"Perovskitization"-assisted electrochemical reduction of solid TiO2 in molten CaCl2

Article abstract of DOI:10.1002/anie.200502318

(Figure Presented) Speed up: In situ formation of perovskite phases (Ca_δTiO_x, x/δ ≥ 2) during electroreduction of porous TiO₂ pellets to Ti metal in molten CaCl₂ kinetically retards the process. Ex situ perovski

Full text of DOI:10.1002/anie.200502318

Communications
Materials Science
size[1] and it has been demonstrated that the removal of
oxygen from a 1–2-mm thick Ti foil takes only a few hours
using the same method.[1b,3d]
DOI: 10.1002/anie.200502318
In all reported studies of electroreduction of solid oxides,
such as TiO2,[1b,d,4a,7,8a] Nb2O5,[4b,c] Cr2O3,[1e,4d,8b] ZrO2,[4e,f]
NiO,[8b] SiO2,[4g,h,8c,d] and various mixed oxides,[1c,f,4e] oxygen
“Perovskitization”-Assisted Electrochemical
Reduction of Solid TiO2 in Molten CaCl2**
ionization (MOx + 2xeQM + xO2ꢁ or MOx + xCa2+
+
2xeQM + xCaO) has been confirmed to be the overall
cathodic reaction, although the kinetic steps involved in the
process were found to be much more complex. In particular,
for TiO2, Nb2O5, ZrO2, and Cr2O3, calcium-enriched oxide or
perovskite phases of various compositions are often observed
in partially reduced cathodes. However, the mechanism of
formation of these intermediate products and their influence
on the reduction process are not fully understood.[1c,4a,7] For
electroreduction of TiO2 to Ti metal, there were also argu-
Kai Jiang, Xiaohong Hu, Meng Ma, Dihua Wang,*
Guohong Qiu, Xianbo Jin, and George Z. Chen*
The present inability to increase on-board hydrogen storage
beyond 6 wt% of the storage materials and/or devices
suggests that a longer time is needed before commercially
viable hydrogen-powered vehicles join the battle against
global warming. This situation would not be quite so
constraining, however, if car manufacturers could replace
steel components by titanium ones as both are equally strong
but titanium is about 40% lighter (and more corrosion
resistant). A lighter titanium car would then allow an
additional load capacity for the heavy hydrogen-storage
system. However, titanium (ꢀ $8000 per ton of sponge from
the Kroll process) is far more expensive than steel (ꢀ $150per
ton of steel from carbothermic reduction; steel producers
consequently emit about 5 ” 1011 kg of CO2 per annum into
the atmosphere).
ments for calcium deposition being a necessary step (Ca2+
+
2eQCa; TiO2 + 2CaQTi + 2CaO),[3,5] although such claims
disagree with results from in-depth cyclic voltammetric
investigations.[1d,7a]
Herein, we demonstrate a novel concept of perovskitiza-
tion-assisted electroreduction of TiO2 to Ti metal in molten
CaCl2. The result of this process is a significantly increased
electrolysis speed and efficiency. In this context, “perovski-
tization” refers to the interaction or reaction between solid
TiOx (1 ꢂ x ꢂ 2) and calcium oxide or cation driven by
chemical or electrochemical potentials, leading to structural
transformation into perovskite phases that may have a
general formula of CadTiOx (x/d ꢃ 2).
The discovery[1a,b] of direct electrochemical reduction of
solid TiO2 (an insulator) to Ti metal in molten CaCl2 promises
an alternative to the Kroll process, which consumes far more
energy (45–55 kWh per kg of Ti) than the industrial produc-
tion of other common metals (for example, 4–6 kWhkgꢁ1 for
steel and 12–14 kWhkgꢁ1 for Al).[1–3] It also suggests innova-
tive approaches to other metallic and semimetallic materials
directly from the respective oxide powders by either electro-
lytic[1,4] or metallothermic means.[3,5] Industrial pilot tests of
the electrolytic route have started,[6] but the most urgent need
is to increase the electrolysis speed and current efficiency,
both of which are currently unsatisfactory, probably owing to
the slow process of removing dissolved oxygen from the metal
phase.[1–6] However, this should not really be the case because
the particles in the produced Ti pellet are usually 10–20 mm in
Due to electric neutrality, the injection of electrons into a
solid oxide can lead to either removal of oxygen anions,[1,4,7]
which is needed for producing metals, or uptake of metal
cations from the electrolyte into the oxide phase.[1a–e,4a,7] The
latter occurs in Li-ion batteries and has been used to
synthesize LiTiO2 and LiTi2O4 in molten LiCl;[8a] similar
processes are to be expected in molten CaCl2.
