Morphology, diameter distribution and Raman scattering measurements of double-walled carbon nanotubes synthesized by catalytic decomposition of methane

Article abstract of DOI:10.1016/S0009-2614(02)00686-3

Double-walled carbon nanotubes (DWNTs) were synthesized by catalytic decomposition of methane in the presence of Fe catalyst at 1373 K. The microstructure of the as-prepared DWNTs was characterized by high-resolution transmission electron microscopy (HRTEM) and resonant laser Raman spectroscopy. HRTEM observations revealed that the dominant type of the as-prepared product was DWNTs, which are mostly bundle-like. A triangular lattice arrangement of DWNTs in a DWNT bundle was observed. The as-prepared DWNTs show corresponding peaks from resonant Raman spectra of the radial breathing mode (RBM), which are considered to be associated with inner tubes as well as outer tubes of the DWNTs. The outer and inner tube diameters of the DWNTs, as determined from HRTEM images, are in the range of 1.6-3.6 and 0.8-2.8 nm, in agreement with the results from the resonant Raman scattering measurements. Moreover, the interlayer spacing of DWNTs is not a constant, ranging from 0.34 to 0.41 nm.

Full text of DOI:10.1016/S0009-2614(02)00686-3

20 June 2002
Chemical Physics Letters 359 (2002) 196–202
www.elsevier.com/locate/cplett
Morphology, diameter distribution and Raman scattering
measurements of double-walled carbon nanotubes
synthesized by catalytic decomposition of methane
a
Wencai Ren , Feng Li , Jian Chen , Shuo Bai , Hui-Ming Cheng
a
b
a
a,
*
a
Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road,
Shenyang 110016, China
b
Sichuan Institute of Light Industry and Chemical Technology, Zigong 643033, China
Received 18 March 2002
Abstract
Double-walled carbon nanotubes (DWNTs) were synthesized by catalytic decomposition of methane in the presence
of Fe catalyst at 1373 K. The microstructure of the as-prepared DWNTs was characterized by high-resolution trans-
mission electron microscopy (HRTEM) and resonant laser Raman spectroscopy. HRTEM observations revealed that
the dominant type of the as-prepared product was DWNTs, which are mostly bundle-like. A triangular lattice ar-
rangement of DWNTs in a DWNT bundle was observed. The as-prepared DWNTs show corresponding peaks from
resonant Raman spectra of the radial breathing mode (RBM), which are considered to be associated with inner tubes as
well as outer tubes of the DWNTs. The outer and inner tube diameters of the DWNTs, as determined from HRTEM
images, are in the range of 1.6–3.6 and 0.8–2.8 nm, in agreement with the results from the resonant Raman scattering
measurements. Moreover, the interlayer spacing of DWNTs is not a constant, ranging from 0.34 to 0.41 nm. Ó 2002
Elsevier Science B.V. All rights reserved.
1. Introduction
be classified into two categories: MWNTs and
single-walled carbon nanotubes (SWNTs). Large
MWNTs with more than five layers and SWNTs
have been successfully synthesized by those tech-
niques.
Double-walled carbon nanotubes (DWNTs), as
a critical type between SWNTs art and MWNTs,
are very important for theoretical and experi-
mental researches. Semiconductor-metallic or me-
tallic-semiconductor tubules can be expected as a
molecular conductive wire covered by an insulator
or as a molecular capacitor in a memory device,
respectively [5]. However, the investigation of
Much attention has been paid to both funda-
mental and applied aspects of carbon nanotubes
(CNTs) since the discovery of multi-walled carbon
nanotubes (MWNTs) by Iijima et al. [1]. Various
synthesis methods such as electric arc discharge [2],
laser vaporization [3] and pyrolysis [4] were de-
veloped to produce CNTs. In general, CNTs can
* Corresponding author. Fax: +1-86-24-2389-1320.
E-mail address: cheng@imr.ac.cn (H.-M. Cheng).
