The trouble with silicon

Hotter than a steelworks: the furnace is running at more than 2,000 degrees Celsius as glittering electric arcs heat up the quartz sand. In terms of chemistry, quartz is an oxide of silicon. When extreme heat forces the oxygen out of a molecule – a process known as thermal reduction – industrial raw silicon remains. To produce these electric arcs, cables as thick as arms carry electricity to the furnaces. The smelter is heated with electricity because oil and gas cannot produce the necessary temperatures. For some time now, silicon manufacturers have been following in the footsteps of the aluminum industry, long regarded as an “electricity guzzler”. Silicon was formerly used mainly as a raw material for microchips, and hardly anyone spoke up about sustainable manufacturing. Now, the solar industry is constantly increasing its production and processing of silicon. Smelting silicon needs to become greener, and not only for the good of the industry’s image. Just to stay competitive, manufacturers are working feverishly on decreasing the use of energy and toxic chemicals.

A kernel of uranium

Power for the metallurgic smelting processes comes from large-scale power plants. In the U.S., for example, aluminum smelters use nuclear reactors and large hydropower plants in the Midwest. In Germany, silicon manufacturers turn to the combination of fossil and nuclear electricity offered by utility companies. A kernel of uranium can therefore be found in every solar panel. Green electricity doesn’t seem to be an alternative yet, for financial reasons and because of the large amounts of electricity required. “About thirty percent of the cost of silicon in Europe comes from electricity expenses,” says Robert Hartung, board spokesperson for Centrotherm Photovoltaics in Blaubeuren, Germany. Centrotherm is one of the largest factory outfitters for the manufacture of solar silicon and crystalline cells. “It’s similar to smelting aluminum. In the future, we believe that new silicon plants will be built where electricity is cheap.” In other words, silicon production and cell manufacturing will migrate to other countries. In addition, because of pricing pressure for silicon, temptation and opportunity are closely related. Hartung describes an example: “Asia Silicon is planning new capacities in a region where hydropower is cheaply available. This way, they have unbeatable costs and production is practically carbon-neutral.” Until now, Chinese manufacturers have made mostly negative headlines, such as when one company disposed of toxic silicon tetrachloride by simply dumping it in the countryside. Norway’s solar corporation REC also uses hydropower. The Scandinavians are currently investing more than a billion dollars to expand their solar capacity in Canada. There, too, hydropower is the main energy source.

Making the best use of materials

Besides electricity consumption, material yield is very important at each production stage. About eighty percent of solar silicon production is done with the Siemens process. The electric arc furnace smelts quartz sand into liquid silicon with a purity of 99 percent. A chemical process using hydrochloric acid, and trichlorosilane then further purifies the raw silicon. Trichlorosilane is extremely corrosive and harmful to the respiratory system. The required safety technology also makes it very expensive. For these reasons, engineers are looking for ways to increase the process chain’s effectiveness. The important factor here is the chemical reaction’s yield: about 82 percent of the raw silicon is absorbed into the process gas. That’s not bad, but it could be much better, especially since the gas is then distilled to remove impurities – aluminum, iron and copper. With the help of hydrogen, trichlorosilane is then dissolved on electrically heated rods of high-purity silicon at a temperature of 1,000 to 1,200 degrees Celsius. The silicon expands along the rods. “However, this process only utilizes 16 percent of the trichlorosilane,” Hartung calculates. Centrotherm therefore offers a converter to use up to 83 percent of the unused trichlorosilane.

