Solar Cells and Photonic Crystals

A place to discuss solar cells and photonic crystals, both in theory and experiment.

Friday, April 24, 2009

fusion

I know this might seem slightly off-topic, but I'm almost dumbstruck by all the recent stories on fusion power. First of all, there was a 60 Minutes story (link), which claimed that since Pons and Fleishmann's original work 20 years ago, which was wholly discredited, significant breakthroughs have been made. However, this doesn't stand up on closer examination. First, there are no theoretical grounds for these claims. Nuclear fusion requires a tremendous amount of energy for a simple reason: every nucleus has a positive electric charge, and nuclei strongly repel one another before they get close enough to fuse, because of Coloumb's law. Putting hydrogen on a palladium matrix, as all these experimentalists do, has virtually no impact on this basic fact. Rather, very high temperatures are needed to allow nuclei to overcome the potential energy barrier. The theoretical explanations I've seen advanced to explain why cold fusion should work are also complete nonsense and are contradicted by the entire field of quantum mechanics. But OK, let's suppose that maybe this is just a novel phenomenon that our theories can't explain yet (even though that's extremely unlikely because we've probed basic physics through careful experiments, up to temperatures of trillions of degrees celsius). In that case, there are still a number of issues with the results. First, consistency with what we know from previous fusion experiments. If fusion takes place, all the following byproducts would be expected: high-energy gamma rays, helium, tritium, and free neutrons. It seems like people have made claims to see possibly all of these components at various times, but never all at once, and in the right ratios. That's a HUGE red flag. Second, that leads me into issues with reproducibility. The "best" cold fusion researchers get about a 70% "success" rate. However, it really ought to be a deterministic process, also reproducible in other labs. The researchers mutter vague statements about the palladium not being cleaved right... but honestly, it sounds to me like an excuse for the embarrassing fact that independent labs can't reproduce their results. The third issue is the lack of consideration of alternatives. The one thing that researchers consistently claim to see is excess heat, evolved over a period of weeks, if not months. However, small errors in the calibration, or simple chemical reactions, could produce identical measurements. Calibration errors or chemical reactions with impurity would also explain the unpredictability of the results. In short, there are a number of alternative, much less exotic explanations for these measurements, fully consistent with the laws of physics and chemistry as we understand them. Extraordinary claims to the contrary require extraordinary evidence, which is WHOLLY lacking here.

Another thing I've seen recently are a number of groups claiming they're going to get hot fusion going as an energy source. First, there was an article by Thomas Friedman in the New York Times (link); more recently, there's a startup, Helion Energy, claiming that it'll just take a bit of VC money and start generating fusion power for cheap. Now, I don't want to lump them in with cold fusion, as they're on solid ground in terms of the basic physics. However, there are still major problems. First of all, in order to get hot fusion going, you need to heat up the plasma tremendously, and at the same time, confine it to a small volume, since it has a naturally tendency to expand into its surroundings. All of this takes energy. A LOT of energy. In fact, there hasn't really ever been a fusion experiment where more energy came out than went in. The closest to date was the JT-60 tokamak in Japan, in which certain channels produced more energy overall than they took in for a short time. However, despite its impressive accomplishments, overall, it's a net energy sink. There are a number of other issues: controlling the plasma is very difficult, mostly due to our lack of complete understanding of the detailed behavior of plasmas (it isn't an easy problem); massive irradiation resulting from production of tritium and neutrons (you might recycle the tritium, but recycling neutrons is mostly a non-starter); not to mention practical difficulties in building and operating infrastructure big and reliable enough to keep fusion going for enough time to make the investment worth it. In short, hot fusion has theoretical potential, but that's all it has for now. A lot of work needs to be done before we take it seriously as an energy source. In the meantime, the basic science is important, and ought to be supported for its own sake. And who knows, maybe there will be a breakthrough eventually to make this sort of thing practical. But it certainly won't be cheap: the NIF has cost at least $4 billion, and the ITER experiment is projected to run $17 billion over 30 years, in a field notorious for tremendous cost overruns.

In summary, cold fusion is almost certainly impossible, and hot fusion is almost certainly impractical. I highly doubt these things will change soon enough for us to put off other low-carbon energy research, so I'd encourage everyone reading this article to keep it in mind when prioritizing energy funding (from either private or governmental choices). And full disclosure: in case you don't know, I'm working on solar energy.