See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312547113

Production of Carbon Nanostructures in Biochar, Bio-oil and Gases from

Bagasse via Microwave Assisted Pyrolysis Using Fe and Co as Susceptors

Article in Journal of Analytical and Applied Pyrolysis · January 2017

DOI: 10.1016/j.jaap.2017.01.018

CITATIONS READS

78 1,888

3 authors, including:

Raja sekhar Reddy Ravikrishnan Vinu

Indian Institute of Technology (ISM) Dhanbad Indian Institute of Technology Madras
15 PUBLICATIONS 467 CITATIONS 158 PUBLICATIONS 5,516 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Raja sekhar Reddy on 28 October 2017.

The user has requested enhancement of the downloaded file.

 Materials Science and Engineering B 119 (2005) 105–118

Review

Characterization methods of carbon nanotubes: a review

T. Belin, F. Epron∗
Laboratoire de Catalyse en Chimie Organique-LACCO UMR6503, CNRS/Université de Poitiers, 40,
Avenue du Recteur Pineau, 86022 Poitiers Cedex, France

Received 19 October 2004; received in revised form 24 January 2005; accepted 9 February 2005

Abstract
Carbon nanotubes due to their specific atomic structure have interesting chemical and physical properties according to those of graphite
and diamond. This review covers the characterization methods of carbon nanotubes which are most employed today. The structure of carbon
nanotubes is first briefly summarized followed by a description of the characterization methods such as STM, TEM, neutron diffraction, X-ray
diffraction, X-ray photoelectron spectroscopy, infrared and Raman spectroscopy. The most interesting features are indexed for each technique.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Carbon; Nanotubes; Characterization methods; Chemical properties; Physical properties; Structure

Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
2 Carbon nanotubes structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3 Morphological and structural characterizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.1 Photoluminescence spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.2 X-ray photoelectron spectroscopy (XPS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.3 Scanning tunneling microscopy (STM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.4 Neutron diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.5 X-ray diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.6 Transmission electronic microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.7 Infrared spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.8 Raman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4 Beyond the characterization techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.1 Special behaviors of the very small diameter nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.2 Separation of semiconducting and metallic nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

1. Introduction the development of studies in the field of physics, chem-

istry and material sciences. Many works were completed
Carbon nanotubes (CNTs) were first described by Iijima on the structure, properties [2,3] and potential applications
[1] and since their discovery, they have contributed to [4,5].
Nanotubes have stimulated interest from a fundamental
∗ Corresponding author. Fax: +33 5 4945 3499. research as well as from an advanced point of view. Many
E-mail address: [email protected] (F. Epron). properties of nanotubes are directly influenced by the way of

0921-5107/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.mseb.2005.02.046
106 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

the graphene sheets are wrapped around. Due to their cylin-

drical geometry and their nanometric scale, CNTs are candi-
dates for potential applications like hydrogen storage [6–9].
However, the amount of hydrogen stored in CNTs is below
the DOE 2015 requirements (9 wt.%) [10]. Nevertheless, hy-
drogen storage properties of nanotubes are not yet fully un-
derstood and explored. Careful investigations are needed to
improve the sorption behaviors [11].
Single wall carbon nanotubes (SWNTs) with their well-
defined atomic structure, their high length to diameter ra-
tio, and their chemical stability constitute one-dimension
molecules [12]. SWNTs exhibit specific electronic proper-
ties and can be either metallic or semiconducting depending
on their geometry.
Production of carbon nanotubes in a controlled way and
in large amount encounters problems, which remains to be
solved. It is needed to identify all properties of these tubes.  and chiral angle θ definition for a (2, 4) nanotube on
Fig. 1. Chiral vector C
However, to fully characterize carbon nanotubes, there are graphene sheet. a1 and a2 are the unit cell vectors of the two-dimensional
hexagonal graphene sheet. The circumference of nanotube is given by the
not so many techniques available. The aim of this paper is
length of chiral vector. The chiral angle θ is defined as the angle between
about the characterization methods of the CNTs. Structure of  and the zigzag axis.
chiral vector C
CNTs are first summarized. Then, morphological and struc-
tural characterization methods are reviewed. The particular
behavior of nanotubes with very small diameter and the sepa- gonized or scrolls graphene sheets [14] (Fig. 2). The coax-
rating methods of the semiconducting/metallic tubes are also ial cylindrical model is widely accepted for CNTs, however
approached before concluding this paper. polygonized tubes are also observed. These are generally re-
stricted to large tube sizes which allows three-dimensionally
correlated regions: low-angle tilt and well-aligned boundaries
2. Carbon nanotubes structure are then observed [4].
The length of the chiral vector C is the circumference of
Carbon nanotubes are sheets of graphite that has been the nanotube and is given by the corresponding relationship:
rolled into a tube [13]. Nanotubes are considered as nearly
one-dimensional structures according to their high length to

 = a (n2 + nm + m2 )
c = |C|
diameter ratio. Most important structures are single walled
nanotubes (SWNTs) and multi walled nanotubes (MWNTs).
where the value a is the length of the unit cell vector a1 or
A SWNT is considered as a cylinder with only one wrapped
a2 . This length a is related to the carbon carbon bond length
graphene sheet. Multi walled nanotubes (MWNTs) are sim-
acc by the relation:
ilar to a collection of concentric SWNTs. The length and di-
ameter of these structures differ a lot from those of SWNTs √
a = |a1 | = |a2 | = acc 3
and, of course, their properties are also very different.
Single walled nanotubes are completely described, except
 (called chiral vector). For graphite, the carbon carbon bond length is acc =
for their length, by a single vector C
0.1421 nm. The same value is often used for nanotubes [8].
Two atoms in a planar graphene sheet are chosen and one is
 is pointed from the first But probably, a slightly larger value such as acc = 0.144 nm
used as origin. The chiral vector C
should be a better approximation due to the curvature of the
atom toward the second one (Fig. 1) and is defined by the
relation:

 = na1 + ma2
C

where n and m are integers. a1 and a2 the unit cell vectors of
the two-dimensional lattice formed by the graphene sheets.
The direction of the nanotube axis is perpendicular to this
chiral vector.
Multi-walled carbon nanotubes are described by different
models which are in good agreement with experiments and in Fig. 2. Various models of carbon nanotubes taking into account the experi-
particular with the electron microscopic images. CNTs may mental measurements: (a) coaxial cylindrically curved, (b) coaxial polygo-
be formed from coaxial cylindrically curved, coaxial poly- nized or (c) scroll graphene sheets (according to [14]).
 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118 107

tube [15–17]. Using the circumferential length c, the diam- 3. Morphological and structural characterizations
eter of the carbon nanotube is thus given by the relation:
D = c/π. Fullerenes, which are carbon compounds and similar
The angle between the chiral vector and zigzag nanotube to CNTs, are closed spheres composed of pentagons and
axis is the chiral angle θ (Fig. 1). With the integers n and m hexagons. Their curvatures are given by the organization
already introduced before, this angle can be defined by: of these pentagons and hexagons. In carbon nanotubes, the
 Stone–Wales transformation is obtained by a reversible di-
√ 
m 3 atomic interchange. The resulting structure is formed of two
θ = tan−1 pentagons and two heptagons in pair. This transformation in-
m + 2n
duces a new defect in CNTs: the heptagon which allows for
concave areas within the nanotube. Thus, there is not only
Nanotubes are only described by the pair of integers (n, m)
straight tubes with hemispherical caps and many equilibrium
which is related to the chiral vector. Three types of CNTs
shapes can be obtained [13].
are revealed with these values: when n = m, the nanotube
Electronic and mechanical properties of CNTs are unique.
is called “armchair” type (θ = 0◦ ); when m = 0, then it is
However, to achieve these properties, the CNTs must be
of the “zigzag” type (θ = 30◦ ). Otherwise, when n = m,
chemically processed in order to purify and bring appropri-
it is a “chiral” tube and θ takes a value between 0◦ and
ate functionalizations. Typical treatments involve the use of
30◦ . The value of (n, m) determines the chirality of the
oxidative methods with nitric acids. The caps at both ends of
nanotube and affects the optical, mechanical and electronic
the CNTs are removed and defects such as carboxylic acid
properties. Nanotubes with |n − m| = 3q are metallic and
groups on surface are revealed by these purification tech-
those with |n − m| = 3q ± 1 are semiconducting (q is an
niques [32]. The properties of the nanotubes are influenced
integer).
by these defects sites on the walls and at the ends [33]. A
The terminating cap of nanotube is formed from pentagons
determination of the concentration of these defects would be
and hexagons. The smallest cap that fits on to the cylinder
helpful for nanotubes properties characterizations. The con-
of the carbon tube seems to be the well known C60 hemi-
centration of the carboxylic acid groups formed by purifica-
sphere. The smallest experimental value of nanotube diame-
tion treatments could be determined by various methods: (i)
ter of 0.7 nm is in good agreement with this cap [2]. However,
the determination of gases concentrations at high tempera-
some authors have recently studied nanotubes at the theoret-
tures could be used as described by Mawhinney et al. [33]
ical limit of 0.4 nm in diameter [18–25]. These nanotubes
with CO(g) and CO2(g) ; (ii) calibrated energy-dispersive X-
are sometimes capped with a C20 dodecahedron [19]. Ac-
ray spectroscopy allows investigation of the atomic oxygen
cording to Liang et al. [20], single-walled 0.4 nm CNTs have
percentage. According to Chen et al. [34], the determination
three possible structures: chiral (4, 2) (diameter = 0.414 nm),
of the atomic oxygen percentage provides an estimation of the
zigzag (5, 0) (diameter = 0.393 nm) and armchair (3, 3)
amount of chemically bonded compounds containing oxygen
(diameter = 0.407 nm) tubes.
such as carboxylic groups; (iii) chemical titration methods are
If we consider MWNTs, there is only a few possible se-
also performed to macroscopically estimate the defect den-
quence of (n, m) tubes to keep realistic intershell distance.
sity of tubes. One of these titration techniques is described
For carbon materials, the intershell spacing d between two
by Hu et al. [32]: a forward titration with sodium bicarbonate
successive tubes is in the range from 0.344 nm [1,26,27] to
(NaHCO3 ) is first used with SWNTs containing carboxylic
0.36 nm [28,29] (perfect to disordered materials). These val-
groups and could be completed with a backward titration us-
ues suggest a possible dependence of intershell distance on
ing HCl on NaHCO3 -treated SWNTs.
the tube size and some authors gives an empirical relationship
In order to investigate the morphological and structural
to fit TEM experimental data [30]:
characterizations of nanotubes, a reduced number of tech-
 c  niques could be used. However, only few techniques are able
d = 0.344 + 0.1 exp − to characterize CNTs at the individual level such as scanning
4π
tunneling microscopy (STM) and transmission electronic mi-
where the term c/4π is the radius of the tube. As a conse- croscopy (TEM). X-ray photoelectron spectroscopy is help-
quence of the tube diameter increase, the intershell distance ful in order to determine the chemical structure of nanotubes
decreases to 0.344 nm. while neutron and X-ray diffraction, infrared and Raman
Using geometrical calculations, Peigney et al. [31] have spectroscopy are mostly global characterization techniques.
determined the theoretical external specific surface area for
CNTs. For one side of graphene sheet, the value obtained is 3.1. Photoluminescence spectroscopy
1315 m2 g−1 but using different multi-walled geometries and
nanotubes bundles the specific area decreases to 50 m2 g−1 . As already described, the SWNTs could be considered ei-
According to their conclusions, specific area being a value ther as metallic or semiconducting type. The gap energy of
easily measured, this should give a good information of CNT the semiconducting tubes is related to the chirality [35,36]
characteristics. and is approximately proportional to the inverse of the tube
108 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

