Production of Fe–Ti Alloys from Mixed Slag Containing Titanium and Fe2o3 Via Direct Electrochemical Reduction in Molten Calcium Chloride

metals

Article Production of Fe–Ti Alloys from Mixed Slag Containing Titanium and Fe2O3 via Direct Electrochemical Reduction in Molten Calcium Chloride

Bo Wang 1,2 , Chao-yi Chen 1,2,*, Jun-qi Li 1,2, Lin-zhu Wang 1,2, Yuan-pei Lan 1,2 and Shi-yu Wang 1,2

1 College of Materials and Metallurgy, Guizhou University, Guiyang 550025, China; [email protected] (B.W.); [email protected] (J.-q.L.); [email protected] (L.-z.W.); [email protected] (Y.-p.L.); [email protected] (S.-y.W.) 2 Guizhou Province Key Laboratory of Metallurgical Engineering and Process Energy Saving, Guiyang 550025, China * Correspondence: [email protected]; Tel.: +86-150-8601-5817

Received: 8 October 2020; Accepted: 27 November 2020; Published: 30 November 2020

Abstract: High-purity intermetallic β-Ti (FeTi4) and FeTi alloys were prepared via molten salt electrolysis from a titanium-containing waste slag and Fe2O3 mixture using molten CaCl2 salt as the electrolyte. The mixed slag powders were pressed into a pellet that served as a cathode, while a graphite rod served as an anode. The electrochemical process was conducted at 900 ◦C with a cell voltage of 3.1 V under an inert atmosphere. The formation process of the alloys and the inﬂuence of the Ti:Fe atomic ratio on the product were investigated. With an increased proportion of Ti, the phase of the product changed from FeTi/Fe2Ti to FeTi/FeTi4, and diﬀerent structures were observed. At a Ti:Fe ratio of 1.2:1 in the raw slag, an alloy with a sponge-like morphology and a small amount of FeTi4 were obtained. During the initial stages of electrolysis, a large amount of intermediate product (CaTiO3) was formed, accompanied by an abrupt decrease in current and increase in particle size. The current then increased and Fe2Ti alloy was gradually formed. Finally, as the reaction process extended inside the pellet, the current remained stable and the product mainly contained FeTi and FeTi phases. The observed stages, i.e., CaTiO (TiO ) Fe Ti(Ti) FeTi(FeTi ), were consistent with 4 3 2 → 2 → 4 the thermodynamic analysis.

Keywords: waste slag containing titanium; molten salt electrolysis; ferro-titanium alloy; FeTi4

1. Introduction Ferrotitanium alloys are well-known hydrogen storage materials [1,2] that can be produced by various processes, such as the carbothermic or metallothermic reduction of oxide ores or concentrates. In carbothermic reductions, carbon in the form of coke, coal, or charcoal is used as a reducing agent. In metallothermic reductions, the main reducing agent is aluminum. Iron and titanium metals are common raw materials used for the synthesis of ferrotitanium alloys [3]. Waste slag containing titanium (50.15%) is generally formed during the production of sponge titanium via the Kroll process. On average, more than 200 kg of titanium-containing waste slag is produced per ton of titanium tetrachloride [4]. Owing to the small size of the slag, addition of titanium-containing waste slag at more than 15% will block the furnace, reducing the recovery eﬃciency. Therefore, the development of an eﬀective method to extract titanium from the titanium-containing slag is crucial and should be investigated.

Metals 2020, 10, 1611; doi:10.3390/met10121611 www.mdpi.com/journal/metals Metals 2020, 10, x FOR PEER REVIEW 2 of 16 Metals 2020, 10, 1611 2 of 15

After nearly 20 years of research, the Fray–Farthing–Chen (FFC) Cambridge process has been successfullyAfter nearly applied 20 yearsto the of preparation research, the of Fray–Farthing–Chentitanium, tantalum, tungsten, (FFC) Cambridge niobium, processand other has metals been successfullyand alloys [5 applied–7]. Several to the studies preparation on the synthesis of titanium, of Fe tantalum,–Ti alloy tungsten,have also niobium,been condu andcted other by the metals FFC andprocess. alloys Zhou [5–7 et]. al. Several [8] successfully studies on prepared the synthesis Fe–Ti ofalloys Fe–Ti using alloy ilmenite have also as a been precursor conducted at 973 byK and the FFC3.2–4.4 process. V. Tan Zhou et al. et [9] al. electrolyzed [8] successfully the mixtures prepared of Fe–Ti TiO alloys2 and Fe using2O3 to ilmenite produce as aferrotitanium precursor at alloys 973 K andcontaining 3.2–4.4 intermetallic V. Tan et al. [ Fe9]– electrolyzedTi phases, such the mixtures as FeTi and of TiO Fe22Ti,and by Fe the2O FFC3 to produce process. ferrotitaniumHowever, the alloysproduced containing alloys generally intermetallic contained Fe–Ti phases,porous suchstructures as FeTi with and numerous Fe2Ti, by thecarbon FFC process.impurities, However, and no theeffective produced method alloys has generally been proposed contained to poroussolve this structures problem. with Panigrahi numerous et carbonal. [10] impurities,prepared a and dense no ehighffectively purified method intermetallic has been proposed solid of to solveβ-Ti (FeTi this problem.4) and FeTi Panigrahi from mixed et al. TiO [10]2 prepared and FeTiO a dense3, at a highlymolar purifiedratio of intermetallic 5.44:1.00. Owing solid to of β the-Ti dense (FeTi4 ) structure and FeTi of from the mixed product, TiO 2 theand quantity FeTiO3 ,of at a carbon molar ratioparticle of 5.44:1.00.inclusions Owing was relatively to the dense low. structure To date, of the ferrotitanium product, the alloys quantity have of carbon primarily particle been inclusions prepared was by relativelyelectrolysis low. of FeTiO To date,3, or ferrotitanium by a mixture alloys of TiO have2 and primarily Fe2O3 [11 been–13]. preparedNo research by electrolysishas been reported of FeTiO on3, orthe by preparation a mixtureof ofTiO Ti–2Feand alloys Fe2 Ofrom3 [11 a– 13waste]. No slag research containing has been titanium, reported because on the waste preparation slag commonly of Ti–Fe alloyscomprises from Si, a wasteCa, and slag Al. containing However, titanium, it is desirable because to use waste the slag titanium commonly-bearing comprises slag as a Si, raw Ca, material and Al. However,to produce it is Fe desirable–Ti alloy to if use the the impurities titanium-bearing contained slag in as waste a raw slag material can be to produce removed Fe–Ti by molten alloy if saltthe impuritieselectrolysis. contained in waste slag can be removed by molten salt electrolysis. TheThe purposepurpose ofof thisthis studystudy isis toto produceproduce aa Fe–TiFe–Ti alloyalloy withwith aa densedense structurestructure toto suppresssuppress carboncarbon impuritiesimpurities byby the the electrochemical electrochemical reduction reduction of of titanium-containing titanium-containing waste waste slag, slag, which which was was realized realized via thevia FFCthe method.FFC method. We aim We to aim explore to explore the influence the influence of the atomic of the ratio atomic of Ti ratio and Feof inTi theand raw Fe materialin the raw on thematerial products on t ofhe electrolysis products of and electrolysis investigate and the investigate iron alloying the process. iron alloying The results process. from The this results study from may providethis study a theoretical may provide basis a andtheoretical general basis guideline and forgeneral the industrial guideline production for the industrial of ferrotitanium production alloy of fromferrotitanium waste titanium-bearing alloy from waste slag. titanium-bearing slag.

