The oxide reduction (OR) process enables the use of used oxide fuel discharged from light water reactors in the electrochemical pyroprocessing which was initially designed for metal fuels in the Experimental Breeder Reactor- II (EBR-II) project . There are two explanations of possible OR process mechanisms. One is a direct electrochemical reduction mechanism, which can be expressed with the following reaction equation, which operates at 2.40 V.
The second mechanism is a chemical reduction mechanism, which suggests that Li2O dissolved in LiCl salt is decomposed to produce metallic Li and oxygen gas at 2.47 V. The metallic Li produced at the cathode reacts with UO2 according to the following reaction equations.
Either way, a noble metal electrode (normally Pt) is employed as the anode material to produce oxygen gas at the operating condition of 650℃ [2-5]. The use of Pt limits the maximum potential applicable to the cell due to the associated dissolution reaction, and degradation of a Pt electrode via repeated OR operations was recently documented . Previously, carbon was proposed as an alternative to expensive Pt; however, the formation of carbonate ions (CO32-) and their decomposition product, carbon dust, was indicated as a key issue that prevented the use of carbon in the OR process .
Recently, the carbon anode was revisited by our group for use in the OR process at a high anode potential to solve abovementioned Pt and carbonate issues [8, 9]. High-potential operation employs the following electrochemical reaction in order to produce metallic Li, which can reduce UO2 into U, as noted above.
This reaction occurs at 3.81 V according to the Gibbs free energy of the reaction. It was experimentally shown that this approach could suppress the accumulation of carbonate ions, presumably owing to the reaction with chlorine gas near the anode [8, 9]. These operations were demonstrated at around 10~15 V, which is several times higher than the level of 3~3.4 V in the Pt case. Although the high-potential carbon anode has brought profound advantages for OR operations, a new problem also arose: chlorine . It is well known that chlorine is a corrosive gas, especially at high temperatures. Severe corrosion of equipment components after several operations of the high-potential carbon anode OR was observed in our previous work .
The gas stream predicted during OR operation with the high-potential carbon anode is a mixture of Cl2, O2, CO, CO2, and Ar. Here, argon is an atmospheric gas in the OR operation environment. Among these gases, we focused on the impact of chlorine and oxygen against corrosion of the equipment, as corrosion by the other gases is likely low. The corrosion behavior of certain commercial alloys under a Cl2 or HCl atmosphere was previously investigated, and it was shown that pure nickel and its alloys, specifically Inconel and Hastelloy, had much lower corrosion rates than those of carbon steels . The corrosion behaviors of iron , nickel , and chromium  in an HCl-O2 mixed atmosphere were investigated by Ihara et al. However, the corrosion behavior of commercial metals under a Cl2- O2 mixed atmosphere is not available. As a starting point, our group investigated the corrosion behavior of stainless steel 316 (SS-316) under a Cl2-O2 mixed atmosphere . Interestingly, it was found that the presence of oxygen significantly reduced the chlorine-induced corrosion rate of SS-316. However, the exponential increase in the corrosion rate with an increase in the reaction temperature suggested limited application of SS-316 at or below 300℃ under a chlorine atmosphere. In the present study, Inconel X-750 was employed to verify its applicability to high-potential carbon anode OR equipment. The effects of the reaction temperature, Cl2-O2 composition, and Cl2 flow rate were investigated.
A quartz tube (40 mm diameter) was employed as the corrosion reactor in this study. The tube was positioned horizontally and the middle part was heated using an electrical furnace (equipped with Kanthal A-1 heating element along 120 mm length of heating zone). One end of the quartz tube was connected to three mass flow controllers (MFCs), while the other end was connected to a dry Cl2 scrubbing system which was designed to adsorb residual Cl2 gas in the exhaust gas stream. Schematic diagram of the experimental set-up can be found in our previous work .
An Inconel X-750 plate (purchased from Nilaco co., Japan, 1.0 mm thickness with major constituents of Ni (~70%), Cr (14~17%), Fe (5~9%), and Ti (2.25~2.75%)) was cut into 30 mm × 10 mm pieces and four of them were utilized in every experiment. The specimens were thoroughly washed with deionized (DI) water and ethanol (Daejung chemicals and metals co., Korea, anhydrous ethanol, 99.9%) before the corrosion experiments. The average weight of the samples (four pieces) was 10.172 g. The flow rates of the feed gases were individually controlled using MFCs (purchased from Kofloc co., Japan) for each gas (Model 5400 for Cl2 and Model 3660 for Ar and O2). It should be noted that the specimens were not polished in order to mimic the actual manufacturing condition of OR equipment, including a native oxide layer.
