INSTITUTO POTOSINO DE INVESTIGACIÓN

CIENTÍFICA Y TECNOLÓGICA, A.C.

POSGRADO EN CIENCIAS APLICADAS

“Synthesis “Títuloand characterization

de la tesis”
of doped
single-walled carbon nanotubes and graphitic
(Tratar de hacerlo comprensible para el público general, sin abreviaturas)
nanoribbons”
Tesis que presenta
Jessica Rosaura Campos Delgado

Para obtener el grado de

Doctora en Ciencias Aplicadas

En la opción de
Nanociencias y Nanotecnología

Codirectores de la Tesis:
Dr. Humberto Terrones Maldonado
Dr. Mauricio Terrones Maldonado

San Luis Potosí, S.L.P., 17 de Diciembre 2009

ii
 Créditos Institucionales

Esta tesis fue elaborada en las instalaciones y con la infraestructura de la División

de Materiales Avanzados para la Tecnología Moderna del Instituto Potosino de
Investigación Científica y Tecnológica, A.C. y el Laboratorio de Investigaciones en
Nanociencia y Nanotecnología, bajo la co-dirección de los doctores
Humberto Terrones Maldonado y Mauricio Terrones Maldonado.

Durante la realización del trabajo el autor recibió una beca académica del
Consejo Nacional de Ciencia y Tecnología (202248).

iii
 Acknowledgments

Sincere thanks to Prof. Mauricio and Prof. Humberto for their support, for believing
in my capacity, and for pushing my limits.

Thanks to Millie and Gene for their hospitality at MIT and for spreading their
motivation to do flag-less science.
I would like to acknowledge the financial support of the SOMENANO during my
stay in Boston and to the SNI for a researcher assistant scholarship (2006-2008).

Thanks to David Smith for hosting me at ASU, and to Ado Jorio and Marcos
Pimenta for welcoming me at UFMG.

I am grateful to the colleagues that I met during these years, for sharing with me
their knowledge and good laughs inside and outside the labs. In particular to Dave
Cullen, Ya-Ping, Xiaoting, Mario, Federico, Alfonso, Hootan, and Indhira.
To Andrés and Lau for their friendship.
Special thanks to Daniel Ramírez, Grisel Ramírez, Gaby Pérez, Karla Gómez,
Aaron, Ana Laura, Lalo Cruz, Dr. Fernando, Dr. Flo, Dra. Yadira, Dr. Emilio and
technicians working at LINAN.

Thanks to my father and mother for their unconditional support and understanding.
To my little sister who lightened my days in spite of being far, far away (de frontera
a frontera).
To my brother Daniel and his amazing family: Nancy, Ale and Carlitos.
To my brother David, my mom’s family and my dad’s family.

I am deeply grateful to Omar whose understanding, ever-lasting love and support

keep me going.

v
 Contents

Constancia de aprobación de la tesis ii

Créditos institucionales iii
Acta de examen iv
Acknowledgements v
List of tables viii
List of figures ix
Appendix xv
Abreviations xvi
Abstract xvii
Resumen xviii

1. Introduction 1
1.1 Carbon nanotubes 2
1.1.1 Structure 2
1.1.2 Electronic properties 5
1.1.3 Doped carbon nanotubes 7
1.2 Graphene and carbon nanoribbons 13
1.3 References 19

2. Production of doped single-wall carbon nanotubes using the aerosol 29

pyrolysis technique
2.1 Description of the synthesis method, materials and equipment 29
2.2 CVD-production of doped single-wall carbon nanotubes 31
2.3 Characterization of CVD-produced doped SWNTs 35
2.3.1 Electron microscopy characterization 36
2.3.2 Raman spectroscopy characterization 47
2.4 Conclusions 60
2.5 Related Articles 62
2.6 References 63

vi
3. Production of graphitic nanoribbons using the aerosol pyrolysis 67
technique
3.1 Description of the synthesis method 68
3.2 Characterization of the pristine material 69
3.3 Annealing treatments of carbon nanoribbons 74
3.3.1 Characterization of the annealed materials 76
3.3.2 Raman spectroscopy study of the annealed materials 82
3.4 Conclusions 93
3.5 Related Articles 94
3.6 References 95

4. In situ reconstruction of graphitic nanoribbons by Joule heating 99

4.1 Conditions of the experiment 107
4.2 Crystallization and sharp edge formation 108
4.3 Loop formation at the edges 116
4.4 Estimation of the achieved temperatures 118
4.5 Conclusions 120
4.6 Related Articles 122
4.7 References 123

5. Conclusions and Perspectives 125

5.1 Future work on doped-SWNTS 128
5.2 Future work on carbon nanoribbons 129

Appendix
Appendix A Carbon nanotubes synthesis methods and characterization 131
tools
Appendix B Raman spectroscopy of sp2 carbon materials 145

vii
 List of Tables

2.1 Precursor compounds used in the doping experiments of 33

SWNTs

2.2 Experimental conditions used in the synthesis of doped SWNTs 34

viii
 List of figures

1.1 Allotropes of carbon 1

1.2 Symmetry classification of nanotubes 3

1.3 Geometry and unit cell of carbon nanotubes 3

1.4 Electronic density of states of two zigzag nanotubes 6

1.5 Kataura plot 6

1.6 Types of doping in carbon nanostructures 7

1.7 Nitrogen doped MWNTs 9

1.8 Phosphorus doping of SWNTs 10

1.9 Silicon doping of SWNTs 11

1.10 Sulfur doping of SWNTs 12

1.11 Structure of armchair and zigzag graphene nanoribbons 13

1.12 Stability of graphene nanoribbons 14

1.13 Methods of production of graphene 16

1.14 CVD and other production methods of carbon nanoribbons 17

2.1 Diagram of the experimental setting for CVD production of 30

SWNTs

2.2 TEM images of CVD-produced pristine SWNTs 30

2.3 SEM and STEM images of materials synthesized with thiophene 36

ix
2.4 TEM images of materials synthesized with thiophene 37

2.5 High resolution EDX point measurements on by-products of 38

thiophene-produced samples

2.6 SEM images of nitrogen-doped SWNTs produced with 38

benzylamine

2.7 SEM and TEM images of nitrogen-doped SWNTs produced with 39

pyrazine

2.8 SEM and STEM images of samples synthesized with 40

triphenylphosphine

2.9 SEM and TEM images of phosphorous-doped SWNTs 41

2.10 HRTEM micrographs of SWNTs produced with different 43

concentrations of MTMS

2.11 HAADF-STEM and EELS analysis of Si-doped SWNTs 44

produced with 0.1 wt. % of MTMS

2.12 HAADF-STEM and EELS analysis of Si-doped SWNTs 45

produced with 0.2 wt. % of MTMS

2.13 Raman spectra of material synthesized with 0.1 wt. % of 48

thiophene

2.14 RBM spectra of materials synthesized with thiophene 49

2.15 Raman spectra of materials synthesized with thiophene 50

2.16 RBM spectra of the samples doped with N, P, and Si at five 51

laser excitation energies

2.17 RBM of N- and Si- doped SWNTs recorded at Elaser=1.95 eV 52

2.18 RBM resonance Raman maps of P-doped SWNT samples 54

x
2.19 RBM and G’ band spectra of N-,P-, and Si- doped SWNTs at 55
Elaser=2.41 eV

2.20 IG'Def /IG'Pris ratios as a function of doping atoms (P, N and Si) per 56

carbon atoms

2.21 G’ band and the corresponding Lorentzian fitting of the samples 57

with maximum IG'Def /IG'Pris ratios

2.22 G’ band splitting of the N-, P- and Si- doped SWNTs 59

3.1 Schematic diagram of the experimental set-up for the production 68

of graphitic nanoribbons

3.2 SEM images and EDX line scan of graphitic nanoribbons 70

3.3 TEM images of graphitic nanoribbons 71

3.4 XRD, Raman spectroscopy, TGA, N2 absorption and XPS of 73

graphitic nanoribbons

3.5 Loop formation at the edges of graphene planes induced by heat 74

treatments in different carbon materials

3.6 Raman spectroscopy of heat treated carbon materials 75

3.7 SEM images of heat treated graphitic nanoribbons 76

3.8 TEM images of the side-edges of heat treated graphitic 78

nanoribbons

3.9 XRD and TGA of heat treated graphitic nanoribbons 79

3.10 Schematic model for the restructuring process achieved during 81

annealing

xi
3.11 Bulk Raman spectra of heat treated nanoribbons at seven laser 83
excitation energies

3.12 Example of the Lorentzian fiitting of Raman spectra 84

3.13 Band positions and ID/IG ratios as a function of Elaser 85

3.14 FWHM of the G band and IG’/IG ratios of HT nanoribbons 87

3.15 SEM and AFM imaging of isolated nanoribbons 89

3.17 Raman spectroscopy line scans of individual nanoribbons at 91

Elaser=2.33 eV

3.18 D and D’ band intensities vs. distance in the individual 92

measurements

4.1 Wall-by-wall breakdown of MWNTs under Joule heating 99

4.2 In situ observation of wall-by-wall breakdown 100

4.3 Superplasticity of carbon nanotubes 101

4.4 Crystallization of a disordered nanotube by Joule heating 103

4.5 Vacancy migration in carbon nanotubes 104

4.6 Schematic diagram of the TEM-STM system used during Joule 107
heating experiments

4.7 IV curve during Joule heating of graphitic nanoribbons 109

4.8 TEM images at different stages of the Joule heating experiment 110

4.9 TEM images of the Joule heating experiment at an applied bias 111
voltage of 1.6 V

xii
4.10 Crystallization and edge formation in graphitic nanoribbons by 112
Joule heating

4.11 Edge motion under Joule heating 113

4.12 Transport on graphene nanoribbons with different edge 114

configurations

4.13 Edge arrays and their time evolution 115

4.14 TEM images of the edges of graphitic nanoribbons annealed in 117

furnace

4.15 TEM images showing loop formation by Joule heating 118

4.16 Sequence of TEM images showing Pt nanoparticles on the 119

nanoribbon surface

A1 Diagram of an arc-discharge chamber 131

A2 TEM images of arc-discharge produced SWNTs 132

A3 Diagram of a laser ablation set-up 133

A4 SEM and TEM images of SWNTs produced by laser ablation 133

A5 CVD-derived SWNTs 134

A6 X-ray diffraction profiles of graphite and SWNTs 138

A7 C1s XPS signal of MWNTs 139

A8 EELS spectra of carbon allotropes 140

A9 Representative Raman spectra of HOPG, SWNTs and graphene 141

xiii
B1 Phonon dispersion of graphene 146

B2 Resonance Raman spectral processes 148

B3 Schematic of the RBM and tangential nanotube vibrations 150

B4 D- and D’- bands double resonance processes 153

B5 Electronic bands of 2D graphite and phonon dispersion curves of 154

2D graphite

B6 G’ band features of CVD-grown SWNTs before and after 157

annealing treatments

B7 G’ band of sp2 carbon materials 158

xiv
 Appendix

Appendix A 130
Carbon nanotubes synthesis methods and characterization tools

Appendix B 143
Raman spectroscopy of sp2 carbon materials

xv
 Abbreviations

Å Angstroms (10-10 m)
Cp Cyclopentadienyl (C5H5)
CVD Chemical vapor deposition
EDX Energy dispersive X-ray analysis
EELS Electron energy loss spectroscopy
e-beam Electron beam
eV Electron volts
HOPG Highly oriented pyrolytic graphite
HRTEM High resolution transmission electron microscopy
HT Heat treatment or heat treated
K Kelvin degrees
MTMS Methoxytrimethylsilane
MWNT Multi-wall carbon nanotube
nm Nanometers (10-9 m)
RBM Radial breathing mode
SEM Scanning electron microscopy
STEM Scanning transmission electron microscopy
SWNT Single-walled carbon nanotube
TEM Transmission electron microscopy
TGA Thermo gravimetric analysis
TPP Triphenylphosphine
µm Micrometers (10-6 m)
XPS X-ray photoelectron spectroscopy
XRD X-ray diffraction

xvi
Synthesis and characterization of doped single-walled carbon
nanotubes and graphitic nanoribbons

Abstract
An aerosol-assisted chemical vapor deposition technique was used to produce
doped SWNTs and graphitic nanoribbons.
The doping elements were introduced to the system by means of precursor
compounds, added to ferrocene (Fe(C5H5)2) – ethanol (C2H6O) solutions at
different concentrations. Thiophene (C4H4S), triphenylphosphine ((C6H5)3P),
benzylamine (C7H7NH2), pyrazine (C4H4N2) and methoxytrimethylsilane
(CH3OSi(CH3)3) were the precursors used in this work. The materials thus
produced were analyzed with SEM, TEM, EDX, EELS and Raman spectroscopy.
Our experiments with low concentrations of thiophene triggered the synthesis of a
new carbon nanostructure: graphitic nanoribbons. Such pristine nanoribbons have
been carefully characterized by SEM, TEM, XRD, XPS, TGA, EDX and Raman
spectroscopy. The nanoribbons were annealed up to 2800 ºC in graphitic
furnaces. Characterization of the heat treated samples includes TEM and SEM
observation, XRD, TGA and bulk and individual nanoribbon Raman spectroscopy.
Our TEM observations revealed that under heat treatments a crystallization
process occurs, and above 1500 ºC, the adjacent graphitic sheets find a more
stable configuration by forming loops. Single, double and multiple loops are formed
at different stages of the heat treatments.
Joule heating experiments were also carried out. Our results show that an intense
irradiaton of the electron beam, prior to the experiment, results in the formation of
sharp zigzag and armchair edges; while the experiment of an as-produced
nanoribbon results in multiple-loop formation due to the high temperatures that are
achieved in the experiment.

KEY WORDS. CVD, doped SWNTs, nanoribbons, loop formation, sharp edges

xvii
Síntesis y caracterización de nanotubos de carbono de pared
sencilla dopados y de nanolistones grafíticos

Resumen

Este trabajo concentra resultados experimentales de la síntesis de nanotubos de

carbono de pared sencilla (SWNTs) y de nanolistones grafíticos, así como una
extensiva caracterización de los mismos. El método de síntesis de SWNTs
utilizado consiste en la deposición química de vapores (CVD) asistido por aerosol
con catalizador flotante. Para el dopaje de SWNTs se eligieron compuestos
precursores que contienen al elemento dopante y se agregaron en diferentes
concentraciones a soluciones de ferroceno (Fe(C5H5)2) y etanol (C2H6O). Los
precursores con los que se trabajó son: tiofeno (C4H4S), trifenilfosfina ((C6H5)3P),
bencilamina (C7H7NH2), pirazina (C4H4N2) y metoxitrimetilsilano (CH3OSi(CH3)3).
Los nanotubos de pared sencilla y los co-productos obtenidos fueron estudiados
por SEM, TEM, EELS, EDX y espectroscopia Raman.
Nuestros experimentos con bajas concentraciones de tiofeno nos condujeron a la
síntesis de nanolistones de carbono. Una completa caracterización mediante las
técnicas SEM, TEM, EDX, XPS, XRD, TGA, y espectroscopia Raman se realizó
para el material puro, es decir sin tratamientos posteriores a la síntesis.
Se realizaron tratamientos térmicos en hornos de grafito hasta temperaturas de
2800 ºC. Los materiales tratados térmicamente fueron evaluados mediante SEM,
TEM, XRD, TGA y espectroscopia Raman en bulto y a nivel individual. Se
encontró que el material sufre un proceso de cristalización y, arriba de 1500 ºC, las
orillas de las hojas grafíticas adyacentes encuentran una configuración más
estable al formar uniones. También se realizaron experimentos de calentamiento
de Joule (Joule heating), los resultados prueban que esta técnica permite formar
tanto uniones como terminaciones armchair y zigzag dependiendo de las
condiciones del experimento.

PALABRAS CLAVE. Dopaje de SWTNs, nanolistones de carbono, uniones

xviii
 1. Introduction

1. Introduction

Carbon is one of the most abundant elements in nature. For a long time, only three
allotropes of carbon were known: amorphous carbon, graphite and diamond. The
crystalline forms (graphite and diamond) vary in physical properties due to their
atomic structure.
The reason why carbon assumes many structural forms is that a carbon atom can
form distinct types of valence bonds, where the chemical bonds refer to the
hybridization of orbitals by physicists1.
In sp2 hybridization, the carbon atoms are arranged hexagonally in honeycomb
lattices parallel to each other, each of these hexagonally arranged carbon sheets is
called graphene (Figure 1.1e). In graphite, the graphene sheets are stacked in an
ABAB… sequence with P63/mmc symmetry (Figure 1.1a). Accepted values of the
lattice constants are a0=0.2462 nm and c0=0.6707 nm at room temperature, so that
the in-plane bond length is 0.1421 nm and the interplanar separation is 0.3354 nm.
Alternative stacking sequences (AA, ABCABC…) can occur for carbon, as well as
the random stacking of graphene layers (turbostratic carbon).2

Figure 1.1 Allotropes of carbon, a) AB graphite, b) diamond, c) C60, d) armchair (5,5) carbon
3
nanotube and e) graphene sheet.

This layered configuration makes graphite a dark, soft material, where as diamond
is transparent and is one of the hardest known materials. In diamond the carbon
atoms are bonded tetrahedrally, in this configuration the atoms exhibit sp3
hybridization (Figure 1.1b).

1
 1. Introduction

The first observation of graphene goes back to the early 60’s, by Boehm4, later in
the 70’s Oberlin 5 reported the growth of filamentous carbon where images of
hollow carbon nanotubes were published. In 1985 the discovery of C60
(buckminsterfullerene, a molecule constituted of 60 carbon atoms arranged in a
truncated icosahedral structure6, see Figure 1.1c) triggered the revolution of carbon
nanostructures, leading to numerous reports of the synthesis and observation of
single-wall carbon nanotubes 7 , nano-onions 8 , nanocones 9 , nanoribbons 10 and
more recently, graphene11.
In the following pages a detailed description of the structure and properties of
carbon nanotubes, graphene and carbon nanoribbons will be presented.

