Study of Potentiometry for Monitoring Activity of GdCl3 in Molten LiCl-KCl Salt

In pyroprocessing spent nuclear fuels by electrorefining in molten LiCl-KCl salt, it is desired to monitor in real time the UCl3 concentration in the salt for safeguards purposes. Current chemical analysis of the highly radioactive salt for electrorefining by an inductively coupled plasma technique is inconvenient and usually time-consuming in generating the salt composition results. In this paper, we evaluated whether a simple potentiometry approach can be used for real-time monitoring the concentration of GdCl3, which was used as a surrogate for UCl3, in LiCl-KCl-GdCl3 salt by measuring the open circuit potential of a Gd metal electrode with respect to a Ag/AgCl reference electrode (RE) when GdCl3 salt was incrementally added to the LiCl-KCl salt. Additions of LaCl3, CeCl3 and NdCl3 salts were used for evaluating the effects of other chloride salts on the selectivity of the Gd metal electrode vs Ag/AgCl RE. While using potentiometry to determine GdCl3 concentrations, Gd metal was unexpectedly observed to be unstable and dissolved in LiCl-KCl salt when GdCl3 is present.

Some flowsheets for pyroprocessing of spent oxide fuels (based on UO 2 or UO 2 /PuO 2 compositions) involve two electrochemical operations: oxide reduction and uranium electrorefining. 1,2 During oxide reduction, oxide fuel is electrochemically reduced to metal. This typically involves a LiCl salt with 1 wt% Li 2 O at 650°C; the oxide fuel is reduced to metal at a cathodic electrode and oxygen gas is liberated from the molten salt at an anodic electrode. During uranium electrorefining, purified uranium metal is electrochemically recovered from the metal produced during oxide reduction. This usually involves a LiCl-KCl eutectic salt with about 6 wt% UCl 3 at 500°C; the reduced metal is oxidized at an anodic electrode and electrorefined uranium metal is recovered at a cathodic electrode. Metals that are more reactive than uranium accumulate in the molten salt as metal chlorides. For safeguards purposes, it is necessary to know the concentrations of metals in the salt, particularly the concentrations of UCl 3 , PuCl 3 and other actinide chlorides.
The composition of the LiCl-KCl salt for electrorefining can be analyzed by using techniques such as inductively coupled plasma optical emission spectrometry (ICP-OES), 3,4 inductively coupled plasma mass spectrometry (ICP-MS), 5 and thermal ionization mass spectrometry. 6 However, because the electrorefining operations are conducted in hot cells, and because electrorefiner salts are highly radioactive, analytical results by these means are inconvenient to obtain. Special facilities and equipment are required just to collect the salt samples and transfer them to the analytical laboratory, where they are digested and diluted prior to undergoing analysis by the instruments. This is typically a very expensive, time-consuming process. Nevertheless, it is the most common means of analyzing radioactive salt samples at many institutions, including Idaho National Laboratory (INL) which has the country's only pyroprocessing facilities for used nuclear fuels.
Potentiometry is a well-established method for analyzing both aqueous and molten salt electrolytes. [7][8][9][10] It is based on the measurement of open circuit potentials (OCPs), also known as electromotive force (EMF) potentials, of an electrochemical cell. The OCP is dependent on temperature and the chemical activities of the species involved in the two half-cell reactions that comprise the overall cell reaction. Typical potentiometry measurements use a reference electrode (RE) and a working electrode (WE) to complete the electrochemical cell. For example, the common EMF series in aqueous electrochemistry is based on OCP measurements of various half-cell reactions (under standardized conditions of temperature and concentration) relative to a standard hydrogen electrode. In simple electrochemical systems, in which the half-cell reactions at the reference and working electrodes are thermodynamically fully characterized, the OCP measurement can be used to determine the activity of electroactive species at the working electrode. A pH meter for measuring hydrogen ion activity in water is one such example. 11,12 A potentiometry approach is seldom used for chemical analysis in molten salt systems, probably due to the fact that molten salt systems are complicated and a good working electrode that selectively responds to the component to be tested is difficult to find. Recently, Bagri and Simpson 9 studied the activity of GdCl 3 in LiCl-KCl salts and the effect of CsCl on the activity of GdCl 3 , using a GdCl 3 solute reference electrode and reported the activity coefficient of GdCl 3 compared to literature data. Their results showed that at the GdCl 3 concentration range of 0.4 to 1.5 mol%, the activity coefficient of GdCl 3 appeared reasonably constant; at a higher GdCl 3 concentration range, the activity coefficient increased with GdCl 3 concentration. However, it is unknown from their study how other elements besides Cs would affect the OCP measurements. It is possible that some elements may affect the potential of a Gd metal electrode, while other elements may not.
