The notable thing here is that it's a molten salt reactor design, where the fuel is dissolved in a molten salt (FLiBe). This allows online continuous processing of the fuel, unlike with solid fuel rods sealed inside a pressure vessel.
This unlocks a lot of options for the fuel cycle, including the use of thorium.
This work builds on a previous molten salt reactor experiment at Oak Ridge, decades ago. There's a whole lore about MSRs.
MSRs have some attractive features, but they also have significant drawbacks.
The most pressing is that fissionable material is spread throughout the fluid, so fission and decay of fission products is occurring right up to the edge of the fluid. The walls and pipes containing the molten salt, and anything dipped into the salt, are exposed to unmoderated neutrons. One can shield using (say) graphite, but then damage to that (and soaking up of radioactive materials) become issues.
The Molten Salt Reactor Experiment at Oak Ridge was near the end of its radiation exposure lifetime when the program ended.
Contrast this to light water reactors. These are designed so that no lifetime component sees unmoderated neutrons. There's a thick barrier of water between the fissioning fuel and the reactor vessel wall and the support structures for the fuel bundles. The bundles themselves are exposed, but they are replaced for refueling and are not lifetime components.
To add to this, even with the shielding provided by water in light water reactors, the neutron exposure is _the_ limiting factor for the reactor vessel.
The metric to look for is called "DPA" (displacements per atom), the number of neutron collisions that a material can tolerate before losing enough structural integrity to fall below the acceptable limits. The best modern reactor steels are at 150-180 DPA.
And a lot of potentially cool reactors like TWR (travelling wave reactor) end up being logistically impossible because lifetime-limited components will be exposed to multiple hundreds of DPAs.
Many old LWR's have had their reactor vessels heat treated during a maintenance break to undo some of that neutron radiation damage and extend the life of the reactor.
Not sure whether it would be possible to do something similar to a liquid fueled reactor, including all the hot pipework. Maybe, but yet another cost. Notably some of the recent MSR projects propose replacing the entire reactor every now and then (Terrestrial or whatever they were called, not sure if they are still around).
Yes, it's called "annealing". Basically, the core is de-fueled and a huge electric resistive heater is put inside it. Then the entire vessel is heated to something like 600C, and kept there for several days.
It helps the atoms displaced by neutron collisions to "snap back" into the correct places in the crystalline structure. But it can never restore the material completely, and over time the annealing breaks will have to be more and more frequent.
It also can't be used for everything. Some pipes will experience large thermal stresses if annealed, and some components can't be heated properly due to complex geometry.
As with everything in engineering, all problems can be solved with additional complexity. It's possible to design LFTR reactors to be more annealable, but it will likely make them impractically complex.
There are also other issues with LFTRs. A significant part of the energy production will happen _inside_ the pipework carrying the molten salt, as delayed fission happens and daughter products decay. This will cause inevitable problems with the reactor power control.
Modern light water reactors are engineering marvels. They are incredibly compact for the amount of power that they generate, and they are now designed with the anticipated 70-100 year operating lifetime. Getting LFTRs to the same level of maturity might be possible, but it'll require literally hundreds of billions (if not trillions) invested, just like with the classic nuclear.
You are describing "dry" annealing there. This has not been applied anywhere, as it requires removing and reinstalling the internal support structures inside the reactor vessel.
A somewhat lower temperature "wet" annealing process has been applied to two test reactors. I don't think it's been applied to any full scale power reactors.
I wonder if it's possible to run it hot enough for the radiation damage (basically a bunch of dislocations, right?) to just anneal itself out continuously, like how Wigner energy is dissipated in graphite when it's hot enough.
No, this is fundamentally impossible with steel. Annealing works by making the material more "plastic", and this necessarily reduces its tensile strength. Which is the limiting factor for the vessel.
You can make the vessel thicker to compensate, but then you can just make it thicker in the first place and skip annealing.
I think it has a key advantage for China specifically though which is it consumes significantly less water and they have a lot of water poor territory.
The oakridge experiment ended and not a lot of R&D has been done on salt reactors. It makes sense that China is still basically in research and testing phases for molten salts.
Oh dear god, no. Graphite is a very good moderator, it is in no way a shield. Those two properties are (sort of) opposites of each other. Lead makes the cheapest and best shield. Also, those parts that are exposed to neutron flux stay radioactive for about 10 years. So it shortens their lifetime in the reactor but the waste isn't a big issue.
> Oh dear god, no [...] Lead makes the cheapest and best shield.
Oh my, definitely no :-) Do not use lead for neutron shielding. You're thinking gamma radiation but then we're talking apples vs oranges then. You want atoms comparable in size to neutrons, so something with plenty of hydrogen. Think water or PET (plastic) when you don't want water to "leak" when transporting a source. For thermal neutrons maybe PET impregnated with boron. Now neutrons may generate gamma when captured by hydrogen, then you may want some lead for secondary effects like that but I am not sure how strong those are.
Lead is fine for shielding of sufficiently energetic neutrons, which can lose energy to lead by inelastic nuclear collisions. But below the threshold for that lead does very little.
Maybe as a special case then as a thin layer before following up with water or PET, or PET impregnated with boron. But would also need an extra layer following it for secondary gamma emission from neutron capture.
Lead is essentially useless as a shield for neutrons that are below the minimum excitation energy of a lead nucleus. Elastic neutron collisions with lead leave the neutron energy essentially unchanged.
