Reduction of titanium dioxide to metallic titanium conducted under the autogenic pressure of the reactants

Article abstract of DOI:10.1021/ic900087c

We report on a reaction to convert titanium dioxide to titanium. The reduction reaction was done under the autogenic pressure of the reactants at 750 °C for 5 h. The MgO, a by- product, was removed by acids to obtain pure metallic titanium.

Full text of DOI:10.1021/ic900087c

7
066 Inorg. Chem. 2009, 48, 7066–7069
DOI: 10.1021/ic900087c
Reduction of Titanium Dioxide to Metallic Titanium Conducted under the Autogenic
Pressure of the Reactants
Michal Eshed, Alexander Irzh, and Aharon Gedanken*
Department of Chemistry, Kanbar Laboratory for Nanomaterials, Nanotechnology Research Center, Institute
of Nanotechnology and Advanced Materials, Bar-Ilan University, Ramat-Gan 52900, Israel
Received January 16, 2009
We report on a reaction to convert titanium dioxide to titanium. The reduction reaction was done under the autogenic
pressure of the reactants at 750 °C for 5 h. The MgO, a by- product, was removed by acids to obtain pure metallic
titanium.
1
. Introduction
synthesis to separate the electrodeposited titanium from the
ionic solutions have been hampered by difficulties. The
Metallic Titanium shows excellent physical and chemical
current paper presents a simple process to convert TiO to
2
properties, such as low density, high corrosion resistance at
high temperatures, good mechanical strength, ductility, and
fatigue resistance. It also exhibits a good biocompatibility
Ti, termed “RAPET” (Reaction under Autogenic Pressure at
ElevatedTemperature). The RAPET reaction takes place ina
closed cell (made of parts of the Swagelok company), and has
1
and bio integration with the human body.
9,10
been described elsewhere. The union (made by Swagelok)
is filled with the reactants, and the reaction is condu-
cted without a catalyst or a solvent at high temperature
and pressure. The pressure in the cell is generated from
the evaporation of the reactants or the products, and the
Swagelok union serves as a relief valve at 160 atm.
Several techniques are currently used to synthesize Tita-
nium. The first, is an electrochemical method for the direct
reduction of solid TiO in which the oxygen is ionized,
2
dissolved in a molten salt, and discharged at the anode,
2,3
leaving pure titanium at the cathode.
The most popular method is the Kroll process. At the first
step TiO is converted to TiCl , after which titanium is
2
4
2. Experimental Details
produced by the reduction of titanium tetrachloride with
an active metal such as magnesium at 800-900 °C according
All chemicals were obtained from Aldrich. A 3 mL closed
vessel cell was assembled from stainless steel parts
4-8
to the following reaction:
0
0
(
Swagelok). A 3/8 union part was plugged from both sides
by standard caps. For this synthesis, 0.5 g of titanium dioxide
anatase) and 0.3 g of magnesium powder were introduced
2
Mg þ TiCl4ðgÞ f 2MgCl þ TiðsÞ
ðlÞ 2ðlÞ
(
into the cell, and the cell was closed tightly at room tempera-
ture under nitrogen atmosphere (in a nitrogen- filled
glovebox). The cell (Swagelok) was placed inside an iron pipe
at the center of the tube furnace. The temperature was raised
at a rate of 10 °C per minute. The closed cell was heated at
750 °C for 5 h. The reaction proceeded under the autogenic
pressure of the precursor. The closed vessel cell (Swagelok)
was gradually cooled (5 h) to room temperature. Magnesium
oxide was removed by treatment with an aqueous solution of
acetic acid for 1 day. The sample was collected by centrifuga-
tion (9000 rpm for 20 min, Hettich Universal 32) and was
washed twice with water, once with ethanol, and then dried.
Since we observed that not all the MgO was dissolved we
continued the treatment by placing the products for an
additional 2 days in formic acid, after which the washing
and drying was repeated.
The byproduct, MgCl , is removed by vacuum distillation,
which complicates this process. The titanium is only partially
2
reduced in the process and lower chlorides such as TiCl and
2
TiCl3 are also obtained. Their removal makes the pro-
cess very costly. Moreover, attempts in the electrochemical
*
To whom correspondence should be addressed. E-mail: gedanken@mail.
biu.ac.il.
(
1) Garbacz, H.; Kurzydlowski, K. J. Macromol. Symp. 2007, 253, 128–
1
33.
(
2) Suzuki, R. O. JOM 2007, 68–71.
3) Jiang, K.; Hu, X.; Ma, M.; Wang, D.; Xianbo Jin, G. Q.; Chen, G. Z.
(
Angew. Chem. 2006, 118, 442–446.
(
(
(
(
(
4) Alexander P. P. U.S. Patent 2,038,402, 1936.
5) Alexander P. P. U.S. Patent 2,043,363, 1936.
6) Alexander P. P. U.S. Patent 2,082,134, 1937.
7) Kroll, W. J. Trans. Am. Electrochem. Soc. 1940, 78, 35–47.
8) Ikeshima, T. In Titanium Science and Technology, Proc. 5th Int.
(9) Pol, S.; Pol, V.; Gedanken, A. Chem.;Eur. J. 2004, 10, 4467–4473.
(10) Pol, S. V.; Pol, V. G.; Kessler, V. G.; Seisenbaeva, G. A.; Sung, M.;
Asai, S.; Gedanken, A. J. Phys. Chem. B 2004, 108, 6322–6327.
Conf. Titanium, Mu ꢀe nchen 1984; Lutjering, G., Zwicker, U., Bunk, W., Eds.;
( 14 DGM-Deutsche Gesellschaft fu ꢀe r Materialurkunde. V., Oberursal, 1985.
3
pubs.acs.org/IC
Published on Web 07/08/2009
r 2009 American Chemical Society

