Nickel formate route to the growth of carbon nanotubes

Article abstract of DOI:10.1021/jp047898p

Nickel formate serves as an ideal catalyst precursor for surface-bound thermal and plasma-enhanced CVD growth of carbon nanotubes. As the method of deposition of the Ni catalyst is solution-based, it benefits from a low-cost, large-area growth, ease of operation and suitability for coating substrates with complex shapes and structures.

Full text of DOI:10.1021/jp047898p

18446
J. Phys. Chem. B 2004, 108, 18446-18450
Nickel Formate Route to the Growth of Carbon Nanotubes
Junfeng Geng, Hongwei Li, Vladimir B. Golovko, Douglas S. Shephard,
David A. Jefferson, and Brian F. G. Johnson*
Department of Chemistry, Cambridge UniVersity, Lensfield Road, Cambridge CB2 1EW, U.K.
Stephan Hofmann, Britta Kleinsorge, and John Robertson
Department of Engineering, Cambridge UniVersity, Trumpington Street, Cambridge CB2 1PZ, U.K.
Caterina Ducati
Department of Materials Science and Metallergy, Cambridge UniVersity, Pembroke Street, CB4 3QZ, U.K.
ReceiVed: May 17, 2004
Nickel formate serves as an ideal catalyst precursor for surface-bound thermal and plasma-enhanced CVD
growth of carbon nanotubes. As the method of deposition of the Ni catalyst is solution-based, it benefits from
a low-cost, large-area growth, ease of operation and suitability for coating substrates with complex shapes
and structures.
Introduction
for growth of carbon nanotubes. In our previous report, we
demonstrated that using this compound, high-purity single-wall
Carbon nanotubes (CNTs) have been considered as ideal
candidates for many applications due to their unique atomic
structure, an inherent small tip radius, high surface area and
aspect ratio, excellent chemical inertness and mechanical
strength.1
1
2
carbon nanotubes were produced in high yield. In fact, nickel
formate may be used not only for single-wall nanotubes but
also for versatile growth of carbon nanotubes with varied
structures. In this paper, we wish to address the issue of why
nickel formate is such an ideal catalyst precursor on one hand,
and to demonstrate surface-bound growth of multiwalled carbon
nanotubes on the other. The choice of the surface growth on
the flat SiO2/Si substrate is based on a common recognition
that such grown nanotubes are potentially significant in micro-
electronic applications, such as for electron field emission, flat
,2
Among the many growth techniques, chemical vapor deposi-
tion (CVD) or plasma-enhanced chemical vapor deposition
(
PECVD) offer controlled surface-bound growth on flat sub-
3,4
strates. In a CVD-based process, catalyst and catalyst prepara-
tions are essential because the growth occurs at the interface of
5
the catalyst and the hydrocarbon vapor. Hence, formation of
an active, uniform catalyst layer is a critical step in a successful
surface-bound CVD process.
1
3-15
panel display and supercapacitors.
Experimental Section
Deposition of catalyst may be realized by a physical method,
as for example, using thermal evaporation or magnetron
sputtering to generate a homogeneous thin metal film, which
in a subsequent growth step, breaks into small metal islands as
Deposition of the Catalyst. For deposition of the nickel
catalyst, silicon wafer was first cleaned by ethanol and acetone
in an ultrasonic bath and rinsed with distilled water. Nickel
formate dihydrate was purchased from Degussa, as a fine green
powder. Deposition of the catalyst was carried out by simply
casting a drop of the aqueous or methanol solution of nickel
formate onto the substrate. The samples were then left to dry
in air. The concentration of the solution may be varied,
according to the growth purpose and catalyst loading. A typical
catalyst.6,7 Although physical methods have proven very effec-
tive, they require a vacuum facility to process catalyst deposi-
tion, and in general, are not suitable for coating a substrate with
complex shapes. Alternatively, a chemical method may be used,
as for example, introducing metal colloidal nanoparticles (Fe,
Co, and Ni) to the catalyst.8,9 Metal colloidal nanoparticles may
-
3
-2
concentration falls in the range 5 × 10 to 2.2 × 10 M. The
thickness of the formate film depends on the amount of solution
on the surface. To keep the film uniform, a trace amount of
anionic surfactant, such as lauric acid, may be mixed into the
solution. The Ni nanoparticles were formed in situ by thermal
decomposition of the formate precursor in the CNTs growth
process.
be pre-controlled in a narrow size distribution in either the
synthesis process or using an after-process technique for size
selection. However, the colloids are frequently air-sensitive,
making their manipulation difficult and, to some extent, limiting
the large-scale production of carbon nanotubes. Another com-
monly used method is through a catalyst precursor, such as an
inorganic or organometallic salt, leading to the desired metal
nanoparticles by the thermal decomposition of the precursor at
Growth of the Carbon Nanotubes. The growth methods
7
,15
high temperatures.1
0,11
are basically the same as described previously.
