We examine laboratory and commercial solar cell efficiencies of Crystalline, CIGs, Amorphous, Cadmium Telluride and Multijunction Concentrator cells.
In solar, it's hard to go a month without hearing news about conversion efficiencies. In September, for example, Oerlikon Solar and its partner, Corning, said they broke the world efficiency record for a lab-created tandem-junction amorphous-silicon cell. The cell, which was tested by the U.S. National Renewable Energy Laboratory, delivered 11.9% stabilized efficiency.
Meanwhile, Conergy said its new selective-emitter technology could boost solar-cell efficiency from one of its German factories by "up to 0.5 percentage points." And scientists at Yonsei University and the Massachusetts Institute of Technology announced new technologies that could one day enhance cell efficiencies by up to 65%, in the case of Yonsei, and that could double cell efficiencies, in the case of MIT.
"Everyone's working on efficiency," said Paul Wormser, senior director of product engineering and system solutions at Sharp Electronics' solar division. "You would be hard pressed to find a manufacturer that hasn't stated publicly at least once a year something about an efficiency improvement program."
Martin Green, a professor and executive research director at the University of New South Wales' Photovoltaics Centre of Excellence, said he's seen cell efficiency improvements in the labs accelerate in the last few years, with about 10 new "highest confirmed efficiencies" for different photovoltaic cell and module technologies every six months.
Commercial cells lag behind the efficiencies achieved by labs in so-called "champion cells," or best of breed cells, and companies don't expect cells produced in high volume to reach the efficiencies of those top lab cells. But the efficiency gains in the labs have translated into commercial gains as well: First Solar, for example, reported that its panels grew from 10.9% efficiency to 11.2% efficiency from the second quarter in 2009 to the same quarter this year, Green pointed out.
The recently published October edition of Progress in Photovoltaics, which tracks the highest confirmed lab efficiencies for solar cells and panels, cites new records for the top concentrator cell, the top large-area crystalline cell, the top copper-indium-gallium-diselenide (CIGS) cells and the top tandem-junction amorphous-silicon cell, among others. Let's take a look at the best efficiencies that the labs - and factories - have produced so far.
For crystalline silicon technologies, efficiencies suddenly become far more crucial starting around 2005, during a worldwide shortage of solar-grade silicon that lasted until 2008. Silicon supplies were limited and expensive, and that gave manufacturers a huge incentive to eke as much power as possible out of each panel. "Pre-2008, getting silicon at all was a challenge, so you needed to squeeze as much as you could out of the silicon," said Jenny Chase, lead solar analyst for Bloomberg New Energy Finance. Because solar panels are sold per Watt of peak capacity, not per panel, manufacturers that boosted their efficiencies could grow their production capacity - and profits - without having to make more panels or access more silicon.
Looking at the natural limits of the materials, crystalline silicon could reach a theoretical efficiency of 28, 29 or 30%, scientists say. Theoretical efficiencies are based on lab conditions that will never be found in commercial production, warned Lars Waldmann, director of public relations for Schott Solar. But in the labs, researchers are three to four percentage points away.
The University of New South Wales, which holds the record for the most efficient crystalline-silicon cell, created a cell with an efficiency of 25%. Sandia National Laboratories tested the cell in 1999, and it was also used in a record-setting solar panel with 22.7% efficiency, according to the university's School of Photovoltaic and Renewable Energy Engineering.
Meanwhile, SunPower makes the most efficient silicon panels on the market with 19.3% efficiency, according to Photon's annual module overview, which came out in February (some efficiency is always lost when companies combine cells into panels). In May, the company announced a new line of panels rated for up to 19.5% efficiency, and in June, SunPower announced it had set a new world record for large-area silicon solar cells with a conversion efficiency of 24.2% measured by the National Renewable Energy Laboratory.
The drive toward higher efficiencies is less critical now than it was a few years ago, during the worldwide solar silicon shortage, Chase said. "Now companies can get as much silicon as they need at reasonable prices." Silicon prices were $59 to $60 per kilogram in September, according to Bloomberg New Energy Finance. Companies were rumored to be paying spot prices as high as $400 per kilogram in 2007, according to the Prometheus Institute at that time.
But when the silicon shortage ended in 2008 and panel prices started falling, companies had another reason to boost efficiencies: to lower costs. Solar panels are sold per Watt of peak capacity, not per panel, which means that a more efficient panel can be sold for more money. If manufacturers can increase their panel efficiencies, the same factory can produce more megaWatts worth of panels without having to add production lines. Aside from the potential to lower the panels' cost per Watt, higher efficiencies can also reduce the cost of installation - including the pieces such as racking and mounting, wiring and inverters - because fewer panels are needed to deliver the same amount of power, Chase said. And the highest-efficiency cells and panels can also sell at a premium, a major advantage as manufacturers' profit margins are being slashed.
