Material Costs Determine c-Si vs. a-Si Solar Race
Lita Shon-Roy, Senior Managing Partner, Techcet Group LLC, Del Mar, Calif., www.techcet.com -- PV Society, 7/10/2009
The solar market hit a speed bump last year that has set back growth for 2009 by 10-15% for both crystalline silicon (c-Si) and amorphous silicon (a-Si) solar cells. The slowdown in the market has greatly eased polysilicon availability and pricing, making the price of c-Si solar cells rival that of a-Si cells. c-Si module prices are now $2.50-$3/Wp, and are expected to be 5-10% lower by the end of the year. In contrast, a-Si prices were reported to be $1.20-$2/Wp. Regardless of the difference in reported costs per watt between c-Si and a-Si modules, it is apparent that a-Si solar cells, while still hoped to be more cost-effective in the long term, will remain <20% of overall solar cell production through 2013.
The material costs associated with c-Si solar cell production are heavily weighted toward the cost of polysilicon; for 2008, ~78% of production material spending was attributed to polysilicon. As a result, the nagging question of how to reduce polysilicon costs has persisted over the past several years. Work being done on all fronts to reduce polysilicon costs will reduce its share of the pie to 67% by 2013 (Fig. 1). These cost reduction efforts include:
• Using upgraded metallurgical-grade (UMG) silicon
• Expanding existing polysilicon capacity, including new producers of polysilicon
• Reducing polysilicon production costs
• Increasing wafer cutting productivity
Solar-grade UMG silicon is a lower-purity polysilicon (6Ns to 7Ns) that is significantly less expensive than the standard-grade poly (11Ns), traditionally used for silicon wafer production. UMG silicon prices are reportedly less than half the cost of standard grade, and UMG silicon manufacturing costs are roughly one-fifth that of standard polysilicon.
UMG silicon cells operate at a lower efficiency than cells produced using standard-grade polysilicon. However, current reports show that UMG silicon cells can operate at up to 15% efficiency (according to reports from Advent Solar and Trina Solar), which is nearly equivalent to current c-Si solar cell efficiency. Most likely, UMG silicon cells will not be able to achieve as high an efficiency as standard polysilicon cells in production, but when silicon cost is factored in, UMG silicon may prove to be more favorable. Several solar fabs are already committed to using UMG silicon in hopes of benefiting from lower silicon costs. Companies that have announced their intention to use UMG silicon include Q-Cells, Trina Solar, Blue Square and CaliSolar, to name a few.
Polysilicon plant expansions and new plant build-outs have continued despite the economic downturn. Last May, Hemlock, the largest polysilicon manufacturer in the world, announced the completion of a new 8500 metric ton polysilicon plant in Michigan. This plant is the first phase of a $1B expansion. The second phase is scheduled to come on line in 2010, increasing the annual capacity to 36,000 metric tons. This expansion is in addition to a new plant being built in Tennessee, scheduled to be operational in 2012.
REC, another leading polysilicon manufacturer, completed its 400 metric ton Plant III expansion in April. The company's current plans call for a total company production output of 10,000-11,000 metric tons of polysilicon this year. This is in support of REC's long-term plans to grow capacity by 3× from 2008 to 2011.
New entrants in the polysilicon business are estimated to account for 4%, or 5.5 metric tons, of the world's polysilicon capacity. There are more than a dozen such companies, including Silicium Becancour, DC Chemical, Isofoton/Endesa, Hoku Scientific and M. Setek. The majority of these polysilicon companies see a strong future in c-Si solar production and continue to persist in their plans to build capacity. As world capacity increases, polysilicon pricing will decline at a steady pace - perhaps 5-10% per year for contract pricing. This will ultimately work in favor of lower cost per watt for c-Si solar cells.
Polysilicon manufacturing processes vary by company. However, there are new methods and technologies being developed to find ways of making less expensive materials. Metallurgical-grade silicon (MGS, 98-99% pure) is usually blended with high-purity (10Ns to 11Ns) polysilicon to make a solar-grade UMG silicon (>6Ns). Other techniques are in the works to produce higher-purity MGS so it need not be blended with higher-purity material to be suitable for solar cell processing. For example, Timminco, which sells polysilicon to Q-Cells, has developed a low-cost method for creating solar-grade silicon directly from MGS. Other companies developing similar processes include 6N Silicon, JFE Steel and Arise. Another example of a potentially lower-cost alternative is the vapor-to-liquid deposition (VLD) developed by Tokuyama, which offers a higher production rate than current production methods (such as fluidized bed reactor, or Siemens' reactor). The Tokuyama method is not yet fully commercialized, but a 200-ton polysilicon pilot plant is now running in Japan.
Wafer cutting processes have also been aimed at reducing silicon costs. This can be done by 1) cutting thinner wafers, and/or 2) losing less silicon during the cutting process. Means of cutting wafers as thin as 130 μm are starting to be used. Although breakage of wafers rises significantly for wafers &150 μm thick, handling and process improvements are being made to counteract this problem. Furthermore, equipment modifications are being developed to minimize the loss of polysilicon (kerf loss) from wire sawing wafers from ingots. The goal here is to maximize ingot yield and minimize cost per wafer. Another innovation is diamond-coated wire, which promises to cut thinner wafers with lower kerf loss.
The price of polysilicon has taken a wild ride from the calm of 2006, when spot market prices were $40-$60/kg, to the rough ride of 2008, when prices reached >$300/kg. As the market fell with the onset of worldwide recession, the price fell to ~$120/kg, just slightly above the average long-term contract price of $113/kg. It is expected that this price will not escalate again over the next few years, but decline.
These conditions create a market obstacle to any other cell technology challenging polysilicon's incumbent position. Hence, we expect c-Si cells to remain the leading solar cell technology for the foreseeable future. And what of thin-film a-Si solar cells? They will continue to grow (Fig. 2), but will be kept down to &20% of solar cell area produced (~12% of gigawatts produced) over the next five to seven years.
|
|
2. The amorphous silicon (a-Si) solar cell market will continue to grow, but will be held to <20% of total solar cell production. (Source: Solar Cell Equipment Consumables Markets, Techcet) |
