Visualization of TiO2 Reduction Behavior in Molten Salt Electrolysis

Shungo Natsui; Takuya Sudo; Ryota Shibuya; Hiroshi Nogami; Tatsuya Kikuchi; Ryosuke O. Suzuki

doi:10.1007/s11663-019-01733-7

Visualization of TiO₂ Reduction Behavior in Molten Salt Electrolysis

Shungo Natsui, Takuya Sudo, Ryota Shibuya, Hiroshi Nogami, Tatsuya Kikuchi, Ryosuke O. Suzuki

IMRAM - Division of Process and System Engineering

Hokkaido University

Research output: Contribution to journal › Article › peer-review

6 Citations (Scopus)

Abstract

An in situ observation technique of the TiO₂ interfacial behavior in molten LiCl-KCl electrolysis was developed. The variation of the thin TiO₂ electrode surface were tracked through the high-speed digital microscopy synchronized with the electrochemical measurement. Two characteristic interfacial behaviors were discovered: physical breakage of the titanium oxide and Li(l) spreading on electrode surface. These electrochemically induced interfacial behaviors affect the current-time curves due to the heterogeneity of the titanium oxide film shape.

Original language	English
Pages (from-to)	11-15
Number of pages	5
Journal	Metallurgical and Materials Transactions B: Process Metallurgy and Materials Processing Science
Volume	51
Issue number	1
DOIs	https://doi.org/10.1007/s11663-019-01733-7
Publication status	Published - 2020 Feb 1

Access to Document

10.1007/s11663-019-01733-7

Cite this

@article{0c5ac6fc913243f8957fe179fc9fe433,

title = "Visualization of TiO2 Reduction Behavior in Molten Salt Electrolysis",

abstract = "An in situ observation technique of the TiO2 interfacial behavior in molten LiCl-KCl electrolysis was developed. The variation of the thin TiO2 electrode surface were tracked through the high-speed digital microscopy synchronized with the electrochemical measurement. Two characteristic interfacial behaviors were discovered: physical breakage of the titanium oxide and Li(l) spreading on electrode surface. These electrochemically induced interfacial behaviors affect the current-time curves due to the heterogeneity of the titanium oxide film shape.",

author = "Shungo Natsui and Takuya Sudo and Ryota Shibuya and Hiroshi Nogami and Tatsuya Kikuchi and Suzuki, {Ryosuke O.}",