In the LiCl-KCl salt used for electrorefining spent nuclear fuels, the concentration of UCl 3 is usually around 6 wt% (or 0.01 molar fraction). We hypothesize that if the activity coefficient of UCl 3 is constant at a relevant range that covers 6 wt% and if the activity coefficient is not significantly affected by other chloride salt components such as NdCl 3 , CeCl 3 , and LaCl 3 in the molten LiCl-KCl salt, it is possible to use a potentiometry method to measure the UCl 3 concentration and, thereby, detect any significant deviation of the normal salt composition during electrorefining. To test the hypothesis, some surrogate work was initially proposed to verify the feasibility of this method.
In this study, we report on the feasibility of a potentiometry approach-using GdCl 3 and Gd metal as surrogates for UCl 3 and U metal respectively-for monitoring the GdCl 3 concentration in LiCl-KCl salt and determining the activity coefficient of GdCl 3 . We also evaluate whether the presence of NdCl 3 , CeCl 3 and LaCl 3 , the three main salt components in the salt for electrorefining, would affect the activity of GdCl 3 . Additional phenomena observed during the experimental potentiometry study are also reported.

Experimental
Chemicals.-The chemicals used in the experiments were as follows. LiCl-KCl eutectic salt and GdCl 3 salt (99.99 wt% pure) z E-mail: guoping.cao@inl.gov were ultra-dry, contained in glass ampules. CeCl 3 , LaCl 3 and NdCl 3 salt were anhydrous and all with 99.9 wt% purity. (Usually, anhydrous salts have a relatively higher moisture content but are less expensive than ultra-dry salts contained in glass ampules.) LiCl-KCl-1 wt% AgCl salt and 99.99% Ag wire were used for fabricating a Ag/AgCl RE, and 99.99% Gd metal foil, 1 mm × 25 mm × 25 mm, was used as a working electrode. All chemicals were purchased from Alfa Aesar, except the Gd metal foil, which was purchased from Sigma Aldrich.
Experimental procedures.-The experimental setup for the potentiometry measurements is schematically shown in Fig. 1. A 1-mm diameter Ag wire inserted in LiCl-KCl-1 wt% AgCl salt contained in a mullite closed end tube (5 mm inside diameter and 1 mm wall thickness) was used as a Ag/AgCl RE. The top of the mullite tube was sealed with epoxy.
203 g of LiCl-KCl eutectic salt was added to an alumina crucible of 93 mm inside diameter, and then the crucible was loaded into a furnace housed in a glovebox filled with argon gas. After the Ag/ AgCl RE was inserted into the salt powder, the crucible was heated to about 769 K and maintained at that temperature for more than 15 h to ensure that the Ag/AgCl RE is in full equilibrium with the molten salt.
To make the Gd metal electrode, a small piece of Gd foil (1 mm thick) was first cut, in the glovebox, using scissors and then clamped using a stainless steel rod with a split end. The surface of the Gd metal foil was lightly filed so that the surface oxide (Gd is sensitive to air and moisture) was removed and a fresh Gd surface contacted the salt. Next, the Gd metal electrode was slowly inserted into the molten LiCl-KCl salt. As shown in Fig. 1, only the Gd metal part was immersed in the molten salt.
