This study focuses on techniques that can be used to fuel next generation reactors. The first two studies are new techniques for recycling used nuclear fuel (UNF) and the third is a method of separating uranium (U) from lithium fluoride (LiF) and thorium fluoride (ThF4) salt also known as FLiTh for a thorium (Th) fuel cycle. The first technique proposed for UNF recycling was to use the cladding as an anode to oxidize the zircaloy and dissolve it into a LiF, sodium fluoride (NaF), zirconium fluoride (ZrF4) salt. Zirconium (Zr) was also reduced and deposited on a tungsten (W) cathode at the same time transporting the Zr through the salt. As commercial zircaloy would be contaminated with UNF oxides, and the oxides will not oxidize as part of the electrochemical process, they would be left at the anode as the Zr is dissolved away. This means the deposited Zr, on the cathode, can be disposed of as low-level waste (LLW) or recycled back into the nuclear industry instead of being stored as high-level waste (HLW). The next technique was fluorination of UNF oxides using ZrF4. Using the same LiF-NaF-ZrF4 salt, uranium oxide (UO2), lanthanum oxide (La2O3), and yttrium oxide (Y2O3) were fluorinated into uranium fluoride (UF4), lanthanum fluoride (LaF3), and yttrium fluoride (YF3). By sampling and recording the change in concentration over time, the reaction rate of all three oxides was determined and a temperature dependent reaction rate was reported from 500°C to 650°C. A zirconium oxide (ZrO2) product layer developed on UO2, but it only slowed down the fluorination process but did not stop it. UO2 and Y2O3 fluorinated entirely but La2O3 did not. The solubility limit of LaF3 in the salt was determined to be the reason the reaction did not go to completion. The last technique was the electrochemical separation of U from FLiTh, to simulate irradiated Th that decays to protactinium (Pa). A constant, albeit small current, was used to deposit U on a W electrode without Th depositing with it. A liquid metal bismuth (Bi) electrode was used as well, and a constant current resulted in Th depositing with the U. To get just U to deposit, the current needed to be applied for a time and then no current applied for a time so the system could reach equilibrium. By cycling these two steps it was possible to get U to deposit in Bi without Th. / Doctor of Philosophy / This study focused on techniques useful to the fabrication of next generation reactor fuels. The first focus was on new techniques for recycling used nuclear fuel (UNF). Nuclear waste currently needs to be stored for hundreds of thousands of years to reach background radiotoxicity levels. If plutonium (Pu) is removed from the waste this time is limited to ten thousand years and if the other transuranics (TRU) are removed the waste only needs to be stored for 300 years to reach background radiotoxicity levels. As recycling UNF can make such a drastic difference, developing techniques for this are of utmost importance. The first technique studied was to show that the zirconium (Zr) in zircaloy cladding could be oxidized and transported through salt. This was done by applying a current between a zircaloy anode and tungsten (W) cathode, dissolving the cladding into the salt. The salt used was lithium fluoride (LiF), sodium fluoride (NaF), and zirconium fluoride (ZrF4) salt called FLiNaZr. This transported Zr through the salt and then deposited it on W. If this process was done with zircaloy contaminated with used nuclear fuel (UNF) oxides, the oxides would not dissolve into the salt as part of the process and would be left behind at the anode as Zr is transported through the salt, effectively separating the two. This alone leads to a 25% reduction in the weight of the UNF that needs to be stored. The next technique studied was converting the UNF oxides into fluorides. This was done by having it react with ZrF4 to make zirconium oxide (ZrO2) and UNF fluorides. The oxides studied here were uranium oxide (UO2), yttrium oxide (Y2O3), and lanthanum oxide (La2O3). UO2 and Y2O3 reacted until no material was left but La2O3 did not. This was due to lanthanum fluoride (LaF3) having a solubility limit in the salt that made it impossible for more to be made and stopping the reacting. The reaction rate for each oxide was found and the order of the reaction rates was Y2O3>UO2>La2O3. This process was a success and should be studied more to ensure it will work with all oxides found in UNF. The last technique studied was electrochemically separating uranium (U) from lithium fluoride and thorium fluoride (ThF4) salt. Thorium (Th) is another nuclear material, and while it cannot fission in a reactor it can be turned into an isotope of U, U-233, that can. To do this Th must be irradiated so it turns into protactinium (Pa) which can then be separated from the salt. In this study U was a surrogate for Pa as it is too radioactive to handle in this lab. First, an inert W electrode was used to deposit U metal, and once it was successful a liquid metal bismuth (Bi) electrode was used. A small constant current was able to deposit U on W without co-deposition of Th. For a Bi electrode, an alternating time of applying current and then letting the system rest was needed to deposit U without co-deposition of Th.

