All fission reactors turn fissile material like uranium or plutonium into an array of fission and activation products (such as plutonium or neptunium). These fission and activation products have a wide variety of physical and chemical properties. Some are noble gases, some are metals, some are volatile at the operation temperature, and others remain part of or dissolved in the fuel matrix (i.e. oxide or fluoride). At the moment of their formation from fission they inherit the neutron-proton ratio of their parent fissile materials, which are generally very different than stable nuclei of the element(s). Thus they are furiously radioactive at the moment of their creation, throwing off beta particles as they rapidly transmute some of their neutrons into protons and altering their neutron-proton ratio. This process of rapid beta decay to stability takes place surprisingly quickly. About 90% of fission products reach a stable nuclear form within a month of their birth from fission.

Regardless of their final form or properties, fission and activation products will accumulate in the fuel of a nuclear reactor. In conventional solid-fueled reactors, the consumption of fuel, and the degradation of cladding material are generally the reasons the reactor must be shut down for refueling rather than the buildup of fission products. In a fluid-fueled molten-salt reactor, the potential exists to refuel the reactor during operation by adding fissile material to the fuel salt. The cladding degradation issue does not apply, on the contrary, molten-salt reactors that use fluoride salts as the chemical medium are impervious to radiation damage in the fuel itself, due to its ionically-bonded nature. This leaves fission product buildup as the only real threat to the long-term operation of the reactor.

A fluid-fueled reactor also has attractive options for the long-term management of fission products. They can be chemically isolated and separated from the fuel salt in a manner analogous to the way that the kidney processes and removes waste products from the bloodstream. A variety of different approaches to the removal of fission/activation products have been proposed for liquid-fluoride nuclear reactors. These include distillation of the fuel salt itself leaving the fission/activation products in the heel. Another proposed method is to selectively precipitate certain neutron-absorbing fission products by overwhelming the salt with another material transparent to neutrons, such as cerium.

Reductive extraction is yet another process, here the fuel salt is contacted with a metallic chemical reductant that will preferentially reduce fission products from fluoride salts to metals.

The MSRE considered distillation as a viable approach since it appears simple in concept. The high temperatures required for the distillation of a carrier salt of LiF-BeF2 (FLiBe) from a fuel salt are challenging. There is also the inefficiency issue of repeatedly attempting to boil a large inventory of carrier salt away in an attempt to concentrate a small inventory of fission products. Inevitably some of the valuable carrier salt will follow the fission products into the waste stream.

Precipitation with cerium is also a challenging method for fission product management because it introduces another chemical species into the carrier salt at ever increasing concentrations rather than actually removing the fission products and leaving no residual behind.

Reductive extraction of fission products increasingly appears to be the most attractive suggested way to manage the long-term buildup of fission products in the fuel salt, especially if lithium metal is used as the reductant. Because lithium is one of the constituents of the FLiBe salt that makes up the solvent into which nuclear fuel is dissolved in the reactor, its addition over time will not be detrimental and more easily managed than a foreign species such as cerium. The metallic lithium can be alloyed with metallic bismuth to carefully manage lithium’s introduction into the fuel salt; bismuth is immiscible with the fluoride fuel salts that are generally favored for molten-salt reactors. Reductive metal extraction is a technique that can either be employed in a “chemical” manner, contacting bismuth containing lithium with the fuel salt, or in an “electrolytic” manner, where an electrical potential is applied to more carefully control the addition of lithium reductant and the removal of metallic fission products.

Ideally, a technique for the removal of fission products from the fuel salt of a liquid-fluoride reactor would be employed that could operate directly on the fuel salt without any pretreatment. But the nature of reductive extraction is that it will tend to remove the most “noble” constituents of the fuel salt first, and the actinides tend to be substantially more noble than the fission products. This figure provides the overview of how extractable various species are from the LiF-BeF2-ThF4 system. Zr, U, Pu, Pa, and other actinides tend to reduce from a salt in preference to the lanthanides. Thorium is a notable exception.

Zirconium, uranium, plutonium, and several other minor actinides not depicted in this graph are all more susceptible to be removed from the fluoride salt mixture than the neodymium and europium, which represent the larger class of lanthanide fission products. Thorium which would be present in the breeder salt blanket has an even lower propensity to be removed from the fuel salt than the lanthanide fission products. But if one can imagine a fuel salt that does not contain thorium and where uranium is the dominant actinide species, a solution to the challenge may present itself.

The approach which we propose to evaluate is the use of a fluorinating/oxidizing agent to convert uranium, typically UF4 found in a liquid-fluoride reactor to its gaseous state UF6. Depending on the fluorination/oxidizing agent and temperature, other actinides will also be fluorinated and/or oxidized from a trivalent or tetravalent state. Neptunium and plutonium do form volatile hexafluorides but plutonium hexafluoride is thermodynamically unstable. ThF4 present in our 233U breeder blanket salt does not oxidize and does not form a volatile species thus facilitating separation of uranium from thorium. If fluorination could be undertaken prior to an attempt at reductive extraction, the uranium, neptunium, many of the transition metals, and non-metals present in the salt could be largely removed and reductive extraction could be employed much more productively to remove fission products.