The OCP between the Gd metal electrode and the Ag/AgCl RE was measured and recorded by a Solartron 1287 potentiostat with a sampling rate of once per second. After the OCP reading was stable or the system was in equilibrium, a small amount of GdCl 3 salt was added. During the potentiometry measurements, it was difficult to know, based on the OCP measurements, whether or when the system has reached equilibrium. Based on experimental experience, typically when the OCP change is less than 0.5 mV in two hrs, we considered the system has reached equilibrium. A stable OCP measurement is important because for GdCl 3 monitoring purposes, the concentration determined by potentiometry is dependent on the OCP measurement and is exponentially proportional to the OCP value. Following the GdCl 3 salt addition and stable OCP reading, another GdCl 3 addition was made. This process continued until the sixth or final GdCl 3 concentration of~0.014 molar fraction in the LiCl-KCl salt was reached. A final GdCl 3 concentration of~0.014 molar fraction was selected in this study so that the molar fraction of GdCl 3 was slightly higher than the typical molar fraction of UCl 3 in the LiCl-KCl salt for electrorefining.
To evaluate the selectivity of a Gd metal electrode or how other typical salt components in the ER salt might affect the results of the potentiometry measurement, after a stable OCP reading was achieved upon the above sixth (final) GdCl 3 addition, small amounts of LaCl 3 , CeCl 3 and NdCl 3 salts (molar fraction for LaCl 3 , CeCl 3 and NdCl 3 salt is 0.0014, 0.0028, and 0.010 respectively) were added, in steps, to the LiCl-KCl-GdCl 3 salt. LaCl 3 and CeCl 3 were added two times and NdCl 3 was added four times. The concentrations of LaCl 3 and CeCl 3 added were similar to those of the salt for electrorefining reported by Karlsson et al. 13 All the above experiments were carried out in a glovebox filled with an inert argon atmosphere in which oxygen and moisture impurities were maintained below 10 ppm and 0.1 ppm, respectively. To achieve the best measurements and repeatability, the salt system (including the Ag/AgCl RE and Gd electrode) was kept undisturbed as much as possible. Furthermore, the salt was not stirred after each addition of GdCl 3 and subsequent CeCl 3 , LaCl 3 and NdCl 3 salts.

Results and Discussion
OCP results in LiCl-KCl salt with different GdCl 3 salt concentrations.- Figure 2a shows the overall OCP measurements vs time for when GdCl 3 salt was incrementally added to the LiCl-KCl salt. It can be seen that after each GdCl 3 salt addition, the OCP value increased. A closer observation of the OCP vs time data revealed that the time to reach a stable potential increased with each subsequent addition of GdCl 3 salt. For example, after the first GdCl 3 salt addition, the time to reach equilibrium was less than one hour, whereas the time to reach equilibrium following the second GdCl 3 addition was about two hours. This trend continued all the way up until the sixth GdCl 3 salt addition, at which point, the time needed to reach equilibrium exceeded six hours.
To more clearly show the OCP changes between the 5th and 6th GdCl 3 additions, the OCP results for that time duration were magnified, as shown in Fig. 2b. Some small fluctuations in the OCP readings were observed. The standard deviation of OCP readings is about ±0.9 mV over the duration of 54 h (from 60 h to 114 hr in Fig. 2b). The small fluctuation was likely due to other factors (such as small temperature changes in the furnace, etc.).
To evaluate whether the OCP results responded to GdCl 3 concentration as expected from the Nernst equation, the OCP data were plotted vs the molar fraction of the GdCl 3 in the LiCl-KCl salt, as shown in Fig. 3. It can be seen that the OCP responded logarithmically to the molar fraction of the GdCl 3 , with a high R 2 value of 0.999. And at the melt temperature of 769 K, the relationship between the OCP (V) and molar fraction of GdCl 3 (X GdCl3 ) can be expressed as where R is the gas  constant, T is the temperature in K, and F is the Faraday constant 96485 C/mol. This suggests that, for the tested GdCl 3 concentrations of 0.0005 to 0.014 molar fraction, the activity coefficient of GdCl 3 in the LiCl-KCl-GdCl 3 salt is constant; therefore, the potentiometry technique can be considered applicable for real-time monitoring of the GdCl 3 concentration of a ternary LiCl-KCl-GdCl 3 salt system and determination of the activity coefficient of GdCl 3 , when no other chloride salts are present.
