Oxygen evolution behavior of La1−xSrxFeO3−δ electrodes in LiCl–KCl melt

Electrochemical reduction processes of oxides in molten salt have been proposed as the carbon-free technologies in order to achieve carbon neutrality. The anodic behavior of La1−xSrxFeO3−δ as an O2 evolution anode in LiCl–KCl at 723 K was investigated. The results suggested that at 723 K, the electrical conductivity of La1−xSrxFeO3−δ tended to increase with the Sr doping. The anodic reactions of the La1−xSrxFeO3−δ electrodes were characterized by electrochemical measurements in LiCl–KCl + Li2O at 723 K. Based on the cyclic voltammograms of the La0.7Sr0.3FeO3−δ electrode, O2 evolution has proceeded between 2.7 and 3.6 V. The potential of the La0.7Sr0.3FeO3−δ electrode during galvanostatic electrolysis has conducted at 39 mA cm−2 for 15 h has remained stable at 2.8 V, indicating that the stable evolution of O2 gas was monitored. The corrosion rate was estimated to have the low value of 8.6 × 10−4 g cm−2 h−1. Electrode surface data obtained after electrolysis indicated that the La0.7Sr0.3FeO3−δ electrode exhibited excellent chemical and physical stability in LiCl–KCl at 723 K. This indicates that the La0.7Sr0.3FeO3−δ electrode is promising candidate material as inert anodes for oxide decomposition. As an application of the La0.7Sr0.3FeO3−δ electrode, the electrolytic reduction of CO2 was also successfully achieved.


Introduction
The electrochemical reduction of metal oxides in molten salt has been proposed as an approach to the production of metals [1][2][3][4][5]. The proposed process of molten salt electrolysis requires the use of an inert anode, as described in the following reactions: or As described above, oxides (MO x ) dissolve into molten salt as cations (M 2x+ ) and oxide ions (O 2− ). Subsequently, by conducting electrolysis, M 2x+ is electrochemically reduced to obtain the relevant element in metallic form at the cathode. This element can be alternatively obtained through the direct reduction of MO x . O 2− is electrochemically oxidized such that O 2 gas evolves at the inert anode. However, certain anodic materials that can stably trigger the evolution of O 2 gas stably in molten salt are required because molten salt is highly corrosive.
Carbon anodes have been used in most experiments on the electrochemical reduction of oxides in molten salt. Notably, these anodes react with the evolved O 2 to mainly form CO 2 . This reaction has several drawbacks. The carbon anode is gradually consumed and will require to replacement. Importantly, the extent of consumption of the carbon anode significantly exceeds the amount of CO 2 produced, as the said gas originating from the anode tends to erode the electrode. This reduces the current efficiency of the process and increases its energy consumption [6][7][8]. In this context, the development of nonconsumable O 2 -evolving anodes for metal extraction or carbon from oxides (including CO 2 ) in molten salt is indispensable.
A few materials can act as nonconsumable O 2 -evolving anodes in molten salts. The characteristics required of such inert anodes include physical stability at the operating temperature, electrical conductivity, resistance to attack by the molten salt, resistance to attack by chlorine and oxygen gases, resistance to attack by chloride and oxide ions, resistance to thermal shock, and their robustness. Three general types of materials have been investigated as anodic materials for the described process: metals, ceramics, and the cermet electrode. Metal electrodes exhibit high (1) In molten salt ∶ MO x → M 2x+ + xO 2− Over all cell reaction ∶ MO x → M + x∕ 2O 2 electrical conductivity and good thermal shock resistance [9][10][11][12]. Although metals react with oxygen at high temperatures, the protective oxide layer formed inhibit additional oxidation reactions [9,10]. Cetain ceramics are known to be stable in molten chlorides. However, the ceramic materials are poor electrical conductors, although a few exhibits semiconducting properties that allow conduction at high temperatures [7,8,13,14]. Studies focusing on ion-conductive ceramics have also been reported [15]. Cermet is a composite material of metal and ceramics, combining the characteristics of both components. The metallic component of cermet has been observed to react with oxygen at high temperatures [16].