2.2. Materials and Methods

2.1.2.1. Materials and CharacterizationCharacterization MethodsMethods TheThe titanium-bearingtitanium-bearing wastewaste residueresidue (slag)(slag) usedused inin thisthis experimentexperiment waswas obtainedobtained fromfrom ZunyiZunyi TitaniumTitanium IndustryIndustry Co., Co., Ltd. Ltd. The The elemental elemental composition composition of of the the waste waste residue residue is listedis listed in in Table Table1. The 1. The Ti contentTi content is 50.15 is 50.15 wt%, wt and%, and the the main main impurities impurities are are Al, Al, Si, Si, and and O. O. The The X-ray X-ray diﬀ diffractionraction (XRD) (XRD) pattern pattern of theof the titanium-containing titanium-containing waste waste slag (Figure slag (Figure1) demonstrates 1) demonstrates that the that main the phases main present phases are present titanium are oxides:titanium FeTi oxides:2O5, TiFeTi3O52O, and5, Ti TiO3O5,2 and. To TiO produce2. To produce a mixed oxidea mixed cathode, oxide cathode, the titanium-containing the titanium-containing slag and Feslag2O and3 were Fe2 mixedO3 were in mixed a ratio in of a 1.2:1. ratio The of 1.2:1. mixture The wasmixture then was ball-milled then ball for-milled 10 h and for sintered10 h and at sintered 950 ◦C forat 950 2 h. °C for 2 h.

Table 1. Elemental composition of titanium-containing slag (wt%). Table 1. Elemental composition of titanium-containing slag (wt %).

ElementsElements TiTi OO AlAl SiSi CaCa Fe Fe Mn Mn Mg Mg ContentContent (wt%) (wt %) 50.15 50.15 40.65 40.65 1.22 1.22 2.012.01 0.650.65 3.61 3.61 1.43 1.43 0.280.28

1 1-FeTi2O5 76.1% 2-Ti3O5 3.1% 3-Rutile-TiO2 20.8% Rwp =10.2295

GoF = 1.2296

1 1 2 Intensity / a.u. Intensity 1 1 1 1 2 2 2 1 2 1 1 1 1 1 1 3 3 2 1 3 2 2 3 2 2 2 3 2 2 2

20 30 40 50 60 70 80 90 2θ/ (°)

FigureFigure 1.1. XRD patternspatterns ofof thethe titaniumtitanium slag.slag.

Metals 2020, 10, 1611 3 of 15

The microstructure and composition of the sintered and electrolyzed pellets were determined by scanning electron microscopy (SEM; EM-30PLUS, Coxem, Korea) coupled with energy-dispersive X-ray spectroscopy (EDS). The pellet phase composition was determined by XRD (X’Pert PRO MPD, PANalytical, The Netherlands). The particle size was determined using the Beckman Coulter LS 13 320 Particle Size Analyzer.

2.2. Experimental Method

The titanium-containing waste residue was mixed with Fe2O3 and ball-milled for 10 h, using 8 wt% liquid paraﬃn as a binder. The mixture was then compressed into pellets with a diameter and height of 10 and 4 mm, respectively, under a pressure of 2 MPa. The pellets were sintered in air at 950 ◦C for 2 h; the sintered pellet was then wrapped with nickel foam and attached to an iron–chromium–aluminum wire (diameter = 1.5 mm) to form a cathode. The electrolyte was anhydrous CaCl2 (>96 wt%) contained in a graphite crucible (inner diameter = 50 mm, height = 110 mm). The graphite crucible containing the molten salt was placed in a vertical cylindrical tube furnace for the electrolysis of the solid pellets. The electrolytic cell was heated to 120 ◦C under a high-purity Ar atmosphere to remove any physical water. Subsequently, the electrolytic cell was heated up to 550 ◦C to remove chemically bound water and then to 782 ◦C to form molten salt electrolyte, and a blank cathode (nickel foam without a pellet) attached with an iron–chromium–aluminum wire was inserted into the molten salt for the pre-electrolysis process. A constant voltage of 2 V was applied between the blank cathode and anode, which were positioned approximately 30 mm apart. The pre-electrolysis process was controlled using an electrochemical workstation (HCP-803, Biologic, Houston, TX, USA) interfaced with a computer, and the current–time curves were recorded. After the pre-electrolysis process, the voltage was adjusted to 3.1 V (between the oxide cathode and anode), and the electrolysis was terminated after 6 h. After the electrolysis, the cathode was removed from the molten salt and cooled to room temperature (20–25 ◦C) under an Ar atmosphere. The cathode pellet was washed with distilled water to remove the solidiﬁed salt, rinsed with 1 N aqueous hydrochloric acid (HCl) to expel the salt residues in the pores of the pellet, and dried in a vacuum dryer. A schematic illustration of the experimental setup is shown in Figure2.

Figure 2. Schematic illustration of the experimental setup. Metals 2020, 10, 1611 4 of 15

2.3. Experimental Principle

The Fe–Ti equilibrium phase diagram is shown in Figure3. When T > 882 ◦C, the crystal Metals 2020, 10, x FOR PEER REVIEW 4 of 16 structureMetals 2020 of, 10 titanium, x FOR PEER is a body-centeredREVIEW cubic β-titanium. In the temperature range from 595 to 10854 ◦ofC, 16 Ti:Fe > 1:1; FeTi and β-Ti (FeTi4) can coexist in equilibrium.

Figure 3. Fe–Ti equilibrium phase diagram. Figure 3. Fe–Ti equilibrium phase diagram. Figure 3. Fe–Ti equilibrium phase diagram. Based on the principle of the FFC method, Ti–Fe alloys can be prepared through the molten salt Based on the principle of the FFC method, Ti–Fe alloys can be prepared through the molten electrochemical deoxidation process. The electrolytic voltage required for the electrochemical salt electrochemicalBased on the principle deoxidation of the process. FFC method, The electrolytic Ti–Fe alloys voltage can be requiredprepared for through the electrochemical the molten salt electrochemicaldeoxidation process deoxidation is 3.1 V, which process. is lower The than electrolytic the theoretical voltage decomposition required forvoltage the of electrochemical CaCl2 (3.18 deoxidation process is 3.1 V, which is lower than the theoretical decomposition voltage of CaCl2 (3.18 V) V) at 900 °C and higher than the theoretical decomposition voltage of TiO2 and Fe2O3. A schematic deoxidation process is 3.1 V, which is lower than the theoretical decomposition voltage of CaCl2 (3.18 atillustration 900 ◦C and of higher the reaction than themechanism theoretical of the decomposition electrolysis process voltage in ofthis TiO study2 and is shown Fe2O3. in A Figure schematic 4. V) at 900 °C and higher than the theoretical decomposition voltage of TiO2 and Fe2O3. A schematic illustrationAccording of to the the reactionrelevant mechanismliterature reports of the [14 electrolysis–16], the intermediate process in products this study generated is shown during in Figure the 4. Accordingillustration to of the the relevant reaction literature mechanism reports of the [14 –electrolysis16], the intermediate process in productsthis study generated is shown duringin Figure the 4. electrolysis of TiO2/Fe2O3 mixtures commonly include Fe, CaTiO3, TiO2, TiO, and Fe2Ti. The entire According to the relevant literature reports [14–16], the intermediate products generated during the electrolysisreduction ofprocess TiO2 /isFe complicated,2O3 mixtures and commonly the valence include state of Fe, titanium CaTiO3 is, TiO gradually2, TiO, andreduced Fe2Ti. during The entirethe electrolysis of TiO2/Fe2O3 mixtures commonly include Fe, CaTiO3, TiO2, TiO, and Fe2Ti. The entire reductionelectrochemical process reduction is complicated, process. and the valence state of titanium is gradually reduced during the electrochemicalreduction process reduction is complicated, process. and the valence state of titanium is gradually reduced during the electrochemical reduction process.

Figure 4. Schematic illustration of the reaction process during electrolysis. Figure 4. Schematic illustration of the reaction process during electrolysis. In this study, initial theoretical analysis was performed using the thermodynamic database and In this study, initial theoretical analysis was performed using the thermodynamic database and thermodynamic software FactSage7.2. For a temperature range from 873 to 1273 K and a step change of thermodynamic softwareFigure 4. SchematicFactSage7.2. illustration For a tempe of therature reaction range process from during873 to 1273electrolysis. K and a step change of 20 K, the ΔGθ and the theoretical decomposition voltages (type 1) that describe the possible electrodeIn this reactions study, initial during theoretical the direct reduction analysis wasprocess performed can be calculated. using the These thermodynamic calculations database are given and thermodynamic software FactSage7.2. For a temperature range from 873 to 1273 K and a step change θ of 20 K, the ΔG and the theoretical decomposition voltages (type 1) that describe the possible electrode reactions during the direct reduction process can be calculated. These calculations are given

Metals 2020, 10, x FOR PEER REVIEW 5 of 16

in Table 2. Origin software was used to linearly fit the calculated values to obtain the relationship between ΔGθ and temperature, as shown in Figure 5.

