Liquid Fluoride Thorium Reactor

[mheslep]

I did find this TED video interesting from lftrsuk' links above, by a couple of MIT nuke eng's on they are calling the "WaMSR", a molten salt designed to burn spent fuel (UOx?) instead of thorium. Good idea from a political / marketing stand point as it plays on the desire to get rid of nuclear waste.

https://www.youtube.com/watch?v=AAFWeIp8JT0

They address one of the advantages discussed above in this thread: the Zirc Alloy metal cladding used in solid fuel reactors has a short life (4 years tops) which forces replacement and limits burnup, increasing the waste stream. Ok, great. But in an MSR, at some place the critical portion of the salt still has to be contained by some solid material (graphite?), that solid material will undergo a high flux and over time have to be replaced. Is this not moving the problem from one place to another? Perhaps the advantage of MS over solid Zirc rods is that, while the graphite (?) moderator might require replacement, the liquid fuel does not and can continue burn up? Can such a moderator be replaced without replacing essentially the entire reactor vessel?

 

[zapperzero]

Can such a moderator be replaced without replacing essentially the entire reactor vessel?

Yes you can. Although, if you fancy fishing stuff out that soup, you're a braver person than I am.

LFTR is, to my mind, a profoundly stupid, dangerous idea and so's any other liquid fluoride salts based scheme. You have a highly corrosive coolant that explodes if it comes in contact with water and burns if it comes in contact with air. Pair that with a burnable moderator. Now imagine what a large-break LOCA looks like.

I could only envision this being safe if it was built on the far side of the moon or something like that, a friendly place that's very cold by default and has no oxygen or water around.

And all that money and brainpower is beng thrown down the drain because lead-moderated, lead-cooled is Not Invented Here. EDIT: and by "here" I mean in the US.

Here, have a peek at the near future.
http://myrrha.sckcen.be/

 

[etudiant]

Yes you can. Although, if you fancy fishing stuff out that soup, you're a braver person than I am.

LFTR is, to my mind, a profoundly stupid, dangerous idea and so's any other liquid fluoride salts based scheme. You have a highly corrosive coolant that explodes if it comes in contact with water and burns if it comes in contact with air. Pair that with a burnable moderator. Now imagine what a large-break LOCA looks like.

I could only envision this being safe if it was built on the far side of the moon or something like that, a friendly place that's very cold by default and has no oxygen or water around.

And all that money and brainpower is beng thrown down the drain because lead-moderated, lead-cooled is Not Invented Here. EDIT: and by "here" I mean in the US.

Here, have a peek at the near future.
http://myrrha.sckcen.be/

This is not correct.
Liquid fluoride salts are essentially inert in air. I worked with them.
There is not enough reactivity in oxygen or nitrogen to displace the fluorine from the salt.

There is perhaps confusion between liquid fluoride salt cooling and sodium cooling.
The latter does indeed tend to explode on contact with water and does burn or at least oxidize very rapidly, with lots of heat, on contact with air, but molten fluorine salts don't.
Separately, the Soviets did deploy lead bismuth cooled reactors on a nuclear submarines, but found them to be a maintenance headache.

 

[zapperzero]

This is not correct.
Liquid fluoride salts are essentially inert in air. I worked with them.
There is not enough reactivity in oxygen or nitrogen to displace the fluorine from the salt.

You've worked with uranium tetrafluoride?? Fine. I must have been imagining things. My apologies to one and all.

 

[etudiant]

You've worked with uranium tetrafluoride?? Fine. I must have been imagining things. My apologies to one and all.

I stand corrected.
Uranium tetrafluoride is indeed nasty stuff, unlike the more stable fluoride salts I've had dealings with.

 

[mheslep]

Though UF4 is toxic, neither the molten salt proposed for the reactor or UF4 alone is explosive in contact with air or water.

http://ibilabs.com/UF4-MSDS.htm

 

[mheslep]

Yes you can. Although, if you fancy fishing stuff out that soup, you're a braver person than I am...

