International Journal of Mineral Processing 149 (2016) 111–118

Contents lists available at ScienceDirect

International Journal of Mineral Processing

journal homepage: www.elsevier.com/locate/ijminpro

Removal of iron from titanium ore by selective chlorination using TiCl4

under high oxygen chemical potential
Jungshin Kang a,b,⁎,1,2, Toru H. Okabe b
a
Department of Materials Engineering, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
b
Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan

a r t i c l e i n f o a b s t r a c t

Article history: With the purpose of developing a new technology for producing high-grade titanium dioxide (TiO2), a funda-
Received 24 March 2014 mental study on the selective chlorination of low-grade titanium ore under a high oxygen chemical potential
Received in revised form 11 December 2015 was carried out. The use of titanium tetrachloride (TiCl4) as a chlorinating agent allowed iron to be directly
Accepted 24 February 2016
removed from various types of Ti ores. In the experiments, titanium ore was first pre-oxidized under air, and
Available online 2 March 2016
then reacted with TiCl4 gas at 1100 K or 1200 K for 9 h under an Ar + 1 ppm O2, Ar + 1% O2, or Ar + 10% O2
Keywords:
gas flow. The iron in the low-grade Ti ore was selectively removed as iron chlorides (FeClx [x = 2, 3]) in a dry
Selective chlorination form, and 98% TiO2 was obtained under certain conditions. This demonstrated that the selective chlorination of
Titanium ore low-grade titanium ore by TiCl4 under a high oxygen chemical potential is a feasible method of producing
Titanium tetrachloride high-grade TiO2 directly.
Up-grading © 2016 Elsevier B.V. All rights reserved.
Oxidation-selective-chlorination
High oxygen potential

1. Introduction
The Kroll process is used for the industrial production of titanium
(Ti) metal (Zhang et al., 2011). High-grade TiO2 (≥90–95% TiO2) obtain-
ed by upgrading low-grade Ti ore (ilmenite) by removing iron (Fe) is
used as feedstock for the Kroll process. In the first step of the Kroll
process, the upgraded Ti ore is chlorinated by chlorine gas (Cl2) in
the presence of carbon to produce TiCl4. Most impurities are also
chlorinated in this carbo-chlorination process to form metal chlorides
(Iida, 2012; Nakamura, 2013). Therefore, if a Ti ore containing a large
amount of impurities is used, then the chlorine loss increases through
excess Cl2 gas consumption (Okabe and Kang, 2012). Simultaneously,
a large amount of chlorides waste is generated by the chlorination of
impurities, and this can lead to the clogging of pipes in the chlorinator.
In addition, the treatment of chlorides waste is quite costly, particularly
in those countries that have stringent environmental restrictions on the
treatment of industrial waste.
⁎ Corresponding author at: Department of Materials Engineering, Graduate School of
Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan.
E-mail addresses: jskang@iis.u-tokyo.ac.jp, jskang@kigam.re.kr, jskang@ust.ac.kr
(J. Kang).
1
Currently, Senior Researcher at Korea Institute of Geoscience and Mineral Resources,
124 Gwahang-no, Yuseong-gu, Daejeon 305-350, Korea.
2
Currently, Assistant Professor at University of Science and Technology, 217,
Gajeong-ro Yuseong-gu, Daejeon 305-350, Korea.
For these reasons, most Ti metal production companies in Japan use
a TiO2 feed with a purity of more than 95% (purities greater than 90% are
used worldwide) (Okabe and Kang, 2012). In recent years, the high cost
of high-grade TiO2 feed has led to some companies in Japan starting to
use low-cost TiO2 feed with a purity of more than 90% (Japan Titanium
Society, 2013). However, such use is limited, owing to the increases
in chlorine loss, pipe clogging, and damage to refractory materials in
the chlorinator.
The industrial processes currently used for upgrading Ti ore are the
Becher (Becher et al., 1965; Hoecker, 1994), the Benilite (Chen and
Huntoon, 1977; Chen, 1976; Chen, 1974), and the slag and UGS process-
es (Guéguin and Cardarelli, 2007; Borowiec et al., 1998). Among these
processes, the slag and UGS processes are the most important when
the production output is taken into consideration. For example, in
2003, 1.4 Mt of TiO2 (titania slag) and 0.2 Mt of TiO2 (upgraded slag)
are produced by the slag and UGS processes, respectively, whereas
only 0.8 Mt of TiO2 (synthetic rutile) was produced by other processes
(Kogel et al., 2006). These production outputs were calculated by multi-
plying the production of feedstock by its proportion of contained TiO2.
As can be seen in the representative flow chart for the slag and UGS
processes in Fig. 1 (a), the processing necessary to remove Fe from Ti ore
is quite complex and time-consuming. One reason for this complexity is
the need to prevent the loss of Ti from the ore, yet also enhance the
removal of Fe during the upgrading process. This is achieved by
repeating the oxidation and reduction reactions. The slag process also
consumes a large amount of energy, because the reaction temperature

