Solar Solutions

By Barbara Maynard

Though solar cells have been around for more than 50 years, researchers are still hard at work trying to improve their efficiency and drive down the costs of making them.

When Bell Labs unveiled the first solar cell powerful enough to run everyday electronics, The New York Times’ front page declared that this development “may mark the beginning of a new era, leading eventually to the realization of one of mankind’s most cherished dreams—the harnessing of the almost limitless energy of the sun for the uses of civilization.”

That was in 1954. Half a century later, solar technology has dropped dramatically in price—from upwards of $100 per watt (W) generated in the 1970s to $3–5 per W today—but it is still considered too expensive to compete with established energy sources. Now, fresh interest in renewable energy is pushing solar from the niche market to a mainstream energy supply. This demand, prompted in part by rising fossil-fuel costs but also from many state government mandates, coincides with exciting technological advances that promise to make energy from the sun more affordable and more versatile than ever.

“The time is finally right for solar energy to make a real contribution to our energy supplies,” says Lawrence Kazmerski, director of the U.S. National Renewable Energy Laboratory’s Center for Photovoltaics. One piece of evidence supporting Kazmerski is the nearly $1 billion that solar-energy companies raised in the financial markets in 2006.

When researchers talk about solar energy, they’re usually referring to one of two ways of actively capturing and using the sun’s energy: photovoltaic cells that convert light directly into electricity, and solar thermal devices that convert light into heat, often to power an electricity-generating turbine. A third strategy, passive solar building design, is another highly effective way of taking advantage of the sun’s energy, with little or no additional equipment required.

Photovoltaics have been synonymous with crystalline silicon since Bell Labs’ first model. More than 90% of the solar cells sold today use crystalline silicon semiconductors, and for good reason. The raw material is abundant, durable, and environmentally benign, and the basic technology is well understood. Crystalline silicon cells are also the most efficient in mass production, converting up to 22% of incoming light to electricity.

Silicon has its limitations, however. Although the raw material is abundant and cheap, producing the high-purity crystalline silicon used by both the solar and computer-chip industries is expensive. The recent boom in demand for solar has outstripped supply and driven up the price of crystalline silicon from about $9 to as much as $100 per kilogram. New billion-dollar fabrication plants now being built should ease this shortage, but the first of these new plants won’t come online until later this year.

Some solar-cell manufacturers are trying to compensate for short supplies by using thinner crystalline silicon wafers in their solar cells. However, crystalline silicon also has a low absorptivity, which means photons can pass through a fair bit of material before finally being absorbed. Therefore, the silicon wafer must be relatively thick, approximately 100–300 micrometers.

Thinking Thin

Thick, high-purity semiconductor materials make for expensive, heavy, and rigid solar cells. To get around these limitations, researchers are exploring the use of thin films of light-sensitive materials. Not only does a thin film use less raw material, but the manufacturing process itself is also less expensive. Because thin films can be sprayed onto just about any substrate, including glass, stainless steel, and plastics, they are also versatile. Indeed, flexible and light thin-film solar cells are used today in a variety of products, including roofing shingles and iPod-charging backpacks.

Amorphous silicon is the best-established thin-film product. It is potentially less expensive to make than crystalline silicon and has a higher absorptivity, which means the cells can be up to 100 times thinner. However, the efficiency of amorphous silicon cells is lower than crystalline silicon, at about 18%.

Thin-film cells based on copper indium diselenide, or CIS, have many of the same advantages as amorphous silicon. With efficiencies approaching 20%, they are still less efficient than crystalline silicon at turning sunlight into energy, but researchers working with CIS are slowly closing this gap.

Cadmium telluride (CdTe) is also used in thin-film solar cells. Like the other thin-film semiconductors, CdTe allows the use of 100 times less material than crystalline silicon, and it can be 100 times less pure. Several companies commercially produce thin-film CdTe cells on a glass substrate, and now one investigator has developed a process to mass-produce the cells much less expensively.