Can you elaborate as to how? To my knowledge it was never attempted on the MSR reactor back in the 1960s.

 

[zapperzero]

Can you elaborate as to how? To my knowledge it was never attempted on the MSR reactor back in the 1960s.

You can build handles or notches into the moderator blocks and move them around with a crane, like they do now with fuel elements. It's relatively easy, mechanically speaking, because you know exactly where they are and you can use sonar if you don't. But what to do with them after you've lifted them out? What if the crane breaks or jams, midway through?

In designs where fuel circulates through channels dug in the moderator, it's "a bit" more complicated.

I don't think graphite would be used, pyrolitic carbon more likely, ideally coated in something that is less porous (although it may get electroplated all by itself, I don't know).

 

[Astronuc]

Though UF4 is toxic, neither the molten salt proposed for the reactor or UF4 alone is explosive in contact with air or water.

http://ibilabs.com/UF4-MSDS.htm

From the website:

UF4 can be readily converted to either uranium metal or uranium oxide. UF4 is less stable than the uranium oxides and produces hydrofluoric acid in reaction with water; it is thus a less favorable form for long-term disposal. The bulk density of UF4 varies from about 2.0 g/cm3 to about 4.5 g/cm3 depending on the production process and the properties of the starting uranium compounds.
Chemical Properties
Uranium tetrafluoride (UF4) reacts slowly with moisture at ambient temperature, forming UO2 and HF, which are very corrosive.

I've been in conversion shops where UF₆ is hydrolized to UO₂F₂ at about 100 C. It also reacts with steam, which is the basis of the 'dry conversion' process. As far as I know, Th fluoride behaves similarly.

 

[mheslep]

You can build handles or notches into the moderator blocks and move them around with a crane, like they do now with fuel elements. ...

That assumes some kind of open top reactor, i.e. solid fuel and water cooled. I don't see how that can be done with a molten salt reactor.

 

[zapperzero]

That assumes some kind of open top reactor, i.e. solid fuel and water cooled. I don't see how that can be done with a molten salt reactor.

Why not? You could have a bucket of molten fluoride salts which keeps hot via fission, with a heat exchanger loop (FLiBe maybe?) running through and pylons made of carbon bricks stacked on top of each other for moderation. You need an inert atmosphere on top, but other than that, what's to keep you from also hanging a crane above the bucket and wrapping the whole package up in a concrete biological shield, like some demonic chocolate egg?

 

[wizwom]

Small cores require higher enrichments, and that's problematic with respect to control/kinetics. The shielding would becomes disproportionately large for small cores. I believe that an LFTR is even more complicated because of the need for a feed and bleed system, which is outside the core, and the need to deposit the fission products in some stable form.

The criticality of a molten salt reactor is controlled by varying the concentration of fissile to moderator, that is, tweaking the k-infinity of the reactor, rather than the control mechanism of a solid element reactor, where you tweak the probability of non-leakage.

Looking at ORNL's report (By L.G. Alexander), though, they are currently steering toward a system where the moderator is separate from the salt; this, of course, is a poor choice. If one uses MgF2 salt as the moderator (about on par with water in moderation) one could do a wholly homogeneous reactor.

To breed, per Lietzke & Stoughton 1957, atom ratios of 17 Mg per Th and 105 F per Th (inclusive) would be needed. This would be a molar ratio of 12.3 MgF2 to 1 UF4.

The scalability issue is that any molten fuel means you are pumping subcritical fissile fuel through your heat exchangers. But if you want to design for higher power, you need larger heat exchangers. The size of each heat exchanger is limited by the need to remain highly subcritical even at your expected highest breeding level. Similarly for pipe size. So, I would imagine a gigawatt range LFTR to have a large number of ~30 cm pipes going to rather small heat exchangers (once-through would be fine, since you don't need to worry about the possibility of over-heating the primary loop). Whether you use the heat exchangers as a NSSS or a brayton cycle heater is immaterial (although a closed Brayton is a definite necessity, there will be fission occurring in all the piping for the molten salt, and thus activation of everything within about a foot of the fluid).

