ARTICLE IN PRESS
1588
H. Ogihara et al. / Journal of Solid State Chemistry 182 (2009) 1587–1592
of elements can be obtained. One of the important features of our
synthesis process is that not only metal alcoxides but also
inorganic metal salts (e.g., metal nitrates) can be used as raw
materials, which lead to the formation of oxide nanotubes
containing a variety of elements. Most metal salts are cheaper,
more available, and more stable than metal alcoxides. In our
process, templates are coated by immersion into solutions
containing the metal precursor, followed by drying of the solution.
Metal oxide layers were deposited on carbon nanofiber templates
through the simple processes, although the coating mechanism
(e.g., the reason why only carbon nanofiber is selectively coated
with oxide layers) remains unclear. Thus, in the present study, the
synthesis and characterization of metal oxides, and the coating
mechanism of carbon nanofiber templates will be discussed on
the basis of TEM, SEM, X-ray diffraction (XRD), thermogravimetry
(TG) and N2 adsorption measurements.
100 nm
100 nm
Fig. 1. TEM images of (a) VGCFs and (b) VGCFs(gr).
Fig. 1 shows TEM images of VGCFs and VGCFs(gr). Both VGCFs and
VGCFs(gr) are straight and have a diameter of 100–150 nm. XRD
patterns of VGCFs and VGCFs(gr) are shown in Fig. 2. Both XRD
patterns exhibit a diffraction line at 20–301, which is attributed to
graphite (002). The diffraction line of VGCFs(gr) is much sharper
than that of VGCFs, indicating that VGCFs(gr) have highly
graphitized structures. It is well known that the graphitization
degree of carbonaceous materials increases upon heat treatment
in an inert atmosphere. Thus, it is reasonable to assume that the
heat treatment increased the graphitization degree of the carbon
nanofibers. Their specific surface areas are almost the same
(18 m2/g for VGCFs and 23 m2/g for VGCFs(gr)). These results
suggest that the shape and the surface area of carbon nanofibers
do not change upon heat treatment, while their graphitization
degree increases dramatically.
2. Experimental
Two types of vapor grown carbon fibers (VGCFs, Showa Denko
Co.) are used as templates. One is treated at 3273 K under an inert
gas atmosphere, while the other is not subjected to this heat
treatment. Carbon fibers obtained with and without heat treatment
will hereafter be referred to as VGCFs(gr) and VGCF, respectively.
Before being used as templates, the carbon fibers were suspended in
a concentrated H2SO4/HNO3 mixture (3:1 v/v) and ultrasonicated for
2 h. After filtration, the carbon fibers were washed with ion-
exchanged water three times, and dried at 393 K. Through the acid
treatment, functional groups such as hydroxyl groups and carboxyl
groups are formed onto the surface of the carbon fibers [23].
A set of VGCF or VGCF(gr) templates was placed in a suction
filtering unit (Buchner funnel with filter paper), and a metal nitrate
solution (Fe(NO3)3 ꢀ 9H2O, Co(NO3)2 ꢀ 6H2O, Ni(NO3)2 ꢀ 6H2O,
La(NO3)3 ꢀ 6H2O, Fe(NO3)3 ꢀ 9H2O or Mn(NO3)2 ꢀ 6H2O), diluted with
ethanol (total metal nitrate concentration: 0.3 M), was dropped
onto the templates. Immediately, the nitrate solution infiltrated
into the surface-connected pore of the fibrous structure. The
excess solution was removed by filtration. The sample obtained
was dried in air at 573 K. During the drying process, the ethanol in
the pores of the fibrous structure evaporated. The metal nitrates
that were left behind adsorbed onto the template surfaces and
transformed into metal oxides through thermal decomposition.
This process was repeated 10 times. Finally, the templates were
removed by calcination in air at 773–923 K for 5 h (the standard
calcination temperature: 773 K).
3.2. Mono metal oxide nanotubes
Iron, cobalt, and nickel oxide nanotubes were prepared by
using VGCFs as templates. The crystallite structure of the
nanotubes was identified by XRD pattern measurements (Fig. 3).
For the iron oxides, the XRD patterns corresponded to
no other characteristic diffraction lines, such as
Fe3O4, were observed. For the cobalt and nickel oxides, diffraction
lines corresponded to Co3O4 and NiO were confirmed. In these
XRD patterns, no diffraction lines attributable to graphite were
observed, indicating that the VGCF templates were removed
completely. Average crystallite sizes estimated on the basis of
Scherrer’s equation were 34.0 nm for Fe2O3, 17.9 nm for Co3O4 and
11.7 nm for NiO. By applying the template method to VGCFs,
nanoscale crystallites of metal oxides were formed.
Fig. 4 shows TEM images of the Fe2O3, Co3O4 and NiO prepared.
Clearly, they all have a nanotubular structure, but the nanotube
walls of NiO are not well formed. Fe2O3 nanotubes seem to be
made up of flat particles ca. 20 nm in thickness. Meanwhile, the
walls of the Co3O4 nanotubes consist of granular particles
10–30 nm in size. In the case of NiO, small particles ca. 10 nm in
diameter have aggregated into a nanotube-like structure. These
TEM images show that the type of constituent elements
influences the nanotube shape. The particle sizes of Co3O4 and
NiO observed in TEM images are almost the same as the average
crystallite sizes estimated from XRD patterns as described above.
On the other hand, in the case of Fe2O3 nanotubes, since the Fe2O3
particles are not spherical but plate-like, a comparison between
particle size (estimated from TEM images) and average crystallite
size (estimated from XRD patterns) is difficult. The crystallite size
estimated on the basis of Scherrer’s equation means the thickness
of layers that are perpendicular to a certain crystal face. Therefore,
for a spherical crystallite, the particle size estimated from the TEM
image often corresponds to the crystallite size estimated by XRD.
In contrast, for a crystallite with anisotropic shape, the particle
size does not always correspond to the crystallite size.
a
-Fe2O3 and
g-Fe2O3 and
Powder X-ray diffraction patterns were recorded with
a
diffractometer (Rigaku, RINT Ultima+) using CuK radiation.
a
TEM images were obtained using a H-800 (Hitachi) operated at
200 kV. SEM images were obtained with a JSM-6300 or JSM-6500F
(JEOL) using an accelerating voltage of 5 kV. Thermogravimetry
analysis was performed with a TG-8120 (Rigaku) thermogravi-
metric analyzer. Samples were placed in an Al2O3 holder and the
temperature was increased at 10 K/min under flowing air.
Nitrogen adsorption measurements were performed on a Auto-
sorb 3 (YUASA IONICS) sorption analyzer. Prior to the sorption
measurements, samples were degassed under vacuum at 573 K for
3 h. Surface areas were calculated by Brunauer–Emmet–Teller
(BET) method. Elemental analysis for C, H, and N atoms were
carried out with a CHN corder MT-6 (YANACO).
3. Results and discussion
3.1. Carbon nanofiber templates
The properties of the carbon nanofiber templates were
characterized by TEM, XRD, and surface area measurements.
An increase in temperature induces sintering of Fe2O3 oxides
above a certain temperature. Thus, the thermal stability of the