Hollow alumina macrotubes

Article abstract of DOI:10.1023/A:1012331227393

Poorly crystallized Al₂O₃macrotubes 50 μm to 6 cm in length, 10 to 300 μm in outer diameter, and 2 to 60 μm in inner diameter were obtained by heating partially hydrolyzed AlCl₃powder to 170-220°C in a flowing inert gas or air. The tubes were characterized by electron microscopy, atomic-force microscopy, IR spectroscopy, x-ray diffraction, and electron probe microanalysis.

Full text of DOI:10.1023/A:1012331227393

Inorganic Materials, Vol. 37, No. 10, 2001, pp. 1037–1040. Translated from Neorganicheskie Materialy, Vol. 37, No. 10, 2001, pp. 1219–1223.
Original Russian Text Copyright © 2001 by Berdonosov, Baronov, Kuz’micheva, Berdonosova, Melikhov.
Hollow Alumina Macrotubes
S. S. Berdonosov, S. B. Baronov, Yu. V. Kuz’micheva,
D. G. Berdonosova, and I. V. Melikhov
Moscow State University, Moscow, 119899 Russia
e-mail: melikhov@radio.chem.msu.ru
Received February 28, 2001; in final form, April 18, 2001
Abstract—Poorly crystallized Al O macrotubes 50 µm to 6 cm in length, 10 to 300 µm in outer diameter, and
2
3
2
to 60 µm in inner diameter were obtained by heating partially hydrolyzed AlCl powder to 170–220°C in a
3
flowing inert gas or air. The tubes were characterized by electron microscopy, atomic-force microscopy, IR
spectroscopy, x-ray diffraction, and electron probe microanalysis.
INTRODUCTION
tubes perpendicular or inclined to the sample surface
became visible (Fig. 1). The tubes continued to exhale
steam and grew rapidly (Ӎ1 mm/s) in length.
Steaming became weaker as the tubes grew longer.
Occasionally, it ceased for several seconds and then
resumed. The tube then continued to grow, but its diam-
eter decreased abruptly (Fig. 2).
When steam exhalation was over, the tube growth
stopped.At a heating rate of 60°C/min, the formation of
tubes took, on average, 30 s.
After the tubes stopped growing, the heater was
switched off, and the boat was taken out of the reactor.
From one to a hundred tubes grew in a run. The largest
length of the tubes was 6 cm, and their average weight
was 20 µg.
Recent findings in fullerene research have sparked
wide interest in the preparation of fullerene-like carbon
tubes [1–4]. Not only nano- and microtubes but also
macroscopic carbon tubes up to several centimeters in
length could be produced [5]. It is reasonable to expect
that macrotubes can also be produced by chemical
means from other materials.
This was confirmed in our experiments by growing
alumina tubes up to 6 cm in length, 10–300 µm in outer
diameter, and 2–60 µm in inner diameter.
TUBE GROWTH
The starting material was anhydrous AlCl prepared
3
by chlorinating high-purity metallic aluminum (as
found later, the purity of Al had no effect on the forma-
tion of macrotubes) and purified by sublimation in an
evacuated ampule to remove nonvolatile contaminants.
CHARACTERIZATION OF MACROTUBES
The macrotubes possessed a relatively high elastic-
ity: the thinnest tubes, ranging up to 2–2.5 cm in length,
did not break when bent through 120°–180°.
AlCl (0.5 g) was placed in a porcelain boat and
3
exposed for 0.25–48 h to humid air (water vapor pres-
sure of Ӎ2.7 kPa). Next, the boat with the partially
Under an optical microscope, the tubes were seen to
hydrolyzed product was transferred to a quartz tube consist of transparent material. Their lustrous surface
≤
4 cm in diameter, fitted with a heating coil.
was covered with a white deposit, which could be
removed mechanically (for instance, with a spatula).
The tubes were examined by optical microscopy
MZ12 Leica microscope, Quantimet 550IW image
Hydrated aluminum chloride is known to be nonvol-
atile, unlike anhydrous AlCl , and to decompose above
3
(
1
50–250°C to form hydrogen chloride, AlCl vapor,
3
and Al O .
2
3
In our experiments, heating the partially hydrolyzed
AlCl powder to 100–150°C gave rise to random
3
motion of powder particles, some of which even
escaped the boat. This was presumably caused by non-
uniform decomposition of the hydroxychlorides pro-
duced by AlCl hydrolysis. At this stage, a rough crust
3
was formed on the sample surface.
Starting at 150°C, small hillocks and then openings
appeared on the crust surface. Some of the openings
exhaled jets of white steam. At 180°C, short transparent
Fig. 1. Boat with tubular alumina fibers.
0
020-1685/01/3710-1037$25.00 © 2001 MAIK “Nauka/Interperiodica”