diameter [37]. Photoluminescence from the recombination real gap (E11 ) followed by a series of sharp interbands tran-
electron–hole pairs at the bandgap has to be expected. How- sitions (E22 as an example) due to Van Hove singularities
ever, as noticed by Obraztsova et al., the SWNTs are generally maxima [40]. These interbands transitions are also called
grouped in bundles. In a bundle, nanotubes interact between “pseudo”-gaps because of the very small, but non-zero DOS,
each others with Van der Waals force. Few nanotubes in these between the Van Hove singularities. The photoluminescence
bundles are metallic and are similar to non-radiative chan- emission is then only possible with E11 transition while a
nels. The luminescence of the semiconducting tubes in these wide range of wavelength are useful for the photolumines-
bundles is relaxed inside these channels. Often, due to this cence excitation. The energy of the Van Hove singularities
interaction between semiconducting and metallic nanotubes, maxima is mainly dependent on the nanotube diameter [40].
no photoluminescence signal is obtained. Thus, different (n, m) SWNTs in a sample will show various
In order to observe the photoluminescence phenomenon, superpositions of distinct Exx transitions which will appear
the bundles must be separated into individual tubes. Some with different wavelengths [43]. As a consequence, the nature
treatments could be used to achieve this separation. Maybe (semiconducting or not), the geometries and the diameters
one of the most popular techniques is the ultrasonication treat- could be accessible using the photoluminescence technique.
ment of the nanotubes with surfactants in water suspension Moreover, the luminescence spectra seems to be very sensi-
such as the SDS (sodium dodecyl sulfate) [36,38–41]. This tive to the presence of chemical defects and to the purity of
process could be done using various CNTs samples. The pho- the samples [36].
toluminescence is also observed using individual nanotubes
grown in channels of zeolite [42]. 3.2. X-ray photoelectron spectroscopy (XPS)
The quasi-one-dimensionality of SWNTs has some elec-
tronic properties described by theory. The carbon atoms in XPS technique can give information about the chemical
nanotubes have p electrons which give a large ␲-electrons structure of carbon nanotubes. But the most widely used data
system. Van Hove singularities are observed in the density of refers to the structure modification of the CNT walls due to
electronic states of these particular ␲-electrons systems (Fig. the chemical interaction with organic compounds or gases
3). Then, Weisman et al. [39] attributes all the optical spec- adsorption. Droppa et al. [44] have studied by XPS the incor-
troscopy of the semiconducting tubes to transitions between poration of nitrogen into carbon nanotubes. The N1s and C1s
the corresponding Van Hove singularities in the valence and peaks have been investigated. Compared to non-nitrogenated
conductance band. These transitions are called E11 , E22 , . . . samples, the C1s peak exhibits both a shift and an asymmet-
in the Fig. 3. An electronic DOS for SWNT has only one ric broadening to higher binding energies. The tip-shift of
this peak, due to the polar character of the carbon nitrogen
bond [45], is an evidence of the incorporation of nitrogen
into the nanotube structure. XPS analysis were also per-
formed by Pham-Huu et al. [46] on carbon nanofibers and
the C1s peak was compared to that obtained from a clean
highly ordered pyrolytic graphite (HOPG) (0 0 0 1) surface.
The graphitic nature of the carbon nanofibers is revealed
by the relative agreement of the maximum of binding en-
ergy and the surface plasmon peak at 291 eV. However, a
wide broadening of the peak is ascribed to defects and C H
terminating surfaces. At about 286.5 eV, a slight broaden-
ing can be attributed to surface oxygen groups with single
bonds. Nevertheless, a contribution at about 289.5 eV de-
scribed surface oxygen groups with multiple carbon oxygen
bonds. These authors conclude that the nanofiber mate-
rial is more like a carbon oxide than a variety of graphite
[46].
By studying the sidewall functionalization of SWNTs by
fluorination, Lee et al. [47] have shown that the C1s of un-
doped SWNTs is composed of three peaks: sp2 carbon (peak
at about 284.3 eV), sp3 carbon (285 eV) and oxygen related
groups (carboxyl) at about 288.5 eV (similar to those ob-
served in nanofibers). In another way, some experiments on
MWNTs (15–20 nm in diameter) were also performed by
Fig. 3. A scheme of photoluminescence excitation for a semiconducting
SWNT. The valence and conductance Van Hove singularities are labeled Chen et al. [48] and revealed the relative differences between
with the sub band index. The dashed lines show non-radiative transitions for the graphite and the carbon nanotubes in: (i) the C1s peak
electrons and holes. Optical transitions E11 and E22 are also represented. energy position, (ii) the full width at half maximum and (iii)
 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118 109

