concentrating PV was installed, yet there are 18 MW in the pipeline for 2006. Several new companies are bringing products to market, while industry giants Sharp and Isofotón are developing concentrating PV systems. What is the appeal? And why now? Is this budding branch of the industry about to flower at last? WHY NOW? Concentrating PV technologies have existed since the late 1970s, and concentrator cells with over 30% efficiency were first described in 1989. If they have not yet penetrated the market, what makes them likely to do so now? When NREL's Bob McConnell and Lewis Fraas of JX Crystals answered this question back in 2002 [REW September-October 2002], they identified two reasons. The first was that the main market for CPV was going to be utilities, but that the market had not yet taken off because system performance and reliability had gone unproven. By 2002, however, systems were finally providing valuable data and experience. Also in 2002, Arizona Public Services (APS) reported that the installed cost of its 300 kW system had been $6000/kW - acceptable for a demonstration plant but by no means competitive. At that time the sector predicted that it could achieve a price of $1500/kW by 2010 given sufficient support for a demonstration programme. Three years on, the industry is talking about an installed system cost of $3000/kW 'within two to five years'. 'Concentrating solar electric power is on the cusp of delivering on its promise of low-cost, reliable, solar-generated electricity at a cost that is competitive with mainstream electric generation systems,'Vahan Garboushian, president of Amonix, Inc. of California, announced this Spring, adding that 'with the advent of multijunction solar cells, PV concentrator power generation at $3 per watt is imminent in the coming few years'. Herb Hayden, Solar Programme Co-ordinator at APS, said in May: 'Our goal is to install PV concentrator systems at $3/W, which can happen soon at production rates of 10 MW per year. Once that happens, higher volumes are readily achievable.' CONCENTRATING PV - THE BASIC PRINCIPLE The principle of concentrating PV (CPV) is quite straightforward. In the familiar 'flat-plate' PV modules, a large area of photovoltaic material (usually crystalline silicon) is exposed to the maximum naturally occurring sunlight. Normally, that maximum is achieved by installing the modules at an incline optimized for the latitude, but sometimes they are installed on moving frames that can follow, or track, the sun as it passes across the sky. The PV cells perform under direct (sunny) or diffuse (cloudy) radiation conditions, but output is at its highest when the maximum amount of light falls on the cells (assuming there are no detrimental effects from overheating). The amount of light that falls on a cloudless day (this varies according to location and season) is regarded as one 'sun', which is defined as 1000 W/m2. Concentrating PV systems use lenses or mirrors to focus sunlight onto a small amount of photovoltaic material. (Usually the Fresnel lens is used, a flat lens that uses a miniature sawtooth design to focus incoming light. When the teeth are arranged in concentric circles, light is focused at a central point. When the teeth run in straight rows, the lenses act as line-focusing concentrators.) The concentration ratio can vary: if the light that falls on 100 cm2 is focused onto 1 cm2 of PV material, the ratio is considered as 100 suns. If the light from 10 cm2 is focused onto that 1 cm2, the ratio is 10 suns. If the concentrated sunlight light falls onto a well designed CPV cell, the cell will produce at least 100 times, or 10 times, the electricity. (In fact the conversion efficiency of cells increases under concentrated light, so the correlation is likely to be greater than one-to-one, depending on the design of the solar cell and the material used to make it).1 While commercial concentration ratios are around 200 to 300 suns, as much as 1000 suns is expected for future concentrating PV systems. As most CPV systems use only direct solar radiation, these installations almost always involve trackers (rotating about either one axis or two axes and therefore called one-axis or two-axis tracking) to keep the sun focused on the solar cell. 10 MW may sound far-off for a sector that installed just 1 MW in 2004. Yet as mentioned earlier, 18 MW is due for installation in 2006, according to Bob McConnell,1 Senior Project Leader at NREL's National Center for Photovoltaics in the US, and convenor of the IEC Working Group for CPV Standards. If that growth curve can be continued, the sector would