Figure 1A presents a cyclic voltammogram (CV) of TiO2
(curve a) in molten CaCl2. Details of the Mo/oxide-powder
working electrode and the Ag/AgCl reference electrode have
been described elsewhere.[8] The CV exhibits three main
reduction peaks at ꢁ0.23 V (c1), ꢁ0.41 V (c2), and ꢁ1.13 V
(c3). Peak c3 possibly includes two unresolved processes, as
implied by the inflexion at ꢁ1.02 V. In recent studies of thin
TiO2 coatings,[1d,7a] these features were attributed to TiO2 !
Ti3O5 (c1)!Ti2O3 (c2)!TiO !Ti(c3), respectively.[1d,7a] In
particular, potentiostatic electrolysis at potentials around
peaks c1 and c2 produced perovskite, which was claimed to
result from the chemical reaction (1) where the O2ꢁ ion
results from partial electroreduction of TiO2.[7a]
[*] K. Jiang, X. H. Hu, M. Ma, Dr. D. H. Wang, G. H. Qiu, Dr. X. B. Jin,
Prof. Dr. G. Z. Chen
College of Chemistry and Molecular Sciences
Wuhan University
Wuhan, 430072 (P.R. China)
Fax: (+86)27-6875-6319
E-mail: wangdh@whu.edu.cn
mel@whu.edu.cn
TiO2 þ Ca2þ þ O2ꢁðor CaOÞ Ð CaTiO3
Prof. Dr. G. Z. Chen
School of Chemical, Environmental and Mining Engineering
University of Nottingham
ð1Þ
In this work, SEM and EDX analyses of the Mo–TiO2
electrode after potentiostatic electrolysis also revealed that
the product at potentials more-negative than peak c3 was Ti
metal. However, at more-positive potentials, particularly
around peaks c1 and c2, various titanate/titanite or perovskite
phases were observed, but their O/Ti atomic ratio remained
close to 2 and the Ca/Ti ratio was ꢂ 0.5. We therefore believe
Nottingham, NG72RD (UK)
Fax: (+44)115-951-4115
E-mail: george.chen@nottingham.ac.uk
[**] The authors thank the National Natural Science Foundation of
China for financial support (Grant No 20125308).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
428
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 428 –432

Angewandte
Chemie
electrochemical reactions (2) and (3) also occur.[1b–e,4a,7a]
TiO2 þ 2ae ! TiOx þ aO2ꢁ ðx ¼ 2ꢁa, 0 ꢂ a ꢂ 0:5Þ
TiOx þ dCa2þ þ 2de Ð CadTiOx ð0 ꢂ d ꢂ 0:5Þ
ð2Þ
ð3Þ
Continuous cycling of the potential to more-negative
regions than peak c3 caused the current to decrease due to
oxygen removal being irreversible in pure molten CaCl2 that
contains very few oxygen ions.[1d,7a] However, if limited to
less-negative potentials, stable CV currents were observed.[1d]
In particular, potential cycling around peaks c1 and c2 caused
the two to merge, accompanied by the appearance of a
reoxidation peak between 0.0 and ꢁ0.1 V, (insert, Figure 1A).
