Current Trends in Research
Recent Studies on Titanium Refining: 2017–2020
2021 Volume 62 Issue 6 Pages 905-913
Details
2021 Volume 62 Issue 6 Pages 905-913
Titanium oxide can be converted directly to its metallic state. This extraordinary technology is realized in a CaCl2-based molten salt that can extract oxygen ions from the cathode. Several discussions have been exchanged in the field of electrochemistry and metallurgy in the last two decades. Recent papers on titanium refining are overviewed by selecting popular papers, especially those published in this journal, Materials Transactions. Basic and scientific studies on titanium and its refining are briefly analyzed, and some practical applications are introduced in related oxide reductions and the corresponding molten salts.
Fig. 6 Proposed routes to prepare Ti metal.
For over 65 years, the main route for the extraction of titanium is the Kroll process.1) Its manufacturing costs have been significantly decreased, eventually surpassing the compatible method, the Hunter process.2) However, the Kroll process is still regarded as an expensive and environmentally unfriendly method,3) and the industrial applications of Ti are still restricted to a few high-end products. Hence Kroll, the inventor of Kroll process, expected that it will be replaced by another method. During the past 20 years, many molten-salt-based electrochemical approaches for the reduction of Ti oxide have been introduced to meet these challenges. These proposals led to novel extraction processes for other reactive metals such as V, Nb, Ta, and rare-earth metals.
In the year 2000, Chen et al. proposed the FFC (Fray-Farthing-Chen) Cambridge process, in which oxygen ions are extracted from the cathodic TiO2 in the molten CaCl2.4) The ionization of the oxide in the chloride melt is understood as the scientific background. This proposal also received strong feedback in practical applications.
Ono et al. at Kyoto University reported in situ calciothermic reduction after the electrochemical deposition of Ca from CaCl2 melt.5–7) This group claimed that the CaO dissolved in the chloride melt contributed as a Ca source for oxide reduction and showed that the impurity level of oxygen as low as 2000 mass ppm can be achieved directly from TiO2 powder (40 mass% oxygen). This idea was later called the OS (Ono-Suzuki) process and is exposed to critical discussions from the supporters of the FFC process.
These two processes generally use molten electrolysis in CaCl2 and start from TiO2. Most researchers have accepted the explanations based on the FFC mechanism. However, when a high voltage is applied between the two electrodes, the potential exceeds the theoretical limit for Ca precipitation from CaO, and the OS process becomes valid and dominant.
Many other ideas have been reported in the fields of Ti refining and molten salt electrochemistry. Because of the high solid solubility of oxygen in Ti and the strong affinity of oxygen with Ti, oxygen removal is a significant task for these new ideas. The improvements in Ti refining have been applied in the reduction of various metals, and many alloy powders have been prepared. This article reviews these interesting papers on electrochemical refining techniques and related topics. These innovative ideas have been published in various journals, and this paper focuses on characteristic papers published in 2017–2020 and mainly in Materials Transactions.
Haarberg in Norway proposed that researchers worldwide gather at a round table conference in Germany in 2008 to discuss the details of these new proposals. After two-year interval meetings, the “5th International Round Table on Ti Refining (TiRT-2016)” was organized in Hokkaido in 2016 by the authors of this article. All presentations and social events at TiRT-2016 were reviewed in Refs. 8–10). This journal, Materials Transactions, first offered the platform for attendants to exchange their contributions in “Special Issue on New Proposals on Titanium Production and Molten Salts” in March 2017.11) This issue is the first milestone of these successive conferences. The 6th round table was held in Reykjavik, Iceland in 2018,12,13) and Materials Transactions also compiled the contributing papers as another special issue in March 2019.13) Most of the proposals are described in a book recently edited by Fang et al. at Utah University.14)
Some proposals have already advanced to the stage of substantial pilot plants, whereas others are still being developed at the laboratory level. The first runners met several technical issues to be solved, and the other researchers need the financial assistance to scale up for practical testing. Currently many researchers thus study to realize their radical proposals. In order to increase the cost advantages, the direct formation of alloy or intermetallic compounds from the oxide mixtures are also studied. All these concurrent trials on titanium refining in 2017–2020 will be reviewed here.
