Energy From Thorium Discussion Forum

It is currently Jan 20, 2018 11:56 am

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 25 posts ]  Go to page 1, 2  Next
Author Message
PostPosted: Aug 21, 2010 2:23 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Anyone looked into chlorides vacuum distillation? Seems like it would be much lower temperature, according to the Wiki on chloride volatility (incomplete though):

http://en.wikipedia.org/wiki/User:JWB/C ... volatility

ThCl4 and UCl4 boil at under 1000 C even at atmospheric pressure. NaCl, arguably the most likely carrier salt, has a much higher boiling point, the question is whats the volatility of NaCl, the most common salt in the world and I couldn't find it :oops:


Top
 Profile  
 
PostPosted: Aug 21, 2010 5:24 am 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
NaCl - melts 801C, boils 1413C (at 1 atm) or 865 C (1 mm) - lower than LiF 845C mp, 1676C bp.

KCl - melts 770C, boils 1420

RbCl - 718/1390

All possible.


Top
 Profile  
 
PostPosted: Aug 21, 2010 8:06 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
865 C, quite reasonable! I guess this also makes MgCl2 more attractive, being more volatile than NaCl.

What about the lanthanide chlorides under high vacuum, will they stay put nicely in the still bottoms?


Top
 Profile  
 
PostPosted: Aug 21, 2010 4:55 pm 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
Interesting paper on distilling chlorides, in context of IFR reprocessing. The lanthanide chlorides will stay down, though not quite as cleanly as the fluorides against LiF.


Top
 Profile  
 
PostPosted: Aug 22, 2010 6:14 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Luke wrote:
Interesting paper on distilling chlorides, in context of IFR reprocessing. The lanthanide chlorides will stay down, though not quite as cleanly as the fluorides against LiF.


Very interesting, thanks! How does this compare to thorium tetrachloride volatility?

PuCl3 looks like trouble, it will stick with Pr, Ce and La. UCl3 looks better, though UCl4 will boil much easier. The vapor pressure of UCl4 is in fact so high that it may present a problem for higher temperature operation, so I wonder if we can operate close to the eutectic point with margin. Perhaps as low as 400 C minimum temp. Visocosity may be a bit of a problem there. Corrosion is more severe with high UCl4/UCl3 ratio but lower temperature operation and a bit of carbon come to the rescue. Interestingly there is a low melting NaCl PuCl3 ThCl4 eutectic. One could start this up and reprocess not at all, until the Pu burns out. According to Taube, chloride reactors can operate at CR>1 with very large fission product inventories.


Top
 Profile  
 
PostPosted: Sep 06, 2010 9:34 pm 
Offline

Joined: Dec 14, 2006 1:01 pm
Posts: 380
Pa(V)Cl boils at 420C.

Maybe we should have a two-fluid MSR with a chloride breeder salt, and a fluoride fuel salt.

The fluoride separates the fuels and many TRUs well. The chloride easily separates Pa.


Top
 Profile  
 
PostPosted: Sep 07, 2010 1:20 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
rgvandewalker wrote:
Pa(V)Cl boils at 420C.

Maybe we should have a two-fluid MSR with a chloride breeder salt, and a fluoride fuel salt.

The fluoride separates the fuels and many TRUs well. The chloride easily separates Pa.


We want the opposite. The blanket must moderate to better absorb neutrons in thorium and reflect some neutrons back into the core. Fuel must be fast to get the benefits of fast fission and TRU burning. Fluoride blanket and chloride core could work. However the moderating effect of the blanket will make the barrier requirements even more severe. I'm much more interested in single fluid fast chloride designs. Pa fissions somewhat in a very fast spectrum, and U234 fissions just fine. These reactors have to be big to keep leakage down to reasonable level, but my thinking is that the fast neutron damage will force lower power (ie neutron) density anyways.