This behavior agrees with reactions (2) and (3).[1d,7a]
The SEM and EDX results also showed intermediate
phases containing Ti and O (O/Ti ꢂ 1), but no Ca, at
potentials slightly more positive than peak c3. This can be
explained by the perovskite decomposing when the oxygen
content is lowered to a certain level, probably O/Ti = 1.[4a]
Therefore, the electroreduction of TiO2 may involve reac-
tions (1)–(3) as well as reactions (4)–(7). Reaction (4) is a
TiOx þ CaO ! CaTiOy ð1 þ x ¼ y ꢂ 3Þ
CaTiOy þ 2 be Ð TiOz þ CaO þ bO2ꢁ ðb ¼ yꢁzꢁ1Þ
CadTiOx þ 2 ge Ð TiOz þ dCaO þ gO2ꢁ ðg > d, z ꢂ 1Þ
TiOz þ 2 z e ! Ti þ z O2ꢁ
ð4Þ
ð5Þ
ð6Þ
ð7Þ
more general form of reaction (1). Obviously, combinations
of reactions (1)–(7) can lead to other reaction types[4a] but
these should have the same overall thermodynamic conse-
quences.
When electrolyzing the large TiO2 porous preforms that
are presently used in industrial tests, in situ perovskitization,
represented by reactions (1), (3), and (4), causes expansion in
the solid phase and slows ion transport in the pores existing
between the oxide particles.[1c,d,7] In previous work[1c,7b] and
herein, it was observed that, before full reduction, electro-
lyzed TiO2 pellets, even at voltages higher than 3.0V, always
have layered structures, as shown in Figure 1Cb. EDX
analyses revealed a large amount of Ca in the inner layers
whose O/Ti ratios were, however, different (core > inter-
mediate layer). The XRD spectra of these partially reduced
TiO2 pellets indicate the presence of a perovskite phase (see
Figure 1Bg and Table 1).
On the basis of these findings and analyses, we can
conclude that in situ perovskitization proceeds inevitably by
both electrochemical and chemical reactions during the
electroreduction of solid TiO2 in molten CaCl2. The perov-
skite phases formed can be further reduced to Ti at more-
negative potentials, but they also retard the electroreduction
kinetically. Therefore, if in situ perovskitization can be
avoided, electrolysis of TiO2 preforms should be faster.
We achieved ex situ partial and full perovskitization by
sintering a pressed mixture of TiO2 and CaO or Ca(OH)2
powders at 13008C (see Figure 2A and Experimental Sec-
Figure 1. A) CVs of the Mo/perovskite-powder electrode at 8508C. The
powder was ground from perovskitized pellets of mixed TiO2 and CaO
with a Ca/Ti ratio of 0 (a), 0.25 (b), 0.5 (c), and 1.0 (d). Scan rate:
20 mVsꢁ1. The insert shows the consecutive CVs of the Mo–TiO2
electrode in a narrower potential range. B) XRD spectra of synthetic
CaTiO3 before (a) and after electrolysis at 3.2 V and 8508C for 10 (b),
30 (c), 60 (d), 120 (e), and 240 min (f). Spectrum (g) was measured
from a partially reduced TiO2 pellet. C) Current–time plots of the
electrolysis of TiO2 (dashed line, 1.0 g pellet) and CaTiO3 (solid line,
1.7 g pellet) at 3.2 V and 8508C. The inserts are (a) the CaTiO3 pellet
wrapped by Mo wires as an assembled cathode, and (b) the cross-
section of a partially reduced TiO2 pellet with the layered interior
structure of (EDX analyses) 1: Ti metal; 2: perovskite, O/Ti=1.49; 3:
perovskite, O/Ti=1.89. All results were obtained in molten CaCl2.
Angew. Chem. Int. Ed. 2006, 45, 428 –432
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
429

Communications
Table 1: XRD phase analyses of different electrolysis products.