note = "Funding Information: The online version of this article (10.1007/s11663-019-01733-7) contains supplementary material, which is available to authorized users. In the refining industries of the oxide-stable metallic elements, the electrolysis of the molten chloride is indispensable. Thus, and efficient electrolysis has been developed, for example, for use in the Kroll process for titanium production. In the Kroll process, TiO 2 is first converted to TiCl 4 by Cl 2 gas. Then, liquid Mg reduces the TiCl 4 such that high-purity metallic sponge Ti is obtained. The liquid Mg and gaseous Cl 2 are regenerated by electrolysis of the byproduct MgCl 2 and recycled. To avoid some complicated steps in the Kroll process, the direct electrochemical decomposition of TiO 2 in molten CaCl 2 has been proposed. In the FFC Cambridge process, the oxide anion cathodically transfers from the solid TiO 2 pellet to the anode in a molten salt bath.[ 1 ] Because the Ti-O binary system contains many suboxides, oxygen in the higher oxide is removed to form a lower oxide upon receiving an electrical charge from the cathode. Another promising method, the OS process, has been proposed, in which the oxide anion transfer in CaCl 2 is better utilized because as much as 20 mol pct CaO that acts as an electrolyte can dissolve into the molten CaCl 2 at 1173 K.[ 2 – 5 ] The electrochemically deposited, liquid Ca at the cathode also dissolves into the molten CaCl 2 , and the dissolved Ca works effectively to reduce the titanium oxide powder. Similarly, LiCl and its binary chloride systems can dissolve oxygen anions at lower temperatures,[ 6 – 8 ] and KCl is sometimes added to lower the temperature further.[ 9 ] Electrochemically deposited liquid Li in molten LiCl-KCl has been observed to form droplets on an attached cathode.[ 10 ] Despite its importance, there is only limited knowledge about the dynamic reducing behavior of TiO 2 by liquid Li. Recently, black, film-like, colloidal Li (in the form of a metal fog) was observed on an electrodeposited thin Li metal in molten LiCl-KCl.[ 11 ] The detailed behavior of reducing TiO 2 ,however, has yet to be clarified. A detailed understanding of the dynamic behavior of TiO 2 reduction is necessary to control and optimize the electrolysis. Besides improving the FFC and OS processes, such knowledge can be applied immediately to the current molten salt electrolysis processes and would bring large energy savings due to increased thermal efficiency in the metal-refining industries. Cyclic voltammograms (CVs) of the oxide electrode in high-temperature molten salts display unique features that the reduction current including multi-interfacial transient dynamic behavior.[ 12 ] Data on the reduction rate, current efficiency, and energy consumption during the electroreduction of oxides under potentiostatic conditions were recorded, and these experimental findings form the basis of the optimization of the electroreduction method.[ 13 , 14 ] For an in-depth discussion, we developed an in situ observation technique to observe the TiO 2 interfacial behavior in molten LiCl-KCl electrolysis by tracking the thin fine TiO 2 electrode surface obtained by the high-speed digital microscopy synchronized with the electrochemical measurement in this study. A schematic diagram of the experimental apparatus is depicted in a Reference 10 and 15 . A vertical quartz glass vessel with a barrel-vaulted (semicylindrical) shape, 100 mm in diameter and 250 mm in height, (Kondo Science, Inc.) was employed. The flat side of the vessel enabled in situ observations. An electric resistance furnace (SiC heater) with a flat quartz window was designed to observe the phenomena within the vessel under controlled temperature. A metal halide light (Photron Co., Ltd., HVC-SL, maximum light flux: 12,500 lm) was used as an auxiliary light source. Changes in the electrode interface were recorded at a rate of 500 fps (0.002 second intervals) with the image size of 640 × 480 pixels using a high-speed digital camera (Ditect Co., Ltd., HAS-D71, monochrome). With a long-distance zoom lens (VS Technology Co., Ltd., VSZ-10100, working distance: 95 mm), the minimum field of view is 666 μ m × 500 μ m and the maximum resolution of 1.04 μ m was obtained. The electrode surface morphology was tracked in each captured image by image processing software (Photron Co., Ltd., PFV Viewer). Reagent-grade LiCl (Wako Pure Chemical Co. Ltd., > 99.0 pct) and KCl (Wako, > 99.5 pct) were used for preparing the melt. The eutectic mixture of LiCl–KCl (59:41 mol pct, melting point = 625 K (352 °C)) was packed in a borosilicate glass crucible with a flat side was settled in the vessel and was dried in vacuum at 573 K (300 °C) for more than 12 hours. Then, it was heated up to 673 K (400 °C, the constant experimental temperature) and maintained for 5 hours to remove residual water. All experiments were conducted in an Ar atmosphere (> 99.9995 pct). The melt temperature was measured with a K-type thermocouple with a glass sheath. After the mixed salt was melted, the suspended electrodes were immersed into the melt. Three types of working electrode were prepared from Ti rod (Nilaco Corp., ϕ 1.5 mm, 99.5 pct). First was as received, second and third were heat-treated for 1 hour under 1073 K and 1173 K, respectively. Figure 1 (a) shows the appearances of prepared titanium oxide electrodes. On the surface of the heat-treated electrodes, a