Electrochemical Removal of Rare Earth Element in LiCl-KCl Molten Salt

This study was carried out to examine the removal of rare earth (RE) elements by electrodeposition for the puriﬁcation and reuse of LiCl-KCl salt after electroreﬁning and electrowinning. The electrochemical behavior of RE elements (Dy and Gd) in LiCl-KCl-DyCl 3 -GdCl 3 at 500 ° C was investigated using the cyclic voltammetry (CV) technique using Mo and Mg electrodes. It was observed that the reduction potential of the RE elements shifted at the Mg electrode owing to the alloy formation with Mg (RE-Mg alloy). Subsequently, a series of potentiostatic electrolysis tests were conducted to remove the RE elements in the salt and check the formation of deposits at the Mg and Mo electrodes. The scanning electron microscopy-energy dispersive X-ray spectroscopy (SEM/EDS) technique was used to conﬁrm that the reduced RE metals were deposited on the surface of the Mg electrode. However, no signiﬁcant deposit on the Mo electrode was observed, and a mud-like deposit was found on the bottom of the electrochemical cell. The salt analysis performed by employing the inductively coupled plasma-optical emission spectrometry (ICP-OES) indicated that the removal eﬃciency of Dy 3+ and Gd 3+ through electrodeposition was 83.5 ∼ 95.2 and 91.6 ∼ 95.2%, respectively.


Introduction
Pyroprocessing is an option which enables to recover the useful nuclear elements from spent fuel and to recycle them for fast reactors [1,2]. Pyroprocessing consists of several unit processes to separate, recover, and store the radioactive elements in the spent fuels [3,4]. After head-end process, the spent oxide fuel is immersed and reduced in 650°C LiCl at the electrolytic reduction process. en, the reduced spent fuel is transferred into an electrorefiner which contains a eutectic LiCl-KCl melt with some portion of UCl 3 at 500°C. Uranium is recovered using solid cathode (i.e., electrorefining) [5][6][7] and uranium and transuranic elements are recovered using liquid cathode (i.e., electrowinning) [8][9][10]. After both of the recovery steps such as electrorefining and electrowinning, there are abundant amounts of rare earth (RE) elements in the salt. Consequently, the molten salt electrolyte containing RE elements should be properly treated to minimize salt waste, to fabricate a durable waste form, and to reuse the purified salt electrolyte in electrorefining and electrowinning. Besides, RE elements are key elements in industries and often used as additives to improve properties of alloy such as thermal and mechanical properties and corrosion resistance [11].
To separate RE elements from the salt, various processes have been employed, such as ion-exchange [12], zone freezing method [13], and phosphate precipitation [14,15]. However, these require a specific device for solid salt detachment and/or layer separation and an additional process such as distillation. Some researchers attempted to extract RE from the molten solvent via electrochemical methods. ey have used an Al electrode in a chloride salt [16][17][18][19], Ni and Cu electrodes in a fluoride salt [20], W electrode in MgCl 2 -containing salt [21], and Mg electrode [22][23][24]. Especially, Yang et al. reported that the selective extraction of Dy was carried out using Mg electrode [22]. We thought that the electrochemical removal of RE in salt enables to the purification and reuse without salt transfer step from electrorefiner/electrowinner to another reactor. e objective of the present study was to examine the removal of RE elements by electrodeposition using Mo and Mg electrode. e use of Mg was considered as it is a reactive material that forms RE-Mg alloys and Mo is an inert electrode widely used in electrochemical experiments. Cyclic voltammetry (CV) measurements were obtained to observe the redox peaks of RE elements by using Mo and Mg electrodes in LiCl-KCl containing DyCl 3 and GdCl 3 at 500°C. Subsequently, a series of constant-potential electrolysis tests was performed to remove Dy 3+ and Gd 3+ in the salt and a cross-sectional analysis of deposit was carried out by SEM/EDS technique.