1.1 Carbon nanotubes

1.1.1 Structure
A single-wall carbon nanotube can be described as a graphene sheet rolled into a
cylindrical shape so that the structure is one-dimensional with axial symmetry1.
Three examples of single-wall carbon nanotubes (SWNTs) are shown in Figure 1.2.
From this figure, it can be seen that the direction of the hexagon can be taken
almost arbitrarily.
The primary symmetry classification of a carbon nanotube is as either being achiral
or chiral. An achiral carbon nanotube is defined by a carbon nanotube whose
mirror image has an identical structure to the original one.
There are only two cases of achiral nanotubes; armchair (Figure 1.2a) and zigzag
(Figure 1.2b) nanotubes. The names of armchair and zigzag arise from the shape
of the cross-sectional ring, as is shown at the edge of the nanotubes in Figure 1.2.
Chiral nanotubes exhibit a spiral symmetry whose mirror image cannot be
superposed on to the original one. In Figure 1.3, the unrolled honeycomb lattice of
the nanotube is shown, in which OB is the direction of the nanotube axis, and the
direction OA corresponds to the equator. By considering the crystallographically
equivalent sites O, A, B, and B’, and by rolling the honeycomb sheet so that points

2
 1.1 Carbon nanotubes

O and A coincide (and points B and B’ coincide), a paper model of a carbon

nanotube can be constructed. The vectors OA and OB define the chiral vector Ch
and the translational vector T of a carbon nanotube, respectively.

8
Figure 1.2 Symmetry classification of nanotubes, a) armchair, b) zigzag and c) chiral nanotubes

The chiral vector can be expressed by the real space unit vectors a1 and a2 (see
Figure 1.3) of the hexagonal lattice defined in Equation 1
Ch = na1 + ma2 ≡ (n, m) (n, m are integers, 0 ≤ |m| ≤ n) (1)

The chiral vector Ch shown in Figure 1.3a is (4,2). An armchair nanotube

corresponds to the case of n=m, that is Ch=(n,n), and a zigzag nanotube
corresponds to the case of m=0, or Ch=(n,0). All other (n,m) chiral vectors
correspond to chiral nanotubes.

Figure 1.3 a) The unrolled honeycomb lattice of a nanotube. The figure corresponds to Ch=(4,2), b)
the unit cell (dotted rhombus), containing sites A and B where carbon atoms are located, and c) the
Brillouin zone (shaded hexagon) of a graphene or 2D graphite layer. ai and bi (i =1,2) are basis
vectors and reciprocal lattice vectors respectively. The high symmetry points, Γ, Κ, and Μ are
indicated.12

3
 1.1 Carbon nanotubes

The diameter of the carbon nanotube, dt, is given by L/π, in which L is the
circumferential length of the carbon nanotube:

dt = L / π , L = Ch = Ch ⋅ Ch = a n 2 + m 2 + mn (2)
where the a is the lattice constant of two dimensional graphite a=1.42 Å x 3=
2.49Å. The C-C bond length of graphite is 1.42 Å. In the case of carbon nanotubes,
the C-C- length is known to be slightly larger than graphite; 1.44 Å.1
The chiral angle θ is defined as the angle between the vectors Ch and a1, with
values of θ in the range 0 ≤ | θ | ≤ 30º, because of the hexagonal symmetry of the
honeycomb lattice. The chiral angle θ denotes the tilt angle of the hexagons with
respect to the direction of the nanotube axis, and the angle θ specifies the spiral
symmetry. The chiral angle θ is defined by taking the inner product of Ch and a1, to
yield an expression for cos θ:
Ch ⋅ a1 2n + m
cosθ = = (3)
Ch a1 2 n + m 2 + nm
2

thus relating θ to the integers (n,m) defined in Equation (1). In particular, zigzag and
armchair nanotubes correspond to θ=0º and θ=30º, respectively.
In Figure 1.3 we show b) the unit cell in real space and c) the Brillouin zone in
reciprocal space of 2D graphite as a dotted rhombus and shaded hexagon,
respectively, where a1 and a2 are basis vectors in real space, and b1 and b2 are
reciprocal lattice basis vectors. In the x, y coordinate system shown in Figure 1.3,
the real space basis vectors a1 and a2 of the hexagonal lattice are expressed as a1
= ( 3 a/2, a/2) and a2 = ( 3 a/2, - a/2), where a =| a1 |=| a2 | = 1.42 x 3 = 2.46 Ǻ is
the lattice constant of a graphene or 2D graphite layer. Correspondingly, the basis
vectors of the b1 and b2 of the reciprocal lattice are given by b1 = (2π/ 3 a, 2π/a)

and b2 = (2π/ 3 a, - 2π/a) corresponding to the graphene lattice constant of

4π/ 3 a in reciprocal space. The direction of the basis vectors b1 and b2 of the
reciprocal hexagonal lattice are rotated by 30° fro m the basis vectors a1 and a2 of
the hexagonal lattice in real space, as shown in Figure 1.3 b) and c). The three
high symmetry points, Γ, Κ, and Μ of the Brillouin zone are shown as the center,

4
 1.1 Carbon nanotubes

the corner and the center of the edge, respectively, of the shaded hexagon that
corresponds to the Brillouin zone of 2D graphite.

1.1.2 Electronic properties

The electronic structure of a carbon nanotube can be obtained from its parent
material, graphite, accounting for quantum confinement of the electronic states in
these 1D materials. The electronic σ bands are responsible for the strong in-plane
covalent bonds with the 2D graphene sheets, while the π bands are responsible for
weak van der Waals interactions between such sheets in graphite.13 In 1992 the
Japanese group of Hamada and co-workers14 published a report on tight binding
calculations of carbon nanotubes, predicting three classes of transport: metallic
nanotubes, semiconducting ones with narrow band gaps and semiconducting
nanotubes with moderate band gaps, with the band gap being tunable by choosing
the nanotube structure. In the same year Prof. M.S. Dresselhaus’s group published
a theoretical study of the electronic structure of carbon nanotubes15, in Figure 1.4
we show an image extracted from that document. In figure 1.4a and 1.4b, the
density of states (DOS) of two zigzag nanotubes with (n,m)= (10,0) and (9,0),
respectively, are plotted in units of states per unit cell of 2D graphite.
The corresponding DOS of 2D graphite (dotted lines) is also plotted in both figures
for comparison. The density of states of the π band becomes singlular by folding
the two-dimensional energy bands of the graphene layer into the one-dimensional
bands of the carbon nanotube, and these singularities are known as van-Hove
singularities.1
In Figure 1.4a, there is an energy gap at the Fermi level (E=0), while we have a
finite density of states for Figure 1.4b. Thus, we can have both semiconducting (Fig.
1.4a) and metallic (Fig. 1.4b) nanotubes by merely changing the nanotube
diameter. The energy gap for the semiconducting nanotubes decreases with
increasing diameter dt and in the limit of d t → ∞ , we obtain the 2D case of a zero-

gap semiconductor.
The condition for a fiber to be metallic is:
2n + m = 3q or n − m = 3q (4)

5
 1.1 Carbon nanotubes

where q is an integer. In particular, all armchair fibers are metallic, and zigzag
fibers are metallic when n is a multiple of three13.

15
Figure 1.4 Electronic density of states for two (n,m) zigzag fibers: (a) (10,0) and (b) (9,0)

Each nanotube chirality has a corresponding electronic structure, the van-Hove

singularities in the density of states of a carbon nanotube give rise to a unique set
of transition energies from the valence to the conduction bands for each (n,m) value,
denoted as Eii (ii= 11, 22, 33, ...).

Figure 1.5 Calculation of the energy

separations Eii(dt) for all (n,m) values as a
function of nanotube diameter between 0.7 <
dt < 3.0 nm. The energy transitions were
calculated using the tight binding model for
carbon nanotubes. The crosses and open
circles denote the peaks of semiconducting
and metallic nanotubes, respectively. Filled
squares denote Eii(dt) values for zigzag
nanotubes which determine the width of
each Eii(dt) curve. Note the points for zero
16
gap metallic nanotubes along the abscissa .
In figure 1.4 we can see such
transitions pointed out for the semiconductor and the metallic cases, the S and M
labels stand for semiconductor and metallic, respectively.
The interpretation of the interband transitions between van Hove singularities for
zigzag nanotubes and chiral nanotubes can be understood by plotting the energies

6
 1.1 Carbon nanotubes

for the transitions between the van Hove singularities and the valence and
conduction bands Eii (dt) of all possible (n,m) nanotubes17. Such a plot of Eii(dt) vs.
dt is shown in figure 1.5 and is known as Kataura plot, this figure is used
extensively to interpret resonance Raman spectra in carbon nanotubes.

1.1.3 Doped carbon nanotubes

In materials science we find three approaches of doping: endohedral, exohedral
and sustitutional, each of them is illustrated in Figure 1.6 in relation to carbon
nanostructures.

20 23 34
Figure 1.6 Endohedral , exohedral and sustitutional doping of carbon nanostructures

Endohedrally doped (filled) nanostructures display a doping method unique to cage

molecules and cylinders in which an ion (or ions) are encaged inside the carbon
framework of a fullerene molecule or a nanotube. These hybrid-systems represent
a completely new type of matter. They are described by the symbol for the
encapsulated atom and the fullerene molecule's formula separated by the @ sign,
which signals the endohedral nature of the complex, for example La@C82.
Endohedral fullerenes containing metal ions (so-called metallofulleres) could act as
vehicles to transport poisonous atoms, can be used as MRI markers in medicine
and can be used to stabilize and passivate reactive species. Metal filled nanotubes
can act as nonreactive passivated nanomagnets.18 Last but not least, the so-called
peapods (fullerenes filled into SWNT, e.g. C60@SWNT) represent a hybrid system
of fullerenes and single wall carbon nanotubes with intriguing new properties.19,20

7
 1.1 Carbon nanotubes

Exohedral doping of carbon nanostructures consists on the intercalation of certain

atoms into available “free” spaces of the carbon arrangement. The intercalation of
lithium in between graphitic layers has been studied for a long time.21 A few years
after the observation of SWNTs by Iijima7, Smalley’s group22 found that the single-
wall variety (SWNT) self-assembled during growth into partially ordered two-
dimensional bundles, or “ropes” with a triangular arrangement. The large one-
dimensional trigonal channels between tubes immediately suggested the possibility
that ropes might afford a new carbon host lattice for intercalation (see Figure 1.6
Exohedral), and that the reaction would likely be accompanied by charge exchange
between host and guest. 23,24
It appeared that the crystal chemistry of such phases would be similar to that of
graphite intercalation compounds and alkali fullerides. 23,25
Sustitutional doping of carbon nanostructures happens when foreign atoms replace
carbon atoms in the ordered network. Hetero-doping refers to the replacement of C
atoms by more than one atom specie. These substitutions alter the electronic
properties of the material.
Tailoring the electronic structure of materials by adding impurities is a long-known
technique for the semiconductor industry.
P-type and n-type doping of carbon nanotubes is possible if we substitute carbon
atoms in the honey-comb lattice with other atomic species with less or more
electrons, respectively.
The most common dopants of carbon nanotubes are boron 26 , 27 , 28 , 29 , 30 and
nitrogen,31,32 which are first neighbors of carbon in the periodic table, in the III and
V groups, respectively.
Nitrogen atoms bond in two fashions to carbon atoms, either double-bonded or
triple-bonded. Triple bonding of nitrogen corresponds to highly coordinated N
atoms replacing C atoms in the graphene sheets; double bonding of nitrogen
atoms is commonly referred to as pyridine-like nitrogen (see Figure 1.6 and 1.7C).
For pyridine-like N, it is envisaged that C vacancies are formed within a
predominantly hexagonal graphene network, thus leading to stable N-rich cavities

8
 1.1 Carbon nanotubes

similar to those predicted in C3N4-layered structures. These vacancies also may

well be responsible for the formation of irreglular graphene sheets. 33

Figure 1.7 Nitrogen doped MWNTs (CNx). (A) a)TEM image of a typical region exhibiting CNx
produced by pyrolysis of ferrocene-melamine solutions, b) picture of bamboos, c) HRTEM image of
34
an aligned CNx exhibiting corrugation, interlinkage and compartments. (B) HRTEM image of a
37
CNx produced by aerosol pyrolysis. (C) Theoretical LDOS associated with a pyridine-like structure
with N-doping carbon nanotubes displaying an armchair (10,10) (upper panel) and a zigzag (17,0)
(lower panel) configurations. The LDOS of doped (black curve) and pure (red curve) carbon
34
nanotubes are compared.

The introduction of nitrogen atoms into multi-wall carbon nanotubes (MWNTs)

induces the so-called “bamboo-like” morphologies34,35,36,37 (see Figure 1.7 A and B).
As mentioned before, impurities in the carbon network alter the electronic structure
of carbon nanotubes. Terrones et al.34 analyzed the DOS associated with the N
atoms arranged in a pyridinic fashion, randomly distributed within armchair and
zigzag carbon nanotubes. The corresponding DOS for both types of chiralities are
depicted in Figure 1.7C. It is clear that pyridine-like sites are responsible for the
prominent donor-like features (shown by arrows in the conduction band) just above
the Fermi energy.
The production of N-doped SWNTs has been achieved by arc discharge 38 and
through the pyrolysis of ferrocene-benzylamine solutions in 2005 by Villalpando-
Paez and co-workers39. We have reproduced these results and we have achieved

9
 1.1 Carbon nanotubes

the synthesis of N-doped SWNTs using a different nitrogen precursor, namely,

pyrazine.40
Although P atoms are larger than C atoms, it has been shown that phosphorus can
form substitutional defects in diamond sp3 thin films 41 . In the case of carbon
nanotubes, P behaves as an n-type donor and modifies the electronic and optical
properties. In Figure 1.8a we show the optimized structure of a zigzag (10,0)
nanotube with a single sustitutional phosphorus atom,42 it is evident that due to the
larger size of the doping atom, it lies outside the cylindrical wall of the nanotube
(consequence of the longer P-C bonds).

Figure 1.8 a) Optimized structure of a hydrogen passivated (10,0) carbon nanotube with a single P
substitutional defect (179 C, 1 P, and 20 H atoms).42 b) Plot of defect formation energy vs.
curvature. Dotted line represents the formation energy for a planar graphene sheet, as the limit for
43
large nanotubes.

43
Maciel and co-authors computed the defect formation energy for carbon
nanotubes from DFT, considering ∆E = ED - EPris - µP - µC where ED (EPris) is the
total energy for the P-doped (pristine) SWNT, and µP (µC) is the total energy of
isolated P (C) atoms. These energies are plotted in Figure 1.8b, showing that the
formation energy decreases as the curvature κ (inverse of the nanotube radius)
increases, with a fitted relation ∆E (eV) ∼ 6.625 - 4.379κ. This reduction reflects
the lower strain in the carbon network required to accept the trigonal bonds of the
phosphorus ion when the nanotube diameter decreases. Therefore, phosphorus
doping is more likely to occur in small diameter nanotubes and in fact can induce
the formation of thinner nanotubes when doping occurs during growth (similar
results have been found for nitrogen doping44).

10
 1.1 Carbon nanotubes

In the literature we can find experimental reports of phosphorous-nitrogen hetero-

doping of multi-wall carbon nanotubes45. We have been able to synthesize for the
first time, phosphorus-doped SWNTs.43 A full description of the synthesis
procedure and materials is given in Chapter II.
Figure 1.9 Two
different views of
nanotubes with one
substitutional Si atom:
a) (6,6) frontal, b)
(6,6) lateral, c) (10,0)
frontal, and d) (10,0)
lateral. Electronic
density of states for
the doped (filled line)
and undoped (dashed
line) nanotubes: e)
(6,6) nanotube and f)
48
(10,0) nanotube.

46 47
Silicon doping of fullerenes and fullerene-like nanostructures has been
achieved experimentally and was reported in the late 90’s. The incorporation of
silicon species into the hexagonal lattice of SWNTs was proposed theoretically by
Baierle et al. in 2001. 48 These calculations showed that Si-doping of SWNTs
introduces donor-like states above the Fermi level (silicon acts as an electron
donor), as pictured in Figure 1.9 e) and f). To this date the synthesis of Si-doped
SWNTs has not yet been reported.
Sulfur had been demonstrated to play a key role in the synthesis of carbon
nanostructures since 1981, when Katsuki et al.49,50 reported that sulfur exerts an
excellent catalytic effect to produce carbon fibers, it has also been reported in the
synthesis of DWNTs51 and nanotube fibers52.
However, it was only very recently that the incorporation of S into the sp2 carbon
lattice was experimentally demonstrated53. From electronic structure calculations,
we note that the inclusion of S into CNT has similar effects as what was found for

11
 1.1 Carbon nanotubes

P. S is also a larger atom than C and tends to induce large corrugation when
incorporated into sp2-like graphitic, fullerene, and nanotube structures.

42
Figure 1.10 a) Optimized geometry for a (8,0) nanotube with a single S substitutional defect. b)
Snapshots from quantum molecular dynamics simulations, showing the effect of widening the CNT
53
diameter in the presence of sulfur.

Figure 1.10a shows the optimized structure for a single S substitutional defect in a
(8,0) nanotube at T = 1000 K. The sulfur atoms promote the formation of a small
bump when they go out of the sp2 carbon lattice plane. This system becomes
unstable at higher temperatures (2,500 K) and bond dissociation occurs at the site
of the S substitution. Addition of a full ring of S atoms into a (8,0) nanotube is
stable at relatively high temperatures (up to 2,500 K) but the substitution causes a
significant widening of the nanotube, as pictured in Figure 1.10b. This widening
can be directly correlated with the formation of heptagonal rings (negative
curvature) and could be caused by the so-called cone-stacked structure and
branching usually observed in carbon fibers when sulfur is present during
synthesis. 53

The interesting properties of doped SWNTs have motivated us to work with an

aerosol assisted CVD technique with floating catalyst to produce single-wall carbon
nanotubes in the presence of sulphur, nitrogen, phosphorus and silicon using
compound precursors for each of the listed elements.
A careful description of the experiments, synthesized materials and
characterization results will be given in Chapter II.

