Synthesis of single-walled carbon nanotubes in vibrationally non-equilibrium carbon monoxide

Article abstract of DOI:10.1016/S0009-2614(01)01473-7

A newly developed method for synthesizing single-walled carbon nanotubes (SWNT), which uses a thermally non-equilibrium flow reactor with extreme molecular vibrational mode disequilibrium of CO gas, the primary feedstock, was presented. The carbon producing CO disproportionation reaction was driven very efficiently in a flow reactor, where extreme disequilibrium between the vibrational and translational mode of the CO gas is maintained even at low translational temperatures by using a powerful and efficient CO gas laser. The vibrationally excited CO reacted in the presence of metal catalysts to form CO₂ and structured carbon molecules, notably SWNT. The individual tubes formed ropes or flat ribbons, which are aligned parallel to each other into larger structures of SWNT material without any post-synthesis treatment. This technique generated SWNT in a continuous (non-batch) process.

Full text of DOI:10.1016/S0009-2614(01)01473-7

6
February 2002
Chemical Physics Letters 352 (2002) 342–347
www.elsevier.com/locate/cplett
Synthesis of single-walled carbon nanotubes in
vibrationally non-equilibrium carbon monoxide
a,
a
b
a
*
Elke Pl o€ njes , Peter Palm , G.Babu Viswanathan , Vish V.Subramaniam ,
c
Igor V.Adamovich , Walter R.Lempert , Hamish L.Fraser , J.William Rich
d
b
a
a
Department of Mechanical Engineering, The Ohio State University, 206 West 18th Avenue, Columbus, OH 43210, USA
b
Department of Materials Science and Engineering, The Ohio State University, Columbus, OH 43210, USA
Department of Aerospace Engineering and Aviation, The Ohio State University, Columbus, OH 43210, USA
Department of Mechanical Engineering and Department of Chemistry, The Ohio State University, Columbus, OH 43210, USA
c
d
Received 17 October 2001; in final form 17 December 2001
Abstract
Single-walled carbon nanotubes (SWNTs) are synthesized in a gas-phase non-equilibrium plasma process.The
carbon producing CO disproportionation reaction is driven very efficiently in a flow reactor, in which extreme dis-
equilibrium between the vibrational and translational mode of the carbon monoxide gas is maintained even at low
translational temperatures by using a powerful and efficient carbon monoxide gas laser.In the presence of metal
2
catalysts, the vibrationally excited CO reacts to form CO and structured carbon molecules, notably SWNTs.The
individual tubes form ropes or flat ribbons and these are aligned parallel to each other into larger structures of SWNT
material without any post-synthesis treatment. Ó 2002 Published by Elsevier Science B.V.
1
. Introduction
worldwide.Most of them are variations of three
principal techniques: arc vaporization [2,3], laser
vaporization [4], and chemical vapor deposition
[5–8].While each of these techniques can provide
high-purity material, scaling to efficient, industrial
quantity production of the pure material required
for applications appears problematical.In this
Letter, we present a novel method for the synthesis
of SWNTs, which uses a thermally non-equilib-
rium flow reactor with extreme molecular vibra-
tional mode disequilibrium of the primary
feedstock, carbon monoxide (CO) gas.The vibra-
tionally excited CO reacts in the presence of metal
catalysts to form CO2 and structured carbon
molecules, notably SWNTs.This method is found
Since their discovery in the early 1990s [1,2],
single-walled carbon nanotubes (SWNTs) have
been a very active topic of study and many exciting
new applications in nanotechnology and material
science have been proposed.Controlled synthesis
of SWNTs can contribute to the fundamental
understanding of macromolecules and their orga-
nization.At present, several methods of producing
SWNTs are being actively pursued in laboratories
*
Corresponding author.Fax: +1-614-292-3163.
E-mail address: ploenjes.1@osu.edu (E. Pl o¨ njes).
0
009-2614/02/$ - see front matter Ó 2002 Published by Elsevier Science B.V.
PII: S0009-2614(01)01473-7

