Titanium powder production by preform reduction process (PRP)

Article abstract of DOI:10.1016/S0925-8388(03)00610-8

To develop an effective process for titanium powder production, a new preform reduction process (PRP), based on the calciothermic reduction of preform containing titanium oxide (TiO₂), was investigated. The feed preform was fabricated from slurry, which was made by mixing TiO₂ powder, flux (e.g. CaCl₂) and binder. Various types of preforms in the form of plates, spheres, or tubes were prepared using a conventional technique, and the fabricated preform was sintered at 1073 K before reduction in order to remove the binder and water. The sintered solid preform containing TiO₂ was then placed in a stainless steel container, and reacted with calcium vapor at a constant temperature ranging between 1073 and 1273 K for 6 h. Titanium powder was recovered from the reduced preform by leaching it with acid. As a result, pure titanium powder with 99 mass% purity was obtained. This process was found to be suitable for producing a homogeneous fine powder when the composition of flux and the size of the preform are controlled.

Full text of DOI:10.1016/S0925-8388(03)00610-8

Journal of Alloys and Compounds 364 (2004) 156–163
Titanium powder production by preform reduction process (PRP)
Toru H. Okabea,∗, Takashi Odab, Yoshitaka Mitsudaa
a
Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo, 153-8505, Japan
b
Department of Metallurgical Engineering, Chiba Institute of Technology, 2-17-1 Tsudanuma, Narashino-city, Chiba, 275-0016, Japan
Received 12 May 2003; accepted 27 May 2003
Abstract
To develop an effective process for titanium powder production, a new preform reduction process (PRP), based on the calciothermic
reduction of preform containing titanium oxide (TiO2), was investigated. The feed preform was fabricated from slurry, which was made by
mixing TiO2 powder, flux (e.g. CaCl2) and binder. Various types of preforms in the form of plates, spheres, or tubes were prepared using a
conventional technique, and the fabricated preform was sintered at 1073 K before reduction in order to remove the binder and water. The
sintered solid preform containing TiO2 was then placed in a stainless steel container, and reacted with calcium vapor at a constant temperature
ranging between 1073 and 1273 K for 6 h. Titanium powder was recovered from the reduced preform by leaching it with acid. As a result,
pure titanium powder with 99 mass% purity was obtained. This process was found to be suitable for producing a homogeneous fine powder
when the composition of flux and the size of the preform are controlled.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Reactive metal; Metallothermic reduction; Metal powder
1. Introduction
purity titanium in molten calcium chloride (CaCl2) [3–6]. As
schematically shown in Fig. 1(b), the FFC process, devel-
oped by Chen et al. [3], is being carried out on a larger scale.
Their activity not only inspired academic materials research,
but has also energized related industries all over the world.
Currently, Ono and Suzuki are also developing a commer-
cial process for titanium production by means of the cal-
ciothermic reduction of TiO2 in molten CaCl2 [7–9]. Their
process, called the OS process, also involves the electrolysis
of molten CaCl2–CaO [7]. The features of the FFC [3] and
OS [8] processes and the differences between them are de-
scribed elsewhere [10]. The application of an electronically
mediated reaction (EMR) [11,12] to the titanium reduction
process is also under investigation [12]. This method has the
potential to prevent impurities from accumulating in metal
deposits [13], and can increase energy efficiency when com-
bined with conventional molten salt electrolysis (MSE) of
CaCl2–CaO molten salt for the production of calcium alloy
reductants.
The application of titanium has been widening and the
demand for titanium has also been increasing in recent years
as it is a lightweight, extremely strong, and highly corrosion
resistant metal. Titanium is currently produced by means of
a metallothermic reduction process called the Kroll process
[1]. In this process, titanium tetrachloride (TiCl4) is reduced
by metal magnesium in an airtight metallic reactor at about
1100 K (Fig. 1(a)). The titanium deposit sticks to the reactor
wall and grows in the form of massive sponge. As a result, it
is rather difficult to recover product titanium. No continuous
commercial process has been successfully established so far;
and currently batch-type processes are employed for mass
production. The chemical reaction during the Kroll process
is a huge exothermic reduction, and producing titanium on
a large scale takes several days [2].