white oxide film formed on each. The scanning electron microscope (SEM) image shown in Figure 1 (b) clearly showed the formation of oxide film on the surface of the electrode, and the X-ray diffraction (XRD) pattern of the fabricated electrode (Figure 1 (c)) identified the oxide film as mainly TiO 2 . The average thicknesses of the oxide films obtained from the SEM images increased with the heat-treatment temperature, namely 0.0, 2.0 and 7.3 μ m; thus, the three samples were named T0.0, T2.0 and T7.3. The immersion depth of the working electrode was fixed at 40 mm and the other part of the electrodes was insulated using protective Al 2 O 3 tube as shown in Figure 1 (a). The counter electrode was a graphite rod (Toyo Carbon Corp., ϕ 10 mm). An Ag + /Ag reference electrode was employed. This electrode, consisted of a silver wire ( ϕ 1.0 mm, 99.99 pct, Nilaco), a LiCl-KCl eutectic melt containing 0.5 mol pct AgCl (Wako, 99.5 pct) and a borosilicate tube, and the silver wire was immersed in the eutectic melt within the tube. Electrochemical measurements were performed using an automatic polarization system (Hokuto Denko Corp., HZ-5000). The inter-electrode voltage and microscope images were synchronized with an error less than 4 μ s using an analog signal synchronous system (Ditect Co., Ltd., DI-SYNC 29 N). Fig. 1 Titanium oxide electrode previously prepared by heat treatment under air environment. ( a ) Photos of the three, different-thickness titanium oxide electrodes. ( b ) Scanning electron microscope (SEM) image of the T2.0 electrode cross-section. ( c ) X-ray diffraction (XRD) pattern of electrodes before and after heat treatment Figure 2 shows the cyclic voltammogram (CV, scan rate: 10 mV/s) obtained by the T7.5 working electrode. Sharp increases in the cathodic and corresponding anodic currents compared to the Ag/Ag + reference electrode were observed at about E = − 2.4 V. These two current changes are thought to be due to the deposition of Li( l ) and the dissolution of the deposits, respectively.[ 16 ] The cathodic current at E = − 0.7 to − 1.0 V is the reported reduction potential of the titanium ion: Ti 4+ + e − → Ti 3+ ( E = − 0.9 V); Ti 3+ + e − → Ti 2+ ( E = − 1.0 V).[ 17 ] Although the solubility of Li 2 O was estimated to be 0.31 mol pct in LiCl-KCl eutectic melt at 673 K (400 °C),[ 8 ] because of the negligible solubility of TiO 2 , the direct electrochemical reduction of the TiO 2 following the mechanism of the FFC process described in Eq. [ 1 ]. 1 TiO x + e - → TiO x - 1 + 1 2 O 2 - \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\text{TiO}}_{x} + e^{ - } \to {\text{TiO}}_{x - 1} + \frac{1}{2}{\text{O}}^{2 - } $$\end{document} Fig. 2 Cyclic voltammogram and corresponding photographs of T7.5 titanium oxide electrode in LiCl-KCl. Images were taken at ( a ) − 0.50 V, ( b ) − 0.70 V, ( c ) −1.00 V, ( d ) − 1.10 V, ( e ) − 1.30 V, ( f ) − 2.50 V, ( g ) − 2.45 V, ( h ) − 0.50 V With the progress of potential, the appearance of the TiO 2 electrode changed from white to black as shown in the photos in Figures 2 (a) through (c). Figure 2 (d) showed the breakage of the black surface, and small pieces broke apart from the electrode. At this moment, the current density became approximately zero, and thus the reactivity of the electrode surface was lost due to mechanical disconnection. In the initial stage of potential progression, the deformation of the titanium oxide occurred due to the volume change with the electrochemical reduction. Next, the deformed part of the surface was partially peeled off from the electrode and dispersed into the melt. In Figures 2 (f) and (g), no metal fog was observed in the Li( l ) electrodeposition while only silver white precipitate was observed. According to a previous study,[ 10 ] when an inactive Mo electrode was used as the working electrode, a blue metal fog was generated, and the amount of the fog generation decreased as the amount of oxide ions increased. The metal fog is also related to the amount of Li( l ) and the time of the dissolution; however, it was likely that the oxide ions already existed in the vicinity of the working electrode due to the titanium oxide reduction. The ratio of the electric charges passing through the anode to one through the cathode, ( q a / q c \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ q_{\text{a}} /q_{\text{c}} $$\end{document} ), in the region of E < − 2.0 V gave a momentary coulombic efficiency of q a / q c = 0.871 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ q_{\text{a}} /q_{\text{c}} = 0.871 $$\end{document} . This showed that the precipitation of the Li( l ) advanced the cathodic reaction of the Ti oxide by the OS process mechanism as described by Eq. [2]: 2a Li + + e - → Li l \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\text{Li}}^{ + } + e^{ - } \to {\text{Li}}\left( l \right) $$\end{document} 2b TiO y + 2 Li ( l ) → TiO y - 1 + Li 2 O ( dissolved ) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\text{TiO}}_{\text{y}} + 2{\text{Li}}(l) \to {\text{TiO}}_{{{\text{y}} - 1}} + {\text{Li}}_{2} {\text{O}}({\text{dissolved}}) $$\end{document} In the CV measurement, mechanical deformation of titanium oxide occurs during the potential sweep, so it is difficult to know the behavior of the TiO 2 electrode that corresponds to the Li( l ) electrodeposition potential. Figure 3 shows the behavior of the TiO 2 electrode surface obtained by the