similar to reaction (3) that deserves more
general investigation.[8a] Third, although
electrolysis of CaTiO3 was faster, the
current efficiency still remained low
(28% for CaTiO3 vs. 15% for TiO2). The
large background current (Figure 1C) due
to electronic conduction in the melt[9] and
cell configuration might be responsible,
and these are being addressed separate-
ly.[1g]
Finally, the reduction of TiO2 preforms
by Ca vapor was recently reported to be
quicker in the presence of CaO or
CaCl2.[3b] However, the additives were
used as a flux (for by-product) and the
trace CaTiO3 phase in the sintered
(8008C) preform was ignored. More
recently, the electrochemical reduction
of solid Nb2O5 in the presence of small
amounts of carbon (< 5 wt%), CaCO3,
TiO2, 9008C
3.0 V, 0.5 h
3.0 V, 2 h
3.0 V, 4 h
2.8 V, 1 h
2.8 V, 2 h
2.8 V, 5 h
2.6 V, 5 h
2.4 V, 11.5 h
perovskite (81.3%) + Ti2O3 (18.7%)
perovskite (83.7%) + Ti2O3 (10.6%) + Ti3O5 (5.7%)
perovskite (89.9%) + TiO (11.1%)
TiO2, 9008C
perovskite (50.7%; 31.2%)[a] + TiO2 (49.3%; 68.8%)[a]
perovskite (65.6%)[a] + Ti3O5 (18.2%)[a] + unknown (16.2%)[a]
Ti (87.2%)[a,b] + unknown (12.8%)[a]
CaTiO3, 8508C
Ti (51.8%)[a,b] + perovskite (38.8%)[a] + unknown (9.4%)[a]
Ti (50.1%)[a,b] + perovskite (37.7%)[a] + unknown (12.2%)[a]
[a] The presence of some unknown phases prevented the XRD spectrum quantitative analysis software
from determining the proportions of the different phases observed. The numbers in italics were
estimated from the area of the strongest XRD peak of each phase and are presented here as approximate
references with caution. Use of the total area of three of the strongest peaks of each phase led to similar
results. [b] The electrolyzed CaTiO3 samples listed in the table fell apart into a powder after washing in
water, and contained up to 7 wt% oxygen as measured by LECO. Their XRD spectra exhibit all the main
titanium peaks, although these were shifted to lower diffraction angles close to those of the standard
TixO (xꢃ2) phases. The negative shifts of the Ti peaks are understandable because of lattice expansion
caused by dissolved oxygen. The XRD spectra of typical fully and partially electrolyzed CaTiO3 samples
are presented in the Supporting Information.
tion). The CVs of the synthetic powder are also shown in
Figure 1A. The currents on these CVs are normalized against
peak c3. It can be seen that peaks c1 and c2, which lead to in
situ perovskitization according to the above discussion, have
been effectively supressed. Figure 1C compares the current–
time plots recorded during electrolysis of similar CaTiO3 and
TiO2 pellets under identical conditions. Interestingly, the
current in TiO2 is obviously larger than that in CaTiO3 for the
first 20min. This may be due to reactions (2) and (3)
occurring with TiO2 but not with CaTiO3. XRD analyses of
the electrolysis products at different times (Figure 1B)
revealed that the metal phase grows at the expense of the
perovskite phase, and at least one early and one later
intermediate phase appear and vanish with time (see the
peaks enclosed in the vertical rectangles in Figure 1B and also
data in Table 1 and Figure S3 in the Supporting Information).
These results may indicate that CaTiO3 is reduced to Ti via
relatively stable intermediate TiIII and/or TiII states, thereby
supporting reactions (5) and (6).[7a] The oxygen content in the
CaTiO3 pellet electrolyzed for 265 min at 3.2 V was lower
than the detection limit of EDX over relatively large areas
inspected by SEM (Figure 2), and was only 2800 ppm
according to fusion elemental analysis. In contrast, an
identical TiO2 pellet electrolyzed for 265 min showed a very
thin metallic surface but its interior still contained layered
perovskite phases similar to those shown in Figure 1Bg and
1Cb. Complete reduction of the TiO2 pellet to Ti metal
(2100 ppm oxygen) took about 600 min. Clearly, the faster
reduction of CaTiO3 occurs because in situ perovskitization is
avoided. The removal of CaO [reaction (5)] may also leave a
more porous structure and very fine TiOz particles, both of
which help to speed up the electrolysis.
and/or CaO (< 20wt%) has been reported. [10] Nonetheless,
the function of the additives was claimed to be the creation of
O2ꢁ vacancies in the oxide phase, and only small amounts of
the calcium niobate phases were detected by XRD in the
sintered oxide pellets. The effect of these calcium niobate
phases was regarded to be detrimental to the reduction.[10a]
In summary, we have demonstrated, for the first time, that
ex situ perovskitization can significantly increase the speed
and efficiency of the electroreduction of solid TiO2 in molten
CaCl2. We anticipate that the fast reduction of CaTiO3 can
also be applied to extract Ti metal directly from natural
perovskite, and from the by-product of steel making from
ilmenite (FeTiO3), which is often enriched in Mg and/or Ca.