look very healthy indeed. BENEFITS OF CPV First among the main benefits of concentrating PV is its efficient use of solar cell material.For a given electrical output, concentrator systems use a far smaller amount of (expensive) semi-conductor or solar cell material than flat-plate systems. The large areas of mirrors or lenses used to provide optical concentration are inexpensive compared with large areas of solar material, so this can keep system costs low. The manufacturing process is lower-tech than for flat-plate modules - Bob McConnell likens it to automobile manufacturing - and it can be scaled up relatively easily and economically, giving concentrating technology a further advantage, says sector insiders.Table 1 shows an estimation of manufacturing investment costs. TABLE 1. Investment required for a 100 MW/year manufacturing plant1 Technology Cost (US$ million) Crystalline silicon PV 150-300 Thin-film PV 150-300 Concentrating PV 30-50 Cell options Most of the systems on, or entering, the market use crystalline silicon cells.With silicon supplies short, and prices high in the current climate, the ability to offer high output from a small quantity of silicon is extremely attractive. Concentrating PV also offers the option of shifting away from crystalline silicon to use the very high-efficiency, nonsilicon cells. Such cells have mostly been developed primarily for space applications.These multi-junction III-V cells (which use elements from columns 3 and 5 of the periodic table, typically gallium and arsenide) are prohibitively expensive for extensive use in large flat panel arrays. Concentrator systems however, because they require smaller and fewer cells, can afford the higher cost of multi-junction cells and yet still be manufactured at lower dollar-per-watt cost than flat plate modules. WHERE DOES CPV WORK BEST? CPV systems function best under clear sky, direct-sun conditions. Early installations were made in Saudi Arabia, Arizona and at Alice Springs, Australia, and while desert conditions are not essential, the technology is likely to make its market entry in countries or locations with a high solar resource.The Spanish market - currently a key growth market for photovoltaics thanks to its feed-in tariff - is demonstrating keen interest in CPV, as the activity around exhibitors at the 20th PV-SEC in Barcelona, such as new market entrant Whitfield Solar, demonstrated. In terms of application, the future of CPV has always seemed most likely to be in ground-mounted, utility-scale plants. If high-efficiency cells are used, such systems can be highly efficient and provide utilities with that all-important low cost per watt. There are economy-of-scale benefits for large systems; in hot climates, systems can generate significant amounts of power precisely when demand peaks due to the air-conditioning load. And while a few years back utility-scale PV was largely untried, in the space of the last two years, in response to Germany's (and more recently, Spain's) market incentives, multi-megawatt ground-mounted flat-plate PV installations have swiftly become the norm.Thanks to their acceptance and the experience gained with them, the pathway for CPV is considerably easier than it would have been a few years ago. What is more, some of these systems use trackers, previously regarded as a 'weakness' of CPV systems. Concentrator systems using trackers were until recently not really considered suitable for rooftop use, due to issues of weight and lack of ability to withstand windy conditions. However, some systems that are designed for rooftop use are now close to market. Some use concentration factors as low as 10 suns, while others use higher concentrations. While some manufacturers have chosen not to include trackers, in order to achieve a slimmer profile,other new products are fully tracking while remaining remarkably easy to handle. SOME OF THE TECHNICAL CHALLENGES AND DEVELOPMENTS In theory, solar concentration ratios of several thousand times are possible. In practice, higher concentrations raise a range of technical issues, including the need for accurate and reliable tracking, adequate heat sinking and cell technology capable of converting the energy efficiently at this intensity. 