Pinching pennies

Cost pressure is rising. Producers of wafers made from silicon blocks are also pinching pennies. One focus is silicon loss at the saws. “Although losses in the sawing of wafers have continuously dropped over the past few years,” Hartung says, “with a 140 to 160-micrometer-thick saw blade, and wafers of 160 to 200 micrometers, 40 percent of the silicon is still lost.” Regardless of falling silicon prices, these losses are a thorn in the side of manufacturers, who put a lot of energy into the process only to throw away almost half of the silicon. Manufacturers of systems therefore spare no expense or effort. For example, Applied Materials introduced its HCT MaxEdge wire saw a year ago. The company says it can decrease production costs for crystalline solar cells by 14 euro-cents per watt and will not only reduce losses, but also enable greater sawing speeds. Some manufacturers want to stop cutting silicon altogether and instead draw the correct shape directly from the liquid silicon. In edge-defined film-fed growth (EFG), the pure silicon liquid is drawn from an electrically heated graphite tub in the shape of octagonal polycrystalline tubes. The tubes grow about one millimeter per second up to six or seven meters. Each edge is ten to 12.5 centimeters long, and the tubes are 280 microns thick. A laser cuts the sides into silicon plates to produce wafers. This process uses 80 percent of the raw material. Wacker Schott Solar used the EFG process until September 2009, when Wacker and Schott terminated their joint venture and announced that they would no longer use the technology.
Evergreen Solar of the U.S. uses the string ribbon process, in which the wafers are pulled directly out of the melt between two wires. This process also leaves less waste than the conventional process using ingots and wire saws. Nevertheless, such alternative processes have yet to become standard in mass production, although the price war on the market for solar cells is forcing production to be more streamlined and organized in compliance with sustainability criteria. The example of Wacker Schott shows that the value chain will be even more starkly divided in the future. Specialized silicon producers supply cell manufacturers, who in turn deliver to module factories. Such conglomerates as Solarworld and REC survive only thanks to their size. The speed of the race is impressive. Schott Solar, for example, more than doubled its cell production in Alzenau within a year. “In 2008, we produced 130 megawatts of crystalline cells,” corporate spokesperson Lars Waldmann confirms. “In 2009, it was about 300 megawatts.” This required between 1,500 and 1,700 metric tons of chemical additives – acids, bases, and salts. These are used after sawing to etch the wafers down to the processing thickness of 180 micrometers and to structure the upper surface.

A corrosive matter

“One of our main focuses for cell production is minimizing the use of acids,” explains Holger Hoppe, corporate representative for environmental management at Schott Solar. The company plans to achieve this goal with longer service lives for the etching baths and lower concentrations in the baths. This plan also decreases the time and money needed to neutralize waste, since the waste water can only be removed once the chemicals are no longer reactive. Polycrystalline wafers are etched with nitric acid and hydrofluoric acid. Alkaline corrosives such as sodium hydroxide and caustic potash are used for monocrystalline wafers. Isopropyl serves as a cleaning agent. Here, too, low yield is a problem: for example, only two percent of the toxic base is used to etch monocrystalline wafers. Meanwhile, 98 percent is used in the expensive follow-up treatment – that is, neutralization and filtering.
These figures are the same for all manufacturers. Despite the laborious process undertaken to purify the waste water, traces of potassium remain. In streams, rivers and lakes, this chemical acts as a fertilizer, nourishing algae and killing fish. The photovoltaics industry thus follows in the footsteps of industrial agriculture – which brings us back to the green image.

Making better use of potassium

These days, the largest plants for solar cells are being built in the Far East. Within a short time, Taiwan has become one of the leading manufacturing nations. Chemical supplier Linde is currently building a pilot plant there that will process the caustic potash waste so it can be immediately fed back into cell production. “There is great value in recycling,” says Dean O’Connor, head of Linde’s solar division. The greatest economic benefit comes from eliminating the expensive after-treatment of toxic waste water. Figures from real-world operation are not yet available, but are expected this summer. Linde is leading the way to where suppliers see their opportunities – with sustainable process technology and intelligent solutions, they can fill the giant market niche opened by price competition and the solar industry’s green reputation. After all, the solar industry still has to perform a number of tasks before it actually achieves a true revolution in energy use. Other points of contention include phosphoric acids and phosphoryl chloride, used to dope silicon wafers with phosphorus. Boric acid, dimethylboron and diborane are used when doping with boron. The antireflective layer made of silicon nitride is created by separating monosilane and ammonia in a vacuum or by sputtering silicon in an ammonia atmosphere. The metallization of front and rear contacts requires pastes made of silver and aluminum. Compared to electricity use in silicon production, trichlorosilane and caustic potash are lesser problems, simply because the demand for these materials is much lower. Nevertheless, every cent counts and could end up tipping the scales – after all, the race for marketable prices has just begun.