“We have streamlined the manufacturing process,” says Walajabad Sampath, a professor of mechanical engineering at Colorado State University. “It is ultimately far more efficient in terms of capital and materials and labor.” As with other thin-film materials, the lower production costs can balance out the lower efficiency of CdTe cells.

“The modules will run about 10% efficient,” Sampath says. “Garden-variety silicon panels are about 12–13%, so there is about a 10–20% penalty in terms of area.” But with projected manufacturing costs running three to four times lower, Sampath’s cells should be competitive against other solar cells and fossil fuels. Future gains in CdTe efficiency, which theoretically could be better than that of silicon, could drive the costs even lower.

Cost is especially pressing for one application Sampath has in mind. He has developed a system of light-emitting diodes (LEDs) and a battery that could be powered by his solar cells in developing countries that don’t have widespread access to electricity. “A small 10-W panel running two 3-W LEDs can provide such a better quality of life than a kerosene lamp,” he says. “And the cost of this whole system is around $50. That’s about what these people spend on kerosene in about a year and a half to two years.”

Therefore, a solar cell’s maximum efficiency is limited by its bandgap. A silicon solar cell, for example, can be no more than about 26% efficient. Manufacturers have almost reached that limit with today’s most efficient silicon cells, which push 22%.

The mention of cadmium can raise health and environmental safety issues, but Sampath says two factors alleviate that concern. “It’s not cadmium, it’s cadmium telluride, and the comparison is like the difference between chlorine and sodium chloride,” he says. “It’s much more strongly bonded than sodium chloride is.” Manufacturers also offer free recycling for cadmium solar panels at the end of their lives.

Building solar cells on glass discounts one of the advantages touted for thin films—the ability to use flexible, lightweight substrates. However, for rooftop applications, glass is difficult to beat. Because the substrate side of the panel faces the sun, it must be transparent, even after 20 or more years of exposure to UV light in the outdoors. “Glass may ultimately be inevitable,” Sampath says. “There isn’t a polymer out there that can fulfill all of these requirements.”

Sampath expects to ramp up production and testing of his design later this year. “Our complete machine will be ready by sometime around September 2007. We will be making 10-W panels at the rate of one every two minutes,” he says.

Multijunction Solar Cells

Different semiconductor materials absorb different wavelengths of light—one of the advantages of CdTe is that it is well matched to the solar spectrum. This tuning is the result of the semiconductor’s bandgap—the energy required to bump an electron from the valence band to the conductance band, where it is available for collection into current. For a semiconductor to convert a photon’s energy into electricity, the photon must carry at least the amount of energy prescribed by the semiconductor’s bandgap. Anything less, and the photon passes through unabsorbed. Anything more, and the excess energy is lost as heat—only one electron is liberated per photon (but read the section about quantum dots for an exception to that rule).

However, if one stacks different semiconductors together, each with a distinct bandgap, then photons not absorbed by one material can be caught by the next. For example, today’s most efficient solar cells use three layers: gallium indium phosphide, gallium arsenide, and germanium. Because this triple-junction solar cell is based on elements from the III and V columns of the periodic table, it is called a III-V multijunction cell.

Such III-V multijunction cells have been powering the Mars rovers Spirit and Opportunity since they landed in early 2004, but they are just beginning to break into commercial use for terrestrial applications. Spectrolab created a stir this past December when it announced it had created a record-breaking 40.7% efficient III-V cell. Dave Garlick of Boeing Space and Intelligence Systems, which owns Spectrolab, says that the new design will undergo a year of rigorous testing before the company begins licensing it to third-party manufacturers.

These III-V cells are not cheap to make, so they are typically used on concentrating photovoltaic (CPV) installations. CPV is essentially a system of mirrors that track the sun’s path across the sky and reflect concentrated light onto small solar cells, cutting the amount of expensive semiconductor materials required to produce a given amount of electricity. Spectrolab III-V multijunction cells have been in use for more than a year in CPV installations in Arizona and Australia.

The Future of Photovoltaics?

While silicon, CIS, and CdTe cells battle for market share and III-V cells make the leap from space to terrestrial applications, researchers are working on the next generation of photovoltaic semiconductors. Everything from dye-sensitized solar cells to organics to quantum dots to nanowires is under study.

