Materials Chemistry
Low Temperature Electrodeposition of Titanium in Fluoride-Added LiCl–KCl–CsCl Molten Salt
2020 Volume 61 Issue 8 Pages 1651-1656
Details
2020 Volume 61 Issue 8 Pages 1651-1656
We investigated electroreduction of Ti2+ to metallic titanium in the LiCl–KCl–CsCl system, which has a eutectic point at 263 °C. Cyclic voltammetry was carried out in the TiCl2-saturated LiCl–KCl–CsCl at 300 °C, and the effect of F− addition on the reduction behavior of Ti2+ was investigated. It was revealed that addition of F− increases the limiting current density of electrodeposition of titanium at 300–500 °C. This might be mainly attributed to the increase in total concentration of soluble titanium ions in the molten salt. Then, we demonstrated the electrodeposition of metallic titanium in the LiCl–KCl–CsCl with and without F− at 350 °C. After the electrolysis in F−-added molten salt and acid leaching, LiF remained on the deposited titanium. This is because the solubility of LiF in aqueous solution is quite low. An efficient method to remove LiF is required when F−-added LiCl–KCl–CsCl is used as electrolyte.
Change of limiting current densities of electrodeposition of Ti with fluoride ion concentration, XF, in LiCl–KCl–CsCl–KF molten salt in TiCl2-saturated condition.
Molten salt electrolysis is one of the promising methods for titanium (Ti) production. For the molten salt electrolysis, mixtures of alkali and/or alkaline earth metal halides are often used as electrolytes at 400–900 °C.1–8) For example, the lithium chloride (LiCl)–potassium chloride (KCl) system is well known due to the low eutectic point, 355 °C, at the composition of LiCl:KCl = 59.5:40.5 / mol%.9) Low-temperature process is preferable from the viewpoint of energy consumption, lifetime of electrolytic cell, and wide choice of materials for the cell. In addition, there is an academic interest in electrochemical behavior of Ti ions at low temperatures. For these reasons, organic solvents10–15) and ionic liquids16–25) are investigated for electrolytes around room temperature. Some researchers claimed that Ti was detected on a cathode after electrolysis in some organic solvents or ionic liquids. However, these deposits were not identified to be pure metallic Ti by X-ray diffraction (XRD) analysis, and it seems extremely difficult to obtain metallic Ti in organic electrolytes as mentioned in literature.15,20,23–25)
As alkali iodides have especially low melting points, some researchers report the electrodeposition of Ti in alkali iodides above 300 °C.26) Our group demonstrated the electrodeposition of Ti in the eutectic lithium iodide (LiI)–potassium iodide (KI)–cesium iodide (CsI) and obtained metallic Ti powder at 250 °C and 300 °C in a previous work.27) It was also revealed that the total concentration of soluble Ti2+ in the eutectic salt is quite low at 300 °C. This low concentration of Ti2+ limits the production rate of Ti. The eutectic LiCl–KCl–cesium chloride (CsCl) also has a low eutectic point, 263 °C, and is a promising candidate for electrolyte at low temperature. Figure 1 shows the LiCl–KCl–CsCl phase diagram.28) In this ternary system, the electrolysis can be carried out around 300 °C. In fact, Tokushige et al. demonstrated electrodeposition of Ti nanoparticles at 300 °C in the eutectic LiCl–KCl–CsCl where 0.1 mol% of K2TiF6 was added.29) They used plasma-induced cathode and rapidly rotating Ti or nickel (Ni) disk anode. The molten salt in thin layer on the disk was electrolyzed between the disk anode and plasma as cathode. The electrodeposited nanoparticles of Ti were flicked with molten salt by centrifugal force and collected. This is the only report for titanium production, as far as we know, in the LiCl–KCl–CsCl system. In the present study, using normal electrochemical set up, the electrodeposition of Ti in the eutectic LiCl–KCl–CsCl and the effect of addition of fluoride ion (F−) to the molten chloride was investigated.