Potential in the back-end

Only a fourth to a fifth of a solar module’s added value results from the back-end and the processes that follow cell production: soldering cell strings, lamination, framing and quality control. Some module manufacturers started to centralize environmental management at the highest corporate level and set uniform manufacturing standards years ago. They realized then what Franz Nieper of Aleo Solar now confirms: “There’s a correlation between profitability and waste.” Aleo’s factory in Prenzlau runs its environmental management according to the ISO 14001 standard, “because our customers ask about certification and environmental criteria,” Steve Pestel explains. He is the environmental manager at Aleo Solar and has plans for his company to be a role model outside of Germany, as well. “We will hold ourselves to the same quality and environmental criteria for our Chinese joint venture as we do for our factory in Prenzlau.”

It pays off

Aleo has significantly expanded production in Prenzlau in the last few years. The company started with a module output of 15 megawatts in 2003. Now, some 190 to 200 megawatts a year comes from three production halls. There is hardly any dangerous waste, but there are many clever ways to save. The company minimizes costs for the disposal of operating and cleaning materials by means of recycling: “In 2008, credits for recycling glass and packaging almost made up for expenses for disposing of other waste, such as cleaning cloths and waste oil,” Pestel says. Aleo tracks every possibility to save pennies. “Our goal is to use five percent less energy each year, taking into account production output.” Steve Pestel describes an example: “Costs for water and waste water are very important.” In principle, water is used at Aleo only to wash glass and could be fed into the municipal waste water system without further processing. “In washing glass, we use water in a circulation system that is cleaned again through an ion filter. We therefore use less water in production than our employees use to shower.” In module manufacturing, packaging, cleaning agents, and films make a dent in pocketbooks.
Other costs include solar glass, aluminum for frames, copper stringers and fluxes for soldering machines. In 2008, 40 percent of Aleo Solar’s production waste consisted of packaging material, such as paperboard and used paper. Clear and tinted foils made of polyethylene contributed another 18 percent. Together, adhesive remains and trimmings from the laminators came to about 15 percent. The share of white solar glass was about ten percent. Broken cells are returned to suppliers, who include Q-Cells and Bosch Solar. One of Aleo’s focal points is the fluxes for soldering machines, in which the cells are connected to stringers. The fluxes are often butyl acetate and isopropyl alcohol with three to four percent solid particles. The actual soldering metal is a combination of tin and lead. Replacing it with a lead-free flux made of copper, tin and silver is difficult, since soldering mistakes increase when there is no lead. Soldering connections are then also insufficiently stable.
Berlin’s Solon is also researching this problem. “We have a register of hazardous substances, for example for the flux in the stringer machines. It contains solvents that emit harmful vapors,” confirms Constantin Gerloff, corporate environmental management officer at Solon. “Our objective is to decrease harmful substances by five percent – in volume and number – every year.” The plant in Berlin uses less than 25 liters of flux each day, about a thousand liters each year. “We see a need for research to further decrease harmful substances in modules,” says Lars Podlowski, representative for technology on Solon’s managing board. “One example is backside foils, which contain halogen. We need to replace them in the long run.”

CFCs in PV films

Halogens are organic bonds that contain, for example, chlorine or fluorine. If they get loose in the environment, they can cause serious damage. Like hydrochlorofluorocarbons (HCFCs), which used to circulate in air conditioning systems and refrigerators as a cooling agent, solar films with fluorine or chlorine can release climate-damaging gases when being recycled or burned. Another problem is that the conventional Tedlar (polyvinyl fluoride) film contains lead. Lead-free back sheets made of heat-resistant polyvinyl butyral (PVB) are therefore increasingly being used. This is not the same as polyvinylchloride (PVC), which also has the problem of containing halogens. In the laminators, which shrink-wrap the modules, reducing energy consumption is essential. The modules stay there for up to twenty minutes in a vacuum at 150 to 200 degrees Celsius until the films turn to liquid and air bubbles are eliminated. Swiss Solar Systems now equips its laminators with hybrid heaters, in which an oil film evenly distributes heat from electrically heated loops across the entire module surface. This process improves the interlacing in the EVA film, which is a polymer of ethylene and vinyl acetate. The large laminators’ waste heat is also used to heat the plants, but this requires large heat exchangers and water-based heating systems that must be considered as early as during a factory’s initial planning stage. String ribbon wafers are drawn directly from the hot liquid silicon, thereby saving material costs.