In order for this approach to be successful, the fluorination/oxidation of the fuel salt must effectively remove uranium in a form that can easily be returned to the salt downstream of the reductive extraction process used to remove the fission/activation products. This reconstitution step involves contacting the “cleansed” FLiBe fuel salt with UF6 from the fluorinator and H2 gas. The UF6 and H2 react with one another to form hydrogen fluoride (HF) and UF4, which is soluble in the FLiBe carrier. Thus the fuel salt is restored to its previous composition, but without fission products, and is introduced back into the liquid-fluoride reactor.

The appeal of fluorination as a technique for the removal of uranium from fluoride fuel salt has been noted for many years and fluorination formed an integral part of most of the chemical processing flowsheets that were developed at Oak Ridge National Laboratory under the Molten-Salt Reactor Program from 1957 to 1976. Fluorinators were envisioned at a variety of locations in the chemical processing, universally under the assumption that they would remove uranium from the fuel salt. Despite the prevalence of fluorination as an envisioned chemical processing technique, the actual amount of development that was undertaken on continuous fluorination was surprisingly small.

Batch fluorination was utilized to remove uranium from the fuel salt of the Molten-Salt Reactor Experiment (MSRE) in 1968, but this was done in the drain tank of the reactor vessel and led to the introduction of a significant amount of corrosion products. Repeated fluorination of the MSRE fuel salt in this manner would have undoubtedly led to the structural failure of the drain tank due to corrosion.

At the heart of the challenge of fluorination is the nature of the gaseous species used. F2 is a very assertive, hazardous, highly corrosive, oxidation agent which effectively converts materials to their fully fluorinated, highest oxidation state. That behavior is simultaneously its advantage and its risk. Even ceramic (oxide) materials can be fluorinated; indeed this is how uranium oxides mined from the earth are prepared for enrichment, by first being fluorinated to uranium tetrafluoride then uranium hexafluoride.

But the aggressiveness of F2 led to many practical engineering challenges in the development of a continuous fluorination system. To protect the fluorinator from F2 attack, ORNL engineers envisioned using an extensive interior cooling system to freeze a layer of salt on the fluorination column’s inner surface. A fuel salt containing fresh fission products has considerable internal heat generation that can be opposed by a cooling system to form a frozen wall on the interior surface of a fluorination column. But a chemically-similar simulant salt, such as LiF-BeF2-UF4, where fission products are replaced with stable isotopes, has no such internal heat generation term. It was necessary to simultaneously heat the salt internally, to simulate the heating effect of fission product decay, while cooling the wall of the fluorinator to generate the frozen wall. Thus testing the frozen wall of the fluorinator under these conditions was very difficult. This was never satisfactorily resolved during the Molten-Salt Reactor Project.

In the years since the MSRE concluded in 1976, alternative fluorination agents have been advanced for consideration. Most notable among these is NF3. NF3 has been considered for rocket propulsion and is extensively used in the electronics industry to clean and etch microelectronic silica chips. It is minimally hazardous and not corrosive at temperatures below 70C and is likely less corrosive than other fluorinating agents. It is not known to react with moisture, is thermally stable at room temperature, and has been demonstrated by PNNL to be an effective, thermally-tunable fluorination/oxidation agent for spent nuclear fuel constituents. By controlling the treatment temperature, NF3 will selectively fluorinate/oxidize spent nuclear fuel constituents. In general, NF3 will cause the formation of their volatile fluorides with the exception of plutonium and rhodium; PuF6 is thermodynamically unstable relative to the non-volatile PuF4, thus requiring an overpressure of F2 to maintain PuF6. The different temperature sensitivities and NF3 concentration effects for the fluorination/oxidation of the different constituents potentially provides mechanisms to effect separations of the volatile fluorides.

At higher temperatures, such as the temperatures at which a liquid-fluoride reactor would operate (600C), NF3 is an effective fluorination/oxidation agent for UF4, NpF4, and many of the transition metal fission products that form volatile fluorides. The alkaline earths, lanthanides, and remaining actinides, including plutonium, will remain dissolved in the molten salt. This is a notable departure from the behavior of F2, which will indiscriminately convert species to their volatile forms (if such a form exists).

The hazard level, and chemical reactivity attributes potentially make NF3 a very attractive fluorinating/oxidizing agent for managing the composition of the fuel salt in a liquid-fluoride reactor where uranium is the dominant or even exclusive fissile material. Fluorination/oxidation of the fuel salt with NF3 would produce UF6 and remove uranium from the salt. Reductive extraction could then be employed to remove non-volatile fission and activation products from the salt. Hydrogen could be used to reduce UF6 back to UF4 and reconstitute the salt for return to the reactor.