For a fluorides reactor, PaF4 and PaF5 also boil quite low under vacuum. UF4 will come out nicely as well. The still bottoms for the blanket will be sent back to blanket salt. You do need to process rapidly since Pa has about a one month half life and fission in the blanket has to be kept low. I don't see why thats a big deal though. Its just a little still. The problem is to sell rapid online processing in terms of development, operational and proliferation risks.


Top
 Profile  
 
PostPosted: Sep 07, 2010 1:39 pm 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3069
Cyril R wrote:
I'm much more interested in single fluid fast chloride designs. Pa fissions somewhat in a very fast spectrum, and U234 fissions just fine. These reactors have to be big to keep leakage down to reasonable level, but my thinking is that the fast neutron damage will force lower power (ie neutron) density anyways.

I think the big challenges with the fast chloride reactor will be:
#1, #2, and #3) the R&D expense of starting over with the materials
#4) the capital expense of the fissile.

Perhaps you could line the inside wall with a neutron absorber (like B4C)? This would not have a strength requirement and might be able to significantly reduce the neutron load seen by the wall.


Top
 Profile  
 
PostPosted: Sep 08, 2010 12:29 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Lars wrote:
Cyril R wrote:
I'm much more interested in single fluid fast chloride designs. Pa fissions somewhat in a very fast spectrum, and U234 fissions just fine. These reactors have to be big to keep leakage down to reasonable level, but my thinking is that the fast neutron damage will force lower power (ie neutron) density anyways.

I think the big challenges with the fast chloride reactor will be:
#1, #2, and #3) the R&D expense of starting over with the materials
#4) the capital expense of the fissile.

Perhaps you could line the inside wall with a neutron absorber (like B4C)? This would not have a strength requirement and might be able to significantly reduce the neutron load seen by the wall.


#1,2 and 3 remains to be seen. Think of the R&D expensive of order of magnitude greater neutron loading on your 'proven' hastelloy N for a high power density LFTR. All the problems with helium embrittlement from boron traces and nickel transmutations are going to be greatly exaggerated. With fast spectra the Hastelloy will do better, the question here is the higher neutron loading per area. Effectively solved by making it a lower power density reactor. Fast neutron flux similar to sodium fast reactors, allowing the materials R&D for nuclear durability on this front. The R&D for sodium cooled fast reactors materials is much bigger than ORNL's Hastelloy N development for fluoride fuelled reactors. General finding for the sodium cooled reactors was that the iron base materials did well under reasonable temperatures (hence my interest in slightly lower operating temperatures) but formed a low melting alloy with bred plutonium metal. A problem we obviously don't have. They had problems with getting the coefficients right, again much easier with fluid fuel chlorides.

Chemically, it seems the free energy of formation difference stuff is well known and only slightly worse than fluorides. The hastelloy would do fine, or the sodium cooled fast reactor iron base materials with pyrolitic carbon, nickel or copper coatings. Why is the R&D huge for this? Seems pretty simple improvement to me. Worry more about the system engineering; in either chloride or fluoride reactor it seems to me there will be huge engineering costs to make a large scale commercial system. ORNL thought 1 billion dollar, which I think is grossly optimistic in today's setting.

As for #4. What is the cost/kWh levelised for the 10x larger fissile startup? From what I've read its very tiny and 10x tiny is still not much. We need to get rid of the TRUs one way or the other. If we put in more its good public relations. Important to market this aspect well.

As for the wall, it is quite big and at the outside of the reactor and will see orders less neutron fluxes. It would be preferable to put the fast reflectors around first and then use B4C to mop up remaining leakage. It is true that the remaining flux would be quite signficant, so maybe you're right and reflectors won't work as well. Then we must live with greater leakage, meaning less complete burning of TRU, or requirement for a small chloride distillation unit.