12
 1.2 Graphene and carbon nanoribbons

1.2 Graphene and carbon nanoribbons

Graphene is the name given to a flat monolayer of carbon atoms tightly packed into
a two dimensional (2D) honey-comb lattice, and is a basic building block for
graphitic materials of all other dimensionalities.54
Whether a strictly 2D crystal can exist, was first raised theoretically more than 70
years ago by Peierls and Landau.55 They showed that, in the standard harmonic
approximation, thermal fluctuations should destroy long-range order, resulting in
the melting of a 2D lattice at any finite temperature.
However, although theory does not allow perfect crystals in 2D space, it does not
forbid nearly perfect 2D crystals in 3D space.
Indeed, a detailed analysis of the 2D crystal problem beyond the harmonic
approximation has led to the conclusion that the interaction between bending and
stretching long-wavelength phonons could in principle stabilize atomically thin
membranes through their deformation in the third dimension. 54,56

57
In 1996 Nakada et al. calculated the electronic properties of graphene
nanoribbons, with different edge configurations and widths using a tight binding
(TB) approximation.

Figure 1.11 a) Schematic of an armchair nanoribbon with Na=11. The empty circles denote
hydrogen atoms passivating the edge carbon atoms, and the black and blue rectangles represent
atomic sites belonging to different sublattices in the graphene structure. The 1D unit cell distance
and ribbon width are represented by da and wa, respectively. (b) Schematic of a Nz=6 zigzag
58
nanoribbon.

Such graphite networks have an electronic structure which is somewhat different

from that of bulk graphite, since quite a large fraction of the carbon atoms sit on the

13
 1.2 Graphene and carbon nanoribbons

edge. There are two basic shapes for graphite edges, namely, armchair and zigzag
edges. The graphene nanoribbons with armchair shaped edges on both sides are
classified by the number of dimer lines (Na) across the ribbon width (Figure 1.11a).
Likewise, ribbons with zigzag shaped edges on both sides are classified by the
number of the zigzag chains (Nz) across the ribbon width (Figure 1.11b).
Their TB results showed that in armchair edge configurations, the ribbon width can
determine the metallic or semiconductor behavior, whereas pure zigzag
configurations or mixtures of zigzag-armchair configurations result in special
localized states near the Fermi level, regardless of the nanoribbon width.
More recent calculations by Louie and co-workers 58 based on a first-principles
approach, demonstrated that both types of edges have non-zero and direct band
gaps. In the case of armchair nanoribbons, a semiconductor behavior was found,
with energy gaps decreasing as a function of increasing ribbon widths (wa). The
variations in energy gap however exhibit three distinct family behaviors.

Figure 1.12 a) Representative configurations of graphene nanoribbons (GNRs) from MD (molecular

dynamics) simulations at different temperatures, T. At T=700 K, two samples of 13-AGNRs and one
of a (6,0) nanotube, for comparison are shown. At T=300 K, two samples of 7-AGNRs are shown.
Ground state configurations, at T=0 K, of the GNRs display a saddle-shape chain and periodic twist
configurations. b) The twist of GNRs disappears for wider ribbons (here zigzag in dark red and
armchair in dark blue) and is replaced by the near-edge undulations, a large pristine graphene
sheet in equilibrium displays a near-edge “frills” pattern; the computed period λ of the edge-ripple is
plotted as a function of inverse edge force, 1/f. (Extracted from Reference 59).

14
 1.2 Graphene and carbon nanoribbons

When the nanoribbons are terminated with zigzag edges, the first-principles results
predict also direct band gaps which decrease with increasing width (wz). However,
when spins are considered, the zigzag nanoribbons are predicted to have a
magnetic insulating ground state with ferromagnetic ordering at each zigzag edge
and antiparallel spin orientation between the two edges.58

In pristine graphene ribbons, disruption of the aromatic bond network results in

depopulation of covalent orbitals and tends to elongate the edge, with an effective
force of fe ~ 2 eV/Å (larger for armchair edges than for zigzag edges, according to
calculations by Yakobson 59 ). This force can have quite striking macroscopic
manifestations in the case of narrow ribbons, as it favors their spontaneous
twisting, resulting in the parallel edges forming a double helix, resembling DNA,
with a pitch λt of about 15 - 20 lattice parameters, see Figure 1.12a. Through
atomistic simulations, Yakobson investigated how the torsion τ~ 1/ λt decreases
with the width of the ribbon, and observed its bifurcation: the twist of wider ribbons
abruptly vanishes and instead the corrugation localizes near the edges (see Figure
1.12b).

Monolayer graphene was presumed not to exist in the free state, although it was
observed in 1962 by Boehm4, unfortunately this work was not widely known.
Graphene was for long described as an ‘academic’ material and was believed to be
unstable with respect to the formation of curved structures such as soot, fullerenes
and nanotubes.
Suddenly, the model turned into reality, when free-standing graphene was
unexpectedly found in 2004 by Novoselov and Geim’s group.11,54
The importance of graphene is attributed to its amazing electronic properties
showing ambipolar electric field effect, quantum hall effect at room temperature,
and the fact that its charge carriers are described by the Dirac equation as
massless fermions.54

15
 1.2 Graphene and carbon nanoribbons

A significant draw-back in this graphene revolution is the lack of control of the

dimensions and the edge structure (either zigzag or armchair) of the crystals, since
both characteristics are crucial on the determination of the electronic properties.

Novoselov’s preparation method11 consists on the production of monolayer and

few-layer graphene flakes by micro-mechanical exfoliation of HOPG, being
worldwide known as the “scotch-tape” method; this friendly production technique
put graphene on the spot light by making it available to any research lab around
the world (see Figure 1.13A).

Figure 1.13 Methods of production of graphene. A) Novoselov graphene flakes obtained by micro-
54 60
mechanical cleavage, B) high temperature heat treatments of diamond nano-particles , C)
60,61
epitaxial growth of graphene over SiC substrates , D) CVD grown graphene films over Ni thin
62
layers (scale bar = 25µm) .

Apart from Novoselov’s micro mechanical cleavage of HOPG to produce graphene,

there are other available methods: high temperature heat treatments of diamond
nano-particles61, the epitaxially grown graphene sheets over SiC substrates62 and
the CVD production of graphene films63,64,65 (Figure 1.13 B-D).
In the literature are reported several ways to obtain carbon nanoribbons:
temperature and pressure treatments of carbon nanotubes 66 , heat treatment of

16
 1.2 Graphene and carbon nanoribbons

diamond nanoparticles deposited on HOPG67, chemical treatments of graphite68,

pyrolysis of hydrocarbons8, 69 , 70 , 71 , using hydrothermal processes 72 , 73
, STM
74 75,76
lithography , and more recently by unfolding or etching carbon nanotubes 77,
see Figure 1.14.

Figure 1.14 CVD and other production methods of carbon nanoribbons. A) Scanning electron
micrographs of Maruyama’s filamentous graphite, where iron particles at the ends of the structures
10
were found , B) SEM images of the graphitic nanoribbons produced by pyrolisis of ethanol-
ferrocene-thiophene solutions69, C) SEM image of the nanoribbons produced by pyrolisis of THF
68
and ferrocene solutions , D) low magnification TEM image of the nanoribbons produced by the ZnS
70
template method , E) AFM image of the nanoribbons obtained through high temperature
66
treatments of diamond nano-particles , F) 3D STM image of an 8-nm-wide graphene nanoribbon
73
patterned by STM lithography , G) AFM image of chemically derived nanoribbons from graphite
67 79
(scale bars = 100 nm) , H) scheme illustrating the structure of a collapsed nanotube , I) and J)
TEM images of an amorphous carbon nanoribbon and graphitic nanobelt, respectively, produced by
71,72
hydrothermal processes .

17
 1.2 Graphene and carbon nanoribbons

Besides the attempts to synthesize carbon nanoribbons, it has been observed that
collapsed nanotubes result in nanobelts or nanoribbons with closed edges as
reported since 1995 (see Figure 1.14H).78,79,80,81,82

Despite the production methods listed above, the synthesis of nanoribbons is still a
challenge, due to drawbacks like the low yield production of the material, high
pressure and temperature required for the synthesis, specific equipment and high
purity materials involved in the processes and most of all, lack of control on the
final structure of the nanoribbons.

The third chapter of this work is devoted to the CVD synthesis of graphitic
nanoribbons. A detailed description of the synthesis technique, complete
characterization and heat treatment results induced by furnace heating will be
presented.
The fourth chapter of this thesis is devoted to the presentation of results of Joule
heating experiments on pristine graphitic nanoribbons.

18
 1.3 References

1.3 References

1
R. Saito, G. Dresselhaus, M.S. Dresselhaus. Physical Properties of Carbon
Nanotubes. Imperial College Press. London, 1998
2
M.S. Dresselhaus, G. Dresselhaus, K. Sigihara, I.L. Spain, H.A. Goldberg.
Graphite fibers and filaments. Springer series in Materials Science 5.
Springer-Verlag, New York, Berlin, Heidelberg. 1988
3
Coordinates generated with S. Maruyama’s free software “Wrapping”, downloaded
from the site http://www.photon.t.u-tokyo.ac.jp/~maruyama/wrapping3/wrapping.html

4
H. P. Boehm, A. Clauss, G. Fischer, U. Hofmann. Surface properties of
extremely thin graphite lamellae. Proceedings of the Fifth Conference on
Carbon. Pergamon Press, Oxford/London/New York/Paris, 1962
5
A. Oberlin, M. Endo, T. Koyama. Filamentous growth of carbon through
benzene decomposition. Journal of Crystal Growth 32, 335-349, 1976
6
H.W. Kroto, J.R. Heath, S.C. O’Brien, R.F. Curl, R.E. Smalley. C60:
Buckminsterfullerene. Nature 318, 162-163, 1985
7
S. Iijima. Helical microtubules of graphitic carbon. Nature 354, 56-58, 1991
8
D. Ugarte. Curling and closure of graphitic networks under electron-beam
irradiation. Nature 359, 707-709, 1992
9
A. Krishnan, E. Dujardin, M.M.J. Treacy, J. Hugdahl, S. Lynum, T.W. Ebbesen.
Graphitic cones and the nucleation of curved carbon surfaces. Nature 388, 451-
454, 1997
10
H. Murayama, T. Maeda. A novel form of filamentous graphite. Nature 345,
792-793, 1990
11
K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos,
I.V. Grigorieva, A.A. Firsov. Electric field effect in atomically thin carbon films.
Science 306, 666-669, 2004
 1.3 References

12
M.S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio. Raman spectroscopy of
carbon nanotubes. Physics Reports 409, 47, 2005
13
A. Jorio, R. Saito, G. Dresselhaus, M.S. Dresselhaus. Determination of
nanotubes properties by Raman spectroscopy. Phil. Trans. R. Soc. Lond. A 362,
2311-2336, 2004
14
N. Hamada, S. Sawada, A. Oshiyama. New one-dimensional conductors:
graphitic microtubules. Phys. Rev. Lett. 68, 10, 1579-1581, 1992
15
R. Saito, M. Fujita, G. Dresselhaus, M.S. Dresselhaus. Electronic structure of
chiral graphene tubules. Appl. Phys. Lett. 60, 18, 1992
16
R. Saito, G. Dresselhaus, M.S. Dresselhaus. Trigonal warping effect of carbon
nanotubes. Phys. Rev. B 61, 4, 2981-2990, 2000
17
M.S. Dresselhaus, P.C. Eklund. Phonons in carbon nanotubes. Advances in
Physics 49, 6, 705-814, 2000
18
K. Lafdi, A. Chin, N. Ali, J.F. Despres. Cobalt-doped carbon nanotubes:
preparation, texture, and magnetic properties. J. Appl. Phys. 79, 8, 6007-6009,
1996
19
B. W. Smith, M. Monthioux, D. E. Luzzi. Encapsulated C60 in carbon nanotubes.
Nature 396, 323-324, 1998
20
Extracted from : www.ifw-dresden.de/institutes/iff/research/Carbon/CNT/
electronic-and-optical-properties/doped/filling
21
M.S. Dresselhaus, G. Dresselhaus, J.E. Fischer. Graphite intercalation
compounds: electronic properties in the dilute limit. Phys. Rev. B 15, 3180-
3191, 1977
22
A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. H. Lee, S.G.
Kim, A.G. Rinzler, D.T. Colbert, G.E. Scuseria, D. Tománek, J.E. Fischer, R.E.
Smalley. Crystalline ropes of metallic carbon nanotubes. Science 273, 483-487,
1996

20
 1.3 References

23
J. E. Fischer. Chemical doping of single-wall carbon nanotubes. Acc. Chem.
Res. 35, 1079-1086, 2002
24
J. Kong, C. Zhou, E. Yenilmez, H. Dai. Alkaline metal-doped n-type
semiconducting nanotubes as quantum dots. Appl. Phys. Lett. 77, 24, 3977-
3979, 2000
25
T. Kanigaki, T.W. Ebbesen, S. Saito, J. Mizuki, J.S. Tsai, Y. Kubo, S.
Kuroshima. Superconductivity at 33 K in CsxRbyC60. Nature 352, 222-223, 1991.
T. Yildirim, O. Zhou, J.E. Fischer, N. Bykovetz, R.A. Strongin, M.A. Cichy, A.B.
Smith III, C.L. Lin, R. Jelinek. Intercalation of sodium heteroclusters into the C60
lattice. Nature 360, 568-571, 1992
26
B.C. Satishkumar, A. Govindaraj, K.R. Harikumar, J.-P. Zhang, A.K. Cheetham,
C.N.R. Rao. Boron-carbon nanotubes from the pyrolysis of C2H2-B2H6 mixtures.
Chem. Phys. Lett. 300, 473-477, 1999
27
X. Blase, J.-C. Charlier, A. De Vita, R. Car, Ph. Redlich, M Terrones, W.K. Hsu,
H. Terrones, D.L. Carroll, P.M. Ajayan. Boron-mediated growth of long helicity-
selected carbon nanotubes. Phys. Rev. Lett. 83, 24, 5078- 5081, 1999
28
W. Han, Y. Bando, K. Kurashima, T. Sato. Boron-doped carbon nanotubes
prepared through a substitution reaction. Chem. Phys. Lett. 299, 368-373,
1999
29
D. Goldberg, Y. Bando, W. Han, K. Kurasshima, T. Sato. Single-walled B-
doped carbon, B/N-doped carbon and BN nanotubes synthesized from single-
walled carbon nanotubes through a substitutution reaction Chem. Phys. Lett.
308, 337-342, 1999
30
K. MacGuire, N. Gothard, P.L. Gai, M.S. Dresselahaus, G. Sumanasekera,
A.M. Rao. Synthesis and Raman characterization of boron-doped single-walled
carbon nanotubes . Carbon 43, 219-227, 2005

21
 1.3 References

31
O. Stephan, P.M. Ajayan, C. Colliex, Ph. Redlich, J.M. Lambert, P. Bernier, P.
Lefin. Doping graphitic and carbon nanotube structures with boron and nitrogen.
Science 266, 5191, 1683, 1994
32
M. Terrones, N. Grobert, J. Olivares, J.P. Zhang, H. Terrones, K. Kordatos,
W.K. Hsu, J.P. Hare, P.D. Townsend, K. Prassides, A.K. Cheetham, H.W. Kroto,
D.R.M. Walton. Controlled production of aligned-nanotube bundles. Nature 388,
52, 1997
33
M. Terrones, P. Redlich, N. Grobert, S. Trasobares, W.-K. Hsu, H. Terrones,
Y.-Q. Zhu, J. P. Hare, C. L. Reeves, A. K. Cheetham, Manfred Rühle, H. W.
Kroto, D.R. M. Walton. Carbon nitride nanocomposites: formation of aligned
CxNy nanofibers. Advanced Materials 11, 8, 655, 1999
34
M. Terrones, P.M. Ajayan, F. Banhart, X. Blase, D.L. Carroll, J.C. Charlier, R.
Czerw, B. Foley, N. Brobert, R. Kamalakaran, P. Kohler-Redlich, M. Rühle, T.
Seeger, H. Terrones. N-doping and coalescence of carbon nanotubes:
synthesis and electronic properties. Appl. Phys. A 74, 355-361, 2002
35
R. Sen, B.C. Satishkumar, A. Govindaraj, K.R. Harikumar, G. Raina, J.-P.
Zhang, A.K. Cheetham, C.N.R. Rao. B-C-N, C-N, and B-N nanotubes
produced by the pyrolysis of precursor molecules over Co catalysts. Chem.
Phys. Lett. 287, 671-676, 1998
36
R. Droppa Jr., P. Hammer, A.C.M. Carvalho, M.C. dos Santos, F. Alvarez.
Incorporation of nitrogen in carbon nanotubes. Jorunal of Non-Crystalline Solids
299-302, 874-879, 2002
37
M. Glerup, M. Castignolles, M. Holzinger, G. Hug, A. Loiseau, P. Bernier.
Synthesis of highly nitrogen-doped multi-walled carbon nanotubes. Chem.
Commun. 2542-2543, 2003
38
M. Glerup, J. Steinmetz, D. Samaille, O. Stéphan, S. Enouz, A. Loiseau, S.
Roth, P. Bernier. Synthesis of N-doped SWNT using the arc-discharge
procedure. Chem. Phys. Lett. 238, 193-197, 2004

