PHOTON International The Photovoltaic Magazine

Solarforce is looking for investors to commercialize silicon ribbon process

While prices for solar silicon are climbing, it seems to be an auspicious moment to dust off past research projects aiming to cut raw material waste during wafer production. That's exactly what newly-founded company Solarforce is up to. The French company plans to start production of 150 µm thin wafers using an innovative crystalline silicon ribbon technology next year.

	© Solarforce
Still far from retiring: 66-year-old Claude Remy plans his comeback on the PV scene. In 1981, Remy founded wafer, cell, and module manufacturer Photowatt, which he led until 1997.

If newcomers to the PV industry take a look into the annual German Status Report Photovoltaics from 1990, they would be astonished by the level of effort put into the development of new silicon solar cell production processes to substitute the raw-material-intensive wafer technology using silicon ingots at that time. In one of the report's papers, Armin Räuber, a former researcher at the Fraunhofer Institute for Solar Energy Systems, listed nine different silicon ribbon methods companies were working on in the US, Japan, Soviet Union, and Germany. Today, only one of these technologies that circumvent wire-sawing has been commercialized successfully: RWE Schott Solar GmbH produces several MW of solar cells based on the Edge-defined Film-fed Growth (EFG) technology once developed at Mobil Solar. The other ribbon technology used for large-scale cell manufacturing by Evergreen Solar was not on Räuber's list, nor was the Ribbon on Sacrificial Carbon Template (RST) process, which already was history in 1990.

In times of a rapidly growing PV market, an imminent feedstock shortage, and the continued pressure to reduce PV system costs, the long foregone ribbon technologies are getting a second chance. After the Energy Center of the Netherlands (ECN) had taken over development of Bayer AG's Ribbon Growth on Substrate process (see PI 12/2000, p. 10), the 2003-founded French company Solarforce SAS is trying to revive RST technology.

Like Evergreen's and RWE Schott Solar's ribbon technologies, the RST process is a vertical silicon growth process, but the silicon crystallizes on a continuous carbon substrate (see graphic). As one pulls the carbon foil upwards through a silicon melt, the silicon deposits on both of its sides. After being cut down with a laser to 125 × 125 mm squares, the foil is burned off – leaving behind for the time being two 150 µm thick wafers, from which solar cells with efficiencies between 14 and 15 percent can be manufactured.

Between 1974 and 1986, different companies worked on the RST process in France: Philips Labs LEP S.A.R.L. was the first to develop RST, before CGE-Alcatel took over in 1981 to continue R&D work through its subsidiary Photowatt International S.A. Claude Rémy, then president of Photowatt, eventually changed direction toward casting of multicrystalline silicon ingots, which is still the company's raw material for its solar cells. The reason to abandon RST was obvious: at that time, off-spec silicon from the electronics industry was available for $2 to $4 per kg, recalls Christian Belouet, who has worked on the RST process for 30 years.

But times have changed and so has Claude Rémy's mind. Not only the price for silicon has dramatically increased, the quality and availability of carbon ribbon, the crucial component for the process, is also making RST attractive now. Today, carbon ribbon is 10 times cheaper, claims Belouet. Moreover, in the early 1980s, nobody was able to process silicon sheets as thin as targeted in the RST method in a production line, adds Belouet.

In Sept. 2003, Rémy and Belouet founded Solarforce S.A. with €100,000 ($119,800) of their private capital in Limonest, France with the purpose of commercializing RST. After Philips' and CGE-Alcatel's patents were abandoned in 1998, Solarforce began preparing a wide portfolio of patents focused on equipment and product. In mid-May, the company revealed that, as of yet, it has applied for two patents, with another one in progress.

The advantage of the RST process is its simplicity and the possibility of manufacturing incredibly thin wafers, claims Belouet and points to the process' impressive pulling speed of 5 cm per minute for a 200 µm thick silicon film, which can be accelerated to 10 cm for a 80 µm film.

Moreover, unlike for common silicon wafer technology, which wastes over 200 µm per wafer (no matter if the wafer is 200 or 300 µm thick) during the sawing process, RST's loss is only around 20 percent when the silicon layers are separated from the carbon substrate and cut into wafers with a laser.

The secret of RST is the carbon foil, which on the one hand forms the meniscus in the transition from the melted to crystallized silicon, and on the other dissipates crystallization heat. The shape of the meniscus, which develops as a result of capillary forces, is crucial in every ribbon growth process: at this point it is decided about possible pulling speed and quality of the crystal. With its high meniscus, Solarforce says its process not only allows for high speeds but also high material quality. While EFG wafers, for example, have very rough surfaces, Solarforce's wafers have a smooth surface, which facilitates cell and module processing. »The difference is intrinsically related to the freezing meniscus environment – free in our process and confined in EFG,« explains Belouet.

	© source: Solarforce; graphic: PHOTON International
Almost waste-free: Solarforce's ribbon process for silicon wafers has a production yield of around 80 percent.

In its very optimistic business plan, Solarforce plans installation of the first production equipment – a ribbon puller, a carbon-cutting machine, and a pyrocarbon deposition furnace in early 2005. And in the same year, the first 55,000 wafers (125 × 125 mm large and 150 µm thick) are slated to hit the market. By 2009, the French researchers want to increase the number of employees to 67, reduce wafers thickness to 80 µm, and produce 7 million slices per year to make possible incredibly low module prices of €1 ($1.20) per watt.

The fundamental issue in getting the business going, however, is to find venture capital. The pilot production equipment alone costs around €830,000 ($994,340) and several million euros
are needed for up-scaling. While Joseph Grabmaier, who was one of the developers of Siemens AG's S-Web silicon ribbon technology failed in his effort to perk up this process in 2000 (see PI 9/2000, p. 18), and ECN struggled long to acquire fresh cash for its second project phase, it will be interesting to see how Solarforce will perform. At least the two French »pensioners« recently won an award in a national contest for innovative start-ups, underlining that they still understand how to advertise for an idea. The next challenge they will face will be to see how well they can convince investors.