To overcome the drawback of the current Kroll process,
Fray et al. conducted extensive research on the direct electro-
chemical reduction of titanium dioxide (TiO2) to obtain high
The direct reduction processes of titanium oxide men-
tioned above have the potential to be the next generation
titanium production processes. However, there are some
drawbacks which have to be overcome before the process
can be commercialized. Specifically, a large amount of
molten salt is required, and the reduction process is slow.
∗
Corresponding author. Tel.: +81-3-5452-6312/14;
fax: +81-3-5452-6313.
E-mail addresses: okabe@iis.u-tokyo.ac.jp (T.H. Okabe),
mitsuda@iis.u-tokyo.ac.jp (Y. Mitsuda).
0925-8388/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0925-8388(03)00610-8

T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
157
This calciothermic reduction of TiO2 is a huge exothermic
reduction process, and in general it is difficult to achieve a
homogeneous reaction. When the reductant calcium comes
into physical contact with TiO2 and reacts with it, the im-
purities in the reductant, such as nitrogen and carbon, are
transferred to the metal phase, and contaminate the titanium
metal. To avoid this problem, calcium vapor is often used
to reduce TiO2 [14,15]. A vapor phase metallothermic re-
duction, as schematically shown in Fig. 1(c), is effective in
preventing contamination from reductant, and it is possible
to produce titanium powder with purity greater than 99%.
However, this process is still very sensitive to contamina-
tion from reactor material, as the TiO2 feed is placed on the
substrate material. Furthermore, it is difficult to scale up the
process because it is difficult to facilitate a homogeneous
reaction when reducing large amounts of feed material in a
single reactor.
To develop an effective process for the titanium powder
production, a new reduction process using feed preform
was investigated in this study. As schematically shown
in Fig. 1(d), this process is based on the metallothermic
reduction of preform, which contains feed material [16].
Advantages of this process are its effective control of purity
and morphology, and its flexible scalability. Contamina-
tion from the reaction container or reductant can easily be
avoided because the feed material in the self-supporting
preform does not come into physical contact with them.
The process is highly flexible in terms of scalability as it is
possible to simultaneously treat multiple pieces of preform
in a single reduction chamber. Furthermore, the amount
of chemicals such as CaCl2 can be reduced, compared to
the direct electrochemical reduction processes mentioned
before.
3. Experimental
Fig. 2 shows the material flowchart for the preform re-
duction process using calcium vapor as a reductant that was
investigated in this study. The experiment consists of three
major steps: preform fabrication, reduction by calcium va-
por, and titanium powder recovery by leaching.
Fig. 1. Various titanium reduction processes.
Furthermore, the process lacks the flexibility to be scaled up,
as it is based on electrochemical reduction. With this back-
ground, a new reduction process for producing fine metal
powder by means of the metallothermic reduction of pre-
form containing feed material was investigated in this study.
The feed preform was fabricated from slurry, which was
made by mixing TiO2 powder, flux (e.g. CaCl2) and a binder.
The purity and form of the starting materials used in this
study are listed in Table 1. About 3–15 g of TiO2 powder
(99.97% purity, Al: < 0.001%, Fe: < 0.001%, Cl: < 0.011%
by mass, average particle size: 0.2 ␮m) was mixed with the
flux and binder using a stirrer, and a homogeneous slurry was
synthesized. About 0.5–6 g of CaCl2, or CaO powder was
employed as flux. A collodion solution, which is a mixture
of 5 mass% nitrocellulose in ethanol and ether, was used as a
binder solution. The viscosity of the slurry was controlled by
varying the amount of flux and binder. The obtained slurry
was then cast into a stainless steel mold and a preform plate
of 3–10 mm thickness was produced. In some experiments,
2. Preform reduction process (PRP)
When producing titanium by means of the metallothermic
reduction of titanium oxide, the feed (TiO2) and reductant
(Ca) are often mixed and heated to elevated temperatures.