0009-2614/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0009-2614(02)00686-3

W. Ren et al. / Chemical Physics Letters 359 (2002) 196–202
197
DWNTs was hindered due to the difficulty of the
controllable synthesis. Recently, Hutchison et al.
[6] have synthesized DWNTs by a hydrogen arc
discharge method. Bandow et al. [7] synthesized
DWNTs through the coalescence of C60 in SWNTs
at high temperature. However, it is difficult to re-
alize controllable synthesis of DWNTs by hydro-
gen arc discharge method because it is not an
equilibrium process, while it is not efficient and
cost-effective to prepare DWNTs through the co-
alescence of C60 in SWNTs. SWNTs and MWNTs
have been successfully synthesized by catalytic
decomposition of hydrocarbons by controlling
reaction parameters [4,8], which shows that this
pyrolysis technique is a suitable method for con-
trollable growth of carbon nanotubes.
In this Letter, DWNTs with high quality were
synthesized by catalytic decomposition of meth-
ane. The as-prepared DWNTs have smaller di-
ameter than those synthesized by hydrogen arc
discharge method. The resonant laser Raman
measurements of the DWNTs were carried out,
and the pair peaks of their radial breathing mode
(RBM) were observed, which is considered to be
associated with inner tubes as well as outer tubes
of the DWNTs. The diameter distribution of the
DWNTs, based on the results from high-resolution
transmission electron microscopic (HRTEM) im-
ages, were obtained, which is in agreement with
the Raman scattering measurements.
rocene vapor carried into the reactor by hydrogen
was controlled by the temperature of ferrocene
and the flow rate of hydrogen. During the prepa-
ration, methane, ferrocene and thiophene vapors
were mixed and carried into the reactor, main-
tained at 1373 K. The vaporized ferrocene was at
first reduced by hydrogen to form atomic iron.
Then, atomic irons were agglomerated into iron
clusters or nanoparticles and were carried by
hydrogen into the reaction zone. The thiophene
vapor was decomposed into sulfur at high tem-
perature.
The experimental conditions were very close to
those for synthesis of SWNTs, except for the
concentration of ferrocene carried into the reac-
tion zone, which is smaller than that of SWNTs.
The morphology and microstructure of the
product were observed using high-resolution TEM
(HRTEM JOEL 2010 200 kV) and Raman spec-
troscopy (Jobin Yvon LabRam HR800, excited by
632.8 nm laser).
3. Results and discussion
3.1. Morphology of the as-prepared DWNTs
Under
optimal
preparation
conditions,
DWNTs were obtained, which adhered to the
thermocouple protecting tube. The bulk mor-
phology of the product is web-like, which looks
semi-transparent and appears to be very light.
We have carried out further HRTEM observa-
tions on the as-prepared samples, which revealed
that more than 70% of the product was DWNTs.
The rest of the product was mostly catalysts cap-
sulated by graphite layers and amorphous carbon.