In the present study, we focused on valency-controlled metal oxide semiconductors as anodic materials because of their high melting point, excellent resistance to oxidation, excellent thermal stability, and low electrical resistivity at high temperatures. Specifically, the p-type La 1−x Sr x FeO 3−δ semiconductor, which is produced under an O 2 partial pressure of 0.2 atm at a high temperature, was selected for nonconsumable O 2 -evolving anodes in molten salt. The p-type La 1−x Sr x FeO 3−δ semiconductor exhibits low electrical resistivity at high temperatures [17]. If parameter x in La 1−x Sr x FeO 3−δ−δ is 0, LaFeO 3 is obtained, which is an insulator at 298 K. The substitution of the trivalent La 3+ ions by a divalent rare-earth metal, such as Sr 2+ ions, forces the Fe ions present in the structure to switch from the stable trivalent to the tetravalent states. The coexistence of the trivalent and tetravalent Fe ions causes to an increase in electronic conductivity compared with LaFeO 3 . La 1−x Sr x FeO 3−δ also exhibits high ionic conductivity due to the formation of additional oxygen vacancies at high temperatures [18,19].
In this study, we synthesized La 1−x Sr x FeO 3−δ electrodes of various compositions (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) and then compared their physical properties. Their electrode reactions were characterized by electrochemical measurements, such as cyclic voltammetry. Electrolysis experiments were conducted under different conditions, and the behavior of the anode was evaluated in molten LiCl-KCl at 723 K.
As an application of the La 0.7 Sr 0.3 FeO 3−δ electrode, CO 2 electrolysis was performed in a molten LiCl-KCl + 3.0 mol% Li 2 O at 723 K. CO 2 decomposition technology will be an effective for combating global warming.

Synthesis of La 1−x Sr x FeO 3−δ
La 1−x Sr x FeO 3−δ powder was synthesized using the solid phase method. La 2 O 3 (99.9%, Kojundo Chemical Laboratory Co., Ltd.), SrO (98.0%, Kojundo Chemical Laboratory Co., Ltd.), and Fe 2 O 3 (99.9%, Kojundo Chemical Laboratory Co., Ltd.) were selected as the oxide precursors. Aliquots of the mentioned metal oxide precursors were weighed to obtain mixtures where parameter x assumed the following values: 0.1, 0.2, 0.3, 0.4, and 0.5. The oxides powders were mixed in a mortar and ground using a bead mill (Easy Nano RMB; Aimex Co., Ltd.) with ZrO 2 balls (φ2 mm) at 2000 rpm in ethanol. The milled powders were examined using a particle size analyzer (SALD-2300; Shimadzu Co., Ltd.) to keep the average particle sizes less than 1 µm. Afterward, the milled powders were then pressed using a 50 mm × 10 mm metal mold at 70 MPa and a cold isostatic press (CIP; Dr.CIP, Kobe Steel Co., Ltd.) at 245 MPa. The formed pellets were sintered in the range of 1473-1723 K for 8 h under the O 2 partial pressure of 0.2 atm. The surfaces of the obtained La 1−x Sr x FeO 3−δ pellets were mechanically polished, first with SiC paper, then with diamond particles (below a size of 6 µm) prior to characterization. The image of a pellet is shown in Fig. 1. A 50 mm × 10 mm × 10 mm metal mold was used. However, the size of the sintered body was subject to shrinking because of the CIP and sintering process. The size of the sintered pellets was measured after the sintering process to be in the range of 46-48 mm × 8-9 mm × 8-9 mm.
