Three independent reactions are connected in Kroll process; chlorination of TiO2, Mg reduction of TiCl4, and molten salt electrolysis of MgCl2. The Kroll Process can produce a highly pure TiCl4, and the final product is clean suitable for the aerospace applications. However, the reduced Ti strongly adhered to the Mg reduction vessel wall and it prohibited from the continuous production. The high energy cost due to the separated operation of the endothermal molten salt electrolysis and the exothermal Mg reduction is another demerit of the Kroll process. This work plans to connect the electrolysis and the reduction in a single bath, using a high purity of TiCl4 produced from the chlorination.
A new method for metallic titanium production is here proposed that TiCl4 is continuously reduced by Ca precipitated by the in-situ electrolysis of molten CaCl2. An outline of this proposal is illustrated in Fig. 1. This method forms liquid Ca and Cl2 gas at the cathode and anode, respectively, by electrolysis of the molten CaCl2 at approximately 1173 K.
|Ca2+(in CaCl2) + 2e = Ca (in CaCl2)||(1)|
|2 Cl- (in CaCl2) = Cl2 (gas) + 2e||(2)|
The evolved Cl2 gas is industrially used for production of TiCl4. The precipitated Ca dissolves in the CaCl2 melt; the solubility of Ca at 1173 K has a minimum value of 2.7 mol% and a maximum value of 4.0 mol%, although some variations in these values have been reported. This dissolved calcium will be hereafter referred to as Ca, which exists especially close to the cathode. When TiCl4 gas is injected to Ca, the reduction occurs as expressed in eq.(3) and the metallic Ti is produced.
|2 Ca (in CaCl2) + TiCl4 (gas) = Ti (solid) + 2 CaCl2 (liquid)||(3)|
Eq. (3) occurs at the interface between the liquid and the gaseous phase, and forms the metallic powder as the product. The powder formation is desirable for the continuous operation, and Ca neither reacts with Ti nor dissolves in Ti. The operation becomes simpler and the plant becomes compact as compared to the plant using the Kroll process where MgCl2 is removed from the hot vessel. We expect that the energy consumed during Ti production can be saved because the exothermal and endothermal reactions are simultaneously conducted in the same cell. If we can realize a method to extract Ti powder from the molten salt, it could be applied to continuous industrial operation.
The purpose of this work is to verify experimentally this proposal combining eqs.(1)(3) in a bath, and to find a suitable condition to form the metallic titanium in this bath. Firstly, prior to the combination of electrolysis with eq.(3), this study verifies the feasibility of whether Ca can reduce TiCl4 gas, as shown in Fig. 2(b). Secondly this study verifies the whole concept shown in Fig. 2(a).
Reduction of TiCl4 by Ca in Molten CaCl2
As shown in Fig. 2(b), 660700 g of CaCl2 was melted at 1173 K, and several Ca grains were added to obtain the desired Ca concentration. TiCl4 gas was then fed at a constant rate for 1.84.8 ks. The end of this pipe was placed 1020 mm above the melt surface. After blowing, the white solidified salt and the residual Ca were removed by water.
A part of the product was piled up as powder at the bottom of crucible, and the other was tightly sintered lumps at the surface of the solidified salt in the vicinity of the crucible wall. The latter was considered to nucleate at the three-phase interface among the molten salt, crucible wall, and gas phase, and to grow toward the melt on its surface. The former powder was a product of the reaction at the free surface between the molten CaCl2 containing Ca and the gaseous TiCl4.
The obtained Ti particles were spherical with a diameter of approximately 1 m, and were sintered slightly. The lumps at the free surface of molten salt were often sintered tightly on surfaces like the plate or bar. However, some samples comprised the same size of Ti particles. No dendritic growth was found in the morphology. Fig. 3 shows the phase identification of the powders by X-ray diffraction (XRD) measurements and summarizes the changes in the Ca concentration during the reaction. Because of Ca consumption in the reaction, the ends of the arrows show the initial and final concentrations evaluated by assuming the complete reaction. Metallic Ti powder was obtained in the experiments from Runs #A to G.
The mixture of -Ti and its lower oxides was obtained in Runs #C, E, F, and G. In these experiments, the final Ca concentration became low because a large amount of TiCl4 was fed, although the initial Ca concentration was high. Only the lower oxides were obtained in Runs #H and I, where the Ca concentration was low and the amount of Ca was also low as compared to the TiCl4 feed.