Organic solar cells could be very inexpensive to make, with bandgaps tailored to the solar spectrum. Stephen Forrest, vice president for research at University of Michigan, heads a team that has made progress against one of the limitations to organics—capturing longer wavelengths of light.

“Organics don’t cover a large fraction of the solar spectrum, particularly the infrared, so there can be low efficiency out at the longer wavelengths of solar radiation where there is actually quite a bit of energy,” Forrest says. “Absorbing in the infrared is necessary to harvest as much of the solar spectrum as possible.”

His group has used tin phthalocyanine in combination with buckyballs and copper phthalocyanine to push the absorption spectrum out to the near-infrared. The resulting cell was only about 1% efficient, which is lower than other organic cells, but the breakthrough demonstrated that organic cells can reach into the longer wavelengths.

Although organic solar cells are still a long way from moving from the laboratory to practical use, Forrest sees a bright future. “We’ve done a lot of theoretical work on them, and we don’t see an obvious limit to the efficiency that one could have,” he says. “I think the potential is very large. Right now the cells are pushing 6%, and it is reasonable to believe we’ll be at 10% before 2008 or 2009.” At that point, he adds, “we’ll be working much more on problems of reliability, packaging, cost-effectiveness, and deposition issues.”

Like organics, quantum dots could provide inexpensive semiconductors designed to match the solar spectrum. “Quantum dots are unique,” says Michael LoCascio, the chief technology officer for Evident Technologies. “They are nanoscale semiconductor particles, and their bandgap is defined by not only their composition but also their actual size. It is possible to tune the bandgap.”

Even more exciting than a tunable bandgap is a phenomenon discovered at Los Alamos National Laboratory called multiple exciton generation, or MEG. It turns out that quantum dots, unlike conventional semiconductors, can use the extra energy carried by a photon that is two or more times the bandgap of the dot.

“Effectively, what you are doing is for each incident photon above a certain energy, you create more than one electron-hole pair, which then can be turned into [electrical] current, and you get another mechanism by which you can eliminate the efficiency losses due to thermalization of the charge carriers,” LoCascio says.

Like organic solar cells, much remains to be done to harness the potential of quantum dots in practical solar cells. “There are a number of challenges left,” LoCascio says. “One of the challenges we’re approaching, and I can’t disclose how we are approaching this, is getting the photogenerated charge carriers out of the quantum dots and then into the surrounding semiconductor polymer and onto the electrodes. Energy transport out of the dot is the breakthrough technology that will make this viable.”

Packaging Matters

Creating an efficient solar cell is only the first step in building an economical photovoltaic system. How the cells are linked together and housed affect how much power is actually generated and the system’s overall efficiency. Therefore, advances in building solar modules, each of which contains a number of connected solar cells, continue apace with the cells themselves.

For example, DuPont makes a variety of products for photovoltaic modules, including a fluoropolymer backing called Tedlar. This backing protects the cell from the damaging effects of ultraviolet light, temperature changes, and moisture, while also providing electrical insulation. DuPont recently improved the reflectance of Tedlar, which helps boost a module’s efficiency.

“There is always some space in between the cells in the modules, so there is some light that falls directly onto the backsheet,” says Hilde Roekens of DuPont’s Fluoropolymer Solutions section. “The higher the reflectance by the backsheet, the more chance you have that light is being reflected back onto the cells.”

DuPont also tweaked Tedlar so that it is less likely to wrinkle during vacuum lamination, which will help keep manufacturing costs down. For further improvements, DuPont is looking at ways Tedlar can boost efficiency by keeping cells cool. “We’re also working on programs to enable the cells to keep lower temperatures,” Roekens says. “With crystalline silicon technology, the cell temperature goes up when the module is exposed to sunlight. The higher the temperature goes, the lower the efficiencies of the cells are going to be.”

