Fundamental study on synthesis and enrichment of titanium subchloride

Article abstract of DOI:10.1016/j.jallcom.2007.02.128

In order to establish a new high-speed/(semi-)continuous titanium production process based on the magnesiothermic reduction of titanium subchlorides (subhalide reduction process), a novel synthetic process for obtaining titanium subchlorides (TiCl_x, x = 2, 3) by the reaction of titanium tetrachloride (TiCl₄) with titanium metal at 1273 K was investigated. It was demonstrated that the efficiency of the TiCl_x formation improved drastically when molten salts were used as the reaction medium as compared with that of the synthesis by employing the direct reaction of TiCl₄ gas with solid titanium. The feasibility of the enrichment process of TiCl_x in molten salt was also demonstrated. The method for producing the titanium subchlorides investigated in this study can be applied to the new high-speed production process of titanium.

Full text of DOI:10.1016/j.jallcom.2007.02.128

Journal of Alloys and Compounds 457 (2008) 376–383
Fundamental study on synthesis and enrichment of titanium subchloride
1
∗
Osamu Takeda , Toru H. Okabe
Institute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan
Received 27 August 2006; received in revised form 22 February 2007; accepted 23 February 2007
Available online 28 February 2007
Abstract
In order to establish a new high-speed/(semi-)continuous titanium production process based on the magnesiothermic reduction of titanium
subchlorides (subhalide reduction process), a novel synthetic process for obtaining titanium subchlorides (TiCl , x = 2, 3) by the reaction of
titanium tetrachloride (TiCl ) with titanium metal at 1273 K was investigated. It was demonstrated that the efficiency of the TiCl formation
improved drastically when molten salts were used as the reaction medium as compared with that of the synthesis by employing the direct reaction
of TiCl gas with solid titanium. The feasibility of the enrichment process of TiCl in molten salt was also demonstrated. The method for producing
the titanium subchlorides investigated in this study can be applied to the new high-speed production process of titanium.
2007 Elsevier B.V. All rights reserved.
x
4
x
4
x
©
Keywords: Titanium; Titanium subchloride; Synthesis; Molten salt; Subhalide reduction process
1
. Introduction
the essential advantage of producing high-purity titanium by
carrying out the reduction process in an oxygen-free system.
Based on this background, the authors are developing a
new high-speed/(semi-)continuous titanium production pro-
cess based on the magnesiothermic reduction of titanium
subchlorides—titanium dichloride (TiCl2) and/or titanium
trichloride (TiCl3) [16–20]. This new titanium production pro-
cess (cf. Fig. 1), named as the subhalide reduction process,
involves three major steps: (1) production of titanium sub-
chlorides (TiClx, x = 2, 3), (2) enrichment of TiClx when the
subchlorides are produced in a magnesium chloride (MgCl2)
medium, and (3) magnesiothermic reduction of TiClx.
The advantages of employing subchlorides as feed materi-
als for the reduction process are as follows: (a) A high-speed
and (semi-)continuous process can be designed using conven-
tional techniques because the heat produced by the reduction
of subchlorides is substantially lower than that produced by the
reduction of TiCl4. In addition, the density of the reaction field
and the heat extraction rate from the reaction container increase
drastically by employing a reduction process in the condensed
phase because titanium subchlorides are stable as a condensed
phase even at elevated temperatures. (b) The iron contamination
of titanium products from the reaction container can be avoided
since metallic titanium can be utilized as the reactor material
under Ti/TiCl2 equilibrium. (c) The MgCl2 reaction product can
be easily removed by vacuum distillation, and high-purity tita-
nium with low oxygen content can be obtained. (d) The obtained
Presently, metallic titanium is produced by the Kroll process
[
1–4] based on the magnesiothermic reduction of titanium tetra-
chloride (TiCl4). A technical drawback in the titanium industry
is the low productivity of the current titanium production pro-
cess due to the slow-speed/batch-type reduction process. In fact,
the rapid increase in the demand for titanium in recent years has
resulted in a serious shortage of titanium in the market [5]; this is
because it is difficult to increase the production speed of titanium
by the Kroll process.