Effects of LaCl 3 , CeCl 3 and NdCl 3 on Selectivity.- Figure 4a shows the OCP measurements vs time for when LaCl 3 and CeCl 3 salts were added into the LiCl-KCl-0.014 GdCl 3 (in molar fraction) salt, after the six incremental GdCl 3 addition. The OCP did not appear to change after LaCl 3 and CeCl 3 additions. Therefore, LaCl 3 and CeCl 3 , in the tested amounts and LiCl-KCl-GdCl 3 composition, can be considered to have no impact on OCP measurements.
Unlike LaCl 3 and CeCl 3 , after each NdCl 3 salt addition, the OCP increased immediately and then decreased over time and reached a stable value, about 2.5 mV higher than that before the NdCl 3 addition, as shown in Fig. 4b. After a total of four NdCl 3 additions were made, an OCP increase of 10 mV was observed. This shows that the presence of NdCl 3 can affect the potentiometry technique for real-time monitoring of LiCl-KCl-GdCl 3 salt.
The reason for the effect of NdCl 3 additions on OCP measurements is not completely understood. There are two possibilities: (1) the NdCl 3 increased the activity coefficient of GdCl 3 or (2) some chemical reaction between the Gd metal and NdCl 3 occurred, causing the formation of GdCl 3 and dissolved Nd metal.
A voltammetry study conducted by Misra et al. 14 in LiCl-KCl-5 wt% LnCl 3 (Ce, Gd, La, and Nd etc.) salt system at 773 K found that the reduction potential (vs Ag/AgCl RE) for Nd 3+ /Nd and Gd 3+ /Gd couples is −1.88 V and −1.93 V respectively. The more positive reduction potential for Nd 3+ /Nd than for Gd 3+ /Gd suggests that the reaction 3NdCl 3 + Gd = GdCl 3 + 3NdCl 2 or NdCl 3 + Gd = GdCl 3 + Nd is possible. The reduction potential for La 3+ /La and Ce 3+ /Ce couples is −1.96 V and −2.11 V, respectively, both of which are more negative than that for Gd 3+ /Gd couple, suggesting the reaction between LaCl 3 or CeCl 3 with Gd is unlikely.
Activity coefficient of GdCl 3 in LiCl-KCl-GdCl 3 salt system.-For the purposes of monitoring the GdCl 3 concentrations, if we have a reliable calibration and the activity coefficient of GdCl 3 is constant in the LiCl-KCl-GdCl 3 salts, we can measure the GdCl 3 concentration based on the OCP measurement without knowing the activity coefficient, because the activity coefficient has already been included in the equation for concentration measurement, as shown in Eq. 1. But it is also useful to know the activity coefficient-an important thermodynamic parameter for molten salt chemistry-and compare this number with others' data. If the activity coefficient of GdCl 3 determined by the OCP measurement method in this study is consistent with literature values that were calculated using other methods such as voltammetry or OCP measurement using similar reference electrodes, that means the potentiometry based on Gd vs Ag/AgCl RE for GdCl 3 concentration measurement is quite reliable. If it is inconsistent with established literature values, we need to understand the underlying reasons for the inconsistency, and the feasibility of potentiometry based on Gd vs Ag/AgCl RE for GdCl 3 concentration measurement may be questionable.
