The oxide reduction (OR) process has been the subject of much research aimed at applying the oxide fuels discharged from pressurized water reactors in the metallic fuel pyroprocessing. This metallic fuel pyroprocessing was successfully demonstrated through the Experimental Breeder Reactor-II project in the USA . Once the oxide fuels are converted into their metallic forms by OR operation, U is selectively recovered via an electro-refining process and transuranic (TRU) nuclides are then recovered along with U via an electro-winning process . The OR process normally employs LiCl as a molten salt and is operated at around 650℃ [3-6]. Oxide fuels are loaded into a metallic cathode basket, with platinum normally employed as the anode owing to its high stability under the harsh condition of high temperatures and a corrosive environment (evolution of O2 gas) [3-6]. However, a recent investigation claimed that repeated OR operations cause noticeable degradation of the platinum anode . Previously, carbon was proposed as an alternative to platinum for the anode [8, 9]. In these reports, the operating voltage was maintained at less than 3.47 V to avoid decomposition of the LiCl salt and the consequent production of chlorine gas at the anode. The reactions employed in this carbon anode system can be summarized as follows:
These carbonate ions are electrochemically decomposed to produce carbon dust at the cathode, according to the following reaction (8), leading to poor process efficiency.
In order to address the carbonate issue, our group recently proposed the use of a carbon anode at a high potential to decompose the LiCl salt itself. Applying higher voltage results in the production of chlorine gas at the anode via the following equation (9):
Here, the reactions mentioned above to produce O2, CO, and CO2 at the anode are still involved in the high-voltage operation due to oxygen ions which are liberated from oxide fuels. Hence, the carbonate formation reaction will proceed even during the high-voltage OR operation, while it was predicted that chlorine gas generated at the anode may react with Li2CO3 or Li2O to produce LiCl [10, 11]. The previous work showed that the concentration of Li2CO3 remained below 0.3wt% after repeated experiments at high cell voltage levels (above 10 V) . This result can be explained by various chemical reactions that can take place between Li2CO3 and Cl2 along with other chemicals, such as Li, Li2O, and C . The direct reaction between Li2CO3 and Cl2 (shown below as reaction (10)) was previously introduced by Nakamura et al. , in which the effects of the reaction temperature and particle size were investigated. The authors could achieve high proportions of LiCl in the temperature range of 350–500℃ with particle sizes smaller than 0.82 mm. However, detail information about the experimental conditions such as gas flow rate, the amount of reactant, and partial pressure of Cl2 was not provided in that report .
In this work, the reaction above was investigated as functions of the reaction temperature, the total flow rate of Ar and Cl2 (Q), and the chlorine partial pressure (pCl2) to identify whether it is a reasonable explanation for the limited accumulation of Li2CO3.
A quartz tube (4 cm diameter) equipped with an electrical furnace in the middle was employed as a reactor. Both ends of the tube were sealed with silicone stoppers connected to Teflon tubes to ensure the flow of gas. The flow rates of the Ar and Cl2 gases were controlled independently using mass flow controllers (MFCs, Kofloc co., Japan). Li2CO3 powder as a reactant was weighed (0.25–1.00 g for each experiment) and then positions in the middle of the quartz tube using an alumina boat. The reactor was purged using Ar gas before the furnace was heated to remove air inside the reactor. After the furnace reached the setting temperature with a ramping rate of 10℃∙min–1, Cl2 gas was fed into it by controlling the MFC. The Cl2 gas flow and the heater were turned off at the end of reaction. The pCl2 value was derived from the inlet flow rate ratio of Cl2 over sum of Ar and Cl2.
The effect of the reaction temperature was investigated by repeating experiments at 300, 400, 500, and 650℃ with 100 mL∙min–1 of Q and 5.07 kPa of pCl2 for various durations. The experiments were conducted at 650℃ for 30 min with pCl2 of 5.07 kPa and various Q (100–300 mL∙min–1) and initial weight of sample (W i, 0.25–1.00 g) to study the effect of Q/Wi ratio. The effect of pCl2 was investigated for various pCl2 conditions of 2.03, 3.04, 5.07, and 10.1 kPa at 650℃ with 100 mL∙min–1 of Q and 1.00 g of W i. The experiments were also conducted for 20 min with 200 mL∙min–1 of Q and 0.25 g of W i.
The progress of the chlorination reaction was calculated by measuing the weight change in the sample before and after the reaction. A structural analysis of the reaction products was conducted by means of the X-ray diffraction (XRD, Bruker D8 Advance).
3. Results and discussion
In order to quantify the progress of the reaction, the conversion ratio of Li2CO3 to LiCl, α, was defined as follows,
where Wf and W i represent the final and initial weight of the samples, respectively, and M.W.LiCl and M.W.Li2CO3 correspondingly indicate the molecular weight of LiCl (= 42.39 g·mol–1) and of Li2CO3 (= 73.89 g·mol–1). Before getting started with chlorination reaction, the stability of Li2CO3 under the gas flow and reaction temperature was verified. This was done to confirm that 1) Li2CO3 is not blown away by the gas flow and 2) Li2CO3 does not react with alumina boat via the following reaction.
After heating Li2CO3 under a 300 mL∙min–1 Ar flow at 650℃ for 2 h, no changes were found in the weight of Li2CO3 indicating that the above issues can be ruled out in this study. This outcome is supported by previous works, in which thermal decomposition temperature of Li2CO3 was reported to be 1,300℃  and beginning of the decomposition reaction was found at 720℃ .
