Energy From Thorium Discussion Forum

Since we're doing crazy-but-fun ideas this week, here's another one.

From Wikipedia's fluoride volatility article, the table of fluoride boiling points says

LiF 1676 C
ThF4 1680 C

A 1958 report gives an equation for the vapour pressure: Log(P/atm) = 7.940-15,270/T (P in atmospheres, T in Kelvin)
equivalently: Ln(P/Torr) = 24.916 - 35,160/T (P in Torr=mmHg, T in Kelvin)

The CRC handbook (75th ed) gives a table for LiF vapor pressures, so we can compare:-

Code:

Pres/mmHg         BPt LiF          BPt ThF4
400                1591              1585
100                1425              1458
10                 1211              1282
1                  1047              1139

ThF4 is higher boiling at low pressures, but that is less of a problem than its melting point of 1110 C. A ThF4 still would have to run above 1200 C to give some margin against freezing, so ORNL could never have considered it for Hastelloy equipment. However, we've had lots of discussions about C-C composites or tungsten for the reactor itself. The still is not subjected to neutron damage or the regs for critical systems, so it should be easier. The best material would probably be a B4C/boron fibre-carbon composite, for criticality reasons. If it could be constructed, a ThF4 still would enable the following reprocessing scheme for a 1 or 1.5 fluid reactor:-

1)Fluorinate out the uranium, and hopefully some of the plutonium as well.

2)If there's much TRU's left, electrolyse them out as in the IFR scheme. Some Th will probably get deposited with the TRUs, but it's all going back to the core, so it doesn't matter. This step is to avoid a criticality risk in the still, so if that could be avoided another way, it is not necessary.

3)Distill the LiF/BeF2 and ThF4 away from the fission products. Some cadmium (but the yield is low) and cerium (but the Xsec isn't too bad) may distill over too, but not much.

4)Now the TRUs are more concentrated, take another pass at them in the electrolyser, using some cheap salt mix like NaF/KF as solvent. Dispose of the fission product mix, together with some residual Th. In 500 years time it will be just another mixed Th/rare earths ore, if anybody want to bother digging it up again. (OK, not quite, but sorta...)


Even crazier scheme

UF4, Bpt 1417 MPt 1036, is actually easier to deal with than ThF4, on a physical properties basis. The difficulty, of course, is that the distillate is pure 233UF4, so you don't want more than a few litres of it in one place. The vapour line/condenser/liquid take off pipework would have to be made to be definitely sub critical even if it got plugged full of frozen UF4 from end to end - hence wanting to make it out of boron composites. Carbon-carbon on its own would be unhelpful!

If a UF4/ThF4 still were possible, you could do without the fluorinator/reducer kit, and its possible corrosion problems, as well as the bismuth liquid extraction system. There are some fission products that might be hard to get rid of with just distillation, but the severe neutron poisons like Sm and other heavier lanthanides will all go, and the rest might be tolerable, or could be removed on a very slow schedule by smaller and so cheaper equipment. Distillation is well understood, simple, reliable, quick, and CHEAP, compared to any reactive process. For a CR=1.0 reactor, it might be enough, 1 fluid or two.

Now, what makes this impossible?
(edited to mark the table as 'code' to force monospace font)
(and again to fix typo in ThF4 vapour pressure equation, 25,160 fixed to 35,160)

What make it impossible: materials!

Theoretically sounds good, but practically:

Today‘s top materials: nickel alloys (Hastelloy, Monel, Inconel etc.) are limited by temperatures around 700 °C (for water steam!). Other materials (boron composites etc.) are still in R&D.

The distillation pot is well understood, simple, reliable and cheap, but:
• You don’t have a reliable material for temperatures around 1000 °C and corosivity of fluorides to make even a pipe (don’t speak about heat exchangers, valves etc.)
• you have to keep all parts above 1100 °C or it gets jammed
• if there is a little bit of H2O ( = humidity in the air) it gets jammed (and only thing you can do is to throw it away)
• you have to cool it with gas (water isn’t good idea) it means that it will be quite big

All this plus temperatures more then 1000 °C plus its toxicity and radioactivity means that it wouldn’t be SIMPLE, RELIABLE and CHEAP but if we’ll find suitable material - it will be POSSIBLE (maybe ).