The plenary lecturers of the 5th round table, Schwandt and Fray at Cambridge,15) reviewed most of the proposals published since 2000. The FFC Cambridge process was well explained there, and their comprehensive review15) was selected for the most cited paper award 2017 in Materials Transactions. They also introduced the development of anodic materials in these processes.15)
Hu et al.16) pointed out the low current efficiency of the FFC process, which may be improved by reducing the electric field intensity in the electronic cell. They claimed that the background current was due to high electronic conductivity. However, the reasons for the background current at the well-proceeded stage have been often discussed since 2000, and the authors believe that a low concentration of CO32− in the melt is a key for improving current efficiency.
The CO32− formation after the anodic reaction with C was suppressed by the solid ZrO2–8 mol% Y2O3 membrane in molten CaCl2.17,18) The usage of membrane is quite effective to suppress the side-reaction, although it decreases the reaction rate due to IR drop at the membrane. It was used to produce Ti alloys/composites from Ti-bearing blast-furnace slag, high titanium slag, and natural ilmenite at 1223–1273 K. Because CaCl2 decomposes at 3.2 V, the applied 3.8 V was high enough to form the metallic Ca from CaCl2. Consequently, both CaTiO3 and Ca12Al14O33 were generated as intermediate products.18) Elemental Fe was generated preferentially if the starting material was ilmenite, which improved the electrical conductivity of the oxide pellet.
CaTiO3 is an inevitable intermediate phase formed preferentially in the cathode at the early stage of TiO2 electrolytic deoxidation in both the FFC and OS processes. Many researchers have extensively discussed the reduction path and kinetics of the low reproducibility in the cooled samples. Hu et al.19–21) showed that CaTiO3 was not formed during TiO electrolysis; however, it appeared in the cathode when TiO2, Ti3O5 and Ti2O3 were used. Chlorine gas evolution was particularly detected except when starting from TiO. Several points remain unclear in the reduction route, because the reduction sequence from Ti oxides was different from the study via in situ XRD analysis by Bahgat et al.22) Because of slow reactions, the authors believe that the reduction route obeys the phase diagram23) and that because of the effect of water impurities in CaCl2, oxygen potential affects the phase stability and reduction sequence in the Ti–O-containing system. The crystal structure of TiO2 (rutile or anatase), the primary and secondary particle sizes in the powder, the degree of calcination for oxide pellets and the water contamination give the significant effects on the sequence of oxygen removal, the current density, the obtained crystalline nature as metallic powder and the morphologies of Ti metal.
2.2 Synthesis of alloysA Ti–13Zr–13Nb alloy24) was prepared by direct electrochemical reduction of a sintered TiO2–ZrO2–Nb2O5 mixture in molten CaCl2 at 1123 K, and the corrosion behavior of dual-phase α + β with a porous structure and designed composition was tested for application as an implant.
Ti–5Ta–2Nb alloy25) was synthesized from a TiO2–Ta2O5–Nb2O5 mixture via the FFC Cambridge process in a molten CaCl2 electrolyte. It changed into dual-phase (α + β)–Ti by subsequent heat treatment. Ti–Nb–Sn biomedical alloys with high porosity26) were also prepared using a mixture of the related oxides.
Ono and Suzuki at Kyoto University proposed a direct oxide reduction from TiO2 powder, which is now called the OS process.5–7) It is based on the electrochemical decomposition of CaO dissolved in CaCl2 melt and on the calciothermic reduction of oxides.27) It was examined from various aspects, and many derived proposals or secondary improvements have been published.
3.1 Applicability at industrial levelIn 2017, Ono et al.28) showed the experiences of the industrial model of the OS process, where a 10-day campaign was conducted with 2 tons of molten salt and a current of 4000 A. Rajulu et al.29) in India studied the kinetic aspects of direct electrochemical reduction using 0.5–1.0 kg TiO2 in 5–8 kg cells. The CO2 (product gas) concentration in the vent gas during electrolysis followed a trend similar to the DC current, as shown in Fig. 1. However, the mass balance of carbon did not match well. This may indicate that a fairly large amount of CO32− remained in the melt, and the background current due to CO32− increased.