The effect of the corrosion reaction temperature was investigated in a temperature range of 300~700℃ by flowing 170 mL·min-1 Ar + 30 mL·min-1 Cl2 for 8 h. The effect of the mixed gas feed was identified by flowing various gas compositions of 30 mL·min-1 Cl2, 20 mL·min-1 Cl2 + 10 mL·min-1 O2, and 10 mL·min-1 Cl2 + 20 mL·min-1 O2 at 600℃ while the flow rate of Ar was kept constant at 170 mL·min-1. During the mixed gas effect experiments, tests were conducted for various reaction durations of 2, 4, 8, and 16 h. Additional experiments were performed for various chlorine flow rates of 5, 10, and 20 mL·min-1 at 600℃ with an argon balance flow for a total flow rate of 200 mL·min-1 without oxygen. After each experiment, the samples were washed using DI water and ethanol to remove any adhering chlorides or oxides.
The effect of the gas compositions on the surface morphology was investigated using a scanning electron microscopy (SEM, Hitachi SU-8020) system with energy dispersive spectroscopy (EDS) for a composition analysis.
3. Results and discussion
The effect of the reaction temperature under a Cl2–Ar flow condition was investigated, and these results are shown in Fig. 1. It should be noted that the corrosion rate was negligible at 300 and 400℃, meaning that the weight change was less than 1 mg (= 0.00983%). Beginning at 500℃, the corrosion rate increased exponentially with an increase in the reaction temperature. The corrosion rate measured at 500℃ was 5.01×10-1 g·m-2·h-1, which is far below the rate of 1.98×101 g·m-2·h-1 for SS-316 at 500℃ . Although Inconel X-750 proved its superior resistance against chlorine gas, the corrosion rate increases too rapidly to be employed at or above 600℃. These results are in agreement with those in previous work that suggested an operation temperature of 538℃ for Inconel and Hastelloy alloys under a Cl2 atmosphere . The exponential relationship between the reaction temperature (in K) and the corrosion rate is displayed in Fig. 1. The following equation was achieved from the fitting result:
It is important to note that this corrosion rate equation is applicable only to the experimental condition in this study.
After the corrosion experiment at 700℃, a significant amount of yellowish corrosion product was observed in both the residual specimens and the end part of the quartz tube, as shown in Fig. 2. The residual specimens were covered by dense dark brown layers with thin yellow flakes. The corrosion products were easily removed by washing to expose the metallic surface, as shown in the figure (After washing). An image of the evaporated corrosion product is also shown in Fig. 2 (Evaporated product); it is composed of yellowish and brown particles which appear similar to those covering the residual specimen. Although the major constituents of Inconel X-750 are Ni (~ 70%), Cr (14~17%), Fe (5~9%), and Ti (2.25~2.75%), only Ni and Cl were detected in the EDS analysis result of the evaporated corrosion product.
