Liquid Fluoride Thorium Reactor - Disadvantages

Disadvantages

LFTR's deviate strongly from today's operating commercial power reactors. A different fuel cycle (thorium rather than uranium), low pressure operation (rather than high pressure), liquid fuel (versus solid fuel), use of molten salts (rather than water or gasses), online refueling and reprocessing using pyroprocesses (opposed to off-site processing using water solvents), LFTRs are different in almost every aspect. This gives rise to a uniquely different set of design challenges and trade-offs with varying levels of design, political and inherent difficulties:

• Mothballed technology. Only a few MSRs have actually been built; those experimental reactors having been constructed more than 40 years ago. This leads some technologists to say that it is difficult to critically assess the concept.
• Startup fuel. Unlike mined uranium, mined thorium does not have a fissile isotope. Thorium reactors breed fissile uranium-233 from thorium, but require a considerable amount of U-233 for the initial start up. Currently there is very little of this material available. This raises the problem of how to start up the reactors in a reasonable time frame. There are a number of ways to start up the reactors. One option is that U-233 could be produced in today's solid fuelled reactors, then reprocessing the U-233 out of the solid fuel to start up a LFTR. A LFTR can also be started up by different fissile isotopes. The two alternative options for LFTR startup are enriched uranium and plutonium from reactors or decommissioned bombs. For enriched uranium startup, a quite high enrichment is needed. Decommissioned uranium bombs have a high enough enrichment, but not enough is available to start up a large number of LFTRs. For plutonium startup, it is more difficult to separate plutonium fluoride from lanthanide fission products. One option for a two fluid reactor is to operate with plutonium or enriched uranium in the fuel salt, breeding U-233 in the blanket, but storing it instead of sending it back to the core. Instead, add makeup plutonium or enriched uranium to continue the chain reaction, similar to today's solid fuel reactors. When enough U-233 is bred, replace the fuel salt with new fuel salt, holding the U-233 as new startup fuel. A similar option exists for a single fluid reactor operating as a converter. Such a reactor would not reprocess fuel online, instead would startup on plutonium with thorium as the fertile, and add makeup plutonium. After many years the plutonium burns out and U-233 is produced in place. At the end of the reactor fuel life, the spent fuel salt can be reprocessed to recover the bred U-233 to start up new LFTRs.
• Salts freezing. The fluoride salt mixtures have high melting points, depending on the mixture it ranges from 300 to over 600 degrees Celsius. The salts, especially those with beryllium fluoride, are very viscous close to their freezing point. This requires careful design and freeze protection in the containment and heat exchangers. Freezing must be prevented in normal operation, during transients, and during extended station blackouts. The primary loop salt contains the decay heat generating fission products, so these help to keep the salt hot and liquid. For the MSBR, ORNL planned on keeping the entire reactor room (the hot cell) at high temperature, like an oven. This avoided the need for individual electric heater lines on all piping and provided more even heating of the primary loop components.. One "liquid oven" concept developed for molten salt cooled, solid fueled reactors, employs a separate buffer salt pool where the entire primary loop is suspended in. Because of the high heat capacity and considerable density of the buffer salt, the buffer salt not only prevents fuel salt freezing, but also participates in the passive decay heat cooling system, provides radiation shielding, and reduces deadweight stresses on primary loop components. This design could also be adopted for LFTRs.
• Beryllium toxicity. The proposed salt mixture FLiBe, contains large amounts of beryllium, a poisonous element. The salt in the primary cooling loops must be isolated from workers and the environment to protect them from beryllium poisoning. This is routinely done very effectively in various industries that handle beryllium. Based on this industrial experience with beryllium, the added cost due to beryllium safety is expected to cost 0.012 cent per kWh ($0.12/MWh), which is very small. Furthermore, after start up, the fission process in the primary fuel salt will produce highly radioactive fission products that produce a high gamma and neutron radiation field. A highly effective containment will therefore be a primary requirement of any molten salt reactor. It is also possible to operate without beryllium fluoride in the salt. It is possible to operate on lithium fluoride-thorium fluoride eutectic without beryllium, as the French LFTR design, the "TMSR", has chosen. This comes at the cost of a somewhat higher melting point, but has the additional advantages of simplicity (avoid BeF₂ in the reprocessing systems), increased solubility for plutonium-trifluoride, reduced tritium production (Beryllium produces lithium-6, which in turn produces tritium) and improved heat transfer (BeF₂ increases the viscosity of the salt mixture). Alternatively various additional solvents are available such as the fluorides of sodium, rubidium and zirconium that allow lower melting points at a tradeoff in breeding.