Materials and Methods
All the experiments were performed in a glove box with an Ar (g) atmosphere (H 2 O < 1 ppm; O 2 < 3 ppm). A mixture of LiCl-KCl eutectic salt (99.99% purity, Sigma-Aldrich) and RECl 3 (RE: Dy and Gd, 99.99% purity, Alfa-Aesar) was prepared in an alumina crucible (30 mm inner diameter, 50 mm height). ree-electrode system was used for the experiment. A Mo wire (1 mm diameter, Nilaco) and Mg plate (2 mm × 5 mm, Alfa-Aesar) was used as a working electrode. e electrode surface area was determined by width, thickness, and dipping height. A glassy carbon rod (3 mm diameter, Alfa-Aesar) and a Ag/AgCl (1 mol%) in the Pyrex tube was used as a counter electrode and a reference electrode, respectively. All the tests were conducted at 500°C. e electrochemical experiments were carried out by Bio Logic SP-300 potentiostat/galvanostat. e electrodeposition tests were carried out by applying a constant potential. e applied potential value for each electrodeposition run was adjusted with respect to be more slightly negative than EMP before run. Table 1 shows the detailed experimental conditions. e cross-sectional image and the elemental analysis of the RE deposit were observed with scanning electron microscopy (SEM; Hitachi, S-8010) in conjunction with energy-dispersive X-ray spectroscopy (EDS; Horiba, EX-250 X-max). e salt composition was analyzed via an inductively coupled plasma-optical emission spectrometer (ICP-OES; ACTIVA M, Horiba Jobin Yvon SAS).