Future nuclear fission reactors (GEN IV) are designed to include fast breeder reactor technologies, which can accept transuranics (elements heavier than uranium) as fuel. This has the potential of being more fuel efficient but requires the closing of the nuclear fuel cycle: full recycling of existing and newly generated nuclear waste to extract uranium and transuranic elements which can be reused as fuel. In the UK a system being investigated is electrochemical pyroprocessing which uses molten LiCl-KCl eutectic (LKE), which aims to recover uranium by electrodeposition on an inert (steel) electrode and the transuranics by electrodeposition as alloys with an active metal electrode (bismuth, cadmium or aluminium). Of the three active metal candidates, aluminium has the best separation efficiency of actinides and lanthanides, which is important as lanthanides are neutron poisons and so are not to be extracted. The development of pyroprocessing requires fundamental understandings of electrochemical alloy formation, as well as on-line monitoring tools to ensure the reprocessing occurs safely and efficiently. To that end, this thesis investigates cerium-aluminium alloying (a non-radioactive model system for plutonium-aluminium) on macro- and microelectrodes to understand the limiting factors during the alloying reaction at each electrode scale and also the circumstances under which the Ce3+ concentration can be reliably determined for on-line monitoring. On a bulk aluminium macroelectrode one cerium-aluminium alloying reaction was observed. This reaction was kinetically limited by the phase change from cerium insertion into the aluminium, and resulted in lattice expansion and progressive roughening of the electrode surface. These factors made it difficult to reliably calculate the Ce3+ concentration. Li+ from the solution was also able to reduce and form alloys with aluminium, approximately 0.3 V more negative than the first cerium-aluminium alloying peak. Since lithium atoms are smaller than cerium, and there is an abundance of Li+ in the salt, lithium-aluminium alloy was found to form preferentially to cerium-aluminium alloy at these more negative potentials. By co-depositing Al3+ and Ce3+ together on a tungsten electrode which is inert under these conditions (it does not alloy), the kinetic barrier to alloy formation by cerium insertion was decreased, which is beneficial to studying the thermodynamics of alloying. Studies of pure aluminium plating and pure cerium plating showed each individual reaction was diffusion limited, with an increased contribution of convection to the mass transport at slow scan rates. Co-deposition on macroelectrodes with a low ratio of [CeCl3]:[AlCl3] showed only one cerium-aluminium alloying peak. The co-deposition currents, and ratio of oxidation peaks charges, showed that co-deposition was occurring with both species under diffusion control, resulting in an amorphous alloy with a Ce:Al ratio that smoothly varied with the [CeCl3]:[AlCl3] ratio. This was in contrast to the alloying behaviour of cerium with liquid bismuth, in which co-deposition occurred at specific ratios determined by the crystal phases that could be formed at the applied potentials, with higher co-deposition ratios being achieved at more negative potentials. Co-deposition on macroelectrodes with a high ratio of [CeCl 3]:[AlCl3] could result in up to five cerium-aluminium alloy peaks, corresponding to all five CexAly crystalline phases predicted by the phase diagram. This phase change from amorphous to crystalline was promoted by the high Ce:Al ratio in the amorphous alloy resulting from the high [CeCl3]:[AlCl3] ratio and by plating pure cerium on the surface, which could then insert into the alloy. Charge analysis of these peaks confirmed the expected stoichiometries of the crystal phase from these in-situ measurements which is important for rapid analysis, whereas all previous literature has relied on ex-situ techniques which cooled the alloy, possibly changing its composition and structure. In all circumstances of alloy formation on macroelectrodes, the rate of reduction of Ce3+ was time dependent and sensitive to convection. This significantly complicated analysis of the electrochemical signal, making it very difficult to reliably calculate the concentration of Ce3+, which is required for on-line monitoring. Co-deposition on in-house microfabricated tungsten microelectrodes resulted in steady state currents for both pure aluminium deposition and cerium-aluminium co-deposition (up to the beginning of lithium-aluminium alloying). Thus, unlike on macroelectrodes, the deposition rate occurred at the flux ratio of each species from solution and only one oxidation peak was observed corresponding to the amorphous cerium-aluminium phase, even at high [CeCl3]:[AlCl3] ratios. The steady state alloying current meant that calculating the Ce3+ concentration was relatively simple from co-deposition on microelectrodes. Co-deposition was highly beneficial for studying alloying, however to avoid the addition of Al3+ to the molten salt, in-house microfabricated thin film aluminium microelectrodes were also used to study alloying. Alloying on microfabricated thin film aluminium microelectrodes was hampered by the formation of a native aluminium oxide layer, which prevented cerium insertion into the aluminium. The oxide layer could be disrupted by reduction of lithium, which showed steady state currents (albeit with significant capacitance) could be achieved for alloying by cerium insertion. However, the full surface area of the microelectrode could not be attained and all microelectrodes lost their aluminium layer after multiple lithiation/de-lithiation cycles. These devices need further development to overcome the oxide layer, or prevent its formation, in order to study alloying in greater detail with aluminium microelectrodes to fully realise their advantages for sensing and monitoring in pyroprocessing.