Production of titanium powder directly from TiO2 in CaCl 2 through an electronically mediated reaction (EMR)

Article abstract of DOI:10.1016/j.jpcs.2004.06.052

Production of titanium powder directly from titanium dioxide (TiO ₂) through an electronically mediated reaction (EMR) using calcium as a reductant was investigated. Feed material (TiO₂ powder or preform) and reductant (Ca-Ni alloy) were charged in electronically isolated locations in a molten calcium chloride (CaCl₂) salt at 1173 K, and the current flow through an external circuit between the feed and reductant locations was monitored during reduction of TiO₂. The voltage between feed electrode and reductant alloy was intermittently measured during the reduction experiment to monitor the reduction process. After the reduction experiment, pure titanium powder with low nickel content was obtained even though liquid Ca-Ni alloy was used as a reductant. These results demonstrate that titanium metal powder can be produced without direct physical contact between the feed and reductant. In some experimental conditions, pure titanium powder with 99.5 mass% purity was successfully obtained.

Full text of DOI:10.1016/j.jpcs.2004.06.052

Journal of Physics and Chemistry of Solids 66 (2005) 410–413
www.elsevier.com/locate/jpcs
Production of titanium powder directly from TiO2 in CaCl2 through
an electronically mediated reaction (EMR)
Il Parkb, Takashi Abikoc,1, Toru H. Okabea, ,2
*
aInstitute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan
bKorea Science and Engineering Foundation (KOSEF), and Institute of Industrial Science, University of Tokyo,
4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan
cDepartment of Materials Engineering, Graduate School of Engineering, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan
Accepted 4 June 2004
Abstract
Production of titanium powder directly from titanium dioxide (TiO2) through an electronically mediated reaction (EMR) using calcium as
a reductant was investigated. Feed material (TiO2 powder or preform) and reductant (Ca–Ni alloy) were charged in electronically isolated
locations in a molten calcium chloride (CaCl2) salt at 1173 K, and the current flow through an external circuit between the feed and reductant
locations was monitored during reduction of TiO2. The voltage between feed electrode and reductant alloy was intermittently measured
during the reduction experiment to monitor the reduction process. After the reduction experiment, pure titanium powder with low nickel
content was obtained even though liquid Ca–Ni alloy was used as a reductant. These results demonstrate that titanium metal powder can be
produced without direct physical contact between the feed and reductant. In some experimental conditions, pure titanium powder with 99.5
mass% purity was successfully obtained.
q 2004 Published by Elsevier Ltd.
Keywords: Titanium; Titanium dioxide; Metallothermic reaction; Reduction process; Molten salt electrolysis; Metal powder
1. Introduction
many researchers. Although various types of titanium
reduction processes were studied in the past, none of them
could be used for industrial purposes. The key factors in
developing a commercial process, which can compete with
the Kroll process, are the establishment of technology for
effective purity control and/or speeding up the reduction
process.
At present, titanium is industrially manufactured by the
Kroll process [1], which is based on the magnesiothermic
reduction of titanium chloride. Although the Kroll process is
suitable for producing high-purity titanium, it is a labor-
intensive, batch-type process with low productivity. Even
when large-scale modern facilities are used [2], It takes over
10 days to produce 10 tons of sponge titanium. With this
background, a new efficient reduction process that can serve
as an alternative to the Kroll process is being investigated by
Fray et al. have recently been trying to develop a new
industrial reduction process, called the FFC Cambridge
process, based on the direct electrochemical reduction of
TiO2 to metallic titanium in molten CaCl2 salt [3]. Their
extensive work has not only inspired academic research
activity on titanium but has also stimulated the related
industries. As a result, a large variety of titanium
reduction processes are currently under investigation
[4–6]. For example, Ono and Suzuki are engaged in
developing a commercial process, based on the cal-
ciothermic reduction of TiO2 powder in molten CaCl2
salt during electrolysis [4].
* Corresponding author. Tel.: C81 3 5452 6314; fax: C81 3 5452 6313.
E-mail addresses: parkil@iis.u-tokyo.ac.jp (I. Park), krparkil@han-
mir.com (I. Park), abiko@iis.u-tokyo.ac.jp (T. Abiko), okabe@iis.u-
tokyo.ac.jp (T.H. Okabe).
1 Present address: KOBE Steel, Ltd. Tel.: C81 3 5452 6314; fax: C81 3
5452 6313.
2 Tel.: C81 3 5452 6312 (direct), C81 3 5452 6314; fax: C81 3 5452
6313.
0022-3697/$ - see front matter q 2004 Published by Elsevier Ltd.
doi:10.1016/j.jpcs.2004.06.052

