The liquid-fluoride reactor is an innovative design for a thermal breeder reactor that was developed from the 1950s through the 1970s at Oak Ridge National Laboratory in Oak Ridge, Tennessee. The reactor utilized a fluid-fuel form, with uranium and thorium fluoride salts dissolved in a matrix of lithium and beryllium fluoride salts. The melted salt was pumped throughout the reactor vessel and generated energy in an interesting manner. As the salt passed through the "core" region of the reactor, moderation provided by solid graphite elements led to neutron thermalization and fission reactions that produced heat. Then as the salt accumulated in a plenum and was pumped out of the core, fission ended and the salt passed through an external heat exchanger where it was cooled and transferred its heat to a secondary salt, and ultimately to a working fluid.
In 1970s, a steam-Rankine cycle was the basic power conversion technique considered for the liquid-fluoride reactor. However, there were a number of problems with this approach, mainly stemming from the fact that the natural temperature range of the fluoride salt was significantly above the typical operating temperatures of steam systems used for nuclear reactors.
More recent work on the liquid-fluoride reactor has focused on using the helium-Brayton (gas-turbine) power conversion cycle for electrical generation. This cycle would offer higher conversion efficiencies (~50% efficiency) through the salt’s unique abilities to take advantage of multiple reheating steps in the Brayton cycle. The high-temperature attributes of the salt also enable other unique applications, such as the thermochemical generation of hydrogen directly from nuclear heat.
The unique attributes of the liquid-fluoride reactor are a consequence of its fuel form. Salts of fluorine and alkali metals are exceptionally stable since they are formed from the most electronegative of elements (fluorine) and the most electropositive (lithium, beryllium, sodium). Due to their exceptional chemical stability, these fluoride salts have low vapor pressures at high temperature (enabling high temperature operation at low pressure) and they do not react with air or water, unlike molten metal coolants such as sodium. The favored combination for a neutronically-efficient liquid-fluoride reactor is a combination of lithium fluoride (highly enriched in the lithium-7 isotope) and beryllium fluoride. Through a proper ratio of these two salts, a solvent with a low melting point can be constructed. The minimum melting temperature of these salts is achieved when a composition of 52 mole % LiF and 48 mole % BeF2 is used. This combination will melt at 356°C. Typical compositions of base salts that have been used in liquid-fluoride reactors are 66 mole % LiF and 34 mole % BeF2.
14 comments:
How small can an LFR be? Could one be used to power a train, bus, or apartment building. Can we make lots of little ones when we want a big one, and thereby get the advanatages of modular system development?
Hi Randall:
The thickness of the shielding on any fission reactor will make it impractical to one in a car bus or train. (In the case of the train, why bother? We have nuclear trains anywhere the rail system in electrified & nuclear provides some of the electricity.)
I would like to know how small (in both physical dimensions & power) the LFR can be made though. Could they be used to run ships, or provide heat & electricity to isolated settlements of a few hundred people?
Have you got any figures on that Kirk?
Could an LFR be made small? Almost certainly. My only hesitation would be if we wanted the fuel to be thorium (which would require a certain level of processing complexity) or simply a fissile (U-233, U-235, Pu-239) stream. If the fuel was pure fissile, the core could be made quite small indeed.
There are a number of physical features of an LFR that give it quite a bit of flexibility in scaling, more so than solid-core reactors. The first is xenon-135 transients. As you go to a high-flux core, you tend to build up more xenon and its iodine-135 precursor in the fuel--they are fission products. Iodine-135 is constantly decaying into xenon-135, and xenon-135 has a huge appetite for neutrons and tends to absorb them and become stable xenon-136.
