by Katherine Derbyshire, Contributing Editor, Solid State Technology
The IC industry owes much of its success to Moore’s Law, which holds that every 18-24 months the number of transistors or memory elements on a chip doubles, while the price of each chip stays the same. Steadily increasing capabilities and steadily shrinking costs have helped ICs spread from computers to cell phones, digital cameras, and singing holiday cards. Moore’s Law powers the Information Age.
Flat-panel displays (FPD) and solar panels free ICs from the tyranny of the cathode ray tube (or teletype) and the power grid, respectively, yet neither benefits from Moore’s Law. FPDs are constrained by the human eye. For comfortable viewing, bigger is better. FPDs made video portable, but they also made it wall-sized, with screens big enough to dominate almost any room. Solar cells, similarly, are constrained by the sun. Conversion efficiency depends on band gap and optical properties, not feature size. Extracting more power requires a larger panel.
Solar cells and flat panels
At the same time, both displays and solar cells are highly cost-sensitive. For solar cells, the ultimate competitor is the power grid, with prices below US$0.10/kWh in the US. While displays are less commoditized than electricity, price is often the only differentiator in a labyrinth of similar products.
Offerings from Applied Materials, one of the few equipment companies serving all three markets, attempt to apply lessons from the display and IC sectors to the solar industry. Display manufacturing seems to be especially applicable.
For FPDs, cost reduction has come by operating on the largest possible scale. To make the arrays of thin film transistors that control an LCD image, fabs deposit silicon on enormous sheets of glass (>2m/side). The cost of transporting such large sheets is mitigated by having glassmaking facilities on-site, linked directly to the fab in a single manufacturing chain.
Similarly, Applied Materials’ SunFab manufacturing line moves 5.7m2 glass sheets between vacuum process stations along automated tracks. According to the company, these sheets are four times bigger than the largest production solar panels, giving more than 20% cost savings. A single SunFab line can produce up to 75MW/year in solar generating capacity. (For comparison, researchers at Solarbuzz estimate that 1744MW of solar capacity were installed in 2007. Large “solar farm” projects typically generate 2-3MW of power.)
Figure 1: SunFab thin-film solar cell production line. (Image courtesy of Applied Materials.)
Figure 2: PECVD deposition system for thin-film solar cells. (Image courtesy of Applied Materials.)
At some point, though, moving such big sheets of glass becomes impractical. Large rolls of flexible material are much easier to handle, and a roll-to-roll batch process is more efficient than a process based on discrete substrates. Large-area devices such as displays and solar cells can be made and transported more efficiently if they can bend.
Efficiency in theory, though, does not necessarily mean more easily in practice. Silicon, the heart of most displays and most solar cells, requires deposition temperatures that are incompatible with most flexible plastics, while metal foils add weight and cost. Both solar cells and FPDs are complex, multilayered structures; aligning and sealing them becomes more difficult when flexible substrates are used. Though vacuum coating of flexible films is a well-established industrial process, flexible electronic devices are still in their infancy. Products like the SmartWeb roll-to-roll deposition system emphasize customization and process diversity.
Solar cells and silicon wafers
Though thin-film and flexible solar cells grab more headlines, bulk silicon remains the market leader. This advantage derives in part from bulk silicon’s higher efficiency, which in turn depends on material quality. Though solar cells are perceived as a less demanding application than ICs, the best solar cells (>20% conversion efficiencies) need high minority carrier lifetimes, explained Charlie Gay, VP and GM of Applied Materials’ solar business group. Their purity and defect specifications are more stringent than many IC applications. For these cells, cutting costs by using less expensive silicon is not an option.
Unfortunately for the photovoltaic industry, the price of silicon has rocketed as market demand has increased. Long-term contracts currently price the material in excess of $70/kg, while spot market prices can well exceed $300/kg. Manufacturers of bulk silicon cells have responded, in part, by trying to use less of the material. Solar cells are less than 200μm thick, compared to 750μm for IC grade wafers.
Though handling of such thin wafers is a challenge, Gay said a typical large plant might handle more than a ton of silicon/hour, while individual process tools might handle 2000-3000 wafers/hour. (A typical IC vacuum deposition system handles around 50 wafers/hour.) Wafer saws such as Applied Materials’ HCT wafering line process up to 15MW equivalents/year.
Without Moore’s Law to help, FPD and solar cell manufacturing approach cost reduction in similar ways. Large-scale manufacturing and heavy use of automation helped FPDs break out of their laptop computer niche. Solar manufacturers hope to use the same methods to compete with fossil fuels. — K.D.