 

[etudiant]

Does the LFTR stability depend on the size of the fuel pool?
It seems logical that a gigawatt unit would be swimming pool sized, so the temperature and the concentration of the fuel might vary materially depending on where in the pool the measurements are taken, even if the fuel is getting pumped past heat exchangers. That seems difficult to control accurately. Is this a concern?
More generally, it is clear after Fukushima that simply meeting a 'design basis' spec is not enough, it is important to have a sense of the possible consequences for a beyond spec accident.
In the case of the various national breeder programs, the accidents that discouraged their proponents were fortunately not catastrophic. The LFTR proponents would enhance their case if they would subject their designs to very critical scrutiny, so that the public gets confidence that hostile eyes have not found cause for alarm.

 

[zapperzero]

Does the LFTR stability depend on the size of the fuel pool?

Size and geometry. Temperature also matters, indeed, and so does the homogeneity of the mix, which is by no means guaranteed.

 

[etudiant]

Size and geometry. Temperature also matters, indeed, and so does the homogeneity of the mix, which is by no means guaranteed.

Thank you for this feedback.
Is it possible to expand on this issue a bit more?
It seems, afaik, a large pool of a 1000*C mixture of thorium fluoride, with substantial amounts of uranium and other transmutation products, where reaction speeds are muted if the temperature rises too much.
Clearly drain plugs are not going to work fast, so preventing excursions, a core requirement, must rely on the thermal effects on reaction rates.
How well proven is that for a range of radionucleide mixtures? Is there a risk of the salt getting vaporized in an excursion?

 

[mheslep]

so preventing excursions, a core requirement, must rely on the thermal effects on reaction rates.

Which clearly they do, right? A substantial expansion of the fluid from heat, much less a vaporization, would cause the area to drop below critical. For the salt to boil, an area would have to somehow rise ~971degC above the freeze plug.

 

[etudiant]

Which clearly they do, right? A substantial expansion of the fluid from heat, much less a vaporization, would cause the area to drop below critical. For the salt to boil, an area would have to somehow rise ~971degC above the freeze plug.

That is the question.
It is not clear to me that a large volume of molten salt would respond quickly to an overtemperature.
Certainly a freeze plug mechanism will take several seconds to work even in a small reactor.
That is an eternity in terms of reaction time.
So the issue is what are the faster acting self limiting elements of the fuel mix and how does this translate to operational management. Is there a risk of prompt excursions in this system?

 

[mheslep]

That is the question.
It is not clear to me that a large volume of molten salt would respond quickly to an overtemperature.
Certainly a freeze plug mechanism will take several seconds to work even in a small reactor.
That is an eternity in terms of reaction time.
So the issue is what are the faster acting self limiting elements of the fuel mix and how does this translate to operational management. Is there a risk of prompt excursions in this system?

The freeze plug would not be instantaneous, but the coefficient of expansion of the liquid salt is ~instantaneous, and so in turn is the reaction rate which is based on density (negatively).

 

[etudiant]

The freeze plug would not be instantaneous, but the coefficient of expansion of the liquid salt is ~instantaneous, and so in turn is the reaction rate which is based on density (negatively).

Thank you for the clarification.
Does this mean that the reaction only stops once the molten salt vaporizes?
Or is there a negative trend as the temperature of the salt rises?

Is there a solid reference which discusses these issues in the context of a review of operational considerations for a MSTR?

 

[zapperzero]

Or is there a negative trend as the temperature of the salt rises?

This.

Of course this doesn't address the issue of how to actually stop the reaction if you feel like it.

 

[mheslep]

I don't follow. Under positive control an operator removes the fluid from the moderator area (graphite i believe?) and thus stops the reaction. If there's failure of control, the operator stops active cooling of the freeze plug (assuming that has not already happened), again the fluid leaves the moderator area and the reaction stops.