This process may also, in principle, be applied for the
electroreduction of other oxides, such as ZrO2, Nb2O5, and
Ta2O5.
Experimental Section
Synthesis: The conventional synthesis of CaTiO3 is by a solid-state
reaction between TiO2 and CaCO3 or CaO at temperatures of about
13508C. This phase has also been prepared by other techniques,
including organometallic, liquid-mix, and mechanochemical methods.
In this work, partial and complete chemical perovskitization of TiO2
was achieved as described below. The TiO2 powder (AR grade;
particle size: 0.1–0.2 mm, Kemiꢀou Reagent Plant, Tianjin, China) was
thoroughly ball-milled (overnight) with a prescribed amount of CaO
or Ca(OH)2 (AR grade, Fengcheng Reagent Plant, Shanghai, China)
in the presence of distilled water (enough to turn the mixture into a
thin slurry). The mixture was dried in air at 1208C (overnight), die-
pressed into small cylindrical pellets (6 MPa), and then sintered at
13008C for 5 h. The color change (white to brown) and SEM and
XRD analyses showed that, depending on the Ca/Ti atomic ratio (0to
1) in the mixture, the sintered pellet (diameter: ꢀ 2.0cm; thickness:
0.2–0.3 cm) had been partially or fully converted into porous CaTiO3
(perovskite), as shown in Figures1Ba and 2A. The porosity was
estimated to be 40–50% on the basis of the pelletꢀs mass and volume,
Three more points are worth mentioning. First, the CVs of
simply mixed TiO2 and CaO powders show some, although
limited, changes of peaks c1 and c2. Second, in the presence of
trace amounts of Mg (< 0.3 wt%), Mg-containing phases
(often crystallites) were detected in partially reduced TiO2
electrodes by SEM and EDX.[8c,d] The highest atomic ratios
were Mg/Ti ꢀ 0.5 and O/Ti ꢀ 1.5, in agreement with a process
and the densities of the oxides (CaTiO3: 4.02 gcmꢁ3
; TiO2:
4.23 gcmꢁ3). Pure TiO2 pellets were also prepared similarly but
were sintered at lower temperatures (900–10008C) for shorter times
430
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 428 –432

Angewandte
Chemie
directly or pre-mixed with the CaO powder before loading onto the
electrode. The perovskite powders were obtained by finely grinding
the respective perovskite pellets that were prepared as described
above.
About 45 g of dried CaCl2 was filled into a small graphite crucible
(inner diameter: 1.8 cm; height: 25 cm), dehydrated, and pre-electro-
lyzed as described above. The graphite crucible also functioned as the
counter electrode. The three-electrode cell was then sealed in a quartz
tube, which was placed in a furnace and heated to the working
temperature (8508C). During the experiments, argon was continu-
ously fluxed through the quartz tube. Cyclic voltammetry was
performed with a CHI 605A Electrochemical System (Shanghai
Chenhua, China). After chronoamperometric experiments for a
prescribed time, the Mo–oxide powder electrodes were washed with
distilled water, dried, and examined by SEM and EDX.
Spectroscopic analysis: The XRD patterns were obtained with a
D8 Advance X-ray diffractometer (Bruker, Germany) with CuKa1
radiation (l = 1.5406 Š) at 40kVand 40mA. The scan rate of 2 q was
4 degminꢁ1. Micrographic and composition analyses were performed
with a FEI Sirion Field Emission Gun SEM system equipped with a
CCD camera and an EDAX GENESIS 7000 energy-dispersive X-ray
spectrometer (EDX). Products were also sent to the Materials Test
and Inspection Research Institute of Wuhan Iron and Steel (Group)
for analysis of their oxygen content with an Oxygen Determinator
(model RO-416DR, LECO, USA).