What happens is that the reactor has a steady rate of xenon production (as a fission product or from iodine decay) and destruction (through neutron absorption). When the reactor shuts down (for whatever reason) the iodine precursor keeps decaying into xenon, but the xenon is now no longer being destroyed by neutrons. The xenon can build to such concentrations that it is not possible to restart the reactor again (because it is absorbing so many neutrons that the reactor can't achieve criticality), at least not until the xenon decays away to a certain level (which happens over a day or two). As you get into high-flux reactors (which many small reactors are) this problem becomes more serious. It's a not a problem at all for fluid-fueled reactors (like the liquid fluoride reactor) because xenon is continously coming out of solution as it is formed, so xenon transients don't preclude you scaling up or down.
Another important consideration is passive decay heat removal. Because the liquid-fluoride reactor relies on a freeze plug to drain the core into a passively-safe, non-critical configuration, you can scale the reactor without concern for in-core decay heat removal. This is a concern for certain gas-cooled reactors that rely on radiation cooling between the core vessel and the surrounding environment for passive decay heat removal. It works, but you can only scale the reactor so big before that trick doesn't work anymore.
The third advantage is continous capability for refueling. With this capability, you don't need to have all your fissile load to carry you between refuelings in the core, like all solid-core reactors must. With all that extra fissile in the core, a solid-core reactor must rely on burnable poisons (such as boron or gadolinium) to "drink" up neutrons at the beginning of the fueling cycle, and then burn out over the course of the fueling cycle until they are almost gone at the end. In addition to being a waste of neutrons, that extra fissile and the extra control required is another system to worry about and another thing that can go wrong. With the fluoride reactor, continuous refueling allows you to precisely control the level of fissile material in the core and achieve a minimum core fissile loading for a given flux requirement.
All of these factors lead me to believe that a liquid-fluoride reactor would be very amenable to a large range of potential operating demands.
What about the following issues:
* U232 hard Gamma emissions (U232 is created in the U233/Thorium fuel cycle)
* caking of salts and precipitants on piping leading to clogging and damaged pumps
* Corrosive properties of floride salts and UF4
* UF4 production of HF gas when exposed to humid air.
* Actinide and other fission product precipitation.
* Health risk associated with BeF2 (water soluble)
* High Cost of Li7
* Water containation resulting in production of hydroflouric acid from UF4 or liberated F2 gas.
* Limited availability of Helium for use with helium-Brayton cycle. Helium production is from Natural Gas fields
* Low breed rate of ~1.1
* Noble metal plate out and Carbon-Carbon Erosion caused by chemical reactions with fission products
What is the estimated lifetime of a commerical reactor?
What are the technical issues decommission a reactor after its lifecycle?
What tools are available to inspect the inner workings of the reactor? With Water reactors most of the reactor can be visially inspected because water is transparent. Other traditional inspection tools (ultrasound, X-ray, etc) also do not work.
How are fission products effectively removed? Actinide have a much smaller neutron delayed fraction and must be affectively removed to keep the reactor under control.
What are the containment challenges with MSRs?
Lack of seed U233 and low breed returns will likely deter large scale commercial development. We would need to first construct many FBR in order to develop large quantities of U233 initially fuel MSRs.
I would guessimate it would require 20 to 25 years before any large scale production of commerical plants could be constructed. We would probably need 2 to 5 years to design a full scale commerical MSR, and another 5 to 10 years to construct and properly test a prototype. Plus another 5 to 10 years to produce enough seed U233, train operators, aquire land for plant sites, etc.
A better idea would to use heavy water reactors, which cannot breed but burn fuel efficiently and there are already commerical designs avail. In my opinion it would be safer to choose a proven plant design for the short term and look at more exotic designs for the long term.
