Rare earth metals are widely used now in different fields of modern technologies when in high purity such as computer memory disks, rechargeable batteries, magnets, and fluorescent lighting. Electrowinning and electrorefining in molten salt are an option to manufacture high purity rare earth metals. Recycling of used nuclear fuel constitutes as another important technology related to rare earths and molten salts. This recycling is called pyroprocessing, and is a developing technology used in the nuclear fuel industry. The process operates at high temperature in a molten salt electrolyte. The goal is to recover actinides and minimize radioactive waste [1-8]. This is commonly referred to as electrorefining. The metallic fuel is dissolved anodically and the actinides are collected cathodically [9-16]. The salt in the electrorefiner (ER) contains most of the special nuclear materials. There is high interest to monitor the composition of this salt from an operational and safeguard perspective [17-21].
There are several options for monitoring the composition of the ER salt such as physical samples, in situ electrochemical sensors, and non-destructive assay. The reason for composition analysis is to ensure attractive material is not being removed from the ER as well as investigating corrosion of structure and operational equipment. These sensors are commonly developed on a bench scale experimental set up before deployment and are heavily impacted by analyte concentration. Such as electromagnetic forces will change if corrosion products are soluble in the ER salt [22-25].
Typically, the crucibles containing these molten chlorides are alumina crucibles [26-31]. Electrochemistry is a useful technique in molten salts to determine diffusion coefficients, activity coefficients, concentrations, and potentials. For electrical isolation of electrodes, ceramic tubes are used such as mullite. It is important to note these ceramics are Al2O3 (alumina) and a mixture of SiO2 and Al2O3 (mullite) and may interact with the molten salts containing rare earth chlorides through ionic exchange to drawdown the concentration of rare earth chlorides. This work focuses on investigating the interaction of alumina and mullite between YCl3, SmCl3, NdCl3, PrCl3, and CeCl3 in LiCl-KCl at 773 K.
The most likely interaction between the ceramics and the salt would be an ionic exchange, such that the rare earth chloride would form a rare earth oxide or a rare earth oxychloride. A thermodynamic database HSC Chemistry 9 was used to evaluate the possibility of ceramic and salt interaction. Table 1 reveals possible chemical reactions.
The oxidation of these rare earth chlorides is possible when there is available free oxygen molecule. The change in free energy is favorable at 773 K . The possibility of oxychlorides is also favorable when there is an available oxygen which could be removed from alumina and/or mullite. Table 2 presents thermodynamic values  from HSC 5.1.
Experimental: All electrochemical experiments were performed in an inert argon atmosphere glovebox (MBraun LABmaster dp). Conditions within the glovebox were maintained at less than 3 ppm O2 and 1 ppm H2O. No fluctuations in atmospheric concentrations were observed during the experiment. The salt mixture contained 24.6 g of LiCl-KCl eutectic (Sigma-Aldrich SAFC, anhydrous beads, 99.99%), 0.5 g of YCl3 (99.99%, anhydrous, Sigma Aldrich), 0.5 g of SmCl3 (99.9%, anhydrous, Sigma Aldrich), 0.5 g of CeCl3 (>99.99%, anhydrous, Sigma Aldrich), 0.5 g of PrCl3 (99.99%, anhydrous, Sigma Aldrich), and 0.5 g of NdCl3 (>99.99%, anhydrous, Sigma Aldrich). The salt was contained in a tantalum crucible (99.9%, 20 ml Low Form, MTI Albany) which was then heated in a Kerr furnace to 773 K. Once at temperature, an alumina tube (5 inches, Open Ended, Advalue Tech) and mullite tube (5 inches, Open Ended, Coors Tech) were inserted approximately 0.5 inches into the molten salt. Fig. 1 presents the experimental set up within the Kerr Furnace.
After exposing the ceramics to the molten chloride mixture for approximately 13 days, the tubes were removed from the furnace and rinsed with tap water to remove any adhered salt; distilled water would have been preferred but was not currently available, any impurities in the tap water should not influence the purpose of these experiments. Approximately 11 mm was cut from the rod at the end immersed in the molten salt. For observation via a Scanning Electron Microscope (SEM) with Wavelength-Dispersive X-Ray Spectroscopy (WDS) the ceramic samples were ground and polished to achieve a 1-micron finish using Buehler discs and polishing suspensions.
