Many experts have commented that PV manufacturing can be viewed as highly synergistic with, say, the semiconductor industry. But why is this and, ironically, could this be behind some of the negative cost issues inherent in PV production?
For starters, both solar PV and electronics share a taste for high-purity silicon as their ‘bread of life’. This is the seat of the semiconductor action that they both rely on (excluding, for present purposes, non-silicon alternatives that have arguably not yet fully entered the mainstream).
But costs are high in electronics, which begs the question, should solar PV seek solutions away from electronics to achieve grid parity and deliver on its potential as a mainstream energy player?
Solar's origins can be clearly seen in the industry's evolution. As far back as the late 1950s, some of the biggest names in electronics spotted solar as an additional use for their expensive fabrication technologies. Japan's Sharp Corporation, best known as a global player in business and consumer electronics, started to develop solar cells as early as 1959. Panasonic has a background in solar cells and modules extending back to the mid 1960s. Kyocera has been in the business since 1974. Canon, Samsung Electronics and the Taiwan Semiconductor Manufacturing Company have many solar PV patents among them, even though they now have few products in the field.
And this process is still happening with a new wave of manufacturers joining the party. For example, microprocessor superstar Intel is putting US$50m into PV cell manufacturing as part of a joint venture, and expects early products to be shipped this year. LG Corporation has said it plans to invest substantially in solar energy, though it has not revealed the detail. Sony is adapting technology first used to manufacture traditional TV sets, for the fabrication of solar roof tiles developed by SolarCentury.
This illustrates the fact that synergies are to be found not only at the silicon/semiconductor level but also in the area of ‘active glass’.
According to SolarCentury's COO Derry Newman, “for us the synergy was obvious. A TV set is basically a piece of glass wrapped in plastic which, in essence, is what a solar panel is. The Sony UK Technology Centre had the experience and a highly-skilled workforce, and SolarCentury had the product and the demand.”
Solar PV production has been able to emerge on the back of the complex fabrication chain and infrastructure evolved for microelectronics. Both require the production of pure silicon ingots, the conversion of those ingots into wafers and the fabrication on the wafers of circuitry for either electronic ‘chips’, or solar cells. Much of the language used by professionals is common to both camps and many skills are transferable between the two.
Ideally, large companies active in both the semiconductor-based disciplines would like full commonality, enabling the same fabrication facilities (fabs) to produce both types of device. However, at some stage, the common fabrication path must divide. High-purity silicon produced by the Siemens Process is polycrystalline (hence polysilicon). Microelectronics incorporate a certain amount of this, but most devices are made primarily from monocrystalline (single crystal) silicon because this has no grain boundaries to hinder the passage of electrical charge carriers.
Electronics producers therefore have to convert polysilicon to the monocrystalline form, using the Czochralski process, Bridgeman technique or float-zone silicon process. Solar PV producers, on the other hand, gain relatively little by using the pricier monocrystalline form and can avoid the associated transition expense.
Much of the cost involved in the conventional fabrication chain lies downstream of the above, in the various processes used to convert pure crystalline silicon into functional microelectronic ‘chips’ or solar cells. These processes tend to be highly technical and expensive, though more so in the case of microelectronics where there is a stronger focus on circuit density and miniaturisation. For solar PV producers, the pressures that electronics manufacturers face in adhering to Moore's Law (which sets a ‘gold standard’ for the miniaturisation progress) are reduced, along with the associated expense.
Even so, solar PV cell manufacturers have to carry out complex fabrication steps including masking, etching, lithography etc. - all highly developed in the electronics arena. Silane may come into the picture once more when solar PV fabs utilise the precursor gas, just as producers of very large scale integrated (VLSI) electronic devices do, to deposit polysilicon onto a semiconductor wafer. Solar wafer producers will require mastery of chemical vapour deposition (CVD), a technology for applying materials in very thin layers - the thinner the better for silicon and other expensive materials – onto various substrates including silicon and glass. Deposition requires costly reactors in which environmental conditions and multiple process parameters can be closely controlled.