'Supercells' - increasing efficiency Steady and impressive improvements in cell efficiency are feeding through to the concentrating PV sector. Some of the current CPV technologies are already making use of crystalline silicon cells with efficiencies as high as 26%.Though most of the multi-junction, high-efficiency cells were developed for the space sector, cell manufacturers are working closely with concentrating PV researchers and manufacturers in optimizing cells for concentrating terrestrial applications (see box on efficiency gains, overleaf). EFFICIENCY GAINS 2004 - Spectrolab announced 37.3% efficiency Apr 2005 - Fraunhofer Institute for Solar Energy Systems (ISE) in Freiburg, Germany, announced 'European' record 35.2% efficiency for a cell measuring 0.031 cm2 Consists of gallium indium phosphide, gallium arsenide, and germanium. Suitable for use in terrestrial concentrator modules May 2005 - NREL announced a new solar cell efficiency record 37.9% at 10 suns June 2005 - Spectrolab announced 39% efficiency at 236 suns Though greater concentration can raise efficiency, the concentration ratio needs to be optimized for each particular cell type (it is not possible for just any cell to function well when exposed to several hundred suns). Beyond a certain concentration, efficiency starts to drop off due to increased resistance losses within the cells. Because of the smaller amount of photovoltaic material involved in CPV, cell efficiency improvements should be able to be quickly taken up by marketable products. Dave Holland, CEO of Solar Systems Australia has said that the 'new solar cell technology from Spectrolab will enable us to upgrade our systems from 24 kW to 35 kW - a 46% increase in output'. FROM LEFT TO RIGHT Three large 35 kW systems used by the Arizona Public Service l Microdish made by Concentrating Technologies using Spectrolab solar cells in Arizona l A series of 20 kW Solar Systems dishes in Alice Springs Australia NREL However, utilizing high-efficiency cells is not quite as simple as just swapping one cell for another and going to market. One important issue to bear in mind is that in terrestrial applications the cells need to be able to cope with environmental conditions that they do not face in space, particularly varying levels of humidity. Thus the cell manufacturers have to put their product through extensive trials to measure performance under varying conditions, over time. CPV manufacturers aim to be able to offer system lifetimes of at least 20 years, so they, and investors, need confidence that the cells will operate well over this timespan. According to NREL researcher Sarah Kurtz,2 the current production capacity for high-efficiency cells is approximately 500 kW/year (at one sun).However,concentrated at 1000 suns, this capacity becomes about 500 MW/year. She estimates that the cost of a cell at 1000 suns could be approximately $0.10/W. The possibility of using the super-efficient solar cells developed for space applications is a source of increasing excitement.NREL's High Performance Project plans to achieve over 41% cell and 33% system efficiencies.Many multi-junction approaches are undeveloped,according to Sarah Kurtz (such as lattice mismatched materials, mechanical stacks, voltage matched), while 5- and 6-junction approaches are being explored.2 Keeping on track CPV needs to be reliable if investors and utilities will buy into it. As Whitfield Solar points out, accurate and reliable tracking of the sun is needed in highly concentrating devices to maintain the focus of the solar energy on the cell and yield the best results - good systems can accurately to within a tenth of a degree. Secondary optical lenses can help minimize tracking accuracy requirements at higher concentration ratios. However, to achieve a low-cost system, the price of the tracker technology has to be kept low. Some critics have said that the need for moving parts is a reason why CPV can never be reliable. However, this myth is disputed by Sarah Kurtz of NREL,2 who comments that there have been very few tracker problems on the existing CPV installations - and that in fact inverters have proved more problematic than trackers. A 100 kW installation in Glendale, Arizona AMONIX All the same, it is important to introduce and maintain high standards across this new sector, and Bob McConnell explains that industry standards for PV trackers are currently being worked on by the IEC International Electrotechnical Commission (see box on Standards, p. 90). Sarah Kurtz cautions that new CPV companies 'almost always underestimate the investment needed to achieve a reliable product.'2 Limits of the lens No lens can transmit 100% of the incident light - the best that lenses can transmit is about 95%, and in practice, most transmit about 85%. In addition, concentrators cannot focus diffuse sunlight,which makes up about 20%-30% of the solar radiation available on a clear day (but this depends strongly on the location).3 To help