by Joe Kwiatkowski, Physicist, Imperial College London
The last quarter of 2007 was an exciting time for the Silicon Valley start-up Innovalight: first a successful finance round that drew US $28 million of new capital, then the accolade of being amongst Red Herring's top one hundred innovators. Why the interest in Innovalight? Because of its remarkable claim to be able to print thin-film silicon solar cells.
Printing is generally a low-cost and high throughput process, in stark contrast to conventional methods used to produce amorphous and crystalline silicon solar cells. As such, Innovalight claims it will be able to substantially reduce the cost of photovoltaics. In a recent interview, CEO Conrad Burke predicted cells may eventually be sold for US $1 per watt — a figure perhaps determined less by technological considerations and more by similar claims made by his neighbors like Nanosolar.
Although details remain tightly guarded secrets, the essential element of Innovalight's process is an ink made of silicon nanocrystals. These nanoparticles can be made in a variety of ways, for example by assembling a group of molecules that contain silicon and then burning off everything except the silicon.
A patent filed in 2005 suggests that Innovalight is using a "radiofrequency plasma" to make its nanoparticles. By blasting silicon rich molecules with an electromagnetic field (at a radio frequency) it is possible to generate a gas in which some of the molecules have lost an electrical charge. Whilst charged, the molecules are extremely reactive and, with a bit of careful chemistry, can be coerced into forming nanoparticles.
By suspending these nanoparticles in a solvent to make an ink, Innovalight can then print silicon films. However, as printed, the nanoparticles are not interconnected and so the film has a high electrical resistance. To lower the resistance, the nanoparticles have to be joined by heating them until their edges are melted, at which point neighboring particles can fuse. The melting point of bulk silicon is over 1400º C and the cost of heating is a substantial cost in the production of crystalline silicon solar cells. However, a fortunate advantage of using smaller particles is that they have lower melting temperatures. Purposefully vague in their descriptions, Innovalight says only that it uses temperatures between 300 and 900º C, (possibly at high pressure and for times that could be anywhere between 5 minutes and 10 hours). Whatever the exact details, the company evidently hopes that a low-temperature printing process could offer substantial savings over conventional silicon solar cells.
It is still unclear what efficiencies Innovalight will achieve. Presumably, because it is working with thin-film solar cells, its silicon is substantially amorphous and would therefore have stabilized efficiencies of about 10%. Whatever the efficiency, and despite the difficulties that are inevitable in developing a new technology, an advantage of Innovalight's manufacturing process is that there is a wonderful number of variables that can be adjusted to get the most out of the cells. For example, nanoparticles can be grown in a variety of shapes and sizes or different nanoparticles can be mixed to determine the exact properties of the printed cell. Or, by adding germanium and tin nanoparticles to the ink, the light absorption properties can be tuned; by printing successive layers with different absorption properties, tandem solar cells could be built that would allow higher efficiencies to be reached.
Though it is probable that Innovalight will have to compromise on cell efficiency in order maintain low costs, it has come up with a phenomenon that might just help make up for its losses. According to a recent paper published in collaboration with the National Renewable Energy Laboratory (NREL), "multiple exciton generation" has been measured in Innovalight's silicon nanoparticles. What this means is that the nanoparticles might be able to produce more electrical charges than would normally be expected from a given amount of sunlight. Without this effect, the highest efficiency that a standard solar cell could ever achieve is 31%; anything else is thermodynamically impossible. However, with multiple exciton generation, the thermodynamic limit is boosted to 44%. If Innovalight could take advantage of this phenomenon it might be able to match, or even exceed, the efficiencies of conventional silicon technologies.
With its new funds Innovalight plans to construct a 3000 square meter manufacturing facility in California, and to triple its workforce over the next year. Although there is no official date on the company's website for the start of production, 2009 has been suggested elsewhere. Until then, all we can hope for from Innovalight's printers are more announcements of funds and awards.
Joe Kwiatkowski is a physicist at Imperial College London, where he works on organic photovoltaics. His current interest is the development of computational methods that can aid the design and optimization of new photovoltaic materials.