Because of the different experimental testing conditions and various standard states used in literature, the derived activity coefficients achieved differ drastically. Bagri and Simpson, in a recent study, summarized the reported activity coefficient of GdCl 3 salt in LiCl-KCl salts, and compared their results using a Gd metal electrode and a reference electrode with Gd metal in a GdCl 3 -saturated LiCl-KCl salt. 9 To calculate the formal potential and activity coefficient of GdCl 3 , its Gibbs free energy of formation must be known, and it is needed to convert the Gd metal electrode potential vs the Ag/AgCl RE to a potential vs the Cl 2 /Cl − RE. For convenience, the potential difference of Ag/AgCl RE with respect to a Cl 2 /Cl − electrode from a Based on the thermodynamics of the electrochemical cell, Gd|LiCl-KCl-xGdCl 3 ||mullite membrane||LiCl-KCl-1wt%AgCl|Ag, the half-cell electrochemical reactions (both in reduction forms) are as follows: is −1.21 at 773 K, which is very close to the measured salt temperature (769 K) in the present study. Using a standard Cl 2 /Cl − reference, with a potential of zero, the Eq. 4 can be simplified as:  where / E GdCl Gd 0 3 is the standard reduction potential of GdCl 3 . In molten salt electrochemistry, a pure salt at liquid state is normally used as the standard state. But because of test temperature at 769 K is below the melting point of GdCl 3 , usually a hypothetical supercooled liquid at the test temperature was used as standard state. Some literature was available on calculating the standard electrode potential for the Gd/GdCl 3 couple using super-cooled liquid state as standard state. For convenience, the calculated standard potential  16 At the test temperature (769 K) in this study, the standard potential of GdCl 3 is −2.810 V.
According to the experimental OCP data in this study, the calculated activity coefficient of GdCl 3 is 1.7 × 10 −4 in the LiCl-KCl-GdCl 3 salt system, at the tested GdCl 3 concentrations of 0.0005 to 0.014 molar fraction. This activity coefficient based on OCP measurements is similar to the reported data of 1.7 × 10 −4 for 2 wt% GdCl 3 at 773 K by Tang and Pesic 17 and 1.5 × 10 −4 for 0.17 mol% GdCl 3 at 771 K by Caravaca et al. 18 with voltammetry technique and Ag/AgCl RE being used in both cases, but higher than that measured (5.3 × 10 −5 for 1.5 wt% GdCl 3 at 773 K) by Bagri and Simpson, 9 using saturated GdCl 3 salt in LiCl-KCl contained in a mullite tube as reference electrode. The comparison with typical literature data in terms of GdCl 3 activity coefficient in LiCl-KCl-GdCl 3 salt suggests that if the same or similar Ag/AgCl RE is used, the calculated GdCl 3 activity coefficient measured by multiple institutions is consistent and quite close. This may also suggest that potentiometry using Gd vs Ag/AgCl RE can be considered reliable.
Stability of the Gd electrode in LiCl-KCl-GdCl 3 salt system.-During the experiments, the Gd metal electrode was found to exhibit dissolution into the salt. Figure 5 shows a photograph of a Gd electrode used before and after the experiment, clearly showing that most of the Gd metal foil was gone. Although the initial weight of the Gd metal was not measured, based on the size estimate and density (7.9 g cm −3 ) of the Gd metal, about 1.5 g Gd metal dissolved into the about 230 g LiCl-KCl-0.014 GdCl 3 −0.0014LaCl 3 −0.0028CeCl 3 −0.010NdCl 3 (in molar fraction) salt, which was the final mass and composition. To our knowledge, the phenomenon of the dissolution of Gd metal into LiCl-KCl-GdCl 3 has not been reported in the literature. In some transient measurements such as cyclic voltammetry (CV), Caravaca et al. studied LiCl-KCl-GdCl 3 salt system and found the very sharp current peak of deposited Gd on a W working electrode, which Caravaca et al. suggested was due to the dissolution of the cathodically deposited Gd. 18 Other researchers similarly identified sharp CV peaks, suggesting the possibility of Gd dissolution in the salt. 17,19 Upon completion of experiments, a small amount of salt was transferred out of glovebox to a fume hood. When water was added into a beaker containing the salt, strong bubbling was observed. Likely due to the reaction between Gd with water, hydrogen gas was generated and result in the bubbling.