The density of the La 1−x Sr x FeO 3−δ pellets was determined using the Archimedes technique. The polished surfaces of the pellets were investigated by scanning electron microscopy (SEM; JSM-7001FD, JEOL Ltd.) and energydispersive X-ray spectroscopy (EDS; JED-2300, JEOL Ltd.). The crystal structures of the pellets were characterized by X-ray diffraction (XRD; RINT-2001 MultiFlex, Rigaku Co, Ltd.). Sample identification was achieved using the relevant diffraction pattern, and the average lattice constant was calculated. The thermoelectric properties of the La 1−x Sr x FeO 3−δ pellets were measured from room temperature to 1073 K using a four-terminal conductivity analyzer (RZ2001i, Ozawa Science Co, Ltd.). Figure 1 shows a schematic view of the electrolytic apparatus used. A quartz holder comprising high-purity silica was used. The working electrode comprised a La 1−x Sr x FeO 3−δ electrode mirror-polished using diamond particles (below a size of 6 µm). Pt wire was used to connect the La 1−x Sr x FeO 3−δ electrode. The counter electrode was Ni wire in a spiral shape (99.35%, φ1.0 mm, Sumiden Fine Conductors Co., Ltd.). The reference electrode was an Ag wire immersed in a eutectic LiCl-KCl melt containing 1 mol% AgCl in a thin-bottomed alumina tube. The alumina tube was sufficiently thin enough to use the membrane of the reference because the alumina intrinsically included a small quantity of various metal ions, which afforded ion conductivity at the experimental temperatures. The potential of the reference electrode was calibrated in reference to that of the Li I | Li electrode, which was prepared by electrodepositing Li metal on a nickel electrode. Lithium chloride (LiCl, 99.0%, Fujifilm Wako Pure Chemical Co.) and potassium chloride (KCl, 99.5%, Fujifilm Wako Pure Chemical Co.) were mixed in a eutectic composition (58.5 mol% LiCl, 41.5 mol% KCl). The obtained mixture was dried under vacuum conditions for more than 24 h at 473 K to remove water from it. The eutectic mixture was prepared to a total weight of 200 g (LiCl, 89.0 g; KCl, 111.0 g). Thereafter, it was melted in a high-purity alumina crucible. Lithium oxide (Li 2 O, 99.5%, Fujifilm Wako Pure Chemical Co.) was used as the source of the oxide ion and was directly added to the melt. The experimental processes were conducted in a glove box under dried argon atmosphere. Thus, the effect of moisture was practically negligible. CO 2 electrolysis was performed in LiCl-KCl + 3.0 mol% Li 2 O at 723 K. The cathode deposits after CO 2 electrolysis were analyzed by Raman spectroscopy analysis (NRS-5500, Jasco Co.) using a YAG laser (532 nm).

Synthesis of La 1−x Sr x FeO 3−δ
La 1−x Sr x FeO 3−δ powder was synthesized using the solid phase method. La 2 O 3, SrO, and Fe 2 O 3 powders were selected as oxide precursors. LaFeO 3 comprises stable trivalent La and trivalent Fe, and it exhibits a perovskite crystal structure. La 1−x Sr x FeO 3−δ was synthesized by replacing the trivalent La ion with the divalent Sr ions, forcing the Fe ion to switch from the stable trivalent to the tetravalent states [18,19].
The oxide precursors of La 1−x Sr x FeO 3−δ were weighed such that values of x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5 were obtained for parameter x, and the obtained mixtures were ground using a bead mill and pressed using a CIP at 245 MPa. Subsequently, the formed pellets were sintered within the range of 1473-1623 K for 8 h under the O 2 partial pressure of 0.2 atm. Figure 2 shows the XRD spectra of the La 1−x Sr x FeO 3−δ samples obtained after the mentioned sintering process. The XRD patterns of the La 1−x Sr x FeO 3−δ samples (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) were indicative of single-phase (La, Sr) FeO 3−δ ceramics. The average lattice constants of the samples were calculated from the relevant XRD spectra calibrated using a reference sample (pure Si). The axis average lattice constants for the a-, b-, and c-axes of the La 1−x Sr x FeO 3−δ samples decreased with an increasing in parameter x. The values of the coefficient (R 2 ) of each axis were calculated using a regression line to be 0.8632, 0.9595, and 0.8830 for the a-, b-, and c-axes, respectively (Fig. S1, supplementary information). The coefficient (R 2 ) exceeded 0.85 for the three axes. The average lattice constants of the synthesized La 1−x Sr x FeO 3−δ semiconductors were confirmed to be correlated with the parameter x of La 1−x Sr x FeO 3−δ .
The coexistence of Fe 3+ and Fe 4+ ion was achieved by substituting La 3+ for Sr 2+ [20]. As the ionic radius of Fe 4+ is smaller than that of Fe 3+ , the average lattice constant decreased with an increase in parameter x (Fig. S1).  Therefore, the formation of a substituted La 1−x Sr x FeO 3−δ solid solution was suggested.