The effects of the gas composition and reaction time on the weight change of specimens are shown in Fig. 3. In the 30 mL·min-1 Cl2 case (Fig. 3(a)), a weight increase was observed at reaction times of 2 and 4 h owing to the formation of chloride scale and the relatively slow evaporation of the scale. With an increase in the reaction time, the gap between the findings before and after washing becomes larger, suggesting that the formation rate of chloride scale is faster than its evaporation rate under the 30 mL·min-1 Cl2 + 170 mL·min-1 Ar condition. In the presence of oxygen, the weight decrease was significantly reduced in both the 20 mL·min-1 Cl2 + 10 mL·min-1 O2 and 10 mL·min-1 Cl2 + 20 mL·min-1 O2 cases (Fig. 3(b) and 3(c), respectively). It is interesting to observe that the gap between the outcomes before and after washing decreases with an increase in the ratio of oxygen. This result suggests that the formation rate of chlorides is suppressed by oxygen, while the impact is relatively low against the evaporation rate. A comparison graph is drawn in Fig. 3(d), and it is clear that the presence of oxygen significantly suppressed the corrosion reaction. However, it needs to note that the effect of oxygen flow rate was subtle. The changes in the corrosion rate with the reaction time and gas composition are summarized in Fig. 3(e). Again, it is clear that the presence of oxygen profoundly reduces the corrosion rate, while the impact of oxygen flow rate is relatively low. A comparison experiment carried out by flowing 20 mL·min-1 Cl2 + 180 mL·min-1 Ar clearly proved the impact of oxygen, as shown in the figure. It is also important to note that the corrosion rate decreased with an increase in the reaction time, especially in the 30 mL·min-1 Cl2 + 170 mL·min-1 Ar case. The formation and relatively slow evaporation rate of chloride scale as discussed above may explain why the corrosion rate diminished with an increase in the reaction time. On the other hand, the effect of the reaction time on the corrosion rate was lessened in the chlorine-oxygen mixed experiments.
The SEM images shown in Fig. 4 were taken to determine the effect of the Cl2-O2 composition on the surface morphologies. In the 30 mL·min-1 Cl2 + 170 mL·min-1 Ar case, a dense surface with clearly distinguishable grain boundaries is shown. Significant changes were not observed in the 30 mL·min-1 O2 + 170 mL·min-1 Ar case except the appearance of shiny spots originated from the formation of oxide grains on the surface. A totally different story was found in the Cl2-O2 mixed gas cases, where the formation of porous surfaces with tunnel-type holes with diameters of approximately 1 μm was noted. This result is quite different from that with SS-316, where the formation of round pits with diameters of several hundred micrometers was observed . These results suggest that the corrosion reaction with chlorine occurs via different mechanisms in Inconel X-750 and SS-316 in the presence of oxygen. However, one thing clear in Fig. 4 is that the presence of oxygen leads the chlorination reaction along the vertical direction to produce porous surfaces. However, additional investigations are necessary to determine in a crystal clear manner how oxygen works to suppress the corrosion reactions in both alloys and to determine the corrosion mechanism that leads to different surface morphologies.
The protective role of oxygen was also observed in our previous work with SS-316 , and a comparison chart is shown in Fig. 5. It is clear in the figure that SS-316 exhibits significantly higher corrosion rates regardless of the Cl2-O2 composition used. The difference in the corrosion rate increased with the chlorine flow rate from 22 (10 mL·min-1 Cl2 + 20 mL·min-1 O2) and 34 (20 mL·min-1 Cl2 + 10 mL·min-1 O2) to 41 times (30 mL·min-1 Cl2). This result verifies that Inconel X-750 is a far better option for a highpotential carbon anode OR application than SS-316, though the relatively high price of Inconel X-750 should also be considered during the design of actual equipment.
Recalling that the corrosion rate in 30 mL·min-1 Cl2 + 170 mL·min-1 Ar was very close to the outcome in the 20 mL·min-1 Cl2 + 180 mL·min-1 Ar case, as shown in Fig. 3(e), the effect of the chlorine flow rate without oxygen became questionable. Experiments were conducted with various chlorine flow rates of 5, 10, 20, and 30 mL·min-1 with an argon balance flow to keep the total flow rate at 200 mL·min-1 without oxygen. These results are shown in Fig. 6, and it is obvious that the Cl2 flow rate has no impact on the corrosion rate given the conditions used here. This result implies that the amount of chlorine generated during OR operation with the high-potential carbon anode can be neglected during equipment design efforts, whereas the impact of low concentrations of chlorine gas should not be underestimated.
Under a chlorine atmosphere, Inconel X-750 was resistant to corrosion up to 400℃, while an exponential increase in the corrosion rate was observed at temperatures greater than 500℃. The presence of oxygen significantly reduced the corrosion rate and produced porous surfaces. The corrosion rates of Inconel X-750 were only 1/22~1/41 of those of SS-316 at 600℃. The effect of the chlorine flow rate was found to be negligible in the range of 5~30 mL·min-1 Cl2. Overall, Inconel is a good candidate as a high-potential carbon anode OR equipment material, although its lifetime should be considered when using it at or above 500℃ with chlorine gas.