Slashdot Mirror
Caltech Makes Flexible, 86% Efficient Solar Arrays
strredwolf writes "Caltech has released a flexible solar array that converts 95% of single-wavelength incandescent light and 86% of all sunlight into electricity. Instead of being flat-panel, they stand thin silicon wires in a plastic substrate that scatters the light onto them. The total composition is 98% plastic, 2% wire — the amount of silicon used is 1/50th that of ordinary panels. So as soon as they can get these to market, solar could be very viable and cheap to produce." Update: 03/01 21:02 GMT by KD : Reader axelrosen points out evidence that the 80%+ efficiency figure is wrong. MIT's Tech Review, in covering the Caltech announcement, says that the new panel's efficiency is in the 15%-20% range — which is competitive with the current state of the art. And the Caltech panel should be far cheaper to manufacture.
http://www.rsc.org/chemistryworld/News/2010/February/14021001.asp
'We have shown the optical absorption efficiency and charge carrier collection efficiency of a silicon wire array cell is comparable to a conventional silicon cell, but a wire array cell uses up to 100 times less silicon due to enhanced light-trapping effects,' says Atwater. Significantly, the wire arrays absorb infrared light more efficiently that conventional silicon surfaces, further improving the performance of the new device.
So the gist is that it's more efficient because it converts infrared, uses some type of clear polymer with alumina "reflector particles" in place of 99% of the expensive (doped) silicon, and is flexible and therefore easier to manufacture.
"I assumed blithely that there were no elves out there in the darkness"
The original article is poorly written. MIT's Technology Review has an article that includes information about efficiency of generating electricity, and it says 15%-20%. http://www.technologyreview.com/energy/24665/?a=f
So the story is really that there might be a way to make cheaper, flexible solar panels by mixing silicon and polymers.
There are many different factors that go into making a high efficiency solar cell. You need a front material that has very high transmittance/low reflectance at incident angles, a high absorption semiconductor, a high photon to carrier generation rate, high/easy carrier collection from the semiconductor, and broad spectrum conversion. These silicon-wire arrays appear to have high absorption and high carrier generation, but thats only part of the story. The other issue is that silicon misses out on a fair chunk of the solar spectrum. Anything after 1100nm is not converted, its simply below the bandgap.
The title of this post and the article is incredibly misleading and very annoying/frustrating to someone who's been working on solar technologies for a while. Don't get me wrong, I think this is a very cool thing, sounds like they have to potential to make very cheap cells, but approach, let alone surpass, current multijunction cells (30-40% eff.) they will not.
Disclaimer: University of Michigan Solar Car Team alum