In order to establish a next-generation reduction process with
high productivity, research and development on the direct reduc-
tion processes of titanium dioxide (TiO2) are actively being
carried out worldwide [6–10]. These new reduction processes
have the potential for producing low-cost titanium by simplify-
ing the titanium production process. However, several technical
problems such as impurity control need to be resolved prior to
establishing a large-scale commercial process. Meanwhile, new
titanium production processes based on chloride metallurgy are
also being investigated [11–15] because chloride metallurgy has
∗
Corresponding author. Tel.: +81 3 5452 6314; fax: +81 3 5452 6313.
E-mail address: okabe@iis.u-tokyo.ac.jp (T.H. Okabe).
Graduate student, University of Tokyo (from October 2006, Assistant Pro-
1
fessor, Department of Metallurgy, Graduate School of Engineering, Tohoku
University, 6-6-2 Aramaki Aza Aoba, Aoba-ku, Sendai 980-8579, Japan).
0
925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.02.128

O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
377
thetic processes (1) and (3) mentioned above), difficulty in the
separation of the reaction product (synthetic process (2)), and
complexity of the reaction apparatus (synthetic process (4)).
For efficiently producing titanium subchlorides on an indus-
trial scale, the reaction of TiCl4 with magnesium metal is
expected to be suitable because the reaction rate is high and
because the reaction product (MgCl2) is not necessarily removed
for the use of titanium subchlorides in the subhalide reduction
process [19–20]. This magnesiothermic reduction of TiCl4 to
form TiCl2 is technically feasible and can be applied to a large-
scale process because this process can be carried out in a reactor
made of steel.
Besides the magnesiothermic reduction of TiCl4, the reaction
of TiCl4 with titanium metal using a simple and compact reac-
tion apparatus is also suitable for subchloride production. When
titanium metal is used as a reductant for producing titanium sub-
chlorides, titanium metal scraps, which will increase in quantity
in the future, can be effectively utilized as the reductant feed
source; consequently, an environmentally sound process can be
designed. In the past, a very small number of studies were con-
ducted on subchloride synthesis that employed titanium metal.
Therefore, a novel synthetic process for producing titanium sub-
chlorides by reacting TiCl4 with titanium metal was investigated
in this study. Furthermore, an experiment for the enrichment
process of TiClx in molten salt was carried out.
2
. Experimental
.1. TiCl synthesis by reaction of TiCl
The materials used in this study are listed in Table 1, and a schematic illus-
Fig. 1. Flowchart of the new titanium production process based on the subhalide
reduction process.
2
x
4
with titanium metal
titanium sponge can be melted together with the reactor because
the reactor is made of titanium metal. Therefore, the procedure
for the mechanical removal of iron-contaminated sponge can
be avoided, and the titanium sponge is directly melted and the
melt is directly cast into an ingot after the vacuum distillation
process. In the current Kroll process, a massive titanium sponge
stuck to the iron container needs to be taken out of the container
and crushed into small particles before melting.
Thus far, the authors have carried out fundamental research
with the aim of establishing the subhalide reduction process
that employs the magnesiothermic reduction of TiCl3 or TiCl2
in a titanium reaction container. Further, we have successfully
demonstrated the feasibility and advantages of the reduction
process [16–20]. In order to carry out a systematic study on
the subhalide reduction process, new synthetic and enrichment
processes for titanium subchlorides, which can be used as feed
materials for the subhalide reduction process, were experimen-
tally investigated in this study.
tration of the representative experimental apparatus for the TiClx synthesis by
the reaction of TiCl4 with titanium metal is shown in Fig. 2. On the basis of
the preliminary experiment, it was found that the efficiency of TiClx synthesis
employing titanium sponge as the feed material was low at 1073–1273 K; there-
fore, titanium powder was used as the feed material for improving the efficiency
of the synthesis reaction in this experiment. Titanium powder (99.4%, average
particle diameter (d50): 25 ␮m, 6.3 g) was placed on molybdenum trays (Mo,
9
9%, octagonal pillar shape, ∅ 52 mm outer diameter, 10 mm height, 0.2 mm
thickness) or stainless-steel trays (SUS304, identical in shape to the Mo tray),
and the trays were inserted into a stainless-steel reaction container (SUS316,
61 mm outer diameter, 90 mm height, 3 mm thickness). The reaction container
filled with argon (Ar) gas was heated to 1273 K in an electric furnace.
