Growth of single-walled carbon nanotubes by a novel technique using nanosized graphite as carbon source

Article abstract of DOI:10.1016/S0009-2614(00)00830-7

Single-walled carbon nanotubes (SWCN) have been produced treating nanosized graphitic powders with atomic hydrogen fluxes in a purpose-designed reactor. The Fe-catalyzed synthesis process yielded rather dense networks of nanotubes forming continuous layers adherent to the substrates. Raman spectroscopy in the 140-250 cm^-1 region, high-resolution transmission electron microscopy and reflection electron diffraction in selected area configuration have been used to define the structural details of the nanotubes. The analysis evidenced a nanotubes population belonging prevalently to the (n,0) achiral symmetry subclass. This new experimental approach for SWCNs growth provides promising conditions for maximizing yields and generating selected tubular configurations.

Full text of DOI:10.1016/S0009-2614(00)00830-7

15 September 2000
Ž
.
Chemical Physics Letters 327 2000 284–290
www.elsevier.nlrlocatercplett
Growth of single-walled carbon nanotubes by a novel technique
using nanosized graphite as carbon source
M.L. Terranovaa,), S. Piccirillo a, V. Sessa a, P. Sbornicchiaa, M. Rossi b,1,
S. Botti c,2, D. Manno d,3
Dipartimento di Scienze e Tecnologie Chimiche and Unita’ INSTM, UniÕersita di Roma ‘Tor Vergata’,
a
`
Via della Ricerca Scientifica I-00133 Rome, Italy
b
Dipartimento di Energetica and INFM, UniÕersita ‘La Sapienza’, Via A. Scarpa 16, 00162 Rome, Italy
`
c
ENEA, Dipartimento InnoÕazione, DiÕisione Fisica Applicata, CR Frascati PO Box 65, 00044 Frascati, Italy
d
Dipartimento di Scienza dei Materiali and INFM, UniÕerita’ di Lecce, Via Arnesano, Lecce, Italy
Received 3 January 2000; in final form 7 July 2000
Abstract
Ž
.
Single-walled carbon nanotubes SWCN have been produced treating nanosized graphitic powders with atomic hydrogen
fluxes in a purpose-designed reactor. The Fe-catalyzed synthesis process yielded rather dense networks of nanotubes forming
1
continuous layers adherent to the substrates. Raman spectroscopy in the 140–250 cmy region, high-resolution transmission
electron microscopy and reflection electron diffraction in selected area configuration have been used to define the structural
Ž
.
details of the nanotubes. The analysis evidenced a nanotubes population belonging prevalently to the n,0 achiral symmetry
subclass. This new experimental approach for SWCNs growth provides promising conditions for maximizing yields and
generating selected tubular configurations. q 2000 Published by Elsevier Science B.V.
1. Introduction
the integers n and m identify the general geometry
of SWCNs. The tubes with nsm are named ‘armc-
hair’; tubes with either ns0 or ms0 are named
A graphene sheet wrapped around the surface of a
seamless cylinder constitutes a single-walled carbon
w x
‘zig-zag’; all other tubes have chiral symmetry 2 .
Ž
.
nanotube SWCN . The helicity of the nanotubes is
Since the first experimental evidence of SWCNs
w x
w
x
usually described by the Hamada vector 1 , which
indicates how the graphene sheet is rolled up along a
lattice vector with components n,m . The values of
3,4 , long-range research efforts have been dedi-
cated to developing synthesis procedures and to opti-
mizing the conditions for enhanced production of
single-walled nanotubes with a controlled geometri-
cal structure. Both scientific interest and perspectives
of technological exploitation motivates the increas-
ing attention applied to this nanomaterial area. Such
a completely artificial form of crystalline carbon
indeed opens a new ‘scenario’ to fundamental sci-
ence and holds great promise for novel applications.
Ž
.
)
Corresponding author. Fax: q39-6-7259-4328; e-mail:
terranova@roma2.infn.it
1 Fax: q39-6-4424-0183; e-mail:
marcorossi@axrma.uniromal.it
2 Fax: q39-6-9400-5400; e-mail: botti@frascati.enea.it
3 Fax: q39-8-3232-0525; manno@axomat.unile.it
0009-2614r00r$ - see front matter q 2000 Published by Elsevier Science B.V.
