Shedding light on solar cell technology

The remaining PV systems are, in one form or another, substantially flat. Flexible cells are included because they become as flat as the surface to which they are attached. They are known as flat-panel or flat-plate and do not depend on mirrors or lenses to concentrate the sun's rays. Most of the panel area is active for energy conversion (except for wiring or other structures that may block incoming light).

Flat-panel PV systems use more active material (unless the active layers are thinned enough to compensate for the larger area) than CPV systems. Therefore, flat-panel systems tend to incorporate less expensive, less efficient PV materials. They have the edge over CPV in many less sunny climates. Flat panels still work as designed in cloudy conditions, extracting power from diffuse sunlight. Concentrating optics in CPV systems require parallel rays. Otherwise, the small cell at the focal point receives only the light falling on that small surface area.

Click on image to enlarge.

But why is the semiconductor industry so interested in PV? The real action takes place in the active layers, where the photons are converted into electrical power. The PV community generally breaks it down into five major cell types, based primarily on commercial importance, with bulk silicon accounting for at least 90 percent of worldwide solar panel production:

» Wafer-based silicon (bulk)

» Thin film (TF)

» III-V compounds

» Dye-sensitized solar cells (DSSC)

» Organic

With the capability and maturity of wafer-based monocrystalline silicon, it might seem strange to see such a diverse set of materials and options. But driving production cost down is what makes these other options so attractive.

Silicon wafer-based devices achieve very high efficiencies--up to 28 percent. However, bulk-silicon solar cells need the thickest-absorbing layer of all the options. This is an indirect bandgap material, after all, and without using clever techniques to trap sunlight in the silicon, the optimum thickness is nearly 1 mm. Either way you slice it (more on that later) bulk-silicon PV requires a lot of feedstock material per panel area or per Watt of output power. Not long ago, people panicked when silicon PV production created a polysilicon supply shortage that may have even driven more investors into some competing technologies. Bulk-silicon technology requires wafering or slicing of silicon ingots, a process in which a lot of silicon is wasted.

Although having the solar cell and a substrate together in one piece is convenient, and thicker materials are easier to handle and assemble into modules, there is a clear need to reduce the total amount of silicon in the solar cell. Avoiding the wafering process has driven the development of Ribbon Growth Silicon (RGS), String Ribbon (STR) and other clever techniques for pulling thin sheets of monocrystalline silicon out of a melt. Silicon ribbon and foil manufacturing processes produce thin, low-volume sheets of silicon without a substrate for the active region. Ribbon silicon offers potential efficiency in the 20 percent range, somewhat lower but still competitive with conventional wafer-based cells. However, ribbon growth techniques are currently quite slow.

Click on image to enlarge.

Thin is in

Compared with bulk-silicon PVs, thin-film solar cells demand much less expensive manufacturing methods. They are hundreds of times thinner than bulk-silicon devices consuming much less material. Thin-film devices based on amorphous silicon (a-Si) and microcrystalline silicon (µc-Si) burst onto the scene in the past couple of years, though they're not new technologies. The idea for a-Si cells dates back to the mid-1970s and the work of RCA Labs scientists David Carlson and Christopher Wronski. Detractors like to talk about a degradation mechanism in a-Si called the Staebler-Wronski effect, in which defects appear in the material after exposure to light. These defects eventually reduce cell efficiency, but the effect has a limit. Efficiency drops by about a third after a few hundred hours of exposure to the sun. A cell that began life at 12 percent will degrade to 8 percent but no further.

But there's a bright side to this a-Si degradation mechanism: The Staebler-Wronski defects can be repaired by annealing. You probably wouldn't want to pull your solar panels off the roof every few months and bake them in the oven to recover optimum cell efficiency. Fortunately, the sun bakes out the defects for you. PV cells from a-Si are the only ones that perform better as the temperature rises. Every other semiconductor prefers to stay cool. That's like a sunbather avoiding summer--it's tough to get tan if you can't take the heat.

Raw material cost is a hot topic when it comes to silicon PV, so the thinner, the better. A-Si cells do not roll out of the factory with everything that was pumped into the deposition chamber. The process uses only 10 percent to 30 percent of the silane gas supplied.

Thin-film amorphous silicon solar cells have been with us for a long time. The first solar powered calculators used this type of cell. But things have really taken off in the last couple of years. This is mainly due to the work of the LCD panel industry which can now produce extremely large displays on glass 5.7 m², so-called 8.5-generation glass. A lot of the recent success of a-Si TFPV (TF is commonly used for thin film but you may also see TFSC for thin film solar cell) can be attributed to Applied Materials work since about 2005.

Other manufacturers like Kaneka of Japan and Uni-Solar (or United Solar Ovonic a subsidiary of Energy Conversion Devices) of the US have produced viable commercial products for many years, but Applied has put a great deal of effort into two key areas of thin film development.

First, they have built large-scale fabrication tools offering turn-key factories capable of producing hundreds of megawatts of panels each year. Second (and I won't debate the relative importance of this point), Applied Materials has been a vocal advocate of this technology. That combination launched a-Si TF into the mainstream for large-scale energy harvesting.