The overall reaction for this process is
TiO2 + 2Ca = Ti + 2CaO
(1)

158
T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
Fig. 3. Schematic illustration of the experimental apparatus for producing
titanium powder by means of the preform reduction process (PRP).
supplied to the feed preform at unit activity through the
gas phase while heating. As shown in Fig. 3, four to ten
pieces of preform were installed in a thick-walled stainless
steel reaction vessel, and were sealed by tungsten inert
gas (TIG) welding. A sponge titanium was also placed at
the bottom of the vessel for gettering nitrogen gas in the
system. The sealed reaction vessel was then heated in an
electric furnace maintained at a constant temperature rang-
ing between 1073 and 1273 K, and the preform was reacted
with calcium vapor. Some representative experimental con-
ditions of the reduction experiments are summarized and
listed in Table 2. After 6 h of reaction, the reaction vessel
Fig. 2. Experimental procedures of titanium powder production by means
of calciothermic reduction of oxides using a preform reduction process
(PRP).
preform in a tubular shape or a spherical shape, 5–20 mm
in diameter, was fabricated using slurry with high viscosity.
Various types of preforms in the form of plates, balls, or
tubes were prepared, and the fabricated preform was heated
at 1073 K before reduction in order to remove the binder
and water.
Fig. 3 is a schematic illustration of the experimental
apparatus for producing titanium powder by means of
the PRP. The sintered solid preform containing TiO2 was
placed in a stainless steel vessel, and about 20 g of calcium
shots were placed at the bottom. The calcium was phys-
ically isolated from the feed preform, and its vapor was
Table 2
Experimental condition of the preform reduction processa
Exp.
no.
Mass of sample wi (g)
Cationic
molar ratio temperature
Reduction
Feed Flux
Binder
b
RCat.
Tred. (K)
/Ti
TiO2 CaO CaCl2 Collodion
A
B
C
4.99
3.93
3.11
0.44
1.60
2.10
–
–
–
6.02
6.32
5.50
0.1
0.5
1
1273
1273
1273
D
E
F
5.40
4.11
3.05
–
–
–
0.47
1.97
2.83
10.50
–
11.15
0.1
0.5
1
1273
1273
1273
Table 1
Starting materials used in this study
G
5.00
–
–
5.00
0
1273
Materials
Form
Purity or
conc. (%)
Supplier
H
I
J
3.98
3.98
3.98
1.40
1.40
1.40
–
–
–
4.63
4.63
4.63
0.5
0.5
0.5
1073
1173
1273
TiO2
CaO
CaCl2
Collodiona
Ca
Powder
Powder
Powder
Solution
Chip
Sponge
Aqueous
Aqueous
99.97
Toho Titanium
99.5 up
99.5 up
5.0b
98.0 up
98.0 up
99.7
35.0b
Kanto Chemicals
Kanto Chemicals
Wako Pure Chemical Industries
Mintech Japan
Toho Titanium
Kanto Chemicals
Kanto Chemicals
K
L
M
3.97
3.97
3.97
–
–
–
1.89
1.89
1.89
6.85
6.85
6.85
0.5
0.5
0.5
1073
1173
1273
N
O
P
3.97
3.97
3.97
–
–
–
–
–
–
4.85
4.85
4.85
0
0
0
1073
1173
1273
Ti
CH3COOH
HCl
a
a
Preform was calcined at 1073 K for 1 h and reacted with Ca for 6 h.
Cationic molar ratio, RCat./Ti = NCat./NTi, where NCat. and NTi are
5 mass% nitrocellulose, 23.75 mass% ethanol, 71.25 mass% di-
b
ethylether solution.
b
the mole amounts of the cation in the flux and titanium, respectively.