Variation of electrolysis current and CO2 concentration in vent gas with time (5–8 kg batch).29)
The addition of Fe to Ti effectively stabilizes and strengthens industrial β-Ti by solution hardening, and the restriction of Fe contamination in Ti became milder.30) However, the specifications for aircraft applications are still strict. In a scheme of α-Ti production from titanium ore (TiFeO3), Fe removal is indispensable because the molten salt electrolysis/reduction of solid oxides cannot preferentially remove Fe from the cathode. Ilmenite can be converted to CaTiO3 through wet chemical treatments. Subsequently, the electrochemical decomposition of CaTiO3 was studied by Suzuki et al.31,32) Inhomogeneous reduction in the cathodic basket is related to the buoyancy of Ca and insufficient dehydration. The rapid reduction of fine particles from the cathodic surface formed a rigid Ti layer, which impeded further reduction of the inner parts of the cathode.32) After optimizing the cooling conditions, the CaO content in the molten salt, TiO2 particle size, and dehydration methods of salt, an industrial level of 0.42 mass% O could be directly achieved as a porous powder starting from TiO2.31–33)
Haraguchi et al.34) studied electrochemical side reactions via cyclic voltammetry. The O2− in the CaCl2 melt reacted with the carbon anode, forming CO and CO2 gas bubbles, which are essential components of the OS process. On the other hand, these bubbles easily disperse or dissolve into the molten salt and formed CO32−, subsequently generating carbon powder. In the cathodic scan, electrochemical reduction of CO32− was not observed; hence, CO32− was spontaneously reduced by Ca during electrolysis. The anodic reactions preceding Cl2 gas generation occurred in three steps: CO and CO2 gas generation related to O2− and successive CO2 generation related to CO32−.
The existence of a small amount of water in CaCl2 causes rapid electrodeposition due to water impurities, resulting in low electrolysis efficiency. The morphological and thermal characteristics of a cathode in a slightly hygroscopic LiCl–KCl–CaCl2 melt were examined using a high-speed camera.35) The measured heat absorption was considerably smaller than the thermodynamically predicted value, which suggested that the generated hydrogen formed a hydride by reacting with the liquid Ca that was electrodeposited on the cathode surface.
Natsui et al.36–38) visualized the TiO2 electrode in molten LiCl–KCl eutectic salt at 673 K and found a periodic arrangement of Li droplets on the cathode. From the observation of colloidal Li deposition behavior for efficient reduction of oxides, increasing the area of the reactive region between TiO2 and the molten salt and avoiding the droplet formation of Li-concentrated liquid are important.
3.3 Oxygen in TiThe mechanical properties of α-Ti or Ti alloys, such as tensile stress and elongation, depend considerably on the oxygen content.39–43) For example, only 0.49 mass% O significantly reduces 0.2% proof stress. Technical improvement to suppress the residual oxygen in Ti refining is certainly required; however, other technical efforts to use highly oxygen-contaminated Ti are expected for the users to realize a practical low-cost high-strength alloy such as Ti–Fe–O–N alloys developed by Fujii et al.44)
3.4 Ca reductionCalciothermic reduction of TiO2 has been studied in Japan, especially for Ti production.45–48) Co-reduction from an oxide mixture or reduction-and-diffusion (RD) process using Ca as the reductant can form Ti-based alloys and Ti-containing intermetallic compounds.49–51) The starting material is normally TiO2 with a crystalline structure of rutile or anatase; however, titanium concentrate, ore, or CaTiO3 have also been used as oxide mixtures.31,52) Most of these trials were successful in producing metallic Ti, although metallic impurities could not be removed.
Thermodynamic evaluation of the expected Ca reduction is based on the quality of the Gibbs free energy of CaO, Ca, and of the residual oxygen in Ti.53) Seki54) experimentally determined the temperature dependence of the equilibrium constant for the reaction “CaO(s, l) = Ca(l, g) + 1/2 O2(g)” by using the slag/alloy equilibrium distribution method or by measuring its oxygen partial pressure using a ZrO2 solid electrolyte. The re-determined Gibbs free energy agreed well with the assessed references within an error of only 14–19 kJ/mol over a wide temperature range.