E. Pl €o njes et al. / Chemical Physics Letters 352 (2002) 342–347
343
to produce SWNTs in a continuous (non-batch)
process.
NiðCOÞ , which decompose in the reactor and
4
form small metal clusters, are used to nucleate the
SWNT formation.
Vibrationally non-equilibrium carbon monox-
ide is produced by initial resonance absorption of
radiation from a continuous wave CO molecular
gas laser into the lowest CO vibrational quantum
states (v ¼ 1 to v ꢀ 10).The vibrational mode
energy is subsequently redistributed into higher
vibrational levels by inelastic collisions among the
CO molecules, so-called anharmonic vibration-to-
vibration (V–V) up-pumping, or ‘Treanor pump-
ing’ [9]
2. Experimental
The experimental setup is shown in Fig.1.Ini-
tial excitation of the vibrational mode of carbon
monoxide at room temperature is achieved with a
CO laser operating on the lowest vibrational
transitions of CO, at a wavelength of about 5 lm.
The laser, designed and fabricated at Ohio State in
collaboration with the University of Bonn [25], has
a c.w. power output of 18 W with a multiline
spectrum containing vibrational transitions 1 ! 0
COðvÞ þ COðwÞ ! COðv ꢁ 1Þ þ COðw þ 1Þ
v < w
ð1Þ
In Eq.(1) CO ðvÞ and COðwÞ represent carbon
monoxide molecules in vibrational quantum states
v and w.This approach produces, in the steady
state, strong vibrational disequilibrium, resulting
in population of high-vibrational states (up to
v ¼ 40), at high-gas densities and low translational
temperatures.Optical pumping by a CO laser has
been previously achieved in gas-phase carbon
monoxide [10–16] at pressures of up to 20 atm [10],
in gas-phase nitric oxide [17,18], in liquid-phase
CO [19,20], and in solid CO and NO matrices
(ꢀ300 mW) to 10 ! 9.The laser is focused into
the reaction cell with a beam diameter of ꢀ1 mm.
The reaction cell is fabricated from Pyrex and
equipped with ports providing access for various
optical diagnostics and sampling assemblies.
Windows at the tube ends, which are purged with
argon to prevent build-up of carbonaceous de-
posits, provide the laser optical access into the
apparatus.Carbon monoxide gas is introduced
into the cell flowing at velocities of about .05 m/s.
The total gas pressure is 100 Torr with equal
amounts of CO and Ar.At these conditions, two
thirds of the laser power is absorbed into the
vibrational mode of CO.
[
21,22].Recently, optical pumping has also been
demonstrated in mixtures of a few percent CO
diluted by infrared inactive gases, such as nitrogen
and air, at atmospheric pressure [23,24].The re-
actor conditions, such as gas temperature, pres-
sure, and composition, can be varied extensively
while still maintaining a strong vibrational non-
equilibrium.
The vibrationally excited CO created by the
above process produces atomic carbon by the
Boudouard disproportionation reaction:
COðvÞ þ COðwÞ ! CO
þ E P 5:5 eV
2
þ C
ð2Þ
E
v
w
Since the activation energy for this reaction is
provided by the vibrational energy mode, the gas
kinetic temperature does not determine the carbon
production rate.Therefore, the temperature can be
kept at a value most suitable for the formation of
SWNTs from the carbon produced.Iron penta-
carbonyl, FeðCOÞ , and nickel tetracarbonyl,
Fig.1 .E xperimental setup.
5