2. Results and Discussion
Before the insertion of alumina and mullite to the molten chloride mixture, the tubes were weighed (6.30 and 8.20 g; alumina and mullite, respectively) and their surfaces documented. Fig. 2 shows photographs of the initial surfaces as received from the manufacturers and will be used as a visual comparison for any surface change.
After approximately 13 days of exposure to the salt, the ceramic tubes were removed from the furnace. Upon removal, several observations were made, the most noticeable is a surface discoloration as shown in Fig. 3. The portion of tube immersed in salt had darkened along with a slight discoloration at the top of the tube (still within the furnace). Due to the vapor pressure of these salts at 773 K, it is possible that some salt vapors may have condensed on the cooler surfaces at the tops of the ceramic tubes. Interestingly, the middle portion of the tubes did not undergo major visual change, possibly due to the heat distribution of the tube within the furnace such that salt vapor preferentially deposited elsewhere. The ceramic tubes exhibited slight increases in mass following salt exposure and water rinse. The alumina tube increased by 2.5% to 6.46 g while the mullite mass increased by 1.1% to 8.29 g.
The tips of the exposed ceramic tubes exposed to the rare earth chloride salt were cut and prepared for observation via SEM (Oxford Instruments) with WDS (Inca wave) capabilities. The first ceramic to be viewed was Alumina. The initial image has a legend of 60 microns while the exposed surface was magnified further with a legend of 20 micron. The reason for additional magnification was to view the salt ceramic interface closer. Fig. 4(a) and 4(b) show a comparison before and after the rare earth chloride exposure of 13 days. Additionally, WDS was performed on the ceramic materials before and after the rare earth chloride exposure bath to investigate elemental differences. Alumina is a ceramic containing aluminum and oxygen (Al2O3) whereas mullite is a mixture of aluminum oxide and silicon oxide (SiO2). Fig. 4(c) shows the WDS (before) and Fig. 4(d) shows the WDS (after) the alumina was exposed to the rare earth chloride mixture for 13 days.
Comparing Fig. 4(c) and 4(d) there does not seem to be an ionically exchanged layer or trace of any rare earths on the surface and radially through the alumina. Fig. 5 shows the SEM images (a and b) and WDS results (c and d) for the mullite. Again, there does not appear to be any interaction at this magnification. The two most detectable elements were aluminum and oxygen, which are the primary components of the ceramic.
SEM comparisons provide crucial visual morphologic characteristics. For both cases (alumina and mullite) the surface appears to be visually similar in crevices and voids. Recall the discussion from the thermodynamic section, the SEM comparison verifies the thermodynamic calculations which suggested no interaction would occur. It is important to note, at higher temperatures or different conditions the result may differ.
Throughout the 13-day period of exposure, samples were periodically taken of the salt, shown in Fig. 6. These samples aid in the visual detection of any changes occurred. Initially the salt mixture was white when frozen (translucent when liquid). After 7 hours of the ceramic being immersed in the rare earth chloride mixture; a blue hue began to appear. After roughly 8 days of ceramic exposure the blue hue began to change to a green color. Visual observation of the salt samples does not specify rare earth concentration or even that changes in the concentration have occurred. However, it is important to note the observed color changes may be due to an interaction with the surface of the ceramic tubes. Fig. 6 shows images of the salt samples taken as a function of exposure time. The last image within the figure represents the salt which was inside the tube at the end of 13 days.
After the experiment was concluded the tantalum crucible was removed from the furnace and a puck of salt was found below the crucible. Fig. 7 shows photographs of the inside, the side, and the underside of the tantalum crucible after the experiment. The movement of the molten salt at this temperature could be attributed to the wetting or surface tension between the molten salt and the ceramics and/or tantalum.