high-speed observation during the chronoamperometry measurement under − 2.75 V for 2.000 seconds. (See also the Electronic Supplementary video files for more detailed understanding. Each video is 6 seconds long.) Fig. 3 Electroreduction behavior of T2.0 and T7.3 TiO 2 electrodes observed by chronoamperometry (at E = − 2.75 V). ( a ) Current–time curves, ( b ) representative photographs of electrode After the initial increase, the current converged to the approximately constant value in all samples. The initial peak current was the largest at T0.0 and showed a tendency to decrease as the TiO 2 film thickness increased. The surface of the T2.0 electrode changed from white to black immediately after the application of potential. This change proceeded non-uniformly and showed mottled pattern grown from black spots. The black part was the titanium oxide reduced by the mechanism of Eq. [ 1 ] (or perovskites like LiTiO 2 and LiTi 2 O 4 [ 18 ]). After the electrolysis at a constant potential, the black-colored titanium oxide surface was also observed in the CV test. In the case of the T7.5 electrode, the black surface rapidly developed, and it became uniform as electrochemical reduction progressed. After t = 0.08 second, the metallic droplets formed, and the metallic liquid spread onto the cathode surface. This phenomenon indicated that the reduction reaction of the TiO x was progressing by the electrochemical deposition mechanism of Eqs. [ 2a ] and [ 2b ]. After t = 2.0 seconds, all metallic droplets disappeared, and the black titanium oxide surface remained. To understand in detail the transient change in the TiO x structure, our future work will involve the chemical analysis of the sample. From the differences in behaviors of the T2.0 and T7.5 electrodes, the following conclusions can be drawn. For the initial TiO 2 with a high electric resistance, the reduction proceeds by the mechanism of Eq. [ 1 ] and the conductivity increases; this reaction should grow heterogeneously on the cathode surface. Therefore, after Li( l ) generated in a region having a relatively high conductivity on the cathode surface, the reaction progresses by spreading over the surface of the electrode. The T7.5 sample has not only thicker but also more porous TiO 2 layer than the T2.0 film.[ 19 , 20 ] The reduction of film TiO 2 did not damage the electrolytic product because of the presence of the electrochemical reaction interface, i.e . highly conductive Li ( l ) and molten salt, on the oxide film. Thus, before electrolysis, the molten salt penetrates into the inside of the porous TiO 2 , and the generated Li( 1 ) droplet causes a lower electrical resistance. The reduction of titanium oxide on the surface progresses by Eq. 2 to wet Li( l ) on the entire electrode surface. Figure 4 shows the voltage behavior between WE and CE, U , and the observed T2.0-TiO 2 electrode surface under various potentials for 2.000 seconds. At E = − 1.0 V, V became almost constant and showed a maximum at t = 0.08 second, but V at E = − 1.5, − 2.0, and − 2.5 V showed unstable behaviors. As seen in the figure, the color of each electrode changed to black after the potential was applied. However, when TiO x broke locally as shown in Figure 2 , the electrical conductivity of the electrode became non-uniform. This is the probable reason for voltage fluctuations. When E = − 2.75 V, some shiny parts that appear to be Li( l ) are found at t = 1.90 seconds. Here, Eq. [ 1 ] was the dominant reaction initially, but Eq. [2] was dominant on some parts of the electrode surface. Once Li( l ) is formed, it becomes the current-carrying site and Eq. [2] proceeds further. Fig. 4 Electroreduction behavior of T2.0 TiO 2 electrodes observed by chronoamperometry (at various potentials). ( a ) Voltage–time curves, ( b ) representative photographs of electrode Summarizing, a system was constructed to visualize titanium oxide reduction based on molten salt electrolysis to better understand its representative interfacial behavior in this study. We focused on reducing TiO 2 in molten LiCl–KCl eutectic salt at 673 K. The observed characteristic interfacial behaviors include (1) mechanical breakage of titanium oxide by the electrochemical reduction and dispersion of titanium oxide into the bath, (2) local electrochemical reductions at the conductive regions, due to surface nonuniformity, and (3) electrodeposited Li( l ) wetting the titanium oxide during the reduction reaction. The results from both the cathode interfacial snapshots and the current-time curves suggest that it is important to increase the area of the reactive region between the titanium oxide and the molten salt for efficient reduction by electrodeposited Li( l ) by avoiding a concentration of Li( l ). This work was made possible by the financial support from the Grant-in-Aid for Scientific Research (KAKENHI Grant No. 18K14036), the Iketani Science and Technology Foundation (Grant No. 0291073-A), Tanikawa Fund Promotion of Thermal Technology, and Amano Institute of Technology. Funding Information: This work was made possible by the financial support from the Grant-in-Aid for Scientific Research (KAKENHI Grant No. 18K14036), the Iketani Science and Technology Foundation (Grant No. 0291073-A), Tanikawa Fund Promotion of Thermal Technology, and Amano Institute of Technology. Publisher Copyright: {\textcopyright} 2019, The Minerals, Metals & Materials Society and ASM International.",