Table 2. ΔGθ and the theoretical decomposition voltage (Eθ) of the possible reactions that occurred during the electrolysis process.

Metals 2020 10 ΔGθ (T = 1173 , , 1611 Electrode Reaction ΔGθ—T 5 ofEθ 15 K) –159.2–0.014T θ 2 TiO2 + 2 [CaO] = 2 CaTiO3 (s) –175.622 − 20 K, the ∆G and the theoretical decomposition voltages (type 1)(1) that describe the possible electrode reactions during the direct reduction process can be calculated. These calculations are given in Table2. 2 FeTiO3 (s) + 2 C (s) = 2 Fe (s) + 2 TiO2 (s )+ 2 CO (g) 597.14–0.64T (2) –153.58 − θ Origin software was used to linearly ﬁt the calculated values472.40 to obtain–0.510 theT relationship between ∆G Fe2O3 + 3 C = 2 Fe + 3 CO –125.677 − and temperature, as shown in Figure5. (3) 2 CaTiO3 (s) + 2 C (s) = 2 TiO (s) +2 [CaO] + 2 CO (g) 737.22–0.34T (4) 338.4 0.89 Table 2. ∆Gθ and the theoretical decomposition voltage (Eθ) of the possible reactions that occurred CaTiO3 (s) + 2 Fe (s) + 2 C (s) = Fe2Ti (s) + [CaO] + 2 during the electrolysis process. 691.48–0.32T (5) 316.12 0.82 CO (g) 2 TiO (s) Electrode+ 2 C (s) Reaction= 2 Ti (s) + 2 CO (g) 851.38∆Gθ—T–0.39T (6)∆ Gθ (T =393.911173 K) Eθ1.02

2 TiO2 + 2 [CaO] = 2 CaTiO3 (s) 159.2–0.014T531.74–0.364T (1) 175.622 - 4TiO2 + 2 C = 2Ti2O3 + 2 CO − − 104.768 0.27 2 FeTiO3 (s) + 2 C (s) = 2 Fe (s) + 2 TiO2 (s )+ 2 CO (g) 597.14–0.64T(7) (2) 153.58 - − Fe O + 3 C = 2 Fe + 3 CO 472.40–0.510T (3) 125.677 - 2 3 622.89–0.326T − 2 CaTiO (s)2+ Ti22 CO (s)3 += 22 C TiO = 4 (s) TiO+2 [CaO]+ 2 CO+ 2 CO (g) 737.22–0.34T (4) 338.4240.492 0.890.62 3 (8) CaTiO3 (s) + 2 Fe (s) + 2 C (s) = Fe2Ti (s) + [CaO] + 2 CO (g) 691.48–0.32T (5) 316.12 0.82 CaTiO3 (s)2TiO + Fe (s) (s)+ 2+ C2 (s)C = 2FeTi Ti (s) (s)+ 2+ CO[CaO] (g) + 2 CO (g) 851.38–0.39T744.60–0.33T (6) (9) 393.91357.51 1.020.93 4TiO2 + 2 C = 2Ti2O3 + 2 CO 531.74–0.364T751.98–0.32T (7) 104.768 0.27 2 TiO (s) + 2 Fe (s) + 2 C (s) = 2 FeTi (s) + 2 CO (g) 376.62 0.98 2 Ti2O3 + 2 C = 4 TiO + 2 CO 622.89–0.326T(10) (8) 240.492 0.62 CaTiO3 (s) + Fe (s) + 2 C = FeTi (s) + [CaO] + 2 CO (g) 744.60–0.33T (9) 357.51 0.93 2 TiO (s) + 2 Fe (s) + 2 C (s) = 2 FeTi (s) + 2 CO (g) 751.98–0.32T645.74– (10)0.31T 376.62 0.98 2 TiO (s) + 4 Fe (s) + 2 C (s) = 2 Fe2Ti (s) + 2 CO (g) 282.11 0.73 2 TiO (s) + 4 Fe (s) + 2 C (s) = 2 Fe2Ti (s) + 2 CO (g) 645.74–0.31T(11) (11) 282.11 0.73

CaTiO 2TiO(s)+2Fe(s)+2C(s)=2FeTi(s)+2CO(g)3(s)+Fe(s)+2C=FeTi(s)+[CaO]+2CO(g) 2TiO(s)+2C(s)=2Ti(s)+2CO(g) CaTiO 400 3(s)+ 2Fe(s)+ 2C(s)=Fe 2TiO(s)+4Fe(s)+2C(s)=2Fe 2Ti(s)+[CaO]+2CO(g)

2Ti(s)+2CO(g) 2Ti O 2 3(s)+2C(s)=4TiO(s)+2CO(g) 4TiO 2(s)+2C(s)=2Ti 2CaTiO 200 (s)+ 2C(s)=2TiO(s)+2[CaO]+ 2CO(g) O 3 2 3(s)+2CO(g)

/ KJ

ΔG

0 2/3Fe O 2 3+2C=4/3Fe+2CO

2FeTiO

3 (s)+2C(s)=2Fe(s)+2TiO 2TiO 2+2[CaO]=2CaTiO3(s) -200 2 (s)+2CO(g)

900 1000 1100 1200 1300 Temperature/ K

Figure 5. Relationship between ∆Gθ and the temperature of the possible reactions that occurred during Figure 5. Relationship between ΔGθ and the temperature of the possible reactions that occurred the electrolysis process. during the electrolysis process. From the calculations presented in Table2, we can determine the reactions that are likely to proceed during heating and the pre-electrolysis process based on the values of ∆Gθ in the given temperature θ range. For reaction 1, ∆G < 0, which implies that CaTiO3 can be generated as an intermediate product through the reaction. The presence of CaO in the sample results from impurities that exist in the form of Ca2+ and O2– ions in the molten salt. When a potential is applied between the cathode and anode, reactions 1, 2, and 3 are likely to occur. Consequently, FeTiO3 will decompose into Fe and CaTiO3, accompanied by the electrolysis of Fe2O3 to Fe (reaction 3). Simultaneously, the electrolytic reduction Metals 2020, 10, x FOR PEER REVIEW 6 of 16