The representative experimental conditions are summarized in Table 2. Liq-
uid TiCl4 (99%, 26 g), which was in stoichiometric mass according to Eq. (1) for
reacting with all the titanium feed powder, was transported by a peristaltic pump
at the rate of 0.12–0.64 g/min and fed into the stainless-steel reaction container
by dropping it from the top of the chamber. After the experiment, the sample was
gradually cooled in the furnace and recovered in a nitrogen (N2) atmosphere.
◦
Ti(s) + TiCl4(g) → 2TiCl2(s); ꢀH = –244 kJ,
◦
ΔG = –18.9 kJ at 1273 K [26]
(1)
Thus far, the following processes have been reported in litera-
ture with regard to the investigations of the synthesis of titanium
chloride: (1) TiCl3 by the reduction of TiCl4 with hydrogen gas
Table 1
Materials used in this study
(H2) [21], (2) TiCl3 by the aluminothermic reduction of TiCl4
[22], (3) TiCl2 by the thermal decomposition of TiCl3 [23], and
(4) TiCl2 by the reduction of TiCl4 with metallic titanium fol-
Materials
Form
Purity
mass%)
Supplier
(
lowed by the vapor transport of TiCl2 [24,25]. These processes
are useful for lab-scale production; however, they are unsuit-
able for industrial-scale production, particularly for the titanium
metal production process due to the slow reaction rate (syn-
TiCl4
Ti
Ti
Liquid
Sponge
Powder
Bulk
≥99.0
≥99.7
≥99.4
≥97
Wako Pure Chemical Industries Ltd.
Toho Titanium Co. Ltd.
Toho Titanium Co. Ltd.
MgCl2
Wako Pure Chemical Industries Ltd.

3
78
O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
Table 2
Experimental conditions for TiClx synthesis by reaction of TiCl4 with titanium metal in a closed-type reaction containera
Experment no.
Mass of sample
TiCl4 feed rate,
Tray material
Mass of recovered
TiClx, wTiClx (g)
Mass of consumed
Ti powder, wTi (g)
−
1
r (g min
)
Ti powder,
TiCl4,
wTi (g)
wTiClx (g)
A
B
C
D
6.31
6.32
6.31
6.30
25.5
25.7
25.5
25.5
0.64
0.17
0.17
0.12
Mo
Mo
SUS 304
Mo
7.97
9.87
7.26
2.69
2.78
2.66
2.84
10.88
Mass of reacted samples after experiments is also listed for reference (see Fig. 7 for the yield of the recovered TiClx and Ti consumption ratio).
a
Temperature on the bottom of the reaction vessel was held at 1269–1277 K.
2
.2. TiCl
x
synthesis in molten salt
for more than 3 days was placed in a stainless-steel reaction container and heated
to 1273 K in an argon atmosphere. MgCl2 was expected to work as a medium that
removes the solid TiCl2 film formed on the surface of the metallic titanium by
dissolving and accumulating TiCl2 in the medium. This is because the solubility
of TiCl2 in MgCl2 at 1273 K is reported to be high (∼83 mol%) as shown in Fig. 4
A schematic illustration of the representative experimental apparatus for
TiClx synthesis in molten salt is shown in Fig. 3, and the experimental conditions
are summarized in Table 3. Anhydrous MgCl2 (97%, 100 g) dried under vacuum
Fig. 2. (a) Schematic illustration of an experimental apparatus for TiClx synthe-
sisbythereactionofTiCl4 withtitaniummetal(aclosed-typereactioncontainer).
Fig. 3. (a) Schematic illustration of the representative experimental apparatus
for TiClx synthesis in the molten salt. (b) Photographs of the stainless-steel
basket and reaction container.
(
b) Photograph of Mo trays tentatively installed in a transparent beaker.

O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
379
Table 3
Experimental conditions for TiClx synthesis in molten salta
Experiment no.