Figure 2. SEM images of the interior of a CaTiO3 pellet before (A) and
after (B) electrolysis (3.2 V, 8508C, 265 min, molten CaCl2). The EDX
spectrum measured over the image area of (B) is also shown.
(2 h). The pellets were used directly in the assembled cathode (see
Figure 1Ca) for constant-voltage electrolysis in a two-electrode cell.
Constant-voltage electrolysis: Molten dried CaCl2 was used as the
electrolyte in the electrolysis experiments and the subsequent cyclic
voltammetry and chronoamperometry studies. Two commercial
CaCl2 salts were used in this work: the anhydrous salt (AR grade,
the content of Mg and alkali metals was about 0.3% counted by
sulfate, Silian Reagent Plant, Shanghai, China) and the dihydrate salt
(ACS grade, the content of Mg and alkali metals was less than
0.005 wt.%, Shanghai Bioengineering Co. Ltd,).
Received: July 3, 2005
Published online: December 12, 2005
The CaCl2 salt in a graphite crucible (inner diameter: 9.0cm;
height: 23.5 cm) was placed in a sealable stainless-steel reactor, which
was assembled in a vertical programmable furnace (Wuhan Exper-
imental Furnace Plant). The furnace temperature was first raised to
and maintained at 200–3008C for more than 12 h, then to 4508C for
2 h under argon, and finally to the working temperature (850–9008C).
After melting, pre-electrolysis of the freshly prepared molten salt
at 2.6–2.8 V was applied to remove some redox-active impurities from
it. A graphite rod (diameter: 2.0cm; length: 20cm) and a steel wire
(diameter: 2.9 mm; length: ꢀ 60cm) were used as the anode and
cathode, respectively. Pre-electrolysis lasted until the current reached
a low and stable value, that is, the background or residual current,
which was found to depend strongly on temperature, cell voltage, and
the purity of the CaCl2 salt.
The oxide pellet was either wrapped with molybdenum wires or
sandwiched between two molybdenum meshes to form the assembled
cathode (Figure 1Ca). After pre-electrolysis, the steel-wire cathode
was replaced by the assembled oxide cathode and electrolysis was
carried out at 2.4–3.2 V. During all pre-electrolysis and electrolysis
processes, the electrolytic cell was sealed in the stainless-steel retort
and was continuously purged with an argon flow.
Keywords: electrochemistry · green chemistry · perovskites ·
solid-state reactions · titanium
.
[1] a) D. J. Fray, T. W. Farthing, Z. Chen, PCT Patent, WO9964638,
1998; b) G. Z. Chen, D. J. Fray, T. W. Farthing, Nature 2000, 407,
361; c) G. Z. Chen, D. J. Fray, Light Met. 2004, 881; d) G. Z.
Chen, D. J. Fray, J. Electrochem. Soc. 2002, 149, E455; e) G. Z.
Chen, E. Gordo, D. J. Fray, Metall. Mater. Trans. B 2004, 35, 223;
f) A. J. Muir Wood, R. C. Copcutt, G. Z. Chen, D. J. Fray, Adv.
Eng. Mater. 2003, 5, 650; g) D. J. Fray, PCT Patent,
WO2004018735A1, 2002.
[2] D. Yan, Energy Saving of Non-ferrous Metall. 2004, 21, 12 (in
Chinese).
[3] a) I. Park, T. Abiko, T. H. Okabe, J. Phys. Chem. Solids 2005, 66,
410; b) T. H. Okabe, T. Oda, Y. Mitsuda, J. Alloys Compd. 2004,
364, 156; c) T. H. Okabe, D. R. Sadoway, J. Mater. Res. 1998, 13,
3372; d) T. H. Okabe, M. Nakamura, T. Oishi, K. Ono, Metall.
Mater. Trans. B 1993, 24, 449.
[4] a) C. Schwandt, D. J. Fray, Electrochim. Acta 2005, 51, 66;
b) X. Y. Yan, D. J. Fray, Metall. Mater. Trans. B 2002, 33, 685;
c) X. Y. Yan, D. J. Fray, J. Electrochem. Soc. 2005, 152, D12;
d) E. Gordo, G. Z. Chen, D. J. Fray, Electrochim. Acta 2004, 49,
2195; e) G. Z. Chen, D. J. Fray, Light Met. 2001, 1147; f) K. S.