Article
Instruments. X- ray diffraction (XRD) patterns were collected
using a Bruker AXS Advance powder X-ray diffractometer (Cu
KR radiation). Transmission electron microscope measure-
ments were carried out using a transmission electron microscopy
Inorganic Chemistry, Vol. 48, No. 15, 2009 7067
characterized with a high-resolution scanning electron micro-
scope (HRSEM) JEOL-JSM 840. Specific surface areas were
measured by the Brunauer-Emmett-Teller (BET) method at
77 K using N gas as an absorbent after heating the sample at
2
(
TEM) instrument FEI Tecnal spirit 120 KV spirit biotwin.
120 °C for 1 h. An Olympus BX41 (Jobin Yvon Horiba) Raman
spectrometer was employed. Oxygen elemental analysis was
measured by CHNS-O analyzer Thermo FlashEA 1112 series.
For some measurements we had special cells that were filled with
the product in the glovebox and carry the sample to the instru-
ment without its exposition to air. Such were the XRD
and RAMAN measurements. On the other hand the TEM,
HRTEM, and SEM measurements were conducted after a
minimal exposure to air. Since in each measurement the expo-
sure time to air was different, the amount of oxygen, and a fire
was or was not observed. In any case the oxygen level, as
measured by elemental analysis, was low even in cases in which
we had to expose the sample to air.
Samples for TEM were prepared by ultrasonically dispersing the
products into absolute ethanol, placing a drop of this suspension
onto a copper grid coated with an amorphous carbon film or
onto a copper plate, and then drying in vacuum. High resolution
TEM (HRTEM) was measured using a JEOL JEM-2100 elec-
tron microscope. The morphologies and nanostructure were
3. Results and Discussion
Figures 1a shows an XRD pattern of the commercial
titania (anatase). Figures 1b and 1c show the XRD patterns
for the as-prepared product before etching the MgO and after
the acid treatment, respectively. The diffraction peaks of
Figure 1b, match well with the PDF tables of Ti and MgO,
for Ti: 00-044-1294 and for MgO: 03-065-0476. The
comparison between Figure 1b and the acid treated product,
Figure 1c, clearly shows that the MgO diffractions have
disappeared. Moreover, the reflection peak at 2θ = 22°
Figure 1. XRD patterns of (a) titania (anatase), (b) the titanium sample
before treatment with acids, and (c) after reacting with acetic acid and
formic acid.
Figure 2. HRSEM picture of (a) titanium dioxide nanoparticles; (b) and (c) titanium nanoparticles after removing MgO at different magnifications and in
different areas in the sample.