Mass of sample i, wi (g)
TiCl4 feed rate,
−
1
r (g min
)
Ti
6.31
TiCl4
MgCl2
W
X
Y
Z
25.66
25.70
25.40
76.15
99.98
100.12
100.07
100.05
0.56
0.16
0.13
0.65
6.31
6.32
18.97
a
Temperature on the bottom of the reaction vessel was held around at 1273 K.
[
27]. After MgCl2 had melted, the titanium sponge (99.7%, 6.3–19 g), which
was placed in a stainless-steel basket, was immersed in the molten MgCl2. In a
manner similar to the previous study, liquid TiCl4 (99%, 26–76 g) was fed into
this reaction container at the rate of 0.13–0.72 g/min by dropping it from the
top of the chamber through a stainless-steel feed tube. It was expected that the
TiCl4 supplied in the form of a gas bubble at elevated temperatures reacted with
the titanium sponge in the molten MgCl2. After the experiment, the stainless-
steel basket was removed from the molten salt. The molten salt in the reaction
container and the residual titanium in the basket were gradually cooled in the
furnace and recovered in the nitrogen atmosphere at room temperature.
Fig. 5. Estimated profile of the vapor pressure of TiCl2 in MgCl2–TiCl2 molten
salt.
86.4 ks (24 h). After the experiment, the capsule was removed from the furnace
and quenched in water. The salt sample was recovered from several positions of
the capsule in a nitrogen atmosphere.
2
.4. Analysis
2
.3. Enrichment of TiCl
x
in molten salt
The phases in the sample were identified by X-ray diffraction analysis
XRD). In order to prevent the samples from reacting with the moisture in the
An estimated relationship between the composition and vapor pressure of
(
TiCl2 in the MgCl2–TiCl2 pseudo-binary system at 1073 and 1273 K is shown
in Fig. 5. The calculation was carried out using thermodynamic data [26] and
solubility data [27] by assuming that Henry’s law can be applied for the compo-
nents of the melt. As shown in the figure, the vapor pressure of TiCl2 at 1073 K
air, they were covered with a polyimide film in a glove box and then analyzed by
XRD. The metallic element content in the sample was determined by inductively
coupled plasma-atomic emission spectrometry (ICP-AES) after dissolving the
sample in acid aqueous solutions. The chlorine content in the sample was deter-
mined by the potentiometric titration method. In this method, the sample was
dissolved in an acid aqueous solution. A silver nitrate (AgNO3) aqueous solu-
tion was then titrated into the solution by monitoring the difference between
the electrochemical potential of a silver electrode and an Ag/AgCl electrode
immersed in the solution.
−
5
−3
(
7.7 × 10 atm) is significantly lower than that at 1273 K (8.9 × 10 atm).
Therefore, it was expected that TiCl2 would be transferred from a higher tem-
perature zone (with a high vapor pressure pH) to a lower temperature zone (with
a low vapor pressure pL) to compensate for the vapor pressure of TiCl2 in the
entire system down to pL when a temperature gradient exists in the MgCl2–TiCl2
molten salt.
The MgCl2–TiClx salt (titanium concentration: 7.36–9.26 mass%, 15 g)
obtained from the experiment for the TiClx synthesis in molten salt was filled
into a tubular stainless-steel capsule (SUS316, ∅ 9 mm inner diameter, 200 mm
length), which was sealed by tungsten inert gas (TIG) welding. In the represen-
tative experiment, the entire capsule, which was filled in a horizontal furnace,
was maintained at 1273 K for 3.6 ks (1.0 h). Subsequently, the temperature at one
end of the capsule was decreased to 1012 K and maintained in this condition for
3. Results and discussion
3.1. TiClx synthesis by reaction of TiCl4 with titanium metal
After the experiment for the TiClx synthesis by the reaction
of TiCl4 gas with titanium metal, pure TiCl3 was condensed at
the lower temperature part of the stainless-steel tube for feed-
ing TiCl4 into the reaction container. The titanium feed powder
placed on the molybdenum trays was sintered, and a black flake
deposit was formed on the sintered titanium powder. By XRD
analysis, it was revealed that the deposit sample obtained mainly
comprised TiCl2, and it also contained some amount of TiCl3
as shown in Fig. 6. The average composition of the deposit
samples obtained from each tray is summarized in Table 4.
