Grid Scale, Solar

EU PVSEC: Photovoltaics getting bigger, faster, thinner, more efficient

Among the highlights from the plenary session of this year’s EU PVSEC, was an examination into the industry’s continued reliance on 1970s-era solar cell designs, and what can be done to evolve the technology to push cell efficiencies way up and costs way down.

September 6, 2011 – PV technology is stuck in the ’70s, according to Dr. Martin Green. The plenary session of the 26th European Photovoltaic Solar Energy Conference (EU PVSEC) featured many highlights, including an announcement of a Greek ?20 billion PV project called “Helios” that aims to sell energy produced in region’s sunny climate to the rest of Europe as solution to the country’s financial woes, and the latest research that could enable cell efficiencies greater than 50% (through five- or six-level multi-junction cells for CPV applications). But perhaps the most intriguing look at the solar PV industry came from veteran Dr. Martin Green of the University of New South Wales in Sydney, Australia, who lays claim to numerous world record efficiencies.

Green said that, in a nutshell, the industry is moving to “bigger, faster, thinner, better and more efficient,” and the bottom line is “much lower cost” — but the industry is still using cell designs that had been developed in the 1970s. Many improvements that have been demonstrated in the laboratory since the late 1970s have yet to find their way into a commercial product, he said, because most manufacturers are still using the same cell design that was known and demonstrated back then. “If you look at a present commercial cell, it’s essentially a black cell. The difference is that the metallization is done by screen-printing,” he explained. “But from the point of view of cell operating principles and design, it’s essentially a cell of the 1970s.”

Historically, the first reasonably efficient solar cells were demonstrated in 1954 (“Vast Power of the Sun is Tapped by Battery Using Sand Ingredient,” heralded the front page of the New York Times on April 26, 1954). This was made possible by the rapid development of crystalline silicon technology for the then fledgling semiconductor industry. By the mid-70s, so-called ‘black’ cells had been developed, where the use of surface etching to add texture gave them the appearance of black velvet.

Green acknowledged that some companies had indeed implemented more advanced designs, including rear contact metallization (a cell design developed by Stanford University). This illustrates that “laboratory efficiencies can ultimately get transferred into production,” he said.

At EU PVSEC, a probable industry roadmap for crystalline-silicon-based cell designs was provided by Applied Materials’ Jim Cushing, senior director of product management at Applied Solar. He said conventional cells would evolve first to double printed (DP) front contact metallization — with screen printing — followed by selective emitter designs, which would go into production across the industry in 2011 (now mostly in pilot line mode, at all but leading-edge customers). Two-sided passivation and back contacts would follow in “Gen 2” designs. Technologies such as selective emitter have enabled an efficiency increase of about 0.4%-0.5%.

Green said the big change in photovoltaics has been the drop in price, and there’s still room for continued improvement. “There might be a perception that it’s all running out of momentum, but there’s plenty of scope for further cost reductions,” he said, adding that the goal of achieving $1/Watt is feasible: “We’ll see it quite soon.” Cost reductions will mostly come from improvements in polysilicon growth and wafering technology — silicon wafers count for about 46% of the total of module costs, with about half of that cost going to the production of the raw polysilicon, and the other half going to the processing that turns the polysilicon into wafer form. “There’s a lot happening in both areas and we’re likely to see perpetually ongoing cost reductions in those areas,” Green said.

Another part of the cost savings will come from suppliers of polysilicon being forced by competitive pressures to accept lower margins. Green expects margins to come more in line with those of a mature industry — i.e., some manufacturers are expected to get their own cost to below $20/kilo this year, and cost to manufacturers will be $30/kilo.

Even greater cost savings will come from process technology improvements that could enable great use of metallurgical grade silicon, which is about 10× less expensive than semiconductor grade silicon. “There’s still hope for manufacturers to use upgraded metallurgical grade silicon,” Green said, suggesting that changes to wafering technologies or in the cell processing technologies could enable use of materials closer to the metallurgical grade than the semiconductor grade. “There are some economies of scale in going to larger ingots (over 1 ton), and improvements in quality are also possible,” he added.

One very recent development is the move to new casting technologies that result in monocrystalline silicon in the center of the block and a multi-crystalline ring around the outside. “It will have quite different properties than CZ material that the industry presently uses,” Green explained, which perhaps will open doors to “perhaps exceed the efficiencies with solidified material from what’s traditionally obtained with Czochralski.” The secret, he revealed, is temperature control: “You have a seed down at the bottom of the ingot. You have to partially melt that seed without melting all of it, so you need to finely control the temperature during the initiation of the growth.”