A few SWNTs and three-walled CNTs were oc-
casionlly found. Figs. 1a–d show typical images of
the isolated DWNT. Most of the DWNTs form
bundles due to the van der Waals attraction be-
tween the tubes, and these bundles always consist
of many well-aligned DWNTs. Moreover, the
number of DWNTs in a bundle is smaller than
the number in a SWNT bundle, possibly due to the
weaker van der Waals force resulting from the
effect of the inner shell of the DWNTs. Fig. 1e
shows a DWNT bundle with a diameter of 15.1 nm
2. Experimental
The apparatus for synthesis of DWNTs by
catalytic decomposition of methane was given in
detail in another letter [4]. In general, the appa-
ratus consists of a furnace fitted with a corundum
tube and a stainless steel jar. The corundum tube
was 60 mm in diameter and 1000 mm in length. In
our experiment, methane was used as the carbon
source, hydrogen as the carrier gas and ferrocene
as the catalyst precursor. Thiophene was carried
into the reactor by methane to enhance the growth
of CNTs, the amount of which was controlled by
the temperature of thiophene and the flow rate of
methane. Ferrocene was taken in the jar and va-
porized at high temperature. The amount of fer-

198
W. Ren et al. / Chemical Physics Letters 359 (2002) 196–202
Fig. 1. HRTEM images of the as-prepared DWNTs by the catalytic decomposition of methane. (a)–(c) Three isolated DWNTs with
equal diameter. (d) An isolated DWNT with large diameter and part of third layers. (e) A bundle of aligned DWNTs and an
isolated DWNT with small diameter and large interlayer distance. (f) A tip of a DWNT and an isolated SWNT. (g) and (h) Two
isolated DWNTs with different inner diameter and the same outer diameter. (i) A DWNT with a small variation in diameter, (j) A
curved DWNT with a variation in shell number of shells and some impurities on it. (k) Two DWNTs with some impurities and an
isolated SWNT with the same diameter. (l) An isolated DWNT and an isolated SWNT with the same outer diameter. (m) Bundles
of DWNTs and an isolated three-walled carbon nanotube.
in which seven parallel DWNTs with an even di-
ameter (outer diameter ꢀ2.0 nm) aligned on-by-
one in the space, and an isolated DWNT with an
outer tube diameter of 1.73 nm, inner tube diam-
eter of 0.89 nm and inter layer spacing of 0.40 nm,
as denoted by an arrow.
Most of the isolated DWNTs that were ob-
served have outer tube diameters ranging from 1.6

W. Ren et al. / Chemical Physics Letters 359 (2002) 196–202
199
to 2.9 nm and inner diameters ranging from 1.0 to
2.0 nm. Figs. 1a–c shows three isolated DWNTs
with same diameter. A few very large or small
DWNTs were also observed, such as those shown
in Fig. 1d (outer diameter 3.5 nm, inner diameter
2.7 nm) and the isolated DWNT in Fig. 1e (outer
diameter 1.73 nm, inner diameter 0.89 nm). A tip
of a DWNT was observed, as shown in Fig. 1f.
The HRTEM image indicates that the tip is
opened and the adjacent graphene layers are
closed in the tip.
The experimental observations show that the
interlayer spacing of DWNTs is not a constant,
ranging from 0.34 to 0.41 nm, in consistent with
the results obtained by other investigations [6,7].
The inner tube diameters are different, even
though the outer tube diameters of the DWNTs
are identical, such as those shown in Fig. 1h (outer
diameter 2.0 nm and inner diameter 1.23 nm) and
in Fig. 1g) (outer diameter 2.0 nm and inner di-
ameter 1.34 nm). Charlier et al. [9] considered that
this phenomenon was related to the different chi-
rality combinations of coaxial tubes.
In some DWNTs, fragments of third and fourth
layers were observed, as shown in Figs. 1b–d. The
diameter variation in one DWNT was also ob-
served (Fig. 1i), similar to the observation by
Hutchison et al. [6]. An interesting phenomenon is
that a DWNT can transit into a four-walled car-
bon nanotube at a junction (Fig. 1j), possibly due
to the change of energy resulting from the for-
mation of defects. Moreover, some DWNTs are
hollow and others include some impurities in
the hollow component, as shown in Figs. 1a
and k.
SWNTs were occasionally observed, whose di-
ameters were in general found to be close to the
outer tube diameter of the observed DWNTs, as
shown in Figs. 1k and l. A few SWNTs were also
occasionally observed in a DWNT bundle also, as
shown in Fig. 1k. A three-walled carbon nanotube
was observed in the sample (Fig. 1m), whose outer
tube diameter is close to the outer tube diameter of
DWNTs and SWNTs.