http://dx.doi.org/10.1016/j.minpro.2016.02.014
0301-7516/© 2016 Elsevier B.V. All rights reserved.
112 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118
Fig. 1. (a) Flow chart of the slag and UGS processes (Guéguin and Cardarelli, 2007; Borowiec et al., 1998). (b) Flow chart for new titanium smelting process adopting the selective
chlorination process investigated in this study.
for reducing iron oxides is very high at around 1923–1973 K (Kogel
et al., 2006; Filippou and Hudon, 2009). The advantage of the slag
process, however, is its speed and waste management, with the Fe
being directly removed as pig-iron (Fe–C alloy) that can be used in the
iron industry.
The purity of the TiO2 feed produced by the slag process is 75–86%,
which is low for it to be used as a feedstock for the Kroll process.
Therefore, the UGS process uses concentrated acid such as 18–20%
hydrochloric acid (HCl) to further remove Fe and other impurities as
metal chlorides (e.g. FeCl2 and MnCl2) or oxychlorides (Guéguin and
Cardarelli, 2007). The downside of this process is the generation of the
aqueous acid waste solution by the use of concentrated acid, which
can be costly to treat in countries with strict environmental regulations
(Wako, 2012).
To resolve these problems, extensive research has been conducted in
the past. One representative method resulting from this is selective
chlorination through pyrometallurgical treatment. For example, selec-
tive chlorination of Ti ore was carried out using Cl2 gas in the presence
of carbon (Rhee and Sohn, 1990a; Fukushima and Kimura, 1975;
Kimura et al., 1979; Fuwa et al., 1978), under a mixed CO and Cl2 atmo-
sphere (Rhee and Sohn, 1990b; Rhee and Sohn, 1990c; Doraiswamy
et al., 1959), using metal chlorides such as CaCl2 or magnesium chloride
(MgCl2) as the chlorine source (Zheng and Okabe, 2006; Matsuoka and
Okabe, 2005; Kang and Okabe, 2013a; Kang and Okabe, 2013b; Kang
and Okabe, 2014), or using Cl2 in the presence of a low-valence
titanium oxide (TiOx, 0 b x b 2) (Ichimura et al., 1994). However, the
direct use of Cl2 gas poses problems from the viewpoint of safety and
environmental concerns. In addition, the use of metal chlorides such
as CaCl2 or MgCl2 also causes problems such as damage to the interior
of the chlorinator, because these metal chlorides are highly corrosive
at high temperatures.
Among the various potential processes, selective chlorination of Ti
ore using TiCl4 was suggested by Othmer (Othmer, 1975; Othmer and
Nowak, 1972). First, Othmer claimed Ti ore containing Fe2O3 could be
selectively chlorinated by TiCl4 under an oxidative atmosphere through
a process called oxidation-selective-chlorination. However, to the best
of the authors' knowledge, experimental results on this process cannot
be found in the literature. Second, Othmer claimed to have carried out
the selective chlorination of Ti ore containing FeO by TiCl4 in the
presence of carbon (Othmer, 1975). In all cases, systematic thermody-
namic analysis considering the chemical potentials of oxygen (pO2)
and chlorine (pCl2) was not conducted.
In this study, a thermodynamic analysis of the selective chlorination
of Ti ore using TiCl4 under a high pO2 was carried out in order to
establish a new process for the selective removal of Fe. The selective
chlorination method used in this study is illustrated in Fig. 1 (b) and
 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118
offers the following advantages. (i) High-grade TiO2 feed can be obtain-
ed directly from various types of Ti ores using a simple pyrometallurgi-
cal method in a single step. (ii) The TiCl4 needed is easy to obtain, as it is
circulated in large amounts in the Kroll process. (iii) There is no acid
waste solution that needs to be dumped out, because concentrated
acid is not used. (iv) Chlorine recovery is possible, because chloride
waste such as FeClx (x = 2, 3) is generated in a dry form.
2. Experimental
Fig. 2 shows a schematic diagram of the experimental apparatus
used in this study, and Table 1 lists the experimental conditions along
with the analytical results for the feedstocks and residues. All samples
used were oxidized at 1100 K for 13 h under an air atmosphere in
advance to increase their oxygen chemical potential. Before the experi-
ments were conducted, the oxidized Ti ore was placed in a quartz cruci-
ble (ϕ = 25 mm, I.D.; d = 120 mm, depth). The quartz crucible was
positioned in a quartz tube (ϕ = 44.5 mm; l = 820 mm, length),
which was plugged with a Viton rubber plug.
Once the samples were prepared, the atmosphere was controlled as
follows. The quartz tube was evacuated for 5 min, and then filled with
an experimental atmosphere of Ar + 1 ppm O2, Ar + 1% O2, or
Ar + 10% O2 until the internal pressure of the tube reached 1 atm.
This procedure was performed three times. After the quartz tube was fi-
nally filled, the filling gas was allowed to flow through the quartz tube
to maintain an internal pressure of 1 atm. All the gas was supplied
using a mass flow controller (MFC) after passing through magnesium
perchlorate (Mg(ClO4)2, Kanto Chemical Co., Inc., purity ≥75%).
After the experimental atmosphere was established, the quartz tube
was positioned in a horizontal furnace that had been preheated to an
experimental temperature of 1100 K or 1200 K. When the quartz tube
was placed, it was tilted by 1.7° relative to the horizontal axis of the
horizontal furnace. The front part of the tube outside the furnace was
cooled by a fan to enhance the circulation of TiCl4 inside the tube. Before
Fig. 2. Schematic diagram of the experimental apparatus.
113
any TiCl4 was fed into the quartz tube, the Ti ore was heat treated in
each atmosphere at each experimental temperature for 1 h to fix the
oxygen chemical potential. Afterward, 25 g of liquid TiCl4 (Wako Pure
Chemical Industries Ltd., purity ≥ 99.0%) was fed into the quartz tube
by a peristaltic pump at a flow rate of 0.126–0.160 g/min. During the
experiments, the exhaust gas discharged from the outlet pipe
was passed through deionized water twice and then through 9.1%
Na2S2O3 aq. solution.
The experiments were generally conducted for 9 h to ensure
sufficient removal of iron from the Ti ore and to confirm the influence
of reaction temperature on the amount of iron removal. After the desig-
nated reaction time had passed, the quartz tube was immediately re-
moved from the furnace and cooled at room temperature. To collect
the residues from the quartz tube, each atmosphere gas used during
the particular experiment was flowed through the tube at room tem-
perature until the TiCl4 inside it was dried. Note that the vapor pressure
of TiCl4 (l) is 0.02 atm at 300 K (Barin, 1995), which is high enough for
evaporation to occur. The collected residues were dissolved in deionized
water with sonication for at least 1 h at room temperature. The chemical
compositions of the residues obtained after the experiments were ana-
lyzed by X-ray fluorescence spectroscopy (XRF: JEOL, JSX-3100RII). The
crystalline phases were identified using X-ray diffraction (XRD:
RIGAKU, RINT 2500, Cu-Kα radiation), while the microstructures and
compositions were analyzed by scanning electron microscopy/energy
dispersive X-ray spectroscopy (SEM/EDS: JEOL, JSM-6510LV).
3. Thermodynamic analysis of selective chlorination using TiCl4
under high pO2
3.1. Conditions for selective removal of iron from Ti Ore
The main complex oxide in the Ti ore when it was oxidized under an
air atmosphere at 1100 K for 13 h was Fe2TiO5 (see Fig. 9 (e) for reference).
From a thermodynamic viewpoint, Fe 2TiO5 can be considered as a
114 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118