CaH2 is often used as an alternative Ca source in lab-type experiments.49) It decomposes to Ca and H2 above the melting temperature of Ca, and a similar reaction with Ca is expected because hydrogen has no reducibility to metallic Ti. However, single-phase intermetallic RENi2Si2 (RE = Y, La) nanoparticles could be prepared at temperatures as low as 873 K with molten LiCl–CaH2.55) Hydrogen-assisted reduction may be another route to remove oxygen from the oxide mixtures.
The removal of solidified CaCl2–CaO from the metallic product has been achieved by aqueous leaching. However, when CaCl2 hydrate is heated, a pyro-hydrolysis reaction occurs, making it difficult to reuse the CaCl2. Inoue and Uda56) proposed the use of nonaqueous solvents, such as dimethyl sulfoxide, for leaching solid CaCl2 from Ca-reduced Ti powder.
3.5 OS process of sulfidesSimilar to the oxide reduction in CaCl2–CaO by the OS process, the reduction of sulfide may occur when CaS dissolves in the CaCl2 melt, and when the electrochemical potential for CaS decomposition is higher than that for CaCl2. The high solubility of CaS is the first criterion for the calciothermic reduction of this metallic sulfide. Molten CaCl2 was used to dissolve 19–21 mol% CaO,57–60) whereas Matsuzaki et al.61) showed 1.5–2.2 mol% CaS as the solubility at 1123–1223 K. Only 0.22–0.31 mol% CaS can dissolve in CaCl2–65 mol% LiCl eutectic salt at 873–973 K.62)
TiS2 was successfully reduced to Ti via either calciothermic reduction or electrolysis in CaCl2–CaS melt.63) Sulfur was easily removed by Ca reduction to a low level of 0.03 mass% S at 1133 K. In the molten salt electrolysis at 1173 K in CaCl2–CaS melt, the S concentration initially decreased significantly and then gradually decreased to 0.01 mass% S when a sufficiently large amount of electric charge was supplied. Sulfide formation from TiO2 was realized by using CS2 gas, the by-product of electrolysis. The overall reaction sequence seems similar to the Kroll process, which uses TiCl4 as an intermediate substance. However, the OS process uses the in situ electrolysis of CaS to form the reductant Ca. This variation in the OS process has the advantage that Ca and S are recycled in an enclosed loop.
Most of the impurity elements in the starting materials and molten salts are absorbed into Ti powder during the operation of OS process. Ahmadi et al.64–67) converted the Ti ore to TiN or TiC prior to the formation of Ti sulfide, which could be achieved to a low level of residual sulfur, carbon, and oxygen in the metallic Ti obtained by the OS process, as shown in Fig. 2. Metallic V was also successfully produced starting from V3S4 in molten CaCl2.68) Oxygen and sulfur contents as low as 3390 ppm and 210 ppm, respectively, were achieved by supplying an electrical charge that was four times more than the stoichiometric amount. The paper written by Matsuzaki et al.68) was awarded the best paper awards of Materials Transactions in 2017 (Fig. 3).
Designed reactions to obtain Ti via TiS2.
OS process for V3S4 in molten CaCl2–CaS.68)
The Ti scraps generated during smelting and fabrication are recycled; high-grade Ti scraps with low O and Fe concentrations are remelted to obtain Ti and its alloys. Low-grade Ti scraps with high O and Fe concentrations are used mainly as a deoxidizer of molten steels and as ferrotitanium.69,70) The technologies for anti-contamination or for efficient removal of O and Fe must be developed for repeated utilization of Ti.
