Bright Future for Thin-Film Solar Cells

The photovoltaic industry is in the middle of one of the biggest booms in its history. Thanks to a combination of high fossil fuel prices and large government incentives for renewable energy, solar installations grew 34% in 2005 and have grown by more than 20% per year for several years now. Yet even such rapid growth isn't enough to meet rocketing demand. Bob MacDonald, director of product marketing and field operations at SolFocus (Palo Alto, Calif.), estimates that demand growth currently exceeds 70% per year.

One of the most important bottlenecks limiting solar cell growth is the global supply of polysilicon. Most of the photovoltaic market — more than 90% — uses silicon wafers as a starting material. Electricity is produced when incoming photons excite free carriers at junctions within the silicon. These solar-grade wafers are made with less demanding specifications than IC-grade wafers; solar-grade wafers can tolerate higher levels of contamination and can be either monocrystalline or multi-crystalline. Still, both IC-grade and solar-grade wafers depend on the same polysilicon raw material.

Until recently, polysilicon suppliers have focused on the much larger IC market, which offers a substantial price premium for material. Solar cell manufacturers have historically purchased IC manufacturing's “leftovers,” using end sections, defective ingots, and other scrap.

Photovoltaic concentrators, like this design from SolFocus, concentrate the sun’s rays up to 500×, dramatically reducing their consumption of semiconductor material. (Source: SolFocus)

The silicon itself is by far the largest contributor to the cost of conventional solar panels: Bulk silicon cost ~$40/kg in 2005, and a typical 1 m square panel might use >2.5 kg of the material to produce ~200 W of power. Since the key parameter for solar cell suppliers and customers is cost per watt, one of the best routes to improved value is to extract more electricity from a smaller quantity of semiconductor.

As in IC wafer manufacturing, solar cell manufacturing loses a significant amount of silicon during the cutting and shaping of the wafer. Reducing this waste cuts manufacturing costs while requiring very little change to the cell design. That's the idea behind Evergreen Solar 's (Marlboro, Mass.) “string ribbon” technique. Two threads draw a thin sheet of silicon out of the melt. The sheet can be used in further manufacturing directly, without the sawing and shaping required for conventional silicon ingots.

Other approaches use modified solar cell designs to reduce the amount of silicon required. Simply using thinner wafers or depositing silicon on glass puts cost and efficiency goals in conflict; however, a thin layer allows more of the incident light to simply pass through without being absorbed and without generating electricity. A porous silicon layer between the two serves as a reflector, capturing light that might otherwise be lost. The porous layer may also help protect the epitaxial layer from contaminants in the underlying metallurgical-grade wafer. However, Jef Poortmans, program director for photovoltaics at IMEC (Leuven, Belgium), explained that more research is needed to see if the epitaxial layer stays clean enough.

Yet, once a solar cell design abandons bulk silicon in favor of thin-film layers, the arguments for using silicon at all start to break down. Current supply constraints notwithstanding, bulk silicon is readily available and relatively inexpensive, especially compared with other semiconductors, although it is not a particularly good generator of photo current. Silicon also absorbs light over only a relatively small fraction of the solar spectrum; its weakest absorption is in the infrared, which is the brightest part of the sun's output.

Multi-junction solar cells exploit more of the spectrum by using materials with different absorption ranges. For example, SolFocus uses a germanium substrate for infrared absorption, with additional GaAs and InP layers for absorption of blue and UV light, used in space applications.

Unfortunately, multi-junction cells are also twice as expensive as silicon. Fossil fuel electricity is not available in space; cost requirements for terrestrial applications are far more demanding. It's difficult to shrink the size of the solar panel itself — the sun's output is ~1000 W/m² — but optical concentrators can reduce the amount of semiconductor material needed for a given panel by using lenses to focus the sun's rays. SolFocus's multi-junction cells absorb about twice as much of the solar spectrum as silicon, and achieves a concentration factor of ~500. They use ~1000× less semiconductor material for a given amount of power than an equivalent silicon panel.

Find more information on emerging technologies.