Concentration of the solution.

T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
159
was taken out of the furnace, and quenched in water. The
preforms in the container were mechanically recovered at
room temperature, and subjected to the following leaching
process.
The titanium powder in the preform obtained after the re-
duction experiment was recovered by leaching the preform
with acid. The reaction product (CaO), flux, and excess re-
ductant in the sample were removed by dissolution in acetic
acid solution. The obtained metal powder was rinsed in a
hydrochloric acid at room temperature. It was then rinsed
with distilled water, alcohol, and acetone, and then dried in
a vacuum.
A chemical analysis of the sample was carried out using an
inductively coupled plasma-atomic emission spectroscopy
(ICP-AES; Seiko Instruments, SPS 4000). The metal powder
was observed using scanning electron microscopy (SEM,
Jeol, JSM-5600LV), and the composition of the sample was
determined both by an energy dispersive X-ray spectrom-
eter (EDS, Jeol, JED-2200) and ICP-AES. Phases in the
sample were identified using an X-ray diffraction analysis
Fig. 5. X-ray diffraction patterns of the samples. (a) Fabricated platy
feed preform (flux, CaO; binder: Collodion). (b) Preform after calcination
(temperature, 1073 K; holding time, 1 h). (c) Preform after reduction
(temperature, 1273 K; holding time, 6 h). (d) Titanium powder obtained
after leaching (exp. no. C, Ti, 99.88%; Fe, 0.10%; Ni, 0.02%; analyzed
by EDS).
Fig. 4. (a) Fabricated platy feed preform (flux, CaCl2; binder, Collodion).
(b) Preform after calcination (temperature: 1073 K, holding time: 1 h).
(c) Preform after reduction (temperature, 1273 K; holding time, 6 h). (d)
Obtained titanium powder after leaching (exp. no. M, Ti, 99.88%; Fe,
0.10%; Ni, 0.02%; analyzed by EDS).

160
T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
Table 3
Analytical results of the titanium powder obtained by means of the preform reduction process
Exp.
no.
Flux
Cationic
Reduction
temperature
Tred. (K)
Impurity content, Cib (mass% )
molar ratio
Titanium
CTi
Calcium
CCa
Oxygen
CO
Iron
CFe
Nickel
CNi
Chromium
CCr
a
RCat./Ti
A
B
C
CaO
CaO
CaO
0.1
0.5
1
1273
1273
1273
99.66
99.62
99.80
0.33
0.11
0.03
0.03c
(0.00)
(0.00)
(0.00)
0.28d
0.00
0.00
0.15
0.17
0.09c
0.01
0.00
0.00
0.66c
0.12
0.23
0.10c
D
E
F
CaCl2
CaCl2
CaCl2
0.1
0.5
1
1273
1273
1273
97.75
99.17
99.84
1.07
0.38
0.07
(1.03)
(0.00)
(0.00)
0.00
0.00
0.09
0.14
0.33
0.00
0.00
0.11
0.01
G
No flux
0
1273
97.11
1.71
(1.08)
0.11
0.00
0.00
H
I
J
CaO
CaO
CaO
0.5
0.5
0.5
1073
1173
1273
87.15
88.73
99.22
2.94
7.61
0.51
(9.69)
(3.41)
0.00
0.06
0.00
0.23
0.00
0.24
0.04
0.17
0.00
0.03
K
L
CaCl2
CaCl2
0.5
0.5
1073
1173
88.79
99.92
6.40
0.02
(4.62)
(0.00)
0.66d
(0.00)
0.57d
0.16
0.00
0.00
0.06
0.01c
0.02
0.02c
0.03
0.00
0.04c
0.00
0.01c
0.10
< 0.01c
0.00
M
CaCl2
0.5
1273
99.88
< 0.01c
< 0.01c
< 0.01c
N
O
P
No flux
No flux
No flux
0
0
0
1073
1173
1273
82.48
78.47
86.41
0.68
0.16
9.31
(16.32)
(21.20)
(3.94)
0.52
0.07
0.17
0.00
0.00
0.00
0.00
0.10
0.16
a
b
c
d
Cationic molar ratio, RCat./Ti = NCat./NTi, where NCat. and NTi are the mole amounts of cation in the flux and titanium, respectively.