In brief, a
The precursor method takes advantages
stainless steel vacuum chamber was used with a base pressure
of low-cost, good stability in air, ease of operation and
commercially available starting materials. It is also suitable for
coating 3D substrates with complex shapes and structures.
We have found that nickel formate may serve as an ideal
catalyst precursor for formation of the desired Ni nanoparticles
-
6
below 10 mbar. A 20 nm thick SiO2 layer was grown by
thermal oxidation or low-temperature electron cyclotron reso-
nance (ECR) onto a polished n-type Si(100) substrate. The
samples were normally heated for 15 min to reach the desired
temperatures using a resistively heated graphite stage. For the
plasma-enhanced growth, samples with freshly deposited catalyst
*
Corresponding author. E-mail: bfgj1@cam.ac.uk.
1
0.1021/jp047898p CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/06/2004

Nickel Formate Route to Carbon Nanotube Growth
J. Phys. Chem. B, Vol. 108, No. 48, 2004 18447
Figure 1. (Upper left): image of the Ni nanoparticles on the surface of silicon substrate under AFM. The particles are formed from thermal
-
3
decomposition of nickel formate precursor at 500 °C. The employed solution of the precursor has a concentration of 5 × 10 M. AFM operation
was carried out by a tapping mode. (Lower left): topographical height measurements for the selected nanoparticles cross-lined in the upper image,
showing the typical sizes of the particles. (Upper right): AFM image of the larger and denser Ni nanoparticles formed at identical decomposition
-
2
conditions but deposited from a higher concentration of the precursor solution (2.2 × 10 M). The solution also contains a trace amount of lauric
acid. (Lower right): AFM topographical height data showing the typical particle size and size distribution.
precursor were first heated typically to 460 °C in 1.1 mbar of
NH3 (100 sccm). A voltage of 800 V between the sample heater
and the gas inlet was employed to generate the dc discharge
current. The feed gas, acetylene (C2H2, grade 1.5, 55 sccm) was
introduced into the chamber as the carbon feedstock immediately
after the discharge started. Growth normally lasted 20 min. The
C2H2/NH3 ratio and the total pressure were varied from one
growth to another depending on growth purpose. A stable
discharge of typically 20-40 mA was maintained for a given
deposition.
Characterization. After growth, SEM observations were
carried out using a JEOL 6340-FESEM system operated at 5
kV. The TEM was performed on a JEOL JEM-3010x operated
at 300 kV. The TEM sample was prepared by scratching the
tubes off the substrate and then dispersed into toluene assisted
by ultrasonication. A drop of such prepared solution was cast
onto a holey carbon film on a copper grid for the examination.
The Raman studies were performed using a microspectrometer
with an incident Ar ion laser at λ ) 514.5 nm and the Renishaw
software. Atomic force spectroscopy (AFM) images were
obtained by the tapping mode using a Digital Instrument with
a nanoscope dimension 3100 controller.
In the presence of hydrogen, the decomposition temperature falls
1
6
to 180-200 °C. Unlike many other nickel precursors, nickel
formate decomposes directly to metallic Ni particles and gas
species (H , H O, CO, and CO ) without forming the oxide
2
2
2
(NiO) intermediate in a vacuum or a gas flow system. This is
attributed to a self-redox reaction in which a nickel atom obtains
two electrons to be reduced to the metallic state, whereas the
two ligands lose two electrons to be oxidized into gaseous
species. This excellent characteristic distinguishes nickel formate
from other nickel precursors and makes the normally required
reducing agents such as H2 no longer necessary in the formation
1
6-18
of phase-pure metallic nickel.