The concentrations of titanium and chlorine in the sample were
36.0–39.1 mass% and 60.1–63.3 mass%, respectively. The stoi-
chiometry of TiClx in the sample is calculated and listed in the
last column of Table 4. From these results, the calculated x val-
ues in TiClx were 2.07–2.38, which showed a good agreement
with the XRD analysis.
As shown in Fig. 7, the yield of the TiClx sample (RTiClx
)
recovered from the reaction container was 23–35%, and it
decreased with an increase in the TiCl4 feed rate. The consump-
tion ratio of the titanium feed powder after the experiment was
Fig. 4. Phase diagram for the MgCl2–TiCl2 pseudo-binary system. The overall
composition after the synthesis experiment (experiments W–Z) is also shown in
the figure [27].
42–45%, and this ratio was independent of the TiCl feed rate
4

3
80
O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
Fig. 7. Yield of TiClx and titanium consumption ratio in the experiment for
TiClx synthesis by the reaction of TiCl4 with titanium metal.
TiCl2 solid film formed on the surface of metallic titanium has
to be removed effectively and collected. Further, the reaction of
TiCl2 with TiCl4 has to be prevented.
◦
TiCl2(s) + TiCl4(g) → 2TiCl3(g); ꢀH = +181 kJ,
◦
ΔG = −28.4 kJ at 1273 K [26]
(2)
3.2. TiClx synthesis in molten salt
From the results obtained in the previous section, the effi-
ciency of the TiClx formation by the direct reaction of TiCl4 gas
with solid titanium was found to be low when a gas/solid phase
reaction at 1273 K was employed. Therefore, a more efficient
synthetic process for titanium subchlorides using molten salt as
a reaction medium was investigated.
Fig. 6. Representative X-ray diffraction patterns of the samples obtained from
each tray after the experiment for TiClx synthesis using a closed-type reaction
container (experiment D). XRD reference pattern: ␣-TiCl3: PDF #29-1357, ␥-
TiCl3: PDF #29-1359, TiCl2: PDF #10-0315 and 73-0751.
After the experiment for TiClx synthesis in molten salt,
the pure TiCl3 was condensed at the lower temperature part
of the stainless-steel chamber that held the reaction container.
The solidified salt in the reaction container had separated into
two layers—the upper and lower parts—when it was observed
after the experiment (experiments W–Y). The concentration
of titanium in the salt obtained from the upper part of the
reaction container was 1.01–2.33 mass% (cf. Table 5, samples
W (a), X (a), and Y (a)), and that from the lower part was
6.59–9.93 mass% (cf. Table 5, samples W (b)–(d), X (b)–(d),
and Y (b)–(d)). The calculated values of x in the composition for-
mula MgCl2–TiClx of the salt in the lower part were 2.04–2.61,
employed in this study. From these results, the efficiency of the
TiClx formation from TiCl4 gas and solid titanium was found
to be low under the experimental conditions (T = 1273 K, 1 atm,
r = 0.13–0.65 g TiCl4/min). One reason for this low efficiency
of the TiClx formation is probably because the TiCl2 solid film
formed on the surface of the titanium feed powder hindered
the chlorination reaction of the metallic titanium, given by Eq.
(
1). Furthermore, the obtained TiCl2 solid was converted into
TiCl3 gas by the reaction with TiCl4 gas (Eq. (2)) during the
experiment; consequently, the nominal TiCl2 yield decreased.
Therefore, for improving the efficiency of TiClx formation, the
Table 4
Analytical results of the samples obtained from the TiClx synthesis experiment by the reaction of TiCl4 with titanium metal in a closed-type reaction container.
Experiment no.
Concentration of element i, Ci (mass%)a
x value in TiClx
Tib
Clc
Feb
Nib
Crb
Mob
0.03
0.04
<0.01
0.04
A
B
C
D
39.1
36.0
38.5
37.5
60.1
63.3
60.5
61.8
0.19
0.16
0.68
0.18
0.49
0.31
0.24
0.42
0.20
0.16
0.17
0.10
2.07
2.38
2.13
2.23
a
Average of six samples obtained from each trays.