I. Park et al. / Journal of Physics and Chemistry of Solids 66 (2005) 410–413
411
In this study, we carried out fundamental research on a
new titanium reduction process with the purpose of
developing new technology for preventing the accumulation
of impurities in titanium deposits using an electronically
mediated reaction (EMR) [7,8]. The EMR is not only
effective in maintaining purity control but is also suitable for
improving energy efficiency when it is combined with the
conventional molten salt electrolysis (MSE) of CaO in
molten CaCl2 salt, for producing reductant calcium alloy.
One of the features of the EMR/MSE process is that the
reduction of TiO2 takes place without any direct contact
with the reductant alloy because titanium is reduced by
electrons released from the reductant. By using the EMR/
MSE process, the reduction of TiO2 and the electric power
input for producing reductant alloy can be carried out
separately. This makes it possible to carry out titanium
reduction during the day and electrolysis for producing
reductant alloy during the night. The possibility of applying
this EMR/MSE process for the reduction of TiO2 was
investigated in this study.
2. Experiment
Fig. 1. Schematic illustration of experimental apparatus for TiO2 reduction
through an EMR.
Experimental conditions of reductions are shown in
Table 1, and a schematic diagram of the reduction apparatus
is shown in Fig. 1. Anhydrous CaCl2 was used as the molten
salt medium, which was dried in a vacuum at 473 K for over
12 h before the experiment. TiO2 powder or a preform
containing TiO2 powder (5w8 g) was kept in a sample
holder and charged in molten CaCl2 salt (1300w1500 g) at
1173 K in an argon atmosphere. The holder containing the
TiO2 powder was placed in the molten salt to prevent it from
making physical contact with the Ca–18 mass% Ni
reductant alloy (50w90 g). TiO2 was thus electrochemi-
cally reduced using electrons discharged from the reductant
alloy. The current flow (iA/C) through an external circuit
between TiO2 and the Ca–Ni alloy was monitored using a
standard resistance of 1 mU. In order to monitor the driving
force of the chemical reaction, the voltage (DEA/C) between
the feed electrode and reductant alloy was intermittently
measured by opening the external circuit during the
reduction experiment.
for 24 h to dissolve the CaCl2. The titanium powder
obtained after the experiment was recovered by leaching it
with acetic acid and hydrochloric acid. The metal powder
obtained was then rinsed with distilled water, alcohol, and
acetone, and finally dried in a vacuum. The morphology of
the metal powder was observed by a scanning electron
microscopy (SEM). Phases in the sample were identified by
an X-ray diffraction analysis (XRD). The composition of the
sample was determined using an energy dispersive X-ray
spectroscopy (EDS), and an inductively coupled plasma-
atomic emission spectroscopy (ICP-AES). The oxygen
concentration of the powder deposit was determined by an
inert gas fusion-infrared absorption spectroscopy.
3. Results and discussion
As shown in Fig. 2(a), an external current (iA/C) of
0.2w0.4 A was observed between the anode and cathode for
the first 500 s. Time zero corresponds to the moment when
the feed TiO2 powder in the sample holder (cathode) was
charged into the molten CaCl2 salt. The reductant Ca–Ni
alloy (anode) was also charged into the molten salt at the
same time. The external current, iA/C, decreased to below
0.1 A after 3 ks. When the external current fell below
0.03 A, the sample holder was taken out of the molten CaCl2
salt. In most cases, the reduction was completed within
7.2 ks (2 h). The value of the electric charge (Qexp) that
passed through the external path was calculated by means of
time integration of the observed external current, and it was
After the reduction experiment, the sample holder was
taken out of the reactor, and was soaked in distilled water
Table 1
Experimental conditions in this study
Exp. no.
Exp. temp. Mass of sample (wi/g)
(T/K)
Form of
feed
CaCl2
Ca–Ni
Feed
molten salt alloy
material
C1
C2
C3
1173
1173
1173
1300
1333
1333
80
72
86
5.02
8.02
7.98
preform
powder
powder