simplify things, IEC standards are under development for CPV systems using lenses or mirrors. Coping with overheating High concentration ratios can potentially introduce a heat problem. Because cell efficiencies decrease as temperatures increase, and higher temperatures also threaten the long-term stability of solar cells, the cells must be kept as cool as possible. However, this has not been problematic in concentrator systems - maintenance of temperatures is generally achieved by using a highly conductive material such as copper directly behind the cells to spread the heat, and some systems use air cooling.According to a rule of thumb, a heat spreader area is needed equal to the aperture area. In practice, concentrator cells operate in the same temperature range as flat plate PV cells. In the case of dish concentrators (such as the Australian Solar Systems design) the cells can actually be cooled to lower temperatures. Minimizing electrical resistance without shadowing A key design consideration is minimizing electrical resistance where the electrical contacts carry off the current generated by the cell.Additional fingers in the contacting electrical grid help achieve low resistance, but their shadow can prevent some light from reaching the cell. One solution to the problems of resistance and shadowing is prismatic covers that direct incoming light to the parts of the cell's surface that lie between the metal fingers of the electrical contact grid. Another solution is a the back-contact cell, which differs from conventional cells in that both the positive and negative electrical contacts are on the back. Placing all the electrical contacts on the back of the cell eliminates power losses from shadowing, but it also requires high-quality silicon material.3 WHO IS INVOLVED IN CPV? The number of companies offering or developing CPV is on the increase. At least two of the new market entrants are spin-offs from research institutes that have been working in the field for many years.Among the current players are: Amonix With more than 15 years in the sector, and five generations of CPV prototypes, California-based Amonix has installed systems in Arizona, and in other southwestern US states. (Arizona has a mandate that 1.1% of retail electricity has to come from renewable sources, and that 60% of that must be from solar electric technologies;Arizona Public Services has contracted with Amonix for over 0.5 MW of installation.) Amonix uses point-focus Fresnel lenses, and its own 26% efficiency silicon cells, with a concentration ratio of 250 suns. Amonix manufactures a 5 kW 'Megamodule' that fits on a flat bed truck.Typically 5-7 Megamodules make up a single unit providing 25-35 kW. This summer, Amonix announced a technology transfer agreement with Guascor of Spain (see below - Guascor). Concentrating Technologies, Inc. This company is working with Spectrolab on the first gridconnected concentrator system that utilizes Spectrolab's GaInP/GaAs/Ge triple-junction solar cells (see below - Spectrolab). Concentrix Solar GmbH A concentrating PV installation in Barcelona. Spain is expected to be a prime market for this technology WHITFIELD SOLAR Based in Freiburg, Germany, this company is a spin-off from Freiburg's ISE (Institute for Solar Energy), and has developed a product it plans to market under the name Flatcon (Fresnel lens all-glass tandem cell concentrator). Fresnel lenses concentrate light at 500 suns onto the 2 mm2 solar cells, using copper as a heat sink. During initial outdoor testing, module efficiencies of 22.7% were achieved using tandem cells with efficiencies of 32%. If higher quality triple cells are applied, module efficiency will rise to above 25%, and may reach as high as 28%, according to MD Hansjörg Lerchenmüller.4 Currently, Concentrix plans to build a pilot production line in Freiburg by the end of 2005. Entech Entech is based in Keller,Texas.An early player in concentrating PV,Entech has several NASA projects and has a new product for terrestrial use in development. Guascor The Guascor Group is investing in new manufacturing plant for CPV for the Spanish market, following an arrangement made with Amonix earlier this year. Guascor Fotón is constructing a factory near Bilbao, Spain, to assemble the systems, apparently using solar cell assemblies shipped from Amonix's California plant. It is understood that Guascor plans to manufacture and install 10 MW of CPV in Spain during 2006, and the capacity of the factory is expected to expand the following year. Isofotón Spanish manufacturerIsofotón is working on the industrialization of a