Costs escalate still further in processes such as plasma-enhanced chemical vapour deposition (PECVD). Producers of ‘active glass’ devices like flat computer and television screens are well versed in these processes, and their expertise marries well with solar PV requirements since most solar PV devices are also active glass-based.
Supplying to both camps
Applied Materials Inc (AMI) is one company which supplies production lines and related services to both camps, and therefore understands the commonalities and synergies. A supplier to semiconductor and display manufacturers for four decades, it entered the solar business a few years ago and now has both crystalline wafer and thin-film interests.
According to Mark Pinto, senior vice president and chief technology officer, solar and electronics both seek to use scarce and expensive high-grade silicon efficiently. This accounts for a common drive to cut ever-thinner wafers from silicon ingots and to minimise breakage when handling ultra-thin wafers. AMI supplies equipment for producing and handling wafers down to 180 microns thick, with 120 microns targeted. Equipment facilitating subsequent fabrication of solar cells onto the wafers has similarly benefited from experience gained in microelectronics.
AMI sees scale and production efficiency as key to bringing down solar PV costs - again closely mirroring what happens in microelectronics. Scale is addressed with ever-larger factories, and fabrication lines are now available to produce up to100MW worth of PV capacity annually, with GW-scale plants being mooted.
The company's own large scale of operations also helps: With hundreds of crystalline silicon systems installed worldwide and a US$1bn order backlog, AMI claims to be the world's largest supplier of such equipment.
Thin film PV is a fast growing area of activity. According to Charles Gay, vp and general manager of the solar business group, AMI's SunFab fabrication lines exploit advanced glass coating, laser scribing, lamination and panel cutting technologies adapted from those used in display production. A key PECVD step is derived from CVD processes used to produce ‘active glass’ for displays. Enabling production of arrays up to 5.7m2, setting a new standard for the industry, also exerts downward pressure on US$/Watt cost.
Gay believes that thin film will be able to deliver better than10% cell efficiency by 2010, up from some 8% currently. Improving optics to increase light capture is primarily a matter for solar specialists, but on issues like microcrystal fraction and uniformity, precise control of layer thickness, and improved contact layer performance, the electronics and solar camps can learn from each other. Having US$3bn in SunFab contracts, including one at GW scale is, admits Gay, helpful, given the current downturn in mainstream solar PV due to the global recession.
Also leveraging display technology is Oerlikon Solar, which was founded in 2003 out of a background in flat panel production. Like AMI, the company supplies production equipment. CEO Janine Sargent agrees that thin film is expanding fast: “We provide a risk-managed approach to help clients get into thin-film solar PV. We have customers ramping up in Europe, Taiwan and China with our innovative production solutions.”
Sargent advocates thin film for its price/performance and sees it eventually driving cost as low as US$0.70 per watt.
While electronics industry methods have adapted well to solar cell production, some in the solar camp believe that the microelectronics heritage can be a mixed blessing and that innovation is required from outside this inherently high-cost sector.
One approach is to replace entire portions of the electronics process chain with something more affordable. For example, several companies aim to purify metallurgical grade silicon as an alternative to the expensive “electronics route” production of pure polysilicon. They say that producing such upgraded metallurgy-grade silicon (UMG-Si) for about a fifth of the cost can more than compensate for the lower energy conversion efficiency of devices made from it (about 14%, though adherents say 18%-22% should be possible.) Cost reductions of this order, they argue, bode well for grid parity.
In fact, SolarWorld AG and Scheuten Solarholding BV in Germany highlighted this approach a couple of years ago by forming a joint venture to produce UMG-Si on an industrial scale. Canada's Timmenco Ltd and 6N Silicon have similar aims. Timmenco advocates metallurgical processes, such as directional solidification, sparging and other gas treatments, as low-cost alternatives to the Siemens Process. Dow Corning, too, says it has developed a novel way to produce solar-grade silicon from metallurgical grade and claims that blending the result with normal polysilicon feedstock can yield solar cells with attractive performance.