year = "2020",

month = feb,

day = "1",

doi = "10.1007/s11663-019-01733-7",

language = "English",

volume = "51",

pages = "11--15",

journal = "Metallurgical and Materials Transactions B: Process Metallurgy and Materials Processing Science",

issn = "1073-5615",

publisher = "ASM International",

number = "1",

}

TY - JOUR

T1 - Visualization of TiO2 Reduction Behavior in Molten Salt Electrolysis

AU - Natsui, Shungo

AU - Sudo, Takuya

AU - Shibuya, Ryota

AU - Nogami, Hiroshi

AU - Kikuchi, Tatsuya

AU - Suzuki, Ryosuke O.

N1 - Funding Information: The online version of this article (10.1007/s11663-019-01733-7) contains supplementary material, which is available to authorized users. In the refining industries of the oxide-stable metallic elements, the electrolysis of the molten chloride is indispensable. Thus, and efficient electrolysis has been developed, for example, for use in the Kroll process for titanium production. In the Kroll process, TiO 2 is first converted to TiCl 4 by Cl 2 gas. Then, liquid Mg reduces the TiCl 4 such that high-purity metallic sponge Ti is obtained. The liquid Mg and gaseous Cl 2 are regenerated by electrolysis of the byproduct MgCl 2 and recycled. To avoid some complicated steps in the Kroll process, the direct electrochemical decomposition of TiO 2 in molten CaCl 2 has been proposed. In the FFC Cambridge process, the oxide anion cathodically transfers from the solid TiO 2 pellet to the anode in a molten salt bath.[ 1 ] Because the Ti-O binary system contains many suboxides, oxygen in the higher oxide is removed to form a lower oxide upon receiving an electrical charge from the cathode. Another promising method, the OS process, has been proposed, in which the oxide anion transfer in CaCl 2 is better utilized because as much as 20 mol pct CaO that acts as an electrolyte can dissolve into the molten CaCl 2 at 1173 K.[ 2 – 5 ] The electrochemically deposited, liquid Ca at the cathode also dissolves into the molten CaCl 2 , and the dissolved Ca works effectively to reduce the titanium oxide powder. Similarly, LiCl and its binary chloride systems can dissolve oxygen anions at lower temperatures,[ 6 – 8 ] and KCl is sometimes added to lower the temperature further.[ 9 ] Electrochemically deposited liquid Li in molten LiCl-KCl has been observed to form droplets on an attached cathode.[ 10 ] Despite its importance, there is only limited knowledge about the dynamic reducing behavior of TiO 2 by liquid Li. Recently, black, film-like, colloidal Li (in the form of a metal fog) was observed on an electrodeposited thin Li metal in molten LiCl-KCl.[ 11 ] The detailed behavior of reducing TiO 2 ,however, has yet to be clarified. A detailed understanding of the dynamic behavior of TiO 2 reduction is necessary to control and optimize the electrolysis. Besides improving the FFC and OS processes, such knowledge can be applied immediately to the current molten salt electrolysis processes and would bring large energy savings due to increased thermal efficiency in the metal-refining industries. Cyclic voltammograms (CVs) of the oxide electrode in high-temperature molten salts display unique features that the reduction current including multi-interfacial transient dynamic behavior.[ 12 ] Data on the reduction rate, current efficiency, and energy consumption during the electroreduction of oxides under potentiostatic conditions were recorded, and these experimental findings form the basis of the optimization of the electroreduction method.[ 13 , 14 ] For an in-depth discussion, we developed an in situ observation technique to observe the TiO 2 interfacial behavior in molten LiCl-KCl electrolysis by tracking the thin fine TiO 2 electrode surface obtained by the high-speed digital microscopy synchronized with the electrochemical measurement in this study. A schematic diagram of the experimental apparatus is depicted in a Reference 10 and 15 . A vertical quartz glass vessel with a barrel-vaulted (semicylindrical) shape, 100 mm in diameter and 250 mm in height, (Kondo Science, Inc.) was employed. The flat side of the vessel enabled in situ observations. An electric resistance furnace (SiC heater) with a flat quartz window was designed to observe the phenomena within the vessel under controlled temperature. A metal halide light (Photron Co., Ltd., HVC-SL, maximum light flux: 12,500 lm) was used as an auxiliary light source. Changes in the electrode interface were recorded at a rate of 500 fps (0.002 second intervals) with the image size of 640 × 480 pixels using a high-speed digital camera (Ditect Co., Ltd., HAS-D71, monochrome). With a long-distance zoom lens (VS Technology Co., Ltd., VSZ-10100, working distance: 95 mm), the minimum field of view is 666 μ m × 500 μ m and the maximum resolution of 1.04 μ m was obtained. The electrode surface morphology was tracked in each captured image by image processing software (Photron Co., Ltd., PFV Viewer). Reagent-grade LiCl (Wako Pure Chemical Co. Ltd., > 99.0 pct) and KCl (Wako, > 99.5 pct) were used for preparing the melt. The eutectic mixture of LiCl–KCl (59:41 mol pct, melting point = 625 K (352 °C)) was packed in a borosilicate glass crucible with a flat side was settled in the vessel and was dried in vacuum at 573 K (300 °C) for more than 12 hours. Then, it was heated up to 673 K (400 °C, the constant experimental temperature) and maintained for 5 hours to remove residual water. All experiments were conducted in an Ar atmosphere (> 99.9995 pct). The melt temperature was measured with a K-type thermocouple with a glass sheath. After the mixed salt was melted, the suspended