Table 1
Experimental conditions and analytical results for initial feedstocks used and for residues.

Exp no.a Source country Particle size of Temp., Atmosphere Reaction Mass percentage of element i, Ci (mass %)b
for Ti ore Ti ore, dore/μm T/K time, t'r/h
Ti Fe Mn Si Al

Feedstock (Initial) Vietnam 45.0 49.7 3.47 0.57 0.33

Australia 48.5 46.7 1.69 1.00 1.02
China 47.2 45.4 2.79 1.65 1.41
131125 Vietnam 74–149 1100 Ar + 1 ppm O2 5 85.7 11.6 1.42 0.38 N.D
131126 Vietnam 74–149 1100 Ar + 1 ppm O2 9 89.8 7.41 1.48 0.40 N.D
131127 Vietnam 74–149 1200 Ar + 1 ppm O2 9 97.2 N.D 1.12 0.40 N.D
131128 Vietnam 149–210 1200 Ar + 1 ppm O2 9 95.5 0.62 2.52 0.39 N.D
131202 Vietnam 210–297 1200 Ar + 1 ppm O2 9 93.3 2.82 2.46 0.42 N.D
131217 Vietnam 297–510 1200 Ar + 1 ppm O2 9 84.5 11.4 2.32 0.60 N.D
131220 Australia 74–149 1200 Ar + 1 ppm O2 9 96.6 0.75 0.64 0.56 N.D
131223 China 74–149 1200 Ar + 1 ppm O2 9 92.0 3.37 1.42 1.66 N.D
131227 Vietnam 74–149 1200 Ar + 1% O2 9 97.2 N.D 1.34 0.53 N.D
140103 Vietnam 74–149 1200 Ar + 10% O2 9 98.0 0.25 0.33 0.51 N.D
a
Experimental conditions; Weight of feedstock used for Ti ore, wore = 1.50 g. Ar flow rate for entire reaction time with fan, fAr = 500 sccm. All feedstocks were oxidized at 1100 K for 13
h under dried air before experiments. All feedstocks were oxidized for 1 h before supplying TiCl4 at each reaction temperature and under each atmosphere. Amount of TiCl4 used, wTiCl4 =
25 g, Feed rate of TiCl4, fTiCl4, = 0.126–0.160 g/min., TiCl4 feeding pattern: continuous feeding without breaks.
b
Determined by XRF analysis (excluding oxygen and other gaseous element). N.D: Not Detected. Below the detection limit of the XRF (b0.01 mass%), Values are determined by average
of analytical results of five samples.