 

[mheslep]

Thank you for the clarification.
Does this mean that the reaction only stops once the molten salt vaporizes?
Or is there a negative trend as the temperature of the salt rises?

Is there a solid reference which discusses these issues in the context of a review of operational considerations for a MSTR?

From the original Oak Ridge MSR work, Fluid Fueled Reactors:
As the salt density falls with increasing temperature, reactivity falls: (1/k) dk/dT ~= -3.8 X 10^-5 / °F
See pg 640-642 here:
http://www.energyfromthorium.com/pdf/FFR_chap14.pdf
If you are inclined there's more here:
http://energyfromthorium.com/pdf/

 

[etudiant]

From the original Oak Ridge MSR work, Fluid Fueled Reactors:
As the salt density falls with increasing temperature, reactivity falls: (1/k) dk/dT ~= -3.8 X 10^-5 / °F
See pg 640-642 here:
http://www.energyfromthorium.com/pdf/FFR_chap14.pdf
If you are inclined there's more here:
http://energyfromthorium.com/pdf/

Hi mheslep,
Thank you for the information and the very helpful references.
The reports, while very informative, are unfortunately more focused on feasibility and economics than on divergences from expected operations. As these are somewhat science advocacy documents, that is not surprising.
As an uninformed observer, it does worry me that the reactivity merely falls with density, because the nuclear reactions are so much faster than any change in density could be. It suggests that local excursions are not ruled out, even if the negative coefficient does preclude a Chernobyl type factor of 1000 power surge.

 

[mheslep]

Hi mheslep,
Thank you for the information and the very helpful references.
The reports, while very informative, are unfortunately more focused on feasibility and economics than on divergences from expected operations. As these are somewhat science advocacy documents, that is not surprising.
As an uninformed observer, it does worry me that the reactivity merely falls with density, because the nuclear reactions are so much faster than any change in density could be. It suggests that local excursions are not ruled out, even if the negative coefficient does preclude a Chernobyl type factor of 1000 power surge.

Could you illustrate by showing how such an excursion is ruled out with a traditional pressure water solid fueled reactor? Clearly control rods insertion is also not instantaneous.

 

[etudiant]

Could you illustrate by showing how such an excursion is ruled out with a traditional pressure water solid fueled reactor? Clearly control rods insertion is also not instantaneous.

Am no expert, but afaik, in conventional reactors, the fuel is in fixed arrays, so the evolution of the nucleides can be allowed for.
In a large pool of thorium fluoride gradually transmuting to U233, it seems at least possible for gradients to form with potentially quite different fuel concentrations and compositions.
I would like to have some idea of how the system would react to such changes in nuclear geometry.
Given that we have had bad experiences with interrupted cooling flows (Fermi reactor most notably) it is reasonable to consider the effect of loss of mixing in the MSTR beforehand. After all, when there is a lot of nuclear material in a small volume, as is the case for the MSTR, belt and suspenders engineering must be the minimum requirement.

 

[wizwom]

Am no expert, but afaik, in conventional reactors, the fuel is in fixed arrays, so the evolution of the nucleides can be allowed for.
In a large pool of thorium fluoride gradually transmuting to U233, it seems at least possible for gradients to form with potentially quite different fuel concentrations and compositions.
I would like to have some idea of how the system would react to such changes in nuclear geometry.
Given that we have had bad experiences with interrupted cooling flows (Fermi reactor most notably) it is reasonable to consider the effect of loss of mixing in the MSTR beforehand. After all, when there is a lot of nuclear material in a small volume, as is the case for the MSTR, belt and suspenders engineering must be the minimum requirement.

The LFTR idea is that the U233 is controlled by gassifying the Pa233 stage, removing the breeding wait from the active reaction mass, and then returning it after it becomes U233 as the reactor needs it.