its energy loss fine structures. The C1s of graphite is usually 3.4. Neutron diffraction
observed at 284.6 eV. However, a negative shift of 0.3 eV ob-
served in the binding energy of these MWNTs was explained Local details such as size, diameter can be obtained by Ra-
by the weaker C C bonds resulting from the curvature of the man or local electron diffraction but, to investigate CNTs at
graphene sheets and by the larger interlayer spacing. With larger scale X-ray diffraction or neutron diffraction is needed.
MWNTs (30 nm in diameter) aligned along their tube axis Neutron diffraction is widely used for determination of struc-
and using spatially resolved XPS, Suzuki et al. [49] have tural features such as bond length and possible distortion of
demonstrated that the shift of the C1s binding energy is only hexagonal network [61]. Moreover, contrary to X-ray diffrac-
about 50 meV between the tips and the sidewalls. The strong tion, a wide range of scattering vector Q can be explored due
influence of structural defects on the electronic structure of to the weak decrease of the atomic factor with Q [62]. Many
MWNTs is called upon by these authors to explain this dif- Bragg reflections and the use of radial distribution function
ference. allow to obtain much more structural features than X-ray.
For the smallest diameters, a distinction between armchair,
3.3. Scanning tunneling microscopy (STM) zigzag and chiral nanotubes can be achieved thanks to the
effect of curvature and chirality on the powder diffraction
To use this technique, carbon nanotubes must be deposited pattern.
on a flat conducting substrate such as HOPG [50–52] or Au Studying carbon carbon bond length in CNTs with neu-
[8,53–55]. STM images give directly the three-dimensional tron diffraction, Burian et al. [61] have shown the same value
morphology of tubes and are consistent with the structure as the bond length of graphite (about 0.141 nm). However,
inferred from scanning electron microscopy [50]. Moreover, Dore et al. [63] noticed two types of bond length (respectively,
STM can resolve simultaneously both the atomic structure 0.1395 ± 0.0002 and 0.153 ± 0.001 nm) with a weighted
and the electronic density of state (DOS). Thus, this technique mean value close to the graphite one (0.142 nm). This weak
is suitable for the nanotube characterization. Experiments to distortion of carbon hexagon is related to the distribution of
determine atomic structure of SWNTs were realized at low single and double bond character over the network. The inter-
temperature by Wildöer et al. [8] (4.2 K), Odom et al. [9] layer spacing in nanotubes is greater than the normal graphite
(77 K), Kim et al. [55] (77 K) and Ichimura et al. [51] (77 K). spacing of 0.3354 nm [61,63]. With MWNT samples, a value
The atomic structure of SWNTs on top of a rope was also of 0.345 nm is obtained by Burian et al. [61].
obtained by various authors [54,56]. These studies confirmed A comparison between neutron diffraction and theoreti-
the predicted behaviors of SWNTs [57–59] and revealed the cal data was made by Giannasi et al. [64]. They noticed the
sensitivity of the electronic properties to the helicity and tube modification of the neutron diffraction pattern due to various
diameter. parameters such as finite size of nanotube bundles, number
Using STM atomically resolved images, the measured of tubes in bundles, and nanotube diameters polydispersivity
chiral angle θexp of nanotubes can be determined by mea- [65]. According to these authors [64], the diffraction pattern
suring the angle between the tube axis and the (n, n) di- can be divided into two Q regions: the bundles packing is de-
rection for θexp < 15◦ . When θexp > 15◦ , the angle θind is scribed for Q < 2 Å−1 and the SWNTs features for Q > 2
measured between the tube axis and the (n, 0) direction Å−1 . Koloczek et al. [66] used the Debye–Waller equation to
and the chiral angle θexp is determined from the relation compute powder diffraction pattern of various single-walled
θexp = 30◦ − θind . This approach as described by Kim et al. and multi-walled CNTs. The Debye–Waller factor has been
[55] limits the contribution of the curved sides of the CNTs included to take into account the thermal vibrations of carbon
which could be distorted by the finite size and the asym- atoms. Neutron experiments were compared to these theoret-
metry of the tip. These authors also noticed that a subtle ical results: diffraction patterns of CNTs were sensitive to the
difference must be made between the measured chiral an- diameter, the size, the nearest neighbor interatomic distance
gle θexp (and tube diameter) and the true chiral angle θ (and and the value of the Debye–Waller factor. However, for the
diameter) due to the fact that the STM images are the con- smallest diameters, it remains possible to distinguish between
volution of the tip shape and samples. The (n, m) indices the three fundamental forms of nanotubes.
were then obtained from the chiral angle θ and the tube di-
ameter using the relations previously described in this re- 3.5. X-ray diffraction
view.
Using computer simulations, Krasheninnikov [60] has This characterization method is not sample destructive and
predicted the STM images of both metallic and semicon- is used to obtain some information on the interlayer spac-
ductor SWNTs with atomic vacancies. On STM images, the ing, the structural strain and the impurities. However, car-
vacancies resulted in an hillock-like features with lateral size bon nanotubes have multiple orientations compared to the
of 0.5–0.8 nm (depending on CNTs chirality) and height up X-ray incident beam. Diameters and chiralities distribution
to 0.1 nm. A comparison of the electronic superstructure near are also observed as well as various number of layers for
vacancy can be done with that observed on graphite near point MWNTs. This leads to a statistical characterization of carbon
defect. nanotubes.
110 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

Due to their intrinsic nature, the main features of X-ray Some studies on the alignment of CNTs have been done
diffraction pattern of CNTs are close to those of graphite (Fig. using X-ray diffraction. Cao et al. [70] have shown that no
4): (i) a graphite-like peak (0 0 2 l) is present and measure- (0 0 2) peak can be measured by X-ray diffraction with well-
ments of interlayer spacing can be obtained from its position aligned straight nanotubes on the substrate surface. In the
using the Bragg law, (ii) a family of (h k 0) peaks due to the case of carbon nanotubes with tube axis perpendicular to the
honeycomb lattice of single graphene sheet. Consequently, substrate surface, the X-ray incident beam is scattered inside
the X-ray diffraction profile is not useful to differentiate mi- the sample and is not collected. Consequently, the intensity of
crostructural details between the CNTs and the graphite struc- the (0 0 2) peak decreases monotonically as CNTs are better
ture [67] but can help to determinate the sample purity (cat- aligned.
alyst, functional groups). A mean diameter of CNTs can be calculated through the
Using the (0 0 2 l) peaks position, the interlayer spacing is use of the Debye–Scherrer relation on the (0 0 2) peak. There
often found to be larger than in HOPG and close to the value is contributions from both the nanotubes and residual graphite
determined in turbostratic graphite [62,27]. The peak position in the (0 0 2) peak so the values are calculated from peak
shifted from 26.5◦ for graphite to ∼26◦ in 2θ for SWNTs. decomposition using pseudo-Voigt profiles.
Moreover, the line shape of the (0 0 2 l) peaks is weakened The simulation of X-ray diffraction patterns were made
and somewhat broadened on its low diffraction angle part possible by the use of computer efficient methods like
compared to graphite ones. The asymmetry is caused by the Cochran–Crick–Vand (CCV) [73]. This method previ-
presence of different crystalline species. Indeed, at least two ously applied to the X-ray diffraction of helical biological
structures are generally present and are difficult to separate: molecules has been used by many authors [14,74–76] to com-
the pure graphite particles consisting of a stack of graphene pute X-ray or electron diffraction pattern of CNTs.
sheet, and the wrapped graphitic plane constituting a carbon The amplitude scattered by multiple carbon atoms is given
nanotube. On the other hand, the decrease of the interlayer by the very well-known relation:
spacing with the increase of diameter of the shells [30] and 
the inner diameter distribution [62] modify also the shape A(k) = fc (k) iq · rj
of the (0 0 2 l) peaks. The intensity and width of the (0 0 2 l) j
peaks are related to the number of layers, to the variations of
interlayer spacing, to the lattice distortions [68,69] and to the where fc (k) is the X-ray atomic scattering factor, q the
carbon nanotube orientation compared to the X-ray incident diffraction vector defined by q − q0 (q and q0 are, respec-
beam [70]. tively, the wave vector of the scattered and incident wave)
The (h k 0) peaks present an asymmetric shape due to the and rj the atomic position. Using this relation adapted to
curvature of the nanotube [62] and the (h k l) reflections only the diffraction of a carbon helix, the diffraction amplitude of
appear in X-ray diffraction patterns with a regular stacking a SWNT can be described as the sum of successive single
of layers (as with flat graphitic layers in polygonized tubes carbon helix [14]. Moreover, a sum of successive diffraction
and with residual carbon particles) [71,72]. amplitude of SWNTs allows to compute MWNTs diffraction
patterns.
The computation of powder diffraction spectra of SWNTs
bundles was obtained by Rols et al. [65] and Kuzmany et al.
[77] using the general formulas for X-ray diffraction. In this
work, they studied the effect of various parameters such as
the mean-tube diameter, the finite size of the bundles and
the diameter dispersivity of the tubes. All these parameters
affect to a significant degree the position and the width of
the (1 0) peak. These authors conclude that these parameters
lead to the systematic overestimation of the tube diameter
with respect to TEM analysis.

3.6. Transmission electronic microscopy

In order to characterize multi-walled nanotubes by TEM,

an image analysis methodology was developed by Gommes
et al. [78]. After numerical treatments, the intensity across
a nanotube section is fitted using a model based on Lam-
bert’s law. Thus, the measurement of outer and inner radius
Fig. 4. XRD pattern of MWNTs synthesized by CVD at CSIC laboratory (di-
ameter of about 60 nm). The incident X-ray wavelength is λ = 0.154056 nm. and linear electron absorption coefficient is allowed. These
The most significant Bragg peaks are noticed with Miller indices. The pres- authors used this method to study MWNTs before and af-
ence of catalysts (Co and Mo) in the CNT sample is shown by stars. ter annealing and have noticed a significant increase of the
 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118 111

Fig. 5. (a) TEM image of a multi-walled nanotube (diameter of about 65 nm) synthesized by CSIC laboratory. These MWNTs were produced by CVD followed
by several oxidation processes. Contrast of the walls are visible. (b) Enlargement of the walls of the nanotube. White lines are used in determination for the
intershell spacing. (c) Mean profile of the intensity levels of the walls showing the fringes of the (0 0 2) layers used in the determination of the intershell spacing.
Here, the value of the intershell spacing is 0.337 ± 0.023 nm and is really close to the graphite one.