mixture of Fe2O3 and TiO2, because the Gibbs energy of formation of shown in Eqs. (2) and (3), where the Δ G°r values are both positive.
Fe2TiO5 at 1200 K is a small negative value, as shown in Eq. (1). There- These calculation results show that the chlorination of Fe2O3 and TiO2
fore, thermodynamic analysis of the selective chlorination of Ti ore by Cl2 gas is not feasible when the pO2 is high. When the pO2 is lowered
under oxidative atmosphere can be performed by utilizing the chemical by using carbon or another reducing agent, Fe2O3 and TiO2 can be
potential diagrams of the Fe–O–Cl and Ti–O–Cl systems at 1200 K. chlorinated by Cl2 gas. Specifically, a chlorination process using Cl2 gas
in the presence of carbon is well known as a carbo-chlorination reaction.
Fe2 O3 ðsÞ þ TiO2 ðsÞ ¼ Fe2 TiO5 ðsÞ ð1Þ However, as discussed above, Fe2O3 can be chlorinated by TiCl4, as
shown in Eq. (4) (or Eq. (5)), which is a combination of the chemical
ΔG°r = −2.4 kJ at 1200 K (Roine, 2006) reactions shown in Eqs. (2) and (3). It is worth noting that the chlorina-
tion reaction of Fe2O3 (s) by TiCl4 (Eq. (4)) is independent of pCl2 and
Fig. 3 shows the chemical potential diagram constructed by overlap-
ping the chemical potential diagrams of the Fe–O–Cl (solid line) and
Ti–O–Cl (dotted line) systems at 1200 K. The hatched region with
solid lines represents the potential region where FeClx (l,g) [x = 2, 3]
and TiO2 (s) are stable. That is, if the pO2 and pCl2 are located in this
hatched region, then any iron oxides will be chlorinated as FeClx (l,g)
[x = 2, 3], whereas any titanium oxides will not be chlorinated and will
be retained as TiO2 (s). When the iron oxides are transformed to FeCl2
(l), it will evaporate, because the vapor pressure of FeCl2 at 1200 K
(0.32 atm) (Barin, 1995) is sufficiently high for evaporation. Therefore,
this hatched region is called the potential region for selective chlorination.
Because an excess amount of TiCl4 was used, the TiO2 (s)/TiCl4 (g) eq.
in Fig. 3 controls the pO2 and pCl2 of the reaction system. This equilibrium
line is the borderline of the potential region for selective chlorination. In
addition, oxygen concentration in the atmosphere controls the pO2 of
the reaction system. Consequently, when the experiment was conduct-
ed under an Ar + 1 ppm O2 atmosphere, the pO2 and pCl2 were fixed at
point a. To analyze the pO2 and pCl2 for the experiments conducted
under the Ar + 1% O2 and Ar + 10% O2 atmospheres, an extrapolated
equilibrium line for the TiO2 (s)/TiCl4 (g) eq. was assumed, as shown
in Fig. 3. On the basis of this assumption, when the experiments were
conducted under the Ar + 1% O2 and Ar + 10% O2 atmospheres, the
pO2 and pCl2 were fixed at points b and c, respectively, where the pCl2
is larger than 1 atm. In these cases, even though the internal pressure
of the quartz tube was maintained at 1 atm during the experiments,
there may be a possibility that the nominal pCl2 in some local reaction
spots instantaneously exceeded 1 atm (chemical potential base) during
the experiment. Therefore, when an excess amount of TiCl4 is used
under Ar + 1 ppm O2, Ar + 1% O2, or Ar + 10% O2 atmosphere and
the reaction system is in an equilibrium state, the Fe2O3 in Ti ore can
be chlorinated and removed as FeCl3 (g), because the pO2 and pCl2 of
the reaction system are fixed at points a, b, or c in Fig. 3, and TiO2 (s)
Fig. 3. Chemical potential diagram of the Fe–O–Cl (solid line) and Ti–O–Cl (dotted line)
can be obtained because the TiO2 in the ore cannot be chlorinated. systems overlapped at 1200 K. The hatched region with solid lines in the diagram is the
Even though pure Cl2 gas is used as a chlorinating agent for the TiO2 (s) and FeClx (l,g) stability region. The hatched region with dotted lines is the TiCl4
chlorination of Fe2O3 and TiO2, these oxides cannot be chlorinated, as (g) stability region (Barin, 1995).
 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118 115