Top
 Profile  
 
PostPosted: Sep 08, 2010 4:08 pm 
Offline

Joined: Mar 07, 2007 11:02 am
Posts: 914
Location: Ottawa
Quote:
As for #4. What is the cost/kWh levelised for the 10x larger fissile startup? From what I've read its very tiny and 10x tiny is still not much. We need to get rid of the TRUs one way or the other. If we put in more its good public relations. Important to market this aspect well.


As an example, the REBUS 3700 chloride fast reactor design (3700 MWth) needed 11 tonnes of fissile Pu per GWe to start up. That is 18 to 20 tonnes of LWR Pu which typically quoted as costing about 100$ a gram to process out of spent fuel. Perhaps cheaper processing methods could be used for molten salt reactors (especially fluoride based) but a high end of 2 billion$ per GWe is certainly not a tiny expense in my view! The IFR typically assumed startup on 20% LEU because of the high cost of Pu (something only found in the fine print!).

David LeBlanc


Top
 Profile  
 
PostPosted: Sep 09, 2010 12:47 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3069
Cyril R wrote:
Think of the R&D expensive of order of magnitude greater neutron loading on your 'proven' hastelloy N for a high power density LFTR. All the problems with helium embrittlement from boron traces and nickel transmutations are going to be greatly exaggerated. With fast spectra the Hastelloy will do better, the question here is the higher neutron loading per area.

Helium embrittlement from boron traces are going to be the same for all reactors in that if the wall lasts for any reasonable time you will burn out all the boron10. We can try to reduce the boron10 content but this is an advantage for all reactors not an argument for fast is better than slow.

Nickel transmutations happen via two mechanisms. With slow neutrons it is a two step process 58Ni -> 59Ni -> He+56Fe and the helium content will grow quadratically in time - lasting >30 years for the MSBR design - but shorter for 1.5 fluid designs and shorter still for true two fluid designs. With fast neutrons there is a significant 58Ni -> He + 55Fe path that dominates even at the French fastish spectrum and will be more pronounced at a very fast spectrum like a chloride reactor. I expect wall life will preclude two fluid chloride designs. A single fluid chloride reactor is still viable though.

The problem can be solved for any reactor by making a lower power density core BUT to double the wall life one would need to also double the fissile inventory - something that is already a cost challenge for fast reactors.

Quote:
As for the wall, it is quite big and at the outside of the reactor and will see orders less neutron fluxes.

The wall I'm concerned about is the first wall that actually touches the chloride salt. I don't understand how it could see orders of magnitude less neutrons. The flux seen is pretty well set by the geometry and power level.

Quote:
It would be preferable to put the fast reflectors around first and then use B4C to mop up remaining leakage. It is true that the remaining flux would be quite signficant, so maybe you're right and reflectors won't work as well.

If you can dig up information on how much a reflect can reduce neutron losses and how thick it is I'd love to see it. This information has been difficult for me to find.
Quote:
Then we must live with greater leakage, meaning less complete burning of TRU, or requirement for a small chloride distillation unit.

I agree with the first part - if we don't find a good reflector then there will be greater leakage which is tough on the wall and means we need a more generous neutron budget. I believe you have a more generous neutron budget due to the fissions in 234U so this portion isn't the problem. I don't see that it has anything to do with how complete the TRU burning is. Leakage will affect your criticality not your fuel evolution.


Top
 Profile  
 
PostPosted: Sep 09, 2010 1:04 am 
Offline
User avatar

Joined: Jan 10, 2007 5:09 pm
Posts: 501
Location: Los Altos, California
Hmm. Spent nuclear fuel has a Kinf < 1, because they put the stuff at the perimeter of the reactor and let the neutrons from the Kinf > 1 central portion flow out into it. So, even though it can't support fission on its own, it produces quite a bit of power when stimulated with quite a lot of neutrons.

But, the stuff has loads of fissile in it and hardly any burnup by MSR standards. If an MSR can get substantially higher burnup, even when keeping around the nonvolatile fission products, then it should be possible to process the SNF into MSR fuel.