Ž
.
PII: S0009-2614 00 00830-7

(
)
M.L. TerranoÕa et al.rChemical Physics Letters 327 2000 284–290
285
w
x
w
x
Arc discharge 3–5 and laser ablation 6–9 were
the methods initially proposed and still are the most
widely used techniques for production of single-
walled tubules. Other methods were later proposed,
such as hydrocarbon decomposition or pyrolysis
reservoir by Ar streams and driven into the reaction
chamber at controlled delivery rates. The nozzle is
designed to provide homogeneity of the powder dis-
tribution across the active area of the substrate facing
Ž
.
the heated Ta filament Ts2200"108C . During
synthesis of the SWCNs described in this Letter, the
substrate temperature was monitored by a thermo-
couple-based system and kept at 900"58C. The
w
x
w
x
10–14 and synthesis by CVD techniques 15–17 .
The present work originated from an attempt to
experiment with novel manufacturing technology.
Our approach utilizes nanosized graphite powders as
carbon feedstock and H2 as reactant gas, making use
of a powder-flowing apparatus connected with a
reactor where atomic hydrogen is produced. The
nanotubes generated by the reaction between atomic
H and nanographite are found to condense into high-
density bundles, forming films on various substrates
Ž
.
pressure in the chamber 40 Torr was produced by
the inert gas carrier and by the hydrogen streams
flowing inside the reaction chamber at a rate of 30
Ž
.
and 200 sccm standard cubic centimeter minute ,
respectively. Due to the geometry of the apparatus,
the powder-containing carrier flux reaches the depo-
sition area intersecting at about 308 the flux of
atomic hydrogen.
Ž
.
pre-coated by transition metal layers Ni, Co, Fe .
The nanosized graphitic powders, used as the
carbon source, were produced by cw-CO2 laser-in-
duced pyrolysis of ethyleneracetylene mixtures. The
mean size of the powders used for the nanotube
synthesis was 20 nm.
2. Experimental
The apparatus used in the present research con-
sists of a reaction chamber equipped with a powder-
flowing system schematically formed by a powder
reservoir, a mixing chamber, a mass flow controller
The layers of SWCNs were deposited by 30-min
Ž
.
lasting runs on iron-coated Si 100 substrates. The
catalyst layers were prepared by dropping a solution
Ž
.
Ž
.
Ž
.
and a final Mo nozzle Fig. 1 . A Joule-heated
filament provides decomposition of the H2 fluxes
into atomic hydrogen. The nanosized graphitic pow-
ders used as carbon source are swept from the
of Fe NO3 in acetone onto the surface of Si 100
3
polished plates. After drying in a furnace, the sub-
strates were positioned on the heated holder of the
reactor and treated for 15 min at 9008C under atomic
Fig. 1. Schematic view of the experimental set up used for SWCNs synthesis.

(
)
286
M.L. TerranoÕa et al.rChemical Physics Letters 327 2000 284–290
hydrogen atmosphere. This process allowed reduc-
tion of the Feq3 ions and subsequent clustering of
metallic iron in nanosized aggregates, homoge-
neously distributed across the substrate surface.
Structural and morphological characterization of
the produced SWCNs have been carried out using
micro-Raman spectroscopy, reflection high-energy
3. Results and discussion
Detailed information about helicity of the nan-
otubes produced by this novel technique have been
obtained by micro-Raman analysis of the frequency
region extending from 140 up to 250 cmy1, where
the signals are produced by the radial breathing
Ž
.
Ž
.
electron diffraction RHEED and high-resolution
mode RBM of carbon atoms in phase along the
direction normal to the tube axis. The analyses have
been performed on as-grown SWCN films supported
on their substrate.
Ž
.
transmission electron microscopy HRTEM .
The Raman analysis was performed in the 140–
1540 cmy1 spectral region, at room temperature in
the backscattering configuration using an Ary ion
Fig. 2 shows the Raman RBM response for a
typical SWCNs deposit. The fine structure of the
spectrum has been evidenced by a lineshape analysis
using Lorentzian curves. As one can see, the more
intense signals are observed at 178, 168 and 155
cmy1, whereas other weak Raman features are de-
tected at 192, 199, 215, 221 and 226 cmy1. By
means of the vrrd relationships presented by sev-
eral research groups 11,18,19 , from the analysis of
the RBM frequencies vr it is possible to obtain
information on the nanotube radial dimensions d .