22
 1.3 References

39
F. Villalpando-Paez, A. Zamudio, A.L. Elias, H.Son, E.B. Barros, S.G. Chou,
Y.A. Kim, H. Muramatsu, T. Hayashi, J. Kong, H. Terrones, G. Dresselhaus, M.
Endo, M. Terrones, M.S. Dresselhaus. Synthesis and characterization of long
strands of nitrogen-doped single-walled carbon nanotubes. Chem. Phys. Lett.
424, 345-352, 2006
40
J. Campos-Delgado, I.O. Maciel, D. A. Cullen, D. Smith, A. Jorio, M.A. Pimenta,
H. Terrones, M. Terrones. Chemical vapor deposition synthesis of N-, P- and
Si-doped single-wall carbon nanotubes. (Submitted to ACS Nano in November
2009)
41
H. Sternschulte, K. Thonke, R. Sauer, S. Koizumi. Optical evidence for 630-
meV phosphorus donor in synthetic diamond. Phys. Rev. B 59, 20 , 12924-
12927, 1999
42
B. G. Sumpter, J. Huang, V. Meunier, J. M. Romo-Herrera, E. Cruz-Silva, H.
Terrones, M. Terrones. A theoretical and experimental study on manipulating
the structure and properties of carbon nanotubes using substitutional dopants.
Int. J. of Quant. Chem. 109, 97-118, 2009
43
I.O. Maciel, J. Campos-Delgado, E. Cruz-Silva, M.A. Pimenta, B.G. Sumpter, V.
Meunier, F. López-Urías, E. Muñoz-Sandoval, H. Terrones, M. Terrones, A.
Jorio. Synthesis, electronic structure, and Raman scattering of phosphorus-
doped single-wall carbon nanotubes. Nanoletters 9,6, 2267-2272, 2009
44
B. G. Sumpter, V. Meunier, J.M. Romo-Herrera, E. Cruz-Silva, D.A. Cullen, H.
Terrones, D.J. Smith, M. Terrones. Nitrogen-mediated carbon nanotube
growth: diameter reduction, metallicity, bundle dispersability, and bamboo-like
structure formation. ACS Nano 1, 4, 369-375, 2007
45
E. Cruz-Silva, D.A. Cullen, L. Gu, J.M. Romo-Herrera, E. Muñoz-Sandoval, F.
López-Urias, B.G. Sumpter, V. Meunier, J.-C. Charlier, D. J. Smith, H. Terrones,
M. Terrones. Heterodoped nanotubes: theory, synthesis, and characterization

23
 1.3 References

of phosphorus-nitrogen doped multiwalled carbon nanotubes. ACS Nano 2, 3,

441-448, 2008
46
T. Kimura, T. Sugai, H. Shinohara. Production and characterization of boron-
and silicon-doped carbon clusters. Chem. Phys. Lett. 256, 269-273, 1996
47
C. Ray, M. Pellarin, J.L. Lermé, J.L. Vialle, M. Broyer, X. Blase, P. Mélinon, P.
Kéghélian, A. Perez. Synthesis and structure of silicon-doped heterofullerenes.
Phys. Rev. Lett. 80, 24, 5365, 1998
48
R.J. Baierle, S. B. Fagan, R. Mota, A. J. R. da Silva, A. Fazzio. Electronic and
structural properties of silicon-doped carbon nanotubes. Phys. Rev. B 64,
085413, 2001
49
H. Katsuki, K. Matsunaga, M. Egashira, S. Kawasumi. Formation of carbon
fibers from naphthalene on some sulfur-containing substrates. Carbon 19, 148-
150, 1981
50
G. Tibbetts, C.A. Bernardo, D. W. Gorkiewicz, R.L. Alig. Role of sulfur in the
production of carbon fibers in the vapor phase. Carbon 32, 4, 569-576, 1994
51
L.Ci, Z. Rao, Z. Zhou, D. Tang, X. Yan, Y. Liang, D. Liu, H. Yuan, W. Zhou, G.
Wang, W. Liu, S. Xie. Double wall carbon nanotubes promoted by sulfur in a
floating iron catalyst CVD system. Chem. Phys. Lett. 359, 63-67, 2002
52
Y.-L. Li, I.A. Kinloch, A.H. Windle. Direct spinning of carbon nanotube fibers
from chemical vapor deposition synthesis. Science 304, 276-278, 2004, and M.
Mottta, Y.-L. Li, I. Kinloch, A. Windle. Mechanical properties of continuously
spun fibers of carbon nanotubes. Nanoletters 5, 8, 1529-1533, 2005
53
J. M. Romo-Herrera, B.G. Sumpter, D.A. Cullen, H. Terrones, E. Cruz-Silva, D.
J. Smith, V. Meunier, M. Terrones. An atomistic branching mechanism for
carbon nanotubes: sulfur as the triggering agent. Angew. Chem. Int. Ed. 47,
2948-2953, 2008

24
 1.3 References

54
A.K. Geim, K.S. Novoselov. The rise of graphene. Nature Materials 6, 183-191,
2007
55
J.C. Meyer, A.K. Geim. M.I. Katsnelson, K.S. Novoselov, T.J. Booth, S. Roth.
The structure of suspended graphene sheets. Nature 446, 60-63, 2007, and
references therein.
56
V.B. Shenoy, C.D. Reddy, A. Ramasubramaniam, Y.W. Zhang. Edge-stress-
induced warping of graphene sheets and nanoribbons. Phys. Rev. Lett. 101,
245501, 2008
57
K. Nakada, M. Fujita, G. Dresselhaus, M.S. Dresselhaus. Edge state in
graphene ribbons: nanometer size effect and edge shape dependence. Phys.
Rev. B. Vol. 54, 24, 1996
58
Y.-W. Son, M.L. Cohen, S. G. Louie. Energy gaps in graphene nanoribbons.
Phys. Rev. Lett. 97, 216803, 2006
59
K.V. Bets, B.I. Yakobson. Spontaneous twist and intrinsic instabilities of pristine
graphene nanoribbons. Nano Research 2, 161-166, 2009

60
C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud, D. Mayou, T. Li, J. Hass,
A. N. Marchenkov, E. H. Conrad, P.N. First, W. A. de Heer. Electronic
confinement and coherence in patterned epitaxial graphene. Science 312, 1191,
2006
61
A.M. Affoune, B.L.V. Prasad, AH.Sato, T. Enoki, Y. Kaburagi, Y. Hishoyama.
Experimental evidence of a single nano-graphene. Chem. Phys. Lett. 348, 17,
2001
62
C. Berger, Z. Song, T. Li, X. Li, A.Y. Ogbazghi, R. Feng, Z. Dai, A. N.
Marchenkov, E.H. Conrad, P.N. First, W. A. de Heer. Ultrathin epitaxial
graphite: 2D electron gas properties and a route toward graphene-based
nanoelectronics. J. Phys. Chem. B 108, 19912-19916, 2004

25
 1.3 References

63
A. Reina, S. Thiele, X. Jia, S. Bhaviripudi, M.S. Dresselhaus, J. A. Schaefer, J.
Kong. Growth of large-area single- and bi-layer graphene by controlled carbon
precipitation on polycrystalline Ni surfaces. Nano Research 2, 509-516, 2009
64
A. Reina, X. Jia, J. Ho, D. Nezich, H. Son, V. Bulovic, M.S. Dresselhaus, J.
Kong. Large area, few-layer graphene films on arbitrary substrates by chemical
vapor deposition. Nanoletters 9, 1, 30-35, 2009
65
K.S. Kim, Y. Zhao, H. Jang, S.Y. Lee, J.M. Kim, K.S. Kim, J.-H. Ahn, P. Kim, J.-
Y. Choi, B.H. Hong. Large-scale pattern growth of graphene films for
stretchable transparent electrodes. Nature 457, 706, 2009
66
M. Zhang, D.H. Wu, C.L. Xu, Y.F. Xu, W.K. Wang. Ribbon-like nanostructures
transformed from carbon nanotubes at high temperature and pressure. Nano
Structured Materials 10, 7, 1145-1152, 1998
67
L.G. Cançado, M.A. Pimienta, B.R.A. Neves, G. Medeiros-Ribeiro, T. Enoki, Y.
Kobayashi, K. Takai, K. Fukui, M.S. Dresselhaus, R. Saito, A. Jorio. Anisotropy
of the Raman spectra of nanographite ribbons. Phys. Rev. Lett. 93, 4, 2004
68
X. Li, X. Wang, L. Zhang, S. Lee, H. Dai. Chemically derived, ultrasmooth
graphene nanoribbon semiconductors. Science 319, 1229, 2008
69
P. Mahanandia, K.K. Nanda, V. Prasad, S.V. Subramanyam. Synthesis and
characterization of carbon nanoribbons and single crystal iron filled carbon
nanotubes. Materials Research Bulletin 43, 3252-3262, 2008
70
J. Campos-Delgado, et. al. Bulk production of a new form of sp2 carbon:
crystalline graphene nanoribbons. Nanoletters 8, 9, 2773-2778, 2008
71
D. Wei, Y. Liu, H. Zhang, L. Huang, B. Wu, J. Chen, G. Yu. Scalable synthesis
of few-layer graphene ribbons with controlled morphologies by a template
method and their applications in nanoelectromechanical switches. J. Am. Chem.
Soc. 131, 11147-11154, 2009

26
 1.3 References

72
Y. Xiong, Y. Xie, X. Li, Z. Li. Production of novel amorphous carbon
nanostructures from ferrocene in low-temperature solution. Carbon 42, 1447-
1453, 2004
73
Z. Kang, E. Wang, B. Mao, Z. Su, L. Gao, S. Lian, L. Xu. Controllable
fabrication of carbon nanotube and nanobelt with a polyoxometalate-assisted
mild hydrothermal process. J. Am. Chem. Soc. 127, 6534-6535, 2005
74
L. Tapasztó, G. Dobrik, P. Lambin, L. P. Biró. Tailoring the atomic structure of
graphene nanoribbons by scanning tunneling microscope lithography. Nature
Nanotechnology 3, 397, 2008
75
A. G. Cano-Márquez, F. J. Rodríguez-Macías, J. Campos-Delgado, C.G.
Espinosa-González, F. Tristán-López, D. Ramírez-González, D. A. Cullen, D.J.
Smith, M. Terrones, Y.I. Vega-Cantú. ExMWNTs: graphene sheets and ribbons
produced by lithium intercalation and exfoliation of carbon nanotubes.
Nanoletters 9, 4, 1527-1533, 2009
76
D. V. Kosynkin, A.L. Higginbotham, A. Sinitskii, J.R. Lomeda, A. Dimiev, B.C.
Price, J.M. Tour. Longitudinal unzipping of carbon nanotubes to form graphene
nanoribbons. Nature 458, 2009
77
L. Jiao, L. Zhang, X. Wang, G. Diankov, H. Dai. Narrow graphene nanoribbons
from carbon nanotubes. Nature 458, 2009
78
N.G. Chopra, L.X. Benedict, V.H. Crespi, M.L. Cohen, S.G. Louie, A. Zettl. Fully
collapsed carbon nanotubes. Nature 377, 135-138, 1995
79
M.-F. Yu, O. Lourie, M.J. Dyer, K. Moloni, T.F. Kelly, R.S. Ruoff. Strength and
breaking mechanism of multiwalled carbon nanotubes under tensile load.
Science 287, 637-640, 2000
80
H.R. Gutiérrez, U.J. Kim, P.C. Eklund. Thermal conversion of bundled carbon
nanotubes into graphitic ribbons. Nanoletters 5, 11, 2195-2201, 2005

27
 1.3 References

81
C.-T. Lin, T.-H. Chen, T.-S. Chin, C.-Y. Lee, H.-T. Chiu. Quasi two-dimensional
carbon nanobelts synthesized using a template method. Carbon 46, 741-746,
2008
82
M. Motta, A. Moisala, I.A. Kinloch, A.H. Windle. High performance fibres from
‘dog bone’ carbon nanotubes. Advanced Materials 19, 3721-3726, 2007

28
 2.1 Description of the synthesis method, materials and equipment

2. Production of doped single-wall carbon nanotubes

using the aerosol pyrolysis technique

The synthesis of carbon nanostructures, in particular single-wall carbon nanotubes,

has been achieved through different techniques, the most popular are: arc
discharge, laser ablation, and pyrolysis of hydrocarbons (see Appendix A).
Thermolysis or pyrolysis of hydrocarbons (e.g. methane, benzene, acetylene,
naphthalene, ethylene, etc.) and of carbon monoxide in the presence of metal
catalysts (e.g. Co, Ni, Fe, Pt, and Pd deposited on substrates such as silicon,
graphite or silica) at high temperatures (>600 ºC) yields graphite filaments, carbon
nanofibers and nanotubes. The hydrocarbon pyrolysis process is also known as
chemical vapor deposition (CVD) 1.
The growth of single-wall carbon nanotubes (SWNTs) requires a catalytic particle
present during the synthesis; some groups have developed methods to deposit
catalytic nanoparticles over substrates for CVD production of SWNTs.
However, an alternative method of providing metal particles as catalysts during the
CVD reaction is through the floating-catalyst-method, in which organometallic
compounds (nickelocene (NiCp2), cobaltocene (CoCp2), ferrocene (FeCp2)) are
mixed with alcohols or hydrocarbons2,3. This method allows the production of
SWNTs in a single-step process, although it presents disadvantages too, like the
low yield and by-products (metal nanoparticles) embedded in the SWNT material.

2.1 Description of the synthesis method, materials and equipment

At IPICYT, the technique used to produce SWNTs is aerosol-assisted-CVD at

atmospheric pressure, based on the decomposition of ferrocene-ethanol solutions
(where the decomposition of ferrocene provides the catalytic iron particles).
The equipment used for aerosol production is an ultrasonic sprayer, which consists
of a glass container where the solution is placed; it contains a piezoelectric
plugged into a variable frequency power supply that enables us to tune the

29
 2.1 Description of the synthesis method, materials and equipment

frequency of vibration in order to generate a vapor cloud of the solution. A special

piece of glass connects the glass container to an end of a quartz tube placed
inside a tubular furnace (as depicted in Figure 2.1), the other end of the quartz tube
is connected to a condenser and an acetone trap for gas release.

Figure 2.1 Diagram of the experimental setting used in the synthesis of single-wall carbon
nanotubes

During the experiment the furnace is set to the desired temperature and an inert
atmosphere is maintained by flowing Ar gas or an Ar-H (95 % - 5 %) gas mixture.

Intense research carried out at

IPICYT by preceding peers allowed
us to know the optimal parameters for
SWNTs production. A concentration
of 1.25 wt. % of ferrocene in ethanol
at 950 ºC in an argon atmosphere
(1.2 l/min) for 30 minutes, yields the
best sample according to Lupo and
co-workers3 (see Figure 2.2).
The SWNTs are deposited in the
zone outside the furnace, where an
Figure 2.2 a) and b) representative TEM important temperature gradient is
images of CVD-produced pristine SWNTs at
found.
IPICYT. Adapted from Reference 3.

30
 2.1 Description of the synthesis method, materials and equipment

After the synthesis, the quartz tube is carefully disconnected and taken out of the
tubular furnace; a metal tool is used to extract the material from the tube walls, as
illustrated in Reference 3.

2.2 CVD production of doped single-wall carbon nanotubes

Tailoring the electronic structure of materials by adding impurities is long known by

the semiconductor industry. P-type and n-type doping of carbon nanotubes is
possible if we substitute carbon atoms with other atomic species containing less or
more electrons, respectively.
The most common dopants of carbon nanotubes have been boron4,5,6,7 and
nitrogen8,9,10,11,12,13, which are first neighbors of carbon, in the III and V groups in
the periodic table, respectively.
Nitrogen doping in multi-wall carbon nanotubes (MWNTs) induces the so-called
“bamboo-like” morphologies5-10 and theoretical calculations of substitutional
nitrogen doping in single-wall carbon nanotubes (SWNTs) have revealed a
different electronic structure, by introducing donor-like features into the conduction
band9. The production of N-doped SWNTs has been achieved by arc discharge11
and through CVD by adding benzylamine to a ferrocene-ethanol solution13. In this
work we demonstrate the use of an alternative precursor (pyrazine) to synthesize
nitrogen-doped SWNTs.
In the literature we can find experimental reports related to phosphorous-nitrogen
hetero-doping of multi-wall carbon nanotubes14. Theoretically, the substitutional
doping of SWNTs by phosphorous atoms is energetically possible15 and in this
section we will describe the experimental conditions to synthesize phosphorous-
doped SWNTs16.
Silicon doping of fullerenes17 and hetero-fullerene18 structures was reported
experimentally in the late 90’s. The incorporation of silicon species in the
hexagonal lattice of SWNTs was proposed theoretically by Baierle et. al. in 200119.
These calculations showed that substitutional silicon-doping of SWNTs introduces

31
 2.2 CVD production of doped SWNTs

donor-like states above the Fermi level. To the best of our knowledge, the
synthesis of Si-doped SWNTs has not been reported hitherto.
The role of sulfur in the CVD synthesis of carbon structures has been investigated
and it seems to have a major effect on the catalytic activity of iron. It increases the
production of carbon fibers20, enhances the synthesis of DWNTs21 and it is
extremely important in the production of carbon nanotube fibers produced by A.H.
Windle and co-workers22,23.

In this section we will describe the CVD-synthesis of SWNTs in the presence of

sulfur, nitrogen, phosphorous and silicon as doping species.
We will present the results of an electron microscopy characterization, which
allowed us to image the produced SWNTs and by-products, and of Raman
spectroscopy characterization, which through the RBM signal confirmed the
presence of SWNTs in each of the experiments performed. Some chemical
analysis techniques were used, like electron energy loss spectroscopy (EELS),
energy dispersive X-ray analysis (EDX).
Although the expected doping level in our samples is below the detection limit of
most elemental analysis techniques (e.g. <1 at. %), we will use the sensitivity of
Raman spectroscopy to probe electronic changes induced by the introduction of
foreign atoms in the hexagonal carbon network through the careful inspection of
the G’ band spectra of the synthesized materials. The appearance of the G’Def
peak, induced by negatively charged defects, and its relative intensity compared to
the G’Pris peak ( IG'Def /IG'Pris ) will provide an insight into the effect of the doping atoms

within the samples24. A comparative detailed analysis will indicate the effects of
the different doping precursors in SWNTs.

We use an aerosol-assisted CVD method with a floating catalyst based on

ferrocene (Fe(C5H5)2), ethanol (C2H5OH) solutions; and we introduce the doping
element by means of an organic compound containing it (hereafter such compound
will be called the precursor compound).