LiCl–KCl–CsCl phase diagram.28)
The experiments were conducted in a glove box filled with argon (Ar) gas (O2 < 0.1 ppm, H2O < 0.1 ppm). LiCl (≥ 99.8%, Nakalai Tesque, Inc.) and CsCl (≥ 99%, Daiichi Kigenso Kagaku Kogyo Co., Ltd.) were gradually heated to 277 °C and kept for 20 h under vacuum. KCl (99.5%, Nakalai Tesque, Inc.) was dried under vacuum at 150 °C for 7 days. Then, the dried KCl was moved into the glove box. LiCl, KCl, and CsCl were mixed at the eutectic composition (LiCl:KCl:CsCl = 51:20:29 / mol%28)) and placed in a nickel (Ni) crucible (OD 82 mm × ID 78 mm) where Ti (99.8%, OSAKA Titanium technologies Co., Ltd.) was settled at the bottom. After drying under vacuum at 177 °C for 120 h, the mixed salt was melted around 600 °C under Ar flow and stirred with a stainless steel (SS) tube. Titanium dichloride (TiCl2) powder was pressed to form a pellet at 106 or 200 MPa under Ar atmosphere, and the TiCl2 pellet was added to the molten salt.
In this study, TiCl2 was prepared from the reagent mixture of aluminum trichloride (AlCl3) and titanium trichloride (TiCl3) (98%, Strem Chemicals, Inc.). The mixture in a Ni container was heated to 450–500 °C under vacuum for 40–48 h in a furnace. The volatile compounds of AlCl3 (sublimation point: 177.8 °C30)) and TiCl4 (boiling point: 136.4 °C30)) were removed due to their high vapor pressures, and TiCl2 was obtained by the following disproportionation reaction.
| \begin{equation} \text{2 TiCl$_{3}$ ($s$)} \to \text{TiCl$_{4}$ ($g$)} + \text{TiCl$_{2}$ ($s$)} \end{equation} | (1) |
The schematic of apparatus used for all experiments in this study is shown in Fig. 2. The electrochemical measurements were carried out by a potentiogalvanostat (VersaSTAT 3F, Princeton Applied Research). The potentials of electrodes vs. reference electrode were recorded by a data recorder (Agilent 34972A, Agilent Technologies).
Schematic of experimental apparatus.
Cyclic voltammetry was carried out at 300 °C before and after addition of TiCl2. A molybdenum (Mo) sheet or a graphite plate was used as working electrode (WE). Another graphite plate and lithium (Li) deposited on tungsten (W) wire were used as counter electrode (CE) and reference electrode (RE), respectively. The scan rate was set at 10 mV s−1.
The effect of F− was also investigated by cyclic voltammetry. A Mo sheet was used as WE, and Ti plates were used as RE and CE. Potassium fluoride (KF) (≥ 95.0%, Nakalai Tesque, Inc.), after drying under vacuum at 235 °C for 60 h, was added to the eutectic LiCl–KCl–CsCl where TiCl2 was saturated. The concentration of F−, XF, was adjusted to be 0, 1, 2, 5, or 7 mol% in anion ratio, and the temperature was kept at 300, 350, 400, or 500 °C.
Electrodeposition of Ti was demonstrated in TiCl2-saturated LiCl–KCl–CsCl with and without F− at 350 °C. A Mo sheet was used as WE, and Ti plates were used as RE and CE. Without addition of F−, −0.79 mA cm−2 of constant current density was applied to WE, and the total electric charge was −36 C cm−2. After addition of 5 mol% of F−, pulse current electrolysis was conducted. −40 mA cm−2 for 1 s (on-time) and 0 mA cm−2 for 1 s (off-time) were applied to WE repeatedly, and the total electric charge was −91 C cm−2. After the electrolysis, the Mo electrodes were immersed into 5 mass% hydrochloric (HCl) acid solution to remove adhered salts. The deposits on Mo electrodes were characterized by an electron probe microanalyzer and energy-dispersion X-ray spectroscopy (EPMA-EDX, JXA-8530F, JEOL Ltd.) and XRD (Cu-Kα, X’pert Pro, PANalytical).
