Chemical reprocessing of nuclear fuel is at the basis of a closed nuclear cycle. We don’t think about it too much in the United States because we just don’t reprocess nuclear fuel! Hence, our spent fuel builds up and people wring their hands about the “unsolved problem of nuclear waste”. In other countries, like France and Japan, where spent fuel is reprocessed, things aren’t too much better. Because thanks to the fact that you can’t sustain the “burning” of natural uranium in thermal-spectrum reactors, separated fuel can only get you so far. But they stockpile it waiting for the glorious day of liquid-metal fast breeder reactors that the nuclear industry has been promising for fifty years. (a day I hope will never come!!!)

You just can’t choose a nuclear fuel, or a fuel cycle, without thinking about how to reprocess the fuel. That’s how we got in the mess we’re in in the first place. Fortunately, when the “Oak Ridge boys” were thinking up liquid-fluoride reactors back in the 1950s, reprocessing was a key consideration and they came up with very attractive ways to do it. Their first and fundamental advantage was the fact they were dealing with a fuel already in fluid form. That immediately eliminated all the complicated steps you have to go through with solid fuel: chopping, decladding, dissolution in nitric acid–and that’s just the front end–then reconstitution of the fuel, which usually implies remote fuel fabrication because you’ve still got a lot more radioactivity in the fuel than fresh fuel.

So all those problems were solved from the beginning, just by working with fluid fuel. But you still needed to get through the basic steps of reprocessing, which is, you exploit chemical and physical differences in the materials in the spent fuel to separate out the things you want from the things you don’t want.

I’ve talked previously about the simple steps of fluorination and distillation that liquid-fluoride reprocessing was based on. Fluorination is especially clever–you take advantage of the fact that uranium will absorb more fluorine to go from a tetrafluoride (four fluorine atoms) to a hexafluoride (six fluorine atoms) and in that conversion, will become gaseous. It’s an incredibly nice feature for trying to separate uranium out from just about anything else (assuming all the other stuff won’t do the same trick).

Fluorination works especially well in our core salt of LiF-BeF2-UF4. When you want to get the uranium out, you bubble fluorine gas through the salt. The lithium won’t take any more–it’s perfectly happy with its one fluorine atom. Neither will the beryllium–it’s happy with two. But the uranium says “more fluorine? I’m outta here!” and converts to the gaseous hexafluoride state, leaving you with just LiF-BeF2.

Now after we run an LFTR for awhile, the core salt will contain not only LiF, BeF2, and UF4, but will contain a number of fission product fluorides generated from the fission of U-233. These fission products are responsible for nearly all the radiation levels in the reactor (when it’s shut down) and they are the materials that pose the greatest biological hazard if released. They also can increasingly interfere with continued nuclear operation, because some of them tend to absorb neutrons that would otherwise be going towards fission or conversion of thorium to uranium. So we want then out.

Distillation appears to be the best way to accomplish that. Distillation takes advantage of the fact that the things we want to keep in the salt (the lithium fluoride and beryllium fluorides) tend to vaporize at lower temperatures that the fission product fluorides. Thus by applying heat at reduced pressure, we can get the LiF and BeF2 to separate from the fission product fluorides, leaving them to accumulate in the bottom of the “still”.

Recently I realized that through the proper choice of isotopes, we could accurately test this entire reprocessing system in a completely non-nuclear, nearly non-radioactive manner. If we used U-238 to stand in for the U-233 fuel, and used stable isotopes of the fission products (such as zirconium, strontium, and barium) we could test a chemically-accurate liquid-fluoride reactor. In the blanket salt, we could add small amounts of U-238 to the salt, simulating the generation of U-233 from thorium. Then that U-238 would be removed from the blanket by fluorination. The U-238 would then be added to the core salt, simulating the continuous refueling of the real reactor. Stable fluorides of the fission products would also be added to the core salt, simulating the accumulation of fission products. Both the U-238 and the stable fission product fluorides could be removed by fluorination and distillation.

Two basic advantages of this approach are: 1. because the reprocessing steps chosen for the reactor are not significantly affected by radiation, the lack of radiation does not compromise the accuracy of the test. 2. It will be much easier and cheaper to “wring” out LFR reprocessing techniques on non-radioactive or very low radioactivity materials rather than on real fuel and blanket salt.

I believe that taking this approach to the development of the reprocessing systems for the reactor would speed development and ultimate fielding of the reactor system.