Following the vr cm
lation proposed by Bandow et al. 18 in the case of
isolated nanotubes, the diameters of our SWCNs
would result in the 1.0–1.5 nm range. Diameters
values ranging between 1.1 and 1.6 are evaluated
from the relationship vr cm
Ž
laser beam wavelength, 514.5 nm; laser power,
10 mW; spectral resolution, 3 cmy1 . The laser spot
.
size on the sample was about 1 mm2.
For the RHEED investigation we used an AEI
EM6G apparatus equipped with a high-resolution
diffraction stage. The measurements were carried out
at 60 keV in selected area SA conditions using a
goniometer stage that permitted to vary the incidence
angle of the ey beam with respect to the sample and
moreover to rotate the sample in a plane parallel to
the beam.
HRTEM investigations have been carried out by a
transmission electron microscope JEOL 2010 operat-
ing at an acceleration voltage of 160 kV, suitable to
obtain a good resolution and to produce as little
radiation damage as possible.
Ž
.
w
x
Ž
.
Ž .
y1
Ž
.
s223.75rd theoretical re-
w
x
y1
Ž
.
w x
s254rd 11 , and
Fig. 2. Raman spectrum showing the radial breathing mode of the deposited SWCNs.

(
)
M.L. TerranoÕa et al.rChemical Physics Letters 327 2000 284–290
287
between 1.0 and 1.6 nm from the phenomenological
the resolution limits, with those associated to arm-
chair achiral subclass or to a chiral symmetry class.
Another confirmation of the tube geometry comes
from the lack of signals in the 650–770 and 1400–
1540 cmy1 regions where, on the basis of calcula-
tions performed in the framework of the zone folding
y1
Ž
.
expression vr cm
s6.5q232rd recently pro-
w
x
posed by Alvarez 19 in order to take into account
interaction effects due to aggregation of individual
tubes. The analysis of both location and relative
intensity of the maxima in the region of the radial
A1g phonon modes is consistent with the HRTEM
observations, which indicate a prevalence of nan-
otubes with diameter values strongly peaked around
1.3–1.5 nm.
w
x
Ž
.
model 22 , the n,0 subclass is not expected to
w
x
exhibit Raman-active modes 23 .
Therefore, even if the contributions of other tube
geometries cannot be ruled out, the Raman scattering
data coherently indicates that SWCNs having a zig-
zag configuration are the dominant component of our
deposits. Moreover, no signals coming from the Si
substrate have been detected in the Raman spectra,
which would indicate that a dense network of nan-
otubes has been generated and deposited in the form
of a rather thick layer during the 30 min run.
The samples for HRTEM observations have been
obtained by stripping parts of the film from the
substrate and depositing it directly on the carbon
covered microgrids. This mechanical process al-
lowed preserving the array’s integrity and did not
alter the original fine structure and morphology of
the deposited material.
The observed peaks have been assigned to the
vibrational modes of specific symmetry types on the
basis of the frequencies obtained for chiral and achi-
w
x
ral tubes using a density functional method 20,21 .
The position of the whole set of observed frequen-
Ž
.
cies is consistent with the presence of n,0 achiral
zig-zag SWCNs, with 13(n(19. In particular, an
excellent agreement is found between the observed
modes at 178 and 168 cmy1 and the frequencies of
177 and 167 cmy1 evaluated, respectively, for the
Ž
.
Ž
.
w
x
17,0 and 18,0 tubules 17,18 . A good correspon-
Ž
dence is also found between the calculated 158
cmy1 and measured 155 cm
signals correspond-
y1
.
Ž
.
Ž
.
ing to 19,0 SWCNs. Just in the case of the weak
Ž
.
Ž
.
signals ascribed to the 16,0 and 13,0 tubes the
lines predicted at 187 and 231 are found up- and
down-shifted of 5 cmy1, respectively. It is worth
noticing that although some coincidence occurs, be-
The micrograph reported in Fig. 3a reveals the
presence of tubular structures with diameters preva-
lently in the 1.3–1.5 nm range and lacking of any
particular alignment.
Ž
.