Densification of La 1−x Sr x FeO 3−δ
The ceramic electrodes used for molten salt electrolysis need to be characterized by high density. In this study, the La 1−x Sr x FeO 3−δ samples with different x values (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) were sintered at various temperatures in the range of 1473-1723 K, to determine the most suitable sintering temperature for the densification of the samples. Notably, the densities (g cm −3 ) were measured using the Archimedes technique. Table 1 shows the relationship between the sintering temperature and density of La 1−x Sr x FeO 3−δ . The densities of the Sr-doped samples with x = 0.1, 0.2, and 0.3 drastically increased at 1623 K. Contrarily, the densities of the Sr-doped samples with x = 0.4 and 0.5 reached a maximum at 1523 and 1498 K. As confirmed from the binary phase diagram of LaFeO 3 -SrFeO 3−δ [21], the melting point tended to decrease as the Sr concentration (x) increased. Thus, the densification temperature decreased as the x value increased. Figure 3 shows data reflecting the electrical properties of the La 1−x Sr x FeO 3−δ samples characterized by the highest density for each composition ratio at various temperatures. The conductivities of each sample were measured using the four-terminal conductivity analyzer from room temperature to 1073 K. When the Sr composition ratio was changed from x = 0 to x = 0.5, at 723 K, the electrical conductivity of the material increased from 5.7 × 10 −1 to 1.6 × 10 4 S/m. The substitution of trivalent La 3+ ion by a divalent metal, such as Sr 2+ ion, forced the Fe ion to switch from the stable trivalent to the tetravalent states. The coexistence of the trivalent and tetravalent Fe ions caused to an increase in electronic conductivity compared with the LaFeO 3 of the parent material. Furthermore, La 1−x Sr x FeO 3−δ (x > 0.1) also exhibits higher ionic conductivity than LaFeO 3 owing to the formation of additional oxygen vacancies at elevated temperatures [19,22].

Electrochemical behavior of oxide ions on La 1−x Sr x FeO 3−δ electrodes
Cyclic voltammetry performed in a LiCl-KCl melt at 723 K to investigate the electrochemical behavior of oxide ions on the La 0.7 Sr 0.3 FeO 3−δ electrode. Figure  the oxidation current density began to increase at a potential of 2.7 V. In previous studies [24,25] conducted on the same electrolyte comprising molten LiCl-KCl and Li 2 O, which included a boron-doped diamond electrode acting as the anode, O 2 and Cl 2 evolution was observed to proceed at 2.5 and 3.6 V, respectively. Therefore, in the present system with a La 0.7 Sr 0.3 FeO 3−δ electrode as the anode, the increase in oxidation current density at 3.7 V observed before the addition of Li 2 O to the LiCl-KCl melt was attributed to the reaction leading to chlorine gas evolution or chlorine compounds. Contrarily, the increase in oxidation current density at 2.7 V observed after the addition of Li 2 O was attributed to the reaction leading to O 2 gas evolution. In the LiCl-KCl melt containing Li 2 O, O 2 gas evolution was assumed to occur between 2.7 and 3.6 V at 723 K.