While the lower panel prices may be a driver for efficiency, they also could limit the amount that efficiencies can grow. Companies will only take steps to raise efficiencies if those steps cost less than it would cost a developer to simply add more panels to achieve the same result.
The new Progress in Photovoltaics edition includes two new results for copper-indium-gallium-diselenide cells, that have demonstrated the highest lab efficiencies of any thin film. In April, a CIGS cell on glass, made by NREL, tested at 19.6% efficiency, replacing an NREL record of 19.4% from January of 2008.
Meanwhile, ZSW Stuttgart produced a cell that delivered 20.3% efficiency when it was tested by Fraunhofer this summer. That cell had an aperture area of only 0.5cm2, making it too small to be accepted as an outright record, Green said. Measurement errors are more likely to happen with small areas and the champion cells can also be less representative of the group because it's possible to make thousands on a single substrate and sort through them for the "flash in the pan," he said.
But the high efficiency - confirmed by Fraunhofer - won the ZSW Stuttgart cell a spot on the publication's "Notable Exceptions" chart, which notes highly efficient cells and panels that don't meet the standards for class records. The cell replaced another ZSW Stuttgart cell, tested by Fraunhofer just six months ago in April, with 20.1% efficiency.
The top commercial CIGS panels, made by Q-Cells using Solibro cells, get up to 11.2% efficiency. Würth Solar also sells a copper-indium-diselenide cell, with no gallium, that delivers 11.8% efficiency.
In August, Oerlion Solar's above-mentioned 11.9% cell, measured by NREL, broke the previous record of 11.7% efficiency set by Kaneka back in 2004. Oerlikon's cell included a new, thin light-trapping glass from Corning. O'Brien said Oerlikon also improved its tandem technology with a better (and cheaper) reflective backsheet -- the sheet at the back of the cell that reflects the photons that get past the silicon layers back into the silicon for another chance to convert those photons into electricity - and a thinner silicon layer, which boosts stabilized efficiency. And the technology has the potential for even higher efficiency: The cell didn't include the usual antireflective coating to enhance the capture of light, O'Brien added.
The top commercial amorphous silicon (a-si) and microcrystalline panels are made by Pramac, an Oerlikon customer, with 9.2% efficiency. Sharp - as well as IBC Solar using Sharp cells - makes the next most efficient panels at 9% efficiency. These types of cells have the potential to reach more than 11% efficiency in 2012, said Thomas Block, production manager in strategy and business development for Schott. Waldmann added that the technology could reach a panel efficiency of up to 12% in the coming years.
Meanwhile, in Progress in Photovoltaics' "Notable Exceptions" section, Uni-Solar produced a tandem-junction cell with 12.5% stabilized efficiency last year. The cell, tested by NREL in March 2009, had layers of a-si and nanocrystalline silicon. But it only had a designated illumination area of 0.27cm2, which is why it isn't listed as the official record. Uni-Solar claims it also holds the world record for flexible a-si, with a cell that demonstrated 15.4% efficiency in the lab.
In production, Uni-Solar's current cell efficiency is 8.2%, with a panel efficiency of 6.7%. Schott has a tandem-junction a-si cell with more than 10% efficiency, Block said. Meanwhile, Uni-Solar plans to deliver commercial panels with 12% efficiency by 2012 and in June released a roadmap anticipating efficiencies could eventually rise above 20%, at a price of 95 cents per Watt.
First Solar, the company that popularized cadmium-telluride, is the world's largest solar manufacturer, with 1.1 gigaWatts of production last year. It owes much of its leadership to having the lowest announced per-Watt costs in the industry. The company, which broke the $1-per-Watt milestone back in February of last year, in the second quarter announced it's producing panels at a cost of 76 cents per Watt. And a big part of the reason for its low costs has been its growing efficiencies, Chase said.
"One of the reasons First Solar is so successful is because they keep improving the efficiency of their modules, which means their effective capacity increases without them spending new money on new lines," she said. "They're doing the same thing, but because they're doing it slightly more smartly, they're getting more Watts and paying slightly more, but can sell for [more money] because the modules are sold per Watt."
In an announcement in September, Sunovia Energy Technologies and EPIR Technologies claimed they've made a technology breakthrough that could raise that potential efficiency. They said the cadmium-telluride technology set a new world record for open-circuit voltage, which correlates to efficiency, and estimate they will be able to make two-junction cells with a production efficiency of 35%.