32
 2.2 CVD production of doped SWNTs

We designed several experiments motivated by the work of Villalpando-Paez et

al.13 and we have also reproduced their results of nitrogen-doped SWNTs using
benzylamine as a precursor.
The precursors were chosen based on their physical properties (melting point and
most of all, good solubility in ethanol). Table 2.1 describes the chemical structure
and physical properties of the precursors that we worked with.
For each of the specific doping elements a solution of ferrocene (Fe(C5H5)2),
ethanol (C2H5OH) and the precursor compound was prepared, and for each
precursor several concentrations were tried.
The concentration of ferrocene (as source of iron) was kept constant for all of the
experiments reported here, at 1.25 % by weight. An aerosol of the solution was
generated using the ultrasonic sprayer.
Argon (or an Ar 95 % - H 5 % mixture) was used as a carrier gas inside a quartz
tube to direct the aerosol to the hot zone of a tubular furnace operated at 950 °C
(during heating and cooling down the gas flow rate was set to 0.2 l/min). The flow
rate for each experiment is specified in Table 2.2.

Compound name Chemical Formula Physical m.p. b.p.

form
Structure

Thiophene C4H4S liquid -38ºC 84ºC

Benzylamine C7H7NH2 liquid 10ºC 185ºC

Pyrazine
C4H4N2 solid 50-56 º C 115-116 ºC

Triphenylphosphine (C6H5)3P solid 79-81ºC 377ºC

Methoxytrimethylsilane CH3OSi(CH3)3 liquid 57-58ºC

Table 2.1 Chemical structure, formula and physical properties of the precursor compounds used in
the doping experiments (Data extracted from the MSDS at www.sigmaaldrich.com)

33
 2.2 CVD production of doped SWNTs

After 30 minutes the aerosol generator was turned off, the system was allowed to
cool down to room temperature and the quartz tube was taken out of the furnace.
A web-like material containing SWNTs and by-products was collected from the
zone outside the furnace.

Nitrogen-doped (N-doped) SWNTs were synthesized reproducing the results

reported by Villalpando-Paez and co-workers13 using benzylamine (C7H7NH2) as N
precursor in the solution at concentrations of 7 wt. % and 11 wt %. Pyrazine
(C4H4N2) was also used as N precursor at concentrations of 0.5 wt. %, 1 wt. %, 1.5
wt. % and 2.5 wt. %.
Phosphorous was inserted to the system using triphenylphosphine (P(C6H5)3) at
concentrations of 0.1 wt. %, 0.15 wt. %, 0.2 wt. % and 0.25 wt. %.
The production of SWNTs in the presence of silicon was achieved by adding
different concentrations of methoxytrimethylsilane (CH3OSi(CH3)3) as Si precursor.
The concentrations used were 0.05 wt. %, 0.1 wt. % and 0.2 wt. %.
In the case of sulfur, we used thiophene (C4H4S) as precursor compound at
concentrations of 0.1 wt. %, 0.12 wt. %, 0.15 wt. % and 0.25 wt. %.

The conditions (temperature, carrier gas and flow rate) for the different
experiments are summarized in Table 2.2.

Doping Concentration Synthesis Carrier Flow

Precursor compound Duration
element (% by weight) temperature gas rate
Thiophene 0.1, 0.12, 0.15,
S 950 ºC 30 min Ar 1.2 l/min
C4H4S 0.25
Benzylamine
N 3, 7, 11 950 ºC 30 min Ar 1.2 l/min
C7H7NH2
Pyrazine
N 0.5, 1, 1.5, 2.5 950 ºC 30 min Ar-H 1.2 l/min
C4H4N2
Triphenylphosphine 0.1, 0.15, 0.2,
P 950 ºC 30 min Ar-H 0.8 l/min
P(C6H5)3 0.25
Methoxytrimethylsilane
Si 0.05, 0.1, 0.2 950 ºC 30 min Ar 0.6 l/min
CH3OSi(CH3)3

Table 2.2 Summary of the precursor compounds and experimental conditions used in the synthesis
of SWNTs in the presence of sulfur, nitrogen, phosphorous and silicon

34
 2.2 CVD production of doped SWNTs

It is worth mentioning that the incorporation of thiophene (even at the lowest

concentration) changed radically the macroscopic appearance of the collected
material. For pristine nanotubes, the web-like material containing SWNTs is
blackish and very fragile, and its extraction from the quartz tube is a delicate task.
When we add thiophene to the ethanol-ferrocene solution, the resulting material is
darker, more abundant and less fragile.
We point out that the nitrogen- and phosphorous- doped materials did not show
remarkable macroscopic differences from their pristine counterparts.
On the contrary, the collected materials that resulted from MTMS addition, showed
a very different appearance. The overall color changed to gray tones and the
material became more flake-like and less web-like.

2.3 Characterization of CVD-produced doped SWNTs

The electron microscopy characterization of the materials was carried out using a
scanning electron microscope (FEI XL-30 SFEG-STEM) operated at 10-15 kV and
a transmission electron microscopy (TEM) was performed on a FEI TECNAI F30
STWIN operated at 300 kV and a (ASU microscope).
The samples were mounted on carbon tape for SEM observation and prepared on
copper grids for STEM and TEM observation. Elemental composition techniques,
such as energy dispersive X-ray analysis (EDX) and electron energy loss
spectroscopy (EELS), were also used to supplement the electron microscopy
characterization.
Micro-Raman spectroscopy measurements of the bulk samples were recorded at
room temperature using a Renishaw InVia equipment. The spectra were recorded
with an Ar line λ=514.5 nm (Elaser=2.41 eV) and a He-Ne λ=633 nm (Elaser=1.95
eV) at a power ~0.3 mW in a back-scattering geometry using a 100x objective lens
to focus the laser beam, using the facilities at LINAN. Analysis of spectra recorded
with many laser excitation energies was carried out at the Raman Laboratory of
Carbon Nanotubes at MIT facilities (λ=532 nm, λ=647 nm and λ=676 nm), and at
the Raman Laboratory of the Universidade Federal de Minas Gerais (UFMG) in

35
 2.3 Characterization of CVD-produced doped SWNTs

Brazil (mapping of the RBMs of the phosphorous-doped samples). No less than

three measurements were recorded per sample, the spectra shown here are the
resulting averages. The sample was mounted on a double-sided scotch tape fixed
on corning glass, and no further preparation of the sample was required.

2. 3. 1 Electron microscopy characterization

Sulfur case
Our microscopy characterization revealed that experiments with low concentration
of thiophene (0.1 - 0.25 wt. %) produced arrays of SWNTs entangled with short,
twisted multi-wall nanotube-like nanostructures, see Figures 2.3 a-d.
In Figure 2.3e, a high magnification SEM image of the by-products is shown. The
high-angle annular dark-field scanning transmission electron microscopy (HAADF-
STEM) image of the same area, depicted in Figure 2.3f, reveals a high contrast at
the tips, suggesting a metallic nature of nanoparticles present at the tips of the
short tubes.

Figure 2.3 SEM images of the material synthesized with thiophene at a) 0.15 wt. % and b) - f) 0.1
wt. %. In figures a), b) and c), single-wall nanotubes appear like threads accompanied by short,
twisted MWNT-like nanostructures. These by-products are imaged at higher resolution in d) and e).
A HAADF-STEM image of e) is shown in f).

36
 2.3 Characterization of CVD-produced doped SWNTs

In order to investigate the structure of the by-products we performed HRTEM of the

materials. We can find a low magnification TEM image in Figure 2.4a, where we
confirm the metallic nature of the nanoparticles at the tips of the short tubes by a
higher contrast image.
In Figure 2.4b an image of two SWNTs and by-products is presented. A multi-wall
nanotube is shown at high magnification in Figure 2.4c, we can see that the body
of the tube is constituted of graphitic layers (not very crystalline) arranged in a
compartment bamboo-like structure catalyzed by the metallic particle at the tip. A
better resolved image of the walls of a different short tube is shown in Figure 2.4d.

Figure 2.4 TEM images of the material synthesized with thiophene. 0.12 wt. % thiophene (a and c)
and 0.25 wt. % of thiophene (b and d).

The carbon composition of the walls was confirmed by high resolution energy
dispersive X-rays (EDX) point measurements on the wall of the tubes, see Figure

37
 2.3 Characterization of CVD-produced doped SWNTs

2.5. In this figure we can also see an EDX point measurement at the metallic tip,
the signal reveals that the nanoparticle is composed of Fe, S and O. The Cu signal
detected comes from the copper grid that the sample was mounted on.

Figure 2.5 High resolution EDX point measurements of a short multi-wall nanotube. At the left,
HAADF-STEM images depicting the zone of measurement by a small-red circle; at the right, the plot
of the corresponding recorded spectra.

Nitrogen case
We reproduced the results published by Villalpando and co-workers13, where
benzylamine was used as precursor compound for nitrogen-doped SWNTs.

Figure 2.6 SEM images of nitrogen-doped SWNTs produced with benzylamine at concentrations of
a) 3 wt. %, b) 7 wt. %, and c) 11 wt. %.

Our SEM observations (Figure 2.6) show that the materials synthesized with
benzylamine contain arrays of SWNTs and small metallic nanoparticles, the

38
 2.3 Characterization of CVD-produced doped SWNTs

amount of these metallic nanoparticles present in the sample increase with

increasing benzylamine concentration and the yield of the produced SWNT
decreases.

Figure 2.7 a), b), c) and d) SEM images of SWNTs synthesized in the presence of pyrazine at
concentrations of 0.5, 1, 1.5 and 2.5 % by wt., respectively. e) and f) two different spots on the
material at 1.5 % by wt. of pyrazine imaged by HRTEM.

39
 2.3 Characterization of CVD-produced doped SWNTs

Recent extensive characterization of these samples through X-ray photoelectron

spectroscopy (XPS)25 has revealed that the synthesized N-doped nanotubes
exhibit a maximum concentration at 0.3 at. % of N. Therefore, EDX or EELS are
not able to detect such low concentrations and Raman characterization becomes
the most sensitive tool to observe the doping effects in SWNTs.

For the production of N-doped SWNTs another precursor was used: pyrazine.
The chemical formula of pyrazine (see Table 2.1) indicates that it contains two
atoms of nitrogen per molecule; we believe that the doping of SWNTs could be
more effective due to this immediate availability of N atoms during the synthesis;
see Figure 2.20. The introduction of this compound in the synthesis of SWNTs led
to the production of nanotubes and iron nanoparticles very similar to the un-doped
counterpart (see Figure 2.2 and 2.7).
Although elemental analysis techniques were used to try to detect N traces in the
walls of the tubes, we failed to find a N signal in any of the samples. However
Raman spectroscopy will prove the tubes were doped.

Figure 2.8 SEM images (left panels), and bright and dark field STEM images (center and right
panels) of the samples synthesized with triphenylphosphine.

40
 2.3 Characterization of CVD-produced doped SWNTs

Phosphorus case
The introduction of triphenylphosphine (TPP) led to the synthesis of SWNTs and
nanoparticles as depicted in Figure 2.8, where SEM and STEM images of the
samples synthesized with 0.1 % by wt. (top panels) and 0.25 % by wt. (bottom
panels) of TPP are presented.
The nanoparticles entangled with the nanotubes, observed in the SEM images (left
panels of Figure 2.8) are clearly heavier when imaged by bright field and HAADF
STEM (center and left panels in figure 2.8, respectively), showing more contrast
than the nanotubes composed of only carbon.
The metallicity of the particles was further confirmed by the observation of lattice
fringes during TEM observation (images not shown here).
A high proportion of SWNTs present in this sample confirms the catalytic nature of
the present metallic nanoparticles.
More details on the morphology of the synthesized materials at the four different
concentrations can be observed on Figure 2.9, where SEM and TEM images of the
materials are presented.

Figure 2.9 SEM images of SWNTs and by-products deposited over copper grids synthesized with
a) 0.1 % by wt., b) 0.15 % by wt., c) 0.2 % by wt. and d) 0.25 % by wt. of TPP; d) and e) TEM
images of the material synthesized with 0.2 and 0.25 % by weight. Scale bars in a) - d) represent
200 nm. Scale bars in e) and f) represent 10 nm and 5 nm, respectively.

41
 2.3 Characterization of CVD-produced doped SWNTs

SEM images of the different concentrations are included in Figure 2.9 a-d, a good
quality of the material and a low proportion of by-products are observed in all of
these cases. Figures 2.9 e and f show HRTEM images of the 0.2 wt. % and 0.25
wt. % samples, we can confirm a good quality and crystallinity of the produced
nanotubes.

Silicon case
Experiments with methoxytrimethylsilane (MTMS) resulted in the formation of
SWNT bundles mixed with by-products, which were further analyzed to determine
their morphology and composition.
Figure 2.10 shows representative TEM images of the nanostructures produced. At
the lowest concentration of MTMS (0.05 % by wt.) catalytic Fe nanoparticles were
present, embedded in the SWNT bundles, very similar to the pristine material (see
Figure 2.2 and Figures 2.10 a and b).
Figure 2.10c shows the material synthesized at 0.1 % by wt. of MTMS, the upper
panel depicts a low magnification image and the lower one a higher magnification
image. Using this MTMS concentration, the SWNTs appear with spherical
nanoparticles clearly embedded in a dense matrix, our elemental composition
analysis helped to investigate the nature of the matrix and spherical
nanostructures.
EELS and HAADF STEM (high-angle annular dark-field scanning transmission
electron microscopy) analysis were performed on this sample. The single-wall
carbon nanotubes and bundles were very difficult to image in the HAADF mode.
They have very low contrast compared to the SiO2 and Fe in the matrix. The
contrast had to be minimized in order to see the tubes, which is why the matrix
material appears washed out in the images of Figure 2.11. The point spectra
shown on the left panels of Figure 2.11 confirm the tubes are carbon (284 eV).
Linescans were not possible due to nanotube movement. We gathered EELS
spectra from single points on several tubes and never found a Si signal (99 eV). If
there is any Si in the tubes, it is below the detectable limit of EELS.

42
 2.3 Characterization of CVD-produced doped SWNTs

Figure 2.10 High Resolution Transmission Electron Micrographs of SWNTs produced with different
concentrations of methoxytrimethylsilane, a) pristine SWNTs, b) 0.05 % by weight, c) 0.1 % by
weight and d) 0.2 % by weight, the arrow in the low panel points to a SWNT embedded in the by-
products. In b), c) and d) the upper panels contain low magnification images and the lower panels
contain higher magnification images, picturing the carbon nanotubes and the morphology of the
accompanying by-products.

43
 2.3 Characterization of CVD-produced doped SWNTs

Analysis of the matrix material is also shown in Figure 2.11. The top spectrum at
the right of the image covers the range for Si (99 eV) and C (284 eV). The bottom
spectrum covers the range for O (532 eV) and Fe (708 eV). As the spectra show,
the matrix is primarily Si and O. We periodically found C in the matrix, which
probably belongs to carbonaceous by-products also present in the sample.
Point spectra from the bright, round particles in the matrix confirm that they are Fe.
Usually weak O, Si, and C peaks also appear in the spectra, but only because the
particle is sitting on or embedded in the silicon oxide matrix.

Figure 2.11 HAADF-STEM images and EELS point spectra of the sample synthesized with 0.1 wt.
% of MTMS. Left panels correspond to point measurements on SWNTs. Top right panel, point
measurement on the matrix, and bottom right panel, point measurement on a metal nanoparticle.

In the Fe-O spectrum below, there is also a weak O signal. There were also weak
Si and C signals at this same point, but we were able to make them disappear by
moving the beam to a region of the particle that wasn’t over the matrix. We thus
conclude that these are pure Fe particles embedded in the Si-O matrix. This is

44
 2.3 Characterization of CVD-produced doped SWNTs

important because for the sample synthesized with a higher MTMS concentration
(0.2 wt. %), the nature of the Fe particles change.
A higher concentration of silicon (0.2 % by wt. of MTMS), promotes the formation
of strikingly different nanostructures and SWNTs. In this sample, short nanorods
with metallic hemi-spherical tips can be found (see Figure 2.10d). In these
samples, it was difficult to find numerous SWNTs, nevertheless, in the lower panel
of Figure 2.10d an individual SWNT can be observed (pointed to by an arrow)
along with the by-products synthesized for this MTMS concentration. We believe
that the formation of pure Fe particles, which serve as SWNT catalysts during the
growth, was not favored at these experimental conditions. The formation of binary
Fe-Si alloys is highly probable at this synthesis temperature; hence, the catalytic
activity of Fe was reduced and thus the formation of nanotubes was not abundant.

O Si
Fe C

Figure 2.12 HAADF-STEM images and EELS line scans on tips of the nanorods, a) Fe-O EELS
profile with before and after images, the tips are primarily Fe but they also contain a significant
portion of O. b) Si-C EELS profile with before and after images. The Si signal gets progressively
stronger as the beam moves from vacuum, through the metal tip, and into the matrix material. The
C signal from the front of the tip is a contamination artifact and not typical of tip chemistry. The
profile shows that the Fe tips contain a significant amount of Si.

45
 2.3 Characterization of CVD-produced doped SWNTs

However, due to the high sensitivity of Raman spectroscopy to identify SWNTs,

their presence was confirmed by the observation of a RBM signal (see Figures
2.17 and 2.19).
While doing EELS and HAADF imaging, the sample was very unstable but we
found a region that was relatively stable under the beam. We still observed some
drift, but tried to compensate for it with the spatial drift correction in the Digital
Micrograph. The matrix material is similar to that of a sample synthesized with 0.1
wt. % MTMS, except for the presence of the small rods with hemispherical tips.

We performed a few line scans on the tips; see Figure 2.12, which were marginally
successful. Because of the drift, we could only do 5-8 points per scan, so the
profiles are pretty rough but they still make the point. Due to the width of the EELS
window, we were not able to capture the Fe and Si signals simultaneously; so on
some tips we scanned for Fe and O (Figure 2.12a) , and on others we scanned for
Si and C (Figure 2.12b).

The sets of images in Figure 2.12 contain EELS profiles with before and after
images. The image taken after the EELS profile was captured to show drift (which
in the last two cases we attempted to correct) and beam damage.
It is clear from both the image contrast and the EELS spectra that the tips contain
Fe. But the tips also contain a significant amount of Si and O, and we never found
a tip that was pure Fe.
Note in the images of Figure 2.12 how the beam drilled through the semi-spherical
Fe particles. From our experience with imaging metals, pure Fe is relatively stable
under the beam. This damage suggests intermixing of the Fe with Si and/or O.