The disappearance of the Gd metal suggests that in a LiCl-KCl salt containing GdCl 3 , the Gd metal may have some solubility in the salt. In other salt mixtures of a metal and a metal's halide liquid, some solubilities of metal in the molten salts have been reported. For example, in an Oak Ridge National Laboratory report, 20 Bredig listed some metal-metal halide phase diagrams, clearly showing the solubility of some metals in their metal halides. It has been known that Li metal has some solubility in LiCl-1wt% Li 2 O salt, 21 and Be has some solubility in LiF-BeF 2 salt. 22 Upon a close examination of the crucible after completing the experiments, and of unpublished behavior of metal uranium in LiCl-KCl-UCl 3 salt for electrorefining, the idea that Gd has some solubility in a LiCl-KCl-GdCl 3 salt system seems perhaps somewhat inaccurate. In the LiCl-KCl-GdCl 3 salt after potentiometry measurements, some black deposits which were likely Gd metal deposits were observed at the bottom of the alumina crucible. For the LiCl-KCl-UCl 3 salt for uranium electrorefining, when uranium dendrites contained in a stainless steel anode basket were in contact with LiCl-KCl-UCl 3 salt at 500°C for quite a long time (about one month or longer), the uranium dendrites had disappeared; however, some uranium dendrites were found at the bottom of the crucible containing the LiCl-KCl-UCl 3 salt.
The underlying reason behind the Gd dissolution in LiCl-KCl-GdCl 3 salt remains unknown; however additional testing is being performed. One possible explanation is that the exchange current might be high. When the Gd metal contacts liquid LiCl-KCl-GdCl 3 salt, an equilibrium between Gd and GdCl 3 in the salt exists: Ideally, when the exchange current is low, the Gd 3+ ions formed from the Gd metal dissolution would deposit back onto the Gd metal, thus forming an equilibrium. But when the Gd exchange current is high, some Gd 3+ ions, instead of depositing back to the Gd metal surface, diffused into the salt and slowly sank at the bottom of the crucible, continuing the dissolution of Gd.
ICP-MS is typically used for analyzing the salt compositions. Upon the completion of the experiments, a salt sample was analyzed by ICP-MS and the result was shown in Table I. The concentration (in molar fraction) of LaCl 3 , CeCl 3 and NdCl 3 measured by ICP-MS is consistent with the calculated values based on the mass of the salts added during experiments. However, the GdCl 3 molar fraction by ICP-MS is 14.2% greater than the calculated value based on salts added in experiments, because of the presence of the dissolved Gd metal in the salt. Given that an estimated 1.5 g of Gd metal electrode was dissolved in the salt, the expected GdCl 3 molar fraction-which includes the contribution from the dissolved Gd metal-measured by ICP-MS should be 0.016, which is close to the actual value (0.0155) measured by ICP-MS measurement. This suggests that ICP-MS is reliable for determining the total Gd concentration in the salt. However, we were more interested in knowing the GdCl 3 concentration, not the total Gd. In the LiCl-KCl-GdCl 3 salt in this study, both GdCl 3 and dissolved Gd metal were present. When a salt sample was analyzed by ICP-MS, the Gd metal (or Gd 0 ) was first converted to Gd 3+ during the sample preparation, and the final GdCl 3 concentration determined by ICP-MS would be greater than the true GdCl 3 concentration. To achieve an accurate analysis of the GdCl 3 concentration in the salt, the contribution of dissolved Gd metal needs to be removed.
Because of the significantly higher GdCl 3 concentration measured by ICP-MS when dissolved Gd metal is present in the salt, ICP-MS can not be used as a reliable technique for monitoring purposes. This is the same for the uranium electrorefining because the uranium metal was found to be dissolved in LiCl-KCl-UCl 3 salt for electrorefining. This is also why the potentiometry herein was proposed for monitoring GdCl 3 or UCl 3 in molten salt and removing the effects from dissolved Gd or U metal. It is worth noting that though some rare earth metals like La, Ce, Pr, and Nd were found to be dissolved in the rare earth halides (chloride, bromide, and iodide), 20 the phenomenon of uranium metal being dissolved in LiCl-KCl-UCl 3 salt was not reported in literature, and we just found this phenomenon in recent years.