Concentrating Solar Power

Whereas photovoltaics directly convert light energy into electricity, solar thermal technologies convert light into heat, which can then be used to generate electricity. Like CPV, concentrating solar power (CSP) systems also use mirrors to focus light onto a small receiver. In today’s CSP systems, the mirrors are usually long rows of reflective parabolic troughs that focus sunlight onto a tube-shaped receiver filled with oil. As the oil heats up, it is pumped to a conventional steam turbine that uses the heat to generate electricity.

For two decades, CSP has been quietly generating electricity in the Mojave Desert. Nine solar electric generating stations range in size from 14 to 80 megawatts (MW), providing a joint capacity of more than 350 MW; a typical coal-powered plant produces 500 MW of electricity. In March 2007, a new CSP plant will go online outside Las Vegas. Called Nevada Solar One, the plant will have a capacity of 64 MW. CSP is a good fit for installations of this size and larger.

“Concentrating solar power will always win out over photovoltaics for a large-scale system,” says Mark Mehos, program manager for the CSP program at the National Renewable Energy Laboratory. “CSP works at large scales. Photovoltaics work very well at small scales. They really don’t overlap that much.” Mehos expects concentrating solar power to be cost-competitive with conventional natural-gas plants within 5–10 years and with coal-fired plants in the longer term.

One big advantage of CSP is that energy can be stored. In the current design, excess heat not needed immediately by the steam turbine is transferred from the oil to molten salt via a heat exchanger. Thermal storage is 99% efficient, Mehos says, but to further reduce costs, researchers are working on ways to use the same fluid for storage as for the heat transfer fluid in the field, eliminating the need for costly heat exchangers and reducing the volume required for storage.

CSP could also provide cheaper electricity if the systems could operate at higher temperatures. “When I say higher temperatures, I mean 550 ºC compared to the roughly 400 ºC that these systems operate at now,” Mehos says. Raising the operating temperature will require advanced receiver materials that can survive and effectively make use of the elevated temperatures without re-radiating the extra energy back to the atmosphere.

Solar Hot Water . . . and More

Solar hot-water heaters are a more familiar form of solar thermal technology, and they are still the most economical way, other than passive solar design, for individuals to put solar energy to use in their homes. With domestic hot water accounting for 13% of residential energy use, the savings can be immediate.

A major limitation of some solar hot-water systems is the susceptibility of copper pipes to freezing. Replacing copper with commodity polymers, primarily polyethylene or polypropylene, could reduce both the freezing hazard and the cost, because copper has become quite expensive. Although the polymers solve the freezing problem, they introduce a new vulnerability: overheating and bursting. Researchers are studying the high temperature resistance of various polymers and designs to prevent overheating.

Jay Burch, a senior scientist at the National Renewable Energy Laboratory, says research on solar hot-water technology is focused on using existing, low-cost materials instead of trying to develop engineered polymers, for example. “We’re trying to make solar water heating an attractive proposition economically on its own two feet without subsidies,” he says. “If polymer systems had sufficient support for research and development, it does appear there are no showstoppers to keep us from getting systems down in the four to six cent per kilowatt-hour range.” Burch says two polymer-based solar hot-water systems will enter pre-production in 2007.

Solar hot-water systems can also be used together with photovoltaics. Such a combination makes use of some of the roughly 70% of light energy not converted to electricity by the semiconductors while also keeping them cool. Burch expects a photovoltaic-solar thermal system to be on the market in 2007.

A third development in solar hot-water technology could eliminate a major limitation of solar energy—the need to store energy for times when the sun is not shining. Burch says that thermochemical storage based on desiccants could provide not just diurnal storage for hot water overnight, but also annual storage to tuck away Btu’s generated during the summer for use during the winter. Such a system could meet all of a home’s year-round thermal needs, including heating, cooling, and hot water, according to Burch.

The future of solar energy is likely to be diverse, with different technologies suited to different situations. The current focus on diminishing fossil-fuel reserves and greenhouse-gas emissions might provide—finally—the push needed to realize the potential of inexpensive, green power from the sun.

Barbara Maynard lives in Missouri, a state with a good solar rating. Her last story for Chemistry was on efforts to replace trans fats in food.