344
E. Pl €o njes et al. / Chemical Physics Letters 352 (2002) 342–347
Gaseous metal carbonyls, FeðCOÞ5 and
NiðCOÞ , a few ppm of which are always present in
4
commercial CO in steel cylinders [26] and nickel-
containing gas regulators, are used as catalyst
precursors.Iron pentacarbonyl and nickel tetra-
carbonyl rapidly decompose at temperatures
above 250 °C; in addition, several of the CO laser
transitions present in the multiline laser spectrum
are absorbed by the metal carbonyls, and therefore
the decomposition of the carbonyls is enhanced by
photodissociation.Organometallic catalyst pre-
cursors have been successfully used previously for
the synthesis of SWNTs in gas-phase continuous-
flow techniques.Typically, a mixture of a carbon
source gas, such as hydrocarbons [5,6], and an
organometallic compound is reacted in a heated
furnace, where the organometallics decompose
and subsequently form metal clusters, on which
the carbon nanotubes grow.Most closely related
to this work, the Rice University Center for
Nanoscale Science and Technology has demon-
strated a high-pressure gas-phase synthesis of
SWNTs in a thermal equilibrium reactor using CO
Fig.2 .R elative population distribution of vibrationally excited
carbon monoxide: (ꢂ) denotes the measured non-equilibrium
vibrational distribution; ð———Þ denotes a calculated Boltz-
mann distribution at 4100 K.The translational temperature is
1
200 K.
feedstock and FeðCOÞ as the catalyst precursor
The extreme departure from a thermal equilibrium
(Boltzmann) distribution of vibrational level pop-
ulations is evident.Vibrational levels extending as
high as v ¼ 30 are substantially populated; the
energy of such levels is ꢀ6.5 eV. The gas-phase
activation energy of the CO disproportionation
reaction is around 5.5 eV. Providing comparable
population of these highly energetic states using
thermal equilibrium excitation would require
temperatures of about 10 000 K.Since vibrational
mode activation substantially contributes to es-
tablishing the reactive complex in the dispropor-
tionation, the vibrational excitation process
strongly enhances the disproportionation reaction.
The enhancement is large even for lower activation
energies, such as the reported ꢀ4 eV activation
energy required to drive the reaction on an FeO
catalytic surface [27], or on other catalytic sur-
faces.
5
[
7].
A Bruker IFS66 Fourier transform spectrome-
ter is employed to measure the spontaneous in-
frared emission from the laser excited CO, which is
used to determine the non-equilibrium vibrational
distribution function (VDF) present in the CO and
to measure the rotational mode temperature [16].
At the reactor operating pressures, rotational and
translational modes are equilibrated.The relative
concentrations of the metal carbonyls were moni-
tored in the exhaust line using FT-IR absorption
spectroscopy.However, a calibration of the FT-IR
absorption measurements for determination of
absolute carbonyl concentrations has not been
performed at this time.
3
. Results and discussion
In our synthesis reactor, the translational/rota-
tional molecular modes remain in equilibrium at a
relatively low temperature, 1200 K under the pre-
sent conditions.The translational temperature in
the excited CO can be controlled in steady state by
conductive and flow-convective cooling of the
Fig.2 shows the measured distribution of
populations among the CO vibrational quantum
levels, when the reactant is prepared in the manner
described above.The measured VDF is compared
with a thermal equilibrium distribution at 4100 K.

E. Pl €o njes et al. / Chemical Physics Letters 352 (2002) 342–347
345
reactor gases, as well as, in some cases, by adding a
non-reactive diluent species (such as argon) to in-
crease the heat capacity of the mixture.
For analysis by transmission electron micros-
copy (TEM), the carbonaceous products are col-
lected on substrate surfaces, such as TEM grids or
silicon wafers, at different locations in the reaction
cell during runs of, typically, 10–90 min duration.
High-resolution TEM has been performed using a
Philips/FEI CM300 FEG STEM.Figs.3–5 show
transmission electron microscope images of the
carbonaceous deposits produced using the non-
equilibrium synthesis technique.The images are
obtained from raw material that has been analyzed
without any post-synthesis purification.Under
these conditions, the products appear to be either
Fig.4. High-resolution TEM image of several ropes of
SWNTs.Scale bar 10 nm.Note that the individual nanotubes
end at a common edge.
Fig.5 .H igh-resolution TEM image of several ropes or ribbons
of SWNTs aligned parallel to each other.Scale bar 10 nm.
flat ribbons or cylindrical ropes approximately 7–
15 nm wide consisting of closely packed SWNTs.
The individual tubes within one rope appear to
have the same diameter with diameters varying
Fig.3 .H igh-resolution TEM image of a rope of SWNTs .S cale
bar 10 nm.Note that the individual nanotubes have a diameter
of 1.04 nm and align parallel into a cylindrical rope or flat
ribbon.