high-concentration PV product, which it aims to bring to market at an installed system price of under E2.5/kW ($3.07 at current exchange rates). Demonstration projects will be installed during 2006.The product uses a highefficiency III-V cell, and a novel optics solution that consists of two lenses (TIR-R), manufactured by injection moulding. The top lens works by total internal reflection (TIR),and the bottom one by refraction (R). According to Isofotón,5 this solution is able to keep an almost fixed spectral optical efficiency,which is advantageous when using multijunction III-V cells. Isofotón's product is designed for high-volume production. The company says5 that tentatively, a level of 10 MW/year is the starting point in terms of achieving a low enough production cost and good price for customers. A production line producing 7000 cell/lens units per hour (three-shift scenario) is under development.5 This scale of production would require the manufacture of at least 1000 trackers/year. MicroPV Inc. With its manufacturing facility in China, and its sales office in the US, this company presented its product, which uses dome Fresnel lenses, and air-cooling, at the European PV-SEC in Barcelona. Cell efficiency is approximately 15% according to company information.The 1 kW systems has a module area of 9 m2, and has two-axis tracking. It is claimed to be suitable for rooftop or utility scale applications. NON-UTILITY APPLICATIONS Although the primary market for concentrating PV is likely to be utility-scale, the technology is also being targeted at offgrid applications, such as water pumping and purification and - potentially - hydrogen production. For example, Solar Systems Pty of Australia is marketing its product for Remote Area Power Supply (such as village systems), as a substitute for diesel fuel in hybrid grids, and for water pumping and purification. Sharp Solar and Daido Steel PV giant Sharp is working with Daido Steel of Japan on a 500-suns 'Sharp Light Concentrator' system based on a 7 x 7 mm gallium arsenide cell,which it presented at this year's Intersolar (Freiburg, June). The company says the system, which uses Fresnel lenses, enables the cell to achieve a conversion efficiency rate of up to 40%. The concentrator system has 90 cells, and uses axis tracking. According to Daido Steel, which is manufacturing the systems in co-operation with Sharp, the domed lenses are optically more efficient than typical Fresnel lenses. They are manufactured using injection-moulding techniques (by a separate company, Daido Metal).The ability to manufacture a high-quality lens at low cost is regarded as something of a breakthrough for the technology. Daido has also developed a lightweight tracker, and a semi-transparent CPV device that has been designed for use in plant breeding. Solar Systems Pty Based in Melbourne, Australia, this company's dish concentrator product, the CS500, uses SunPower cells which have 26% efficiency under concentrated light. Dish units can be combined to give a range of sizes, from 20 kW to multimegawatt installations.Amongst its existing installations is a 14-module plant at White Cliffs, New South Wales. Solar Systems has recently put in a number of installations on aboriginal lands in Australia, and plans to install over 5 MW in 2006. Spectrolab Inc. Based in Sylmar, California, this Boeing subsidiary has been providing solar cells to the space industry for 40 years. It makes its high-efficiency multi-junction solar cells available for use in high-concentration PV modules. Spectrolab has continued to produce world-record concentrator cells - 34.2% in 2001, 36.9% in 2003 and 37.3% in 2004. In June 2005 the company announced 39% efficiency at 236 suns. Just as important was that (terrestrial) field testing over the course of one year had shown no degradation in the solar cells. Spectrolab is currently being funded by NREL to produce a 41% concentrator cell under the High Performance Photovoltaics programme. CPV modules mounted on Daido Steel's two-axis tracker DAIDO STEEL Spectrolab is working with many concentrator PV companies around the world. It has teamed up with Concentrating Technologies, Inc for a demonstration at the Arizona Public Service (APS) to develop and deploy the first grid-connected concentrator system that utilizes Spectrolab's GaInP/GaAs/Ge triple-junction solar cells. This module is currently operational at the APS Solar Test and Research (STAR) facility in Tempe,Arizona.The data collected from this module shows reliable performance of the solar cells under 500-sun concentration. Additional reliability data have been