From the calculations presented in Table 2, we can determine the reactions that are likely to proceed during heating and the pre-electrolysis process based on the values of ΔGθ in the given Metalstemperature2020, 10, 1611 range. For reaction 1, ΔGθ < 0, which implies that CaTiO3 can be generated as an6 of 15 intermediate product through the reaction. The presence of CaO in the sample results from impurities that exist in the form of Ca2+ and O2– ions in the molten salt. When a potential is applied between the of thecathode intermediate and anode, products reactions (CaTiO 1, 2,3 andand 3 TiO are2 ,likely as shown to occur. in reactions Consequently, 4–9) will FeTiO occur.3 will The decompose Fe generated willinto react Fe with and titaniumCaTiO3, accompanied to form Fe–Ti byalloys, the electrolysis as shown of inFe reactions2O3 to Fe (reaction 10 and 11, 3). whichSimultaneously, include the the FeTi andelectrolytic Fe2Ti alloys. reduction It should of the be intermediate noted that the products alloy phases,(CaTiO3 suchand TiO as FeTi2, as shown and Fe in2Ti, reactions will be 4 a–9)ﬀected will by the contentoccur. The of theFe generated enriched will Fe around react with CaTiO titanium3 [17 ].to form Fe–Ti alloys, as shown in reactions 10 and 11,During which theinclude electrolysis the FeTi and process, Fe2Ti oxygenalloys. It ionsshould will be migratenoted that to the the alloy anode phases, to produce such as FeTi CO and or CO 2. TheFe electrode2Ti, will be potentials affected by of the content impurity of the elements enriched are Fe listed around in CaTiO Table33 .[17]. In our previous research [ 18], we provedDuring that impurities,the electrolysis such process, as the Ca,oxygen Al, andions Siwill compounds, migrate to the comprising anode to theproduce titanium-containing CO or CO2. The electrode potentials of the impurity elements are listed in Table 3. In our previous research [18], waste residue were reduced to their corresponding elements, and most of the impurities were removed we proved that impurities, such as the Ca, Al, and Si compounds, comprising the titanium-containing as molten salts. Some of the remaining impurities were observed by washing the molten salt with waste residue were reduced to their corresponding elements, and most of the impurities were hydrochloricremoved as acid. molten The salts. corresponding Some of the remaining chloride was impurities dissolved were in observed HCl washing by washing solution, the molten but some impuritiessalt with may hydrochloric remain in acid. the productsThe corresponding after this chloride process, was owing dissolved to the in blockage HCl washing of the solution, pores. but some impurities may remain in the products after this process, owing to the blockage of the pores. Table 3. Electrode potentials of impurity elements. Table 3. Electrode potentials of impurity elements. Elements Electrode Reaction Electrode Potential/V CaElements ElectrodeCa2+ + R2eeactionCa Electrode Potential/V2.866 − → − Si Ca Ca2+Si 4++ 2е+– 4e→ Ca Si –2.866 0.843 − → − Al Si SiA14+ +3+ 4e+– 3e→ Si Al –0.843 1.662 − → − Al A13+ + 3e– → Al –1.662 3. Results and Discussion 3. Results and Discussion 3.1. Mixed Oxide Cathode 3.1. Mixed Oxide Cathode The XRD patterns in Figure6a,b illustrate the main phases present in the sample after high energy The XRD patterns in Figure 6a,b illustrate the main phases present in the sample after high ball milling and high temperature heat treatment, respectively. Fe2O3, TiO2, FeTi2O5 and Fe2TiO5 energy ball milling and high temperature heat treatment, respectively. Fe2O3, TiO2, FeTi2O5 and were identiﬁed in XRD patterns, and the particle size was reduced from 48.11 to 4.55 µm after milling, Fe2TiO5 were identified in XRD patterns, and the particle size was reduced from 48.11 to 4.55 μm after as listedmilling, in Tableas listed4 and in Table Figure 4 7and. Figure 7.

2 1-Fe O 35.5% (a) 1-Fe2O3 50.6% (b) 2 3 2 2-TiO 11.0% 2-TiO2 51.8% 3 2 1 1 3-FeTi O 38.4% 3-Fe TiO 12.7% 2 2 5 2 5 1 Rwp=3.727 3 Rwp = 3.9994 1 2 GoF=0.9764 GoF =1.0079 1 1 1 1 2 2 1 1 1 2 3 1 Intensity / a.u. Intensity 2 1 3 2 / a.u. Intensity 1 2 3 3 32 31 1 2 1 3 1 3 2 1 3 1 3 2 2 23 1

20 40 60 80 20 40 60 80 Metals 2020, 10, x FOR PEER 2REVIEWθ / (°) 2θ / (°) 7 of 16

FigureFigure 6. 6.(a )(a XRD) XRD pattern pattern before before highhigh energy ball ball milling; milling; (b ()b XRD) XRD pattern pattern after after high high energy energy ball ball milling; SEM images of the titanium-containing waste residue and Fe2O3 (c) before heat treatment and milling; SEM images of the titanium-containing waste residue and Fe2O3 (c) before heat treatment and (d) after heat treatment. (d) after heat treatment.

3.5 Ball milling time: 0 h Ball milling time: 2 h 6 Volume: 100% Volume: 100% 3.0 Mean: 140.6 μm Mean: 21.40 μm Median: 48.11 μm 5 Median: 19.57 μm 2.5 Mean/median ratio: 2.922 Mean/median ratio: 1.094 Mode: 31.51 μm 4 Mode: 21.70 μm 2.0

3 1.5

Volume(%)

Volume(%) 2 1.0

0.5 1

0.0 0 0 1 3 7 20 55 148 403 1097 1 3 7 20 55 148 403 1097 Particle Diameter (μm) Particle Diameter (μm)

5 Ball milling time: 6 h Ball milling time: 8 h 5 Volume: 100% Volume: 100% Mean: 12.16 μm 4 Mean: 9.615 μm Median: 6.729 μm 4 Median: 10.78 μm Mean/median ratio: 1.129 Mean/median ratio: 1.429 Mode: 19.76 μm 3 Mode: 13.61 μm 3

2 Volume (%)

Volume (%)

1 1

0 0 1 3 7 20 55 148 403 1097 1 3 7 20 55 148 403 1097 Particle Diameter (μm) Particle Diameter (μm)

Metals 2020, 10, 1611 7 of 15

Table 4. Particle size of the products after being milled for diﬀerent durations.

Time Raw Ore 2 h 6 h 8 h 10 h Granularity (Median) 48.11 µm 19.57 µm 10.78 µm 6.73 µm 4.55 µm

3 1.5

Volume(%)

Volume(%) 2 1.0

0.5 1

0.0 0 0 1 3 7 20 55 148 403 1097 1 3 7 20 55 148 403 1097 Particle Diameter (μm) Particle Diameter (μm)

2 Volume (%)

Volume (%)

1 1

0 0 1 3 7 20 55 148 403 1097 1 3 7 20 55 148 403 1097 Particle Diameter (μm) Particle Diameter (μm)

6 Ball milling time: 10 h Volume: 100% Mean: 6.544 μm 5 Median: 4.56 μm Mean/median ratio: 1.437 4 Mode: 13.61 μm

Volume (%) 2

0 1 3 7 20 55 148 403 1097 Particle Diameter (μm)

Figure 7. Particle size distribution of the products after being milled for diﬀerent durations. Figure 7. Particle size distribution of the products after being milled for different durations. The SEM images in Figure6c,d illustrate the e ﬀect of the high-temperature heat treatment on the cathode material. After the sintering of the cathode material at high temperature, the porosity increased and the particles were interconnected with each other. The electrical conductivities of the sintered 7 6 4 2 cathode pellets increased from 10− to 10− S/cm before sintering and up to 10− to 10− S/cm afterward.

Metals 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/metals Metals 2020, 10, x FOR PEER REVIEW 8 of 16

6 Ball m illing tim e: 10 h Volume: 100% Mean: 6.544 祄 5 Median: 4.56 祄 Mean/median ratio: 1.437 4 Mode: 13.61 祄

3 Volume (%) 2

0 1 3 7 20 55 148 403 1097 P article D iam eter (祄 )

Figure 7. Particle size distribution of the products after being milled for different durations.

Table 4. Particle size of the products after being milled for different durations.

Time Raw Ore 2 h 6 h 8 h 10 h Granularity (Median) 48.11 μm 19.57 μm 10.78 μm 6.73 μm 4.55 μm

The SEM images in Figure 6c,d illustrate the effect of the high-temperature heat treatment on the cathode material. After the sintering of the cathode material at high temperature, the porosity increased and the particles were interconnected with each other. The electrical conductivities of the sintered cathode pellets increased from 10–7 to 10–6 S/cm before sintering and up to 10–4 to 10–2 S/cm

Metalsafterward.2020, 10 , 1611 8 of 15

3.2. Effect of Raw Material Ratio on the Morphology of the Product and Its Elemental Composition 3.2. Eﬀect of Raw Material Ratio on the Morphology of the Product and Its Elemental Composition The microstructures of the products obtained using different ratios of the raw material are shownThe in microstructures Figure 8a–e. As of the the propor productstion obtainedof Ti in the using raw di materialsﬀerent ratios was ofincreased, the raw materialthe particles are shown of the inproduct Figure appeared8a–e. As the more proportion densely ofpacked, Ti in the which raw is materials consistent was with increased, the apparent the particles increase of thein solubility product appearedfrom the moremacroscopic densely packed,appearance which in isFigure consistent 8f. Table with the5 contains apparent a increase summary in solubility of the elemental from the macroscopiccompositions appearanceof the products in Figure for which8f. Table the atomic5 contains ratio a of summary Ti and Fe of in the the elementalproduct is compositionssimilar. As the ofcontent the products of Ti in the for electrolytic which the atomicproduct ratio increases, of Ti and the quantity Fe in the of product the β-Ti(FeTi is similar.4) phase As theincreases, content and of Tithe in β the-Ti phase electrolytic takes productthe form increases, of a metastable the quantity alloy Fe of1– thexTixβ (x-Ti(FeTi < 0.8) 4[19],) phase which increases, is a relatively and the denseβ-Ti phasealloy. This takes could the form be explained of a metastable by the increased alloy Fe1 densixTix (tyx < of0.8) the [cathode.19], which At is an a atomic relatively ratio dense of Ti:Fe alloy. = − This1.2:1, could the micromorphology be explained by the of increased the product density is ofmore the cathode.uniform Atand an demonstrated atomic ratio ofa Ti:Fesponge-like= 1.2:1, thestructure. micromorphology of the product is more uniform and demonstrated a sponge-like structure.