Okabe et al.71–73) developed a new method of dissolved O removal from Ti. The thermochemical reactivity of Mg with oxygen is weaker than that of Ca, and magneciothermic reduction could not achieve a low residual oxygen in Ti; however, Mg is cheaper than Ca. His group chose Mg as reducing agent under assistance of HoCl3, LaCl3 or YCl3 in MgCl2. For example, the deoxidation of Ti via the reaction O (in Ti) + Mg + HoCl3 → HoOCl + MgCl2 was proposed using Ho-containing MgCl2.74,75) The activity of the deoxidized product, aMgO, can be effectively maintained at a low level by forming HoOCl. The E–pO2− diagram predicts that the electrochemical deoxidation of Ti in the MgCl2–HoCl3 system is more effective than that in Mg–MgO. Iizuka et al.76,77) controlled the O concentration in Ti in the range of 200–2000 ppm O by varying the activity of Y, aY, in the Y/Y2O3 equilibrium at 1300 K in NaCl–KCl, as illustrated in Fig. 4.77) The residual oxygen can be reduced to 30–60 ppm O in the presence of Y/YOCl/YCl3. They claimed that the price of Y metal is considerably low in recent years. Zheng et al.78,79) used molten Mg–MgCl2–YCl3 for deoxidation.
Reduction and deoxidation using Y–YCl3–YOCl equilibrium.77)
Tanaka et al.80–82) reduced TiO2 to directly produce Ti with an ultralow oxygen concentration by using Y or La as the reductant. When the solvent was YCl3 or CaCl2, the TiO2 pellets were reduced to metallic Ti. Owing to the large solubility of O2− in the molten salts, the oxygen concentrations in Ti after reduction in YCl3 and CaCl2 were only 90 and 350 ppm O, respectively.
Metallic Mg can reduce TiO2 directly to Ti; however, producing Ti with 1 mass% O or less is thermodynamically impossible when pure MgO equilibrates with pure Mg. Zhang et al.82–85) at the University of Utah proposed a two-step thermochemical reduction using Mg. First, Mg removes oxygen from TiO2 to lower oxygen level. Hydrogen atmosphere probably forms TiH2. The second step is the decomposition of TiH2 to Ti. The powder was irregular or spherical in shape with various heat treatments and it contained 1100 mass ppm oxygen. These properties match with the demand from the industries.86) As shown in Fig. 5, the reduction of TiO2 to metallic state consists of several steps, even if Ca is taken as the reductant.87) The effective reaction rate of Mg reduction in the molten salt is a key for mass production, because the solubility of Mg in the chloride is considerably lower than that of Ca. MgO and Mg(OH)2 were formed as by-products;88) hence, the prices of Mg and hydrogen are of great concern.
The USTB process at University of Science and Technology Beijing89,90) consists of two steps: the carbon reduction of TiO2 to form TiC0.5O0.5, and its subsequent anodic dissolution and cathodic precipitation as the Ti film in the molten NaCl–KCl. Ti(C,O) is electronically conductive; hence, this consumable anode seems suitable for electrolysis. Although the idea of electrorefining of TiOX has long history,91–94) and the electrochemical dissolution remains slow, a large-scale plant was constructed in China and a practical test has been started.
Mu et al.95) and Jiang et al.96) studied the basic information on the synthesis of the Ti(C,O) phase. After sintering TiO2 and carbon at the composition of Ti:C:O = 2:1:1, several phases were formed: gray-white titanium carbide, yellow titanium oxycarbide, and aubergine titanium oxycarbide. They did not dissolve uniformly as the anode during electrolysis. Considering the slow kinetics of the synthesis of the Ti(C,O) phase from the oxide, careful verification might be needed to achieve equilibrium. The dissolution potential of the arc-melted Ti2CO is more negative than that of conventional powder sintering.97) A constant-current electrolysis indicated that titanium ions from the anode were stably dissolved into the melts; subsequently, metallic Ti was deposited on the cathode.
