Harvesting the Sun: Renewable Power Generation from Photovoltaic Solar Cells
Shortages of Silicon Wafers Have Led to Temporary Growing Pains
Moslehi, PhD, is chief technology officer and senior vice president,
semiconductor technology research, for The Noblemen Group, a boutique
investment banking, strategic advisory, and business development firm.
Moslehi has 20 years' experience working in the semiconductor and semiconductor
equipment industries. He can be reached at email@example.com.
In recent years, global energy demand has been steadily increasing, fueled in part by tight supplies and the rapid growth of China, India, and other emerging economies. Energy prices have soared, with the price of oil skyrocketing to around $60 per barrel, some three times more expensive than a barrel cost in early 2002. This trend, as well as the need for environmentally friendly and reliable energy sources, has promoted a strong interest in wind, solar, biomass, and other emerging renewable sources.
The sun provides more energy to the earth in one hour than the annual energy consumption of the entire world combined, yet solar power supplies <0.01% of the globe’s total energy needs. However, this picture is changing rapidly. Spurred in part by government subsidies and tax incentives in Germany, Japan, the United States, and other countries, the solar cell market has grown between 30 and 40% each of the past five years, reaching an estimated market size of more than $5.3 billion in 2005. More than 1600 MW of new solar cell capacity was added worldwide last year, a 70%+ boost from 2004. According to the European Renewable Energy Council, renewable energy could provide the sources for more than half of the estimated world electrical power needs by 2030, with photovoltaic power generation accounting for >8%.
In 2002, nearly 90% of solar electricity demand was provided by crystalline silicon; of that, single-crystal supplied 33% and polycrystalline 55%. Amorphous (noncrystalline) silicon, which is used in Japan for calculators, watches, and other consumer products, accounted for an additional 5%. But only a small portion of incident solar energy gets converted into electricity. The typical conversion efficiency of commercial single-crystal cells is 15–17%, with a theoretical maximum of ~28% and laboratory-demonstrated efficiency of ~24%. Typical conversion efficiency for polycrystalline is 13–15%, while amorphous silicon ranges from 5 to 8%.
The emerging thin-film technologies—based on the deposition of very thin films or layers of special photovoltaic materials on glass, plastics, and other low-cost substrates—could offer much-lower-cost alternatives to silicon-based cells. A large group of materials are being considered for thin-film solar cells, including cadmium telluride (which holds the largest market share), cadmium sulphide, and copper indium diselenide. A fourth, compound semiconductor–based approach uses single-crystal gallium arsenide and its alloys, such as gallium–indium phosphide. Other technologies being pursued include low-cost plastic solar cells.
For some time, Germany and Japan have had government programs encouraging the widespread adoption of solar power. Germany pays panel owners for excess solar power generated. Spain, Italy, and other European countries have established similar plans. Japan has the highest per-capita installed capacity of solar power generation. In the United States, in addition to newly enacted federal tax credits, many states offer tax credits for installation of solar panels. Over the next decade, the U.S. government plans to increase the installed solar power capacity to as much as 10,000 MW (versus the existing 175 MW), and California plans to offer rebates of nearly $2.9 billion for 3000 MW of solar panels.
The solar cell and panel market is dominated by the Japanese, which control about half the sector, followed by U.S. and European companies. In 2004, Germany, Japan, and the United States accounted for more than 80% of the total worldwide solar cell installations.
The estimated capital investment in the United States for crystalline-silicon solar cell panels is in the range of $3.7–$5.3 per watt, plus installation costs of ~$3.5 per watt, which adds up to $7–$9 per watt, way down from the nearly $60-per-watt numbers from 30 years ago. With government incentives, this cost could be lowered to $3–$4 per watt, or about 10–15 cents per kilowatt-hour. This is quite competitive with local electricity rates in California and many other markets, particularly for households with at least $100 of monthly electricity costs.
By 2010, solar electricity is expected to become cost-competitive in most U.S. regions. The long-sought target for economic viability is $1–$2 per watt (or 3–7 cents per kilowatt-hour), which should be achieved over the next decade. This will be made possible with the historical annual average price reductions of 4%, continued conversion-efficiency improvements, and the economies-of-scale impact of large-scale manufacturing. In parts of the world with very high energy costs or in remote areas, solar electricity has become cost-effective without government support and incentive programs.