• Loss of delayed neutrons. The production of the abovementioned delayed neutrons in the heat exchanger and other external piping introduces another design constriction. In order to be gently, slowly and predictably controlled, nuclear reactors rely critically on the delayed neutrons. In other words, they require the additional slowly evolving neutrons from fission product decay to continue the chain reaction normally. Because the delayed neutrons evolve slowly, this makes the reactor very gently controllable. In a LFTR, the presence of the fission products outside of the core in the heat exchanger and piping means a portion of these delayed neutrons are also lost. They do not participate in the critical chain reaction of the core, which in turn means the reactor will be less gently behaving during changes of flow, power etc. Approximately up to half of the delayed neutrons can be lost to retain the advantage of easy reactor behavior. In practice, it means that the heat exchanger must be compact so that the volume outside the core is as low as possible. The more compact (higher power density) the core is, the more important this issue becomes. Having more fuel outside the core in the heat exchangers also means more fissile fuel is needed to start up the reactor, which is expensive. This makes a fairly compact heat exchanger an important design requirement for a LFTR.
• Waste management. There is also a need to manage the waste, which is still very radioactive, even though it is hazardous for a shorter period. Because some fission products, in their fluoride form, are highly water soluble, fluorides are a less suited long term storage form. For example, cesium fluoride has a very high solubility in water. For long term storage, conversion to an insoluble form such as a glass, could be desirable.
• Decommissioning costs are uncertain. Cleanup of the Molten-Salt Reactor Experiment was about $130 Million, for a small 8 MW(th) unit. Much of the high cost was caused by the unpleasant surprise of fluorine and uranium hexafluoride evolution from cold fuel salt in storage that ORNL did not defuel and store correctly, but this has now been taken into consideration in MSR design. In addition, decommissioning costs don't scale strongly with plant size based on previous experience, and costs are incurred at the end of plant life, so a small per kWh fee on the electricity produced is sufficient to pay for it. For example, a GWe reactor plant produces over 300 billion kWh of electricity over a 40 year lifetime, so a 0.2 cent per kWh decommissioning fee delivers $600 million at the end of the plant lifetime.
• Noble metal buildup. Some of the radioactive fission products, such as noble metals, don't form salts, but deposit somewhere in the piping. This includes most of the reprocessing part, that is not yet well tested. Equipment, such as nickel wool sponge cartridges, will have to be developed to filter and trap the noble metals to prevent them from building up excessively in the piping, reprocessing plant, and heat exchanger over time.
• Limited graphite lifetime. Compact designs have a limited lifetime for the graphite moderator and fuel / breeding loop separator. Under the influence of fast neutrons, the graphite first shrinks, then expands indefinitely until it becomes very weak and can crack, leading to mechanical problems and causing the graphite to absorb more fission products that poison the nuclear reaction. A replacement of this central part can be challenging and needs to be done using remote equipment. However this must be compared to today's solid fuelled reactors, which must typically replace 1/3 of the entire core, including all of the highly radioactive fission products therein, every 12 to 24 months. The 1960 two-fluid design had an estimated graphite replacement period of four years. Eliminating graphite from sealed piping was a major incentive to switch to a single-fluid design. Most MSR designs arrange for it to be easy to replace. In a molten salt reactor, virtually all of the fuel and fission products can be drained out of the core to a holdup tank, when the graphite must be replaced. Only a fraction of a percent of the fission products end up in the graphite, primarily due to fissions occurring close to the bare graphite and the resulting fission products slamming into the graphite. However, this does make the used graphite surface itself radioactive, and without recycling/removal of at least the shallow surface layer, this will create a fairly bulky waste stream. Removing the thin surface layer of the graphite that contains the embedded fission products, and recycling the remainder of the bulk graphite, would solve this issue. Several techniques exist to recycle or dispose of nuclear moderator graphite. Graphite is inert and immobile at low temperatures, so it can also be readily stored or buried if required. At least one design used graphite balls (pebbles) floating in salt, which could be removed and inspected continuously without shutting down the reactor. Reducing the power density of the reactor design increases graphite lifetime.
• Graphite causes positive reactivity feedback. When graphite heats up, it increases the amount of fission of U233, thus causing an undesirable positive feedback. The LFTR-Design must avoid certain combinations of graphite and salt, and certain core geometries. This problem can be solved by employing either a well thermalized spectrum with lots of graphite, or using little or no graphite at all resulting in a faster spectrum. However this comes at the price of reduced breeding or increasing fissile inventory.
• The solubility for plutonium is limited. The fluorides of plutonium, americium, and curium, occur as trifluorides, which means they have three fluorine atoms attached (PuF₃, AmF₃, CmF₃). Such trifluorides have a limited solubility in the FLiBe carrier salt. This makes startup on these transuranic wastes more difficult especially for a compact design that uses a smaller primary salt inventory. This solubility can be increased by operating with less or no beryllium fluoride (which has no solubility for trifluorides) or by operating at a higher temperature (as with most other liquids, the solubility goes up with temperature). A thermal spectrum, lower power density core does not have issues with plutonium solubility.