electron absorption coefficient. They concluded that a more energy difference between the two adjacent occupied and/or
structured arrangement of the wall material can be an expla- unoccupied orbitals is expected. However, the nature of inter-
nation of this increase. layer interaction remains the same. In graphite, the graphene
The intershell spacing of MWNTs were studied by Kiang layers planes are bonded by a weak Van der Waals force.
et al. [30] using high resolution TEM images. The inter- The arrangement of the tubes in MWNTs is similar to the
shell spacing is found to range from 0.34 to 0.39 nm varying graphene layers observed in turbostratic graphite. The dif-
with nanotube diameter (Fig. 5a–c). However, these values ferent honeycomb sheets of various diameters which form
are somewhat greater than the graphite interplanar distance the MWNTs are therefore uncorrelated. In MWNTs, dopant
(0.336 nm) [79]. This increase in intershell spacing is proba- atoms or molecules could be introduced between adjacent
bly due to the curvature of the graphene sheets which is mod- nanotubes as already observed with intercalation compounds
ified by the tube radius. This curvature leads to an increase in graphite [81]. This leads to an increase of the intershell
of the repulsive force and the size effect is more intense in spacing too.
the small diameter (below 10 nm). These authors noticed also Two or several graphitic layers are crossed by the inci-
that the physical and chemical properties will be changed by dent beam of electron when the nanotube is perpendicular
the modification of the intershell spacing. This can be use- to this one [82–84] (Fig. 6a). These layers are misaligned
ful to increase reactivities or to develop useful features for when the tube is helical. Therefore, the angle of this mis-
storage media. A theoretical approach of the interlayer inter- alignment is equal to twice a helical angle. The resulting
actions in double wall nanotubes has been done by Tanaka et electron diffraction pattern is then splitted into two sets. As
al. [80]. Their conclusions are that an increase of the diameter a first approximation, the curvature of the nanotube can be
of the tube gives a weaker interlayer interaction. A decrease of neglected and the half of the misalignment angle can be used
112 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

as the true helical angle. Qin et al. [82] have demonstrated et al. [89], this distribution can be explained by the range of
that the true helicity of CNTs can be deduced from electron various diameters of tubes within the bundles. For SWNTs
diffraction patterns. A curvature correction factor is deter- samples produced by CCVD method, no preferred helicity is
mined by a Bessel function easily calculated from the tube revealed. However, due to the small number of tubes within
indices (n, m). The chiral indices of carbon nanotubes can be the bundles, one or two well-defined helicities is highlighted
measured by the selected-area electron diffraction (SAED) [89].
technique. This method is based on the semi-quantitative
analysis of the diffraction pattern of nanotubes. The intensi- 3.7. Infrared spectroscopy
ties of the diffraction patterns can be compared to simulations
using kinematic theory [85]. Then, the chiral indices can be There is 7–9 infrared active modes in SWNTs which de-
retrieved for SWNTs or MWNTs up to radii of about 4 nm. pend on the symmetry: chiral, zigzag and armchair [91]. The
The uncertainties on diameter and apparent helicity are, re- main active modes for carbon nanotubes in infrared spec-
spectively, of ±12% and ±0.6◦ . Contrary to spectroscopic troscopy are the A2u and E1u modes [92]. With MWNTs,
methods, no empirical parameter such as an exact value of these phonon modes are observed by Kastner et al. [93,94]
carbon carbon bond length is required. at about 868 and 1575 cm−1 , respectively. Kuhlmann et al.
As widely observed, nanotubes appear sometimes iso- [91] noticed that these modes (around 850 and 1590 cm−1 )
lated, but are most often packed together by Van der Waals appear in all CNTs symmetry independently of the diameters.
forces to constitute bundles. The nanotubes within a bundle In samples containing mostly SWNTs, these authors showed
are packed in a hexagonal close-packed array [86]. This struc- two structures around 874 ± 2 and 1598 ± 3 cm−1 . How-
ture can be described as a two-dimensional hexagonal lattice ever, the frequencies deviate from the graphite frequencies to
(Fig. 6b). The determination of the structures of CNTs within higher values by 5 and 8 cm−1 , respectively. In CNTs char-
the bundles can be obtained using TEM coupled with elec- acterization, infrared spectroscopy is often used to determine
tron diffraction. Various works have been performed to de- impurities remaining from synthesis or molecules capped on
termine the helicity of SWNTs within the bundles [76,84,86– the nanotube surface. Numerous works are performed on or-
88]. Colomer et al. [89] compared SWNT bundles resulting ganic molecules and CNTs: infrared spectroscopy exhibits
from three different methods: laser ablation, electric arc dis- all the modification of the CNTs structure and reveals the na-
charge and CCVD. They confirmed previous works about the ture of compounds added to the CNTs. As an example, Saito
distribution of helicities within the bundles. In most case, the et al. characterized the reaction products of MWNTs chem-
helicity is randomly distributed between the “zigzag” (he- ically modified by amino compounds [95]. These works can
licity about 0◦ and m index = 0) and “armchair” (helicity be classified into two groups: characterization of molecules
about 30◦ and index n = m) structures. On samples synthe- attached to CNTs [96–101], catalytic properties of CNTs
sized by laser ablation, Colomer et al. [89] found numerous [102,103].
well-defined helicities for SWNTs within bundles of small di- Pereira et al. [103] studied the catalytic activities of multi-
ameters. However, the distribution of helicities is attributed walled carbon nanotubes in the oxidative dehydrogenation of
to the coalescence of small bundles to constitute larger bun- ethylbenzene (ODE). They noticed the performance of CNTs
dles. A wide range of helicities is also obtained by other compared to activated carbon and graphite samples: this ma-
results using SAED [88,84]. Using SAED technique, studies terial is more stable under oxidative conditions due to its
on electric-arc produced SWNTs showed a uniform distribu- original structure. The highest catalytic activity is obtained
tion of helicity in the bundles [90]. According to Colomer with MWNTs which are oxidized before the catalytic exper-
iments. The authors revealed that the key parameter is the
concentration of oxygenated surface groups. In another way,
Wang et al. [102] investigated the possibilities of the use of
CNTs for catalytic removal of NOx . The modifications of
the NO, NO2 , NO3 and CNTs carbon signals were followed
by FT-IR. These experiments were carried out over 1 wt.%
Pd/CNTs. In this system, the palladium particles acts as a
catalyst for CNTs hydrogenation. The reducing agent for the
provision of hydrogen and carbon in NO reduction is the hy-
Fig. 6. (a) Cross-section of a multi-walled nanotube. Scheme of the condi- drogenated CNTs.
tions of observation: the nanotube axis is perpendicular with respect to the
incident electron beam. The contrast observed on the TEM images is ob- 3.8. Raman spectroscopy
tained from the portions of the layers colored in white gray. The portions of
the layers colored in dark gray are involved in the measurement of helicity
Raman spectroscopy is one of the most powerful tools
of the nanotubes. The misalignment of these layer lead to the existence of
an helical angle. (b) Definition of the two-dimensional hexagonal lattice of for characterization of carbon nanotubes. Without sample
SWNTs within a bundle. The lattice constant D is given by the diameter of preparation, a fast and non-destructive analysis is possible.
the nanotube and the inter-tube spacing. All allotropic forms of carbon are active in Raman spec-
 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118 113

troscopy [104]: fullerenes, carbon nanotubes, amorphous car- Two different frequencies are attributed to A1g symme-
bon, polycrystalline carbon, etc. The position, width, and rel- try according to Liu et al. [110] and Saito et al. [111]:
ative intensity of bands are modified according to the carbon a low-frequency part (RBM) where atoms move in the
forms [105]. The most characteristic features are summarized same direction and a high-frequency part where atoms
(Fig. 7) as following: (i) a low-frequency peak < 200 cm−1 move in opposite directions. These atomic movements
(or bunch of peaks for polydisperse samples when resonating take place in the plane perpendicular to the nanotube
conditions are met) characteristic of the SWNT assigned to axis.
a A1g “breathing” mode of the tubes, whose frequency de- The “radial breathing mode” (RBM) is directly dependent
pends essentially on the diameter of the tube (RBM: radial on the nanotube diameters through the relation
breathing mode); (ii) a large structure (1340 cm−1 ) assigned
to residual ill-organized graphite, the so-called D-line (D: A
ωRBM = +B
disorder); (iii) a high-frequency bunch (between 1500 and d
1600 cm−1 ) called G band also characteristic of nanotubes,
corresponding to a splitting of the E2g stretching mode of where A and B parameters are determined experimentally
graphite [106]. This bunch could be superimposed with the [112]. However, the frequency of RBM is not related to
G-line of residual graphite [107]; (iv) a second order ob- the chiral angle θ of the nanotubes [113–115]. As an ex-
served mode between 2450 and 2650 cm−1 assigned to the ample, for an isolated SWNT on Si/SiO2 substrate, the
first overtone of the D mode and often called G mode; (v) experimental value of A is found to be 248 nm cm−1 and
a combination mode of the D and G modes between 2775 B = 0 cm−1 [116]. Using a force constant model with inter-
and 2950 cm−1 . actions to the fourth neighbor, Bandow et al. calculated the
As described by Kuzmany et al. [92] and other authors value of A = 224 nm cm−1 [117]. For SWNT bundles, the
[108,109], the active modes for a carbon nanotube are: tube/tube interactions lead to values of A = 234 nm cm−1
and B = 10 cm−1 [118]. A calculated value of RBM pa-
γRaman = 2A1g + 3E1g + 3E2g , γIR = A2u + 2E1u (1)
rameters (A = 232 nm cm−1 and B = 6.5 cm−1 ) is given by
γRaman = 2A1g + 2E1g + 4E2g , γIR = 3E1u (2) Alvarez et al. using a crystal of identical and infinite nan-
otubes and taking into account the Van der Waals interac-
γRaman = 3A1 + 5E1 + 6E2 , γIR = A2 + 5E1 (3) tions between tubes with a Lennard–Jones potential [119].
for zigzag (equation (1)), armchair (equation (2)) and chi- Summarized in Bandow et al. [120], other values of param-
ral tubes (equation (3)) where g- and u-modes are Ra- eter A are given according to the models used: zone fold-
man and infrared active, respectively. Raman frequencies ing method (A = 223 nm cm−1 ) [121], force constant model
and all symmetries described by Kuzmani as well as sev- (A = 218 nm cm−1 ) [113], local density approximation (A =
eral combinations and harmonics are given in Table 1. 234 nm cm−1 ) [114], pseudopotential density functional the-
ory (A = 236 nm cm−1 ) [115] and elastic deformation model
(A = 227 nm cm−1 ) [122]. Fig. 8 illustrates the dependence
of the RBM on the nanotube diameters as described by some
authors. The direct environment of the nanotubes causes also
a modification of the RBM mode. Thus, the correction due
to the bundling (tube/tube interaction) is an up-shift of about
14 cm−1 for (9, 9) nanotubes [123]. The tube/tube interaction
is not the only one to be taken into account, the interaction be-
tween graphene sheets in the case of MWNTs also contributes
to the RBM modification [120]. This phenomenological re-
lation is valid in a weak range of size (1 nm < d < 2 nm).
Indeed, with nanotube diameters below 1 nm, a chirality de-
pendence of ωRBM appears due to the distortion of nanotube
lattice: the relation is thus not valid in this range of diameters.
Moreover, with diameters greater than 2 nm, the intensity of
RBM is weak [112].
For chiral tubes, the G-Band is composed of six modes
with symmetry A1g , E1g and E2g (two of each) [124,125].
However, for achiral tubes such as armchair and zigzag tubes,
Fig. 7. Raman spectrum showing the most characteristic features of CNTs: the G band consists of only three modes which are, respec-
radial breathing mode (RBM), the D band, G band and G band. Sec-
tively, by order of decreasing frequencies: E2g , A1g , E1g and
ond order modes are also observed. Spectrum obtained from SWNTs sam-
ple (diameter of about 1.07 nm) mixed with KBr using a Elaser = 1.16 eV A1g , E1g , E2g [126]. The G band is also non-dispersive and
(λ = 1064.5 nm) excitation. This sample is produced by electric arc- its frequency is not sensitive to the excitation wavelength
discharge method followed by air oxidation at CSIC laboratory. [116].
114 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