It has been established that the catalytic properties of nickel
strongly depend on the size of the particles. In our previous
12
report on the growth of single-wall carbon nanotubes, the
nearly monodispersed Ni nanoparticles are extremely small, with
a typical size of ∼1.0 nm. In that case, the small metal particles
result from the well-dispersed nickel formate precursor in the
fine and mesoporous silica powder. As to the flat SiO2/Si
substrate, however, we find that Ni particles may be formed in
a large-size scope, typically between 1 and 30 nm, depending
on experimental conditions. Under the appropriate conditions,
a specific particle size with narrow size distribution may be
obtained. This has been confirmed by the atomic force
spectroscopy (AFM) measurements and from the results of the
nanotubes growth.
Results and Discussion
Formation of the Ni Nanoparticles. In the solid state, nickel
formate typically exists in the form of a dihydrate, Ni(HCOO)2‚
Figure 1a shows the AFM topographical image of small and
discrete Ni nanoparticles prepared by annealing the silicon
substrate coated with a thin layer of nickel formate. The
calcination was conducted at 500 °C, for 2 min and under an
argon atmospheric pressure. As usual for any AFM operation,
the accurate particle sizes are measured from their topographic
heights. Some typical heights are shown in Figure 1a. In this
case, the Ni particles are extremely small, with a typical size
between 1 and 2 nm. These particularly small metal particles
may be used as ideal catalytic seeds in the growth of single-
wall carbon nanotubes. In contrast, Figure 1b shows larger and
2
H2O. Our measurements using thermal gravimetric analysis
(TGA) indicate that the dihydrate loses the crystallized water
at a temperature of ca. 160 °C and thermally decomposes
between 230 °C and 260 °C, which is consistent with literature
reports using TGA analysis17 and the differential scanning
calorimetry (DSC) measurements.1
8

18448 J. Phys. Chem. B, Vol. 108, No. 48, 2004
Geng et al.
Figure 2. SEM images showing growth behavior of the carbon nanotubes with different catalyst densities on the SiO
2
/Si substrate: a1, nanotubes
grown from a low surface coverage of the Ni nanoparticles; a2, image of the same sample at a higher magnification; b1, nanotubes grown from
identical CVD conditions as those shown in (a) but with a higher catalyst density on the surface; b2, image of the sample b1 at a higher microscope
magnification. The CVD growth was performed at 530 °C with NH at 100 sccm, C H at 55 sccm, and a total pressure of 60 mbar.
3 2 2
denser Ni particles produced from a higher concentration of
the solution (see the figure caption for deposition conditions).
The topographic height profile indicates that the particles are
fairly uniform, with a typical size of about 20 nm and a size
distribution from 15 to 30 nm. Note that (a) and (b) of Figure
1
may represent two extreme examples of the prepared Ni
nanoparticles with respect to the particle size. We find that the
particles with a size between these two extremes may also be
prepared by adjusting the concentration of the solution, the
quantity, and the method of deposition and annealing.
Growth of Carbon Nanotubes by CVD. In Figure 2 are
shown the morphologies of the carbon nanotubes produced by
thermal CVD. Figure 2a shows the SEM images of the
nanotubes from a low density of Ni nanoparticles. In compari-
son, Figure 2b shows the nanotubes grown with a higher catalyst
loading, which yields a greater abundance of the tubes under
identical CVD conditions, and indicates a strong relationship
between the productivity and the catalyst loading. The tubes
are, in general, uniform, and the diameter distribution falls in
the range 20-30 nm. The tubes contain metal nanoparticles in
the tip, suggesting the possible tip growth mechanism. The
higher magnification images reveal that the morphology and
quality of the tubes obtained at the higher catalyst density are
very much similar to those obtained at the lower one, suggesting
that the increase of the productivity takes place without sacrifice
of the tube quality. Laser Raman examination (Figure 3) gives
the feature of the multiwalled carbon nanotubes, with the D
Figure 3. Raman profile of the CVD nanotubes showing the graphitic
structure of the tubes (G peak) and the structural defect (D peak). The
test was carried out using an Ar ion laser at λ ) 514.5 nm.
reveals that the Ni nanoparticles are sitting at the tube tips,
indicating the tip growth mechanism. Examination by laser
Raman spectroscopy indicates the graphitic structure, with the
characteristic G and D scattering peaks centered at 1600 and
-
1
1375 cm , respectively. TEM observation was carried out
because the SEM images did not provide significant information
about the sidewalls of those vertically aligned tubes. The result
shows that the nanotubes are multiwalled and most of them have
an inner diameter of about 3-6 nm. This implies that most of
the Ni nanoparticles are ca. 3-6 nm in size. Although defects
or dislocations are frequently observed in the graphitic structure,
the tubes are, in general, well crystallized in terms of the
concentric graphene layers. Many tubes also contain a Ni particle
encapsulated in the cavity. One of the encapsulated examples
-
1
-1
peak centered at 1360 cm and the G peak at 1603 cm .