electrodes were immersed into the melt. Three types of working electrode were prepared from Ti rod (Nilaco Corp., ϕ 1.5 mm, 99.5 pct). First was as received, second and third were heat-treated for 1 hour under 1073 K and 1173 K, respectively. Figure 1 (a) shows the appearances of prepared titanium oxide electrodes. On the surface of the heat-treated electrodes, a white oxide film formed on each. The scanning electron microscope (SEM) image shown in Figure 1 (b) clearly showed the formation of oxide film on the surface of the electrode, and the X-ray diffraction (XRD) pattern of the fabricated electrode (Figure 1 (c)) identified the oxide film as mainly TiO 2 . The average thicknesses of the oxide films obtained from the SEM images increased with the heat-treatment temperature, namely 0.0, 2.0 and 7.3 μ m; thus, the three samples were named T0.0, T2.0 and T7.3. The immersion depth of the working electrode was fixed at 40 mm and the other part of the electrodes was insulated using protective Al 2 O 3 tube as shown in Figure 1 (a). The counter electrode was a graphite rod (Toyo Carbon Corp., ϕ 10 mm). An Ag + /Ag reference electrode was employed. This electrode, consisted of a silver wire ( ϕ 1.0 mm, 99.99 pct, Nilaco), a LiCl-KCl eutectic melt containing 0.5 mol pct AgCl (Wako, 99.5 pct) and a borosilicate tube, and the silver wire was immersed in the eutectic melt within the tube. Electrochemical measurements were performed using an automatic polarization system (Hokuto Denko Corp., HZ-5000). The inter-electrode voltage and microscope images were synchronized with an error less than 4 μ s using an analog signal synchronous system (Ditect Co., Ltd., DI-SYNC 29 N). Fig. 1 Titanium oxide electrode previously prepared by heat treatment under air environment. ( a ) Photos of the three, different-thickness titanium oxide electrodes. ( b ) Scanning electron microscope (SEM) image of the T2.0 electrode cross-section. ( c ) X-ray diffraction (XRD) pattern of electrodes before and after heat treatment Figure 2 shows the cyclic voltammogram (CV, scan rate: 10 mV/s) obtained by the T7.5 working electrode. Sharp increases in the cathodic and corresponding anodic currents compared to the Ag/Ag + reference electrode were observed at about E = − 2.4 V. These two current changes are thought to be due to the deposition of Li( l ) and the dissolution of the deposits, respectively.[ 16 ] The cathodic current at E = − 0.7 to − 1.0 V is the reported reduction potential of the titanium ion: Ti 4+ + e − → Ti 3+ ( E = − 0.9 V); Ti 3+ + e − → Ti 2+ ( E = − 1.0 V).[ 17 ] Although the solubility of Li 2 O was estimated to be 0.31 mol pct in LiCl-KCl eutectic melt at 673 K (400 °C),[ 8 ] because of the negligible solubility of TiO 2 , the direct electrochemical reduction of the TiO 2 following the mechanism of the FFC process described in Eq. [ 1 ]. 1 TiO x + e - → TiO x - 1 + 1 2 O 2 - \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\text{TiO}}_{x} + e^{ - } \to {\text{TiO}}_{x - 1} + \frac{1}{2}{\text{O}}^{2 - } $$\end{document} Fig. 2 Cyclic voltammogram and corresponding photographs of T7.5 titanium oxide electrode in LiCl-KCl. Images were taken at ( a ) − 0.50 V, ( b ) − 0.70 V, ( c ) −1.00 V, ( d ) − 1.10 V, ( e ) − 1.30 V, ( f ) − 2.50 V, ( g ) − 2.45 V, ( h ) − 0.50 V With the progress of potential, the appearance of the TiO 2 electrode changed from white to black as shown in the photos in Figures 2 (a) through (c). Figure 2 (d) showed the breakage of the black surface, and small pieces broke apart from the electrode. At this moment, the current density became approximately zero, and thus the reactivity of the electrode surface was lost due to mechanical disconnection. In the initial stage of potential progression, the deformation of the titanium oxide occurred due to the volume change with the electrochemical reduction. Next, the deformed part of the surface was partially peeled off from the electrode and dispersed into the melt. In Figures 2 (f) and (g), no metal fog was observed in the Li( l ) electrodeposition while only silver white precipitate was observed. According to a previous study,[ 10 ] when an inactive Mo electrode was used as the working electrode, a blue metal fog was generated, and the amount of the fog generation decreased as the amount of oxide ions increased. The metal fog is also related to the amount of Li( l ) and the time of the dissolution; however, it was likely that the oxide ions already existed in the vicinity of the working electrode due to the titanium oxide reduction. The ratio of the electric charges passing through the anode to one through the cathode, ( q a / q c \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ q_{\text{a}} /q_{\text{c}} $$\end{document} ), in the region of E < − 2.0 V gave a momentary coulombic efficiency of q a / q c = 0.871 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ q_{\text{a}} /q_{\text{c}} = 0.871 $$\end{document} . This showed that the precipitation of the Li( l ) advanced the cathodic reaction of the Ti oxide by the OS process mechanism as described by Eq. [2]: 2a Li + + e - → Li l \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\text{Li}}^{ + } + e^{ - } \to {\text{Li}}\left( l \right) $$\end{document} 2b TiO y + 2 Li ( l ) → TiO y - 1 + Li 2 O ( dissolved ) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\text{TiO}}_{\text{y}} + 2{\text{Li}}(l) \to {\text{TiO}}_{{{\text{y}} - 1}} + {\text{Li}}_{2} {\text{O}}({\text{dissolved}}) $$\end{document} In the CV measurement, mechanical deformation of titanium oxide occurs during the potential sweep, so it is difficult