pO2. This thermodynamic analysis suggests that the chlorination of Ti
ore can proceed regardless of the pCl2 and pO2 when TiCl4 is used as a
chlorinating agent.
2=3 Fe2 O3 ðsÞ þ 2 Cl2 ðg Þ ¼ 4=3 FeCl3 ðg Þ þ O2 ðg Þ
ΔG°r = 36.0 kJ at 1200 K (Barin, 1995)
TiO2 ðsÞ þ 2 Cl2 ðg Þ ¼ TiCl4 ðg Þ þ O2 ðg Þ
ΔG°r = 108.6 kJ at 1200 K (Barin, 1995)
2=3 Fe2 O3 ðsÞ þ TiCl4 ðg Þ ¼ 4=3 FeCl3 ðgÞ þ TiO2 ðsÞ
ΔG°r = −72.6 kJ at 1200 K (Barin, 1995)
2=3 Fe2 TiO5 ðsÞ þ TiCl4 ðg Þ ¼ 4=3 FeCl3 ðg Þ þ 5=3 TiO2 ðsÞ
ΔG°r = −72.0 kJ at 1200 K (Barin, 1995; Roine et al., 2011)
3.2. Actual potential region for the selective chlorination process
As mentioned earlier, if the reaction system is in equilibrium with an
excess amount of TiCl4, the pCl2 and pO2 of the system are fixed at points
a, b, or c depending on the oxygen chemical potential. However, the ac-
tual pCl2 and pO2 for the selective chlorination of Ti ore under high pO2
can be in a wider chemical potential area and not restricted to just
these three points when several factors such as the redox potential of
iron oxide are considered. This area is called the potential region for
oxidation-selective-chlorination, as shown in Fig. 4.
Fig. 4 shows the equilibrium lines of the FeOx (s)/FeCly (l,g) eq. and
TiO2 (s)/TiCl4 (g) eq. at 1200 K and the potential region for oxidation-
selective-chlorination under high pO2 as the hatched area. When the
redox potential of solid Fe2O3 in Ti ore is considered, the log pO2 value
ranges from the log pO2 determined by Fe3O4 (s)/Fe2O3 (s) eq. to 0
(pO2 = 1). In addition, when the chlorination of Ti ore particles by
TiCl4 is considered, the outer boundary of the ore particle reacting
with TiCl4 is controlled by the TiO2 (s)/TiCl4 (g) eq., and the center of
the ore particle not yet reacting with TiCl4 is controlled by the FeOx
(s)/FeCly (l,g) eq. Then if the chemical potential gradient of pCl2 and
pO2 between the two equilibriums is considered, the pCl2 and pO2 can
be located at any position between those two equilibrium lines. Thus,
under these two conditions and the oxidative atmospheres used in
this study, the hatched region in Fig. 4 becomes the potential region
for the oxidation-selective-chlorination of Ti ore under high pO2.
As shown in Fig. 4, the potential region for oxidation-selective-
chlorination exists in the stability region of FeCl3 (g) and TiO2 (s).
Furthermore, the vapor pressure of FeCl3 (g) is high enough for evapo-
ration at 1200 K, and the vapor pressure of TiO2 (s) is negligible. There-
fore, when the TiCl4 is used as a chlorinating agent for oxidation-
selective-chlorination at 1200 K, the iron oxide in Ti ore can be removed
as FeCl3 (g), and TiO2 (s) can be obtained.
discussed in Section 3. However, the selective chlorination of Ti ore
under a high pO2 did not proceed under equilibrium, because FeCl2
was also identified. The generation of FeCl2 is presumed to be due to
insufficient oxidation of the center parts of the Ti ore particles during
ð2Þ pretreatment and experiment, because FeTiO3 was identified in the
center of the particle after the experiment (see Fig. 8). Iron chlorides
were identified in the form of FeClx·y(H2O)z, and it is expected that
H2O attached to the iron chlorides when they were exposed to air
ð3Þ
upon removal of the plug to collect the residues. In addition, because
TiO2 was identified by XRD analysis, it is assumed that Cl2 gas was gen-
erated during the experiment by the oxidation of TiCl4 under high pO2.
ð4Þ Fig. 6 (a), (b), and (c) shows photographs taken when 1 min, 2 min,
and 4 min had elapsed after the quartz tube was removed from the fur-
nace, respectively, when the experiment was conducted at 1100 K
under an Ar + 1 ppm O2 atmosphere. As shown in Fig. 6 (a), (b), and
ð5Þ (c), as the time elapsed, the thick yellow fog inside the quartz tube
became thinner, and the fog had disappeared when 6 min had elapsed.
This yellow fog also demonstrates that Cl2 gas was generated during
experiment.
Na2S2O3 aq. solution is frequently used as a dechlorinating agent
(Ohkawa, 2005; American Water Works Association, 2006). The color
of the Na2S2O3 aq. solution through which the exhaust gas was passed
changed from transparent to white and opaque when the experiments
were conducted under Ar + 1 ppm O2 and Ar + 1% O2 atmospheres.
The color of the Na2S2O3 aq. solution changed from transparent to bright
yellow (and white) and opaque when the experiments were conducted
under the Ar + 10% O2 atmosphere. After the Na2S2O3 aq. solution was
filtered, powders obtained were dried under air at room temperature
and identified as sulfur by XRD analysis. This sulfur is produced by the
reaction between Na2S2O3 aq. solution and Cl2 gas.
The SEM images in Fig. 7 show the surface of the Ti ore particles
obtained before and after the experiment conducted at 1200 K

4. Results and discussion

4.1. Selective removal of iron from Ti Ore as iron chloride

Fig. 5 (a) shows the brown and yellow deposits condensed in the
low-temperature part of the quartz tube after the experiment was con-
ducted at 1200 K under Ar + 1 ppm O2 atmosphere (Exp. no. 131128).
After the quartz tube was dried, the color of the deposits changed from
brown to yellow because the remaining TiCl4 evaporated and was
removed from the tube. Fig. 5 (b) shows the XRD results of these
deposits after the tube was dried, in which the main compounds
identified were FeCl2·4H2O, FeCl3·6H2O, and TiO2. Fig. 4. Comparison of potential regions for oxidation-selective-chlorination in this study,
If the reactions in the system proceeded under equilibrium, then all carbo-chlorination, and carbo-selective-chlorination using a modified chemical potential
of the Fe in the Ti ore should be removed as FeCl3 under a high pO2, as diagram of the Fe–O–Cl and Ti–O–Cl systems at 1200 K (Barin, 1995).
116 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118