The electrochemical behavior of oxide ions on the La 1−x Sr x FeO 3−δ electrode (x = 0.3, 0.4, and 0.5) was also investigated by comparing the cyclic voltammograms obtained at each x value, to determine the composition of the La 1−x Sr x FeO 3−δ electrode that affords the highest reactivity leading to O 2 evolution. Figure 5 shows the cyclic voltammograms of La 1−x Sr x FeO 3−δ electrodes (x = 0.3, 0.4, and 0.5) in the LiCl-KCl melt containing 3.0 mol% Li 2 O at 723 K. The electrode surface area was defined as the geometric surface area. The different background currents shown in Fig. 5 were attributed to the differences in the real surface area. As shown in Fig. 3, these three electrodes (x = 0.3, 0.4, and 0.5) exhibited the highest electrical conductivity. The La 0.5 Sr 0.5 FeO 3−δ electrode exhibited the highest oxidation current density between 2.8 and 3.1 V, but the La 0.7 Sr 0.3 FeO 3−δ electrode exhibited the highest oxidation current density between 3.2 and 3.6 V (Fig. 5). The current density (Fig. 5) of La 0.5 Sr 0.5 FeO 3−δ electrode practically remained constant from 2.9 to 3.4 V. In a study by Kado et al. [25], the oxidation current between 2.5 and 3.6 V in LiCl-KCl at 723 K was due to O 2 generation. Therefore, the shape of the cyclic voltammograms of each electrode suggested that the O 2 evolution reaction on La 0.7 Sr 0.3 FeO 3−δ and La 0.6 Sr 0.4 FeO 3−δ were governed by the charge-transfer step, whereas that on La 0.5 Sr 0.5 FeO 3−δ was governed by the diffusion step. The differences in the shape of the cyclic voltammograms were due to differences in the O 2 evolution processes on the electrode/the molten salt interface. The different amounts of Sr added to La 1−x Sr x FeO 3−δ affected the catalytic activity of the O 2 evolution reaction.
We decided to use the La 0.7 Sr 0.3 FeO 3−δ electrode in electrolysis tests. This electrode exhibited the lowest electrical conductivity during the cyclic voltammetry experiments. This confirmed that an increase in electrical conductivity did not necessarily enhance the O 2 evolution reaction.

Electrolysis tests performed using the La 0.7 Sr 0.3 FeO 3−δ electrode
Based on the cyclic voltammetry results, the O 2 evolution process was concluded to have occurred between 2.7 and 3.6 V in the LiCl-KCl melt containing 3.0 mol% Li 2 O at 723 K. Figure 6 shows the time transition of current densities measured during potentiostatic electrolysis experiments conducted at 2.9 V, 3.3 V, and 3.5 V. The anodic currents  Fig. 6 was caused by altering the electrode surface area because of the adsorption and desorption of O 2 gas bubbles evolved at the electrode/molten salt interface during the experiment. The electrolysis conditions were selected based on Fig. 4. The O 2 evolution process was observed at the potentials ranging from 2.7 to 3.6 V in the LiCl-KCl melt containing 3.0 mol% Li 2 O at 723 K. The current density corresponding to the potential range of the gas evolution was selected from 10 to 100 mA cm −2 depending on the experiment. Figure 7 shows data on the time transition of the potential during galvanostatic electrolysis experiments conducted at current densities of 49, 46, 39 mA cm −2 for 2, 5, 15 h, respectively. The potential at which the O 2 gas evolution process occurred fluctuated between 3.1 and 3.3 V during the experiments conducted at 49 and 46 mA cm −2 for 2 and 5 h, respectively. For the galvanostatic electrolysis at 39 mA cm −2 , the potential remained stable at 2.8 V during the electrolysis (Fig. 7). Oppositely, as the currents density exceeded 39 mA cm −2 , Fig. 8 Gas chromatogram of gas samples collected before and after galvanostatic electrolysis conducted at a current density of 39 mA cm −2 for 15 h Fig. 9 Scanning electron microscopy images and energy-dispersive X-ray spectroscopy results of the surface of the La 0.7 Sr 0.3 FeO 3−δ electrode before and after electrolysis at 39 mA cm −2 for 15 h the potential began to fluctuate. Therefore, an applied current density of 39 mA cm −2 was chosen for the experiment.
The mechanism of the fluctuating potential shown in Fig. 7 was unclear. A possible explanation was that the effect of potential dependence on the behavior of the nucleation, growth, and detachment of O 2 gas on the electrode might have caused the difference in the current time transition.
In addition, the gas generated on the La 0.7 Sr 0.3 FeO 3−δ electrode during the electrolysis conducted at a current density of 39 mA cm −2 for 15 h was analyzed by gas chromatography (GC-2014, Shimadzu Co, Ltd.). Figure 8 shows the gas chromatograms of gas samples collected before and after galvanostatic electrolysis. It was confirmed in advance that the peak attributed to O 2 gas had a retention time of approximately 3 min based on the results of injecting O 2 standard gas. Figure 8 shows that the intensity of the peak due to O 2 in the gas sample collected after electrolysis was increased. These intermediates formed by the anodic reactions suppressed the current efficiency and the subsequent reactions should occur.