Table 1
Symmetry modes and Raman frequencies for carbon nanotubes from different references [94,116]
Notations Positions (cm−1 )
A1g RBM A/d + B (see text) [116]
D D band ∼ 1340a,b ∼ 1350 [116]c
A1g G band
E1g G band 1585 [110]d 1500–1605 [116]c
E2g G band 1591 [110]d
E2g – 1620 [93]a –
2 × 1220 Second order mode 2440–2455 [93,107,140]a,b –
2×D G band 2500–2700 ∼ 2700 [116]c
E2g + D Combination 2925 [93]b –
2E2g Overtone ∼ 3250a –
a Excitation wavelength (742 nm).
b Excitation wavelength (514 nm).
c Values for SWNTs.
d Theoretical values.

SWNTs (diameters and metallic/semiconducting features) The relation describing this dependence is given, for semi-
+ −
can be characterized by studying the G band line shape al- conducting tubes, by ωG − ωG = 47.7/d 2 and for metallic
+ −
though the information obtained is less accurate than that tubes by ωG − ωG = 79.5/d where d is the tube diame-
2
from RBM. The Raman line shape differs between metal- ter in nm. For semiconducting tubes, Jorio et al. [112] gives
lic and semiconductor nanotubes thus allowing to distin- a value of 45.7 nm2 cm−1 instead of 47.7 nm2 cm−1 [116].
guish the two types. The line shape of G band is com- These relations and some data from authors are presented in
posed of up to 6 peaks but only the two most intense peaks Fig. 9.
are useful for analysis [112]. The lower and higher fre- The disorder-induced D band and G band show, for
− +
quency components (ωG and ωG ) are, respectively, asso- graphite or nanotubes, a strong dispersion as a function of
ciated with the atomic displacements along circumferen- laser excitation energy [116,128]. Indeed, graphite-like forms
tial direction and along the tube axis [116]. For semicon- present a strong linear scatter coefficient with energies of
−
ducting nanotubes, the profile of the two components ωG about 53 cm−1 eV−1 for D band and about 106 cm−1 eV−1
+
and ωG is narrow [127] and fits to a Lorentzian line shape for G band [112,116]. However, the various forms of carbon
[112,116]. However, for metallic tubes, the profile of the can be distinguished by the position and the linewidth of the
−
lower frequency component ωG is broad, asymmetric and D band [104]: amorphous carbon has a very broad linewidth
can be described by a Breit–Wigner–Fano (BWF) line shape (> 100 cm−1 ), D band position at 1285–1300 cm−1 [129]
+
[112,116,127]. There is no dependence of the frequency ωG and a linewidth of 10–30 cm−1 [105] are characteristic of
−
(about 1591 cm−1 ) on the nanotube diameters contrary to ωG . SWNTs, crystalline graphite-like forms have a typical posi-

+ −
Fig. 8. Phenomenological relations between RBM Raman mode Fig. 9. Dependence of G band frequency components ωG and ωG on
(ωRBM ) and the nanotube diameter as determined by several authors nanotube diameter for SWNTs. Data obtained from various authors
[54,116,117,119,141–143]. [116,127,144–148].
 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118 115

Indeed, these special nanotubes gives some additional and

interesting behaviors. The 0.4 nm CNTs are somewhat unsta-
ble because of their large curvature [18]. However, numer-
ous studies have been reported concerning the new properties
[18–23]. These single-walled nanotubes could be destroyed
at much lower temperature than larger ones [18]. Proper-
ties like strong anisotropic optical responses [20], large split
of the G-like Raman mode [18], superconductivity at rela-
tively high temperature (15 K) [23] are revealed by the 0.4 nm
CNTs.
Tang et al. [18] have summarized the various features of
Raman spectra for 0.4 nm CNTs: (i) the RBM modes were
identified at about 515, 548 and 584 cm−1 , respectively, for
the 0.4 nm-sized (5, 0), (4, 2) and (3, 3) tubes. The experi-
mental relation between RBM frequency and the diameter
of the CNTs could be applied for these nanotubes. How-
ever, the calculated values (ωRBM = 620 cm−1 for a 0.4 nm
Fig. 10. Fitted values of simulated RBM frequency for bundles and iso- tube) is higher than the experimental ones. Indeed, the rela-
lated nanotube as a function of the radius. The RBM of an isolated tube is tion is defined in the regime of elastic approximation. Other
splitted into two hybrid RBM (RBM1 and RBM2) due to the lowering of RBM frequencies are calculated using an ab initio method
the symmetry when nanotubes are packed in bundles (from Henrard et al.
[133]).
and are reported in Table 2. (ii) The disordered-induced D
band is observed at about 1350 cm−1 . The strong curvature
effect of the 0.4 nm nanotubes lead to a broad shape for the
tion of 1305–1330 cm−1 and a width of about 30–60 cm−1 D band. (iii) The graphite-like G band near 1600 cm−1 is
[130], and MWNTs are similar to crystalline graphite-like also defined in these nanotubes. The splitting of the G band
− +
forms [129]. into the ωG and ωG components can be large which leads
Effect of packing on Raman spectra has been reported −
to small ωG values. The experimental relation described in
by various authors [123,131,132]. However, Raman spec- the Raman spectroscopy section is generally used to fit the
− +
tra simulations of SWNT bundles were obtained by Hen- separation between ωG and ωG (about 1600 cm−1 ) for CNTs
rard et al. [133]. Finite and infinite bundles of SWNTs were with diameter larger than 0.9 nm. However, this relation is
studied in the frequency range of breathing mode (RBM) still verified with 0.4 nm tubes and the values at about 1300
(about 100–300 cm−1 ). For infinite and homogeneous bun- and 1100 cm−1 could be, respectively, assigned to the (4, 2)
dles, they predicted a second mode of breathing character for semiconducting tubes and to the metallic (5, 0) and (3, 3)
tubes with a radii greater than 0.7 nm. There is a strong ef- tubes.
fect of the intertube interaction on the RBM and, due to the The optical responses of 0.4 nm single-walled nanotubes
lowering of the symmetry, two hybrid modes (noted RBM1 have been measured and calculated by Liang et al. [20]. Com-
and RBM2) appears when SWNTs bundle. The frequency mon CNTs such as SWNTs (with diameter > 0.9 nm) and
of RBM1 mode relative to the RBM of isolated tubes has MWNTs have absorption spectra which are sensitive to the
an upward shift of 15 cm−1 within a bundle (Fig. 10). The diameters. However, absorption spectra are also sensitive the
larger the radii of the single-walled nanotubes, the lower this chiralities for small diameter nanotubes. Calculations done
frequency shift (about 0 for radii = 1.0 nm). For inhomoge- by Machon et al. [25] have also demonstrated the strong
neous bundles (bundles with a wide range of nanotube di- anisotropy of the optical responses. If the light is polarized
ameters), only global RBM mode is observed and the anal- perpendicular to the tube axis, the 0.4 nm tubes are transpar-
ysis of breathing modes is not going to be able to resolve ent in the energy range up to 4 eV [20].
individual tube features. However, evidences of individual
tubes are found for small size of bundles with very different
radii. Table 2
The RBM frequencies in cm−1 in (5, 0), (4, 2) and (3, 3) nanotubes with
diameter of about 0.4 nm
Nanotubes (5, 0) (4, 2) (3, 3)
4. Beyond the characterization techniques
Experimental RBM [18] 515 548 584
Ab initio method used by 527 531 541
4.1. Special behaviors of the very small diameter Tang et al. [18]
nanotubes Ab initio method used by 526 527 532
Agrawal et al. [24]
As previously described in this work, recently the research The experimental values are compared to those obtained using ab initio
have been focused on nanotubes with a very small diameter. methods.
116 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