under an Ar + 1 ppm O2 atmosphere. As shown in Fig. 7, small pores

were present on the surface of the Ti ore particles after the
experiment. From the results, the Fe in the Ti ore was removed
as FeClx (l,g) [x = 2,3] and subsequently condensed at the low-
temperature part of the quartz tube, because these small pores are
(or are expected to be) produced by the removal of iron in the ore as a
result of the high vapor pressure of FeClx (l,g) [x = 2,3] at 1200 K.
In this study, oxidation of the Ti ore was performed before its use as
feedstock for the experiments. This pretreatment of the Ti ore guaran-
tees that the log pO2 value is between the log pO2 determined by Fe3O4
(s)/Fe2O3 (s) eq. and 0 in the y-axis (log pO2) in Fig. 4. In addition, the
generation of Cl2 gas shows that selective chlorination using TiCl4
under high pO2 proceeded in the high pCl2 region. Furthermore, the
porous structure on the surface of the Ti ore particles implies that the
pair of pCl2 and pO2 was located between the TiO2 (s)/TiCl4 (g) eq. and
FeOx (s)/FeCly (l,g) eq. On the basis of these results, it is expected that
selective chlorination using TiCl4 under high pO2 proceeded in the
potential region for oxidation-selective-chlorination in Fig. 4, as
discussed in Section 3.