In addition, certain intermediates decomposed to form O and O 2− [26], and the atomic oxygen easily reacted with the metal wire in the cell.
The jump in potential led by the small increase in current density (Fig. 7) might also have been caused by changes in side reactions associated with O 2 ions, such as those in Eqs. (7)- (12). The relatively low electric conductivity of the La 0.7 Sr 0.3 FeO 3−δ (Fig. 3) may be another reason for the suppressed the current efficiency.
The sample subjected to electrolysis at 39 mA cm −2 current density for 15 h had the longest time and highest total electrolysis charge in this study. The morphological surface changes exhibited by the La 0.7 Sr 0.3 FeO 3−δ electrode because of the electrolysis experiment were investigated by SEM, EDS, and XRD. Figure 9 shows the SEM images of the electrode surfaces and the results of the EDS elemental analyses conducted on the electrodes before and after electrolysis. The SEM images revealed tiny holes (several tens of nanometers) in the surface morphology after electrolysis. It was suggested that the tiny holes were produced by eroding the anode during O 2 evolution. The EDS results confirmed that there were practically no change in the elemental composition. The detected carbon   Figure 10 shows the XRD patter ns of the La 0.7 Sr 0.3 FeO 3−δ electrode recorded before and after galvanostatic electrolysis experiments. A single-phase diffraction pattern attributed to (La,Sr)FeO 3−δ was observed in all the samples of before and after galvanostatic electrolysis experiments. The lack of change in the diffraction pattern before and after galvanostatic electrolysis indicated that the crystal structure of the (La,Sr)FeO 3−δ electrode had not changed. The La 0.7 Sr 0.3 FeO 3−δ electrode showed high corrosion resistance. As a mechanism for the high corrosion resistance, it is suggested that the La 0.7 Sr 0.3 FeO 3−δ electrode itself functions as an oxide protective layer. The La 0.7 Sr 0.3 FeO 3−δ electrode exhibited sufficient chemical and physical stability for use in molten salt electrolysis experiments.
The corrosion rate γ is another important property that defines the suitability of nonconsumable O 2 -evolving anodes. In this study, the value of parameter γ (g cm −2 h −1 ) was estimated using Eq. (13) [8]: where ΔW is the difference between the weight (g) of the anode before and after electrolysis, t is the electrolysis time (h), and S is the surface area of the part of the electrode that is immersed in the electrolyte. The corrosion evaluation of the anode during galvanostatic electrolysis was conducted in a LiCl-KCl melt containing 3.0 mol% Li 2 O at 723 K. The conditions are summarized in Table 2.
As shown in Table 2, the corrosion rate of the La 0.7 Sr 0.3 FeO 3−δ electrode was calculated to have a value between 0 and 8.6 × 10 −4 g cm −2 h −1 . Considering that the weight loss of La 0.7 Sr 0.3 FeO 3−δ was not confirmed after immersion for 10 h without electrolysis in LiCl-KCl at 723 K (Table S1), the calculated values in Table 2 were the corrosion rate due to electrolytic corrosion. The weight loss of this anode was caused by eroding the anode during O 2 evolution because tiny holes (several tens of nanometers) in the surface morphology was observed in the SEM images for the sample after electrolysis at 39 mA cm −2 for 15 h (Fig. 9). Fray et al. [8] deemed an anode inert as long as the value of the corrosion rate was below 1.5 × 10 −3 g cm −2 h −1 . Therefore, La 0.7 Sr 0.3 FeO 3−δ is an extremely promising candidate in the manufacture of (13) = ΔW∕ t ⋅ S

CO 2 decomposition using the La 0.7 Sr 0.3 FeO 3−δ electrode
Cyclic voltammetry experiments were performed in a LiCl-KCl melt + 3.0 mol% Li 2 O after CO 2 bubbling for 5 h at 723 K. Figure 11 shows the cyclic voltammograms recorded before and after the CO 2 bubbling for 5 h. The oxidation current density after CO 2 bubbling was smaller than before CO 2 bubbling. The reaction shown in Eq. (14) proceeded by CO 2 bubbling in an electrolytic bath.