to know the behavior of the TiO 2 electrode that corresponds to the Li( l ) electrodeposition potential. Figure 3 shows the behavior of the TiO 2 electrode surface obtained by the high-speed observation during the chronoamperometry measurement under − 2.75 V for 2.000 seconds. (See also the Electronic Supplementary video files for more detailed understanding. Each video is 6 seconds long.) Fig. 3 Electroreduction behavior of T2.0 and T7.3 TiO 2 electrodes observed by chronoamperometry (at E = − 2.75 V). ( a ) Current–time curves, ( b ) representative photographs of electrode After the initial increase, the current converged to the approximately constant value in all samples. The initial peak current was the largest at T0.0 and showed a tendency to decrease as the TiO 2 film thickness increased. The surface of the T2.0 electrode changed from white to black immediately after the application of potential. This change proceeded non-uniformly and showed mottled pattern grown from black spots. The black part was the titanium oxide reduced by the mechanism of Eq. [ 1 ] (or perovskites like LiTiO 2 and LiTi 2 O 4 [ 18 ]). After the electrolysis at a constant potential, the black-colored titanium oxide surface was also observed in the CV test. In the case of the T7.5 electrode, the black surface rapidly developed, and it became uniform as electrochemical reduction progressed. After t = 0.08 second, the metallic droplets formed, and the metallic liquid spread onto the cathode surface. This phenomenon indicated that the reduction reaction of the TiO x was progressing by the electrochemical deposition mechanism of Eqs. [ 2a ] and [ 2b ]. After t = 2.0 seconds, all metallic droplets disappeared, and the black titanium oxide surface remained. To understand in detail the transient change in the TiO x structure, our future work will involve the chemical analysis of the sample. From the differences in behaviors of the T2.0 and T7.5 electrodes, the following conclusions can be drawn. For the initial TiO 2 with a high electric resistance, the reduction proceeds by the mechanism of Eq. [ 1 ] and the conductivity increases; this reaction should grow heterogeneously on the cathode surface. Therefore, after Li( l ) generated in a region having a relatively high conductivity on the cathode surface, the reaction progresses by spreading over the surface of the electrode. The T7.5 sample has not only thicker but also more porous TiO 2 layer than the T2.0 film.[ 19 , 20 ] The reduction of film TiO 2 did not damage the electrolytic product because of the presence of the electrochemical reaction interface, i.e . highly conductive Li ( l ) and molten salt, on the oxide film. Thus, before electrolysis, the molten salt penetrates into the inside of the porous TiO 2 , and the generated Li( 1 ) droplet causes a lower electrical resistance. The reduction of titanium oxide on the surface progresses by Eq. 2 to wet Li( l ) on the entire electrode surface. Figure 4 shows the voltage behavior between WE and CE, U , and the observed T2.0-TiO 2 electrode surface under various potentials for 2.000 seconds. At E = − 1.0 V, V became almost constant and showed a maximum at t = 0.08 second, but V at E = − 1.5, − 2.0, and − 2.5 V showed unstable behaviors. As seen in the figure, the color of each electrode changed to black after the potential was applied. However, when TiO x broke locally as shown in Figure 2 , the electrical conductivity of the electrode became non-uniform. This is the probable reason for voltage fluctuations. When E = − 2.75 V, some shiny parts that appear to be Li( l ) are found at t = 1.90 seconds. Here, Eq. [ 1 ] was the dominant reaction initially, but Eq. [2] was dominant on some parts of the electrode surface. Once Li( l ) is formed, it becomes the current-carrying site and Eq. [2] proceeds further. Fig. 4 Electroreduction behavior of T2.0 TiO 2 electrodes observed by chronoamperometry (at various potentials). ( a ) Voltage–time curves, ( b ) representative photographs of electrode Summarizing, a system was constructed to visualize titanium oxide reduction based on molten salt electrolysis to better understand its representative interfacial behavior in this study. We focused on reducing TiO 2 in molten LiCl–KCl eutectic salt at 673 K. The observed characteristic interfacial behaviors include (1) mechanical breakage of titanium oxide by the electrochemical reduction and dispersion of titanium oxide into the bath, (2) local electrochemical reductions at the conductive regions, due to surface nonuniformity, and (3) electrodeposited Li( l ) wetting the titanium oxide during the reduction reaction. The results from both the cathode interfacial snapshots and the current-time curves suggest that it is important to increase the area of the reactive region between the titanium oxide and the molten salt for efficient reduction by electrodeposited Li( l ) by avoiding a concentration of Li( l ). This work was made possible by the financial support from the Grant-in-Aid for Scientific Research (KAKENHI Grant No. 18K14036), the Iketani Science and Technology Foundation (Grant No. 0291073-A), Tanikawa Fund Promotion of Thermal Technology, and Amano Institute of Technology. Funding Information: This work was made possible by the financial support from the Grant-in-Aid for Scientific Research (KAKENHI Grant No. 18K14036), the Iketani Science and Technology Foundation (Grant No. 0291073-A), Tanikawa Fund Promotion of Thermal Technology, and Amano Institute of Technology. Publisher Copyright: © 2019, The Minerals, Metals & Materials Society and ASM International.