4.2. Influence of reaction temperature on removal of iron

As shown in Table 1, when the experiment was conducted at 1100 K
for 9 h under an Ar + 1 ppm O2 atmosphere (Exp. no. 131126), the mass
percent of Fe in the Ti ore decreased from 49.7% to 7.41%. On the other
hand, the authors reported that the amount of Fe in Ti ore decreased
from 49.7% to 0.50% when selective chlorination using TiCl4 was
conducted at 1100 K for 4 h in the presence of carbon (low pO2) (Kang
and Okabe, 2014). The difference in the mass percent of Fe per reaction
time between high and low pO2 conditions results from the oxidation
state of the iron in the Ti ore. The oxidation state of Fe in Ti ore under
high pO2 is 3 +, whereas that under low pO2 is 2 +. Generally, the
reaction rate for the chlorination of Fe2O3 is slower than that for the
chlorination of FeO, which is why industrial processes for upgrading Ti
ore have a reduction step. Therefore, when the experiment was
conducted at 1200 K for 9 h under an Ar + 1 ppm O2 atmosphere
(Exp. no. 131127) with the identical particle size of Ti ore and reaction
time to those of Exp. no. 131126, the mass percent of Fe in the ore
Fig. 5. (a) Photograph of the deposits at low-temperature part of the quartz tube after
the experiment (Exp. no. 131128) and (b) XRD results for the deposits after the tube
was dried.
Fig. 6. Photographs taken after the quartz tube was removed from the furnace after
(a) 1 min, (b) 2 min, and (c) 4 min had elapsed (Exp. no. 131125).
decreased from 49.7% to below 0.01%, because the reaction rate was
enhanced at the higher reaction temperature.
4.3. Influence of particle size on removal of iron
The influence of particle size on the removal of Fe from Ti ore by
selective chlorination was also investigated. Table 1 shows that as the
particle size is increased, less Fe was removed from the Ti ore for a
given set of experimental conditions. For example, when a Ti ore parti-
cle with sizes in the range of 297–510 μm was used, 11.4% Fe remained,
and 87% TiO2 was obtained. On the other hand, when Ti ore particles of
74–149 μm were used, less than 0.01% Fe remained, and 98% TiO2 was
obtained. The purity of TiO2 was calculated by converting all the
elements measured by XRF to their nominal simple oxides. The XRF
analysis results do not include gaseous elements such as oxygen. The
concentration of a metal oxide in the sample can be evaluated by as-
suming that each metal element in the sample exists as a corresponding
pure metal oxide (i.e., not a compound).
Fig. 8 presents an SEM image and the EDS results for a cross section
of a Ti ore particle obtained when the experiment was conducted using
Ti ore particles of 297–510 μm (Exp. no. 131217). As can be seen from
the EDS results in Fig. 8, FeTiO3 was detected in the center part, and
TiO2 was produced at the outer part of Ti ore particle. The detection of
FeTiO3 demonstrates that selective chlorination of Ti ore did not pro-
ceed completely and that oxidation of the Ti ore before its use was insuf-
ficient when the particle size of the Ti ore is large.
Fig. 9 shows the results of XRD analysis of the feedstock and residues
obtained when the experiment was conducted using Ti ore with various
particle sizes at 1200 K under an Ar + 1 ppm O2 atmosphere. As shown
in Fig. 9, TiO2 was obtained when Ti ore particles of 74–149 μm and
149–210 μm were used, because almost all of the Fe was selectively
removed from the Ti ore. However, when Ti ore particles of
210–297 μm and 297–510 μm were used, FeTiO 3 and Fe 2 TiO 5
were identified, respectively, in addition to TiO 2. Therefore, it is
considered that when Ti ore particles of 297–510 μm were used, the cen-
ter parts of the Ti ore particles that were not oxidized by pretreatment
(FeTiO3, Fig. 8) and portions of the rest of the Ti ore particles that were
oxidized by pretreatment (Fe2TiO5) were not completely chlorinated
by TiCl4 under high pO2.
 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118
Fig. 7. SEM images of the surface of a Vietnamese Ti ore particle obtained (a) before the
experiment (after being oxidized at 1100 K for 13 h under dried air) and (b) after the
experiment (Exp. no. 131127).
4.4. Influence of various types of Ti ores and atmospheres on removal of iron
The influence of various types of Ti ores and atmospheres on the re-
moval of Fe from Ti ore by selective chlorination was also investigated.
As shown in Table 1, when Australian and Chinese Ti ores were used
as the feedstock, the mass percent of Fe in the Ti ores decreased from
46.7% to 0.75% and from 45.4% to 3.37%, respectively. As a result, 97%
and 93% TiO2 were obtained from the Australian and Chinese Ti ores,
respectively. In addition, when the experiments were conducted
under Ar + 1% O2 and Ar + 10% O2 atmospheres, the mass percent of
Fe in the Ti ores decreased from 49.7% to below 0.01% and from 49.7%
to 0.25%, respectively, and 98% TiO2 were obtained in both atmospheres.
Therefore, selective chlorination using TiCl4 can proceed for various
types of Ti ores under various high oxygen chemical potentials to
produce high-grade TiO2.
5. Conclusions
In order to produce high-grade TiO2 directly from low-grade Ti ore in
a single step, selective removal of iron from Ti ore using TiCl4 under high
pO2 (pO2 ≥ 10−6 atm) was investigated. The appropriate chemical poten-
tial region for the selective chlorination under high pO2 was studied by
utilizing the chemical potential diagram considering pCl2 and pO2.
When Ti ore was reacted with TiCl4 at 1200 K under high pO2, the iron
was directly removed as FeClx (l,g) [x = 2,3], and 98% TiO2 was obtained
in a single step under certain conditions.
In the case of selective chlorination of Vietnamese Ti ore particles of
74–149 μm by TiCl4 at 1100 K for 9 h under an Ar + 1 ppm O2 atmo-
sphere, 7.41% of the Fe remained in the residues. However, mass percent
117
Fig. 8. (a) SEM image of a cross section of a Ti ore particle obtained after the experiment
was conducted using Ti ore particles of 297–510 μm under the Ar + 1 ppm O2
atmosphere (Exp. no. 131217) and (b) EDS results for the Ti ore particle.
of Fe decreased to b 0.01% when the reaction temperature was increased
to 1200 K. When the Vietnamese Ti ore particles of 74–297 μm was
chlorinated at 1200 K under an Ar + 1 ppm O2 atmosphere, the mass
