Fusion reactors have gotten a lot of attention over the last few years. The recent excitement seemed be kicked into high gear when Commonwealth Fusion Systems announced in late 2021 that they had closed on a $1.8B funding round led by Tiger Global Management. Over the next few years, many other fusion companies were born and followed suit, with large seed rounds and early-stage funding rounds. Nearly all of these fusion efforts were spinoffs from various university programs, although some more recent ones were not.

Most of these fusion efforts have also centered around the deuterium-tritium (D-T) approach for fusion. There’s a good reason for this. The D-T reaction has the highest cross section at the lowest energy, meaning it is the most likely fusion reaction to proceed at a given plasma temperature. But the D-T reaction has two big drawbacks. First of all, it is dependent on a fusion fuel (tritium) that is very rare and very expensive. Secondly, the D-T reaction releases 80% of its energy as a high energy (14.6 MeV) neutron that does not contribute to the heating of the plasma. Since neutrons are uncharged particles, they cannot be confined by electric or magnetic fields and sail right out of the plasma.

The strategy to mitigate these challenges is a “breeding” blanket in the fusion reactor, and in more recent D-T reactors designs, and especially those that plan to use high-temperature superconducting magnets, that breeding blanket is planned to be based on FLiBe salt. Dr. Charles Forsberg of MIT gave a very interesting talk about this in 2019 at Oak Ridge:

I took a fusion engineering class at Georgia Tech during my graduate studies, and knew about many of the challenges of magnetic confinement fusion, and of D-T fusion in particular. Since learning about molten-salt reactors, I had often wondered if lithium-beryllium fluoride salts had been considered as the breeding blanket for a fusion reactor. I reasoned that LiF-BeF2 already contained two of the important ingredients for that blanket. It had lithium, needed to breed new tritium, and it had beryllium, which could serve as a neutron-multiplying material. But it did not seem like LiF-BeF2 had been favored as a fusion blanket material.

After I watched Dr. Forsberg’s presentation, it seemed that not only would FLiBe be a suitable breeding blanket material, it might be the ONLY suitable breeding blanket material for these newer reactor designs that had higher magnetic field density.

If that was the case, as he was pointing out in his lecture, then there would be some remarkable opportunities for joint technology development between those of us who wanted to use FLiBe in molten-salt reactors (Flibe Energy), those who wanted to use FLiBe in molten-salt-cooled, TRISO-fueled reactors (Kairos Power), and those who wanted to use FLiBe in their breeding blankets (pretty much any D-T fusion company). Each approach heats the FLiBe differently, but at the end of the day, each approach ends up with hot FLiBe whose enthalpy needs to be converted into electricity and a minimum of waste heat.

Our lithium-fluoride reactor technology would heat FLiBe through direct fission reactions in the fuel salt. The fuel salt (LiF-BeF2-UF4) would contain fissile material that when properly moderated in a lattice of graphite, would sustain fission reactions that would directly heat the fuel salt, mostly through the slowing-down of fission products, but also a bit through conduction from the graphite moderator (where neutrons are slowing down). The fuel salt would heat the coolant salt (probably also FLiBe or perhaps FLiNaBe) and that heated coolant salt would then heat supercritical carbon dioxide, which would flow through turbomachinery to produce electrical power.

In the flibe-cooled reactor technology (Kairos), the FLiBe salt would be heated by conduction from the hot graphite spheres that contained thousands of coated particles of enriched uranium fuel. There would be a small amount of heating from neutrons, but the overwhelming majority would be heating through conduction. The FLiBe salt wouldn’t contain any nuclear fuels and would pass out of the reactor carrying enthalpy to another salt, probably a nitrate salt, which would then raise steam in a steam generator. That steam would likely drive a supercritical steam turbine to generate electrical power.

In a D-T fusion reactor, a blanket filled with FLiBe would be heated several ways. The largest heating term would come from the slowing-down of high-energy (14.6 MeV) neutrons from the D-T reaction. The intent of the design would be for these neutrons to strike lithium-6 nuclei and fission them into helium-4 and tritium (hydrogen-3). This would produce replacement tritium for the D-T fusion reactor. Each fission of lithium-6 would release another 4.6 MeV into the FLiBe blanket. There would also be a need for neutron multiplication, since not every fusion neutron will end up creating another triton. Beryllium is envisioned as a neutron multiplier in the reactor blanket. A high-energy neutron would strike beryllium-9 and fission it into two nuclei of helium-4 (alpha particles) and release another neutron. Thus the overall process in the D-T reactor blanket is the fissioning of lithium-6 and beryllium-9 to release neutrons, tritium, and helium. The hot FLiBe would then be pumped out of the blanket region and heat another salt, perhaps, in order to raise steam, or perhaps directly heat supercritical CO2. That is apparently yet to be determined for most D-T fusion designs.

Most people do not think about the fusion reactor having a fission blanket, but that is what is really going on. With that high-energy (14.6 MeV) neutron, certain reactions are feasible in the D-T reactor blanket that are a lot less common with the lower-energy neutrons that are born from the fission of uranium or plutonium (2 to 6 MeV). That simple fact may end up being very important.