Figure 3 shows cyclic voltammograms of Mo and graphite electrodes at 300 °C. The theoretical decomposition voltage of pure LiCl (l) is calculated to be 3.66 V at 327 °C from the standard Gibbs energy of formation of LiCl (l), −353.174 kJ mol−1.32) Therefore, waves A and B in Fig. 3(a) correspond to reaction (2) and (3), respectively.
| \begin{equation} \text{Cl$^{-}$} \to \text{1/2 Cl$_{2}$ ($g$)} + \text{e} \end{equation} | (2) |
| \begin{equation} \text{Li$^{+}$} + \text{e} \to \text{Li ($l$)} \end{equation} | (3) |
| \begin{equation} \text{Ti$^{2+}$} + \text{2 e} \to \text{Ti ($s$)} \end{equation} | (4) |
(a) Cyclic voltammograms of Mo and graphite at 300 °C before and after addition of TiCl2 and (b) enlarged view of (a).
Figure 4 shows cyclic voltammograms at 350 °C at various concentration of F− in TiCl2-saturated condition. Before F− was added, the potential of WE was scanned from 0 V to −1.0 V vs. Ti reference electrode. A small cathodic wave, indicated by black line, was observed, and this reaction reached diffusion-controlled condition below around −0.4 V vs. RE. This wave is assumed to correspond to reduction of Ti2+ to Ti.
(a) Cyclic voltammograms at 350 °C at various concentrations of F−, XF, in TiCl2-saturated condition and (b) enlarged view of (a).
At XF = 1, 2, 5, and 7 mol%, the potential of WE was firstly scanned to anodic direction from 0 to 0.8 V vs. RE, and very low anodic current density of about 1 mA cm−2 was observed. After that, potential was reversely scanned to −0.8 V vs. RE. In the reverse scan, cathodic waves below 0 V vs. RE and diffusion-controlled behavior were observed. And, the potential was again scanned to anodic direction to 0.8 V vs. RE. In the 2nd cycle, large anodic waves appeared above 0 V vs. RE. This is because the metallic Ti deposited below 0 V vs. RE in the 1st cycle was anodically dissolved in the 2nd cycle.
At XF = 5 and 7 mol%, there are multiple cathodic waves below 0 V vs. RE. It is well known that addition of F− to molten chloride leads to stabilization of higher valence Ti ions such as Ti3+ and Ti4+.36–40) In this experiment, metallic Ti coexists and can reduce Ti4+ to Ti3+, and there is no obvious partial reduction wave of Ti4+ to Ti3+ that should be at higher potential than Ti reference.39,40) Therefore, the multiple cathodic waves should be synthetic waves of the reduction of Ti2+ and Ti3+. It is reasonable to assume that Ti3+ forms by the following disproportionation reaction of added Ti2+.
| \begin{equation} \text{3 Ti$^{2+}$} \to \text{2 Ti$^{3+}$} + \text{Ti ($s$)} \end{equation} | (5) |
In this work, the current density between −0.8 and −0.6 V vs. RE was adopted as the limiting current density of electrodeposition of Ti, iL. The nominal compositions of salts and iL are summarized in Table 1. At the same composition, the higher the temperature is, the higher iL is. It is well known that maximum concentration of soluble Ti2+ 41) and diffusion coefficient of Ti2+ 4) in molten chlorides such as LiCl and LiCl–KCl increase as the temperature increases. Therefore, it seems that the increase in iL is attributed to the increase in these factors. Figure 5 shows the change of iL with the concentration of F−, XF. Because the cation ratio is almost constant as seen in Table 1, it is concluded that the increase in F− concentration from 0 to 7 mol% leads to the increase in iL. For example, iL at XF = 5 mol% is twelve times as high as that without F−. Possible reasons of the change of the limiting current density are changes in total concentrations of soluble Ti ions and/or diffusion coefficients of them by F− addition. According to the work by Song et al., diffusion coefficient of Ti2+ in F-added AlCl3–NaCl at 200 °C decreases as the F− concentration increases.42) Therefore, the increase of the limiting current density is probably attributed to the increase of total concentration of Ti ions.
