The way I read that was that they're saying they've found a way to use the ionizing energy from a radioactive pile to ionize an intermediate element which will then emit light upon returning to its ground state. Not sure if this has been done before, but that's the mechanism that stands out for me.
Yes, that's the impression I had as well.
@HerraTohtori: Batteries that run off of thermal energy from a nuclear source already exist: Radioisotope Thermoelectric Generators. Only drawback is they're incredibly inefficient. http://en.wikipedia.org/wiki/Radioisotope_thermoelectric_generator
Yes, I know that, and the reason they are so inefficient is because thermal radiation is by and large black body radiation, which by definition spreads the radiation on a wide spectrum.
This is a problem with photoelectric phenomenon, because it can only utilize photons that have high enough energy levels, and thermal radiation has very low energy levels by definition compared to visible light even. And semiconductors with low enough barrier energy to utilize thermal radiation in a photoelectric reactionare, as far as I know, rather rare and expensive...
This means that the panels that can utilize thermal radiation are first of all expensive, and secondly a large part of the radiation is just lost in translation because it has too low energy levels per photon to trigger the photoelectric reaction.
Converting large portions of ionizing radiation into visible light of narrow spectrum means the photoelectric cells can be optimized for those wave lengths, and there's much less of those photons that can't be converted to electricity.
All in all it sounds interesting and promising, but large scale utilization reeks of practical problems I already mentioned (mainly the fact that photoelectric panels can only handle limited intensity (power per surface area) before they heat up, melt and/or ignite), so the larger the desired power out put of a single unit, the bigger the photon collecting sphere around the reactor and photon generator needs to be.
Anyway, the point is indeed in reducing complexity. Photons are much safer and efficient way to transfer energy than high pressure steam or liquid. Requires much less maintenance, too. The reactor can be run at much lower temperatures during normal operations. There's no need for extremely heavy duty pressure vessel for the reactor core. There's no need for pumps for cooling.
All this means the reactor will be very much lighter than current systems, and that means you could hypothetically use it as energy source for, say, airplanes. The biggest block of mass involved would be radiation shielding, but because the process itself would be much more effective, the actual amount of radiation would be smaller than from a current reactors of equal power output. Assuming the dimensional size limitations don't hamper it's use as a mobile energy source, that is.
So, yeah, it's definitely interesting.
But it won't beat fusion.