Fig. 2. A bundle of DWNTs observed by HRTEM, which
contains nine DWNTs, arranged in a triangular lattice (with
lattice constant 2.47 nm). Eight of them have a uniform di-
ameter of ꢀ2.15 nm, whose cross-section in the bundle shows a
concentric circle structure. A larger DWNT, as denoted by an
arrow, is flattened against to the adjacent smaller DWNTs,
forming an elliptical cross-section structure.
tance and diameter of the DWNTs are easily
obtained by precise measurements on the con-
centric circle, being ꢀ0.36 and ꢀ2.15 nm. The
HRTEM image shows that within the bundle,
the DWNTs form a two-dimensional triangular
lattice with a lattice constant of 2.47 nm, through
van der Waals intertube bonding, and an inter-
tube separation of 0.324 nm at closest approach
within a bundle, as similar to SWNT bundle [3].
Another characteristic of the bundle is that most
of the DWNTs in the bundle have a uniform
diameter, except for one larger DWNT, as de-
noted by an arrow. Moreover, the DWNTs with
a small diameter in the bundle remain almost
perfectly as rigid concentric cylinders, while the
DWNT with larger diameter is flattened against
to the adjacent smaller DWNTs, forming an el-
liptical cross-section structure. This is likely the
result of the effect of subtle competition between
elastic and van der Waals force on the structure
of DWNTs in the bundle, similar to SWNT
bundles [10].
3.2. Diameter distribution of the as-prepared
DWNTs
The diameters of 57 isolated DWNTs were
measured, with an accuracy of about ꢁ0.1–0.15
nm, based on HRTEM images similar to those
shown in Fig. 1. From these measurements a di-
A HRTEM view of the tip of a DWNT
bundle (shown in Fig. 2) was observed. The
cross-section of each DWNT in the bundle shows
a concentric circle structure. The interlayer dis-

200
W. Ren et al. / Chemical Physics Letters 359 (2002) 196–202
3.3. Raman scattering measurements of the as-
prepared DWNTs
ameter distribution for our DWNTs was deduced,
and the results are shown in Fig. 3. We can see that
more than 90% of the DWNTs have outer diam-
eter ranging from 1.6 to 2.8 nm, with a Gaussian
mean diameter of 2:26 ꢁ 0:01 nm, and inner di-
ameter ranging from 1.0 to 2.0 nm, with a
Gaussian mean diameter of 1:52 ꢁ 0:01 nm. These
results show that the DWNTs synthesized by our
methane catalytic decomposition method have
smaller diameters than those grown by the hy-
drogen arc discharge method, which have outer
diameters ranging from 1.9 to 5.0 nm, inner di-
ameter ranging from 1.1 to 4.2 nm. Moreover, the
diameter distribution of our DWNTs is narrower
than those prepared by the hydrogen arc discharge
method. This point is beneficial to Raman scat-
tering study associated with the radial breathing
mode (RBM) related to the small diameter of the
DWNTs.
It is well known that resonant Raman scattering
measurements associated with the RBM of small
diameter CNTs is a particularly valuable tool for
detecting SWNTs and measuring the diameter of
SWNTs [11]. The RBM can not be observed for
the MWNTs. However, the RBM resonance of
DWNTs, a critical type between SWNTs and
MWNTs, may possibly be observed. In fact, the
RBM was detected for those DWNTs derived
from the chains of fullerenes in single-walled car-
bon nanotubes [7], while it was not observed for
the DWNTs with large diameter synthesized by
hydrogen arc discharge method [6]. Carbon na-
notubes with large diameter are not expected to
exhibit a large (resonantly enhanced) Raman
cross-section, their RBM band is therefore much
more difficult to be detected. Because of smaller
diameter of our DWNTs, the RBM of those is
expected to be detected.