There are two catches to this simple back-of-envelope analysis:
- PWR fuel includes burnable poisons. BF3 is a gas, and will leave when the SNF is fluorinated. The PWR guys just use boron-10, which produces Li-7 when it absorbs a neutron, which is fine with us.
- The other burnable poison I found was Gadolinium. GdF3 has a melting point of 1231 C, and I have no idea if it precipitates out or stays in the salt, but I'll guess it stays. When hit with a neutron, Gadolinium looks like it just turns into other Gadolinium isotopes, so this stuff is more of a permanent poison than something burnable, and will be a pain in the ass.
- Any SNF that we get to use will probably be aged a few years at least before it can be moved. So we'll have lost some of the Pu-241 and the reactivity that comes with that.

I'm going to guess that we can directly use SNF which has been declad and fluorinated as fissile feed. That should be a lot cheaper than separated Pu.

In particular, I like Lars' idea of locating the DMSR inside a decommissioned PWR. If the SNF fluorinator can be made reasonably small, we can install that at the spent fuel pool and that pool becomes the fissile make-up feed for the DMSR.

-Iain


Top
 Profile  
 
PostPosted: Sep 09, 2010 2:00 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3069
According to TM-7207 the makeup fuel is LEU20 except for the additions in years 2 and 3. The fissile content of SNF is too low for us. We want to add thorium as our fertile feed, not 238U so when we need to add makeup fissile we want it to be as rich as possible hence LEU20. It is possible that SNF/Pu would work even better as makeup fissile and might be the extra boost needed to get DMSR to be break even eventually.


Top
 Profile  
 
PostPosted: Sep 09, 2010 8:21 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Re Boron: B-10. Its neutron, gamma is orders of magnitudes lower at fast energies compared to epithermal/thermal. That matters a lot. Neutron, alpha is even more orders of magnitudes. Neutron, proton is similar in very fast compared to thermal but is small compared to previous two absorptions. It may be the difference between a 10 month lifetime and a 1000 month lifetime. Which is the difference between a practical design and an infeasible one. Boron control rods work because boron-10 is poison in thermal energies.

Re Nickel: Ni58 neutron, alpha barely shows on the nndc databases. Less than 0.001 barn @ 2 MeV and downhill from there.

Re the wall. It is big because of single fluid design being a big reactor (roughly similar total core power density as PWR).

Re neutron leakage effect. My thinking is that leakage affects burnup in this reactor because more leakage means the once through cycle has to stop earlier as the conversion ratio will drop sooner and we don’t want to add fissile. So indirectly it affects fuel evolution with less pu being burned out. That makes sense to you?

Re reflectors. Cannot find much at all for fast reflectors.


Top
 Profile  
 
PostPosted: Sep 09, 2010 10:00 am 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
iain wrote:
.........The other burnable poison I found was Gadolinium. GdF3 has a melting point of 1231 C, and I have no idea if it precipitates out or stays in the salt, but I'll guess it stays. When hit with a neutron, Gadolinium looks like it just turns into other Gadolinium isotopes, so this stuff is more of a permanent poison than something burnable, and will be a pain in the ass.........
Why? Gd is a neutron absorbing rare earth, like many other fission products. If you have on-line reprocessing, you can feed the top-up spent fuel in via the reprocessing unit. For start-up, or if you are not processing as you go, you need a big reprocessor to split the spent fuel into waste, TRUs, and U. The U is too low in fissile to be used, unless you are running U/Pu cycle and it is just a source of fertile. It might make reasonable CANDU fuel. We'll keep the TRUs for chloride (or DMSR) make-up.

Since we're supposed to be on chloride reactors here:-
GdCl3 melts at 609C, boils 1580C, c/f UCl4 melts 590C, boils 791C, so distilling the UCl4 out is easier than LiF distillation.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 25 posts ]  Go to page 1, 2  Next

All times are UTC - 6 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group