4.2. Separation of semiconducting and metallic 5. Conclusions

nanotubes
Useful techniques for nanotubes characterization are re-
The separation of semiconducting and metallic nanotubes viewed. X-ray photoelectron spectroscopy (XPS) allows to
is required for many applications such as nanoscale electron- determine the functionalization of the nanotubes. Binding
ics. However the nanotubes are typically grown in bundles. energy and FWHM are modified in a significant way. New
Few papers were published in the past 3 years about the sep- peaks appears according to the bonds formed between car-
aration of the nanotubes [134–137]. bon atoms and compounds added. Scanning tunneling mi-
The selective destruction of metallic nanotubes could be croscopy (STM) is a powerful technique used to obtain three-
realized with bursts of electricity as shown by Collins et dimensional images and electronic states of the nanotubes.
al. [137]. However, this technique is only useful for tran- Nanotubes helicities could be determined with high resolu-
sistor geometries and can not be extended for bulk separa- tion images. The features of the nanotubes are explored at a
tion. The ropes containing both semiconducting and metal- larger scale using neutron diffraction. This diffraction tech-
lic tubes are deposited on a flat surface of silicon oxide. nique allows to investigate a large range of scattering vector
Then, electrodes on the top of ropes are fabricated by lithog- Q compared to X-ray diffraction. Thus, with a statistical ap-
raphy technique. A voltage is applied to the tubes through proach, XRD is used to obtain some information on the inter-
the electrodes. The metallic CNTs are destroyed by current- layer spacing, the structural strain and the impurities. Numer-
induced oxidation and only the semiconducting tubes remain ous features such as intershell spacing, chiral indices, and he-
[138]. licity are also obtained by transmission electron microscopy
The bulk separation of various CNT types could be (TEM). Infrared spectroscopy is often used to determine im-
done using one of the following ways: (i) Krupke et al. purities remaining from synthesis or molecules capped on the
[134] have demonstrated a new method to extract the metal- nanotube surface. But the most powerful technique to char-
lic tubes in solutions based on alternating current dielec- acterize nanotubes remains the Raman spectroscopy: without
trophoresis. When submitted to an electric field, the SWNTs sample preparation, a fast and non-destructive analysis is pos-
have an induced dipole moment which is used to selec- sible. Moreover, Raman spectra simulations are performed
tively move these tubes. As well as for photoluminescence on various nanotubes geometries. For a correct characteri-
spectroscopy, the bundles must be separated in individ- zation of nanotubes, all these techniques described here can
ual tubes using an ultrasonication treatment of the nan- not be used separately but must be used in complementary
otubes with surfactants such as the SDS (sodium dode- ways.
cyl sulfate). The dielectrophoresis is then applied to this
solution: the metallic tubes are deposited onto the elec-
trodes while the semiconducting ones remain in suspen- Acknowledgments
sion. (ii) CNT types are also separated using a novel pre-
cipitation route as described by Chattopadhyay et al. [136]. The authors would like to thanks the European program for
The SWNTs with octadecylamine (ODA) in tetrahydrofu- Advanced Hydrogen Storage Material (HYMOSSES-ENK6-
ran (THF) formed a stable suspension. The authors assume CT-2002-00653). We would like to thank Dr. Ana Benito,
that the stability is done by the physisorption and the or- Dr. Wolfgang Maser, Dr. Manuel Perez and Cristina Vallés
ganization of the ODA molecules along the SWNT side- of CSIC-ICB (Zaragoza, Spain) for the supply of raw and
walls. Whereas the metallic tubes properties are insensi- treated carbon nanotube materials. Thanks also to Dr. Y. Ba-
tive, the electrical properties of semiconducting nanotubes tonneau and Dr. I. Fechete for helpful discussions. The au-
are modified by the adsorption of alkylamines [139]. Thus, thors acknowledge E. Gautron, Dr. D. Mesnard for their help
the ODA physisorption on the SWNTs might improve the in TEM and Raman experiments.
stability of the semiconducting tubes in the suspension in-
stead of the metallic ones. In this case, there is a separa-
tion between ODA-stabilized semiconducting and metallic
tubes. (iii) An original way of nanotubes separation is the References
DNA-assisted method used by Zheng et al. [135]. The nu-
[1] S. Iijima, Nature 354 (1991) 56.
cleic acid polymers forms an hybrid material with CNTs.
[2] M. Dresselhaus, G. Dresselhaus, R. Saito, Carbon 33 (1995) 883.
The phosphate groups (from DNA) on a DNA–CNT hybrid [3] J. Mintmire, C. White, Carbon 33 (1995) 893.
provide a negative surface charge density which is depen- [4] P. Ajayan, T. Ebbesen, Rep. Prog. Phys. 60 (1997) 1025.
dent of the DNA sequence and the electronic property of the [5] R. Ruoff, D. Lorents, Carbon 33 (1995) 925.
tube. On the contrary, the metallic tubes in DNA–CNT hy- [6] K. Murata, K. Kaneko, F. Kokai, K. Takahashi, M. Yudasaka, S. Iijima,
Chem. Phys. Lett. 331 (2000) 14.
brid contributes to a weak surface charge density. Finally, the
[7] A. Dillon, K. Jones, T. Bekkedahl, C. Kiang, D.S. Bethune, M.J.
ion-exchange liquid chromatography with DNA–CNT hybrid Heben, Nature 386 (1997) 377.
suspensions is used by the authors in order to separate the [8] J. Wildoer, L. Venema, A. Rinzler, R. Smalley, C. Dekker, Nature 391
nanotubes. (1998) 59.
 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118 117