Why couldn't the still be machined from a block of graphite doped with boron?

How about a small still? The heat comes from the radioactive decay of the fission products, which is pretty intense if you get them in the first year or so. The problem would be supporting the graphite, since it's fairly conductive.

Hot fluorine I understand to be very corrosive. I wasn't under the impression that fluorides were so corrosive.
The ORNL work had a challenge with plugged fuel lines but I don't recall any discussions of corrosion in the still.

No one, can machine valves, measuring and regulation devices and other more complicated devices at this moment. Or for example how to make a conection of two graphite tubes? etc.



I doubt that heat from decay can hold the temperature constantly above 1100 °C. Acctually I am not sure that it can even reach this temperature.

Fluorides are no as corrosive as gasseous fluorine, but there are still corrosive. First problem is in residual fluorine from fluorination. Second problem is temperature - corrosive damage grows with temperature.

If residual fluorine is a problem, it can be removed by adding something for it to react with. ORNL added berrylium metal for redox control, lithium or thorium would also work, and we've discussed controling the reactor redox potential electrochmically, so why not the feed tank for the still?

Only the still base has to be near-pure ThF4, and so held at ~1200C. If you feed a still pot of ThF4 with fuel salt - FLiBeTh or FLiTh for fast neutron designs - the still base composition will adjust so that the distillate will have the same composition as the feed, less the low volatility components (fission products). From the volatility data, that composition will be between 80% and 90% ThF4 by moles (~99% by mass, as its 232Th vs 7Li). The distillate can then be handled at ~600-700C, as can connections and plumbing.

Connecting graphite tubes should be fairly straight forward, using Grayloc-type mechanical couplings: the only requirement is that the usual steel product material be replaced by a high-temperature one, such as tungsten.

For pumps, it has been suggested here that electro-magnetic devices installed outside piping could be a viable alternative to complex mechanical gizmos with labyrinth seals, etc.....

For valves, since we're proposing ambient-pressure fluid circuits, one should be able to replace difficult mechanical devices with simple things like siphon breakers and freeze valves.

Conclusion: With a bit of thoughtful design effort, high-temp salt operation should not be a show-stopper by any means !

I don't think it can be done at ambient pressure, that would require 1680C just to boil the LiF!. However, that doesn't mean you need to get a vacuum seal on red-hot graphite. The still doesn't have to be very big, the processing rate is only 1 m^3/day per GW(e) for thermal designs, and less than 100 l/day for fast designs. The still, a furnace (induction?) to heat it and some insulation around that would all fit in 10m^3. You can put a vacuum vessel, made of Hastelloy or maybe even 316 stainless steel as it need never be exposed to very high temperatures and can be cooled on the outside, around the whole thing, and it would be no bigger than commonplace industrial equipment. The vacuum vessel would also form the initial barrier to the escape of radioactive material

I believe ORNL used freeze valves to separate the fuel tank from the vacuum side - they did not mention any problems with pressure leakage from the fuel tank. In fact, it doesn't even seem to me you need a freeze value, a simple plumbers trap should do even if the salt is still molten.

As far as the pump goes, I think they used pumps operating at mundane temperatures to move inert gases to accomplish the pumping. So the pumps saw neither the LiF gas nor the 1000C temperatures.

Luke,

The use of vacuum distillation as described in the ORNL reports is to vaporize the relatively low boiling point LiF and BeF salts at approximately 1000C and condense them for reuse while discarding the the still bottom containing fission products. Uranium has already been removed at this stage by fluorination, the noble metals by plate-out and thorium shouldn't be there in the first place in a two-fluid design.