412
I. Park et al. / Journal of Physics and Chemistry of Solids 66 (2005) 410–413
Fig. 2. Variation of the external current and electromotive force. (Exp.
no. C1).
Fig. 3. Titanium powder obtained after reduction. (a) SEM image. (Exp. no.
C3). (b) X-ray diffraction pattern (Cu Ka). (Exp. no. C3).
about 650 C for the case shown in Fig. 2(a). It was found
that this value was about 3% of the theoretical electric
charge (Qtheo) necessary for reducing 5 g of TiO2 as feed
material. The percentage of the obtained charge in relation
reductant. The detailed mechanism of TiO2 reduction in
molten CaCl2 using an EMR is currently under
investigation.
to the theoretical charge, EMR ratio (XEMR (Z100!Qexp
/
To understand the local reaction, a three-dimensional
(3D) chemical potential diagram of the Ti–Ca–O system at
1173 K was constructed, and is shown in Fig. 4(a) [9,10]. As
the Ca–Ni alloy was used in the reduction experiment, the
activity of Ca was not unity in theory, therefore, the stability
plane of Ca (l) in Fig. 4(a) shifts to lower activity region.
But this shift can be neglected considering the scale of the
diagram. The diagram provides an insight into the phase
relationship during the EMR. Point ‘a’ and ‘b’ in Fig. 4(a),
corresponding to Ti/Ca/CaO and Ti/TiO/CaO equilibrium
can be connected through the metal/molten salt interface
Qtheo)), varied between 1w9% depending upon experimen-
tal conditions. Fig. 2(b) shows the change in the voltage
between the anode and cathode obtained by opening an
external circuit intermittently. Until 2 ks after beginning the
reduction experiment, the voltage was above 0.05 V, and
DEA/C decreased with time. The theoretical electromotive
force of the reaction, DEtheo, calculated from the standard
Gibbs energy change (DG8r ZK240:6 kJ [9]) in the reaction,
TiO2(s)C2 Ca (l)/Ti (s)C2 CaO (s), was 0.62 V at
1173 K. The value of DEA/C observed during the reduction
experiment was substantially smaller than that of DEtheo
.
The reasons for the values of EMR ratio and DEA/C are
smaller than the theoretical values were not clear at this
stage, but it is possible that the dissolution of calcium in the
molten CaCl2 salt caused an electric conduction through the
CaCl2. Fig. 3(a) and (b) show the SEM image and XRD
pattern of the obtained titanium powder, respectively. It was
found that pure titanium powder was produced in the sample
holder after the reduction experiment. The analytical results
of the obtained titanium powder and EMR ratio are listed in
Table 2. Titanium with high purity of 99.5% or more was
obtained using an EMR. In a certain experimental condition,
titanium powder with 3500 mass ppm of oxygen was
successfully obtained. This result clearly shows that the feed
material (TiO2) was reduced without any physical contact
with the reductant (Ca–Ni alloy). It is worth noting that
the nickel content in the titanium was quite low at just 0.2
mass%, even though a Ca–Ni alloy was used as the
Table 2
Analytical results of the titanium powder obtained using an EMR and EMR
ratio at 1173 K
Exp.
no.
Composition, Cia (mass%)
EMR ratio,
b
XEMR (%)
Ti
Ca
Fe
Ni
Cl
O
C1
C2
95.6
99.9
4.16
0.10
0.05
0.00
0.16
0.00
0.06
0.04
(0.00) 2.8
(0.00) 3.6
0.57d
0.16c 0.27c 0.05c
C3
99.9
0.01
0.00
0.00
0.00
(0.00) 5.1
0.35d
0.15c 0.20c 0.04c
a
Determined by EDS analysis. Values in parenthesis are listed for
reference.
b
XEMRZ100!Qexp/Qtheo, where Qexp is amount of electric charge
passed through external circuit, and Qtheo is the theoretical electric charge
necessary for reducing TiO2.
c
Determined by ICP-AES analysis.
d
Determined by inert gas fusion-infrared absorption spectroscopy
analysis.