Analyzed by EDS. Values in parenthesis are listed for reference.
Determined by ICP analysis.
Determined by inert gas fusion-infrared absorption spectroscopy (LECO).
(XRD, Rigaku, Rint 2000). Oxygen analysis was carried
out by inert gas fusion–infrared absorption spectroscopy
(LECO).
It was not necessary to pulverize the reduced preform be-
fore the leaching treatment. Titanium powder obtained after
rinsing and drying is shown in Fig. 4(d). The EDS analy-
sis revealed the purity of the obtained titanium powder to
be 99.9 mass% Ti, 0.10 mass% Fe, and 0.02 mass% Ni
(exp. no. M). A summary of the analytical results related
to the powder as well as to the other samples is listed in
Table 3.
4. Results and discussion
By regulating the amount of binder, it was possible to
control the viscosity of the slurry and obtain preform in dif-
ferent shapes. A preform plate, 3–6 mm thick could easily
be produced by casting the slurry into the mold. Fig. 4(a)
shows fabricated platy preform, 50 × 20 × 4 mm in dimen-
sion. This was made by casting slurry that was produced
by mixing TiO2, CaCl2 and collodion solution (exp. no.
M). As shown in Fig. 4(b), the shape of the preform af-
ter calcination at 1073 K for an hour remained unchanged.
Fig. 4(c) shows the preform after reduction at 1273 K for
6 h. The preform became slightly deformed after reduction,
and was covered with metallic calcium. It was easy to re-
cover the preform from the reaction vessel because it is
physically isolated from it. These results indicate that this
preform reduction process is suitable for separating feed
material from reactor material, and it is easy to control
contamination from the reaction vessel. When the preform
obtained after reduction was put into an acetic acid solu-
tion, its shape was lost and titanium powder was obtained.
Fig. 5 shows representative XRD patterns of the samples
obtained during the experiment (exp. no. C). The analysis
revealed that the fabricated feed preform was a mixture of
TiO2 and CaO, and some portion of these phases reacted to
form CaTiO3 after calcination. Fig. 5(c) shows an XRD pat-
tern of the preform after reduction at 1273 K for 6 h. Excess
calcium reductant was found in the mixture of titanium and
CaO after reduction. One can see from Fig. 5(d) that pure
titanium was produced after leaching.
By increasing the amount of flux and using CaCl2 as
a flux, it was possible to synthesize high-viscosity slurry,
and fabricate it to form spherical or tubular-shaped pre-
form. Fig. 6(a) shows fabricated spherical feed preform, and
Fig. 6(b) shows preform that has been calcined at 1073 K
for 1 h. Fig. 6(c) shows the preform that has undergone re-
duction at 1273 K for 6 h. One can see from the sectioned
image [Fig. 6(c) right] that the reduced preform is homo-
geneous within the sensitivity of the optical observation. A

T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
161
Fig. 6. (a) Fabricated spherical feed preform. (Flux, CaCl2; binder, Col-
lodion). (b) Preform after calcination (temperature, 1073 K; holding time,
1 h). (c) Preform after reduction (temperature, 1273 K; holding time, 6
h). (d) Titanium powder obtained after leaching (Ti, 99.70%; Ca, 0.06%;
Fe, 0.10%; Ni, 0.10%; analyzed by EDS).