Another critical metric is the energy-payback period. It refers to the amount of time the solar cell panel takes to generate the same amount of energy that was spent fabricating it. Beyond the energy-payback-period point, the panels will be energy-positive, generating superclean energy for the rest of their service life (typically 25–30 years). For crystalline silicon, the energy-payback period has shrunk to about four years; with thinner wafers, the time is projected to drop to two years. Thin-film panels could reduce the payback period to less than a year.
Polysilicon remains the raw material for single-crystal silicon wafers, which are used for fabricating both crystalline solar cells and microchips. Solar cell silicon-wafer consumption now exceeds 30% of the total supply, with the rest used by the chip industry. Since the near-term annual growth rate is projected at 30% over the next two years, the amount of silicon consumed by solar cells could exceed that of the chipmaking industry, with some forecasts showing the solar cell industry as the largest poly consumer by 2008.
Solar cells’ rapid growth rate, coupled with the semiconductor sector’s healthy appetite, has resulted in severe poly shortages. In less than two years, the contract price of polysilicon has increased from around $30/kg to more than $60/kg; in 2006, the price could exceed $80/kg, with even higher spot prices. This situation has led to significant jumps in the price of wafers, which account for 40–45% of the total fabrication cost of solar cell panels. (Glass, frame, interconnection, and other module expenses account for another 30–40%, while diffusion, metallization, process materials, and other fab and processing costs constitute the remaining 20–25% of the panel’s total costs.)
These price hikes are welcome news for materials suppliers, which have long suffered from pricing pressures and very low margins. On the other hand, solar cell suppliers have been hit by price increases and a significant slowdown in their growth rate, projected to be around only 5% in 2006. IC manufacturers are also expected to be adversely affected by higher wafer prices and potentially slower growth, particularly those manufacturers that lack long-term wafer supply contracts.
The phenomenal growth of solar photovoltaics also presents new opportunities for equipment and materials suppliers. However, the solar cell industry is primarily based on 150-mm and smaller wafer sizes. It also has significantly fewer processing steps and a much shorter fabrication cycle time than typical chip manufacturing. Older wafer fabs are usually well suited for cost-sensitive solar cell production.
In the near term, Hemlock, MEMC, Wacker, Tokuyama, Mitsubishi, Sumitomo, and other polysilicon suppliers cannot meet the increasing demand. They have been reluctantly examining the investment risks associated with adding new capacity and have decided to do so only after securing long-term contracts from potential customers. Some industry reports indicate that the next two to three years of poly supplies have been sold out. However, the capital-intensive poly manufacturing operations require a few-hundred-million-dollar investment for a typical factory, which then takes two to three years to build. Therefore, over the next two years, polysilicon supplies are not likely to keep up with demand. The projected shortfall is estimated to grow from about 6000 metric tons in 2006 (out of the forecast supply of ~29,000 tons) to about 20,000 tons in 2008.
Many solar cell suppliers plan to further thin down the wafers to less than the current range of 200 µm, thereby reducing silicon consumption. When the planned new poly capacity comes on line in 2008, the rapid growth of solar cells should resume for the foreseeable future. But in the intervening few years, thin film and other emerging technologies for direct generation of solar electricity will have a golden opportunity to compete, a prospect further enhanced by increasing private and public investments.
The march toward the increased adoption of solar-based renewable forms of energy—including photovoltaics, solar thermal, wind, biomass, and hydro—is moving into high gear. This will have major economic implications, such as the creation of new jobs, a cleaner environment, and growing markets for the semiconductor manufacturing sector.
For more on the emerging polysilicon shortage and solar power, see the Lead News story in MICRO’s March 2006 edition.
Search | Current Issue | MicroArchives
Buyers Guide | Media Kit
Questions/comments about MICRO Magazine? E-mail us at firstname.lastname@example.org.
© 2007 Tom Cheyney
All rights reserved.