1
038
BERDONOSOV et al.
rial was also characterized by x-ray diffraction (Debye–
Scherrer and Laue photographs) and selected-area diffrac-
tion (SAD) on a JEM-100B electron microscope.
The tubes were found to have rather smooth inner
and outer surfaces (Fig. 3), typically with a beaded or
bell-mouthed free end (Fig. 4) and without encapsu-
lated material. Each tube was, most likely, made up of
several fragments.
According to microprobe analysis data, the tubes
consisted of oxygen, aluminum, and trace amounts of
chlorine (table). If a small amount (≤5 mol %) of FeCl3
was added to AlCl , and the hydrolyzed material was
heated in a hydrogen atmosphere, the resulting α-Fe
3
1
mm
(
identified by x-ray diffraction) was concentrated
mainly at the free tube end (table).
Fig. 2. Macrotube with a varying outer diameter.
Similar tubes grew if ≤40 mol % NH Cl was added
4
to AlCl . Attempts to produce macrotubes via partial
3
hydrolysis of anhydrous FeCl or ZrCl , followed by
3
4
heat treatment, were unsuccessful.
AFM examination showed that the outer surface
layer consisted of irregularly shaped particles, arranged
with no periodicity (Fig. 5a).
The inner surface was formed by tightly packed
elongated cylindrical particles up to 2.5 µm in radius
(
Fig. 5b), occasionally with bumps Ӎ250 nm across
(
Fig. 5c), or it was smooth, with height variations
2
within 10–15 nm over a 20-µm area (Fig. 5d). The
smooth surface was typical of the lower part of the tubes,
and the wavy surface was typical of the upper part, with
a sharp transition from one surface morphology to the
other.
1
µm
(
a)
The x-ray pattern (CuK radiation) from ground
α
tubes (Fig. 6) showed three weak peaks (2θ = 43.7°,
5
1.2°, and 56.66°) attributable to corundum. From
SAD patterns, the phase composition of the tubes could
not be determined.
Comparison of the IR spectrum of the tubes (Fig. 7)
with earlier data [6] led us to conclude that the tubes
consisted of Al O .
2
3
GROWTH MECHANISM
Based on the present results, Al O macrotubes
2
3
2
0 µm
seem to grow by the following mechanism: After expo-
(
b)
sure to water vapor, the hydrolyzed surface layer of
AlCl particles contains OH groups and water mole-
3
Fig. 3. SEM micrographs illustrating the (a) external and
b) internal morphology of macrotubes.
cules. During subsequent heating, the hydrolysis prod-
ucts decompose, and water rapidly vaporizes from the
top layer of the sample. The resulting gas-tight crust
(
analysis system), scanning electron microscopy (SEM) consists of Al O and aluminum hydroxychlorides. The
2
3
(
JEOL and Amray instruments), electron microprobe underlayer consists of undecomposed aluminum
analysis (ISIS Oxford Instruments microanalyzer oper- hydroxychlorides and unreacted AlCl . Further heating
3
ated at 20 kV), atomic-force microscopy (AFM) gives rise to the decomposition of the hydroxychlorides
(
P47-SPM-MDT Solver instrument), and IR spectros- and AlCl hydrolysis under the crust, leading to the for-
3
copy (Spectrum-2000 Perkin-Elmer Fourier-transform IR mation of water vapor and hydrogen chloride. In addi-
spectrophotometer equipped with an IR microscope and tion, the AlCl vapor pressure becomes noticeable at
3
liquid-nitrogen-cooled HgCdTe detector). The tube mate- 150–160°C and increases rapidly at higher temperatures.
INORGANIC MATERIALS Vol. 37 No. 10 2001

HOLLOW ALUMINA MACROTUBES
1039
1
00 µm
200 µm
(
a)
(b)
1
00 µm
100 µm
(
c)
(d)
Fig. 4. SEM micrographs of typical (a–c) beaded and (d) bell-mouthed macrotubes: (a, b, d) external view, (c) inner surface.
(
a)
(b)
nm
nm
2
1
00
00
1000
5
00
8
00
4
000
6
00
3000
4
000
1
000
2000
3
000
4
00
8
00
1
000
2000
nm
6
nm
00
1000
2
00
0
0
4
00
2
00
0
0
(
c)
(d)
nm
00
1
nm
400
8
00
00
00
00
7
6
2
00
5
1200
1000
0
8
nm
00
6
00
4
00
200 400
0
3
00
00
00
2
400
00
nm
3
1
200
00
1
0
0
Fig. 5. AFM topographic images of the (a) outer and (b–d) inner surfaces of macrotubes.
INORGANIC MATERIALS Vol. 37 No. 10 2001