Growth of Carbon Nanotubes by PECVD. A SEM image
of the PECVD nanotubes is shown in Figure 4. It is seen that
the abundant, vertically aligned carbon nanotubes cover most
of the substrate surface. The image at a higher magnification

Nickel Formate Route to Carbon Nanotube Growth
J. Phys. Chem. B, Vol. 108, No. 48, 2004 18449
formed from the formate precursor. First, we believe that among
the many factors, the relatively low temperature of dissociation
of the precursor has played an important role. Apart from the
result presented in this paper, this characteristic may also be
understood from other investigations. Nickel formate has been
used as a precursor to metallic nickel in the preparation of nickel
1
6
tetracarbonyl, where the requirement of the purity of the
formed metal is extremely high. In the area of so-called thick-
technology, this compound has been used to study the formation
of ideally spherical, agglomeration-free Ni particles at microme-
ter scale for applications in hybrid electronic products such as
1
9,20
ceramic capacitor and batteries.
In those studies, nickel is
replacing noble metals such as Pt, Pd, Ag, and Au, for use as
electrode and thick-film material because of its good conductiv-
2
1
ity and lower cost. Xia et al. have employed nickel formate
as a precursor to produce micrometer scale Ni particles by an
ultrasonic spray pyrolysis technique. They conclude that phase-
pure nickel can be obtained at temperatures as low as 350 °C
by the thermal decomposition of the precursor, which is the
lowest temperature compared with using other nickel salts in
the spray pyrolysis. They also find that nickel can be obtained
at a residence time as short as 0.1 s at 600 °C, which is 1-2
orders of magnitude smaller than the times reported earlier using
2
2
Figure 4. Vertically grown carbon nanotubes on the SiO
by the plasma-enhanced CVD method using Ni nanoparticles as catalyst.
The growth was conducted at 460 °C, with 1.1 mbar of NH (100 sccm)
and C (55 sccm).
2
/Si substrate
Ni(NO3)2 and NiCl2 (1-100 s). Stopic et al. have reported
the effect of nickel formate on the decomposition temperature
of dehydrated NiCl2 in their differential thermal analysis (DTA)
studies. DTA curves reveal that the reduction of the dehydrated
NiCl2 by hydrogen starts at 375 °C. While in the presence of a
small amount of nickel formate (equivalent to 0.1 wt % Ni),
this temperature drops to 275 °C, which matches the starting
decomposition temperature of nickel formate. In comparison,
our results indicate that much smaller Ni particles (1-30 nm)
can also be formed from nickel formate for growth of carbon
nanotubes.
3
2
H
2
In addition, the decomposition of nickel formate tends to
produce Ni nanoparticles at a narrow size distribution. This may
be seen from the AFM measurements and the growth results.
In calcinations, it appears that the rapidly formed small
nanoparticles (e3 nm) at short calcination times and lower
temperatures (e.g., e2 min at 500 °C) tend to adsorb some other
species around them. This is particularly the case when a trace
amount of surfactant with a long carbon chain, such as lauric
acid, was mixed into the formate solution. This observation
allows us to speculate that in the formate-derived Ni nanopar-
ticles, some carbon-containing species formed from the thermal
decomposition of the precursor might have helped to passivate
the newly formed Ni particles from agglomeration, at least at
the early stage of the nucleation of the particles.
Besides the temperature and size effects, the geometry of the
particles is another factor in the determination of the growth
behavior. In this work, at the growth temperatures, 530-545
°C for CVD and 460 °C for PECVD, the Ni particles (3-30
nm) are thought to be in solid state (compared with the melting
point of the bulk nickel at 1455 °C). For the solid-state particles,
they do not necessarily take a spherical shape. Instead, nano-
particles with mixed geometries were formed. This geometrical
effect may not be particularly essential for growth of multiwalled
nanotubes, as the produced Ni particles are relatively large (e.g.,
Figure 5. HRTEM image of a carbon nanotube vertically grown by
the PEVCD method. This image shows the crystal fringes of the
encapsulated nickel single crystal and the graphene layers. Using the
graphene layers (d ) 0.344 nm) as a reference, the fringe spacing of
the metal is measured as 0.203 nm, corresponding to Ni(111) plane.
is shown in Figure 5. In this case, the metal nanoparticle is
elongated along the tube axis. Using the graphene layers (d )
0
.344 nm) as a reference, we measured the fringe spacing of
the encapsulated metal particle at 0.203 nm, corresponding to
the Ni(111) crystallographic facet. This image unambiguously
indicates that the Ni particle is a pure single crystal, as evidenced
by the lack of grain boundaries within the particle.