AFAIU, what you are proposing is to allow distillation of the core salt of a single fluid reactor by using higher temperatures to vaporize the thorium, too. This would probably have the effect of vaporizing some of the fission products, too, and therefore require condensing it at multiple condensation temperatures to separate them from the thorium. This is a full-fledged fractional distillation column with multiple outputs rather than a simple pot-and-condenser setup. As you point out, if you go through all this trouble you can probably skip the uranium fluorination step by simply distilling the UF4 along with everything else and mix it back into the salt while carefully avoiding criticality.

Is this a good description of what you are proposing?

Perhaps there are fundamental reasons why this cannot work. Certain fluorides might form inseparable azeotropes, for example. Or maybe the separation equilibrium achievable by distillation is just not good enough. Perhaps this is possible from a physical point of view but was considered by ORNL and rejected as impractical because of complexity and material engineering issues. If that is the case, it is quite possible that material technology has since advanced enough to make it feasible.

Boiling points of thorium and the lanthanides are so high that the distillation may be difficult. You just have to separate them electrolytically. This could be done after other components including Lithium and Beryllium have been removed reducing the quantities.

First, you don't need to boil all the lanthanides. You can discard them with the still bottom. You only need to boil the highest boiling material you want to keep i.e. the thorium (B.P. 1680C at 1atm, lower in vacuum)

Could you be more specific about what is particularly difficult here?

Building a furnace for these temperatures is off-the-shelf technology. Building the still structure from some carbide composite is more or less off-the-shelf technology. Coating it with CVD graphite or diamond is off-the-shelf technology. Keeping a thermal gradient across this structure with sub-degree accuracy is just heating elements, pyrometers and control loops.

Distillation will be done at infinitesimally slow rates compared to, say, an oil refinery. With an oversized still and very low throughput you can probably use a radiative heat sink. No plumbing for cooling the condenser. No valves. No pumps. Just gravity and heat. The still stack doesn't even need to be perfectly gastight. Any vapor leaking will condense on the radiative heat sink around it and drip back to the bottom.

I'm starting to think that this is significantly simpler than handling hot fluorine gas for the uranium fluorinator!

The distillation of the actual MSRE fuel salt is described in ORNL-4577 (3.2 MB pdf in Kirk's archive). The fission products they expected to have most difficulty separating from LiF were cerium, promethium, yttrium and strontium. For all of them, the separation factors found were much worse, buy at least one order of magnitude, than they had expected from their previous work on measuring relative volatilities. They attributed the discrepancy to sample contamination, but as they were unable to repeat the experiment, there is no way to be sure.

If the measurements for this one run are correct after all, the relative volatlities are ~100 against LiF, or about 10 against ThF4. That would require 2 stills, one to distil ~80-90% of the material out for return to the core, and a 2nd one about 1/5th the size to finish the job, returning its distillate to feed the main still and discharging something like 66% ThF4-33% fission products to waste. If their earlier data is correct, a single still is probably adequate. The lanthanide fluorides all boil at least 500C above ThF4, and we can afford to throw out some Th. It is much cheaper than 7Li. I don't think we need real fractional distillation - and if we do, then the idea is not workable. Given the extreme temperatures and radiation, the equipment has to be simple to build, and have nothing that can go wrong or need fixing.

ORNL did not cool their condenser - in fact they had to heat it, so i suspect radiative cooling will be plenty, especially as T^4 scaling will make the losses twice as high for the extra 300K I am proposing. If we want active cooling, I would use direct contact of the vapour with molten salt (FLiBe or FLiTh) at 500C-600C pumped up a pipe inside the vapour outlet pipe from the still and sprayed onto the inner wall. The simplest system would just take a spur off the primary salt loop itself, probably after the heat exchanger, use it to condense/cool/dissolve the still output, and feed it back into the core. The flow would be comparatively tiny, a few litres per minute, or a few parts in 10,000 of the primary loop flow. Such a system would ease the worry over accidental criticality of UF4, by ensuring it was diluted as it was condensed.

Any idea of the boiling point for UF4 or PuF3 or ThF4 under low pressure?