The effects of the reaction temperature on the reaction kinetics were investigated by repeating the experiments at 300, 400, 500, and 650℃, and these results are shown in Fig. 1. It is obvious in the figure that lowering the reaction temperature profoundly diminishes the reaction rate. One interesting outcome was that the reactions did not complete at low temperatures of 300 and 400℃. This result suggests that the diffusion of chlorine atoms within the Li2CO3 particle is too slow to achieve complete conversion at this temperature range. According to this result, it is beneficial to operate the OR process at 650℃ with regard to the removal of Li2CO3 in the salt. However, the trend was a little bit different in previous work done by Nakamura et al. . The authors could achieve high conversion ratio even at 350℃ and proposed a temperature range 400–506℃ for high conversion ratio. It is interesting that the authors reported an abrupt decrease in the conversion ratio at 550℃. This discrepancy might have come from the different reaction conditions such as reactor design, gas flow rate, and chlorine partial pressure.
An XRD measurement was taken to confirm the reaction product, and the result is shown in Fig. 2. The peaks could be assigned using the three phases of Li2CO3 (JCPDF no. 01-087-0729), LiCl (JCPDF no. 01-074-1972), and LiCl-H2O (JCPDF no. 01-070-9971). Here, it would be reasonable to assume that the LiCl-H2O phase was formed during the handling and measurement of the reaction product in air. This result confirms that the reaction between Li2CO3 and Cl2 proceeded without any unexpected by-products and that the assumptions employed during the calculation of α are reasonable. In addition, the results also propose that the direct chlorination reaction using Li2CO3 and Cl2 is a viable means of synthesizing high-purity LiCl.
The effects of Q/W i ratio were investigated by repeating identical experiments for various W i (0.25, 0.50, and 1.00 g) and Q (100, 200, and 300 mL∙min–1) values, and the results are shown in Fig. 3. The experimental runs were conducted for 30 min while pCl2 and the reaction temperature were held constant at 5.07 kPa and 650℃, respectively. It is clear in the figure that an increase in the Q/W i ratio leads to an increase in α, which proves that the chlorination reaction is under control of Cl2 mass transfer in the gas phase. A change in the Q/W i ratio should not affect the reaction rate when the reaction is under control of chemical reaction. In order words, under the experimental conditions of this work, the consumption rate of Cl2 via the reaction with Li2CO3 is faster than the transfer rate of Cl2 at the surface of Li2CO3.
Fig. 4(a) shows the experimental results for various pCl2 conditions with a constant Q of 100 mL∙min–1 and W i of 1.00 g (Q/W i ratio 100 mL∙min–1·g–1). It is clear in the figure that a higher pCl2 leads to a higher reaction rate as expected from the above results. A linear relationship between pCl2 and α was found as shown in Fig. 4(b). As discussed above, these results are not available for analysis of reaction kinetics due to dominant influence of Cl2 mass transfer rate. However, this data is very promising in terms of high-voltage OR operation, as it proves fast chemical reaction between Li2CO3 and Cl2 at the OR operation temperature. In addition, an average utilization value of 37.9% was derived at α = 0.5 proving that Cl2 is an efficient reactant for this reaction. The utilization ratio of Cl2 was calculated by dividing the amount of Cl2 consumed during the reaction by the supplied amount of Cl2.
The effect of pCl2 was also investigated with a constant Q of 200 mL∙min–1 and W i of 0.25 g (Q/W i ratio 800 mL∙min–1·g–1). Fig. 5(a) shows the experimental results with a corresponding linear fitting result. This result indicates that the reaction rate is still under control of chlorine mass transfer. Here, it should be noted that the slope of the linear fitting increased by 27% when the Q/W i ratio increased 8 times from 100 to 800 mL∙min–1·g–1. This outcome reveals that an increase in the Q/W i ratio does not result in a proportional increase in α, and it means lower Cl2 utilization at higher Q/W i ratio. The reaction time–α relationship at Q of 200 mL∙min–1 and W i of 0.25 g (Q/W i ratio 800 mL∙min–1·g–1) is shown in Fig. 5(b), which was conducted at 5.07 kPa of pCl2 and 650℃. It is interesting to observe a linear relationship between the α value and the reaction time over the entire range, because the reaction rate decreased with increasing reaction time at Q/W i ratio of 100 mL∙min–1·g–1 (in Fig. 4(a)). The effect of Q/W i ratio on α is well displayed in Fig. 5(c), in which the results in 5(b) are shown along with the results of Fig. 4(a) with pCl2 of 5.07 kPa. This result suggests that the slowing down of the reaction rate at Q/W i ratio of 100 mL∙min–1·g–1 came from insufficient supply of Cl2 owing to its slow mass transfer rate, whereas this phenomenon was eliminated by increasing the Q/W i ratio to 800 mL∙min–1·g–1. The utilization ratio of Cl2 at α = 0.5 was 20.0% with Q/W i ratio of 800 mL∙min–1·g–1, which was significantly lower than 37.9% with Q/W i ratio of 100 mL∙min–1·g–1.
The chemical reaction between Li2CO3 and Cl2 was investigated for various reaction temperatures, Q/W i ratios, and pCl2 values. The reaction temperature was identified as a key parameter that determines the maximum α value, and at least 500℃ was required for complete conversion of Li2CO3 into LiCl. The relationship between the Q/W i ratio and α value revealed that the overall reaction rate is under control of Cl2 mass transfer in the gas at 650℃. The key outcomes of this work show that the reaction between Li2CO3 and Cl2 is fast at 650℃, and this reaction is strongly suggested as a mechanism which takes place at the anode during the high-voltage OR operation to limit the accumulation of Li2CO3 in the salt.