5-30 nm), causing the intake and precipitation of carbon atoms
to occur at a larger area of the grain boundaries or stage edges.
Also, the normally found tip-growth mechanism allows the
change of the shape of the particles during the growth
5
Discussion. The effectiveness of the CNTs growth in this
procedure. However, for growth of single-wall nanotubes, the
work is mainly attributed to the highly active Ni nanoparticles
Ni nanoparticles are required to be extremely small (1-3 nm)

18450 J. Phys. Chem. B, Vol. 108, No. 48, 2004
Geng et al.
and the nanotubes normally nucleate from the substrate-anchored
particles by the base growth mechanism.23 This situation makes
only a small surface part of a nanoparticle exposed to the gas
phase and hence very limited surface area is available for carbon
uptake at the gas-nanoparticles interface. In this case, the
geometry factor becomes more important. Gordon et al. have
studied small Ni nanoparticles with atomic numbers particularly
Acknowledgment. We thank financial support from Thomas
Swan Co. & Lt., U.K., and the EU project CARDECOM GRD1-
2001-41830. We also thank Dr. Ian A. Kinloch, Professor Alan
H. Windle and Dr. Yali Li (Department of Materials Science,
University of Cambridge), Mr. Stephen Cash and Dr. Charanjeet
Singh (Thomas Swan Co. & Ltd., UK), and Dr. Milo S. P.
Shaffer (Department of Chemistry, Imperial College, London)
for helpful discussions.
between 55 and 147 corresponding to a particle size of 1.0 and
4,25
1
.5 nm, respectively.2
In comparison of two geometrical
configurations, cubooctahedra or icosahedra, they conclude that
the cohesive energy between the nanoparticles of the same size
and different geometry is very small (∆E e 500 cal/mol),
indicating that the coexistence of both structures is possible.
Considering the catalysis process in the growth of single-wall
nanotubes, it is reasonable to assume that both structures would
assist the nucleation of the tubes but lead to different tube
structures or chiralities. The geometrical difference of metal
particles may be a main responsible reason of why single-walled
nanotubes obtained from a catalyst precursor route in CVD are
always mixed in structure, even though the tube diameter may
be controlled within a narrow range.
References and Notes
(
1) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Science 2002,
97, 787.
(2) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. C. Science of
Fullerenes and Carbon Nanotubes; Academic Press: San Diego, 1996.
2
(3) Ural, A.; Li, Y.; Dai, H. Appl. Phys. Lett. 2002, 81, 3464.
(4) Ren, Z. F.; Huang, Z. P.; Xu, J. W.; Wang, J. H.; Bush, P.; Siegal,
M. P.; Provencio, P. N. Science 1998, 282, 1105.
(5) Helveg, S.; Lopez-Cartes, C.; Sehested, J.; Hansen, P. L.; Clausen,
B. S.; Rostrup-Nielsen, J. R.; Abild-Pedersen, F.; Norskov, J. K. Nature
2004, 427, 426.
(6) Chhowalla, M.; Ducati, C.; Reppesinghe, N. L.; Teo, K. B. K.;
Amaratunga, G. A. J. Appl. Phys. Lett. 2001, 79, 2079.
(7) Hofmann, S.; Ducati, C.; Kleinsorge, B.; Robertson, J. Appl. Phys.
Lett. 2003, 83, 4661.
(
8) Kleinsorge, B.; Golovko, V. B.; Hofmann, S.; Geng, J.; Jefferson,
Conclusions
D.; Johnson, B. F. G. ChemComm 2004, 1416.
(9) Ago, H.; Komatsu, T.; Ohshima, S.; Kuriki, Y.; Yumura, M. Appl.
Phys. Lett. 2000, 77, 79.