1
040
BERDONOSOV et al.
stitute the white deposit forming on the wall of the
5
4
3
2
quartz reactor). Since the submicron particles have a
high surface energy, they readily become attached to the
crater edge, feeding the further growth of the macrotube.
4
3.70
As a result, the tube grows rapidly in length,
whereas its outer diameter remains unchanged. Since
5
1.20
56.66
the exhalation of AlCl vapor from the sample is
3
1
unsteady, the tubes consist of microscopic regions dif-
fering in dimensions. If part of a tube breaks off acci-
dentally, subsequent growth is often accompanied by a
reduction in tube diameter (Fig. 2).
0
2
4.5
34.5
44.5
54.5
64.5 74.5
θ, deg
2
If the starting charge contains FeCl , the dopant
Fig. 6. X-ray pattern of ground macrotubes.
3
vaporizes in the final stage of heat treatment, because
the vapor pressure of FeCl is much lower than that of
T, %
3
AlCl . If the process is run in hydrogen and the temper-
9
9.3
3
(
a)
9
8
7
6
5
4
3
0
0
0
0
0
0
0
ature is high enough for rapid reduction of FeCl , we
3
find metallic iron at the free tube end.
1637.8
2
309.6
Al O tubes grown by the method described above are
2
3
potentially attractive as catalyst supports and substrates
for conducting films. This method is probably also appli-
cable for producing macrotubes of other oxides.
3
350.8
690.2
2
3.1
1
21.1
CONCLUSION
(
b)
1
1
10
00
Poorly crystallized alumina macrotubes 50 µm to
6 cm in length, 10 to 300 µm in outer diameter, and 2 to
0 µm in inner diameter were produced by thermal
decomposition of partially hydrolyzed AlCl in a gas-
1
635.1
2
361.8
6
9
8
7
0
0
0
3
eous atmosphere.
3385.9
6
65.7
6
1.8
4
000
3000
2000
1500
1000
600
ACKNOWLEDGMENTS
Wavenumber, cm–1
We thank E.V. Antipov, A.V. Bogdanov, A.G. Bog-
danov, I.A. Budashov, Yu.A. Velikodnyi, I.N. Kuroch-
kin, F.M. Spiridonov, and M.A. Prokof’ev for their
assistance with this study and useful discussions and
Fig. 7. IR spectra of (a) macrotubes and (b) Al O [6].
2
3
As a result of the rise in internal pressure, the vapor N.N. Oleinikov for valuable suggestions.
breaks through the crust, producing a “microvolcano”
on the sample surface, with the elevated crater edge
This work was supported by the Russian Foundation
for Basic Research, grants no. 99-03-32646 and
no. 00-03-32644.
consisting of Al O .
2
3
The release of water, AlCl , and HCl vapors is
3
accompanied by further formation of Al O , which is
2
3
REFERENCES
nonvolatile under the conditions of this study. As a
1
2
. Ting, J.-M. and Lan, J.B.C., Beaded Carbon Tubes, Appl.
Phys. Lett., 1999, vol. 75, no. 21, pp. 3309–3311.
. Dagani, R., Much Ado about Nanotubes, Chem. Eng.
News, 1999, vol. 77, no. 2, pp. 31–34.
3. Rakov, E.G., Preparation of Carbon Nanotubes, Usp.
Khim., 2000, vol. 69, no. 1, pp. 41–59.
result, aggregates of Al O particles less than 1 µm in
2
3
size are formed in the vapor phase (such particles con-
Microanalysis data for macrotubes grown from AlCl (I) and
3
9
5 mol % AlCl + 5 mol % FeCl (II)
3
3
4
. Ma, X., Wang, E., Zhou, W., et al., Polymerized Carbon
Nanobells and Their Field-Emission Properties, Appl.
Phys. Lett., 1999, vol. 75, no. 20, pp. 3105–3107.
Weight percent
Element
I
II
5
. Ajayan, P.M., Nugent, J.M., Siegel, R.W., et al., Growth
of Carbon Micro-Trees, Nature (London), 2000,
vol. 404, no. 6775, p. 243.
Al
O
32–42
44–60
3–7
17–26
22–30
1–3
Cl
Fe
6. Hummel, D.O. and Schol, F., Atlas of Polymer and Plas-
tics Analysis, Vienna: Henzer, 1982, vol. 2, part A, II,
p. 891, Spektrum 4444.
–
32–75
INORGANIC MATERIALS Vol. 37 No. 10 2001