Metals 2020, 10, x FOR PEER REVIEW 9 of 16

Figure 8. MicrostructureMicrostructure of of the the product product of of the electrolys electrolysisis of the raw materials with different diﬀerent Ti:Fe element ratios: ( (aa)) 1:2; 1:2; ( (b)) 1:1; 1:1; ( (cc)) 1.2:1; 1.2:1; ( (d)) 1.5:1; 1.5:1; ( (e)) 2:1; 2:1; ( (ff)) macroscopic macroscopic appearances appearances of the product of electrolysis corresponding to (a–e), respectively.respectively.

The electrolytic product was analyzed by EDS for three times and the average values are shown in Table 5. With an increase in Ti, the C content in the product decreased, but impurities such as Si and Ca increased. A higher iron content, e.g., Ti:Fe = 1:2, will increase the electrical conductivity of the cathode and speed up the electrolytic process. Owing to the lower decomposition voltage of Fe, it was preferentially reduced compared to Ti, making the cathode material porous (Figure 8f). As the proportion of Fe atoms decreased, the content of FeTi4 in the product increased, producing a dense cathode structure and limiting the infiltration of carbon powder; this explains the gradual decrease in the carbon content. Based on our previous research [20], the lower porosity limits the transfer of electrons and oxygen ions, and this affects the electrolysis process. This is also the primary reason for the gradual increase in impurities (O, Si, and Ca). The microstructure of the product is sensitive to the ratio of Ti:Fe. Based on these results, the optimum raw material composition for the production of Fe–Ti alloys is Ti:Fe = 1.2:1.

Metals 2020, 10, 1611 9 of 15

Table 5. Elemental contents (wt%) of the electrolytic products produced from raw materials with diﬀerent Ti and Fe atomic ratios.

Ti:Fe O Ti Fe C Si Ca 1:2 1.95 24.53 65.3 7.65 0.57 - 1:1 2.08 45.63 45.82 5.04 1.43 - 1.2:1 0.69 49.34 45.43 2.67 1.87 - 1.5:1 2.05 44.62 47.69 2.02 2.1 1.52 2:1 6.49 57.4 29.92 2.06 2.04 2.09

The electrolytic product was analyzed by EDS for three times and the average values are shown in Table5. With an increase in Ti, the C content in the product decreased, but impurities such as Si and Ca increased. A higher iron content, e.g., Ti:Fe = 1:2, will increase the electrical conductivity of the cathode and speed up the electrolytic process. Owing to the lower decomposition voltage of Fe, it was preferentially reduced compared to Ti, making the cathode material porous (Figure8f). As the proportion of Fe atoms decreased, the content of FeTi4 in the product increased, producing a dense cathode structure and limiting the inﬁltration of carbon powder; this explains the gradual decrease in the carbon content. Based on our previous research [20], the lower porosity limits the transfer of electrons and oxygen ions, and this aﬀects the electrolysis process. This is also the primary reason for the gradual increase in impurities (O, Si, and Ca). The microstructure of the product is sensitive to the ratio of Ti:Fe. Based on these results, the optimum raw material composition for the production of Fe–Ti alloys is Ti:Fe = 1.2:1.

3.3. Effect of Raw Material Ratio on the Electrolysis Process The XRD patterns of the products obtained from the raw materials with different elemental ratios are shown in Figure9. FeTi, FeTi 4, TiO, and Fe2Ti phases were identified in the XRD patterns. When the atomic ratio is Ti:Fe = 1:2, the primary phase of the product is Fe2Ti. As the Fe atomic ratio decreased, the Fe2Ti diffraction peak in the product gradually decreased. When Fe:Ti (atomic ratio) is 1, the primary phases in the products are FeTi and FeTi , and as the proportion of Ti atoms ≤ 4 increased, the primary diffraction peak of FeTi4 increased, and an intermediate product (TiO) was observed. Based on the thermodynamic relationship shown in Figure5, Fe–Ti can be produced via two routes. One route is the direct formation of Fe2Ti and FeTi through the reaction of CaTiO3 with Fe in a single step, as expressed by reactions 5 and 9 in Table2. The other route is the stepwise deoxidation of CaTiO3 to produce TiO and Ti, which can further react with iron to form Fe–Ti alloys. According to a previous study [21], CaTiO3 can be directly reduced to form Fe2Ti or FeTi in the Fe-rich region. When the Fe content is low, Fe–Ti is primarily generated via the second route. The phases in the products produced from the raw materials with different ratios of Ti and Fe are summarized in Table6, which is consistent with the observations in a previous study [22].

Table 6. Atomic ratio of Ti and Fe and the primary phase in the electrolytic products.

Atomic Ratio Ti Fe Main Phase Rwp GoF (Ti:Fe)

1:2 27.11 61.50 Fe2Ti: 73.2%; FeTi: 26.8 3.8886 0.9955 1:1 41.39 35.88 Fe2Ti: 39.2%; FeTi: 58.8%; TiO: 2.0% 5.9063 1.1227 1.2:1 46.81 36.89 FeTi4: 16.0%; FeTi: 84% 4.1006 1.0048 1.5:1 40.02 36.66 FeTi4: 18.7%; FeTi: 44.3%; TiO: 37% 7.4289 1.1794 2:1 40.72 18.01 FeTi4: 76.1%; FeTi: 12%; TiO: 11.9% 5.7525 1.0404 Metals 2020, 10, x FOR PEER REVIEW 10 of 16

Table 5. Elemental contents (wt %) of the electrolytic products produced from raw materials with different Ti and Fe atomic ratios.

Ti:Fe O Ti Fe C Si Ca 1:2 1.95 24.53 65.3 7.65 0.57 − 1:1 2.08 45.63 45.82 5.04 1.43 − 1.2:1 0.69 49.34 45.43 2.67 1.87 − 1.5:1 2.05 44.62 47.69 2.02 2.1 1.52 2:1 6.49 57.4 29.92 2.06 2.04 2.09

3.3. Effect of Raw Material Ratio on the Electrolysis Process The XRD patterns of the products obtained from the raw materials with different elemental ratios are shown in Figure 9. FeTi, FeTi4, TiO, and Fe2Ti phases were identified in the XRD patterns. When the atomic ratio is Ti:Fe = 1:2, the primary phase of the product is Fe2Ti. As the Fe atomic ratio decreased, the Fe2Ti diffraction peak in the product gradually decreased. When Fe:Ti (atomic ratio) is ≦ 1, the primary phases in the products are FeTi and FeTi4, and as the proportion of Ti atoms increased, the primary diffraction peak of FeTi4 increased, and an intermediate product (TiO) was observed. Based on the thermodynamic relationship shown in Figure 5, Fe–Ti can be produced via two routes. One route is the direct formation of Fe2Ti and FeTi through the reaction of CaTiO3 with Fe in a single step, as expressed by reactions 5 and 9 in Table 2. The other route is the stepwise deoxidation of CaTiO3 to produce TiO and Ti, which can further react with iron to form Fe–Ti alloys. According to a previous study [21], CaTiO3 can be directly reduced to form Fe2Ti or FeTi in the Fe- rich region. When the Fe content is low, Fe–Ti is primarily generated via the second route. The phases Metalsin the2020 products, 10, 1611 produced from the raw materials with different ratios of Ti and Fe are summa10rized of 15 in Table 6, which is consistent with the observations in a previous study [22].