• Potential proliferation risk from reprocessing. If reprocessing of the salt mixture works well, this technology like any advance in reprocessing technology can pose a proliferation-risk. As an alternative method of reprocessing this could be used to separate plutonium from other reactors as well. However, as stated above, plutonium is chemically difficult to separate from thorium, and plutonium cannot be used in bombs if diluted in large amounts of thorium. In addition, the plutonium produced by the thorium fuel cycle is mostly Pu-238, which creates high levels of spontaneous neutrons and decay heat that greatly complicate fission bomb design and reduces the bomb yield drastically by predetonation due to Pu-238 starting the chain reaction before the fissile material in the bomb is fully assembled (causing a "fizzle" rather than a fission bomb explosion). Finally, all of the reprocessing steps involve automated, non-human handling in a fully closed and contained hot cell, this is due to the high temperatures and radiation levels, which makes diversion of bomb material very difficult. Compared to today's water solvent extraction methods such as PUREX, the pyroprocesses are inaccessible and produce impure fissile materials, often with large amounts of fission products contamination. This is not a problem for an automated reactor-hot cell system, but poses severe difficulties for would-be bomb makers.
• Proliferation risk from protactinium separation for some specific designs. Compact designs can breed only using rapid separation of protactinium, a potential proliferation-risk, since this gives possible access to high purity 233-U without the U-232 contamination that is in the core. This is not easy, but still a possible new path to weapons-grade material. Because of this possibility, commercial power reactors may have to be designed without separation of protactinium. In practice, this means either not breeding or operating at a lower power density. For a two fluid design, this could simply operate with a bigger blanket, and keep the high power density core (which has no thorium and therefore no protactinium).
• Proliferation of Neptunium-237. In designs utilizing a fluorinator, Np-237 comes out with uranium as gaseous hexafluoride and can be easily separated using solid fluoride pellet absorption beds. Theoretically, it should be possible to use Np-237 as fission bomb material. No one has ever successfully produced a bomb with this material, but it should be theoretically possible to use it, because of its considerable fast fission cross section and low critical mass. When the Np-237 is kept in the reactor, it will transmute to Pu-238, an extremely high value fuel for space radioisotope thermal generators. A single gram is worth thousands of dollars. Pu-238 is itself an excellent proliferation deterrent, as explained earlier. Because of this, the Np-237 will likely be sent back to the reactor to be transmuted to Pu-238. In addition, it is possible to use vacuum distillation instead of fluorination, which does not separate neptunium at all. It should be noted, that all reactors, not just thorium reactors, produce considerable amounts of neptunium, which is always present in high (mono)isotopic quality, and it is just as easily extracted chemically. This is therefore not a distinguishing issue for LFTRs in particular. In fact, americium could also be theoretically used for nuclear weapons, and LFTRs do not produce meaningful quantities of americium, indeed they are one of the few reactor types that can burn existing stockpiles of americium and neptunium with high efficiency.
• Neutron poisoning and tritium production from lithium-6. Lithium-6 is a strong neutron poison; using LiF with natural lithium, with its 7.5% Lithium-6 content, will not even allow the reactor to start up. The high neutron density in the core rapidly transmutes lithium-6 to tritium, losing precious neutrons that are required to sustain at least break-even breeding, and also producing tritium in the process. Tritium is a radioactive isotope of hydrogen, which is nearly identical, chemically, to ordinary hydrogen. In hot fluoride salts, hydrogen will be present as elemental tritium. In the MSR the tritium is quite mobile because, in its elemental form, it rapidly diffuses through metals at high temperature. So some effort is needed to keep the emissions low. If the lithium is isotopically enriched in lithium-7, and the isotopic separation level is high enough (99.995% lithium-7), the amount of tritium produced is only a few hundred grams per year for a 1 GWe reactor. This much smaller amount of tritium comes mostly from the lithium-7 - tritium reaction reaction and from beryllium, which can produce tritium indirectly by first transmuting to lithium-6, thus regenerating the tritium producing lithium-6. Because of these reasons, if a LFTR design uses a lithium salt, it uses the lithium-7 isotope, via enrichment of natural lithium, to reduce tritium formation. In the MSRE, tritium formation was also reduced by this removal of lithium-6 from the fuel salt via isotopic enrichment. Since lithium-7 is at least 16% heavier than lithium-6, and is the most common isotope of lithium, the lithium-6 is comparatively easy and inexpensive to extract from naturally occurring lithium. Vacuum distillation of lithium achieves efficiencies of up to 8% per stage and only requires heating of raw lithium in a vacuum chamber. The aforementioned methods of chemistry control and lithium isotopic separation worked in preventing hydrogen corrosion and excessive tritium generation in the MSRE. Practical MSRs also operate the salt under a blanket of dry inert gas, usually helium. Fortunately, in a LFTR, there is also a good chance to recover the tritium, since it is not diluted in a large amount of water as it is in CANDU reactors. Various methods exist to trap tritium, such as hydriding it to a stable metal hydride using titanium, oxidizing it to less mobile (but still volatile) forms such as by using sodium fluoroborate or molten nitrate salt as coolant loops, or trapping it in the turbine power cycle gas and offgas using copper oxide pellets . For future systems, ORNL developed a secondary loop coolant system that would chemically trap the few hundred grams of residual tritium to a less mobile form, so that it could be trapped and removed from the secondary coolant rather than diffusing into the turbine power cycle. This technique by itself would reduce tritium emissions to the actual environment to acceptable levels, ORNL calculated.