Anonymous, you've given me a long list of questions, many of which will be answered by reading some of the numerous documents attached to this site. Nevertheless, I will attempt brief answers:
What about U232 hard Gamma emissions (U232 is created in the U233/Thorium fuel cycle)
The U-232 contamination in U-233 is only a problem for solid-fueled reactors, which have gotten used to using relatively non-radioactive fresh fuel. Reprocessed fuel from the thorium cycle is "hot" because of U-232. But this very "contaminant" is a fundamental benefit of the cycle, since it is the primary deterrent against using U-233 in nuclear weapons. I'm incredibly glad the U-233 is contaminated by U-232--the more the better! And in a fluoride reactor where fuel fabrication is remote and as simple as reducing a UF6 stream in a column of salt, it's no problem.
caking of salts and precipitants on piping leading to clogging and damaged pumps
Are you sure you're not confusing this with an aqueous salt that could precipitate? This is a liquid anhydrous salt, and won't be caking on anything once you've melted it. None of these problems were encountered in the operation of a real liquid-fluoride reactor (the MSRE) over the course of four years.
Corrosive properties of floride salts and UF4
As I've stated on here repeatedly, and can be verified by examining the actual test documents, which are here as well, the right salt combination with the right material doesn't lead to corrosion. This is an issue that keeps coming up even though the data is there to show that it isn't a problem.
UF4 production of HF gas when exposed to humid air.
We're not planning to expose UF4 to humid air. We're keeping it in the reactor. The reaction rate you mention is very very small anyway. UF4 is more stable as a fluoride than an oxide.
Actinide and other fission product precipitation.
The actinides don't precipitate in the fluoride salt--this isn't the aqueous homogenous reactor (which had issues with precipitation). Again, verified in actual operation. Some of the noble metals can "plate out" onto surfaces--this is reduced though distillation of the fuel salt to remove fission products.
Health risk associated with BeF2 (water soluble)
We're not planning to eat beryllium or disperse it to the environment. I'm sure I can find chemicals in any nuclear reactor that shouldn't be eaten, ingested, or dispersed to the environment. BeF2 is glassy, and if it's water soluble, that's news to me.
High Cost of Li7
Everything has its price--isotopically separated Li-7 is worth it.
Water containation resulting in production of hydroflouric acid from UF4 or liberated F2 gas.
Again, no plans to dump this stuff in water.
Limited availability of Helium for use with helium-Brayton cycle. Helium production is from Natural Gas fields
We don't need that much--this is a closed-cycle system. I'm sure we won't have to institute draconian measures like banning helium from children's balloons.
Low breed rate of ~1.1
I don't even plan to get that high--I want to run this thing at a CR of 1.0. That way no fissile with come in (after the start charge) and no fissile will leave. We can start the reactor on LEU, HEU, U-233, reactor-grade Pu, weapons-grade Pu, or any combination. U-233 is the best, but the others can be accommodated.
Noble metal plate out and Carbon-Carbon Erosion caused by chemical reactions with fission products.
Again, reprocessing to remove fission products. No carbon erosion was detected during operation of the MSRE.
You seem to have a whole bunch more questions, Anonymous. Why don't you identify yourself and I'll answer the rest.
Hi Kirk,
Thank you for you response. Below are my comments:
>But this very "contaminant" is a fundamental benefit of the cycle, since it is the primary deterrent against using U-233 in nuclear weapons
Unfortunately this isn't really true. While the gammas make detection easier (compared to U-235 and to a limited degree Pu239) explosive devices, it does not include the potential of U-232 used in dirty bomb devices which would be far more hazardous and more likely to occur than a fission bomb.
To be fair, use of Thorium/U-233 for WMDs is not my primary concern. The issue I have with it is the potential hazards with reactor maintaince and shielding. For instance the MSRE reactor has been permenently shutdown for more than two decades but it has not yet been completely dismantled because of the high decommisioning costs and exposure risks.
>None of these problems were encountered in the operation of a real liquid-fluoride reactor (the MSRE) over the course of four years..This is a liquid anhydrous salt, and won't be caking on anything once you've melted it.
This is not true. I had discussion with an engineer that worked on the MSRE. One of the reasons why support was discoutinued was because of the liquid salts solidify on various tubing and containment walls. Large chucks of solidified salts would break off and flow into the circulation pumps causing damage. Unfortunately this information was never publically made available.