Mohandas, D. J. Fray, poster presentation at EUROMAT 2005,
5 – 8 Sept. 2005, Federation of European Materials Societies,
Prague, Czech Republic; g) T. Nohira, K. Yasuda, Y. Ito, Nat.
Mater. 2003, 2, 397; h) K. Yasuda, T. Nohira, K. Amezawa, Y. H.
Ogata, Y. Ito, J. Electrochem. Soc. 2005, 152, D69.
[5] a) K. Ono, R. O. Suzuki, JOM 2002, 54, 59; b) R. O. Suzuki, K.
Teranuma, K. Ono, Metall. Mater. Trans. B 2003, 34, 287; c) R. O.
Suzuki, J. Phys. Chem. Solids 2005, 66, 461.
[6] a) http://www.darpa.mil/dso/thrust/matdev/titanium.htm; b) S.
Ashley, Sci. Am. 2003, 289, 38.
[7] a) K. Dring, R. Dashwood, D. Inman, J. Electrochem. Soc. 2005,
152, E104; b) K. Dring, M. Jackson, R. Dashwood, H. Flower, D.
Inman in Cost-Affordable Titanium—A Symposium Dedicated
to Professor Harvey Flower (Eds.: F. H. Froes, M. A. Imam, D.
After termination of the constant-voltage electrolysis at
a
prescribed time, the cathode was removed from the molten salt,
cooled in argon, and then washed with water in air to dissolve the
solidified salt on the cathode. The electrolyzed pellets were visually
inspected for color and dimension changes, and then broken into two
halves whose cross-sections were examined under optical and
scanning electron microscopes. For powder X-ray diffraction and
fusion elemental analyses, the electrolyzed pellets were manually
ground into a powder with a mortar and pestle, further washed with
water, and dried at about 1208C in air or vacuum before further
analysis.
Cyclic voltammetry and chronoamperometry: The recently
reported Mo–oxide powder electrodes[8a,b] and a quartz-sealed Ag/
AgCl reference electrode[8c] were used to perform cyclic voltammetry
and chronoamperometry (potentiostatic electrolysis) in molten
CaCl2. Typically, the Mo-cavity powder electrode[8b] and the oxide-
powder-modified Mo electrode[8a] were loaded with 0.1–1.0 mg of
oxide powder. The sub-micrometer TiO2 powder was used either
Angew. Chem. Int. Ed. 2006, 45, 428 –432
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
431

Communications
Fray), Minerals, Metals & Materials Soc., Warrendale, 2004,
p. 95.
[8] a) K. Jiang, X. H. Hu, H. J. Sun, D. H. Wang, X. B. Jin, Y. Y.
Ren, G. Z. Chen, Chem. Mater. 2004, 16, 4324; b) G. H. Qiu, M.
Ma, D. H. Wang, X. B. Jin, X. H. Hu, G. Z. Chen, J. Electrochem.
Soc. 2005, 152, E328; c) P. Gao, X. B. Jin, D. H. Wang, X. H. Hu,
G. Z. Chen, J. Electroanal. Chem. 2005, 579, 321; d) X. B. Jin, P.
Gao, D. H. Wang, X. H. Hu, G. Z. Chen, Angew. Chem. 2004,
116, 751; Angew. Chem. Int. Ed. 2004, 43, 733.
[9] a) U. Stohr, W. Freyland, Phys. Chem. Chem. Phys. 1999, 1, 4383;
b) G. Z. Chen, D. J. Fray, J. Appl. Electrochem. 2001, 31, 155.
[10] a) Q. Xu, L. Q. Deng, Y. Wu, T. Ma, J. Alloys Compd. 2005, 396,
288; b) L. Q. Deng, Q. Xu, T. Ma, B. Li, Y. C. Zhai, Acta Metall.
Sin. (Chin. Ed.) 2005, 41, 551.
432
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 428 –432