 

[etudiant]

The LFTR idea is that the U233 is controlled by gassifying the Pt233 stage, removing the breeding wait from the active reaction mass, and then returning it after it becomes U233 as the reactor needs it.

You are suggesting the LFTR design envisages bubbling up Plutonium vapor for recycling after it decays back to U233?
This is news to me.
Imho, it does not seem a good idea.

 

[Astronuc]

You are suggesting the LFTR design envisages bubbling up Plutonium vapor for recycling after it decays back to U233?
This is news to me.
Imho, it does not seem a good idea.

Higher order fluorides, UF₆, are volatile. In the gaseous diffusion and centrifuge enrichment processes, UF6 gas is used as a carrier from which U(235)F₆ is separated from U(238)F₆. Similarly, different fluorides have different stability domains and volatilies, so one tailors the process to favor a particular element. One would take advantage of differences between PaF₄/PaF₅ and UF₄ (Boiling point: 1417°C) / UF₆ (Boiling point: 56.5°C).

The element is a dangerous toxic material and requires precautions similar to those used when handling plutonium. Protactinium is one of the rarest and most expensive naturally occurring elements.

http://www.webelements.com/protactinium/

The attraction of the Th-based fuel cycle is the lack of transuranic elements, although some quantity of U-235 or Pu-239 is required to initiate a Th-based system.

 

[wizwom]

You are suggesting the LFTR design envisages bubbling up Plutonium vapor for recycling after it decays back to U233?
This is news to me.
Imho, it does not seem a good idea.

Protactinium, not Plutonium. A LFTR never gets to any significant amount of Plutonium.
The chain is 232Th->233Th->233Pa->233U->fission
The 233Pa has an absorption cross section about 14 times that of 232Th, so you want to get it out of the way of neutrons if you can, and LFTR does exactly that as the molten salt passes through the flouridizer.

 

[etudiant]

Thank you very much, Astronuc and wizwom. Very helpful input.
That even the initial LFTR design prototype included a fairly capable fuel reconditioning element to remove undesirable fission products is entirely logical, but a new wrinkle to me.
It is certainly not a much discussed feature of this class of designs.

 

[zapperzero]

It is certainly not a much discussed feature of this class of designs.

I dunno. I harp on it every chance I get. "A reprocessing plant near every power station! La Hague in your own back yard!" etc etc

 

[mheslep]

As far as I know most of the reticence about reprocessing comes about from the fact that Plutonium processing goes on with U235 fuel cycles. That's not an issue with a Thorium fuel cycle.

 

[wizwom]

I dunno. I harp on it every chance I get. "A reprocessing plant near every power station! La Hague in your own back yard!" etc etc

Except its not "a reprocessing plant" - its an integral part of the reactor, and it never ships fuel out, and, in fact, should have trivial amounts of waste flow (just the fission products, about 1 gram per MWd).

And on the plus side: no "spent fuel" to store. No refueling shutdowns.

 

[zapperzero]

Except its not "a reprocessing plant" - its an integral part of the reactor, and it never ships fuel out, and, in fact, should have trivial amounts of waste flow (just the fission products, about 1 gram per MWd).

And on the plus side: no "spent fuel" to store. No refueling shutdowns.

Oh it is a reprocessing plant, only it's co-located with the reactor and is integral to its functioning, unlike current reprocessing plants.

Are you including gasses in your gram/MWd?

 

[mheslep]

1 MW-day/gram is ~100% burn up of a fissionable fuel, so yes, IF 100% is achievable w/ a LFTR, 1 gram includes the mass of all fission products. Natural thorium is ~100% Th232, the fertile isotope.

By contrast, a traditional 5% LEU reactor would produce the same mass of fission products per unit energy, but 20X the waste mass at 100% burnup. However, as I understand the current process, largely because of fission product poisoning solid fuel reactors typically achieve on the order of 10% burnup, adding another 10X of waste mass (not fission product) per unit energy. So we might expect a LFTR to produce 200X less waste than a traditional LEU solid fuel reactor.