They were studying these sorts of devices at Western Washington University when I was there back in the late 90's. They were using simple black-body radiators and the efficiency was pretty poor, even with a recuperator pre-heating the combustion air. They also had to run coolant behind the solar cells to keep them operating, so some of the energy was lost through that route.
One thing I was never able to quite pin down is what the theoretical efficiency limit would be for a device like this, specifically whether it was analogous to a heat engine or could do better than the Carnot limit. Any physicists/physical chemists care to chime in?
EDIT: Wikipedia says they're limited to Carnot efficiency.
Apparently, these panels work on the thermoelectric effect (different from the photoelectric effect), they're Peltier piles working in reverse. Thermoelectrics are Carnot-limited, but photovoltaics are not.
The article suffers a bit from the fog that covers most popular science articles but I think I get the picture. As others have said this particular technology is based on photovoltaics, in particular narrow-bandgap photovoltaics typically referred to as Thermophotovoltaics [1][2].
Bad news first, unfortunately the Carnot limit to efficiency still applies [3] and as an aside the limit for simple, single-junction photovoltaics under one sun illumination is about 31%. Since we're not dealing with either the sun nor necessarily single junction TPVs there's more to say on the topic.
The scheme proposed here seems to be very similar to a design proposed by one of the awards of the DARPA RIMS [4] (later MIPS [5]) program. Specifically they're proposing a combustion cell for the butane, a tungsten photonic crystal [6] emitter to reduce long wavelength "waste" emissions and THEN a TPV. A TPV consists of one or more pn junctions consisting of narrow bandgap semiconductors, typically horrendously complicated structures of ternary or quaternary materials like InGaAsP. [8] Please note that they don't even try to address the structure of the TPV in the article.
Some practical engineering difficulties include the fact that the nanosctructured tungsten emitter is quite reactive in oxygen at elevated temperatures (probably still true for their proposed structure). I should know, I reduced an incredibly expensive photonic crystal emitter to a very pretty layer of rust while trying to measure its emission spectrum while it was under argon. It turns out that it's incredibly difficult to purge oxygen from this structure after it's been exposed even ONCE. If you're using this as a battery replacement for soldiers you have to ensure that it can really take a beating and since the TPV is quite brittle (yep, killed a few of those too) you have to have an enclosing structure that is quite rigid, adding to your weight. The combustion cell is also going to be an engineering challenge since you have to maintain pretty good temperature control to keep your conversion efficiency up. How do you control your fuel-air mixture? What happens at high altitudes? How are you minimizing parasitic heat loss?
The reason I think a lot of people are enamored with this fairly niche piece of engineering is due to an erroneous Sandia Labs announcement that they had excess emission in a particularly favorable wavelength region. [7] That and the fact that DARPA is throwing an increasing quantity of money at a problem space with relatively few solutions.
Despite all the problems I listed this sort of TPV structure is still a very attractive option in at least one market namely powering deep space missions like JUNO or New Horizons [9]. Currently these are powered by something called an RTG [10] using thermoelectrics, but more on thermoelectrics another day. Below I have linked to a very good article on the subject of TPVs for space written by some of my former co-workers. [11]
Sorry for the lengthy reply I got excited that I have something to contribute to one of these discussions!
The wikipedia article (http://en.wikipedia.org/wiki/Thermophotovoltaic) has some excellent applications for this technology including off-the-grid power generation from camp fires or earthen heat stores.
Imagine a CPU which integrated this technology into the package so that it could re-capture a portion of its generated heat as power. Current designs could be made more efficient, and lower battery usage.
My impression is that they utilize a second conversion stage that is essentially a conventional solar cell. So the sequence is: heat => light => electricity. The novelty is the design of the first conversion stage, which is specifically tuned to produce wavelengths that can be efficiently converted by the second stage solar cell. As opposed to just burning the butane and running a solar panel off of the firelight.
from the article: "Imagine using the heat generated from your laptop to charge it as well? Or charging your car’s hybrid battery by using the heat off of the engine?"
Wish the author had rephrased these in view of The Law of Conservation of Energy. Don't want to give the impression of developing a perpetual motion machine.
hey protomyth, author here. Haha, I see your point abt the conservation of energy. I assumed that everyone would know that law, so I didn't address it specifically. Obviously, the return energy would be less than the expended.
I guess to assume really does make an ass out of u and me.
It's not so much for the HN audience, but the next generation of article down the chain will use that paragraph as a basis for their headline. It is by far the catchiest and I don't think they will be explicit about conservation of energy with their readers.
Well wait, they're not talking about using the car's momentum to power the car, just its heat. The heat is waste, this just makes it somewhat more efficient.
Oh I see what you're saying, sorry. You mean that the writing would be ambiguous to those not as familiar with the concepts, they might take it to mean that all the energy would come from the heat.
yeah, as my down-voted comment mention, I figure the HN folks would get what the author meant (once again, good article), but it might give the false impression to someone less familiar with the concept. They might take the sentence as suppling all the energy, and someone picking up the article might use those sentences as a basis for a headline (catchy = headline grab).
One thing I was never able to quite pin down is what the theoretical efficiency limit would be for a device like this, specifically whether it was analogous to a heat engine or could do better than the Carnot limit. Any physicists/physical chemists care to chime in?
EDIT: Wikipedia says they're limited to Carnot efficiency.