>As I've stated on here repeatedly, and can be verified by examining the actual test documents, which are here as well, the right salt combination with the right material doesn't lead to corrosion. This is an issue that keeps coming up even though the data is there to show that it isn't a problem.
>Again, reprocessing to remove fission products. No carbon erosion was detected during operation of the MSRE.
Carbon wasn't used to protect the pipes in the MSRE and thats how the plating problem was discovered. Although the reactor did use a graphite core (moderator)and carbon filters (to trap gases produced during operation) Liberated Florine gase reacted with carbon to produce carbon-florine compounds, some which are sensitive explosives. If I recall correctly, in 2000 or 2001 some work was done on the MSRE to stabilize these compounds.
If you read carefully the reports are based upon clean salts, that did not contain fission products or UF4.
In the 1970s, ORNL did some non-reactor tests using Carbon CVD pipes and Floride salts (with no fuel). Florine is relatively stable in salts but it can be readily liberated from UF4. The objective was to test erosion first in a labotory eviroment. The next set of test was to try it an working reactor but these test were never carried out.
Since UF4 fuel was not used in tests, and did not include fission products either, we cannot verify conclude the test will reflect a working reactor.
>The actinides don't precipitate in the fluoride salt--this isn't the aqueous homogenous reactor
Sorry, but actindes do precipitate in anhydrous florine salts.
>We're not planning to expose UF4 to humid air. We're keeping it in the reactor
Over the expected operational lifetime of the reactor, water will almost certainly containment the florine salts, this will occur whenever the salt is exposed to air (during maintance, inspections, filtering, etc). To consider than no contamination will ever occur is a serious oversight.
>Again, no plans to dump this stuff in water.
No one ever plans to go to war either, but is happens. To make the assumption that it will never occur would be a bad idea. Water is everywhere in our environment and salts have an affinity for water.
>We don't need that much--this is a closed-cycle system. I'm sure we won't have to institute draconian measures like banning helium from children's balloons.
Helium is virtually impossible to prevent leakage. It will not be possible to operate a turbine with out replenishing it daily. It will easily leak through turbine bearings and any pipe joints.
>I don't even plan to get that high--I want to run this thing at a CR of 1.0
Why bother with the additional technical barriers and complexity of MSR when 0.9 CR is easily obtainable using Heavy water reactors. The only disadvantage is that is easier to create fuel for WMDs, which isn't a concern for nations that already have WMDs (US, China, Russia, India, etc). For countries that we would be concerned with Nuclear prolifieration, they would not have the technical and financial resources to construct MSRs anyway. In my opinion, it would be far better to simply remove the political barriers instead of trying to develop complex technological solutions to get around them.
I would beg to differ in regards to the "additional technical barriers" of MSR versus HWR:
The fueling machines for efficient HWRs (requiring on-line refuelling) are fiendishly complex machines. They must be capable of opening and closing pressurized fuel channels repetitively, without leaking expensive heavy water contaminated with tritium. Its comparable to removing & reinstalling the pressure vessel head on a PWR repeatedly for refueling.
By contrast, the MSR simply refuels on-line by circulating the fuel salt.
Also, for truly efficient fuel utilisation in an HWR, its better to have short fuel bundles instead of long fuel rods, and to shuffle them around the core by using fueling machines at both ends of the core - one at each end of the fuel channel, such that one machine loads new fuel (or fuel from the core periphery), while the other takes old fuel out, to move it somewhere else, or into the spent fuel storage bay.
This usually means that the fuel channels must be horizontal, to accomodate the fueling machines at each end. This in turn means that you end up with fuel channel sagging problems.
No such problem with MSR, since the fuel pipes can run vertically through the core.
I would agree however, that the breeding ratio should be maximized. This should be done by designing the MSR with a cooled HW moderator, instead of graphite. This also eliminates the graphite aging (cracking, etc.) problems, while at the same time sharply reducing the radwaste problem, as C-14 has a half-life of 5000years, compared to 12y for tritium.
