Thursday, November 1, 2007

Fusion Reactors and Proliferation

As a part of my ongoing participation in the Cult of Bussard, I posted the following as a thread on
Here's the issue:

Today, there are two ways to make stuff for nuclear weapons:

1) You can enrich uranium to produce highly enriched U-235 (often abbreviated HEU). This is hard because you have to have centrifuge cascades or other high-tech, expensive, and easily detectable industrial equipment.

2) You can put U-238 into a breeder reactor to produce Pu-239, then chemically purify the Pu-239 to produce bomb-grade stuff. This is actually a lot easier to do than producing HEU, but it requires a nuclear reactor, which is in turn, large, high-tech, expensive, and easily detectable.

The US, EU, Russia, China, and even the good ol' IAEA all know how to find states that are attempting to get weapons through either of these methods, because the industrial footprint is unmistakable.

But notice that method #2 works equally well with any high-density neutron source. Fortunately, the only neturon source that's dense enough to be workable is a fission reactor.

Or at least it is until you get a high-power D-D or D-T fusion reactor going. Then you've got plenty of neutrons. Instead of a lithium thermal blanket, you jacket your fusion reactor with a water blanket to slow down the neutrons to thermal energies, then a U-238 blanket on top of that. Marinate for a while, then take your U-238 + Pu-239 away to be refined.

Now, nobody's going to worry too much about somebody sneaking a tokamak or a laser inertial confinement plant into the outskirts of Islamabad or Khost. Such a facility would cost billions of dollars, take up at least a city block, and require all sorts of exotic metallurgy. Again, it's highly detectable.

But what about a polywell or even a focus fusion plant? It seems reasonable that you could fit such a plant into an ordinary warehouse or even a large garage. We think that the materials to construct such a plant (especially one where you weren't generating electric power) are fairly mundane. Would a polywell be detectable from a non-proliferation standpoint?

So, two problems emerge that we need to consider. First, what does a compact fusion reactor mean for non-proliferation enforcement? Are there actually ways of detecting them? Is there some sort of technological secret sauce that will be easily traceable or put the technology out of the hands of rogue states and non-state actors? Will a polywell be so cheap to deploy and so simple that it's acquirable by a rogue state? By a terrorist group?

But the second problem may be more interesting to this group: If I were a legislator that was faced with funding research into a technology that might have huge power applications but also came with a substantial proliferation risk, I might equivocate, especially if that technology wasn't yet a slam-dunk. This might be silly, but legislators are in the business of getting re-elected, which doesn't always translate into doing the right thing. We obviously want the funding to occur. (I think the proliferation risk is real, but it's dwarfed by the possibility of a permanent solution to our energy problems.)

What are some of the arguments that we'd put in front of legislators (or any other potential funders) that would assuage their proliferation fears?

Minor disclaimer: I really don't know if Bussard's polywell fusion technology is viable. Similarly I don't know if focus fusion is viable. But I strongly suspect that something will be viable some time in the next 15 years. The incentives to solve energy problems are just too large to avoid lots of fresh eyes taking a look at the fusion problem. Somebody's going to come up with something. We should ask questions like this now. We should also ensure that funding doesn't get restricted based on people being scared of the lesser of two evils.


M. Simon said...

Let me repeat what I said at talk.polywell.

1. purifying the Pu239 will be radiation intensive.

2. Production reactors should be designed for pB11 only.

TheRadicalModerate said...

1. Agree that you'll get isotope hits from purification, but you probably won't off of the fusion machine itself. I'm still worried that separation is simple enough to be done simultaneously in multiple locations, which could defeat sniffing to a large extent. (Note that I'm much more worried about sophisticated terrorists than I am about rogue states. With rogue states, we can probably blow the whistle on them and apply pressure. No such luck for terrorists.)

2. If there's some sort of non-obvious tuning that will prevent a p-B11 machine from being dual-used as a D-D or D-T machine, that would be a really good argument to have in the bag of tricks.

As I mentioned on the other thread, I don't have a clue how to do the neutron flux calculation. It may be that the number of neutrons thrown off even by a scaled-up fusion reactor is so much lower than that in a fission/breeder system that the fusion machine makes a lousy breeder.

M. Simon said...

Fusion reactors burning deuterium will produce 10X to 20X as many neutrons as an equivalent fission reactor.

Extracting Pu is not going to be done by terrorists. Too much science, too much eqpt., too much radioactivity.

The power supply for a non-power producing (neutrons only) 100 MW reactor will consume 1 to 10 MW. Not a lot. Not trivial.

Proliferation will be about as hard as doing it with a fission nuke.

A distributed system increases the chances of detection by accident if nothing else. Plus terrs are sloppy. Radioactives blowing in the wind will get detected.