Results and Discussion
e reduction potential of Dy 3+ and Gd 3+ in LiCl-KCl at 500°C was identified using a Mo and Mg electrode. Figure 1 shows the cyclic voltammogram using Mo electrode (surface area: 0.165 cm 2 ) with a scan rate of 100 mV/s. e cathodic peaks at about −1.92 V and −2.04 V, represent the reduction reactions of Dy 3+ to Dy and Gd 3+ to Gd, respectively. In case of Mg electrode (surface area: 0.502 cm 2 ), the reduction peaks were not clearly observed, as shown in Figure 2. It is because of the formation of RE-Mg alloy. At a scan rate of 3 mV/s, low height of two peaks were found at −1.88 and −1.94 V, which were assumed to be the formation of Dy-Mg alloy and Gd-Mg alloy, respectively [22]. e phase diagram shows that one Dy-Mg alloy (DyMg) and four Gd-Mg alloys (GdMg, GdMg 2 , GdMg 3 , and GdMg 5 ) can exist at 500°C [25], but one peak of Dy-Mg alloy and one peak of Gd-Mg alloy were observed in this study.
A series of potentiostatic electrolysis tests were subsequently conducted to remove the RE elements (Dy 3+ and Gd 3+ ) in LiCl-KCl-0.15 wt% DyCl 3 -0.59 wt% GdCl 3 at 500°C using Mg and Mo electrode. Figure 3 shows the results of CV measurements before and after each electrolysis run. In run 1, the Mg plate (surface area: 1.04 cm 2 ) was used as a working electrode to recover the RE elements in the form of RE-Mg alloy and to remove these elements from the salt by applying −1.9 V for approximately 12 h. As shown in Figure 3, the redox peak heights of Dy 3+ and Gd 3+ clearly decreased after the run 1. However, another anodic peak appeared at −1.82 V, likely, because of the presence of Mg 2+ in the salt, which was contributed by the Mg electrode. In particular, owing to the small size of the employed experimental cell, the distance between the anode and cathode (10∼15 mm) was not sufficient to ensure that the cathodic reaction was independent of the anodic reaction. Because the Cl 2 (g) generated on the graphite anode surface may affect the Mg cathode, it results in the producing of MgCl 2 in the salt [24].
is was confirmed by the salt analysis by ICP-OES technique, as shown in Figure 4. Initially, Mg 2+ did not exist in the salt, but after run 1, the concentration of Mg 2+ increased to 0.16 wt%. To prevent the dissolution of Mg electrode, the cathode was changed to inert Mo electrode in the subsequent electrolysis runs. In run 1, Dy 3+ was decreased from 0.15 to 0.12 wt% and Gd 3+ was decreased 0.59 to 0.40 wt%. Even though the applied potential was −1.9 V, which is slightly positive than the reduction peak potential of Gd 3+ on Mg electrode, those two elements were reduced. is result was not consistent with the study of Yang et al. [22]. In their  report, they presented the extraction of Dy from LiCl-KCl melt containing DyCl 3 and GdCl 3 using Mg electrode. However, here, selective reduction of Dy was not examined. In run 2, the applied potential was −1.90 V, which is the same potential value with run 1, using Mo wire electrode (surface area: 0.32 cm 2 ). After 11 h, the cyclic voltammetry was measured to check the removal of rare earth elements in the salt. As shown in Figure 3, the cathodic peaks decreased slightly, while the anodic peaks distinctly decreased. After the electrolysis for 11 h, the salt composition of Dy 3+ , Gd 3+ , and Mg 2+ was decreased as 0.08, 0.22, and 0.08 wt%, respectively. In the runs 3∼5, the cathode was replaced from Mo wire to Mo coil to increase the electrode surface area (3.37 cm 2 ) and to remove the elements in the salt as much as possible. e applied potentials of run 3∼5 were negatively increased from −1.90 to −1.94 V because the EMP value before each run was shifted. For example, the EMP values before run 3∼5 were −1.88, −1.91, and −1.93 V vs. Ag/AgCl. During the run 5, much higher current flowed so the electrolysis was stopped at 5 h of operation. After the electrodeposition runs, the redox peaks of the rare Earth elements in the salt were clearly diminished (Figure 3). rough the electrodeposition runs 3∼5, the concentration of Dy 3+ and Gd 3+ was decreased to 0.016 and 0.040 wt%, respectively, and the Mg 2+ was not detected. e removal efficiency of Dy 3+ and Gd 3+ was calculated to be about 83.5∼95.2 and 91.6∼95.2%, respectively, even though the current efficiency was very low. To overcome the low current efficiency, the anodic reaction such as the chlorine gas evolution needs to be separated from the cathodic reaction such as the electrodeposition of metal ions in the salt. Also, an addition of Cl 2 (g) capturing unit may work to prevent the rechlorination of the metal deposits and to protect the Mg electrode. e photographs of Mg electrode before and after the electrodeposition run 1 were presented in Figure 5. As shown in Figure 5(b), the RE deposit on Mg electrode was visually examined. e cross-sectional area of the RE deposit on Mg electrode was analyzed by SEM-EDS. Figure 6 is the cross-sectional SEM image of RE deposited Mg electrode, as shown in Figure 5(b). Two layers were observed in the crosssectional image by SEM (Figure 6(a)) and the layer consists of Dy, Gd, and Mg confirmed by EDS. It indicates the RE deposits were formed on the Mg surface ( Figure 6(b)). e outer layer had slightly higher amount of RE, but it was not significant. It may be due to the interdiffusion of RE elements through Mg bulk phase. Also, there was no big difference in the concentration profiles of Dy and Gd in the cross-sectional area in Figure 6(b), which is expected from the salt analysis result. In case of Mo electrode, a distinguishable deposit on the electrode surface was not found, but black-mud like compound was observed on the bottom of the cell and it was difficult to collect as mentioned in the previous study [26]. erefore, it needs to be consider to use of the alloy forming material and its ion for the attainment of RE recovery.

Conclusions
In this work, the electrochemical behavior of DyCl 3 and GdCl 3 in a LiCl-KCl eutectic melt at 500°C was examined by performing CV and electrodeposition tests using Mo and Mg electrodes. e CV results exhibited a shift in the reduction potentials of Dy 3+ and Gd 3+ when using the Mg electrode, owing to the formation of RE-Mg alloys. Furthermore, potentiostatic electrolysis tests were performed to remove Dy and Gd from the salt. e ICP-OES analysis of the salt samples indicated that Dy 3+ and Gd 3+ in the salt is removed using both of the Mg and Mo electrode with the removal efficiency of Dy 3+ and Gd 3+ was obtained as 83.5∼95.2 and 91.6∼95.2%, respectively. e low current efficiency could be improved by adopting a Cl 2 (g) venting system to prevent the rechlorination, protect the cathode material, and separate the electrodeposition reaction from the anodic reaction. e SEM-EDS analysis of the deposited RE layer on the Mg electrode clarified that the reduced Dy and Gd existed in the form of RE-Mg alloy.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.