346
E. Pl €o njes et al. / Chemical Physics Letters 352 (2002) 342–347
between 0.7 and 1.2 nm in different bundles. Re-
markably, the individual tubes in these structures
appear to cut off at a common edge (Figs.3 and 4).
This could be caused by either a fracture of a
longer SWNT structure or it could be a feature of
the SWNT formation mechanism in the non-
equilibrium plasma.Typically, the ropes or rib-
bons appear to be in a length range of up to
relying on the presence of free carbon, such as
laser ablation, have yielded best results using Ni or
mixtures of Ni and other metals [30].Multi-walled
carbon nanotubes have yet to be detected in our
experiments.
The material deposited in a two-hour run by a
3
laser excited volume of about 1 cm was collected
and weighed.This measurement gives a total de-
position rate of 10–20 mg/h.It must be emphasized
that the results shown in Figs.3–5 were obtained
with a CO partial pressure of 50 Torr.Neverthe-
less, the non-equilibrium vibrational up-pumping
necessary to drive the nanotube synthesis process
also has been demonstrated to occur at CO pres-
sures in excess of one atmosphere [10].Both the
rate of the gas-phase CO disproportionation reac-
tion and the laser power required to sustain the
same level of vibrational non-equilibrium scale
quadratically with CO pressure.Therefore, the
yield of the disproportionation reaction scales lin-
early with the available laser power so that much
higher throughputs can be achieved using a higher
power laser to excite CO at high pressures.High-
power CO lasers have been demonstrated with c.w.
powers of several hundred kW and efficiencies of
30% [31–33].Consequently, the non-equilibrium
CO disproportionation method should be scalable
to large quantity production of SWNTs.
0
:5–1 lm.In some of the material samples, also
the bundles are oriented in a preferred direction
and line up parallel to each other (Fig.5).This
orientation of the ropes or ribbons as well as the
alignment of the individual SWNTs into bundles
with a common cut-off is possibly due to the in-
teraction of the strongly polarizable SWNTs with
the linearly polarized CO laser field during the
growth process.This may present a relatively
simple means for forming larger structures of
aligned SWNT materials.Other methods for pro-
ducing aligned arrays of SWNTs involve many
steps for the production of suitable catalyst tem-
plates [28,29].
For the deposited material shown in Figs.3–5,
the purity of the SWNT material was estimated
from TEM images to be on the order of 70%.
However, purity levels were found to vary, and
this is attributed to changes in the metal carbonyl
concentrations in the CO feedstock.The natural
abundance of FeðCOÞ in CO steel cylinders is
The overall energy efficiency of the laser-in-
duced non-equilibrium SWNT synthesis process
based on the CO disproportionation reaction
critically depends on the fraction of energy going
into carbon production by this reaction that sub-
sequently forms carbon nanotubes in the presence
of metal catalyst.This fraction may well depend
on a number of process parameters such as tem-
perature, pressure, catalyst partial pressure, etc.
Further studies of the SWNT synthesis process in
optically pumped CO plasmas are needed to pro-
vide insight into this critical issue.The non-equi-
librium SWNT synthesis technique reported here
enables mechanistic studies of the processes in-
volved in nanotube formation, since the synthesis
reactor is equipped with optical access ports into
the reaction volume and process parameters, most
notably the gas temperature, can be varied over a
wide range.This technique for the synthesis of
SWNTs represents a new type of laser induced
5
known to vary significantly with cylinder gas
pressure and temperature [26].The same is true for
NiðCOÞ , which forms to some extent in nickel
4
containing gas regulators.The results from Figs.
3
–5 were obtained with a mixture of NiðCOÞ and
4
FeðCOÞ .Similar results were obtained with only
5
nickel carbonyl precursor.From the currently
available data we cannot conclude whether a
mixture of NiðCOÞ and FeðCOÞ has any advan-
4
5
tages over a pure NiðCOÞ catalyst precursor.Iron
4
pentacarbonyl alone did not produce aligned
SWNTs, instead only very few individual single-
walled nanotubes of about 1 nm diameter have
been observed together with large amounts of
amorphous carbon.These results are in contrast to
the Rice University high-pressure gas-phase CO
(
HiPCO) process [7], which produces pure carbon
nanotube material using Fe as the catalyst.How-
ever, other synthesis methods, particularly those