In general, from our electron microscopy characterization two things are worth
mentioning. First of all, the experiments carried out, successfully produced SWNT
material synthesized in the presence of S, N, P and Si for all the concentrations
reported in Table 2.2. Our second conclusion is related to the nature of the by-
products. The presence of sulfur and silicon dramatically modified the morphology

46
 2.3 Characterization of CVD-produced doped SWNTs

and structure of the accompanying materials; while phosphorous and nitrogen did
not induce changes in the morphology of the nanoparticles when comparing them
to the by-products obtained in the production of pristine SWNTs.

2. 3. 2 Raman spectroscopy characterization

Raman spectroscopy is a technique very sensitive to the electronic structure of

carbon materials. A complete appendix to the Raman spectroscopy of carbon
materials is included in this work (Appendix B), in such pages the reader will find a
description of the double resonance process, first and second order Raman
scattering, characteristic nanotube vibrations (radial breathing mode and tangential
modes), disorder induced modes and information related to the Raman analysis
that will be discussed in future paragraphs.

Sulfur case
We have performed Raman scattering measurements with Elaser = 2.41 eV (λ =
514.5 nm). As mentioned before, in a typical routine of Raman spectroscopy, two
or three measurements are recorded per sample and the average of the spectra is
reported. When doing the analysis of these samples we recorded several spectra
per sample. We noticed that for the same sample the shape of the signal changed
from one spot to the other. Figure 2.13 depicts spectra recorded at two different
spots of the sample synthesized with 0.1 wt. % of thiophene.
In Figure 2.13a we show the spectrum of a zone where the carbon-related signal
was found above 1000 cm-1 (D, G, D’, G’ and D+G bands). It is worth mentioning
that at low-frequencies other modes are clearly present (see inset in Figure 2.13a).
Care should be taken when assigning these modes. Since we are dealing with
SWNTs, a first impulse would be to interpret them as RBM peaks. Nevertheless in
the spectrum there are no other SWNT-related Raman features. For example, the
high intensity of the disorder-related modes (D, D’ and D+G bands) is not
characteristic of SWNTs (compare to cyan line in Figure 2.13b), and we do not
observe a multi-peak feature around ~1580 cm-1 originating from the quantum

47
 2.3 Characterization of CVD-produced doped SWNTs

confinement of the wavevector along the circumferential direction (see Appendix

B). Figure 2.14a contains a plot of the RBM region with spectra of all the different
concentrations of thiophene. The yellow lines are guides to the eye to show that
the two main peaks are located at the same frequencies for all the concentrations
(~219 cm-1 and ~285 cm-1).

Figure 2.13 Raman spectra of material synthesized with 0.1 wt. % of thiophene. a) spectrum
showing sulfur-related low frequency features (see inset) and carbon-related features (D, G, D’, G’
and D+ G bands). b) spectrum of nanotubes present in the sample (blue line) showing RBM signal,
a detail on the G’ band region is depicted in the inset. We include the spectrum of pristine SWNTs
heat treated at 400 ºC (cyan line) for comparison

When we analyze the RBM signal frequencies in relation to the laser excitation
energy used (Elaser), we should be able to assign the ωRBM to a certain (n,m)
chirality (see Introduction). However, it is clear from Figure 2.14b that the
intersection points between the green and yellow lines are deserted, indicating that
no chirality is excited at this energy at those particular frequencies; hence the
signal can not be attributed to resonance with carbon nanotubes.
When we correlate these results with the information obtained from the elemental
composition characterization (Figure 2.5), we find that the signal at low frequencies
belongs to sulfur26,27 (peaks around 216 cm-1 and 285 cm-1), which is indeed
present in the by-products. Thus, we conclude that the spectrum shown in Figure
2.13a has a contribution primarily from the bamboo-like multi-wall nanotubes with

48
 2.3 Characterization of CVD-produced doped SWNTs

metallic tips. The low crystallinity of the walls (Figure 2.4d) is responsible for the
high intensity of the disorder induced modes (D, D’ bands) and the sulfur content is
responsible for the features found at low-frequencies.
In Figure 2.13b we plot the spectrum of a zone were contributions of SWNT and
by-products can be identified from the sample with 0.1 wt. % of thiophene. Again,
contribution from the carbonaceous by-products is associated with the high
intensity of the D band and the presence of the D’ band. However, in this spectrum
a single RBM peak was found, which denotes the presence of SWNTs; besides,
the splitting of the G’ band (G’Def and G’Pris bands, see Appendix B) implies n-
doping due to the presence of electron-donor substitutional sulfur atoms.
Unfortunately this last conclusion can not be confirmed because we were not able
to reproduce such a spectrum in other spots of the sample.

Figure 2.14 a) RBM region showing spectra of materials synthesized with thiophene, the spectrum
-1 -1
of pristine SWNTs is also included. In this plot, the yellow lines at 219 cm and 285 cm high-light
28
the position of the peaks present for all the concentrations of thiophene. b) Katarua plot , the green
line is placed at 2.41 eV (energy corresponding to laser λ=514.5 nm), the yellow lines are placed at
the same frequencies as in a).

The stories for the samples with higher content of thiophene (0.12, 0.15 and 0.25
wt. %) are very similar. The gray line in Figure 2.15 represents a typical spectrum

49
 2.3 Characterization of CVD-produced doped SWNTs

of the 0.15 wt. % sample, where no RBM signal was found and the existing
features are attributed to the other sp2 carbons present in the sample. Note the
shape of the G band, isolated in the inset, no multi-peak features are present and
the D’ band intensity is high, while in the pristine SWNT spectrum we find a G- and
G+ splitting of the G band and the D’ band is absent.

Figure 2.15 Plot of Raman spectroscopy of materials synthesized with thiophene (black and gray
lines). The spectrum of pristine nanotubes is included for comparison (cyan line). A detail on the G
band region is pictured in the inset.

We can conclude from the Raman analysis of the thiophene-produced samples

that, although the microscopy characterization showed clearly and abundantly the
presence of single-wall nanotube species, the recorded spectra did not show
predominant features of SWNTs. On the contrary, the by-products signal
overcame the nanotube signal, making it difficult to extract information on the effect
of substitutional sulfur-doping of the single-wall carbon nanotubes.

The discussion of the Raman characterization of the rest of the doped-nanotube

samples (synthesized with TPP, benzylamine, pyrazine and MTMS) will be done as
a whole, since in all the cases we were able to record good Raman spectra.

50
 2.3 Characterization of CVD-produced doped SWNTs

The presence of SWNTs in all our materials was positively proved by electron
microscopy images(see Figures 2.3 - 2.11) and by Raman spectroscopy traces,
through the RBM signal.

Nitrogen, phosphorus and silicon

In Figure 2.16 we include a plot of the RBM of the samples synthesized with TPP,
pyrazine, benzylamine, MTMS recorded with five different laser excitation energies:
2.41 eV (λ=514 nm), 2.33 eV (λ=532 nm), 1.96 eV (λ=633 nm), 1.91 eV (λ=647
nm), and 1.83 eV (λ=676 nm).

Figure 2.16 Raman radial breathing mode (RBM) spectra of the different doped samples recorded
with five laser excitation energies: 2.41 eV (λ=514 nm), 2.33 eV (λ=532 nm), 1.95 eV (λ=633 nm),
1.91 eV (λ=647 nm), 1.83 eV (λ=676 nm)

51
 2.3 Characterization of CVD-produced doped SWNTs

It has been theoretically and experimentally demonstrated that the incorporation of

heavier elements in the hexagonal carbon lattice is energetically favored in
narrower nanotubes which exhibit higher radii of curvature.15,16,29
In our experiments, the production of narrower nanotubes when N and Si atoms
are substitutionally doping the SWNTs is illustrated in Figure 2.17, where the RBM
region of the Raman spectra recorded at Elaser=1.96 eV is plotted. The red arrows
help to visualize the RBM features that significantly decrease in intensity as the
precursor concentration in the ferrocene-ethanol solution is increased. We could
correlate this decrease to the absence of large diameter doped SWNTs.

Figure 2.17 RBM spectra recorded at Elaser=1.96 eV of a) N-doped SWNTs synthesized with
different concentrations of benzylamine and pyrazine and b) Si-doped SWNTs synthesized with
different concentrations of MTMS. The spectra were normalized to the G band and the red arrows
point to the RBM features that decrease as the precursor concentration is increased.

We carried out a many laser excitation energies study of the phosphorous-doped

SWNT samples. Mapping of the RBMs as a function of the excitation laser energy
was performed with excitation energies ranging from 1.61 to 2.71 eV. An Ar-Kr
laser (1.91, 2.18, 2.33, 2.41, 2.47, 2.49, 2.53, 2.60, 2.62 and 2.71 eV) and a dye
tunable laser (16 lines from 1.80 to 2.01 eV, 16 lines from 2.03 to 2.19 eV and 5
lines from 2.23 to 2.29 eV) were used. The RBM peak position is related to the
tube diameter by
227
1 + Cd t
2
ωRBM = (1)
dt

52
 2.3 Characterization of CVD-produced doped SWNTs

where ωRBM is the RBM frequency shift in cm-1, dt is the tube diameter in
nanometers, and C is a constant factor accounting for environmental effects. By
measuring the RBMs with different laser energies, it is possible to build a two-
dimensional map of the RBM profiles as a function of the excitation laser energy.
Figure 2.18 shows such maps in the 1.9-2.3 eV energy range for the 0, 0.1, 0.15,
0.2 and 0.25 wt. % of P-doped SWNT samples. The maps show the resonance
patterns related to different (2n+m)=constant families belonging to the ES22, EM11,
ES33, transition energies, (n,m) being the nanotube indices. No ωRBM or optical
transition energies (Eii) shift can be seen from the maps for different doping levels,
thus indicating that the local nature of P doping does not change either ωRBM or Eii
of the SWNTs.
Although the RBM Raman cross section depends on (n,m) and Eii, a qualitative
description of the diameter distribution of the samples can be obtained by adding
all the RBM spectra for each sample, as shown in Figure 2.18f. The diameters vary
from 0.7 to 2.2 nm for the different samples. However, the samples with higher
doping levels (0.2 wt. % and 0.25 wt. %) have larger RBM intensities with higher
RBM frequencies, meaning that either the presence of phosphorous induces the
formation of tubes with smaller diameters or the relative Raman scattering of
narrower tubes has been enhanced by doping.
In both cases, this result and the two-peak G’ feature (see G’ band analysis in
future paragraphs), confirm substitutional phosphorous doping.
Finally, this accurate information on the diameter distribution, based on the ωRBM
analysis, was used to analyze the ωG’ peaks and we can rule out any diameter or
resonance effects as responsible for the behavior of the ω G'Pris − ω G'Def splitting.

The change in the RBM intensity toward high ωRBM frequencies above 0.15 wt. % of
TPP indeed confirms the effect shown by the ωG’ analysis, where doping seems to
be more effective above this doping level.

53
 2.3 Characterization of CVD-produced doped SWNTs

Figure 2.18 (a-e) Radial breathing mode (RBM) resonance Raman maps of the phosphorous-
doped SWNT samples with 0, 0.1, 0.15, 0.2, 0.25 wt. % of TPP, respectively. The optical
transmission energies (Eii) were inserted on the map of the undoped sample a) to indicate the (n,m)-
dependent RBM resonances by using equation (1), with the best fit to our data given by C=0.035. f)
Diameter distribution of the samples obtained by summing all the RBM spectra measured with the
different excitation laser lines. The upper x-axis was obtained by the inverse of the ωRBM–dt relation
cited above.

54
 2.3 Characterization of CVD-produced doped SWNTs

In Figure 2.19 we include a plot of the RBM and G’-band regions, recorded at
Elaser=2.41 eV, of the samples synthesized with TPP, pyrazine, benzylamine,
MTMS and of pristine SWNTs annealed at 400 °C (HT P ristine). This plot reveals
that SWNTs were successfully grown in the different environments and at every
single concentration used. We can also note the evolution of the G’Def peak as the
precursor concentration is increased. This increase in intensity is quantified by
means of the IG'Def /IG'Pris relative intensity.

Figure 2.19 Raman spectra of the synthesized materials, acquired with Elaser = 2.41 eV, showing the
RBM and G’ band regions. The spectrum of pristine SWNTs annealed at 400 ºC as described in
Reference 15 is included for comparison

We have performed a Raman spectroscopy comparative analysis, based on the

IG'Def /IG'Pris relative intensity, to quantify the effect of doping on the electronic and
vibrational structure of SWNTs, relatively among the different precursors.

A two-peak Lorentzian fitting of the G’ band of all the spectra recorded at

Elaser=2.41 eV was carried out, according to previous studies16,30. This procedure
has to be made with care, since the G’Pris frequency depends on tube diameter,
which changes upon doping. The ratio of the relative intensities of the G’Def band

55
 2.3 Characterization of CVD-produced doped SWNTs

and G’P band ( IG'Def /IG'Pris ) was computed, the resulting values are plotted in Figure

2.20 as a function of doping atoms available per carbon atom during the synthesis
in the precursor-ferrocene-ethanol solution.
For all of the precursors, an increase in the number of doping atoms per carbon

atom available in the synthesis results in the increase of the IG'Def /IG'Pris relative

intensity, suggesting higher doping levels as we increase the precursor

concentration.

Figure 2.20 Plot of the IG'Def /IG'Pris relative intensities as a function of doping atoms (P, N and Si)
per carbon atoms introduced in the synthesis environment (calculated from the precursor
concentration. Such concentrations are also included next to each corresponding symbol), the solid
squares represent the set of P-doped samples synthesized with triphenylphosphine, the solid circles
and solid triangles represent the N-doped samples synthesized with pyrazine and benzylamine,
respectively, and the solid inverse-triangles stand for the Si-doped materials synthesized using
methoxytrimethylsilane as a precursor.

The lowest IG'Def /IG'Pris G'Pris relative intensity values were obtained for the Si-doped

samples. However, these are also the samples synthesized with the lowest amunt
of doping atoms per carbon atom. We are confident that higher Si doping will
56
 2.3 Characterization of CVD-produced doped SWNTs

result from the use of higher concentrations of MTMS during the synthesis.
However, other experimental conditions should be varied to avoid the suppression
of the catalytic effect of Fe when Fe-Si alloys are formed. We believe that the
increase of flow rates and temperatures could speed up the reactions and promote
the SWNT formation while the catalyst particle is active.

On the contrary, although the highest concentration of benzylamine used in this

work (11 wt. %) provided the highest number of doping atoms per carbon atom

during the synthesis, the IG'Def /IG'Pris relative intensity values are below those of 1.5

wt. % and 2.5 wt. % of pyrazine, which proves a more efficient N-doping when
pyrazine is used as the nitrogen precursor.

Figure 2.21 G’ band Raman spectra and the corresponding Lorentzian two-peak-fitting of the G’

band (G’Def and G’Pris bands) of the samples with maximum IG'Def /IG'Pris relative intensities for each

precursor (corresponding to the highest concentrations used in this work). The spectrum of
annealed pristine SWNTs at 400 ºC is included for comparison. The blue and cyan lines represent
the Lorentzian fitting of the G’Def and G’Pris bands, respectively.

57
 2.3 Characterization of CVD-produced doped SWNTs

XPS analysis of N-doped benzylamine SWNTs proved that the content of nitrogen
reaches a maximum at 0.3 at. %, and our Raman results indicate that pyrazine-
produced N-doped nanotubes could surpass this limit. A careful XPS
characterization of our pyrazine N-doped materials is underway.

Further inspection of this figure leads us to conclude that the most efficient doping
is attributed to phosphorous, since the amount of P atoms available per C atom is

relatively small (below the amount of N atoms available) and still the IG' Def /IG'Pris

relative intensity shows its highest values.

However, it is not yet clear whether the IG'Def /IG'Pris relative intensities can be

directly compared for different dopants. As discussed in Ref. 16 and 25, different
atoms disturb the SWNT lattice differently. To address this point we selected the

samples with highest IG'Def /IG'Pris relative intensities from Figure 2.20 (i.e. 0.25 wt.

% of TPP, 2.5 wt. % of pyrazine, 11 wt. % of benzylamine and 0.2 % of MTMS).

In Figure 2.21 we show the Raman spectra in the G’ band region of the above
mentioned samples and their corresponding fitting. In this graph it is very easy to
visualize the G’Def bands (blue lines) and the G’Pris bands (cyan lines). As we go
from the pristine sample to the P-doped sample we see an increase in intensity of

the G’Def band, directly measured in the IG'Def /IG'Pris relative intensities.

58
 2.3 Characterization of CVD-produced doped SWNTs

Figure 2.22 Plot corresponding to the G’ band splitting ( ω G'Pris − ωG'Def ) of the P-, N- and Si-

doped SWNTs as a function of precursor concentration.

The information of the fitted spectra was further used to calculate the frequency

splitting ω G'Pris − ω G'Def between the two G’ peaks for the doped samples, see

Figure 2.22. From this figure it is clear that there is a range of values particular to
each doping element. For phosphorous the splitting values are in the range of ~31-
33 cm-1, for nitrogen the splitting values are in the range ~35-40 cm-1 and Si
reported the highest values, in the range of ~41-45 cm-1. These results suggest
that the G’ band splitting is more related to the nature of doping than to the doping
level.

59
 2.4 Conclusions

2. 4 Conclusions

The synthesis of single-wall carbon nanotubes in sulfur, nitrogen, phosphorous and

silicon environments was achieved via aerosol-assisted-CVD.
Precursors containing the target doping element were mixed in ethanol-ferrocene
solutions at different concentrations, thiophene for sulfur incorporation,
triphenylphosphine was used in the case of phosphorous, methoxytrimethylsilane
was used as a silicon precursor, and benzylamine and pyrazine were used as
nitrogen precursors.

Electron microscopy of the samples as well as RBM Raman signals confirmed the
presence of carbon nanotubes in the synthesized materials.

Electron microscopy analysis revealed that most of the materials consisted of

doped SWNTs entangled with metallic nanoparticles. In the case of S and Si,
interesting morphologies were formed, such as short bamboo-like MWNTs and Si
nanorods with metallic tips, respectively.