The decrease in oxide ions in the electrolytic bath led to a decrease in current density due to the O 2 evolution reaction (Eq. 4).
Potentiostatic electrolysis was performed at 3.23 V, the peak of oxidation current density. Figure 12 shows the time transition of current densities measured during potentiostatic electrolysis experiments at 3.23 V. The anodic currents measured in association with the O 2 gas evolution process practically remained steady at 1 mA cm −2 .
Black deposition observed on the Ni cathode after the 15 h electrolysis was analyzed by Raman spectroscopy. Figure 13 shows the Raman spectra of the electrodeposition. Two peaks attributed to carbon were confirmed at 1350 cm −1 and 1584 cm −1 . The following cathodic reaction proceeded. CO 2 electrolysis was successfully achieved. However, the current density of the electrolysis was low in the present experimental system; therefore, the electrolysis bath requires improvement.

Conclusion
Dense La 1−x Sr x FeO 3−δ ceramic materials of different compositions (x = 0.1, 0.2, 0.3, 0.4, and 0.5) were obtained through a sintering process between 1498 and 1673 K for 8 h under an O 2 the partial pressure of 0.2 atm. The electrical conductivity of the La 1−x Sr x FeO 3−δ ceramics increased from 5.7 × 10 −1 to 1.6 × 10 4 S/m as x increased from 0 to 0.5.
The suitability of the La 1−x Sr x FeO 3−δ electrodes for use as O 2 -evolving anodes was investigated through a series of electrochemical measurements in LiCl-KCl melt containing Li 2 O at 723 K. The cyclic voltammetry experiments were conducted on the La 1−x Sr x FeO 3−δ electrodes with x = 0.3, 0.4, and 0.5. The La 0.5 Sr 0.5 FeO 3−δ electrode exhibited the highest oxidation current density between 2.8 and 3.1 V, whereas the La 0.7 Sr 0.3 FeO 3−δ electrode exhibited the highest oxidation current density between 3.2 and 3.5 V. The current density of the La 0.5 Sr 0.5 FeO 3−δ electrode shown in Fig. 5 practically remained constant from 2.9 to 3.4 V, indicating that the O 2 evolution was governed by a diffusion control. As the applied potential range was equal to from 0.4 to 1.0 V expressed by overpotential, the applied potentials were far from the equilibrium potential of O 2 generation [25]. The La 0.7 Sr 0.3 FeO 3−δ electrode was subjected to electrolysis tests and exhibited the lowest electrical conductivity among the electrodes used in the cyclic voltammetry experiments. This confirmed that increases in electrical conductivity were not necessarily associated with enhancements of the reaction leading to O 2 gas evolution.
During galvanostatic electrolysis, the potential of the La 0.7 Sr 0.3 FeO 3−δ electrode remained stable between 2.8 and 3.3 V, regardless of the reaction conditions. Based on the gas chromatography results, the stable evolution of O 2 gas was confirmed to have occurred, and the current efficiency of the electrolysis process conducted at 39 mA cm −2 current density for 15 h was estimated at 30.4%. The corrosion rate of La 0.7 Sr 0.3 FeO 3−δ was estimated at a very low value of 8.6 × 10 −4 g cm −2 h −1 . The XRD, SEM, and EDS results of the La 0.7 Sr 0.3 FeO 3−δ electrode before and after the electrolytic process showed that the electrode exhibited excellent chemical and physical stability in the LiCl-KCl melts containing Li 2 O at 723 K. This study shows that La 0.7 Sr 0.3 FeO 3−δ is a promising candidate material for the manufacture of nonconsumable O 2 -evolving anodes.
As an application of the La 0.7 Sr 0.3 FeO 3−δ electrode, CO 2 electrolysis was successfully achieved. However, the current density of the CO 2 electrolysis was low in the present experimental system; therefore, the electrolysis bath and electrode require improvement.
Author contributions Shunichi Kimura wrote the main manuscript text. Takashi Fukumoto and Yuta Suzuki prepared figures 11-13. Takuya Goto planned the concept of this study. All authors reviewed the manuscript.

Declarations
Competing interests On behalf of all authors, the corresponding author states that there is no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.