Calcium and TiCl4 react stoichiometrically with a molar ratio of 2:1, as shown in eq.(3). When the molar ratio of the amount of Ca and that of TiCl4 (NCa/NTiCl4) is larger than 2, it is expected that the entire amount of fed TiCl4 can be reduced to -Ti and some amount of residual Ca. The experiments could approximately confirm this requirement. In terms of thermodynamics, this means that a strong affinity of Ca with Cl can decompose TiCl4 even if Ca dissolves as Ca. In Runs #C and E, however, the supplied TiCl4 could not be reduced to titanium because NCa/NTiCl4 was smaller than 2. In contrast, the lower oxides of titanium were formed when the Ca concentration decreased during the reaction to a level of <2 mol%, although -Ti could be formed when the initial concentration of Ca was high. This shows that the thermochemical activity of Ca should be maintained high because no additional charge of Ca was provided.
The reasons for the formation of the lower oxides of Ti were the small leakage of air in Ca feed, the ceramic crucible, and a small amount of water in the CaCl2 used. TiCl4 gas blown onto the molten salt reacts with Ca to form lower chlorides such as TiCl3 and TiCl2. However, they were thermodynamically unstable under the given experimental conditions and react with oxygen in the form of lower oxides.
TiCl3 or TiCl2 (in CaCl2) + CaO (in CaCl2) = Ti2O3 or TiO + CaCl2 (liquid) (4)
It is not easy to remove all the possible oxygen sources in the lab-scale test, but it may be possible to minimize oxygen contamination on a larger scale.
Reduction by Ca from Electrolysis of CaCl2
The apparatus as illustrates in Fig. 4 was used for the electrolysis combined with eqs.(1)-(3). The cathode and the anode were made of Ti and carbon bars, respectively. In several experiments, the anode was set inside the mullite pipe in order to exhaust the evolved Cl2 gas efficiently to the outside of the vessel. The electrolysis was started by applying 4.0 5.0 V constantly between two electrodes. Because the theoretical voltage for decomposition of CaCl2 is 3.21 V at 1173 K, the applied voltage was set enough high by considering an over-voltage for Cl2 gas evolution. TiCl4 gas was constantly blown to the melt surface near the cathode for 4.1 25.0 ks after a certain amount of Ca forming by the preliminary electrolysis.
Resultantly -Ti was successfully obtained partially as the dendrite form adhered to the cathode, as shown in Table 1. This is because Ti2+ or Ti3+ ions were electrochemically precipitated after long electrolysis, or because they were preferentially reduced to the metallic Ti powder by meeting with the larger amount of Ca near the cathode. Especially in Run#ED and #EE, -Ti porous powder was recovered as the slightly sintered particles at the bottom of the MgO crucible. The primary particle size was smaller than 1 m, and the secondary particles were about 15 m at largest.
Assuming that all the supplied current was used only for the electrolysis of CaCl2, the total amount of Ca during the electrolysis, NCa, was evaluated. Using the amount of blown TiCl4 gas, NTiCl4, the ratio of NCa/NTiCl4, was listed in Table 1, in addition to the final Ca concentration in the molten salt after the reaction, XCa. Because TiCl4 and Ca react by the molar ratio of 1:2 to form Ti according to eq.(3), all the amount of blown TiCl4 should be reduced completely when NCa/NTiCl4 > 2.
In the experiments of both Run#ED and #EE, NCa/NTiCl4 > 2 and XCa > 2 mol%. These conditions should form only -Ti. However, the intermediate product, TiO, was included as well as -Ti. This shows that a significant large difference in Ca concentration in the melt was generated depending on the position. This concentration profile of Ca is formed because the dissolution and diffusion of precipitated Ca at the cathode toward the bulk is slow, or more probably because Ca is consumed by the back-reaction of eq.(2) that the Cl2 gas evolved at the anode reacts with Ca generated at the cathode.
The reduction became incomplete, and remained at the mixing level of Ti2+ or Ti3+, when the formed amount of Ca was smaller than the twice amount of TiCl4 or when the Ca concentration was lower than 2 mol%. Resultantly these products were recovered as the lower Ti oxides such as TiO and Ti2O3, which was similar with the case of Ca reduction as mentioned in the previous section 3.
In summary, a new approach was proposed to continuously produce metallic Ti from Ca dissolved in CaCl2 and gaseous TiCl4. As fundamental confirmation of this proposal, metallic Ca was dissolved in molten CaCl2 in advance, and it was reacted with TiCl4 gas blown onto the free surface of molten CaCl2. As a result, -Ti powder or its sinter was recovered when the Ca concentration was above 2 mol%. In addition to -Ti, the lower oxides of Ti were obtained when the Ca concentration decreased to below 2 mol%. Combining with the molten salt electrolysis of CaCl2, -Ti was formed as the powder at the bottom of MgO crucible, when NCa/NTiCl4 > 2 and XCa > 2 mol%.