• Corrosion from tellurium. The reactor makes small amounts of tellurium as a fission product. In the MSRE, this caused small amounts of corrosion at the grain boundaries of the special nickel alloy, Hastelloy-N used for the reactor. Metallurgical studies showed that adding 1 to 2% niobium to the Hastelloy-N alloy improves resistance to corrosion by tellurium. One additional strategy against corrosion was to keep the fuel salt slightly reducing by maintaining the ratio of UF₄/UF₃ to less than 60. This was done in the MSRE by continually contacting the flowing fuel salt with a beryllium metal rod submersed in a cage inside the pump bowl. This causes a fluorine shortage in the salt, reducing tellurium to a less aggressive (elemental) form. This method is also effective in reducing corrosion in general from the fluoride salt, because the fission process produces more fluorine atoms freed from the fissioned uranium that would otherwise attack the structural metals.
• Radiation damage to nickel alloys. The standard Hastelloy N alloy, a high nickel alloy used for constructing the primary fuel salt loop, was found to be embrittled by the neutron radiation from the reactor core. This was caused by neutrons reacting with nickel to form helium. This helium, being a gas, concentrated at specific points inside the alloy. There, it caused stresses to build up. ORNL addressed this problem by adding 1–2% titanium or niobium to the Hastelloy N. This small titanium addition changed the internal structure of the alloy, so that the helium produced inside it would not concentrate in specific areas but would be finely divided in it. This relieved the stress and allowed the new modified Hastelloy N to withstand considerable neutron flux. However the maximum temperature is limited to about 650 °C. A number of other alloys also showed promise. The outer vessel wall that contains the salt can have neutronic shielding, such as boron carbide, to effectively protect it from neutron damage.
• Long term fuel salt storage issues. If the fluoride fuel salts are stored in solid form over many decades, radiation can cause the release of corrosive fluorine gas, and uranium hexafluoride. This was due to radiolysis of the salt from remaining fission products, when colder than 100 degrees Celsius. The salts should be defueled and wastes removed before extended shutdowns. The fluorine and uranium hexafluoride evolution can be prevented by storing the salts at a temperature above 100 degrees Celsius. Because some of the fission product fluorides have high solubility in water, fluorides are less suitable for long term storage anyway. For longer term storage, fluoride containing wastes could go through a vitrification process to be encased in insoluble borosilicate glass suitable for long-term disposal.
• Business model. Today's solid fuelled reactor vendors make long term revenues by making profit on the fuel fabrication. Without any fuel to fabricate and sell, a LFTR would have to adopt a different business model.
• Development of the power cycle. Developing a large helium or supercritical carbon dioxide turbine is needed for the highest efficiency designs. These gas cycles offer numerous potential advantages for use with molten salt fuelled or molten salt cooled reactors. These closed gas cycles are not yet fully developed though, and face some design challenges and engineering upscaling work for a commercial turbine-generator set. A standard supercritical steam turbine could be used at least in initial designs, at a small penalty in efficiency (the net efficiency of the MSBR was designed to be approximately 44%, using an old 1970s steam turbine). However, even in that case, a molten salt to steam generator would still have to be developed. Currently, molten nitrate salt steam generators are being used in concentrated solar thermal power plants such as Andasol in Spain. Such a nitrate salt loop and steam generator could also be used for a molten salt reactor, as an additional, third circulating loop, where it would allow the use of existing molten nitrate salt steam generator equipment, and as an added bonus effectively trap any tritium that diffuses through the primary and secondary heat exchanger

Dr. Charles Forsberg of ORNL discussed the technological challenges for a commercial MSR in a 2006 paper.

Dr. David Leblanc from Carleton University discussed the basics, trade-offs, problem areas and possible design remedies in a presentation.