I. Park et al. / Journal of Physics and Chemistry of Solids 66 (2005) 410–413
413
the reduction process. The reduction of TiO2 proceeds
without direct physical contact with the reductant Ca–Ni
alloy, and contamination of titanium can effectively be
prevented by utilizing the EMR. This approach may be
applied to develop a (semi-)continuous production of
titanium powder.
Acknowledgements
The authors are grateful to Prof. Masafumi Maeda and
Prof. Yoshitaka Mitsuda of University of Tokyo, and Mr S.
Kosemura of Toho Titanium Co., Ltd for their support
during this study. Thanks are due to Toho Titanium Co., Ltd
for sample preparation and analysis, and Messrs Itaru
Maebashi, Yoshihiro Kayanuma, and Shusuke Iwata for
their experimental assistance. A part of this research is
financially supported by TEPCO Research Foundation, and
it also includes the results of the program, “Development of
Electrochemical Reduction Process for Oxide Nuclear
Fuel,” conducted by the Central Research Institute of the
Electric Power Industry, which is financially supported by
the Ministry of Education, Culture, Sports, Science and
Technology (MEXT). One of the authors (I.P.) was
supported by the Postdoctoral Fellowship Program of the
Korea Science and Engineering Foundation (KOSEF).
Fig. 4. (a) Three-dimensional chemical potential diagram of the Ti–Ca–O
system at 1173 K. (b) Depiction of the reaction pathway consistent with the
chemical potential diagram.
(solid two-phase equilibrium shown by the line between
point a and b). Fig. 4(b) is a schematic representation of the
EMR pathway, consistent with the chemical potential
diagram. The metal/molten salt interface, which corre-
sponds to the line connecting points a and b in Fig. 4(a),
connects the two three-phase fields. In this situation, the
reductant calcium need not necessarily make physical
contact with the titanium oxides; rather electron and ion
transfer can be considered to be the major step especially in
the process with molten CaCl2.
References
[1] W. Kroll, Tr. Electrochem. Soc. 78 (1940) 35–47.
[2] A. Moriya, A. Kanai, Journal of Mining and Materials Institute of
Japan 109 (1993) 1164–1169.
[3] Z. Chen, D.J. Fray, T.W. Farthing, Nature 407 (2000) 361–364.
[4] K. Ono, R.O. Suzuki, Journal of Metals 54 (2002) 59–61.
[5] T.H. Okabe, Abstracts of EUCHEM /Molten Salts Conference, Royal
Society of Chemistry, Oxford, 2002, p. PL2.
[6] T.H. Okabe, T. Uda, Titanium Japan 4 (2002) 325–330.
[7] T.H. Okabe, D.R. Sadoway, Journal of Materials Research 12 (1998)
3372–3377.
[8] T.H. Okabe, I. Park, K.T. Jacob, Y. Waseda, Journal of Alloys and
Compounds 288 (1999) 200–210.
4. Conclusion
[9] I. Barin,, Third ed. Thermochemical Date of Pure Substances, VCH
Verlagsgesellschaft mbH, Weinheim, Federal Republic of Germany,
1995.
The experimental results for calciothermic reduction of
TiO2 in molten CaCl2 salt showed that the electronically
mediated reaction (EMR) plays an important role in
[10] P. Wanlder, G. Eriksson, Calphad 2 (1999) 189–218.