percent of Ti in the residues was ≥ 93.3%, however, this dropped to
84.5% with ore particles of 297–510 μm due to Fe in the center of the
particles not being chlorinated.
The chlorination of Vietnamese, Australian, and Chinese Ti ores of
74–149 μm at 1200 K for 9 h under Ar + 1 ppm O2 atmosphere
produced a mass percent of Ti in the respective residues of 97.2%,
96.6%, and 92.0%. Furthermore, when the selective chlorination of the
Vietnamese Ti ore of 74–149 μm was conducted at 1200 K for 9 h
under Ar + 1 ppm O2, Ar + 1% O2, or Ar + 10% O2, the mass percent
of Ti in the residues was 97.2%, 97.2%, and 98.0%, respectively. These
results show that selective chlorination process under high pO2 can be
applied to various types of Ti ore feedstocks, and with any oxygen
concentration in the atmosphere from 1 ppm to 10%.
Therefore, the selective chlorination of low-grade Ti ore using TiCl4
under high oxygen chemical potential is a feasible process for producing
high-grade titanium dioxide.
Acknowledgments
The authors are grateful to Professors Kazuki Morita and Takeshi
Yoshikawa, The University of Tokyo; Professor Tetsuya Uda, Kyoto
University; Professor Osamu Takeda, Tohoku University; and Messrs.
Susumu Kosemura, Masanori Yamaguchi, and Yuichi Ono, Toho
Titanium Co., Ltd., for their valuable discussions and the supply of
118 J. Kang, T.H. Okabe / International Journal of Mineral Processing 149 (2016) 111–118
Fig. 9. XRD results for the residues obtained when the experiment was conducted
under Ar + 1 ppm O2 atmosphere Ti ore particles with various sizes: (a) 74–149 μm,
(b) 149–210 μm, (c) 210–297 μm, and (d) 297–510 μm. (e) Results for the Vietnamese
Ti ore feedstock (74–149 μm) before the experiment.
samples throughout this study. Furthermore, the authors thank
Dr. Katsuhiro Nose and Dr. Yu-ki Taninouchi for their valuable
suggestions and technical assistance. This research was partly
funded by a Grant-in-Aid for the Next Generation of World-Leading
Researchers (NEXT Program) for the Research Project for Development
of Environmentally Sound Recycling Technology of Rare Metals
(NEXT Program #GR019), and Grant-in-Aid for Scientific Research
(S) (KAKENHI Grant #26220910), KIBAN S, by JSPS. Jungshin Kang
is grateful for the financial support that was partly provided by the
national research council of science & technology (NS grant by
the Korea government (MSIP) (No. CRC-15-06-KIGAM)), the MEM
International Graduate Program from MEXT, Japan, and the Grants
for Excellent Graduate Schools from MEXT, Japan.
References
American Water Works Association, 2006. Water Chlorination/Chloramination Practices
and Principles. second ed. American Water Works Association, Denver, Colorado, USA.
Barin, I., 1995. Thermochemical Data of Pure Substances. third ed. VCH Verlagsgesellschaft
mbH, Weinheim, Germany.
Becher, R.G., Canning, R.G., Goodheart, B.A., Uusna, S., 1965. Proc. Aust. Inst. Min. Metall.
21 pp. 21–44.
Borowiec, K. Grau, A.E. Gueguin, M. Turgeon, J-F. United States Patent 5830420 (1998).
Chen, J.H. United States Patent 3825419 (1974).
Chen, J.H. United States Patent 3967954 (1976).
Chen, J.H. Huntoon, L.W. United States Patent 4019898 (1977).
Doraiswamy, L.K., Bijawat, H.C., Kunte, M.V., 1959. Chem. Eng. Prog. 55, 80–88.
Filippou, D., Hudon, G., 2009. JOM 61, 36–42.
Fukushima, S., Kimura, E., 1975. Titan. Zirconium 23, 67–74.
Fuwa, A., Kimura, E., Fukushima, S., 1978. Metall. Mater. Trans. B Process Metall. Mater.
Process. Sci. 9, 643–652.
Guéguin, M., Cardarelli, F., 2007. Miner. Process. Extr. Metall. Rev. 28, 1–58.
Hoecker, W. European Patent EP0612854 (1994).
Ichimura, K. Oka, S. Takahashi, Y. Japan Patent 1994–191847 (1994).
Iida, T., 2012. Kinzoku Mater. Sci. Technol. 82, 218–221.
Japan Titanium Society, 2013. News & Update, Titan. 61 p. 84.
Kang, J., Okabe, T.H., 2013a. Metall. Mater. Trans. B Process Metall. Mater. Process. Sci. 44,
516–527.
Kang, J., Okabe, T.H., 2013b. Mater. Trans. 54, 1444–1453.
Kang, J., Okabe, T.H., 2014. Mater. Trans. 55, 591–598.
Kimura, E., Fuwa, A., Fukushima, S., 1979. Nippon Kogyo Kaishi 95, 821–827.
Kogel, J.E., Trivedi, N.C., Barker, J.M., Krukowski, S.T., 2006. Industrial Minerals & Rocks
Commodities, Markets, and Uses. seventh ed. Society for Mining, Metallurgy, and
Exploration, Inc. (SME), Littleton, Colorado, USA, pp. 987–1013.
Matsuoka, R., Okabe, T.H., 2005. Iron removal from titanium ore using selective
chlorination and effective utilization of chloride wastes. Proc. Symp. on Metallurgical
Technology for Waste Minimization, 134th TMS Annual Meeting, San Francisco,
United States (http://www.okabe.iis.u-tokyo.ac.jp/japanese/for_students/parts/pdf/
050218_TMS_proceedings_matsuoka.pdf).
Nakamura, N., 2013. Molten Salts 56, 121–124.
Ohkawa, C., 2005. New Extraction Process of Precious Metals from Scrap by Chemical
Vapor Treatment (Master thesis) The University of Tokyo.
Okabe, T.H., Kang, J., 2012. Future prospect of titanium smelting technology. In: Okabe,
T.H., Nose, K. (Eds.), The Latest Technological Trend of Rare Metals. CMC Publishing
Co., LTD., Tokyo, pp. 83–94.
Othmer, D.F. United States Patent 3859077 (1975).
Othmer, D.F., Nowak, R., 1972. AICHE J. 18, 217–220.
Rhee, K.I., Sohn, H.Y., 1990a. Metall. Mater. Trans. B Process Metall. Mater. Process. Sci. 21,
341–347.
Rhee, K.I., Sohn, H.Y., 1990b. Metall. Mater. Trans. B Process Metall. Mater. Process. Sci. 21,
331–340.
Rhee, K.I., Sohn, H.Y., 1990c. Metall. Mater. Trans. B Process Metall. Mater. Process. Sci. 21,
321–330.
Roine, A., 2006. HSC Chemistry 6.1 Outokumpu HSC Chemistry for Windows, version 6.1.
Outokumpu Research Oy Information Center, Finland.
Roine, A., et al., 2011. HSC Chemistry version 7.11. Outotec Oy Information Center, Finland.
Wako, T., 2012. Conference of WEPA Dialogue in Sri Lanka, Industrial Wastewater
Management in Japan. Ministry of the Environment, Japan (http://www.env.go.jp/
en/focus/docs/files/20120801-51.pdf).
Zhang, W., Zhu, Z., Cheng, C.Y., 2011. Hydrometallurgy 108, 177–188.
Zheng, H., Okabe, T.H., 2006. Selective chlorination of titanium ore and production of
titanium powder by preform reduction process (PRP). In: Yamaguchi, S. (Ed.), Proc.
16th Iketani Conf., Masuko Symposium, The 16th Iketani Conference Organizing
Committee, Japan, pp. 1005–1010.