Determined by inductively coupled plasma-emission spectroscopy (ICP-AES).
Determined by potentiometric titration method.
b
c

O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
381
Table 5
Analytical results of the samples obtained from the reaction container in the experiment for TiClx synthesis in molten MgCl2
Experiment no.
position)
Mass of recovered
salt, wsalt (g)
Concentration of element i, Ci (mass%)
x value in
TiClx
(
Tia
Clb
Mga
Fea
0.02
0.02
0.02
0.07
Nia
Cra
0.05
0.05
<0.01
<0.01
W (a)
W (b)
W (c)
W (d)
22.8
30.7
23.2
34.9
1.58
75.5
72.4
71.6
71.4
22.9
20.2
19.5
18.6
<0.02
<0.02
<0.02
<0.02
7.45
2.47
2.24
2.33
7.42
8.90
9.93
X (a)
X (b)
X (c)
X (d)
16.6
22.0
27.3
17.6
2.33
7.07
8.87
9.58
74.6
72.3
71.9
71.3
23.0
20.6
19.2
19.0
0.03
0.01
0.01
0.08
<0.02
<0.02
<0.02
<0.02
<0.01
0.04
0.03
0.04
4.35
2.31
2.39
2.26
Y (a)
Y (b)
Y (c)
Y (d)
12.6
30.1
24.4
23.9
1.01
6.59
8.16
8.96
74.9
72.7
71.5
71.1
24.1
20.6
20.3
19.7
0.02
0.02
<0.01
0.13
<0.02
<0.02
<0.02
0.03
<0.01
0.07
0.02
0.06
6.38
2.61
2.04
2.06
Z (a)
Z (b)
Z (c)
Z (d)
29.8
33.6
40.3
45.7
16.1
69.0
70.8
71.0
68.7
14.8
17.0
15.0
12.4
0.05
0.01
0.02
0.02
<0.01
<0.01
<0.01
<0.01
0.01
0.01
0.01
0.01
2.17
2.36
2.63
2.33
12.2
14.0
18.9
a
Determined by inductively coupled plasma-emission spectroscopy (ICP-AES).
Determined by potentiometric titration method.
b
which was close to the composition of TiCl2. The x values of the
salt in the upper part were greater than 4. These large x values
are probably due to an analytical error, but the oxidation state of
a titanium ion is probably high at 4. As shown in Table 5 (experi-
ment Z), a one-layer solidified salt, in which the concentration of
titanium was 12.2–18.9 mass% (cf. Table 5, samples Z (a)–(d)),
was obtained when three times by mass of TiCl4 and titanium
sponge were fed in the experiment. The calculated values of x in
MgCl2–TiClx of the salt were 2.17–2.63, which was also close
to 2 (Table 5).
When MgCl2 molten salt was employed as the reaction
medium, the yield of TiClx was 46–64%, and the consumption
ratioofthetitaniumspongefeedachievedwas80–93%(Table6).
From these results, it was demonstrated that the efficiency of
the TiClx formation was drastically improved by using molten
MgCl2 and that MgCl2 can be used as the reaction medium for
the synthesis of the TiClx feed for the subchloride reduction
process.
nism is schematicallyillustratedas Route1 in Fig. 8. Meanwhile,
the gaseous TiCl4 could react with the TiCl2 dissolved in the
molten MgCl2 to form TiCl3 dissolved in molten salt (cf. Eq.
(3)), which in turn might react with the metallic titanium (cf.
Eq. (4)). The detailed mechanism of the reaction of TiCl4 with
metallic titanium in molten salt is under investigation.