Fig. 7. (a) Fabricated tubular feed preform. (Flux, CaCl2; binder, Collo-
dion). (b) Preform after calcinations (temperature, 1073 K; holding time,
1 h). (c) Preform after reduction (temperature, 1273 K; holding time, 6
h). (d) Titanium powder obtained after leaching (Ti, 99.93%; Ca, 0.04%;
Ni, 0.03%; analyzed by EDS).
metallic phase was observed at the center of the preform,
which confirms that the reduction was completed in this ex-
perimental condition. The EDS analysis revealed the purity
of the obtained titanium powder after leaching to be 99.7%.
Fig. 7 shows the results for the tubular preform. One can see
from the sectioned preform after reduction [Fig. 7(c) right]
that a uniform cake was obtained after reduction. The EDS
analysis revealed the titanium powder that was obtained af-
ter leaching to be 99.9 mass%Ti, 0.04 mass%Ca, and 0.03
mass%Ni.
Fig. 8 shows SEM images of the titanium powder ob-
tained after reduction at 1273 K for 6 h followed by leach-
ing (exp. no. J, M, P). When flux was not used, the size
of the titanium particles could be measured only in sub-␮m
amounts. However, when flux was used, this size increased
to 1–2 ␮m. Interestingly, it was found that homogeneous ti-
tanium powder was obtained when CaCl2 was used as flux.
Furthermore, the purity of the titanium powder obtained
was more than 99% when flux was used (see exp. no. A–F
in Table 3), whereas it decreased when flux was not used
(exp. no. G and P). The detailed mechanism of the reduction
process and its influence on the morphology of the deposit
are currently under investigation.
As shown in Fig. 9, the particle size of the obtained tita-
nium powder increased as the reaction temperature increased
(exp. no. K, L, and M). Analytical results of the titanium
powder obtained are listed in Table 3. The purity of the
sample was low when the reaction temperature was 1073 K
(exp. no. H, K and N). It was found that the vapor pressure of
calcium at this temperature (pCa = 8.5 × 10−4 atm. at 1073

162
T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
Fig. 9. SEM images of the titanium powder obtained after various re-
duction temperatures (reduction conditions: flux = CaCl2, cationic molar
ratio, XC/Ti = 0.5, reduction temp., T = 1073, 1173, 1273 K, Holding
time, t’ = 6 h; exp. no. K, L, M).
Fig. 8. SEM images of the titanium powder obtained after reduction
using various fluxes. (Reduction conditions: flux = CaO, CaCl2, no flux,
Cationic molar ratio, XC/Ti = 0.5; reduction temp., T = 1273 K; holding
time, t’ = 6 h; exp. no. J, M, P).
K [17]) was too low for reduction by vapor. When the pre-
form containing CaCl2 as flux was exposed to calcium va-
por at temperatures above 1173 K (pCa = 4.4 × 10−3 atm.
at 1173 K [17]), titanium powder with more than 99% purity
by mass was successfully obtained (exp. L and M).
Fig. 10 shows an enlarged SEM image of the titanium
powder obtained after reduction at 1273 K for 6 h using
CaCl2 as flux (exp. no. L). The LECO and ICP chemical
analyses revealed that the oxygen content in the powder
was 0.66 mass%, and the iron and nickel content was 0.01
mass%. In some cases in this study titanium powder with
0.28 mass% oxygen was obtained (see exp. C in Table 3). It
was found that the preform reduction process was effective
in producing titanium powder with 99% purity. This method
can be applied to the production of various metals and al-
loys because it is based on a very simple metallothermic
reduction.
Fig. 10. SEM image of the titanium powder obtained by means of the
preform reduction process (PRP). (Reduction conditions: flux = CaCl2,
cationic molar ratio, XC/Ti = 0.5, reduction temp., T = 1273 K, holding
time, tꢀ = 6 h; exp. no. M).