Further Discussion on Potentiometry for Monitoring Uranium Electrorefining Operation
In the potentiometry measurement in this study, the OCP between the Gd/GdCl 3 in equilibrium and Ag/AgCl RE is the basis for GdCl 3 concentration monitoring. Any factors that affect either Gd/GdCl 3 equilibrium potential and the Ag/AgCl RE could lead to OCP changes. Generally, based on our experience, the potential of Ag/AgCl RE is quite stable in LiCl-KCl salt system for long term (e. g. 4 to 5 weeks) operation. But some small variations (e.g. about 20 mV) in different Ag/AgCl REs with the same LiCl-KCl-AgCl salt composition in the RE were sometimes observed. Because each potentiometry sensor will be calibrated by measuring the OCP of different LiCl-KCl-GdCl 3 salts with known compositions to obtain the equation (in a general form) of OCP vs GdCl 3 concentration: where A and B are constant at a given temperature and X GdCl 3 is the molar fraction of GdCl 3 , a small variation between different Ag/ AgCl REs is expected not to significantly affect the GdCl 3 concentration measurement. In this study, the addition of NdCl 3 likely affected the Gd/GdCl 3 equilibrium by reacting with the Gd metal "dissolved" in the salt or reacting with the Gd electrode, and change the GdCl 3 concentrations. This may have some implications for potential application of potentiometry measurements. Since one goal of the potentiometry study using GdCl 3 salt as surrogate is for monitoring the UCl 3 concentration in LiCl-KCl-UCl 3 salt for uranium electrorefining. Based on the successful results using GdCl 3 surrogate salt, it is hypothesized that by measuring the OCP between a U electrode and Ag/AgCl RE, the UCl 3 concentration can be measured. Generally, in the LiCl-KCl-UCl 3 salt for uranium electrorefining, there are no other salts that have higher Gibbs free energy of formation than UCl 3 . But if some salts that have higher Gibbs free energy of formation than UCl 3 were added to the molten salt vessel for electrorefining, they will react with the uranium (anode and cathode) in the vessel, and the U electrode for the potentiometry sensor as well, leading to UCl 3 concentration increase, and thus OCP increase. This scenario is a deviation of the normal electrorefining operation, which should be avoided. Luckily, this scenario can be detected by the potentiometry using U electrode vs Ag/AgCl RE, suggesting that the potentiometry sensor can also be used as a monitor for ensuring normal operation of the uranium electrorefiner.
In this study, the GdCl 3 concentration range in LiCl-KCl-GdCl 3 salt is 0.0005 to 0.014 molar fraction. At this range, the activity coefficient of GdCl 3 is constant, but caution is needed when extending the application beyond this range.

Summary
In a LiCl-KCl-GdCl 3 salt system, when there are no other active chlorides present, the potentiometry measurement using Gd metal and a Ag/AgCl RE can be used to measure the GdCl 3 concentration and calculate the activity coefficient of GdCl 3 salt in the tested GdCl 3 concentration range of 0.0005 to 0.014 molar fraction.
The addition of LaCl 3 and CeCl 3 had no apparent effect on the activity of GdCl 3 in the LiCl-KCl-GdCl 3 salt system. However, for reasons not completely understood, NdCl 3 did affect potentiometry measurements conducted on the LiCl-KCl-GdCl 3 salt, though the reaction between Gd electrode and NdCl 3 , which probably occurred after the NdCl 3 addition, is considered the likely cause.
The Gd metal can be dissolved in LiCl-KCl-GdCl 3 salt. To achieve an accurate analysis of the GdCl 3 concentration in the salt by ICP-MS, the contribution of dissolved Gd metal needs to be removed.
The potentiometry measurements in the present study cover only LiCl-KCl-GdCl 3 salts with 0.0005 to 0.014 molar fraction GdCl 3 , so caution is needed when extending beyond that range. While the main purpose of this study was to evaluate the feasibility of using potentiometry for GdCl 3 concentration monitoring in a LiCl-KCl-GdCl 3 salt system, the results herein may also be helpful in understanding molten salt chemistry and thermodynamics.