Ti:Fe=1:2 1 2 1-FeTi 2-Fe2Ti 3-FeTi4 4-TiO 2 2 2 2 2 1 2 2 2 1 Ti:Fe=1:1 1 1 4 2 2 1

Ti:Fe=1.2:1 1

3 1 1

Ti:Fe=1.5:1 1 4 4 1 1 3 3 4 3

intensity / a.u. intensity Ti:Fe=2:1 1 3 1 4 4 4 3 1 3

PDF#96-152-2976-FeTi4

20 30 40 50 60 70 80 90 2θ / (°)

FigureFigure 9. XRDXRD patterns patterns of ofthe the product product obtained obtained from from the raw the materials raw materials with different with diﬀ Tierent and TiFe andatomic Fe atomicratios. ratios. 3.4. Eﬀect of the Duration of the Electrolysis Process on the Physical Properties of the Alloys

Considering the optimum composition of the raw materials with a Ti:Fe ratio of 1.2:1, the effects of the duration of the electrolysis process on the physical properties and electrochemical behavior of the alloys were investigated. The SEM images and XRD diffraction patterns of the products obtained after different durations of electrolysis are shown in Figure 10. In the early stage of electrolysis (0–1 h), owing to the deoxidation of Fe2O3, the CaO content in the electrolyte increased rapidly (the maximum solubility of CaO in CaCl2 molten salt is 19.4% [23]); this could promote reaction 1. The cathode comprised a large amount of CaTiO3 produced from TiO2 and CaO, resulting in particle fragmentation and blockage of the pores by the broken particles; this led to the formation of a dense cathode structure. 2 During the electro-deoxidation process, O − was electrolyzed from the solid particles and diffused into the electrolyte through these pores; the electrolyte also penetrated these pores. Therefore, the porosity is also important for the deoxidation of CaTiO3. As the electrolysis progressed (1–3 h), TiO2 was reduced to Ti or low-valence oxides; the intermediate product CaTiO3 gradually transformed to Fe2Ti and FeTi, which led to a decrease in the particle size and an increase in the bulk pores. This was primarily because of the formation of Fe–Ti alloy solid solution. Consequently, vacancies were formed in the bulk of the original CaTiO3 particles, thereby increasing the porosity of the pellet. Finally, with the formation of FeTi and FeTi4, the structure of the product became dense and sponge-like. These observations are consistent with the results obtained from the thermodynamic analysis. The electrochemical behavior of the cathode pellet in molten CaCl2 at 900 ◦C was evaluated by cyclic voltammetry measurements at a scan rate of 20 mV/s. Three distinct reduction peaks at C1 ( 0.82 V), C (0.26 V), and C (1.6 V) were observed (Figure 11, curve A). The XRD analysis (Figure9) − 2 3 demonstrated that the components of the cathode pellets after molten salt immersion are primarily CaTiO3, TiO2, and Fe2O3. The peak potentials of C1 and C2 correspond to the theoretical decomposition voltages of reactions 5 (Eθ = 0.82 V) and 7 (Eθ = 0.27 V) respectively; hence, they could be attributed to the reduction of CaTiO3 and TiO2 to generate Fe2Ti and Ti2O3, respectively. This observation indicates that the two alloying routes occurred simultaneously. According to thermodynamic calculations, Metals 2020, 10, x FOR PEER REVIEW 11 of 16

Table 6. Atomic ratio of Ti and Fe and the primary phase in the electrolytic products.

Atomic Ratio (Ti:Fe) Ti Fe Main Phase Rwp GoF 1:2 27.11 61.50 Fe2Ti: 73.2%; FeTi: 26.8 3.8886 0.9955 1:1 41.39 35.88 Fe2Ti: 39.2%; FeTi: 58.8%; TiO: 2.0% 5.9063 1.1227 1.2:1 46.81 36.89 FeTi4: 16.0%; FeTi: 84% 4.1006 1.0048 1.5:1 40.02 36.66 FeTi4: 18.7%; FeTi: 44.3%; TiO: 37% 7.4289 1.1794 2:1 40.72 18.01 FeTi4: 76.1%; FeTi: 12%; TiO: 11.9% 5.7525 1.0404

3.4. Effect of the Duration of the Electrolysis Process on the Physical Properties of the Alloys Considering the optimum composition of the raw materials with a Ti:Fe ratio of 1.2:1, the effects of the duration of the electrolysis process on the physical properties and electrochemical behavior of the alloys were investigated. The SEM images and XRD diffraction patterns of the products obtained after different durations of electrolysis are shown in Figure 10. In the early stage of electrolysis (0–1 h), owing to the deoxidation of Fe2O3, the CaO content in the electrolyte increased rapidly (the maximum solubility of CaO in CaCl2 molten salt is 19.4% [23]); this could promote reaction 1. The Metals 2020cathode, 10, 1611comprised a large amount of CaTiO3 produced from TiO2 and CaO, resulting in particle11 of 15 fragmentation and blockage of the pores by the broken particles; this led to the formation of a dense cathode structure. During the electro-deoxidation process, O2– was electrolyzed from the solid θ the E particlesof the Feand2O diffused3 reduction into the to produceelectrolyte Fe through is negative, these pores; which the indicates electrolyte that also this penetrated reaction these proceeds spontaneouslypores. Therefore, at a temperature the porosity of is 1173also K.important When afor potential the deoxidation is applied, of CaTiO Fe O3. (EAs =the 0.8486electrolysis V) can be 2 3 − electrochemicallyprogressed (1–3 reduced h), TiO2to was Fe reduced in a short to Ti time; or low-valence therefore, ox noides; reduction the intermediate peak was product detected. CaTiO Peak3 C3 gradually transformed to Fe2Ti and FeTi, which led to a decrease in the particle size and an increase could be attributed to the cathodic formation of Ca in molten CaCl2 [24,25]. As the electrolysis progressed,in the bulk titanium pores. oxides This was inthe primarily sample because surface of were the formation reduced toof TiFe–Ti or low-valencealloy solid solution. oxides with Consequently, vacancies were formed in the bulk of the original CaTiO3 particles, thereby increasing a higher decomposition voltage; these oxides were easily oxidized, which caused a decrease in the the porosity of the pellet. Finally, with the formation of FeTi and FeTi4, the structure of the product oxidationbecame peak dense potential. and sponge-like. In this study,These observations the potential ar ofe consistent the A1 peak with decreasedthe results obtained to –0.21 from V (Figure the 11, curvethermodynamic B). analysis.

Metals 2020, 10, x FOR PEER REVIEW 12 of 16

3 (e) 1 1-CaTiO 2-Fe 3-FeTi 4-Fe O 4 3 2 3 5-TiO 6-FeTi 7-SiO 8-Fe Ti 7 2 4 2 2 9-Ti O 2 3 3 3 6 6 6 6 h 1 1 1 1 1 1 8 2 4 9 8 2 7 5 81 8 8 9 5 3 h

1 h

intensity / a.u. intensity 7 30 min After soaking in molten sale 4 4 1 1 4 7 4 1 4 7 4 5 1 20 40 60 80 2θ / (°)

FigureFigure 10. SEM 10. SEM images images of theof the product product electrolyzed electrolyzed for: ( (aa) )30 30 min; min; (b () b1) h; 1 h;(c) (3c )h; 3 ( h;d) 6 (d h;) 6(e h;) XRD (e) XRD diffractogram of the product electrolyzed for 30 min, 1, 3, and 6 h. diﬀractogram of the product electrolyzed for 30 min, 1, 3, and 6 h.

The electrochemical behavior of the cathode pellet in molten CaCl2 at 900 °C was evaluated by cyclic voltammetry measurements at a scan rate of 20 mV/s. Three distinct reduction peaks at C1 (– 0.82 V), C2 (0.26 V), and C3 (1.6 V) were observed (Figure 11, curve A). The XRD analysis (Figure 9) demonstrated that the components of the cathode pellets after molten salt immersion are primarily CaTiO3, TiO2, and Fe2O3. The peak potentials of C1 and C2 correspond to the theoretical decomposition voltages of reactions 5 (Eθ = 0.82 V) and 7 (Eθ = 0.27 V) respectively; hence, they could be attributed to the reduction of CaTiO3 and TiO2 to generate Fe2Ti and Ti2O3, respectively. This observation indicates that the two alloying routes occurred simultaneously. According to thermodynamic calculations, the Eθ of the Fe2O3 reduction to produce Fe is negative, which indicates that this reaction proceeds spontaneously at a temperature of 1173 K. When a potential is applied, Fe2O3 (E = −0.8486 V) can be electrochemically reduced to Fe in a short time; therefore, no reduction peak was detected. Peak C3 could be attributed to the cathodic formation of Ca in molten CaCl2 [24,25]. As the electrolysis progressed, titanium oxides in the sample surface were reduced to Ti or low-valence oxides with a higher decomposition voltage; these oxides were easily oxidized, which caused a decrease in the oxidation peak potential. In this study, the potential of the A1 peak decreased to –0.21 V (Figure 11, curve B).