A DWNT can be considered as two coaxial
SWNTs separated radically by the interlayer dis-
tance of graphene sheet, about 0.34 nm in a geo-
metric shape. Theoretical calculation indicates that
the interlayer interaction does not open the energy
gap. Thus it is concluded that the outer and inner
layer retain the basic electronic properties of each
constituent graphene monolayer tubule, even
though the detailed energy dispersion relations are
affected by the interlayer interaction [5]. We con-
sider that the effects of coupling between adjacent
graphene layers on the vibrational and electronic
states are not of sufficient strength to significantly
affect the physical properties of SWNTs due to the
weak van der Waals interaction, as similar to the
intra-bundle and inter-tube coupling in a SWNT
bundle. Therefore, it is not expected that many
new Raman modes actually be observed, and the
spectral intensity should still be dominated by the
selection rules for the two isolated SWNTs con-
stituting a DWNT. As an inference, the frequency
upshift phenomenon should be observed due to
the interaction between the outer and inner layers
of the DWNT.
Fig. 3. Diameter distribution of the as-prepared DWNTs using
data obtained from HRTEM images. The Gaussian fit to these
data gives a mean diameter of 2:26 ꢁ 0:01 nm in outer diameter
and 1:52 ꢁ 0:01 nm in inner diameter.
To realize the in situ Raman scattering mea-
surements, the as-prepared sample in a grid was

W. Ren et al. / Chemical Physics Letters 359 (2002) 196–202
201
Table 1
Raman peak positions and the calculated diameters for the as-
prepared DWNTs
used, which was shown to contain more than 70%
of DWNTs by HRTEM observation. The Raman
spectra taken from the sample are shown in Fig. 4.
The RBM band similar to that of SWNTs was
observed, which mainly consists of two compo-
nents at low frequency region forming pair peaks.
The two RBM bands centered at 112 and
197 cmꢂ1. Another characteristic of the RBM
band is that each RBM band has some Raman
shifts with a small frequency interval, which is
thought to be caused by the two-graphene layers in
the DWNT and their charity.
Because the as-prepared DWNTs are mostly
bundle-like, the formula: xrðdtÞ ¼ 254 ðcmꢂ1
nmÞ=dt ðnmÞ [12], where xr is the Raman shift,
and dt is the diameter of tubes, was adopted to
calculate the diameter of the outer tubes and inner
tubes. Table 1 summarizes the peak positions and
their calculated tube diameters. It is found that the
diameters corresponding to the lower frequency
band obtained by the calculation are almost equal
to the outer diameters obtained by the measure-
ments based on HRTEM images, and the higher
frequency band correspond to the inner diameter.
Moreover, HRTEM observation shows that few
SWNTs with diameter about 2.3and 1.3 nm were
found. Therefore, we can conclude that the two
mean peaks are the RBM of the two coaxial
SWNTs constituting a DWNT. Tubes with large
Outer diameter of the as-
prepared DWNTs,
xrðdtÞ cmꢂ1
Inner diameter of the as-
prepared DWNTs,
xrðdtÞ cmꢂ1
(nm)
(nm)
92.3 (2.75)
103.9 (2.44)
112.9 (2.25)
128.3 (1.98)
130.9 (1.94)
146.2 (1.74)
155.2 (1.64)
164.1 (1.55)
170.5 (1.49)
175.6 (1.45)
190.9 (1.33)
197.2 (1.29)
197.2 (1.29)
207.4 (1.22)
212.5 (1.20)
207.4 (1.22)
212.5 (1.20)
246.6 (1.03)
275.6 (0.92)
284.4 (0.89)
119.3 (2.13)
128.3 (1.98)
130.9 (1.94)
139.8 (1.82)
146.2 (1.74)
diameter (>3 nm) exhibit a small (resonantly en-
hanced) Raman cross-section, their RBM band is
therefore much more difficult to be detected.
Therefore, the lower frequency corresponding to
larger diameter CNTs were not detected. More-
over, another phenomenon was found that the
higher frequency RBM band is centered about
197 cmꢂ1, which give an average diameter of 1.3
nm, smaller than the value obtained by HRTEM
observations. We believe that the smaller upshifts
may be the result of an additional tube-wall re-
storing force due to the effect of outer–inner tube
interactions.