Multi-layered cells with concentrators are the world's most efficient cells, bar none. And in a test in September, Spire broke the record for this type of cell. Its multijunction concentrator cell, boasting 41.3% efficiency, included layers of indium-gallium-phosphide, gallium-arsenide and indium-gallium-arsenide, as well as a concentrator that magnifies the sunlight 406 times (or, in industry language, 406 suns).
Spectrolab had set the previous record with a 364-sun concentrator cell made of layers of gallium-indium-phosphide, gallium-indium-arsenide and germanium in August of last year. Sharp also announced in September that it broke the record, but not by as much as Spire, which boasts a 42.1-% efficient cell developed in partnership with Tokyo University.
"Congratulations to Spire. It's certainly a good result and it reflects the progress being made in the [concentrating photovoltaic] industry on tech improvements," said Lillington, who added that the record has been broken about every six to nine months. "Fortunately, we've been able to be the world record holder for a number of years and it's probably unrealistic to expect to maintain the record 100% of the time, but we have a roadmap that will hopefully put us back in the lead."
These cells are targeted at concentrating-photovoltaic projects, which use mirrors or lenses to concentrate sunlight into the cells. The projects require far fewer - and smaller - cells than those used in traditional solar panels, enabling companies to use more expensive, but more efficient, materials. CPV cells use 400 times less semiconductor material than conventional PV cells, Lillington said.
Typically, Spectrolab has been able to turn lab cells into commercial cells in about two years, Lillington said. "These are more than laboratory curiosities," he said. The company is already producing cells with 38.5% average efficiency today, and plans to launch a new cell with 40% average efficiency in the first quarter of 2011 and a 41.5% efficiency cell in late 2012 or early 2013, he added. These cells could theoretically reach 55 to 60% efficiency, but realistically, Lillington said he thinks they will top out in the 45% range somewhere between 2015 and 2020.
The company plans to grow its production from 40 MW of CPV cells this year to roughly 120 MW next year.
When does the drive to efficiency end?
The push toward higher efficiency is likely to peak at some point when efficiency gains are no longer commercially viable, O'Brien said. "That's always been the case with PV where you want to make sure you're not pursuing the perfect efficiency to the point where [it's no longer cost-effective,]" he said. "Many times increasing the efficiency will increase the cost of the cell of module, and you want to make sure that the added value outweighs the cost of getting that higher efficiency."
Of course, efficiencies are hardly the only important metric to the industry. Cost - both of the panels and of installation and other "balance of system" components, such as wiring and inverters -- is obviously another big factor, which is why companies often speak in terms of their cost per Watt.
Another key metric is the amount of electricity that the panels deliver, which isn't necessarily reflected in their efficiencies. That's because efficiencies measure power, or the highest amount of Watts that can be produced in peak conditions, rather than energy, or the number of kiloWatt-hours of electricity that the system can produce. Amorphous-silicon technologies can usually produce more electricity in diffuse light or in high temperatures, for example, than crystalline-silicon technologies with the same peak capacity, O'Brien said. "Customers shouldn't just look at the efficiency label on the module, but also should pay attention to the expected energy output on the module," he said. At least in the first few years of testing, amorphous-silicon panels consistently produce 5 to 10% more kiloWatt-hours than crystalline panels with the same rated capacity, Green said.
Companies have shifted to talking about the levelized cost of energy or cost per kiloWatt-hour in addition to the cost per Watt, and many companies also are discussing the payback time or return on investment that customers can expect, Wormser said. The different metrics are important because they help customers pick the right technologies for their specific projects and locations, he added. For some projects, cost per Watt will be more important; for others, cost per kiloWatt-hour. In some cases, it makes sense to buy a less efficient system because it comes at a better price, Wormser said.
In some locations, it may be simpler to connect a project to the grid if the capacity is under, say, 20 MW, than if it's above 20 MW, he said. Customers in that case may want to keep their capacity below 20 MW, but will want to get as many MW-hours as possible within that capacity, he explained. Because of all the variations that come with different locations, policies, capital expenses and types of projects, different technologies will prove to be ideal in different conditions, he said. "That's why we have so many technologies from so many companies on the market today."
1. Marktübersicht, Solarmodule 2010, Korrektur zu PHOTON Profi 2-2010, https://photon.de:448/download/marktuebersicht_solarmodule_2010.pdf
Jennifer Kho is a freelance reporter and editor based in Oakland, Calif. She has more than a decade of journalism experience and has been covering green technology since 2004.
Past PVWorld Issues