E. Pl €o njes et al. / Chemical Physics Letters 352 (2002) 342–347
347
[
10] J.W. Rich, R.C. Bergman, M.J. Williams, final contract
report AFOSR F49620-77-C-0020 (November 1979).
11] J.W. Rich, R.C. Bergman, Chem. Phys. 44 (1979) 53.
12] R.L. DeLeon, J.W. Rich, Chem. Phys. 107 (1986) 283.
13] C. Flament, T. George, K.A. Meister, J.C. Tufts, J.W.
Rich, V.V. Subramaniam, J.-P. Martin, B. Piar, M.-Y.
Perrin, Chem.Phys.163 (1992) 241.
mode selective chemistry in collision-dominated
environments.
[
[
[
Acknowledgements
[
[
[
[
[
14] I.Adamovich, S.Saupe, MJ.Grassi, O.Schulz, S.
Macheret, J.W. Rich, Chem. Phys. 173 (1993) 491.
15] H.L. Wallaart, B. Piar, M.-Y. Perrin, J.-P. Martin, Chem.
Phys.196 (1995) 149.
16] E.Pl o€ njes, P. Palm, A.P. Chernukho, I.V. Adamovich,
J.W. Rich, Chem. Phys. 256 (2000) 315.
This research was partially supported by the
Director of Defense Research and Engineering
(
DDR&E) within the Air Plasma Ramparts
MURI Program managed by AFOSR, and by the
AFOSR Space Propulsion and Power Division.
Partial support from the Ohio Board of Regents
Investment Fund and the National Science
Foundation is gratefully acknowledged.We would
also like to express our gratitude to Prof.Richard
E.Smalley and Prof.Robert H.Hauge from the
Center for Nanoscale Science and Technology at
Rice University and to Dr.Wade Adams of the Air
Force Research Laboratories at Wright-Patterson
Air Force Base for useful discussions.
17] H.D u€ nnwald, E. Siegel, W. Urban, J.W. Rich, G.F.
Homicz, M.J. Williams, Chem. Phys. 94 (1985) 195.
18] S. Saupe, I. Adamovich, M.J. Grassi, J.W. Rich, Chem.
Phys.174 (1993) 219.
[
[
19] D.S. Anex, G.E. Ewing, J. Phys. Chem. 90 (1986) 1604.
20] R.S. Disselkamp, G.E. Ewing, J. Phys. Chem. 93 (1989)
6
21] J.P. Galaup, J.Y. Harbec, R. Charneau, H. Dubost, Chem.
334.
[
Phys.Lett.120 (1985) 188.
[22] I.Hadj Bachir, R.Charneau, H.Dubost, Chem.Phys.177
1993) 675.
(
[
23] E.Pl o€ njes, P. Palm, W. Lee, M.D. Chidley, I.V. Adamo-
vich, W.R. Lempert, J.W. Rich, Chem. Phys. 260 (2000)
3
53.
24] W. Lee, I.V. Adamovich, W.R. Lempert, J. Chem. Phys.
14 (2001) 1178.
References
[
1
[
[
1] S.Iijima, T.Ichihashi, Nature 363 (1993) 603.
2] D.S. Bethune, C.H. Kiang, M.S. De Vries, G. Gorman, R.
Savoy, J.Vasquez, R.Beyers, Nature 363 (1993) 605.
3] C. Journet, W.K. Maser, P. Bernier, A. Loiseau, M.L.
Delachapelle, S.Lefrant, P.Deniard, R.Lee, J E. .Fischer,
Nature 388 (1997) 756.
[25] S.B u€ scher, O.Schulz, A.Dax, H.Kath, W.Urban, Appl.
Phys.B 64 (1997) 307.
[26] R.K. Tepe, D. Vassallo, T. Jacksier, R.M. Barnes, Spec-
trochim.Acta Part B 54 (1999) 1861.
[
[
[27] H. Cheng, D.B. Reiser, S. Dean Jr., Catal. Today 50 (1999)
579.
4] A.Thess, R.Lee, P.Nikolaev, H .J .Dai, P.Petit, J.Robert,
C.H. Xu, Y.H. Lee, S.G. Kim, A.G. Rinzler, D.T. Colbert,
G.E. Scuseria, D. Tomanek, J.E. Fischer, R.E. Smalley,
Science 273 (1996) 483.
[28] R.R. Schlittler, J.W. Seo, J.K. Gimzewski, C. Durkan,
M.S.M. Saifullah, M.E. Welland, Science 292 (2001) 1136.
[29] N. Wang, Z.K. Tang, G.D. Li, J.S. Chen, Nature 408
(2000) 50.
[
[
[
5] H.M. Cheng, F. Li, G. Su, H.Y. Pan, L.L. He, X. Sun,
M.S. Dresselhaus, Appl. Phys. Lett. 72 (1998) 3282.
6] C.N.R. Rao, A. Govindaraj, R. Sen, B.C. Satishkumar,
Mater.Res.Innovat.2 (1998) 128.
[30] C.Laurent, E.Flahaut, A.Peigney, A.Rousset, New J.
Chem.11 (1998) 1229.
[31] A.P. Averin, N.G. Basov, E.P. Glotov, V.A. Danilychev,
L.N. Drachuk, O.M. Kerimov, I.N. Matveev, A.M.
Sorkin, N.D. Ustinov, N.V. Cherburkin, V.I. Yugo, Sov.
J.Quantum Electron.13 (1983) 1391.
7] P. Nikolaev, M.J. Bronikowski, R.K. Bradley, F. Roh-
mund, D.T. Colbert, K.A. Smith, R.E. Smalley, Chem.
Phys.Lett.313 (1999) 91.
[32] S .- I.Sato, T.Abe, T.Fujioka, Paper FM4, CLEO June 19–
22, 1984.
[
[
8] H.Dai, Physics World, vol.43, June 2000.
9] C.E. Treanor, J.W. Rich, R.G. Rehm, J. Chem. Phys. 48
[33] I.B. Kovsh, Lebedev Physical Institute, Moskow, 1990
(private communication to J.W. Rich).
(1968) 1798.