Hi Jaro--I definitely agree with you about the simplicity of fluoride reactor refueling, but our friend Anonymous does love his CANDUs...
I've given some thought to the idea of using heavy-water as a moderator for a fluoride reactor. There is the issue of the temperature differential, but as you have pointed out, there is no pressure differential. One concern I had, however, had to do with oxygen diffusion into the salt. Oxide formation in the fluoride, while not chemically favored, does have a small equilibrium constant, and uranium oxides can precipitate. This is why the Molten-Salt Reactor Experiment (MSRE) carried a small amount of zirconium fluoride. The ZrF4 served to "mop up" any oxygen that made its way into the core. Later, they developed superior approaches to excluding oxygen that obviated the need for ZrF4 in the salt.
With heavy-water moderator, radiation will split the hydrogen and oxygen in the moderator, and some will diffuse into the salt and could cause oxide precipitation. I'm not sure if it is a problem that could be solved. What do you think?
I don't see the oxygen diffusing into the salt across pipe walls. Remember that there are two concentric pipes with a void (possibly filled with high-temperature insulation) separating them.
Hydrogen diffuses much easier, and can cause hydride cracking or blistering problems in zirconium alloys.
But this is much less of a concern in low-pressure systems, and in any event the much higher temp. MSR fuel pipes would need to be made of some other material.
Kirk, what's a practical business-model that would motivate commercial development of a fused-salt reactor? Most reactor manufacturers derive most of their income from fuel fabrication, so I can't see interest from that end. The -absence- of fuel fabrication seems like an important -advantage- of the salt reactor.
The only model I've thought of is to combine reactor development and power production. That seems like a bad bet, because power production is regional, and disperses capital.
Hi Kirk. I have a few questions related to safety and reliability. You mentioned the melting point of the flouride solutions was 356 degrees Celsium. Does that mean that liquid flouride reactors will suffer from the same problem of lead-bismuth reactors in that if the core tempurature falls (in the event of an accident) the coolant would solidify and render the reactor inoperable?
Also, how hazardous is the waste xenon that must be removed from the reactor between startups? Will it neccessitate removal in shielded tanks and storage like other nuclear waste?
Lastly, in applications where long intervals between refueling would be desirable, as in developing countries where simple, low maintenance reactors would be needed, what do you think is the longest interval between refueling possible for these reactors? Would it be possible, for example, to build a plant in developing Africa that requires minimal maintenance and 5 to 15 years between refuelings?
Hi Kirk, congratulations on this very informing blog!
No doubt you've thought of this already, or someone else has mentioned it, but maybe consider:
- Passive safety operation stead of active pumps etc. (perhaps something like the new AP design has?)
- Future designs would still incorporate emergency dump tanks for the primary salt, right? How about using armored dry casks instead? These could also be used to permanently store the salt (cold so it would be solid which is great for long term storage)after the plant's useful life. Could make storage and transportation (if needed) a lot easier?
- Someone mentioned elsewhere that helium shortage and leakage could be a problem. However, I think that argon or even carbon dioxide would also be suitable for Brayton cycles with a bit of modification.
Hi,
I'm an engineer and would like to ask some more questions about the technical aspects.
Why do you plan to make the reprocessing of the fuel within the power plant. It will increase complexity and increase investment cost significantly. An alternative is to build a central processing plant somewhere in the country. Each plant will ship 1/4 of the used fuel every 3 month to it and will get in return fresh fuel???
How is the process of reprocessing done?
Are there any new concepts of such a reactor of let me say 2000 - 3000MW?
What are the inlet and outlet temperatures?
What is the energy density KW/m3 or ft3 of it?
What kind of materials will you use for the reactor?
How many tons of fuel, moderator will you use?
What are the rough dimensions of a 2000MW reactor?
If you can work with temperatures above 360C supercritical water could be used for a secondary cycle to make the electricity out of it
vandale
Post a Comment