The USTB method was applied to form various Ti–Al alloy powders using KCl–LiCl–MgCl2 melt and Al anode.98)
Recent progress in Mg reduction of TiCl4 via the Kroll process in Japan has not been reported since 2007,99) although this process is the only practical process for Ti refining. Nakamura et al.100) summarized the current technical issues based on the developments at Osaka Titanium Technologies (established in 1952 with an annual production capacity of 40,000 ton Ti). The premium-quality (PQ) Ti sponge could be produced only by manufacturers targeting aircraft parts, because it requires high reliability. The demand for PQ Ti sponge is increasing; thus, cost reductions within the Kroll process are conducted using middle-to-low-grade raw materials, thereby producing the world’s largest sponge cake and improving productivity. The cost for TiCl4 preparation can be decreased through low-temperature chlorination via TiN from ilmenite101) or through recycling of polyethylene terephthalate.102)
Considering the production of titanium powder from titanium sponge as the raw material, the shuttle reactions due to disproportionation (the backward reaction of Ti2+ to Ti3+ and Ti metal) were studied at 1023 K in molten NaCl–KCl by Lu et al.103) and in molten MgCl2–LiCl by Wu et al.104) Their results showed that Ti2+ is more stable in these molten salts than in fluoride melts. A Ti-rich Ti–V alloy with a cauliflower-like morphology was deposited from LiCl–KCl at 700 K.105)
Kumamoto et al.106) chose the eutectic LiCl–KCl–CsCl system to conduct the electroreduction of Ti2+ to Ti at low temperatures such as 573 K. The addition of F− increases the current density of electrodeposition because of the increase in the total concentration of soluble Ti ions in the molten salt. The solubility of LiF in aqueous solution is considerably low; hence, an efficient method to remove LiF is required. Kumamoto et al.107) also studied the LiI–KI–CsI ternary system with a eutectic point at 477 K. The Ti2+ concentration was as low as 0.18 mol% at 573 K under TiI2-saturated conditions, and metallic Ti was obtained on the Mo cathode at 523 and 573 K. Hence, a new electrolytic process to produce bulk metal Ti at below 573 K is possible.
Xu et al.108) studied the electrodeposition of Ti in an AlCl3-1-butyl-3-methylimidazolium chloride (BMIC) ionic melt. No metallic Ti was obtained in the presence of Ti4+ and Ti2+ at 353 K. Therefore, further efforts are needed to realize low-temperature electrolysis using ionic liquids.
In both the FFC and OS processes, TiO2 and the product Ti are not melted, and the solid TiO2 at the cathode does not dissolve into the chloride melt at approximately 1173 K because it is difficult to hold the reactive liquid Ti in the ceramic crucible or the metallic vessel at above the melting temperature. Samal et al. examined carboaluminothermic reduction to form Fe–Ti alloy from ilmenite by using thermal plasma and graphite crucible.109) For low temperature operation, CaO was added to optimize CaO/Al2O3 ratio. However, molten electro-slag is expected to cover the molten Ti from open air, and lighter-than-Ti slag such as CaF2–CaO–TiO2 melt has been searched.110,111) Ozawa et al.112) precisely measured the density of liquid Ti as 4193 kgm−3 at its melting point (1941 K) by using an electrostatic levitator.
The solubility of TiO2 in molten CaCl2 remarkably increased with the addition of CaO because a titanate ion (TixOyZ−) was preferentially formed.113) The solubility of CaTiO3 in CaCl2 melt was as high as approximately 7 mol% at 1573 K,114) and the electrolysis was favorable in a bath consisting of Ca3Ti2O7.115) The reduction mechanism consists of a three-step reduction of Ti in the melt at a molar ratio of RCaO/TiO2 < 1.5 and a two-step reduction in Ca-rich melt.116)
Once metallic Ti is formed, the next problem is residual oxygen. Oxygen removal from Ti–Si melts during arc melting was investigated by Watanabe et al.117) The oxygen in the Ti–(23–30 mass%)Si ingots was removed as gaseous SiO during melting. Enhanced uniformity of oxygen distribution in the metallic liquid was achieved under the flow of He gas rather than Ar gas because the melted region was extended toward the depth direction when using He gas. A two-step plasma-arc-melting process (the first step was under Ar–30% H2 gas flow, and the second step was under Ar gas flow) was effective for the deoxidation of Ti melts.117) The oxygen content after arc melting decreased from 1.5 to 0.7 mass% at the surface of the melt.