Ž
.
w x
tween tubes with 17,0 and 10,10 indices 20 , the
other detected signals have no correspondence, within
The observed relatively poor contrast of the
HRTEM image is determined by the particular mor-
Ž .
Fig. 3. a HRTEM image obtained by a very thin film stripped from the substrate and containing tubular structures without any particular
Ž .
Ž
alignment. b Schematic view of the position of some tubular structures unambiguously determined with the aid of locally applied digital
filtering.

(
)
288
M.L. TerranoÕa et al.rChemical Physics Letters 327 2000 284–290
phology of these samples, which appear constituted
by tangles of tubular structures randomly interlaced.
To increase the contrast and decrease the background
noise, different digital filterings of the image have
been locally applied. Fig. 3b schematically shows the
position of some tubular structures that have been
consequentially and unambiguously identified.
As far as it concerns the diffraction measure-
ments, two methods to model the diffraction pattern
of carbon nanotubes have been reported in the litera-
w
x
ture. Zhang et al. 24 used the geometrical recipro-
cal space model to localize the reflection points,
taking into account tilting of the nanotube with
y
w
x
w
x
respect to the e -beam. Quin 25 and Lucas 26
applied the kinematical approximation to calculate
the diffraction intensity, obtaining additional infor-
Ž
mation about the fine structure broadening and in-
.
tensity distribution of the reflections. On the basis
Fig. 4. RHEED pattern taken from the surface of as-deposited
layers containing SWCNs.
of the proposed models, the predicted diffraction
patterns of a single perfect carbon nanotube are
characterized by the presence of reflections streaked
along directions perpendicular to the tube axis and
away from it. In our case the RHEED patterns, taken
in reflection conditions from a very rough surface,
appear as a transmission pattern; moreover the
diffraction signals are not generated by an individual
carbon nanotube but by a very large amount of
non-parallel tubes, as evidenced by HRTEM images,
and one must consider the reciprocal space corre-
sponding to a film constituted by a dense tangle of
enrolled single crystalline sheets. The diffraction pat-
terns result therefore constituted by the superposition
of the diffraction signals corresponding to single
carbon nanotubes. In this situation, the incidence
angle between ey-beam and tube axis of each SWCN
is totally random and the superposition leads to the
broadened rings of Fig. 4, characterized by an inten-
sity distribution coherent with that calculated for the
streaked reflections produced by an individual de-
fect-free nanotube.
techniques. The explanation for this situation very
likely lies on the difference among reaction mecha-
nisms acting during the first stages of the various
processes. Instead of carbon sublimation or vaporiza-
tion, as in the case of laser-ablation or arc-discharge
depositions, under our experimental conditions the
initial step of the SWCNs growth involves a process
of carbon etching by atomic H. Some of our experi-
w
x
ments 27–29 demonstrated that the interaction of
atomic H with solid carbon phases is capable to
produce a variety of carbon-based structures. In the
light of the previous findings, the etching by hydro-
gen at the onset of the process of SWCN synthesis is
expected to cause a decrease in the size of the
graphite nanoparticles by disrupting larger structures,
and to produce excited sp2-carbon atoms and
graphene clusters which are swept towards the sub-
strate by the gas stream. A catalyzed structural rear-
rangement of graphene sheets into a cylindrical ge-
ometry is thought to take place at this stage.
Overall, the Raman, HRTEM and RHEED mea-
surements consistently point out that the proposed
synthesis methodology favors the formation of thick
layers constituted by densely packed networks of
SWCNs. Using the present approach the nanotube
bundles can be produced at temperature values in the
range 850–9508C, which are remarkably lower than
those associated to the laser ablation or arc-discharge
The final step of the process is likely to be
dominated by a mechanism similar to those, which
are effective in nanotube growing, by other synthesis
methodologies based on condensation of carbon-
dominated vapours. The necessary presence of some
Ž
.
transition metals Fe, Co, Ni represents indeed a
common key for the production of single-sheet nan-

(
)
M.L. TerranoÕa et al.rChemical Physics Letters 327 2000 284–290
289
w
x
otubes 30 and to explain this fundamental issue
some hypothesis have been raised. It has been argued
that chemisorption of carbon on the surface of transi-
tion metals would help in modifying the orientation
of carbon p-orbitals and in producing curvature ef-
conditions generated SWCNs belonging prevalently
Ž
.
to the n,0 achiral symmetry subclass. However,
some preliminary results of experiments still in
progress suggest that a different choice of process
parameters can favor the selective formation of nan-
otubes with a different geometry.
w
x
fects in flat graphene structures 31 . It has been also
observed that stabilization of C dangling bonds by
metal atoms inhibits tube closure and formation of
fullerene-like molecules 6,32,33 . It should be em-
phasized, however, that the theory of SWCNs growth
is still under discussion, and no exhaustive interpre-
tation of the role played by the metals has still been
given.