TiCl (s, l) + TiCl (g) → 2TiCl (g, l);
2
4
3
◦
◦
ꢀH
= +181 kJ, ΔG = −28.4 kJ at 1273 K [26]
(3)
(4)
TiCl3(g, l) + (1/2)Ti(s) → (3/2)TiCl2(s, l);
◦
◦
ΔH = −212 kJ, ΔG = +4.75 kJ at 1273 K [26]
Since the solubility of TiCl4 in MgCl2 was reported to be low
(
0.55 mol at 1073 K) [28], it was expected that the TiCl4 bub-
ble injected into the molten MgCl2 physically contacted with
the metallic titanium to form TiCl2; and subsequently, the TiCl2
formed dissolved into the molten MgCl2. This reaction mecha-
Table 6
Yield of TiClx and Ti consumption ratio in the experiment for TiClx synthesis
in molten MgCl2
Experiment no.
TiCl4 feed rate,
Yield of TiClx,
RTiClx (%)
Ti consumption
−
1
ꢀ
ratio, R (%)
Ti
r (g min
)
W
X
Y
Z
0.56
0.16
0.13
0.65
64
46
49
60
80
93
84
75
Fig. 8. Schematic illustration of an expected reaction system for TiClx synthesis
by the reaction of TiCl4 with titanium metal in molten MgCl2.

3
82
O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
the magnesiothermic reduction of titanium subchlorides (sub-
halide reduction process). It was demonstrated that the efficiency
of TiClx formation was drastically improved when molten salt
(MgCl2) was used as the reaction medium as compared with
that of the synthesis by employing the direct reaction of TiCl4
gas with solid titanium. An experiment for the enrichment
process of TiClx in molten salt was carried out, and the fea-
sibility of the process was also demonstrated. The method for
producing the titanium subchlorides investigated in this study
can be applied to the new high-speed production process of
titanium.
Acknowledgment
The authors are grateful to Professors Masafumi Maeda,
Susumu Nanao Yoshitaka Mitsuda, and Kazuki Morita of the
University of Tokyo and Professor Tetsuya Uda of Kyoto Uni-
versity for their valuable discussions throughout this project.
The authors also thank Mr. Susumu Kosemura, Mr. Masanori
Yamaguchi, and Mr. Yuichi Ono of Toho Titanium Co. Ltd. for
their valuable discussions and for providing the samples. This
work was financially supported by a Grant-in-Aid for Scientific
Research (B) from the Ministry of Education, Culture, Sports,
Science and Technology, Japan (MEXT, Project ID #18360365).
One of the authors, O. Takeda, is grateful for the financial sup-
port provided by a Grant for the 21st Century COE Program
Fig. 9. Profile of the Ti concentration after the experiment for the enrichment of
TiClx in molten MgCl2 by utilizing the temperature dependence of the solubility
of TiClx in MgCl2.
3
.3. Enrichment of TiClx in molten salt
Since the concentration of TiClx synthesized in the molten
“
Human-friendly Materials based on Chemistry” from MEXT
salt was low, TiClx has to be enriched in order to improve the
efficiency of its subsequent reduction by magnesium. For this
reason, a fundamental experiment for investigating the feasibil-
ity of the enrichment process of TiClx in molten salt was carried
out.
and the financial support provided by Research Fellowships of
the Japan Society for the Promotion of Science (JSPS) for Young
Scientists.
The analytical results for the concentration of titanium in the
salt extracted from several positions of the capsule are shown
in Fig. 9 along with the temperature profile before quenching.
A black deposit was observed at the lower temperature part of
the capsule, and its titanium concentration was 17 mass%, which
was higher than the initial titanium concentration of the feed salt
References
[
[
1] W. Kroll, Tr. Electrochem. Soc. 78 (1940) 35–47.
2] T. Ogasawara, Titanium Jpn. (Jpn. Titanium Soc.) 53 (2005) 103–108.
[3] A. Moriya, A. Kanai, Shigen-to-Sozai 109 (1993) 1164–1169.
[
[
[
4] T. Fukuyama, M. Koizumi, M. Hanaki, S. Kosemura, Shigen-to-Sozai 109
1993) 1157–1163.
5] S. Nakamura, Industrial Rare Metals, Arumu Publishing Co., Tokyo, Japan,
004, pp. 52–55.
6] Z. Chen, D.J. Fray, T.W. Farthing, Nature 407 (2000) 361–364.
