Nanosolar

I'm not always aware of which topics are hot at any given time since I don't own a television (hmm - a future blog topic?) although I do get the daily Dagens Nyheter. But I've been curious about a company called Nanosolar for a while and I've decided get some perspective about their technology and claims in fabricating CIGS solar cells, the same kind that I research. The suggestion to do so at this time came from Bengt Axmacher's comment to my post about building a solar charger. If I'm surmising correctly, he's wondering how a company such as Solibro that uses expensive vacuum deposition equipment in a relatively slow and materials inefficient manner to painstakingly produce solar modules one glass pane at a time could possibly compete with Nanosolar's cheap ink printing method using a metal foil roll-to-roll process with 100% materials utilization to produce solar cells that are - apparently - just as good.

Some fact gathering is needed here and I started at Nanosolar's website. Before getting into their technology, I first was curious about the end result: the efficiency of their solar panels (I'm one of those people who likes to find out how the story ends before reading the book). That's apparently a secret - I couldn't find it anywhere even in a broader web search. But they do advertise the efficiency of their absolute best cell as being 14.6% in a paper published March 2006. It was a half square centimeter in size and was made on a glass substrate. Incidentally, March 2006 was also when I fabricated my own record cell of 18.5% efficiency using a process identical to Solibro's. Unfortuneately, it can be very tricky to compare results between two different cells. In this case, the Nanosolar cell was actually measured to be only 13.95% efficient but was calculated to be 14.6 had it not had a top metal contacting grid on it. Similarly, my cell would have been 19.1% if I'd used the same trick of calculating away the contact area. Other things can also affect efficiencies such as antreflective coatings or whether the cells are encased for durability. When making large cells there will be more inherent losses and they are also more difficult to manufacture uniformly. Modules compound the losses due to the interconnects of the individual cells and problems in cell matching. Trying to calculate what the efficiency of a fully assembled solar panel might be based on the results of a weentsy little lab cell is not straightforward. If you take a look at my posting about a solar module you'll see it was 14.8% efficient, i.e. not even close to the 18.5% I got for the single cell. So here goes: I'm going to guess wildly and predict 8 to 9% efficiencies for Nanosolar's fully assembled panels.

On to some technology. The copper-indium-gallium-selenium ink that Nanosolar uses to print their solar cells is proprietary, but apparently it's made of nanoparticles of metal selenides which are printed onto a metal foil which forms one of the electrical contacts. It then goes through a heat treatment of 4 or 5 hundred degrees. Additionally, solar cells require a buffer layer and a top contact to form a structure that I described in a CIGS solar cell. I couldn't find out which method is used to deposit these layers, but Nanosolar emphasizes that vacuum deposition is not it. Hopefully, I'll get around to describing the vacuum deposition method someday since it's the way I make my solar devices.

I'm undeniably intrigued by Nanosolar's process. Vacuum deposition methods are indeed wasteful. Materials usage can only be minimized when huge glass panels are processed since the waste is mostly at the edges. Applied Materials makes thin film silicon on glass panes that are 5.4 square meters - I think it's the biggest in the industry. But vacuum equipment is terribly expensive, takes a tremendous amount of energy to run, corrosive selenium gets into everything and, in my experience, the heaters for the metals keep burning out - I don't just melt copper, I evaporate it at about 1500C. It's difficult to get uniform coating and control the respective quantities of copper, indium and gallium in the resultant films. And this is just for one of the layers needed to make cells. Modules also require patterning steps to form the interconnects between cells. Nanosolar doesn't require this in their process, but instead they have to assemble cells into modules and make interconnects. As far as esthetics go, I think Solibro has the upper hand. I suspect their sleek black panels are easier for an architect to place on buildings.

Back to the original question of whether Solibro can compete with Nanosolar: I conclude that it would be entirely premature to predict the demise of Solibro.

But neither company has yet made a product available to the public.