What would be really interesting is if you could tune how much IR (edit: near-IR, actually) light goes through, and how much gets absorbed... that way you get solar heating when needed.
Edit: thinking more about it, why not have the whole visible spectrum be tunable, so that you can create window blinds any time you want them, and increase energy production in the process?
First off, normal glass also blocks IR light (which is a good thing) when the outside produces more IR light you want to cool the inside, and when it's producing less you want to heat the inside.
Anyway, the energy from sunlight is for the most part visible light so you can't really use most of the energy without also blocking visible light. The best option is automatic shades which make the outside look dimmer but don't totally block the view.
It's an interesting idea, I am simply mentioning the practical limitations.
At this point it's a question of how cheaply they can scale it up and how much you can improve the efficiency. Looking at some of their numbers they block ~30+% of the light to gain ~1.3% of the energy you are better off with smaller windows and a high efficiency solar cell around the window. AKA 10% smaller windows block 10% of the light + a 20% efficient solar cell = 2% efficiency. Granted, there is probably a market even in it's current state.
I think the news article exaggerates the purpose of the device described in the actual paper. It isn't intended as a competitor for opaque power generation cells. The idea is that, since you already need an anti-reflective coating to block IR light, it might as well be a coating that can scavenge a bit of the incident power instead of sending it all back into space.
This seems a bit weird. They claim that they think they will reach an efficiency of 12%, like normal solar cells, but that's of course neglecting the fact that only a fraction (max half) of the sunlight is in the IR. So the amount of energy you'd get is still only half (not that that is insignificant).
The 12% efficiency quote seems to originate in an interview with the authors. I read the actual paper,[1] and it only mentions a projected efficiency of 2-3%.
You are right that there simply isn't enough energy in the near-infrared to achieve a 12% efficiency. As in a lot of scientific reporting, it's hard to know whether the journalist misinterpreted the authors, or if the authors were indulging in a bit of hyperbole to promote their work. =)
The device described in the paper is intended for energy scavenging rather than as a competitor for standard opaque cells. The idea is that, since you already want an anti-reflective coating to block IR light and minimize your air conditioning bill, why not use a coating that also incorporates a solar cell?
If that's true then regular solar cells only get a shot a half the energy too. So it doesn't seem so weird to me.
More weird to me, after spending some time in the business, is all these solutions which only require "some engineering" are perpetually around the corner. People, stop taking "some engineering" or granted! :)
Well, it depends on what you mean by efficiency. I would say that if I say "I have these solar cells that take sunlight and create electricity with 12% efficiency", I would expect that I get 12% of the total insolation in electricity. If I instead say "these solar cells take infrared sunlight and create electricity with 12% efficiency", then I expect it to be 12% of the infrared sunlight, not 12% of the total sunlight. It depends on what the hypothetical "fully efficient" case is.
Besides of windows, one of the first real-world applications of this technology might be in our smartphones in order to gain in battery life.
That's what WYSIPS ( http://www.wysips.com/ ) has planned to do with their similar technology. And they seem a bit ahead.
I find this fascinating. Slightly off topic, but do you think we'll see in 10-20 years most new buildings generating a significant % of their energy independently (their own solar, geothermal, waste re-use)?
I stopped reading when they said it is only 1.7% efficient but expect it to get up to 12%. That's almost an order of magnitude! It seems unlikely that their prototypes are so bad that they can get an order of magnitude improvement easily. And if it takes years of work to get to 12%, well, everybody else is claiming great performance in the nebulous future, too.
Considering that you currently get 0%, that's not bad. If the cost is low enough and people could reduce they're electric bill by a small amount why not dual-purpose your windows? It might even pay for the upgrade over ten years.
Edit: thinking more about it, why not have the whole visible spectrum be tunable, so that you can create window blinds any time you want them, and increase energy production in the process?