PY - 2020/2/1

Y1 - 2020/2/1

N2 - An in situ observation technique of the TiO2 interfacial behavior in molten LiCl-KCl electrolysis was developed. The variation of the thin TiO2 electrode surface were tracked through the high-speed digital microscopy synchronized with the electrochemical measurement. Two characteristic interfacial behaviors were discovered: physical breakage of the titanium oxide and Li(l) spreading on electrode surface. These electrochemically induced interfacial behaviors affect the current-time curves due to the heterogeneity of the titanium oxide film shape.

AB - An in situ observation technique of the TiO2 interfacial behavior in molten LiCl-KCl electrolysis was developed. The variation of the thin TiO2 electrode surface were tracked through the high-speed digital microscopy synchronized with the electrochemical measurement. Two characteristic interfacial behaviors were discovered: physical breakage of the titanium oxide and Li(l) spreading on electrode surface. These electrochemically induced interfacial behaviors affect the current-time curves due to the heterogeneity of the titanium oxide film shape.

UR - http://www.scopus.com/inward/record.url?scp=85076213971&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85076213971&partnerID=8YFLogxK

U2 - 10.1007/s11663-019-01733-7

DO - 10.1007/s11663-019-01733-7

M3 - Article

AN - SCOPUS:85076213971

SN - 1073-5615

VL - 51

SP - 11

EP - 15

JO - Metallurgical and Materials Transactions B: Process Metallurgy and Materials Processing Science

JF - Metallurgical and Materials Transactions B: Process Metallurgy and Materials Processing Science

IS - 1

ER -

Visualization of TiO₂ Reduction Behavior in Molten Salt Electrolysis

Abstract

Access to Document

Other files and links

Fingerprint

Cite this