Change of limiting current densities of electrodeposition of Ti with F− concentration in TiCl2-saturated condition.
Conditions of salt surfaces at various experimental temperatures are summarized in Table 2. At 400 °C and 500 °C, solid precipitate was not observed, and the salts seemed to be melted completely. It should be noted that TiCl2 pellets and particles always existed in the salts, but these were not observed at the surface of salts. This is because the TiCl2 sinks in the salt as the heavier density of TiCl2, 3.13 g cm−3 30) than that of the eutectic LiCl–KCl–CsCl, 2.21–2.34 g cm−3 at 300–500 °C.43) On the other hand, some solid precipitates were observed at the surface of the F−-added molten salt at 300 °C and 350 °C. It is assumed that the precipitates are fluorides such as lithium fluoride (LiF), KF, and/or cesium fluoride (CsF). Therefore, the additive amount of F− is limited at these temperatures.
Electrodeposition of Ti at 350 °C was demonstrated at XF = 0 and 5 mol%. Figure 6 shows the potentials of WE and CE vs. RE during the electrolysis. As the current density in the experiment using the F−-added molten salt is higher, the electrolysis time is short.
Potentials of WE and CE during electrolysis in the molten salt at the composition of (a) XF = 0 and (b) XF = 5 at 350 °C.
Figure 7 shows XRD patterns of deposits on Mo electrode. In both of the XRD patterns, the peaks corresponding to hcp-Ti are observed. In addition, Ti was detected in both deposits by EDX. We therefore confirmed that metallic Ti was obtained in the LiCl–KCl–CsCl molten salt with and without F− at 350 °C. The current efficiency of electrolysis without F− was calculated to be 83% from the weight change of Mo electrode, assuming the reduction of Ti2+ to Ti. On the other hand, in the XRD pattern of deposits obtained in the F−-added molten salt, peaks corresponding to LiF are also observed as seen in Fig. 7(b) even after the leaching. Thus, the current efficiency can not be calculated. Remaining of LiF is due to its low solubility in aqueous solution, 0.27 g / 100 g-H2O at room temperature.30)
XRD patterns of (a) deposits and Mo electrode after the electrolysis at XF = 0 and (b) deposits recovered from Mo electrode after the electrolysis at XF = 5 at 350 °C.
Optical and SEM images of deposits and Mo electrodes are shown in Fig. 8. As seen in Fig. 8(a) and (c), only gray porous Ti was obtained by the electrolysis without F−. On the deposits obtained in the F−-added molten salt, some white parts due to electron charge-up are observed as seen in Fig. 8(d). These regions should correspond to LiF because fluorine was also detected by EDX. Therefore, an efficient method for removal of LiF is required when F− is added to the eutectic LiCl–KCl–CsCl.
Optical images of deposits and Mo electrodes after electrolysis at 350 °C at (a) XF = 0 and (b) XF = 5, and SEM images of deposits obtained in the molten salt at (c) XF = 0 and (d) XF = 5.
In this study, we focused on the F−-added LiCl–KCl–CsCl as electrolyte for the low temperature electrodeposition of metallic Ti. Cyclic voltammetry was carried out at XF = 0–7 mol% in TiCl2-saturated condition. It was revealed that the addition of F− drastically increases the limiting current density of electrodeposition of Ti at 300–500 °C. This might be because the total concentration of soluble Ti ions is increased by F− addition. Metallic Ti was then obtained by electrolysis in the molten salt at XF = 0 and 5 mol% at 350 °C. The current efficiency of electrolysis without F− was calculated to be 83%. A small amount of LiF remained on the Ti deposit obtained in the F−-added molten salt after leaching. An efficient leaching method to remove LiF should be therefore developed when adding F− to the eutectic LiCl–KCl–CsCl.
This study was financially supported by New Energy and Industrial Technology Development Organization (NEDO) under the Innovative Structural Materials Project (Innovative Structural Materials Association (ISMA)) (project number: 100180900026). The authors appreciate Mr. Hideki Fujii and Mr. Matsuhide Horikawa at Toho Titanium Co., Ltd. for their valuable comments and support.