Thin-film devices began to show real promise in the early eighties when prototype cells broke through the 10 percent efficiency barrier. Considering the manufacturing flow to create a complete solar panel product, thin-film devices skip one step. Bulk-silicon panels are assembled from individual cells. Each individual wafer slice is fabricated separately, then assembled and connected together in a series string to make one module. But for TF processing, the cell material is coated directly onto the module-level substrate. Individual cell processing is done afterwards.

Just as supply scarcity was an issue for bulk wafer production, material sourcing might creep into non-silicon thin-film technologies as well. Selenium " one of the chalcogenide or periodic table group VI elements " was the first solid-state solar energy conversion material discovered. But we're not running low because widespread manufacturing of selenium solar cells started right then in 1883. Chalcogenide elements like tellurium used in the promising cadmium-telluride cells are quite rare. Depending on the growth of the PV market, tellurium supply could be a bottleneck. Leading cadmium-telluride manufacturer, First Solar claims it will not slow them down, and just in case, they have multi-year supply contracts.

Rarity is just one issue for chalcogenide PV. Although small demonstration cells have been reported with outstanding efficiencies in the range just shy of 20 percent, there are challenges to maintaining this over large modules that will be required for commercial use.

Process control remains a big challenge. You might wonder why people bother with these rare and difficult materials when silicon is abundant and can be made into integrated circuits with close to a billion transistors on wafers 300 mm in diameter. Then again, you may not because the answer once again is, "Cost." If easy means cheap (and it usually does), then it's hard to top cadmium-telluride (CdTe).

One common method of manufacturing cells is more like spreading peanut butter on toast than semiconductor manufacturing. This promising second-generation solar cell may have more to fear than pure technology challenges. In a field so driven by concerns for protecting the environment, the use of toxic cadmium in these cells is often mentioned (especially by the competition). Installation costs for CdTe panels need to factor in the de-commissioning and toxic disposal.

First Solar plans to reclaim the panels they sell which serves the dual purpose of recycling the rare materials into new panels. The best research cell efficiency reported for CdTe is about 16 percent. They are sometimes referred to as "CDT" cells which is pretty confusing considering there is a polycrystalline silicon panel supplier called CDT Solar.

The other thin-film compound solar cell is also based on chalcogenide compounds. Widely known as CIGS cells, they are copper-indium-gallium-diselenide. It's the selenide that gives the chalcogenide tag to CIGS. Recipes for this type of cell tend to vary the amounts of the indium and gallium, so they sometimes appear as C(IG)S or more scientifically as Cu(InGa)Se₂. You could eliminate gallium altogether if you don't mind a slightly poorer alignment of the material's band gap to solar radiation. Without gallium, it's known as CIS. Thankfully, you need to get well outside the solar field to confuse this acronym which also stands for CMOS image sensor.

CIGS are considered to be the more promising of the two chalcogenide contenders. The best cell efficiency reported is 19.9 percent which is almost into bulk silicon territory. However, processing challenges remain for depositing CIGS cells onto large area panels. Full size CIGS panel efficiencies tend to drop to around 12 percent. There's a lot of research into improving this and protecting the intellectual property of the new techniques.

By the way, CIGS don't completely escape the spectre of deadly cadmium that plagues cadmium-telluride marketing. CIGS PV cells typically include a heterojunction with cadmium-sulfide (CdS). Fortunately, very little cadmium is required to create an operational device, but there is still a lot of research into alternatives to the CdS layer. Better for the green energy alternative to be built on green materials, I suppose.

Speaking of green, the last category of photovoltaics is based on organic materials. Organic and dye-sensitized cells might also be called photoelectrochemical solar cells " or PSC's " depending on who you talk to. The PV community generally sees them as separate top-level categories, but dye-sensitized cells are actually a subset of organics. Just in case it was getting too easy, these devices are also called Gratzell cells after one of the inventors.

Dye-sensitized cells are expected to be a very important third-generation technology. The organic dyes used to sensitize these cells to sunlight suggest we can learn a lot from nature's own solar cell " plants. Organic PV might one day reach the cost level of supermarket produce as well. The promise of these cells is that they will be so cheap and easy to manufacture that they can be integrated into many of the things we use every day right down to the clothes we wear.

New developments in photovoltaics, or at the very least the press announcements, are coming at an ever-increasing rate. The PV market is characterized by rapid growth and a diverse range of technologies with no clear winner on the horizon. Our future could be a time where cheap electricity fuelled by only the sun is taken completely for granted. Consider the lowly and ubiquitous solar calculator powered by amorphous silicon cells. It was certainly the first practical application of PV technology for the masses. It's hard to remember back to the time when I had to worry about replacing batteries in a calculator. Think about the utility of this simple device and imagine what other things might one day be powered essentially for free. Maybe it doesn't even take much imagination considering you can already buy a solar-powered Bluetooth headset for your phone. There is a lot at stake in photovoltaics, so it can be tempting to twist information to gain an advantage over competing technology. Hopefully, this brief introduction will make it easier to traverse this sometimes treacherous landscape.