This frozen puck remaining in the furnace was easily removed from the graphite liner by turning the liner upside down. Fig. 8 shows photographs of the top, side, and bottom of the recovered salt. The weight of the frozen salt puck was 15.94 g. The bottom picture of the frozen puck shows a darkened region, which is graphite from the liner.
X-Ray Fluorescence (XRF) was performed on the salt samples to investigate concentration as a function of ceramic exposure time. It is important to note the accuracy associated with X-ray fluorescence; the machine used was from Malvern PANanalytical (Epsilon model) having a resolution of 135 eV at an elemental range of C-Am. Fig. 9 presents the concentrations determined from XRF, note Aluminum and Silicon are in units of parts per million (ppm). Inferring from the results, there does not appear to be a change in the rare earth concentration in the frozen salt samples. It is true the weight percent values are less then expected but the XRF was not calibrated whereas ICP was calibrated using standard solutions. Lastly, a comparison of two salt samples was investigated; the last bulk sample taken was compared to a sample of the salt taken from inside the mullite tube. If there was a reaction to occur between the rare earth chloride salt and the ceramic tube, then the highest concentration difference should be inside the tube compared to outside the tube. Table 3 compares the concentrations found via XRF. The results between the two samples are similar, inferring no reaction occurred inside or outside the tube. It is important to note the XRF was calibrated for these elements. The main point was to observe if a difference in “relative” concentrations could be determined. However, the differences were observed.
Another piece of equipment which would accompany the XRF would be X-Ray Diffraction (XRD). XRD utilized the atomic structure of the crystal to scatter incoming xrays to produce a unique pattern. Known patters of known crystals are placed into large databases so that unknown crystal structures being analyzed can be determined. A known pattern of rare earth oxides and oxychlorides have been published [33-34]. Fig. 10 could be compared with the samples generated if XRD was performed. This method would validate oxides or oxychloride formation on the surface with the ceramics.
Throughout the experiment there was a continual change in salt depth. When each salt sample was taken, the salt level was simultaneously measured and is plotted as a function of time (Fig. 11). The mechanism for movement of the molten salt was not investigated, however the phenomena could possibly be attributed to the surface tension of the salt at the tantalum crucible surface or capillary effects.
It is evident the salt is moving out of the tantalum crucible. Recall, the salt is comprised of the rare earth chlorides and the eutectic LiCl-KCl. It is a possibility that the entire mixture is creeping up the tantalum, alumina, or mullite surface. This is analogous to a wetting effect. Further investigations could be performed between the tested surfaces and the components of the mixtures to determine a wetting angle. One such method is the Sessil drop method to determine the solid surface energies. A liquid drop is placed onto a flat surface, a contact angle can then be measured such that a higher repulsion will result in a higher contact angle. This is a relatively simple method but may be difficult when working with high temperature molten salts. It would be difficult to incorporate a transparent window into the vessel at these high temperatures.
Molten salts are gaining popularity in a variety of fields such as heat transfer fluids, thermal batteries, solvents, and coolant or fuel in molten salt reactors. As experiments and investigations are conducted in molten salts, it is important to understand the effects of the materials containing these salts. There could be possible interactions such as ionic exchange, galvanic corrosion, and other interactions which may affect the purity of these salts, which also impacts the conclusion of the experiments. This work sought to determine the interaction of alumina and mullite, which are commonly used to contain molten LiCl-KCl, at 773 K with a nominal concentration of around 9 weight percent rare earth chlorides. After 13 days of exposure to this salt the ceramics and collected salt samples showed visual changes. Intuitively, this color change would suggest an interaction between the ceramics and the salt. However, after viewing the ceramic under a scanning electron microscope with wave dispersion spectra capabilities, no interactions were observed. If a film were to have formed from ionic interaction or by other means, it was not observed in the prepared samples. An interesting phenomenon occurred during the experiment such that majority of the molten salt did not remain inside the tantalum crucible. Instead, the molten salt covered the outside of the crucible and deposited on the graphite furnace liner. This may be due to a higher than expected surface tension of this specific salt mixture and/or an interaction between the salt and the tantalum crucible or ceramic tubes.