In this context, the feasibility of the presented
methodology to synthesize in an effective way sin-
gle-walled carbon might be used to explore its poten-
tiality for the large scale production of SWCN with
selected tubular configuration.
w
x
The present results indicates that the presence of
Fe, Ni or Co metals in form of submicrometer-sized
particles is necessary also for SWCNs growth by our
technique employing nanographite and atomic hy-
drogen as C feedstock and reactants, respectively.
Bonding of carbon species, produced by H-etching,
to the metal coordination sites is thought to reduce
localized C supersaturation and to avoid formation of
thick carbon clusters and subsequent coalescence
into graphite layers, as experienced when the pro-
cesses were carried out in the absence of a supported
metallic catalyst.
Acknowledgements
Ž
The authors are indebted to Dr. Ruani CNR,
.
Ž
.
Bologna and A. Tagliapietra Renishaw, Italy for
the Raman spectra.
References
w x
1
N. Hamada, S. Sawada, A. Oshiyama, Phys. Rev. Lett. 68
Ž
.
1992 54.
w x
2
M.S. Dresselhaus, G. Dresselhaus, P.C. Eklund, Science of
Fullerenes and Carbon Nanotubes, Academic Press, New
York, 1996.
w x
Ž
.
3
w x
4
S. lijima, T. Ichihashi, Nature 363 1993 603.
4. Conclusions
S. Bethune, C. Kiang, M.S. de Vries, G. Gorman, R. Savoy,
Ž
.
J. Vasquez, R. Beyers, Nature 363 1993 605.
w x
5
C. Journet, W.K. Maser, P. Bernier, A. Loiseau, M. Lamy de
La Chapelle, S. Lefrant, P. Deniard, R. Lee, J.E. Fisher,
The results indicate that SWCNs can be success-
fully generated using a powder flowing apparatus
coupled to a reactor, where conditions are estab-
lished for the occurrence of effective reactions be-
tween nanosized graphite and atomic hydrogen in the
presence of substrate-supported Fe dispersions. With
respect to other synthesis techniques, this new proce-
dure presents some advantages, such as to reach
relatively high production yields at temperatures re-
markably lower than those related to the arc-dis-
charge and laser ablation methods.
It is worth noting, moreover, that the nanotubes
produced by this approach are found to be arranged
in dense bundles forming continuous films adherent
to the subsrates, and can therefore be easily collected
without material loss.
Ž
.
Nature 388 1997 756.
w x
6
A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C.
Xu, Y. Hee Lee, S. Gon Kim, A.G. Rinzler, D.T. Colbert,
G.E. Scuseria, D. Tomanek, J.E. Fisher, R.E. Smalley, Sci-
´
Ž
.
ence 273 1996 483.
w x
7
M. Yudasaka, T. Komatsu, T. Ichihashi, S. lijima, Chem.
Ž
.
Phys. Lett. 278 1997 102.
w x
8
Ž
.
Y. Zhang, H. Gu, S. lijima, Appl. Phys. Lett. 73 1998
3827.
w x
9
A.C. Dillon, P.A. Parilla, J.L. Alleman, J.D. Perkins, M.J.
Ž
.
Heben, Chem. Phys. Lett. 316 2000 13.
w
x
Ž
.
10 R. Rao, A. Govindaraj, C.N.R. Rao, Chem. Mater. 9 1997
2708.
w
x
11 H.M. Cheng, F. Li, X. Sun, S.D.M. Brown, M.A. Pimenta,
A. Marucci, G. Dresselhaus, M.S. Dresselhaus, Chem. Phys.
Ž
.