[9] T. Odom, J. Huang, P. Kim, C. Lieber, Nature 391 (1998) 62. [42] J. Guo, C. Yang, Z.M. Li, M. Bai, H.J. Liu, G.D. Li, E.G. Wang,
[10] DOE FreedomCAR program, http://www.eere.energy.gov/vehiclesa- C.T. Chan, Z.K. Tang, W.K. Ge, X. Xiao, Phys. Rev. Lett. 93 (2004)
ndfuels. 017402-1.
[11] E. Tzimas, C. Filiou, S.D. Peteves, J.B. Veyret, EUR 20995 EN- [43] Y. Miyauchi, S. Chiashi, Y. Murakami, Y. Hayashida, S. Maruyama,
hydrogen storage: state of the art and future perspective, European Chem. Phys. Lett. 387 (2004) 198.
Communities, 2003. ISBN92-894-6950-1. [44] R. Droppa, P. Hammer, A. Carvalho, M. Dos Santos, F. Alvarez, J.
[12] H. Gommans, J. Alldredge, H. Tashiro, J. Park, J. Magnuson, A.G. Non-Cryst. Solids 299–302 (2002) 874.
Rinzler, J. Appl. Phys. 88 (2000) 2509. [45] P. Hammer, N. Victoria, F. Alvarez, J. Vac. Sci. Technol. 18 (2000)
[13] E. Thostenson, Z. Ren, T. Chou, Comp. Sci. Technol. 61 (2001) 2277.
1899. [46] C. Pham-Huu, N. Keller, V. Roddatis, G. Mestl, R. Schlogl, M.J.
[14] S. Amelinckx, A. Lucas, P. Lambin, Rep. Prog. Phys. 62 (1999) 1471. Ledoux, Phys. Chem. Chem. Phys. 4 (2002) 514.
[15] Y. Murakami, Y. Miyauchi, S. Chiashi, S. Maruyama, Chem. Phys. [47] Y. Lee, T. Cho, B. Lee, J. Rho, K. An, Y. Lee, J. Fluorine Chem. 120
Lett. 377 (2003) 49. (2003) 99.
[16] R. Saito, G. Dresselhaus, M. Dresselhaus, Phys. Rev. B 6 (2000) 2981. [48] P. Chen, X. Wu, X. Sun, J. Lin, W. Ji, K. Tan, Phys. Rev. Lett. 82
[17] A. Jorio, R. Saito, J. Hafner, C. Lieber, M. Hunter, T. McClure, G. (1999) 2548.
Dresselhaus, M. Dresselhaus, Phys. Rev. Lett. 86 (2001) 1118. [49] S. Suzuki, Y. Watanabe, T. Ogino, S. Heun, L. Gregoratti, A. Barinov,
[18] Z.K. Tang, N. Wang, X.X. Zhang, J.N. Wang, C.T. Chan, P. Sheng, B. Kaulich, M. Kiskinova, W. Zhu, C. Bower, O. Zhou, Phys. Rev. B
New J. Phys. 5 (2003) 146.1. 66 (2002) 035414.
[19] L.C. Qin, X. Zhao, K. Hirahara, Y. Miyamoto, Y. Ando, S. Iijima, [50] K. Sattler, Carbon 33 (1995) 915.
Nature 298 (2000) 50. [51] K. Ichimura, M. Osawa, K. Nomura, H. Kataura, Y. Maniwa, S. Suziki,
[20] W.Z. Liang, G.H. Chen, Z. Li, Z.K. Tang, Appl. Phys. Lett. 80 (2002) Y. Achiba, Phys. B: Cond. Matter 323 (2002) 230.
3415. [52] N. Lin, J. Ding, S. Yang, N. Cue, Carbon 34 (1996) 1295.
[21] N. Sano, M. Chhowalla, D. Roy, G.A.J. Amaratunga, Phys. Rev. B 66 [53] A. Hassanien, M. Tokumoto, Carbon 42 (2004) 2649.
(2002) 113403. [54] E. Obraztsova, V. Yurov, V. Shevluga, R. Baranovsky, V. Nalimova,
[22] X. Zhao, Y. Liu, S. Inoue, T. Suzuki, R.O. Jones, Y. Ando, Phys. Rev. V. Kuznetsov, V. Zaikovskii, Nanostruct. Mater. 11 (1999) 295.
Lett. 92 (2004) 125502-1. [55] P. Kim, T. Odom, J. Huang, C. Lieber, Carbon 38 (2000) 1741.
[23] Z.K. Tang, L.Y. Zhang, N. Wang, X.X. Zhang, J.N. Wang, G.D. Li, [56] W. Clauss, D. Bergeron, A. Johnson, Phys. Rev. B 58 (1998) R4266.
Z.M. Li, G.H. Wen, C.T. Chan, P. Sheng, Synth. Met. 133–134 (2003) [57] N. Hamada, S. Sawada, A. Oshiyama, Phys. Rev. Lett. 68 (1992) 1579.
689. [58] J. Mintmire, B. Dunlap, C. White, Phys. Rev. Lett. 68 (1992) 631.
[24] B.K. Agrawal, S. Agrawal, R. Srivastava, J. Phys.: Condens. Matter [59] R. Saito, M. Fujita, G. Dresselhaus, M. Dresselhaus, Appl. Phys. Lett.
16 (2004) 1467. 60 (1992) 2204.
[25] M. Machon, S. Reich, C. Thomsen, D. Sanchez-Portal, P. Ordejon, [60] A. Krasheninnikov, Solid State Commun. 118 (2001) 361.
Phys. Rev. B 66 (2002) 155410. [61] A. Burian, J. Koloczk, J. Dore, A.C. Hannon, J.B. Nagy, A. Fonseca,
[26] X.F. Zhang, X.B. Zhang, G. Van Tendeloo, S. Amelinckx, M. Op de Diam. Relat. Mater. 13 (2004) 1261.
Beeck, J. Van Landuyt, J. Cryst. Growth 130 (1993) 368. [62] P. Lambin, A. Loiseau, C. Culot, L. Biro, Carbon 40 (2002) 1635.
[27] Y. Saito, T. Yoshikawa, S. Bandow, M. Tomita, T. Hayashi, Phys. Rev. [63] J. Dore, M. Sliwinski, A. Burian, W. Howells, D. Cazorla, J. Phys.:
B 48 (1993) 1907. Condens. Matter 11 (1999) 9189.
[28] M. Bretz, B. Demczyk, L. Zhang, J. Cryst. Growth 141 (1994) 304. [64] A. Giannasi, M. Celli, J. Sauvajol, M. Zoppi, D. Bowron, Physica B
[29] X. Sun, C. Kiang, M. Endo, K. Takeuchi, T. Furuta, M. Dresselhaus, 350 (2004) e1027.
Phys. Rev. B 54 (1996) R12629. [65] S. Rols, R. Almairac, L. Henrard, E. Anglaret, J. Sauvajol, Eur. Phys.
[30] C. Kiang, M. Endo, P. Ajayan, G. Dresselhaus, M. Dresselhaus, Phys. J. B 10 (1999) 263.
Rev. Lett. 81 (1998) 1869. [66] J. Koloczek, Y. Kwon, A. Burian, J. Alloys Compd. 328 (2001) 222.
[31] A. Peigney, C. Laurent, E. Flahaut, R. Bacsa, A. Rousset, Carbon 39 [67] W. Zhu, D. Miser, W. Chan, M. Hajaligol, Mater. Chem. Phys. 82
(2001) 507. (2003) 638.
[32] H. Hu, P. Bhowmik, B. Zhao, M.A. Hamon, M.E. Itkis, R.C. Haddon, [68] D. Reznik, C. Olk, A. Neumann, J. Copley, Phys. Rev. B 52 (1995)
Chem. Phys. Lett. 345 (2001) 25. 116.
[33] D.B. Mawhinney, V. Naumenko, A. Kuznetsova, J.T. Yates Jr., J. Liu, [69] A. Burian, J. Dore, H. Fischer, J. Sloan, Phys. Rev. B 59 (1999) 1665.
R.E. Smalley, Chem. Phys. Lett. 324 (2000) 213. [70] A. Cao, C. Xu, J. Liang, D. Wu, B. Wei, Chem. Phys. Lett. 344 (2001)
[34] J. Chen, A.M. Rao, S. Lyuksyutov, M.E. Itkis, M.A. Hamon, H. Hu, 13.
R.W. Cohn, P.C. Eklund, D.T. Colbert, R.E. Smalley, R.C. Haddon, [71] M. Liu, J. Cowley, Carbon 32 (1994) 393.
J. Phys. Chem. B 105 (2001) 2525. [72] D. Bernaerts, S. Amelinckx, P. Lambin, A. Lucas, Appl. Phys. A 67
[35] M. Ouyang, J.L. Huang, C.L. Cheung, C.M. Lieber, Science 292 (1998) 53.
(2001) 702. [73] W. Cochran, F. Crick, V. Vand, Acta Cryst. 5 (1952) 581.
[36] J.S. Lauret, C. Voisin, G. Cassabois, P. Roussignol, C. Delalande, A. [74] P. Lambin, V. Meunier, L. Henrard, A. Lucas, Carbon 38 (2000) 1713.
Filoramo, L. Capes, E. Valentin, O. Jost, Physica E 21 (2004) 1057. [75] A. Lucas, V. Bruyninckx, P. Lambin, Europhys. Lett. 35 (1996) 555.
[37] X. Liu, T. Pichler, M. Knupfer, M.S. Golden, J. Fink, H. Kataura, Y. [76] J. Cowley, F. Sundell, Ultramicroscopy 68 (1997) 1.
Achiba, Phys. Rev. B 66 (2002) 045411. [77] H. Kuzmany, W. Plank, M. Hulman, Ch. Kramberger, A. Gruneis, Th.
[38] J. Lefebvre, J.M. Fraser, Y. Homma, P. Finnie, Appl. Phys. A 78 (2004) Pichler, H. Peterlik, H. Kataura, Y. Achiba, Eur. Phys. J. B 22 (2001)
1107. 307.
[39] R.B. Weisman, S.M. Bachilo, D. Tsyboulski, Appl. Phys. A 78 (2004) [78] C. Gommes, S. Blacher, K. Masenelli-Varlot, C. Bossuot, E. McRae,
1111. A. Fonseca, J.B. Nagy, J.P. Pirard, Carbon 41 (2003) 2561.
[40] M.J. O’Connell, S.M. Bachilo, C.B. Huffman, V.C. Moore, M.S. [79] A. Charlier, E. McRae, R. Heyd, M. Charlier, D. Moretti, Carbon 37
Strano, E.H. Haroz, K.L. Rialon, P.J. Boul, W.H. Noon, C. Kittrell, J. (1999) 1779.
Ma, R.H. Hauge, R.B. Weisman, R.E. Smalley, Science 297 (2003) [80] K. Tanaka, H. Aoki, H. Ago, T. Yamabe, K. Okahara, Carbon 35
503. (1997) 125.
[41] S. Lebedkin, F. Hennrich, T. Skipa, M.M. Kappes, J. Phys. Chem. B [81] M.S. Dresselhaus, M. Endo, Top. Appl. Phys. 80 (2001) 11.
107 (2003) 1949. [82] L. Qin, Chem. Phys. Lett. 297 (1998) 23.
 118 T. Belin, F. Epron / Materials Science and Engineering B 119 (2005) 105–118

