Kraig Farrar and Matthew Memmott, Chemical Engineering

Molten salt reactors (MSR) are a relatively unexplored but promising technology for civilian nuclear power. Original experiments with this technology was performed by Oak Ridge National Laboratory (ORNL) in the 1950s and 1960s. While this research was successful, it ultimately failed to receive additional funding and the program was cancelled as a result. Recently interest in MSR technology has resurged due to its promise of producing nuclear power at atmospheric pressure, high efficiency, and without the production of long lived waste. The purpose of this project is to determine the ideal material to use as both a fuel and coolant.

Both coolant and fuel fluids must be liquid at roughly atmospheric pressure and reactor operating temperatures (roughly 500-1000 °C). A fluoride salt is preferred as a fuel carrier due to its ability to avoid the release of gaseous fission products. The fuel carrier must have low neutron capture cross sections, low vapor pressure at reactor operating temperatures, low enough boiling point to provide a safe margin between reactor operation and freezing. The use of salts that exist in relatively reducing environments is strongly preferred because it allows for the use of less expensive container materials and allows for fission products to be more easily separated. Finally, the fuel carrier salt must also have a reasonable solubility of uranium. In the case of fluoride salts this would be UF4 or UF3. The coolant fluid will also act as a carrier for thorium. As such, the coolant fluid must have a low neutron capture cross section and solubility of thorium. The coolant also has the same requirements of low vapor pressure and relatively low melting point. It should be compatible with economically viable alloys and allow for separation of uranium. Both fuel and coolant fluids should ideally have high thermal conductivity, high volumetric heat capacity, and low viscosity. It is also ideal to avoid highly toxic materials and those that require isotopic separation. MCNP6 was utilized to compare neutronic performance, while several figures of merit have been chosen to compare overall heat transfer capabilities of each salt. Corrosivity of salts will be determined based on standard electrochemical potentials.

After determining the criteria, an initial set of potential fuel salts was produced based on elements that are neutronically acceptable. LiF, BeF2, MgF2, CaF2, NaF, KF, ZrF4, PbF2, AlF3, SnF2, ZnF2, RbF, BiF3, YF3, BaF2, SrF2, CeF2. Potential coolant fluid materials include all these salts as well as several molten metals: magnesium, lead, aluminum, and bismuth. LiF is eliminated as a prospective salt constituent due to its need for isotopic enrichment. Beryllium fluoride is extremely toxic and has a high viscosity, so it was eliminated from consideration as well. Lead, tin, and zinc fluorides are extremely corrosive and have poor heat transfer properties. While there are ways to mitigate the corrosivity with more expensive container materials, their heat transfer performance does not merit the expense necessary to do so. YF3, BaF2, SrF2, and CeF2 all have exceptionally high melting points and have mediocre heat transfer performance. The refractory alloys necessary to contain such a high melting salt appear to not be worth the mediocre performance that the salt offers. Rubidium fluoride is too expensive to use without justification and its performance as a heat transfer fluid is not exceptional. Aluminum fluoride and zirconium fluoride both offer the possibility of reducing the melting point of a salt mixture and have no disqualifying characteristics. NaF, KF, MgF2, and CaF2 offer the best heat transfer capabilities of salts which have not been otherwise disqualified. While they are all inferior to LiF in this regard, they do not require isotopic enrichment, making them orders of magnitude less expensive. Essentially all the criteria above apply to coolant selection as well, except for the need that it be a salt. Molten lead and bismuth must be excluded from consideration due to insufficient solubility of thorium at operating temperatures. Magnesium must also be excluded due to violent reactions with air and water at operating temperatures. This leaves only molten aluminum as a potential replacement for the fluoride salts. While molten aluminum has vastly superior heat transfer properties, calculations show that this is not relevant to the overall heat transfer due to great resistances in the fuel salt. This means that the extra consideration that must be given to corrosion with molten aluminum may provide very little benefit in terms of overall heat transfer. Of the salts that are not disqualified from use, neutronic calculations show that sodium fluoride can provide the highest rate of uranium production.

While many fluids are technically possible to use in a fluid fueled nuclear reactor, the numerous constraints placed on any such fluid allow for easy elimination of most of the salts. Neutronic considerations alone provide a small set of acceptable materials to work with initially. Such a small set of fluoride salts can easily be narrowed down to alkali fluorides and alkaline earth fluorides exclusively based on materials compatibility and heat transfer properties. Despite the apparently low number of options it is important to remember that this analysis makes several key assumptions: 1) high toxicity in the salt is unacceptable 2) isotopic separation for any of the constituents is unacceptable. This eliminates the use of lithium fluoride, as well as chloride salts. 3) temperatures will be above 500 °C. This eliminates the use of nitrate salts and hydroxide salts, which have fairly good properties if one is working in a lower temperature range. Even uranium containing nitric acid has been considered for use in a fluid fueled reactor. 4) release of gaseous fission products cannot be effectively managed without a non-volatile salt to contain them. This eliminates the use of several molten metals that could be superior in some ways to salts. Ultimately, none of these constraints are insurmountable and therefore this analysis is specific to a specific set of design choices. Other designs may reasonably choose to relax some of the constraints of this analysis to meet other objectives.

After individual salts were identified that could meet the design criteria for a molten salt reactor, lower melting mixtures, or eutectics were found that would be acceptable as both fuel carrier salt and coolant salt. A mixture of NaF-KF-MgF2 (34.5-59-6.5 mol%) has been identified as a promising fuel salt. The most promising coolant salt that has been identified is a mixture of sodium fluoride and thorium fluoride. There are three eutectics possible eutectics containg 22.5, 37, and 41 mol% ThF4 and the balance NaF. These eutectics have melting points of 618 °C, 690 °C, and 705 °C respectively. A design choice between the three eutectics would be determined by the relative importance of melting point and the rate of uranium production.