The high temperature melting such as arc-welding in open air normally increases the oxygen content in Ti, even if a protective gas shielding is applied. Because only the surface of welded parts is oxidized, the FFC process claimed that the effective cathodic protection could be achieved.15)
The use of TiN as an intermediate has been proposed because TiN can be easily converted from TiO2. Seki118) investigated the kinetics of TiO2 carbonitrization in N2 gas. The reaction consists of the adsorption of carbon and the main combustion of carbon to form TiN. Seki and Yamaura119) expected to decompose TiN thermally by heating to over 3500 K using an electric arc furnace.
Ahmadi et al. converted TiN to TiS2 and produced Ti via the OS process.65,67) After the removal of Fe from ilmenite, Ti(C,O,N), TiN, and TiC are the intermediate products. Experimentally, TiN is remarkably better than TiC with regard to residual impurities.66) The Ti waste containing nitrogen and carbon can be used as the starting material if it is easily converted to TiS2.
The processes that start from oxides require highly purified TiO2 extracted from ilmenite. Shin et al.120) studied hydrochloric acid leaching and hydrolysis from Na2TiO3. The leaching efficiency of Ti increased to 96% Ti at 343 K. The leaching reaction was analyzed by using the shrinking core model of the heterogeneous reaction. The chemical reaction was the rate-determining step. The residue after hydrolysis at above 353 K consisted mostly of TiO2. Guo et al.121) synthesized Li4Ti5O12 in molten salt and examined the ion doping.
11.2 Corrosion by saltThe metallic materials for electrodes, reaction vessels, and sensors are exposed to a severe corrosion environment. Various efforts have been made to minimize the damage during molten salt electrolysis. Izzuddin et al.122) found that Ni–20Cr–xFe alloys were tough in a mixture of oxidizing air and chlorine-containing vapor from NaCl–KCl–CaCl2. A protective Cr2O3 scale could be maintained for a longer corrosion period at 843 K by adding Fe.
The deposition of Al from molten fluoride is commonly performed during the electrolysis of Al oxide. Fluorides are fairly stable in open air, and their vapor pressure is sufficiently low for continuous operation. Although the environmental impact of fluorine is significant, a solution containing Ti ions is favorable for stable applications.123)
Norikawa et al.124,125) investigated the electrochemical deposition of Ti in a KF–KCl molten salt at 923 K. Approximately 0.1 mol% K2TiF6 and an excess amount of Ti sponge were added to this water-soluble salt, indicating the electrochemical deposition of Ti metal from Ti3+. Dense and smooth films of Ti with 20 µm thickness were obtained. The comproportionation reaction between the metal and metallic ions has been severely controlled, particularly in rare-metal deposition in fluoride salts.126,127)
The electrochemical Dy-alloying behaviors of Inconel and Hastelloy in molten LiF–CaF2–DyF3 were studied by Watanabe et al.128) The diffusion of Ti3+ was investigated in LiF–LiCl–0.50 mol% Li2TiF6–0.33 mol% Ti at 923 K. The potentials for Ti(III)/Ti(0) and Ti(VI)/Ti(III) in LiF–LiCl were reported to be more positive than those in KF–KCl.129) Electrodeposition on the Ni substrate was conducted using K3TiF6 from K2TiF6 and Ti in the water-soluble KF–KCl molten salt at 923 K. The deposited film was adherent, compact, and smooth without cracks or voids.130)
In the last two decades, many researchers have proposed transformative ideas on Ti refining, and these ideas were further developed in 2017–2020, as illustrated in Fig. 6. The direct reduction from Ti oxide has been extensively studied, and some proposals proceeded to the large-scale testing level. Fundamental surveys have been thoroughly conducted and have been discussed by several researchers, especially those in electrochemical societies and metallurgical societies. The Japan Institute of Metals and Materials has shared important positions and serves as the information center.
Proposed routes to prepare Ti metal.
This work was supported by the Grants-in-Aid for Scientific Research (KAKENHI) Program under No. 17H03434 and 20H0249110, by Japan Mining Industry Association, and by Innovative Structural Materials Association (ISMA)-NEDO.