From Table 1, we can see that an outer diam-
eter incorporates with several possible candidates
of inner tubes, which is consistent with our
HRTEM observations, as discussed above. The
tube diameters cannot take continuous values but
discrete due to the effect of their chirality [9].
Therefore, several secondary inner tubes with dif-
ferent diameters can be grown near an ideal tube
diameter determined only by the van der Waals
interaction with parent outer tube, even slightly
increasing the total electronic energy of
DWNTs.
Fig. 4. Raman spectrum, using an excitation laser energy of
1.92 eV, of the RBM for the as-prepared DWNTs.

202
W. Ren et al. / Chemical Physics Letters 359 (2002) 196–202
4. Conclusions
References
[1] S. Iijima, Nature 354 (1991) 56.
1. DWNTs were synthesized by catalytic decom-
position of methane in the presence of Fe cata-
lyst, at 1373 K. The morphology of the product
was observed carefully by HRTEM, showing
most of the as-prepared product is bundles con-
sisting of well-aligned DWNTs. Moreover, A
triangular lattice arrangement of DWNTs in a
DWNT bundle was observed.
2. The diameters of the as-prepared DWNTs,
which were determined from HRTEM images,
mostly range from 1.6 to 2.9 nm in outer diam-
eter and 1.0–2.0 nm in inner diameter, having a
mean value of 1.52 and 2.26 nm from a Gauss-
ian distribution analysis. Moreover, the inter-
layer spacing of DWNTs is not a constant,
ranging from 0.34 nm to 0.41 nm.
[2] C. Journet, W.K. Maser, P. Bernier, A. Loiseau, M. Lamy
de la Chapelle, S. Lefrant, P. Deniard, R. Lee, J.E. Fischer,
Nature 388 (1997) 756.
[3] A. Thess, R. Lee, P. Nikolaev, H.J. Dai, P. Petit, J. Pobert,
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.
[4] H.M. Cheng, F. Li, G. Su, H.Y. Pan, L.L. He, X. Sun,
M.S. Dresselhaus, Appl. Phys. Lett. 72 (1998) 3282.
[5] R. Saito, G. Dresselhaus, M.S. Dresselhaus, J. Appl. Phys.
73 (1993) 494.
[6] J.L. Hutchison, N.A. Kiselev, E.P. Krinichnaya, A.V.
Krestinin, R.O. Loutfy, A.P. Morawsky, V.E. Muradyan,
E.D. Obraztsova, J. Sloan, S.V. Terekhov, D.N. Zakharov,
Carbon 39 (2001) 761.
[7] S. Bandow, M. Takizawa, K. Hirahara, M. Yudasaka,
S. Iijima, Chem. Phys. Lett. 337 (2001) 48.
[8] L.J. Ci, J.Q. Wei, B.Q. Wei, J. Liang, C.L. Xu, D.H. Wu,
Carbon 39 (2001) 329.
3. The as-prepared DWNTs show corresponding
RBM Raman shift associated with their inner
tubes as well as outer tubes. The diameters of
DWNTs obtained from the Raman measure-
ments are in good agreement with the result
from HRTEM observations.
[9] A. Charlier, E. Mcrae, R. Heyd, M.F. Charlier, D.
Moretti, Carbon 37 (1999) 1779.
[10] J. Tersoff, R.S. Ruoff, Phys. Rev. Lett. 73 (1994)
676.
[11] M.S. Dresselhaus, P.C. Eklund, Adv. Phys. 49 (2000)
705.
[12] H.M. Cheng, F. Li, X. Sun, S.D.M. Brown, M.A. Pimenta,
A. Marucci, G. Dresselhaus, M.S. Dresselhaus, Chem.
Phys. Lett. 289 (1998) 602.
Acknowledgements
The authors acknowledge NSFC Grants No.
50025204 and State Major Basic Research Project
of MOST, China for support of this work.