Others have focused their attention on improving on electronics-derived processes. In one notable example, Evergreen Solar Inc. has developed its String Ribbon technology as an alternative method for producing PV wafers. In essence, a set of special parallel ‘strings’ are pulled through a pool of molten polysilicon inside a custom furnace. A thin wafer is created between the two strings as the polysilicon cools and recrystallises after emerging from the pool. By achieving a continuous production flow, the method enhances efficiency and reduces cost.
Another example comes from Sovello AG (formerly EverQ), a joint venture formed to manufacture wafers at a new 30MW factory in Germany, using the same technology. Sovello is owned by Evergreen Solar, Norway's Renewable Energy Corporation (REC) and Q-Cells AG in Germany.
REC is notable in its own right, having combined process innovation with a wise business strategy in its thrust to drive down solar PV costs. On the innovation front, the company uses a proprietary fluidised bed process to manufacture high-grade silicon for both solar PV and electronics use. It claims it is cheaper than established methods, not least because the process is continuous and lends itself to automation. Managers believe that the company's improved silicon production can drastically reduce payback times for solar modules.
REC also operates a patented continuous-flow process for producing ultra-pure silane gas, a key silicon precursor. This involves the gasification of metallurgical-grade silicon, then reduction/distillation of the resulting trichlorosilane to silane. Starting materials are obtained from independent sources so that there is no dependence on by-products or intermediates from other industries. REC has noted an increased requirement for its silane gas resulting from the ramping up of solar PV demand alongside that from the semiconductor industry.
Strategically, REC has developed a presence right across the supply chain, from silane and polysilicon production through to solar cell and module fabrication. Extensive vertical integration has given it substantial control of the necessary material resources as well as costs, quality and delivery times. Moreover, it has made the company a desirable partner and this, in turn, has enabled it to access promising low-cost technologies that it has not developed itself. For instance, REC has partnered with German company CSG Solar, which uses a proprietary silane-on-glass technique to produce thin-film modules, to help further its thin-film ambitions.
Another independent forging ahead without benefit of electronics parentage is PV Crystalox Solar PLC, which began life in Great Britain 27 years ago as a manufacturer of ingot furnaces. From 1990, the emergence of a solar PV market provided an opportunity to diversify into polysilicon ingot production. Like REC, this company has shown that a sound business strategy, scale and a strong focus on process improvement are not the sole prerogatives of companies sired by electronics. Today it claims to be world's largest “pure play” producer of solar-grade polysilicon and, with an assured supply of the material available at favourable rates, it has extended its operations downstream into wafer and solar cell production.
The latest hike in production scale will come this year as a new polysilicon factory comes on stream in Germany. Complementing this will be a new facility dedicated to producing precursor chlorosilane gas. This is being commissioned on the same site by Evonik-Degusa, with whom Crystalox has a 10-year contract to supply the gas. This further fortification of supply arrangements should help avoid the periodic gas shortages that have been a feature of the existing supply chain.
Some of the Asia-Pacific PV start-ups are similarly rooted outside electronics. One of the largest, China's Suntech Power Holdings, founded by PV scientist Dr Zhengroung Shi in 2001, is now among the world's top five manufacturers of PV cells and has ambitions to be the lowest cost per watt producer of PV solutions. Given China's technology, its rapidly expanding presence in solar PV, and its relatively low-wage economy and entrepreneurial drive, it could succeed.
Others who have started with a “clean sheet” include energy companies and utilities, which can use large financial resources to buy into the business and acquire the necessary technologies. BP Solar and Shell Solar exemplify this. Some, like Timmenco in Canada, have light metals as their founding discipline, or other materials, as with Dow Corning.
Clearly, an electronics background, though having been a necessary foundation for solar PV initially, is no longer essential. At the present stage of solar PV evolution, enterprises without it can do just as well as their electronics-based counterparts.
Independence encourages creativity, new ways of thinking, new supply relationships, and ‘home-grown’ skills. These skills can be precisely tailored to meet current needs and can also relate without heritage ‘baggage’ to entirely different end-user needs. Today it seems that players can exploit the synergies but avoid the dependence.