by Katherine Derbyshire, Contributing Editor, Solid State Technology
On paper, silicon doesn’t look like a particularly good photovoltaic material. It’s expensive, its optical properties are poor, and its coverage of the solar spectrum is limited. Yet bulk silicon solar cells, based on single crystal or multicrystalline wafers, dominate the market with more than 90% of sales.
That dominance may be changing, though. Rapid growth in the photovoltaic market has contributed to rising polysilicon prices and serious supply constraints. These in turn have sparked considerable interest in thin-film solar cells. While bulk silicon cells tile the module area with silicon wafers, thin film cells coat large substrates as single units. Not only do they offer substantially lower costs/m2 of cell area, but thin-film cells can potentially conform to curved or flexible substrates, allowing many more cell installation options.
One of the most promising of the thin film technologies is copper indium gallium diselenide (CIGS). The name encompasses a wide range of miscible quaternary and ternary alloys, with indium, gallium, and occasionally aluminum freely substituting for each other, and sulfur sometimes used instead of or in addition to selenium. The electrical properties depend on the exact composition, giving CIGS materials a bandgap anywhere from 1.0 to 3.5 eV.1
The flexible bandgap is helpful for specialized installations, such as for indoor use, where the available spectrum can differ from that of normal sunlight. CIGS compounds can also create a stack of tandem cells, covering a wider fraction of the optical spectrum than a single junction can.
Unfortunately, while properties change dramatically with composition, the miscibility of CIGS alloys makes film composition very difficult to control. The most common manufacturing methods evaporate or sputter copper, indium, and gallium sequentially or simultaneously, and reacting the resulting film with selenium vapor establishes the final film composition. Simultaneous deposition of the three elements results in a composition that varies with location in the process chamber; with sequential deposition, segregation and preferred Ga-Se reactions can lead to composition nonuniformity. The composition of the finished film depends on the thermal profile and diffusion from the substrate (usually Mo-coated glass), as well as on the initial deposition. Indeed, composition uniformity over large areas has so far been a major obstacle to CIGS commercialization.
Vapor deposition of CIGS thin films is undesirable for cost reasons as well. Vacuum chambers and multisource sputtering guns are expensive, especially for the large process areas desired for solar panels. Sputtering and evaporation have relatively poor materials utilization, coating the entire process chamber as well as the target substrate. Not only do wasted materials add cost, but CIGS films may face materials supply constraints once they achieve significant sales. Indium is already an important component of the transparent conducting oxide used in many flat-panel displays and photovoltaic cells. Research at the National Renewable Energy Laboratory estimates that indium availability is likely to constrain manufacturing once CIGS production reaches the tens of gigawatts level.2 (Note that this is a long-term consideration; the current photovoltaic market produces <3GW/year, considering all cell technologies.)
An ideal CIGS deposition method would serve all three objectives: improved materials utilization, improved uniformity, and reduced process costs. One possible approach, being pursued by International Solar Electric Technology (Chatsworth, CA), Nanosolar (Palo Alto, CA), and others, uses printing processes to apply suspensions of metal oxide particles. In particle suspensions, the viscosity and other flow properties depend on particle size, particle concentration, and the suspension medium used. By varying these factors, manufacturers can adapt their “ink” to a wide variety of printing methods, from screen printing to ink jet deposition. Printing methods in general are adaptable to a wide variety of substrates, including metal foils, glass, and plastic. They achieve materials utilization rates in excess of 90%, while using far less expensive equipment than needed for vacuum processing.
Particle deposition leaves behind a porous film, which a thermal step densifies and reduces from oxide to metal. The high surface area of the initial powder layer allows the reduction reaction to proceed very quickly, much faster than either a bulk or a thin-film reaction could. Though the final composition still depends in part on the film’s thermal history, controlling the quantities of precursor powders gives manufacturers uniform initial compositions and nearly unlimited access to the full CIGS alloy system.
So far, non-vacuum CIGS processes have achieved very promising results. Nanosolar has demonstrated a cell efficiency of 14% 3, comparable to the 18.8% efficiency achieved by the best vapor-deposited CIGS on glass cells, and superior to the 9.8% achieved by commercial amorphous silicon cells.4 Still, other CIGS commercialization efforts have run aground, or at least seen substantial delays, due to uniformity issues. Until someone actually produces high-efficiency CIGS cells in commercial quantities, the technology will remain one of many intriguing possibilities. — K.D.
References
[1] Joseph D. Beach and Brian E. McCandless, “Materials Challenges for CdTe and CuInSe2 Photovoltaics,” MRS Bulletin, vol. 32, pp225-229 (March, 2007).
[2] Rommel Noufi and Ken Zweibel, “High-efficiency CdTe and CIGS Thin-film Solar Cells: Highlights and Challenges,” Proc. 4th World Conf. On Photovoltaic Energy Conversion, vol. 1, pp 317 – 320 (May, 2006).
[3] J. K. J. van Duren, et al., “The Next Generation of Thin-film Photovoltaics,” Mater. Res. Soc. Symp. Proc., vol. 1012, paper 1012-Y05-03 (2007).
[4] M. A. Green, et. al., “Solar cell efficiency tables (version 30),” Prog. in Photovoltaics: Research and Applications, vol.15(5), pp 425-430 (2007)