In particular, the incorporation of sulfur promoted a very dense formation of by-

products. The enhancement of the catalytic activity of iron by sulfur addition was
confirmed through the copious presence of short, twisted bamboo-like MWNTs
along with SWNTs. We believe that the modification of the experimental
parameters (e.g. reduction of ferrocene concentration) could yield to the
optimization of S-doped SWNTs.

Our RBM analysis showed that as the doping precursor was increased in the
sprayer solution, narrower diameter tubes were favored. The latter is consistent
with theoretical calculations indicating that dopants of heavier elements embedded
in the hexagonal carbon lattice are more energetically favored in narrower tubes,
exhibiting higher radii of curvature.

60
 2.4 Conclusions

We have used the IG'Def /IG'Pris relative intensities as a direct doping index for the

amount of N, P and Si incorporated in the structure. Our samples synthesized with

sulfur could not be included in this analysis due to the poor quality of the recorded
Raman spectra. Hence we do not have strong evidence for the sulfur doping of

SWNTs. Si-doped samples showed low IG'Def /IG'Pris relative intensity values which

are directly related to the small amount of silicon atoms available per carbon atoms
during the synthesis. Nitrogen doping is more effective when pyrazine is used
instead of benzylamine and phosphorous doping is very effective even at low TPP
concentrations.

Our Raman records show that the increasing precursor concentration leads to

higher doping levels, increasing the IG'Def /IG'Pris relative intensities, and that the

frequency splitting ω G'Pris − ω G'Def between the two G’ peaks depends more on the

doping element than on the doping amount. We are carrying out more synthesis
experiments in order to have a higher number of samples per doping element, to
be able to run Raman spectroscopy measurements and elucidate the nature of the
G’ band splitting peculiarities observed in this work.

61
 2.5 Related Articles

2.5 Related Articles

1 . J. Campos-Delgado, I.O. Maciel, D. A. Cullen, D. Smith, A. Jorio, M.A.

Pimenta, H. Terrones, M. Terrones. Chemical vapor deposition synthesis
of N-, P-, and Si- doped single-wall carbon nanotubes (Submitted to ACS
Nano).

2 . A.L. Elías, P. Ayala, A. Zamudio, M. Grobosch, E. Cruz-Silva, J.M. Romo-

Herrera, J. Campos-Delgado, H. Terrones, T. Pichler, M. Terrones.
Spectroscopic characterization of N-doped single-walled carbon nanotube
strands: a X-ray photoelectron spectroscopy and Raman study. Journal of
Nanoscience and Nanotechnology (in press).

3 . I.O. Maciel, M.A. Pimenta, M. Terrones, H. Terrones, J. Campos-

Delgado, A. Jorio. The two peaks G’ band in carbon nanotubes. Phys.
Stat. Sol. (b) 245, 10, 2197-2200, 2008

4 . I.O. Maciel, N. Anderson, M.A. Pimenta, A. Hartschuh, H. Qian, M.

Terrones, H. Terrones, J. Campos-Delgado, A. Rao, L. Novotny, A. Jorio.
Electron and phonon renormalization near charged defects in carbon
nanotubes. Nature Materials 7, 878-883, 2008

5 . I.O. Maciel, J. Campos-Delgado, E. Cruz-Silva, M.A. Pimenta, B.G.

Sumpter, V. Meunier, F. López-Urías, E. Muñoz-Sandoval, H. Terrones,
M. Terrones, A. Jorio. Synthesis, electronic structure, and Raman
scattering of phosphorus-doped single-wall carbon nanotubes.
Nanoletters 9,6, 2267-2272, 2009

62
 2.6 References

2.6 References

1
M. Terrones. Carbon nanotubes: synthesis and properties, electronic devices
and other emerging applications. Int. Mat. Rev. 49, 6, 325-377, 2004
2
C.N.R. Rao, A. Govindaraj, R. Sen, B.C. Satishkumar. Synthesis of multi-walled
and single-walled nanotubes, aligned-nanotube bundles and nanorods by
employing organometallic precursors. Mat. Res. Innovat. 2, 128-141, 1998
3
F. Lupo, J.A. Rodríguez-Manzo, A. Zamudio, A.L. Elías, Y.A. Kim, T. Hayashi,
M. Muramatsu, R. Kamalakaran, H. Terrones, M. Endo, M. Rühle, M. Terrones.
Pyrolytic synthesis of long strands of large diameter single-walled carbon
nanonotubes at atmospheric pressure in the absence of sulphur and hydrogen.
Chem. Phys. Lett. 410, 384-390, 2005
4
X. Blase, J.-C. Charlier, A. De Vita, R. Car, Ph. Redlich, M Terrones, W.K. Hsu,
H. Terrones, D.L. Carroll, P.M. Ajayan. Boron-mediated growth of long helicity-
selected carbon nanotubes. Phys. Rev. Lett. 83, 24, 5078- 5081, 1999
5
K. McGuire, N. Gothard, P.L.Gai, M.S. Dresselhaus, G. Sumanasekera, A.M.
Rao. Synthesis and Raman characterization of boron-doped single-walled
carbon nanotubes. Carbon 43, 219-227, 2005
6
W. Han, Y. Bando, K. Kurashima, T. Sato. Boron-doped carbon nanotubes
prepared through a substitution reaction. Chem. Phys. Lett. 299, 368-373, 1999
7
D. Goldberg, Y. Bando, W. Han, K. Kurashima, T. Sato. Single-walled B-doped
carbon, B/N-doped carbon and BN nanotubes synthesized from single-walled
carbon nanotubes through a substitutution reaction. Chem. Phys. Lett. 308,
337-342, 1999
8
R. Sen, B.C. Satishkumar, A. Govindaraj, K.R. Harikumar, G. Raina, J.-P.
Zhang, A.K. Cheetham, C.N.R. Rao. B-C-N, C-N and B-N nanotubes produced
by the pyrolysis of precursor molecules over Co catalysts. Chem. Phys. Lett.
287, 671-676, 1998

63
 2.6 References

9
M. Terrones, P.M. Ajayan, F. Banhart, X. Blase, D.L. Carroll, J.C. Charlier, R.
Czerw, B. Foley, N. Brobert, R. Kamalakaran, P. Kohler-Redlich, M. Rühle, T.
Seeger, H. Terrones. N-doping and coalescence of carbon nanotubes:
synthesis and electronic properties. Appl. Phys. A 74, 355-361, 2002
10
R. Droppa Jr., P. Hammer, A.C.M. Carvalho, M.C. dos Santos, F. Alvarez.
Incorporation of nitrogen in carbon nanotubes. Journal of Non-Crystalline Solids
299-302, 874-879, 2002
11
M. Glerup, M. Castignolles, M. Holzinger, G. Hug, A. Loiseau, P. Bernier.
Synthesis of highly nitrogen-doped multi-walled carbon nanotubes. Chem.
Commun. 2542-2543, 2003
12
M. Glerup, J. Steinmetz, D. Samaille, O. Stéphan, S. Enouz, A. Loiseau, S. Roth,
P. Bernier. Synthesis of N-doped SWNT using the arc-discharge procedure.
Chem. Phys. Lett. 238, 193-197, 2004
13
F. Villalpando-Paez, A. Zamudio, A.L. Elias, H.Son, E.B. Barros, S.G. Chou, Y.A.
Kim, H. Muramatsu, T. Hayashi, J. Kong, H. Terrones, G. Dresselhaus, M.
Endo, M. Terrones, M.S. Dresselhaus. Synthesis and characterization of long
strands of nitrogen-doped single-walled carbon nanotubes. Chem. Phys. Lett.
424, 345-352, 2006
14
E. Cruz-Silva, D.A. Cullen, L. Gu, J.M. Romo-Herrera, E. Muñoz-Sandoval, F.
López-Urias, B.G. Sumpter, V. Meunier, J.-C. Charlier, D. J. Smith, H.
Terrones, M. Terrones. Heterodoped nanotubes: theory, synthesis, and
characterization of phosphorus-nitrogen doped multiwalled carbon nanotubes.
ACS Nano 2, 3, 441-448, 2008
15
B. G. Sumpter, J. Huang, V. Meunier, J. M. Romo-Herrerea, E. Cruz-Silva, H.
Terrones, M. Terrones. A theoretical and experimental study on manipulating
the structure and properties of carbon nanotubes using substitutional dopants.
Int. J. of Quantum Chem. 109, 97-118, 2009

64
 2.6 References

16
I.O. Maciel, J. Campos-Delgado, E. Cruz-Silva, M.A. Pimenta, B.G. Sumpter, V.
Meunier, F. López-Urías, E. Muñoz-Sandoval, H. Terrones, M. Terrones, A.
Jorio. Synthesis, electronic structure, and Raman scattering of phosphorus-
doped single-wall carbon nanotubes. Nanoletters 9,6, 2267-2272, 2009
17
T. Kimura, T. Sugai, H. Shinohara. Production and characterization of boron-
silicon-doped carbon clusters. Chem. Phys. Lett. 256, 269-273, 1996
18
C. Ray, M. Pellarin, L.L. Lermé, J.L. Vialle, M. Broyer. Synthesis and structure
of silicon-doped heterofullerenes. Phys. Rev. Lett. 80, 24, 5365, 1998
19
R.J. Baierle, S. B. Fagan, R. Mota, A. J. R. da Silva, A. Fazzio. Electronic and
structural properties of silicon-doped carbon nanotubes. Phys. Rev. B 64,
085413, 2001
20
G. Tibbetts, C.A. Bernardo, D. W. Gorkiewicz, R.L. Alig. Role of sulfur in the
production of carbon fibers in the vapor phase. Carbon 32, 4, 569-576, 1994
21
L. Ci, Z. Rao, Z. Zhou, D. Tang, X. Yan, Y. Liang, D. Liu, H. Yuan, W. Zhou, G.
Wang, W. Liu, S. Xie. Double wall carbon nanotubes promoted by sulfur in a
floating iron catalyst CVD system. Chem. Phys. Lett. 359, 63-67, 2002
22
Y.-L. Li, I.A. Kinloch, A.H. Windle. Direct spinning of carbon nanotube fibers
from chemical vapor deposition synthesis. Science 304, 276-278, 2004.
23
M. Mottta, Y.-L. Li, I. Kinloch, A. Windle. Mechanical properties of continuously
spun fibers of carbon nanotubes. Nanoletters 5, 8, 1529-1533, 2005
24
I.O. Maciel, N. Anderson, M.A. Pimenta, A. Hartschuh, H. Qian, M. Terrones, H.
Terrones, J. Campos-Delgado, A. Rao, L. Novotny, A. Jorio. Electron and
phonon renormalization near charged defects in carbon nanotubes. Nature
Materials 7, 878-883, 2008
25
A.L. Elías, P. Ayala, A. Zamudio, M. Grobosch, E. Cruz-Silva, J.M. Romo-
Herrera, J. Campos-Delgado, H. Terrones, T. Pichler, M. Terrones.
Spectroscopic characterization of N-doped single-walled carbon nanotube

65
 2.6 References

strands: an X-ray photoelectron spectroscopy and Raman study. Journal of

Nanoscience and Nanotechnology (in press)
26
C.S. Venkateswaran. The Raman spectrum of sulphur in the solid and liquid
states. Proceedings Mathematical Sciences 1, 3, 120-121, 1934.
27
R. Norris. The Raman spectrum and the specific heat of crystalline sulphur.
Proceedings Mathematical Sciences 13, 4, 291-299, 1941
28
Extracted from http://www.photon.t.u-tokyo.ac.jp/~maruyama/kataura/kataura.html

29
B.G. Sumpter, V. Meunier, J.M. Romo-Herrera, E. Cruz-Silva, D.A. Cullen, H.
Terrones, D.J. Smith, M. Terrones. Nitrogen-mediated carbon nanotube growth:
diameter reduction, metallicity, bundle dispersability, and bamboo-like structure
formation. ACS Nano 1, 4, 369-375, 2007
30
I. O. Maciel, M.A. Pimenta, M. Terrones, H. Terrones, J. Campos-Delgado, A.
Jorio. The two peaks G’ band in carbon nanotubes. Phys. Stat. Sol. (b) 245, 10,
2197-2200, 2008

66
 3. Production of carbon nanoribbons using the aerosol pyrolysis technique

3. Production of carbon nanoribbons using the aerosol

pyrolysis technique

Graphene is a single sheet of carbon atoms organized in a honey-comb structure.

Before 2005, graphene was just a concept used by theorists to describe the
formation of carbon nanotubes or to predict amazing electronic properties1;
nevertheless in that year a graphene boom took place when Novoselov et al.2
reported their “scotch-tape” method to obtain single layer and double layer
graphene.
This method gave scientists all around the world the opportunity to work with
graphene as long as they had at their disposal HOPG, scotch-tape, SiO2
substrates and an optical microscope. There are other ways to obtain graphene
flakes and graphene nanoribbons, from heat treatment of diamond nanoparticles
deposited on HOPG3, epitaxially grown4, with chemical treatments of graphite5, and
more recently from treatments of carbon nanotubes6,7,8,9.
Despite the production methods listed above, the synthesis of this material is still a
challenge. Other drawbacks are the low yield production of the material, no control
on its shape, the high pressure and temperature required for its synthesis, specific
equipment, and high purity materials involved in the processes.

We have found a route to synthesize large amounts of graphitic nanoribbons, this

material consists of stacked graphene sheets forming a ribbon-like structure of 10-
20 nanometers in thickness, 20-300 nm in width and several microns in length10.
In this chapter we describe in detail the production method and we include results
of the characterization carried out, which includes SEM, TEM, XRD, EDX and XPS.

The thermal stability of the graphitic nanoribbons has been tested through furnace
heating. The resulting materials have been characterized by electron microscopy.
In particular a detailed Raman spectroscopy study of the heat treated nanoribbons
is included.

67
 3.1 Description of the synthesis method

3.1 Description of the synthesis method

The experimental set-up used in the synthesis of graphitic nanoribbons is shown

schematically in Figure 3.1, where all the necessary equipment is placed inside a
fume hood. A quartz tube (∼1.1 m long), connecting an ultrasonic sprayer and a
gas condenser/acetone trap for the exiting gases, is placed inside two tubular
furnaces (see Fig. 3.1). The length of each furnace is 40 cm and the separation
between the two furnaces is ∼15 cm. An inert atmosphere is maintained by flowing
argon gas through the system. A solution containing 280 ml of ethanol, 2.8235 mg
of ferrocene (FeCp2) and 0.266 ml of thiophene (C4H4S) is prepared and placed
inside the ultrasonic sprayer. During the heating of the furnaces, the ultrasonic
sprayer is kept turned off and the flow of Ar is set to 0.2 l/min. When both furnaces
reach 950 ºC, the ultrasonic sprayer is turned on. The aerosol thus produced is
carried to the hot zone of the furnaces by raising the flow rate to 0.8 l/min. These
conditions are maintained for 30 min, after which the sprayer is turned off, the flow
rate is decreased to 0.2 l/min, and the two furnaces are allowed to cool down to
room temperature.
Once the system has cooled down, the quartz tube is taken out and the material of
interest containing the nanoribbons is scraped from the walls of the tube. The
nanoribbons are deposited as a powder within furnace 1 in a region between 12
and 20 cm relative to the left edge position of the first furnace (see Figure 3.1).

Figure 3.1 A schematic diagram of the experimental set-up used in the synthesis of graphitic
nanoribbons. The yellow marked area in the tube within Furnace 1 corresponds to the region where
11
the graphitic nanoribbon material is deposited during synthesis.

68
 3.1 Description of the synthesis method

The powder containing the nanoribbons has been shown to be efficiently

dispersible in various alcohols. We have used ethanol and methanol followed by
sonication to create solutions of the graphitic nanoribbons. These solutions have
further been used to prepare samples mounted on copper grid sample holders for
use in scanning electron microscope (SEM) and transmission electron microscope
(TEM) observations, and to prepare isolated nanoribbons samples.

3.2 Characterization of the pristine material

The morphology of the initial black powder consisted of ribbon-like structures

exhibiting lengths of several microns, widths ranging from 20-300 nm and
thicknesses of 10-20 nm (Figure 3.2). It is interesting to note that the initial ribbons
revealed both flat regions as well as rippled areas (Figure 3.2b).
The edges of the as-prepared ribbons also displayed relatively sharp cuts that
could be related to the presence of either zigzag or armchair edges (Figure 3.2b-c).
Since the ribbons were extremely thin, SEM images (FEI-field emission SEM XL30
SFEG) almost suggested transparency when observed at 10-15 keV; note the
presence of the holey carbon grid underneath the ribbons in Figure 3.2c.
Bright-field and dark-field images of the ribbons using scanning transmission
electron microscopy (STEM), and are depicted in Figures 3.2d & 3.2f, respectively.
The ribbons displayed only one type of contrast and Fe catalyst particles were
never observed in these structures (note that carbon nanotubes containing metal
catalyst particles at the nanotube tips were usually produced in the second furnace
while the graphitic nanoribbon material was extracted from the first furnace).
By performing detailed elemental energy dispersive X-ray (EDX, using a JEOL
JEM-2010F operated at 200 keV and equipped with a thin-window light-element-
sensitive X-ray detector) spectroscopy line-scans along the ribbon surface (Fig. 3.
2e), it was found that the nanoribbons consisted of C, while S was notably absent.
Even surface-sensitive XPS could not detect any trace of S. Although graphene
nanoribbons consist of only carbon, S appears to be crucial for growing the ribbons
and it could well act as a catalyst. However, the presence of sulphur inserted in the

69
 3.2 Characterization of the pristine material

system during growth was positively confirmed in by-products found in a different

zone of the quartz tube (i.e. in the region between furnace 1 and 2. See Chapter II
for the description of similar experiments aimed to produce S-doped SWNTs,
Figures 2.4 and 2.5). Figure 3.3 depicts ribbons under TEM and HRTEM imaging
conditions (JEOL JEM-2010F operated at 200 keV). At low magnification, the
material consisted only of carbon ribbons (Fig. 3.3a). HRTEM images of the
ribbons (Figs. 3.3b-c) revealed the presence of hexagonal patterns, which were
confirmed after obtaining the fast Fourier transform (FFT; see Fig. 3. 3c inset). In
order to confirm the graphitic structure and layer stacking, and to identify the edge
structure, electron diffraction patterns were recorded from different ribbons (Fig.
3.3d). Interestingly, all the analyzed ribbons consisted of ABAB… stacked graphite
(since all reflections from 3D graphite are visible in Fig. 3.3d), and the edges
exhibited armchair and zigzag (or close to zigzag) morphologies (see below).