Metals 2020, 10, 1611 12 of 15 Metals 2020, 10, x FOR PEER REVIEW 13 of 16 Metals 2020, 10, x FOR PEER REVIEW 13 of 16

1000 A1 1000 A1 A A2 500 A A2 500 B B 0 0 C C2 -500 2

I/ mA -500

I/ mA -1000 -1000 C3 C1 C3 C -1500 1 -1500

-2000 -2000 -1.6 -1.2 -0.8 -0.4 0.0 -1.6 -1.2 -0.8 -0.4 0.0 Ewe/ V Ewe/ V FigureFigure 11. 11.Cyclic Cyclic voltammogram voltammogram ofof thethe cathodecathode pellet in molten molten CaCl CaCl2 at:at: (A (A) 900) 900 °C;C; (B) ( Bafter) after testing testing Figure 11. Cyclic voltammogram of the cathode pellet in molten CaCl2 2at: (A) 900 °C◦; (B) after testing for some time. Scan rate: 20 mVs–11. forfor some some time. time. Scan Scan rate: rate: 2020 mVsmVs−–1. . 3.5.3.5. E ffEffectect of of the the Ratio Ratio of of Raw Raw Material Material onon thethe ElectrolysisElectrolysis C Currenturrent 3.5. Effect of the Ratio of Raw Material on the Electrolysis Current FromFrom the the current–time current–time curvescurves recordedrecorded during the the electrolysis electrolysis process, process, as asshown shown in inFigure Figure 12, 12, From the current–time curves recorded during the electrolysis process, as shown in Figure 12, therethere are are three three distinct distinct stages. At At the the initial initial stage stage of electrolysis, of electrolysis, the cathode the cathode was in was close in contact close contactwith there are three distinct stages. At the initial stage of electrolysis, the cathode was in close contact with withnickel nickel foam. foam. Electroreduction Electroreduction reactions reactions primarily primarily occurred occurred on the on the surface surface of the of the cathode cathode pellet; pellet; nickel foam. Electroreduction reactions primarily occurred on the surface of the cathode pellet; therefore, the initial current was high (1.57 A). Within the next 100 min, a significant drop in current therefore,therefore, the the initial initial current current was was highhigh (1.57(1.57 A). Within the the next next 100 100 min, min, a asignificant significant drop drop in incurrent current was observed, and oxygen on the surface of the cathode were gradually expelled. The O2– 2–diffusion waswas observed, observed, and and oxygen oxygen on on the the surface surface ofof thethe cathodecathode were gradually gradually expelled. expelled. The The O2 O– diffusiondiffusion distance increased, while the oxygen element in the low-valence titanium oxide was expelled from distancedistance increased, increased, while while the theoxygen oxygen elementelement in the low low-valence-valence titanium titanium oxide oxide was was expelled expelled from from the lattice of the metal titanium. A high driving force was required to expel oxygen from the cathode, thethe lattice lattice of of the the metal metal titanium. titanium. AA highhigh driving force force was was required required to to expel expel oxygen oxygen from from the the cathode, cathode, and this led to a decrease in the current [26]. In the case where a small amount of oxygen was to be andand this this led led to to a a decrease decrease inin the current [26]. [26]. In In the the case case where where a small a small amount amount of oxygen of oxygen was to was be to ionized, the semiconductor cathode became conductive. The surface of the cathode became fully beionized, ionized, the the semiconductor semiconductor cathode cathode became became conductive. conductive. The The surface surface of of the the cathode cathode became became fully fully metalized when the current reached the lowest point. metalizedmetalized when when the the current current reached reached thethe lowestlowest point.

1.6 Ti:Fe=1:2 1.6 Ti:Fe=1:2 Ti:Fe=1:1 Ti:Fe=1:1 1.4 Ti:Fe=1.2:1 1.4 Ti:Fe=1.2:1 Ti:Fe=1.5:1 1.2 stage Ⅰ Ti:Fe=1.5:1 1.2 stage Ⅰ Ti:Fe=2:1 Ti:Fe=2:1 1.0 ¬ 1.0 ¬ stage Ⅱ

stage Ⅱ stageⅢ

¬ stageⅢ 0.8 ¬ 0.8

Current/ A

0.6 ¬ 0.6 ¬ 0.4 0.4 0.2 0.2 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350 Time/ min Time/ min Figure 12. Current–time curves display the three stages of electrolysis of the raw materials with FigureFigure 12. 12.Current–time Current–time curves curves display the three three stages stages of of electrolysis electrolysis of of the the raw raw materials materials with with different Ti:Fe element ratios. diﬀdifferenterent Ti:Fe Ti:Fe element element ratios. ratios.

Metals 2020, 10, 1611 13 of 15

In the second stage, from 100 min to 5 h, the current increased, indicating that the electrolysis extended from the surface to the interior of the cathode, and the interior of the cathode was gradually metalized. The current plateaued when the electrolytic process was almost complete. Based on the results from a previous study [27], the current–time curve during electroreduction of metal oxide in molten CaCl2 is dependent on the solubility of the oxygen in the metal. In the presence of dissolved oxygen, a peak was observed in the current profile; it gradually decreased to a stable current value, which was attributed to the background current. Current efficiency is one of the technical indicators of an electrolytic process. The formula for calculating the current efficiency is as follows:

Q η = theoretical 100% (1) Qactual ∗

Z F mFeTi Qtheoretical = · · (2) MFeTi Q = I t (3) actual ∗ where η is the current eﬃciency; Qtheoretical is the theoretically required quantity of electricity, the unit is C; and Qactual is the actual electricity consumed, the unit is C. Z is the stoichiometric number of electrons required for the electrolysis reaction; F is Faraday constant, 96485 C/mol; mFeTi is the mass of the actual product of FeTi, the unit is g; MFeTi is the relative atomic mass of FeTi, 104 g/mol; I is the current used during the electrolytic process, the unit is A; and t is the duration of electrolysis, the unit is s. Table7 lists the calculated values for current e ﬃciency.

Table 7. Current eﬃciency of the electrolysis process with the raw material containing diﬀerent Ti:Fe atomic ratios.

Atomic Ratio (Ti:Fe) Current Eﬃciency 1:2 15.3243% 1:1 17.2082% 1.2:1 20.2135% 1.5:1 18.5583% 2:1 13.6615%

From the current–time curve and the stepwise reduction process of CaTiO3, a slow reduction rate of CaTiO3 was observed, which became a limiting step during the entire electrolytic process. From the thermodynamic analysis, the reduction of CaTiO3 was diﬃcult. Simultaneously, because the crystal structure of CaTiO3 is a body-centered cubic structure, this structure is stable, and more energy is required during the reduction process, thereby reducing the current eﬃciency.

4. Conclusions The following conclusions were drawn from this study:

1. High-purity FeTi and FeTi4 alloys were successfully prepared from the mixed titanium-containing waste slag and Fe2O3 by electrolysis at 900 ◦C and 3.1 V for 6 h in molten calcium chloride. As the atomic ratio of Ti increased, the Fe2Ti phase decreased while that of FeTi4 increased. The morphology of the product gradually became dense, and impurities such as carbon particles were absent in the dense structure of the FeTi and β-Ti alloys. A lower porosity of the pellet limited the transfer of electrons and oxygen ions, which would further limit the removal of oxygen ions and other impurity elements from the cathode pellet. 2. The alloying process of ferrotitanium can be divided into two routes, depending on the proportion of the iron atoms in the raw materials. When the atomic ratio of Ti and Fe is 1.2:1, both alloying Metals 2020, 10, 1611 14 of 15

routes coexist simultaneously. At this atomic ratio, the microscopic morphology of the cathode is uniform and it displayed a sponge-like structure. 3. The current–time curve of the electrolytic process can be divided into three main stages. As the proportion of the iron atoms decreased, the time required for the start of the second stage and the current eﬃciency gradually decreased.