[83] L. Qin, T. Ichihashi, S. Iijima, Ultramicroscopy 67 (1997) 181. [117] S. Bandow, S. Asaka, Y. Saito, A. Rao, L. Grigorian, E. Richter, P.
[84] L. Qin, S. Iijima, H. Kataura, Y. Maniwa, S. Suzuki, Y. Achiba, Chem. Eklund, Phys. Rev. Lett. 80 (1998) 3779.
Phys. Lett. 268 (1997) 101. [118] M. Milnera, J. Kurti, M. Hulman, H. Kuzmany, Phys. Rev. Lett. 84
[85] M. Kociak, K. Hirahara, K. Suenaga, S. Iijima, Eur. Phys. J. B 32 (2000) 1324.
(2003) 457. [119] L. Alvarez, A. Righi, T. Guillard, S. Rols, E. Anglaret, D. Laplaze, J.
[86] R. He, H. Jin, J. Zhu, Y. Yan, X. Chen, Chem. Phys. Lett. 298 (1998) Sauvajol, Chem. Phys. Lett. 316 (2000) 186.
170. [120] S. Bandow, G. Chen, G. Sumanasekera, R. Gupta, M. Yudasaka, S.
[87] J. Cowley, P. Nikolaev, A. Thess, R. Smalley, Chem. Phys. Lett. 265 Iijima, P. Eklund, Phys. Rev. B 66 (2002) 075416.
(1997) 379. [121] A. Rao, R. Richter, S. Bandow, B. Chase, P. Eklund, K.A. Williams,
[88] D. Bernaerts, A. Zettl, N. Chopra, A. Thess, R. Smalley, Solid State S. Fang, K.R. Subbaswamy, M. Menon, A. Thess, R. Smalley, G.
Commun. 105 (1998) 145. Dresselhaus, M. Dresselhaus, Science 275 (1997) 187.
[89] J. Colomer, L. Henrard, P. Lambin, G. Van Tendeloo, Eur. Phys. J. B [122] G. Mahan, Phys. Rev. B 65 (2002) 235402.
27 (2002) 111. [123] U. Venkateswaran, A. Rao, E. Richter, M. Menon, A. Rinzler, R.
[90] L. Henrard, A. Loiseau, C. Journet, P. Bernier, Eur. Phys. J. B 13 Smalley, P. Eklund, Phys. Rev. B 59 (1999) 10928.
(2000) 661. [124] A. Jorio, A. Souza-Filho, V. Brar, A. Swan, M. Unlu, B.B. Gold-
[91] U. Kuhlmann, H. Jantoljak, M. Pfander, P. Bernier, C. Journet, C. berg, A. Righi, J.H. Hafner, C. Lieber, R. Saito, G. Dresselhaus, M.
Thomsen, Chem. Phys. Lett. 294 (1998) 237. Dresselhaus, Phys. Rev. B 65 (2002) R121402.
[92] H. Kuzmany, B. Burger, A. Thess, R. Smalley, Carbon 36 (1998) 709. [125] M. Dresselhaus, P. Eklund, Adv. Phys. 49 (2000) 705.
[93] J. Kastner, T. Pichler, H. Kuzmany, S. Curran, W. Blau, D.N. Weldon, [126] J. Maultzsch, S. Reich, C. Thomsen, E. Dobardzic, I. Milosevic, M.
M. Delamesiere, S. Draper, H. Zandbergen, Chem. Phys. Lett. 221 Damnjanovic, Solid State Commun. 121 (2002) 471.
(1994) 53. [127] J. Colomer, J. Benoit, C. Stephan, S. Lefrant, G. Van Tendeloo, J.B.
[94] P. Eklund, J. Holden, R. Jishi, Carbon 33 (1995) 959. Nagy, Chem. Phys. Lett. 345 (2001) 11.
[95] R. Saito, K. Matsushige, K. Tanaka, Phys. B: Condens. Matter 323 [128] R. Vidano, D. Fischbach, J. Willis, T. Loehr, Solid State Commun. 39
(2002) 280. (1981) 341.
[96] B. He, W. Sun, M. Wang, S. Liu, Z. Shen, Mater. Chem. Phys. 84 [129] S. Brown, A. Jorio, M. Dresselhaus, G. Dresselhaus, Phys. Rev. B 64
(2004) 140. (2001) 073403.
[97] Z. Yang, H. Chen, L. Cao, H. Li, M. Wang, Mater. Sci. Eng. B 106 [130] Y. Wang, D. Alsmeyer, R. McCreery, Chem. Mater. 2 (1990)
(2004) 73. 557.
[98] D. Jung, Y.K. Ko, H. Jung, Mater. Sci. Eng. C 24 (2004) 117. [131] L. Henrard, E. Hernandez, P. Bernier, A. Rubio, Phys. Rev. B 60
[99] M. Xu, Q. Huang, Q. Chen, P. Guo, Z. Sun, Chem. Phys. Lett. 375 (1999) R8521.
(2003) 598. [132] D. Kahn, J. Lu, Phys. Rev. B 60 (1999) 6535.
[100] L. Liu, Y. Qin, Z. Guo, D. Zhu, Carbon 41 (2003) 331. [133] L. Henrard, V. Popov, A. Rubio, Phys. Rev. B 64 (2001) 205403.
[101] M. Aizawa, M. Shaffer, Chem. Phys. Lett. 368 (2003) 121. [134] R. Krupke, F. Hennrich, H. Löhneysen, M.M. Kappes, Science 301
[102] S. Wang, W. Zhu, D. Liao, C. Ng, C. Au, Catal. Today 93–95 (2004) (2003) 344.
711. [135] M. Zheng, A. Jagota, E.D. Semke, B.A. Diner, R.S. McClean, S.R.
[103] M. Pereira, J. Figueiredo, J. Orfao, P. Serp, P. Kalck, Y. Kihn, Carbon Lustig, R.E. Richardson, N.G. Tassi, Nat. Mater. 2 (2003) 338.
42 (2004) 2807. [136] D. Chattopadhyay, I. Galeska, F. Papadimitrakopoulos, J. Am. Chem.
[104] S. Arepalli, P. Nikolaev, O. Gorelik, V. Hadjiev, W. Holmes, B. Files, Soc. 125 (2003) 3370.
L. Yowell, Carbon 42 (2004) 1783. [137] P.G. Collins, M.S. Arnold, P. Avouris, Science 292 (2001) 706.
[105] A. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 14095. [138] Carbon nanotubes transistors. http://www.research.ibm.com.
[106] A.A. Mamedov, N.A. Kotov, M. Prato, D.M. Guldi, J.P. Wicksted, A. [139] J. Kong, H. Dai, J. Phys. Chem. B 105 (2001) 2890.
Hirsch, Nat. Mater. 1 (2002) 190. [140] W. Bacsa, W. De Heer, Phys. Rev. B 50 (1994) 15473.
[107] H. Hiura, T. Ebbesen, K. Tanigaki, H. Takahashi, Chem. Phys. Lett. [141] K. Williams, M. Tachibana, J. Allen, L. Grigorian, S.C. Cheng, S.L.
202 (1993) 509. Fang, G.U. Sumanasekera, A. Loper, J. Williams, P. Ekund, Chem.
[108] R.R. Bacsa, A. Peigney, C. Laurent, P. Puech, W.S. Bacsa, Phys. Rev. Phys. Lett. 310 (1999) 31.
B 65 (2002) 161404. [142] A. Rao, J. Chen, E. Richter, U. Schlecht, P. Eklund, R.C. Haddon, U.D.
[109] M. Damnjanovic, Phys. Lett. A 94 (1983) 337. Venkateswaran, Y. Kown, D. Tomanek, Phys. Rev. Lett. 86 (2001)
[110] P. Liu, L. Liu, Y. Zhang, Phys. Lett. A 313 (2003) 302. 3895.
[111] R. Saito, T. Takeya, T. Kimura, G. Dresselhaus, M. Dresselhaus, Phys. [143] M. Hulman, R. Pfeiffer, H. Kuzmany, New J. Phys. 6 (2004) 1.
Rev. B 57 (1998) 4145. [144] P. Bernier, W. Maser, C. Journet, A. Loiseau, M.L. De la Chapelle, R.
[112] A. Jorio, M. Pimenta, A. Souza-Filho, R. Saito, G. Dresselhaus, M. Lee, J.F. Fischer, Carbon 36 (1998) 675.
Dresselhaus, New J. Phys. 5 (2003) 139.1. [145] R. Bacsa, C. Laurent, A. Peigney, W. Bacsa, Th. Vaugien, A. Rousset,
[113] R. Jishi, L. Venkataraman, M. Dresselhaus, G. Dresselhaus, Chem. Chem. Phys. Lett. 323 (2000) 566.
Phys. Lett. 209 (1993) 77. [146] Y. Zhang, Z. Shi, Z. Gu, S. Iijima, Carbon 38 (2000) 2055.
[114] J. Kurti, G. Kresse, H. Kuzmany, Phys. Rev. B 58 (1998) R8869. [147] P. Corio, P. Santos, V. Brar, G. Samsonidze, S. Chou, M. Dresselhaus,
[115] D. Sanchez-Portal, E. Artacho, J. Soler, A. Rubio, P. Ordejon, Phys. Chem. Phys. Lett. 370 (2003) 675.
Rev. B 59 (1999) 12768. [148] Y. Park, K. Kim, H. Jeong, W. Kim, J. Moon, K.H. An, D.J. Bae, Y.
[116] M. Dresselhaus, G. Dresselhaus, A. Jorio, A.G. Souza-Filho, R. Saito, Lee, G. Park, Y. Lee, Synth. Met. 126 (2002) 245.
Carbon 40 (2002) 2043.

View publication stats