T.H. Okabe et al. / Journal of Alloys and Compounds 364 (2004) 156–163
163
5. Conclusion
study. Thanks are due to Mr. S. Kosemura of Toho Titanium,
for sample preparation and analysis, and Messers Itaru
Maebashi, Shusuke Iwata and Naoto Sato for experimental
assistance. A part of this research 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 MEXT (Ministry of Education,
Culture, Sports, Science and Technology).
A new reduction process for producing fine titanium pow-
der by the metallothermic reduction of preform containing
feed material was developed. Various types of preforms in
the form of plates, spheres, or tubes were prepared using a
conventional technique and were supplied as feed material
for metallothermic reduction. The slurry obtained by mix-
ing TiO2 powder, flux (e.g. CaCl2, CaO. . . ) and binder (e.g.
collodion), was cast into a mold and then dried to obtain
the preform. Before the reduction process, the fabricated
preform was heated at 1073 K in order to remove the binder
and water present in it. The sintered solid preform con-
taining TiO2 was then placed in a stainless steel container,
and reacted with calcium vapor at constant temperatures
ranging from 1073 to 1273 K for 6 h. The titanium pow-
der was recovered from the reduced preform by leaching it
with acid. As a result, pure titanium powder in sponge mor-
phology with 99 mass% purity was obtained. This process
was found to be suitable for producing fine, homogeneous
powder when the composition of flux and the size of the
preform are controlled. Another advantage of this process
is purity control. Contamination from the reaction container
and reductant, can easily be avoided because the feed mate-
rial in the self-supporting preform is suitable for minimizing
physical contact with the reaction container and reductant.
Furthermore, the developed process is highly scalable as
demonstrated by its ability to simultaneously treat multiple
pieces of preform in a single reduction chamber.
References
[1] W. Kroll, Tr. Electrochem. Soc. 78 (1940) 35–47.
[2] A. Moria, A. Kanai, Shigen-to-Sodai (J. Mining and Materials In-
stitute of Japan) 109 (1993) 1164–1169.
[3] Z. Chen, D.J. Fray, T.W. Farthing, Nature 407 (2000) 361–364.
[4] D.J. Fray, J. Metals 53 (2001) 26–31.
[5] Z. Chen, D.J. Fray, T.W. Farthing, Metall. Mater. Trans. B 32B
(2001) 1041–1052.
[6] D.J. Fray, Can. Metallurg. Quart. 41 (2002) 433–439.
[7] K. Ono, R.O. Suzuki, J. Metals 54 (2002) 59–61.
[8] K. Ono, R.O. Suzuki, Materia (Bull. Jpn. Inst. Metals) 41 (2002)
28–31.
[9] K. Ono, R.O. Suzuki, Titan (Japan Titanium Assoc.) 50 (2002) 105–
108.
[10] T.H. Okabe, T. Uda, Titan (Japan Titanium Assoc.) 50 (2002) 325–
330.
[11] T.H. Okabe, D.R. Sadoway, J. Mater. Res. 12 (1998) 3372–3377.
[12] T.H. Okabe, Y. Waseda, J. Metals 49 (1997) 28–32.
[13] T.H. Okabe, I. Park, K.T. Jacob, Y. Waseda, J. Alloys Comp. 288
(1999) 200–210.
[14] K. Ono, A. Miyazaki, J. Jpn. Inst. Metals 49 (1985) 871.
[15] K. Ono, T.H. Okabe, M. Ogawa, R.O. Suzuki, Testu-to-Hagane (J.
Iron Steel Institute of Japan) 76 (1990) 86–93.
[16] S. Iwata, T.H. Okabe, M. Imagunbai, M. Maeda, in: Proceeding of
Autumn Meeting of Japan Institute of Metals, Osaka, 2–4 November,
2002, p. 397.
Acknowledgements
The authors are grateful to Professor Masafumi Maeda
of the University of Tokyo and Dr. Masana Imagunbai of
CBMM Asia for their support and useful input during this
[17] I. Barin, Thermochemical Data of Pure Substances, 3rd Edition,
VCH, Weinheim, 1995.