collected by NREL on the Spectrolab cells working in a Fresnel lens module, which has been operational since April 2004. Whitfield Solar A newcomer in commercial terms, this UK-based company was set up in April 2004 to develop and manufacture a CPV system resulting from 30 years of research at the University of Reading under the direction of Dr George Whitfield. The new company presented its prototype low-cost two-axis solar concentrator for the first time at the European PV Solar Energy Conference in Barcelona in June this year. CEO Clive Weatherby reported that the company had met with an overwhelming response. The design philosophy has been to produce the most effective, yet lowest-cost, system with minimal material usage, minimal silicon usage, and ease of manufacture.The resulting product is suitable for ground or roof installation, and features intelligent two-axis tracking to within 0.1 degrees. Each unit measures approximately 4 m x 1 m, with 24 troughs of 1 m dimension per unit, in series. The product uses Fresnel lenses and silicon cells. Current electrical performance is 250 Wp, with 300 Wp expected for 2006, and 350 Wp for 2007. ACHIEVING LOWER COSTS The big driver for concentrating PV - as for the industry as a whole - is bringing down the cost per watt installed, and ultimately per kWh produced. A rule-of-thumb cost for flat-plate PV is an installed system cost of $5-6/W. The CPV industry is looking at installed systems at $3/W before long. THE IMPORTANCE OF STANDARDS For concentrator products to be successful in the marketplace, three issues are critical: high performance, low cost and long-germ stability. Standards help ensure this. Robert McConnell of NREL is Convenor for the working group within the International Electrotechnical Commission (IEC) developing standards for CPV. The goal is to develop an agreed-upon set of tests that 'stress' samples of CPV systems by cycling them between extreme temperatures (110°C and -40°C), exposing them to high humidity, high voltages and currents, and assessing them for impact of wind and hail. McConnell is 'personally committed to the importance of standards to help companies bring their products to market', he explains. Industry standards provide manufacturers, developers and investors with confidence in products and their performance over time. IEEE standards for CPV were introduced in the US in 2001, IEC standards are under development for CPV using lenses or mirrors, and industry standards for PV trackers are also being currently worked on by the IEC. Because of the 'multiplier effect' of using concentrated light, the performance benefits of high efficiency cells are enhanced in CPV; according to calculations by Sarah Kurtz of NREL,2 the system cost can fall dramatically as cell efficiency climbs (see Figure 1). Comparisons of the cost CPV in relation to flat plate go out of date quickly, as the cost of flat plate PV has been falling steadily. So it is important to compare the expected price of matured CPV (expected to be near $3 per watt in 2-3 years) with the projected price of flat plate PV in those 2-3 years - possibly $3-4. Hansjörg Lerchenmüller of Concentrix explains that several independent cost assessments (including his own company's calculations) have resulted in something around Û2.5/W ($3.07 at current exchange rate) for concentrating systems with III-V cells. He also points out the importance of using a realistic economic lifetime and interest rate when making levelized electricity cost calculations. 20-25 years and 5%-8% would seem more reasonable than some of the ambitious figures sometimes used, he warns. THE STAGE IS SET For decades, concentrating PV has been not so much waitingin the wings, as workingin the wings. Now it looks as if technical advances and market opportunities are coming together for CPV to play a big role at last.People will be watching its performance in the marketplace closely. Jackie Jones is Editor of Renewable Energy World e-mail: rew@jxj.com With thanks to Prof. Bob McConnell and Dr Roger Bentley for their advice and support. REFERENCES McConnell, R. and Symko-Davies, M. PV FAQs. (2005) What's new in concentrating PV?US Department of Energy. EERE. Kurtz, Sarah (NREL). Presentation at World Renewable Energy Congress, Denver, Colorado, US, 31 August 2004 US Department of Energy, Factsheet from the Solar Energy Technologies Program EERE website. ISE Fraunhofer/Concentrix press release Diaz, V., Alonso, J., Alvarez, J.L., Mateos, C. (Isofotón) Presentation at the 3rd International Solar Concentrators Conference in Scottsdale, Arizona, US, in May 2005.