10) Edwards, A. B.; Garner, C. D.; Roberts, K. J. J. Phys. Chem. B
997, 101, 20.
11) Emmenegger, C.; Bonard, J. M.; Mauron, P.; Sudan, P.; Lepora,
A.; Grobety, B.; Zuttel, A.; Schlapbach, L. Carbon 2003, 41, 539.
12) (a) Geng, J.; Singh, C.; Shephard, D. S.; Shaffer, M. S. P.; Johnson,
B. F. G.; Windle, A. H. ChemComm 2002, 2666. (b) U.K. patent application
No. 0216654.4.
13) Choi, Y. S.; Cho, Y. S.; Kang, J. H.; Kim, Y. J.; Kim, I. H.; Park,
S. H.; Lee, H. W.; Hwang, S. Y.; Lee, S. J.; Lee, C. G.; Oh, T. S.; Choi,
J. S.; Kang, S. K.; Kim, J. M. Appl. Phys. Lett. 2003, 82, 3565.
14) Niu, C.; Sichel, E. K.; Hoch, R.; Moy, D.; Tennent, H. Appl. Phys.
Lett. 1997, 70, 1480.
(15) Hofmann, S.; Ducati, C.; Robertson, J.; Kleinsorge, B. Appl. Phys.
Lett. 2003, 83, 135.
16) Gilliland, W. L.; Blanchard, A. A. In Inorganic Syntheses; Fernelius,
W. C., et al., Eds.; McGraw-Hill Book Co. Inc.: New York, 1946; Vol. II,
Chapter 75, pp 234.
(17) The Merck Index, 13th ed.; Merck Research Laboratory, Division
of Merck & Co.: Whitehouse Station, NJ, 2001; p 1166.
18) Edwards, A. B.; Garner, C. D.; Roberts, K. J. J. Phys. Chem. B
997, 101, 20.
19) Messing, G. L.; Zhang, S. C.; Jayanthi, G. V. J. Am. Ceram. Soc.
1993, 76, 2707.
(20) Patil, P. S. Mater. Chem. Phys. 1999, 59, 185.
21) Xia, B.; Lenggoro, I. W.; Okuyama, K. J. Am. Ceram. Soc. 2001,
4, 1425.
22) Stopic, S.; Nedeljkovic, J.; Rakocevic, Z.; Uskokovic, D. J. Mater.
Res. 1999, 14, 3059.
(23) Li, Y.; Kim, W.; Zhang, Y.; Rolandi, M.; Wang, D.; Dai, H. J.
Phys. Chem. B 2001, 105, 11424.
In this paper we have presented the following results. (1)
Nickel formate serves as an ideal precursor for formation of Ni
nanoparticles with a controllable size and a narrow size
distribution. These freshly formed nanoparticles are catalytically
very active, leading to a highly efficient growth of carbon
nanotubes. (2) Growth of randomly distributed and vertically
aligned carbon nanotubes on flat surface can be realized by
either CVD or PECVD using the nickel formate route. As our
method of deposition of the catalyst is solution-based, it has
the merits of low-cost, ease of operation, large-area growth, and
suitability for coating 3D substrates with complex shapes and
structures. (3) The reason nickel formate is an ideal catalyst
precursor is analyzed from a chemistry point of view. The self-
redox feature of the decomposition reaction, together with the
relatively low decomposition temperature (230-260 °C), dis-
tinguishes nickel formate from other common nickel precursors.
In the case of nickel formate, it is not necessary to have external
reducing agents such as H2 for formation of the desired Ni
nanoparticles. (4) The formed Ni nanoparticles can be nearly
monodispersed in size, possibly arising from a tendency of the
newly formed nanoparticles to chemically adsorb a thin layer
of carbon-containing species produced from the precursor. This
protective layer may passivate the particles from agglomeration
particularly at the early stage of nucleation. (5) Although the
particles size may be controlled within a narrow range, different
geometrical structures coexist. We believe that this structural
difference of metal leads to the structural difference of nano-
tubes, particularly in the case of growth of single-wall carbon
nanotubes.
(
1
(
(
(
(
(
(
1
(
(
8
(
(
24) Gordon, M. B.; Cyrot-Lackmann, F.; Desjonqueres, M. C. Surf.
Sci. 1977, 68, 359.
25) Gordon, M. B.; Cyrot-Lackmann, F.; Desjonqueres, M. C. Surf.
Sci. 1979, 80, 159.
(