(
(
8.9 mass%). The concentration of titanium in the salt around
the black deposit was also high (13 mass%), although the tita-
nium concentration in the salt at the higher temperature part
remained almost unchanged from the initial value. The mecha-
nism of the enrichment of TiClx is currently under investigation,
but at this stage, its details are unclear. From these results, how-
ever, the feasibility of the enrichment process of TiClx in molten
salt by utilizing the temperature gradient was demonstrated,
and this method can be utilized in the subhalide reduction pro-
cess shown in Fig. 1. Currently, technologies for enhancing the
enrichment efficiency of titanium subchlorides are also being
developed.
2
[7] K. Ono, R.O. Suzuki, JOM (J. Met.) 54 (2002) 59–61, February.
[
8] T. Abiko, I. Park, T.H. Okabe, Proceedings of 10th World Conference on
Titanium, Ti 2003, Hamburg, July 13–18, 2003, pp. 253–260.
[
9] T.H. Okabe, T. Oda, Y. Mitsuda, J. Alloys Compd. 364 (2003) 156–163.
[
10] G.D. Rigby, I.P. Ratchev, R.I. Olivares, K. Mukunthan, S.A. Bliznyukov,
A.A. Shook, Presentation at 21st Annual ITA Conference, Titanium 2005,
Scottsdale, Arizona, USA, September 25–27, 2005.
[
[
[
11] G.R.B. Elliott, JOM (J. Met.) 50 (September) (1998) 48–49.
12] G. Crowley, Adv. Mater. Processes 161 (November) (2003) 25–27.
13] S. Takaya, M. Tsumura, A. Fuwa, E. Fukasawa, N. Nakahara, Abstracts
of the Mining and Materials Processing Institute of Japan (MMIJ) 2005
Spring Meeting II, 2005, pp. 87–88.
4
. Conclusions
[14] A. Fuwa, S. Takaya, JOM (J. Met.) 57 (October) (2005) 56–60.
[
[
[
15] M.V. Ginnata, Proceedings of 10th World Conference on Titanium, Ti 2003,
Hamburg, July 13–18, 2003, pp. 237–244.
16] O. Takeda, T.H. Okabe, Abstracts of EUCHEM 2004 Molten Salts Con-
ference, Poland, June 20–25, 2004, p. 186.
17] O. Takeda, T.H. Okabe, Abstracts of the Mining and Materials Processing
Institute of Japan (MMIJ) 2004 Autumn Meeting II, 2004, pp. 329–330.
A fundamental study on new synthetic and enrichment
processes for titanium subchlorides (TiClx, x = 2, 3) at
273 K was carried out in order to establish a new high-
speed/semi-continuous titanium production process based on
1

O. Takeda, T.H. Okabe / Journal of Alloys and Compounds 457 (2008) 376–383
383
[
18] O. Takeda, T.H. Okabe, Proceedings of 2005 TMS Annual Meeting, San
Francisco, California, February 13–17, 2005, pp. 1139–1144.
19] O. Takeda, T.H. Okabe, Mater. Trans. 47 (2006) 1145–1154.
20] O. Takeda, T.H. Okabe, Metall. Mater. Trans. 37B (2006) 823–830.
21] W.C. Schumb, R.F. Sundstrom, J. Am. Chem. Soc. 55 (1933) 596–604,
February.
[24] N. Okudaira, Titanium Zirconium 12 (1963) 179–200.
[25] T. Uda, T.H. Okabe, Y. Waseda, K.T. Jacob, Metall. Mater. Trans. B 31
(2000) 713–721.
[26] I. Barin, Thermochemical Data of Pure Substances, VCH Verlagsge-
sellschaft mbH, Weinheim, Germany, 1989.
[27] K. Komarek, P. Herasymenko, J. Electrochem. Soc. 105 (1958) 210–215.
[28] M.V. Smirnov, V.S. Maksimov, Electrochem. Mol. Sol. Electorolytes 7
(1969) 37–41.
[
[
[
[
[
22] K. Yokoyama, J. Mining Metall. Insti. Jpn. 84 (1968) 958–967.
23] W.F. Kriev, S.P. Vango, D.M. Mason, J. Chem. Phys. 25 (1956) 519–523.