This article on fast breeder reactors is expertly timed to coincide with everyone forgetting what I said about regular nuclear reactors a couple of months back.
A brief recap: you start with a big chunk of uranium mined out of the ground. This chunk of uranium will be 0.3% U-235, which is fissile, and 99.7% U-238, which is fertile. The difference between fissile and fertile nuclear material is that fissile uranium can be induced to split fairly easily by bombarding it with what are called slow “thermal” neutrons. In order to slow neutrons down so that they trigger nuclear fission nuclear reactors cores are surrounded by a moderating material such as graphite or water which rob neutrons of their energy as they repeatedly bounce off of water/graphite molecules. Once they have been bled dry of energy the slow neutrons then interact with U-235 nucleii, producing energy, fission byproducts and more neutrons which themselves are slowed by the moderator and continue the chain reaction. At this point the reaction is self sustaining your main challenge is in controlling the rate of neutron absorption/fission so that your nuclear reactor doesn’t melt into a pool of highly dangerous radioactive sludge, usually with control rods composed of a neutron poison such as boron that stops the neutrons dead in their tracks and halts the nuclear reaction.
Now, if we were to try to use fast neutrons to start a chain reaction in our fissile material we’d run into a big problem. Fast neutrons have what’s called a neutron absorption cross-section (basically the chance that a neutron fired at a U-238 atom will be absorbed by it) that is far lower than that for thermal neutrons and U-235, which means you’d need a very high density of fissile material – in other words, a big critical mass – in order to induce a chain reaction. While a light water reactor using slow neutrons needs less than 5% of its fuel to be enriched U-235, a fast neutron reactor needs a much larger 20% plus quantity of enriched fuel. Fuel enrichment is one of the most expensive parts of the whole nuclear reactor energy generation process, so using highly enriched fuel in a reactor like this is both costly and inefficient. So why would we bother?
The answer is: so that we can get at the energy locked up in the 99.7% of natural uranium that is merely fertile, and not fissile. Fast breeder reactors are called “breeder” reactors because they convert the otherwise-useless fertile U-238 that comprises 80% of the fuel material into fissile P-239 by hitting them with fast high-energy neutrons. Set it up right and a fast breeder reactor can potentially create more useable fissile fuel than it consumes; the concept of fast breeder reactors in general would increase our useable stocks of uranium by a factor of about 100. However, there are some drawbacks to the fast breeder process that have prevented the practical development of the concept beyond a few test reactors.
The main stumbling block currently facing fast breeder reactors is an economic rather than a technological one: with the high upfront cost involved in building a breeder reactor (25% more than an equivalent light water reactor) and the additional ongoing costs involved in enriching nuclear fuel to the point where it can sustain a chain reaction using fast neutrons, creating new nuclear fuel out of fertile material using a fast breeder reactor is currently more expensive than just digging a lump of uranium out of the ground and enriching it the old fashioned way. This is just a temporary barrier, though, as sooner or later (assuming no discovery of previously unknown uranium deposits, between 50-100 years from now) the supply of naturally-occurring fissile U-235 is going to start drying up and the cost of producing fuel that way is going to exceed the cost of making it out of U-238 in a breeder reactor. So at some point we are going to be doing this; it’s just a matter of time.
Then you have the technological and political challenges inherent in building a fast breeder reactor. The technological hurdle is that you can’t use water in a fast breeder reactor. Like, at all. It’s a very good neutron moderator which makes it great for light water reactors, but it’s absolutely what we don’t want when we’re trying to saturate our reactor core in fast neutrons. This means that you have to rely on an exotic coolant like liquid sodium/molten salt, described in my post on LFTR reactors. Liquid sodium actually has some advantages over using water as a coolant, but it also makes reactor maintenance a bit trickier as it’s rather corrosive, not to mention potentially catastrophic if the reactor ever sprung a leak as liquid sodium will ignite on contact with air.
And then there’s the political angle, which anyone who knows what the letters “P-239” should have picked up on. P-239 is plutonium 239, otherwise known as the primary constituent of nuclear weapon designs the world over. Widespread deployment of fast breeder reactors would vastly increase the worldwide availability of plutonium and would almost certainly lead to increased proliferation of nuclear weapons, especially since the fast breeder fuel cycle involves regularly reprocessing the fuel cores in order to extract the converted P-239. My attitude towards this is that there are far cheaper and easier ways of creating widespread devastation than detonating a ghetto nuclear weapon, and that terrorists/rogue states would be better off looking at those before trying to develop nuclear weapons and their associated delivery systems, but then I don’t live in a world where such weapons are widely available. One suitcase nuke isn’t much of a threat, but a hundred of them could seriously muss up a country’s hair.
I should finish by saying that unlike LFTR reactors, fast neutron reactors do exist even if they’re not specifically set up to breed nuclear fuel. They’re used because fast neutron reactors have a very efficient neutron economy (that is, very few neutrons are lost to absorption or other processes allowing the reactor to run a on a comparatively small amount of fissile material) which means they can be used either in compact applications like nuclear submarine reactors, or else in commercial reactors that don’t want to bother with actually enriching nuclear fuel a la the CANDU series. Actual fast breeder reactors have also been built, although further construction was discontinued after it became apparent they were not economically efficient. Expect that to change at some point in the near future; many Generation IV nuclear reactor designs – that is, the ones that we will be building over the next half-century – are of the fast breeder type, and eventually we’re going to have to switch over to breeder reactors if we want to continue using uranium as an energy source.