Lett. 289 1998 602.
w
x
12 M. Terrones, N. Grobert, J. Olivares, J.P. Zhang, H. Ter-
rones, K. Kordatos, W.K. Hsu, J.P. Hare, P.D. Townsend, K.
Prassides, A.K. Cheetham, H.W. Kroto, D.R. Walton, Nature
The structural and morphological investigations
consistently point out that the described experimental
Ž
.
388 1997 52.

(
)
290
M.L. TerranoÕa et al.rChemical Physics Letters 327 2000 284–290
w
w
w
w
w
w
w
w
w
w
w
x
13 Z.R. Ren, Z.P. Huang, J.W. Xu, J.H. Wang, P. Bush, M.P.
Thess, E.E. Smalley, G. Dresselhaus, M.S. Dresselhaus, Sci-
Ž .
ence 275 1997 187.
Ž
.
Siegal, P.N. Provencio, Science 282 1998 1105.
x
w
x
14 R. Boese, A.J. Matzger, K.P.C. Vollhardt, Am. Chem. Soc.
24 X.B. Zhang, X.F. Zhang, S. Amelinckx, G. Van Tendeloo, J.
Ž
.
Ž .
Van Landuyt, Ultramicroscopy 54 1994 237.
1119 1997 2052.
x
w
w
x Ž .
15 H.J. Dai, A.G. Rinzler, P. Nikolaev, A. Thess, D.T. Colbert,
25 L.C. Quir, J. Mater. Res. 9 1994 2450.
Ž
.
x
26 A.A. Lucas, V. Bruyninckx, Ph. Lambin, Europhys. Lett. 35
R.E. Smalley, Chem. Phys. Lett. 260 1996 471.
x
Ž
.
Ž
.
16 J. Kong, A.M. Cassel, H. Dai, Chem. Phys. Lett. 292 1998
1996 355.
w
w
w
w
w
w
w
x
567.
27 M.L. Terranova, V. Sessa, V. Rigato, M. Braglia, G. Cocito,
x
Ž .
17 S.L. Sung, S.H. Tsai, C.H. Tseng, F.K. Chiang, X.W. Liu,
Thin Solid Films 232 1993 21.
Ž
.
x
28 M.L. Terranova, M. Rossi, V. Sessa, G. Vitali, Phys. Stat.
H.C. Shih, Appl. Phys. Lett. 74 1999 1297.
x
Ž .
Sol. 154 1996 127.
18 S. Bandow, S. Asaka, Y. Saito, A.M. Rao, L. Grigorian, E.
Ž
.
x
Richter, P.C. Eklund, Phys. Rev. Lett. 80 1998 3779.
29 M.L. Terranova, M. Rossi, V. Sessa, A. Alippi, Adv. Mater.
x
Ž .
3 1997 301.
19 Alvarez, A. Righi, T. Guillard, S. Rols, E. Anglaret, D.
Ž
.
x
Laplaze, J.-L. Dauvajol, Chem. Phys. Lett 316 2000 186.
30 J. Bernholc, C. Brabec, M. Buongiorno Nardelli, A. Maiti, C.
x
Ž
.
Ž .
Roland, B.I. Yakobson, Appl. Phys. A 67 1998 39.
31 E. Richter, K.R. Subbaswamy, Phys. Rev. Lett. 79 1997
20 J. Kurti, G. Kresse, H. Kuzmany, Phys. Rev. B 58 1998
¨
x
Ž
.
R8869.
x
21 H. Kuzmany, B. Burger, M. Hulman, J. Kurti, A.G. Rinzler,
2738.
¨
Ž
.
x
32 A. Maiti, C.J. Brabec, C. Roland, J. Bernholc, Phys. Rev. B
E.E. Smalley, Europhys. Lett. 44 1998 518.
x
Ž .
52 1995 14850.
22 R.A. Jishi, L. Venkataraman, M.S. Dresselhaus, G. Dressel-
Ž
.
x
haus, Chem. Phys. Lett. 209 1993 77.
33 T. Guo, P. Nikolaev, A. Thess, D.T. Colbert, R.E. Smalley,
x
Ž .
Chem. Phys. Lett. 243 1995 49.
23 A.M. Rao, E. Richter, S. Bandow, B. Chase, P.C. Eklund,
K.A. Williams, S. Fang, K.R. Subbaswamy, M. Menon, A.