Author Contributions: Conceptualization, C.-y.C. and J.-q.L.; methodology, L.-z.W.; software, Y.-p.L.; validation, C.-y.C., J.-q.L. and Y.-p.L.; formal analysis, B.W.; investigation, S.-y.W.; resources, C.-y.C.; data curation, B.W.; writing—original draft preparation, B.W.; writing—review and editing, B.W.; visualization, B.W.; supervision, C.-y.C.; project administration, J.-q.L.; funding acquisition, C.-y.C. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by “National Natural Science Foundation of China, grant number 51664005, 51774102, 52074096 and 51804088” and “Talent Team Giant [2017] 5626” and “Platform Talent KY (2015) 334” supported by the Talents of Guizhou Science and Technology Cooperation Platform.” Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Kuenen, J. Ferro-Alloy Production. In EMEP/EEA Emission Inventory Guidebook; EEA: Copenhagen, Denmark, 2009; pp. 3–7. 2. Smirnov, L.A.; Grabeklis, A.A.; Demin, B.L. Processing slag from ferroalloy production. Steel Transl. 2009, 39, 93–96. [CrossRef] 3. Tripathy, P.K.; Gauthier, M.; Fray, D.J. Electrochemical Deoxidation of Titanium Foam in Molten Calcium Chloride. Met. Mater. Trans. A 2007, 38, 893–900. [CrossRef] 4. Zhao, K.; Wang, Y.; Gao, F. Electrochemical extraction of titanium from carbon-doped titanium dioxide precursors by electrolysis in chloride molten salt. Ionics 2019, 25, 6107–6114. [CrossRef]

5. Li, J.; Zhang, X.Y.; Liu, Y.B.; Li, Y.G.; Liu, R.P. Electrochemical behavior of tungsten in (NaCl-KCl-NaF-WO3) molten salt. Rare Met. 2013, 32, 512. [CrossRef] 6. Chen, G.Z.; Fray, D.J.; Farthing, T.W. Direct Electrochemical Reduction of Titanium Dioxide to Titanium in Molten Calcium Chloride. Nature 2000, 407, 361–364. [CrossRef] 7. Han, W.; Li, M.; Zhang, M.-L.; Yan, Y.-D. Progress in preparation of rare earth metals and alloys by electrodeposition in molten salts. Rare Met. 2016, 35, 811–825. [CrossRef] 8. Zhou, Z.; Hua, Y.; Xu, C.; Li, J.; Li, Y.; Gong, K.; Ru, J.; Xiong, L. Preparation of Ferrotitanium from Ilmenite by

Electrolysis-Assisted Calciothermic Reduction in CaCl2-NaCl Molten Salt. JOM 2015, 68, 532–539. [CrossRef] 9. Tan, S.; Örs, T.; Aydınol, M.K.; Anikina, E.; Karakaya, I. Synthesis of FeTi from mixed oxide precursors. J. Alloys Compd. 2009, 475, 368–372. [CrossRef] 10. Panigrahi, M.; Iizuka, A.; Shibata, E.; Nakamura, T. Electrolytic reduction of mixed (Fe, Ti) oxide using molten calcium chloride electrolyte. J. Alloys Compd. 2013, 550, 545–552. [CrossRef] 11. Shi, R.M.; Bao, L.F.; Zhang, B. Eﬀect of Temperature on the Ti-Fe Alloy Prepared by Molten Salt Electrolysis. Adv. Mater. Res. 2014, 936, 1189–1194. [CrossRef] 12. Du, J.; Li, Q.; Yang, S.; Yang, C.; Li, Z.; Xi, Z. Preparation of TiFe alloy by ilmenite electrolysis in molten salt. Rare Met. Mater. Eng. 2010, 39, 2247–2250.

13. Guo, X.; Guo, Z.; Wang, Z. Direct preparation of TiFe alloy by electrolytic reduction from TiO2 and Fe2O3. J. Univ. Sci. Technol. Beijing 2008, 30, 620–624. 14. Ding, M.T. Ratio of magnesium in prepared high titanium ferroalloy with titanium concentrate and titanium slag by therite reduction. Adv. Mater. Res. 2012, 581, 1092–1095. [CrossRef] 15. Lu, X.; Zou, X.; Li, C.; Zhong, Q.; Ding, W.; Zhou, Z. Green Electrochemical Process Solid-Oxide Oxygen-Ion-Conducting Membrane (SOM): Direct Extraction of Ti-Fe Alloys from Natural Ilmenite. Met. Mater. Trans. A 2012, 43, 503–512. [CrossRef] 16. Mohanty, J. Electrolytic Reduction of Titania Slag in Molten Calcium Chloride Bath. JOM 2012, 64, 582–584. [CrossRef] 17. Li, Z.-Q.; Ru, L.-Y.; Bai, C.-G.; Zhang, N.; Wang, H.-H. Eﬀect of sintering temperature on the electrolysis

of TiO2. Int. J. Miner. Met. Mater. 2012, 19, 636–641. [CrossRef] Metals 2020, 10, 1611 15 of 15

18. Shen, Y.Y.; Chen, C.Y.; Li, J.Q. Preparation of titanium by SOM electrolytic process from ultraﬁne high titanium slag. Rare Met. Mater. Eng. 2019, 48, 1671–1676. 19. Panigrahi, M.; Shibata, E.; Iizuka, A.; Nakamura, T. Production of Fe-Ti alloy from mixed ilmenite and titanium dioxide by direct electrochemical reduction in molten calcium chloride. Electrochim. Acta 2013, 93, 143–151. [CrossRef]

20. Chen, C.; Yang, X.; Li, J.; Lu, X.; Yang, S. Direct Electrolytic Reduction of Solid Ta2O5 to Ta with SOM Process. Met. Mater. Trans. A 2016, 47, 1727–1735. [CrossRef] 21. Zhou, Z.; Hua, Y.; Xu, C.; Li, J.; Li, Y.; Zang, Q.; Zang, Y.; Kuang, W. Synthesis of micro-FeTi powders by

direct electrochemical reduction of ilmenite in CaCl2-NaCl molten salt. Ionics 2017, 23, 213–221. [CrossRef] 22. Wang, S.L.; Li, S.C.; Wan, L.F.; Wang, C.H. Electro-deoxidation of V2O3 in molten CaCl2-NaCl-CaO. Int. J. Miner. Met. Mater. 2012, 19, 212. [CrossRef]

23. Wang, S.L.; Wang, W.; Li, S.C.; Cao, S.H. Cathodic behavior of molten CaCl2-CaO and CaCl2-NaCl-CaO. Int. J. Miner. Met. Mater. 2010, 17, 791. [CrossRef]

24. Rong, L.; He, R.; Wang, Z.; Peng, J.; Jin, X.; Chen, G.Z. Investigation of electrochemical reduction of GeO2 to Ge in molten CaCl2-NaCl. Electrochim. Acta 2014, 147, 352–359. [CrossRef] 25. Kartal, L.; Daryal, M.B.; ¸Sireli, G.K.; Timur, S. One-step electrochemical reduction of stibnite concentrate in molten borax. Int. J. Miner. Met. Mater. 2019, 26, 1258–1265. [CrossRef] 26. Song, J.X.; Wang, Q.Y.; Hu, G.J.; Zhu, X.B.; Jiao, S.Q.; Zhu, H.M. Equilibrium between titanium ions and high-purity titanium electroreﬁning in a NaCl-KCl melt. Int. J. Miner. Met. Mater. 2014, 21, 660. [CrossRef]

27. Chen, S.; Liao, C.F.; Lin, J.Y.; Cai, B.Q.; Wang, X.; Jiao, Y.F.Electrical conductivity of molten LiF-DyF3-Dy2O3-Cu2O system for Dy-Cu intermediate alloy production. Int. J. Miner. Met. Mater. 2019, 26, 701–709. [CrossRef]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).