I assume this is why an alloy of lead is used in practice. Still doesn't change the fact that graphite is a moderator not a shielding material. Also, structural materials in reactors are usually invisible to neutrons and a sandwich of materials is often used. Different layers do different things. Usually, one layer of shielding and one layer of a material that isn't impacted (much) by neutron flux for structural strength.
There is a rabbithole for almost all of these material choices, especially in nuclear. Not going down that rabbithole in a discussion targeted at folks who don't spend their lives working in nuclear doesn't make that person wrong. It makes them an effective communicator.
PS Lead is a very very common shielding material in nuclear.
A moderator is a neutron shielding material, since it removes energy from the neutrons. That's what moderation is all about. Water is a much better moderator, but graphite still performs the function.
It breeds thorium to fissionable uranium from a starting fissionable uranium starter fuel. It doesn't directly use thorium for fuel.
What people need to understand about the cycle efficiency is that when you mine uranium, the fissionable part of uranium (U-235) is only 1% of that uranium, the rest is nonfissionable U-238.
Thorium is about twice as abundant as Uranium (all isotopes). The MSR uses Thorium to create U-233, a fissionable but not naturally occurring Uranium isotope.
So the "unlimited energy aspect" is that about 200-300x more breedable Thorium exists than fissionable U-235.
A MSR nation could also try to breed U-238 into plutonium, which would provide another 100x more breeding stock, although LFTR never talked about U-238 breeding. IIRC the plutonium may be difficult to handle because of gamma rays, but I don't recall exactly.
While I don't have confidence that even LFTR/MSR reactors can get economical enough to challenge gas peakers, it may be possible to make truly price-competitive MSR electricity with the right modular design. I wish the Chinese the best of luck, because if they do it will spur the rest of the world to adopt this about-as-clean-and-safe-as-it-gets nuclear design.
China has thorium, and while less than others [1], it’s better than they do with uranium [2].
> it may be possible to make truly price-competitive MSR electricity with the right modular design
Yes. But probably not in the near term with thorium. This isn’t designed to be cheaper. It’s designed to be more available to China than being dependent on Russian deposits.
Eh, U-235 is .7%, not 1%, but also U-238 can be bred into Plutonium. What makes Thorium interesting -besides its abundance- is that U-233 is very difficult to work with, so proliferation concerns are mitigated.
Thorium 232 is the thorium in the cycle yes. And all kinds of nonsense is correct for the daughter products. But in general, to actually use do anything with thorium you need excess neutrons.
Even the daughter uranium 233 only produces on average 2.48 neutrons per fission, so it’s very difficult even in a combined lifecycle process to have enough - thorium doesn’t produce uranium 233 immediately (takes almost 30 days), neutron capture with that low a ratio requires a LOT of thorium, which is going to mostly just suck up all neutrons and you won’t have any extra for addition uranium 233 fissions, etc.
It’s quite difficult (impossible?)to have actually work without a source of a large amount of additional neutrons.
> to actually use do anything with thorium you need excess neutrons
Unless 100% of those neutrons is being absorbed by the thorium, this means you'll have neutron flux at the boundary. Which, for a liquid moderator, means all the pipes and tanks and pumps.
Sure, if you ignore all the parts of the neutron economy that make it possible to work. The part everyone missed in this discussion is that all of the numbers of neutrons (and their barns) aren't constants. Since the fuel is a fluid, you can use density and shape to improve the neutron economy in the reactor core. Basically, when the atoms are closer together, the economy improves. You can also use a better moderator like graphite since the basic design is safer and the rate of fission is just easier to control.
And considering that people made these things work 60 years ago without modern computers, the idea that its impossible or needs 40 years of research seems pretty far fetched. What is left of the nuclear industry wants to build current designs like the AP1400. That is a great idea, but there are things you can do with a LFTR that you can't do with an AP1400. The biggest of them is making synthetic fuel. The other advantages are the amount of waste produced and the fact that you can make a LFTR into a waste burner consuming the spent fuel rods from a AP1400. The downside is you actually have to fix nuclear regulations to do this and getting politicians to do that has proved impossible.
There are no technological barriers, this is entirely political.
> That you’re even discussing graphite moderated (?!!) makes this pretty clear.
And why would this be? Is graphite expensive? No it isn't. Also, we created a working one of these designed in the 1960's without computers. You seriously think this is hard compared to other types of engineering we do today?
A LFTR can also do things that a PWR or BWR can't and has several major advantages. But since it uses pencil lead apparently we can't even try it.
Because it has dangerous behavior in real reactors due to the void co-efficient behavior, to the point of… being the cause of the largest nuclear disaster in recorded history?
Not OP but he maybe referring to the new gas cooled gen 4 reactors not Soviet RBMKs. The ones I heard are working with sealed beads of uranium, encased in porous carbon, then some other layers, including some carbide (silicon?). The porosity of carbon absorbs gases but they ultimately stay sealed. The whole thing is helium cooled.
Yeah. I was listening to David Ruzic's video [1] about them getting one of those reactors on campus and when he showed the structure of beads, that's the first thing that popped in my head - at that size how are they going to ensure every single bead has an intact surface.
This unlocks a lot of options for the fuel cycle, including the use of thorium.
This work builds on a previous molten salt reactor experiment at Oak Ridge, decades ago. There's a whole lore about MSRs.