7
068 Inorganic Chemistry, Vol. 48, No. 15, 2009
Eshed et al.
Figure 3. TEM images of (a) titanium dioxide nanoparticles, (b) titanium nanoparticles after the removal of MgO, and (c) overall EDS spectra of the
sample presented in Figure 3b.
(
Figures 1b and 1c) is due to the plastic layer that surrounds
rupture the bond between the titanium and the oxygen.
Therefore, when the MgO is removed, cavities are left
inside the particles, causing the porosity of the titanium
product. The average diameter of the titanium nanoparti-
cles calculated from Figure 3b was found to be 157 nm. The
increase in the size of the product, as compared with
the special inert XRD cell to avoid the oxidation of the Ti
sample. The peak at 2θ = 22° does not appear in Figure 1a
because there was no need to isolate TiO from ambient air.
In addition, the reduced intensity of the titanium in Figure 1c
as compared to Figure 1b can be explained as follows: The
acid according to our interpretation damages the crystali-
2
the size of the TiO (70-200 nm) reactant is due to the
2
þ
nity of the product, because part of the H ions are incorpo-
sintering of the product at the high temperature of the
reaction. The size distribution of the titanium particles is
not homogeneous. Some particles break up to form smaller
particles of larger size than that of the titania’s original
size.
rated in the Ti unit cell. Indeed minor shifts in the peak
position are observed. According to elemental analysis, the
oxygen was found to be in a very low concentration of
0
.0157% (wt). This substantiates our XRD results, demon-
strating that the fabricated material is indeed Ti, and not an
oxide.
Table 1 presents results of EDS measurements taken on
various particles exhibited in Figure 3b. No traces of Mg are
detected. It means that MgO was removed upon the acid
treatment. At one point when Mg was found its concentra-
tion was very low. However, oxygen is detected in the EDS
measurements. This is because the transfer of the sample to
the EDS machine exposed the sample to air, and the Ti
burned for a few second upon this exposure. The fire that
bursts upon exposure to air is by itself an indication for the
formation of zero-valent Ti.
Figure 2a depicts a HRSEM photograph of the reactant
titania before the RAPET reaction. The titania particles
reveal a smooth surface, and their diameter ranges from 70
to 200 nm. Figures 2b, and 2c illustrate the product after the
acid treatment. Figure 2b shows that some particles retain
there smoothmorphology, whereas Figure 2cshowthat some
of the particles became rough and porous.
The porous nature of the product formed, i.e, metallic
Ti, is further confirmed by the TEM images presented in
Figure 3. The comparison of the TEM image of the titania,
Figure 3a, to that of titanium, Figure 3b, indicated clearly
the porous nature of the Ti product. The porous nature can
be explained as follows: when the reaction occurs at 750 °C
Figure 4 exhibits a high magnification TEM image of
titanium. The well- defined electron diffraction (ED) spots
are shown as an inset in Figure 4. The ED pattern confirms
the crystalline nature of the metallic titanium. The dominant
˚
intensity spots relate to a d- space of 2.28 A, assigned to
the liquid atoms of Mg diffuse into the TiO structure and
diffractions from the (101) plane of Ti which fits the PDF
2

Article
Inorganic Chemistry, Vol. 48, No. 15, 2009 7069
Table 1. Results of the EDS Measurements of the Sample after the Removal of
a
the MgO
area in the
sample
detector
element weight % atomic % uncert. % correction
1
O (k)
Ti (K)
O (k)
Ti (K)
Mg (K)
O (k)
15.69
84.3
32.64
64.0
3.35
35.78
64.21
58.05
38.01
3.92
0.49
0.55
1.33
0.62
0.14
0.34
0.4
0.49
0.98
0.49
0.98
0.88
0.45
0.98
2
3
17.67
82.32
39.12
60.87
Ti (K)
a
The concentration of the elements was measured at three different
places on the grid.
table(No. 00-044-1294). Clear fringesare observed in Figure 4
and can assist in assigning the product. The measured spacing
˚
was found to be 2.24 A
BET analysis of the nitrogen adsorption- desorption iso-
therms of titanium (after the treatment with the two acids)
shows a small hysteresis behavior. The surface area is 34.4
Figure 4. HRTEM image of titanium.
2
m /g and the pore volume is 0.1027 cc/g, whereas the surface
2
area of pure titania was found to be only 8.5 m /g. This
-1
of 200-700 cm . We could not detect the scattering peak at
increase in the surface area is the result of the pores formed in
the reaction.
-1
1
37 cm assigned to metallic titanium because of instrument
limitations. On the other hand, we can state clearly that Ti
oxides are absent in the product mixture.
We have conducted Raman measurements in an inert cell
which has avoided the exposure of the sample to air. The
purpose of the Raman scanning was to monitor the presence
of residual Ti oxides. Titanium oxides are known to show
scattering featuresatthe following wavenumbers: Ti O has 4
4. Conclusion
2
3
-1
The paper reports on an easy and simple method for the
peaks at the range of 200-400 cm , and TiO has a peak at
00 cm . No Raman peaks have been detected in the range
2
-1 11
preparation of titanium. During the reduction process the
morphology has changed and led to an increase of the surface
area. As we have seen, the by-product, MgO, was removed
completely to obtain pure titanium.
6
(
11) Nemanich, R. J.; Tsai, C. C.; Connell, G. A. N. Phys. Rev. Lett. 1980,
4
4, 273.