Figure 3.2 (a-c) SEM images of graphitic nanoribbons at different magnification. Note that the
ribbons are very thin (~ 10nm) and could be transparent to the SEM beam (see 1c); (d) HAADF
image of a nanoribbon and its corresponding elemental EDX line-scan (see d), indicated by the line
shown in (d). The elemental profile in (e) shows the absence of S and indicates that the ribbon
mainly consists of C; (f) dark field STEM image of bulk nanoribbons displaying rippled regions within
the ribbons.

70
 3.2 Characterization of the pristine material

The average bulk structure of the ribbon material was further studied by XRD
(Brucker D8-advanced equipped with a Cu anode (λ=1.5406 Ǻ) operated at 35 kV,
25 mA and 293 K). It was found that the nanoribbons exhibited a highly crystalline
graphite-like structure, with the presence of strong (002), (100), (101), (004) and
(110) reflections (Fig. 3.4a). Moreover, the linewidth of the (002) diffraction line
gave an average Lc crystallite size of ca. 10-14 nm, in good agreement with SEM
observations.

Figure 3.3 (a-b) Low and high magnification TEM images of graphitic nanoribbons; c) HRTEM
image of a ribbon edge (the arrow indicates the ribbon edge) displaying a hexagonal pattern, and its
corresponding FFT (inset); d) indexed electron diffraction pattern of an individual thin graphene
nanoribbon (ca. 10 nm thick) showing the ABAB… stacking of the graphite structure with a 3D order

Bulk Raman spectroscopy measurements on the ribbons (Renishaw InVia system

equipped with laser excitation energy Elaser = 2.41 eV) revealed the presence of the
D and G bands, located at 1355 and 1584 cm-1, respectively (Fig. 3.4b). In general,
it was found that the D-band exhibited the highest intensity, probably due to the
high proportion of edges and ripples within the ribbons. The disorder-induced
combination mode (D+G) at about 2940 cm-1 is also exceptionally strong. A more

71
 3.2 Characterization of the pristine material

detailed Raman spectroscopy study of pristine and heat treated nanoribbons is

included in future pages.
In order to confirm that the nanoribbons were highly crystalline, we performed TGA
studies (Thermo Haake, Cahn VersaTherm HS heating at 5 K/min to 1173 K in air).
It was found that the decomposition temperature of the ribbons in air corresponded
to 702 ºC (Fig. 3.4c). This value is almost the same as that observed in highly
crystalline carbon nanotubes produced by arc discharge techniques24. Adsorption
isotherms of N2 and H2 were measured volumetrically at 77 K after preheating in
vacuo. Nitrogen (N2) adsorption measurements on the carbon nanoribbons (Fig.
3.4d) revealed a BET surface area of 59 m2/g, which was similar to the surface
area of acetylene black (86 m2/g). The adsorption data indicated that the
nanoribbon material corresponds to a flat surface, which was not porous to N2
molecules. The interaction of an N2 molecule with the surface of the nanoribbon is
weaker than that with highly-crystalline carbon black, which is in agreement with
the presence of predominant edge-like surfaces. However, H2 adsorption at 77 K
indicated the presence of rather strong sites for supercritical H2 adsorption
corresponding to ca. 15 % of the monolayer capacity of N2. Consequently, the
nanoribbon should have a unique nanostructural fit for the adsorption of
supercritical H2.
XPS studies revealed the nature of the carbon bonds present in the sample (Figs.
3.4e-f) (JEOL JPS-9010MX using Mg Kα radiation at 10-6 Pa. XPS spectrum being
deconvoluted with a Gaussian-Lorentzian mixed function after correction). The
material contained sp2 and sp3 hybridized carbon atoms (39% sp2 and 39% sp3;
Fig. 3.4e), and the rest of the carbonaceous material was bonded to O (Fig. 3.4f)
and consisted of carbonyl groups (C=O) and carboxylic groups (COO); note that 85
at. % corresponded to C and 15 at. % to O. We believe that the large number of
sp3 hybridized carbon atoms was caused by the exposed edges and the rippled
(highly curved) regions within the nanoribbons. The 1:1 ratio of sp3:sp2 carbon
atoms could also explain the large intensity of the D-band observed in Raman
spectroscopy, because the material was indeed highly crystalline and showed an
AB… stacking of the graphene layers.

72
 3.2 Characterization of the pristine material

Figure 3.4 (a) XRD pattern of a bulk nanoribbon sample; b) typical Raman spectrum of bulk
nanoribbons showing the presence of the D and G band as well as the overtone and combination
mode features taken at 514 nm laser excitation; c) TGA plot and its first derivative presented in the
inset showing the behavior of the DTA peak in an oxygen atmosphere for the ribbon material at high
temperature; d) A Typical N2 adsorption isotherm of the nanoribbon sample; e) XPS data of
2 3
graphene nanoribbons for the C binding energies; it is clear that sp and sp hybridized carbon
atoms are included in the sample with a 1:1 ratio, as well as carbonyl groups (C=O) and carboxylic
groups (COO), and f) XPS data for graphene nanoribbon material corresponding to the O binding
energies, confirming the presence of different O terminated edges.

73
 3.3 Annealing treatments of carbon nanoribbons

3.3 Annealing treatments of carbon nanoribbons

Historically, post-synthesis thermal treatments of carbon materials have been used

with different objectives, including the achievement of higher degrees of
graphitization12, elimination of amorphous carbon in carbon nanotubes13,14,
graphitization of diamond particles3, transformation of SWNT bundles into graphite
nanoribbons15, and the elimination of metal particles from the starting materials16,17.
During these thermally-induced structural transformations processes like
graphitization and a process called loop formation at the edges of graphitic
materials occurs18-22 (see Figure 3.5), whereby the chemically reactive open edges
of adjacent graphene sheets join with one another to form a semi-cylindrical loop
with a radius larger than the sheet-to-sheet separation distance.

Figure 3.5 Loop formation at the edges of graphene planes induced by heat treatments in different
18 20
carbon materials, a) filamentous graphite , b) cup-stacked fibers , c) graphite polyhedral
19 21
crystals , and d) powders of pyrolytic graphite .

These loops, which serve to eliminate the reactive edges and satisfy the dangling
bond requirements of the edge structures, have been observed for diverse
graphitic materials after heat treatments.
In Figure 3.5 we show some representative images of such materials: filamentous
graphite18, graphite polyhedral crystals19, cup-stacked nanofibers20 and powders of

74
 3.3 Annealing treatments of carbon nanoribbons

yrolytic graphite21. José-Yacamán et al.22 previously reported an elongated

graphitic structure with bent ends, and although closed ends were also observed in
this prior report, their bent ends do not resemble the loops observed in the present
work.
On the other hand, a sensitive technique to the graphitization process that occurs
during annealing is Raman spectroscopy, through the evolution of the D, G and G’
bands we can have an idea of the degree of graphitization of the studied sample.
Figure 3.6 pictures heat treatment experiments found in the literature for different
carbon materials studied by Raman scattering.

In order to study the thermal stability of our nanoribbons, high temperature heat
treatments were conducted by collaborators in Japan, using a graphite furnace for
30 minutes under an Ar atmosphere. The heat treatment temperatures (THT) used
in the study were 1000, 1500, 2000, 2500 and 2800 ºC, the resulted samples will
be denoted by “HT **** ºC”. SEM imaging, TEM imaging, X-ray powder diffraction,
and thermogravimetric analysis (TGA), were used to characterize our ribbon
samples upon increasing the heat treatment temperature23. A Raman
spectroscopy study at the bulk and individual level was also performed.

Figure 3.6 Raman spectroscopy of heat treated carbon materials, a) benzene-derived carbon
24 25 20
fibers , b) activated carbon fibers , c) cup-stacked carbon fibers

75
 3.3 Annealing treatments of carbon nanoribbons

3.3.1 Characterization of the annealed materials

The morphology of the heat treated (HT) ribbon samples was studied by SEM (FEI-
field emission SEM XL30 SFEG).
The distinctive morphology features of the pristine sample, such as ripples on the
surface and un-continuous edges, are still present in the HT samples.
In general these samples do not seem to show dramatic changes, as can be
observed in Figure 3.7, which shows SEM micrographs for the pristine sample, HT
1000 ºC, HT 2000 ºC, 2500 ºC and HT 2800 ºC. Sample HT 1500 ºC, not shown
here, also shared the pristine morphology characteristics.

Figure 3.7 Different magnification SEM images of HT carbon nanoribbons

In order to study the structure of the HT samples we performed high-resolution

TEM (JEOL JEM-2010F operated at 200 keV), with special interest on the edges of
the nanoribbons, Figure 3.8 depicts the side-edge structure for the different THT
values.
As the THT is increased it is evident that transformations take place, leading to
structures with many striking differences that will be discussed carefully in the
following paragraphs.

76
 3.3 Annealing treatments of carbon nanoribbons

It is clear that the pristine sample contains many active end planes (edges).
As the samples are annealed, interesting features can be pointed out:

1) Small structural changes are observed for the pristine and the 1000 ºC THT
sample, both exhibit a well developed lattice fringe structure. We attribute
these small changes to the fact that the synthesis temperature (950 ºC) of the
pristine material is very close to the first annealing temperature (1000 ºC).

2) At 1500 ºC, the lattice fringes straighten up and single-loops start to appear at
the edges, although open edges are still present. Here, it is clear that point
defects present in the individual layers start to anneal and disappear.

3) At 2000 ºC, the disappearance of open edges and double-loop formation are
observed.

4) For the highest annealing temperatures (2500 ºC - 2800 ºC), we observe

increasingly straighter lattice fringes, multi-loop formation covering an
increasing number of layers and the absence of open edges.

It is evident that a graphitization process, defect annealing and loop termination of

edges due to high temperature exposure take place in the samples.
The loop formation observed in the edges of the ribbons has already been reported
when temperature increasing is involved in the experiments18-21.

X-ray powder diffraction patterns (Philips Diffractometer model 12014424, Cu Kα

radiation λ=1.54 Ǻ) plotted in Figure 3.9a, reveal a match to the graphite pattern for
several reflections, not just for the (002) reflection, as is done in most studies of
CVD-produced MWNTs26.
The analysis of the (002) peak position revealed an average interlayer spacing of
3.36 Å (close to perfect graphite) for the pristine sample, HT 1000 ºC, HT 2500 ºC
and HT 2800 ºC samples, and 3.37 Å for the HT 1500 ºC and HT 2000 ºC samples.

77
 3.3 Annealing treatments of carbon nanoribbons

Figure 3.8 Transmission electron micrographs of the different HT carbon nanoribbons (scale bar =
23
5 nm).

78
 3.3 Annealing treatments of carbon nanoribbons

Small changes are seen in all the reflections as the annealing temperature
increases (see Fig. 3. 9b). It is noteworthy that in the HT samples no iron carbide
peaks are present, as has been reported for the pristine ribbons10.

Figure 3.9 a) X-ray diffraction characterization of each HT treated sample, the numbers in
parenthesis indicate the crystallographic reflections, b) interlayer distances between the (002)
planes as a function of THT, c) first derivative with respect to temperature of the weight loss (DTA)
curves of the heat treated samples, and d) Td vs. THT plot for the peak temperature in c) for the
23
different HT samples (the growth temperature for the pristine sample is 950ºC).

79
 3.3 Annealing treatments of carbon nanoribbons

In Fig. 3.9c, we show the first derivative with temperature (DTA) of the weight loss
curves obtained from the TGA analysis (Thermo Haake, Cahn VersaTherm HS
system, heating the samples at 5 ºC/min to 900 ºC in air).
Interesting phenomena can be deduced from these measurements:

• As THT is increased, the TGA analysis reveals an up-shift in the decomposition

temperature Td of the ribbon materials in air, from 702 ºC for the pristine sample
to 780 ºC for the HT 2800 ºC sample; this behavior suggests a restructuring
taking place during the heating process (replacement of open edges by loops),
which makes the material less reactive, thereby increasing the decomposition
temperature.

• The HT 1000 ºC sample, however, has a lower Td (665 ºC) than the pristine
sample (702 ºC). As suggested by previous work in carbon nanotubes27, defect
migration and vacancy coalescence induced by heat treatment effects could be
taking place at this stage, forming large area defects which turn the sample into
a more reactive material. It is also possible that at 1000 °C in an Ar
atmosphere, the ribbon sample detaches some passivation atoms (e.g. H, O)
that recombine with defective areas creating larger reactive and broken regions
around defects. The above effects, combined with the presence of open edges
passivated by oxygen or hydrogen groups, could contribute further to the
decrease of the decomposition temperature. It is noteworthy that at 1000 °C the
ribbon sample is not able to completely anneal defects into energetically stable
sp2 hybridized carbon regions, thus leaving a metastable carbon network that is
still quite reactive.

• A difference of 107 ºC in Td for the HT 1500 ºC and HT 1000 ºC samples

suggests that the number of defects and reactive sites decreases significantly
in this temperature range. We suggest that competing processes are here in
play. Firstly, the high temperature promotes the annealing of defects as

80
 3.3 Annealing treatments of carbon nanoribbons

revealed by the straightening of the lattice fringes observed in Figure 6, and the
termination of the reactive edges by loop formation are responsible for the Td
change.

• For the HT 1500 ºC samples and up, the Td range is from 772 ºC to 780 ºC (see
Fig. 3.9d). This stability in Td suggests that no major restructuring of the sample
is occurring among these different temperature stages, which agrees with our
TEM observation in that the edges form single to multiple loops in this
temperature range, thereby reducing chemically active sites and reducing the
overall reactivity.

In order to have a clear picture of the restructuring taking place during annealing
treatment, we have identified the transformations that occur for each THT. Figure
3.10 shows a diagram imaging such transformations.

Figure 3.10 Schematic model for the restructuring process achieved by the thermal annealing
23
treatment.

For THT below 1500 ºC the sample consists of relatively well ordered planes and
well-developed crystallinity (confirmed by XRD and electron diffraction10) but still

81
 3.3 Annealing treatments of carbon nanoribbons

contains point defects (vacancies), out-of-plane interstitial atoms, Stone-Wales and

other defects, in addition to the bare edges of the ribbons where some oxygen
groups are present10.
By 1500 ºC, defect annealing starts to take place and we observe straighter lattice
fringes, approaching a defect-free planar structure, and some open edges become
passivated by single-loop formation driven by surface energy minimization.

For HT 2000 ºC, double loop formation is evident, and as the temperature is
increased further, multi-loops with increasing loop diameters are observed and
open edges are no longer present.

The zipping mechanism proposed by Rotkin28 seems to elegantly describe the loop
formation phenomenon. According to this model, two adjacent graphene sheets
will tend to join their edges to form a loop, and the radius of curvature will
systematically be greater than the separation of the sheets.
When single and double loops are formed (see Figure 3.8) an increase in the
radius of curvature is evident, confirming this aspect of the proposed model26.
This local increase in the interlayer distances at the loops may well explain the
presence of slightly higher interlayer spacings observed by analyzing the graphite
(002) peak from the X-ray powder diffraction patterns of samples treated at
different temperatures.
Note that as the loop radius is increased, the loops start forming more facetted
structures, as is commonly seen in other carbon nanostructures10,29.

3.3.2 Raman spectroscopy study of the annealed materials

Raman spectroscopy is a powerful tool to analyze carbon nano-systems, due to

the sensitive response in Raman signal of C-C bonds and defects in the structure
(see Appendix B).
Bulk micro Raman spectra with laser excitation energies in the range 1.91 eV -
2.61 eV were acquired at room temperature for 2 minutes with laser powers of 1

82
 3.3 Annealing treatments of carbon nanoribbons

mW. The samples were analyzed in the backscattering geometry using a 100x
objective lens which focused the laser beam on a spot of ~0.5 µm in diameter. At
least two measurements were recorded per sample and the spectra shown here
are the resulting averages. All the spectra were normalized to the G band
intensity. Seven different laser energies were used: Krypton (1.91 eV), Argon
(2.41, 2.54, and 2.61 eV), YAG (2.33 eV) and Rhodamine dye (2.13 and 2.17 eV).
Spectra were recorded for six samples (a pristine sample and samples heat treated
at 1000 ºC, 1500 ºC, 2000 ºC, 2500 ºC and 2800 ºC).

Figure 3.11 Bulk Raman spectra of the investigated samples using seven different laser excitation
energies in the range 1.91 eV - 2.61 eV.

In Figure 3.11 we show the spectra of the bulk pristine sample, and the bulk heat
treated samples at 1000, 1500, 2000 ºC, 2500 and 2800 ºC (HT 1000 ºC, HT 1500
ºC, HT 2000 ºC, HT 2500 ºC and HT 2800 ºC, respectively).
A quick inspection of these measurements reveals differences among the three
samples for each of the different Elaser. The main Raman features (D, G, D’ and G’

83
 3.3 Annealing treatments of carbon nanoribbons

bands) show interesting behavior as the heat treatment temperature (THT) and
Elaser are increased.

In order to make a quantitative analysis of the Raman features, we have fitted

those main peaks for the pristine, HT 1500 ºC, HT 2000 ºC and HT 2800 samples
for each of the seven laser excitation energies. We used a program to
automatically identify and fit the peaks with Lorentzians called “Peak Fit”, with the
Residuals method.

Figure 3.12 Example of the fitting carried out for the bulk Raman measurements. The spectrum of
the sample HT 2000 ºC for Elaser = 2.33 eV was fitted with 5 Lorentzian peaks (plot on the left), the
numerical values derived from the fitting (right) were further used for analysis of the frequency
positions and the integrated area ratios.

Each spectrum was fitted with 4 main peaks (D, G, D’ and G’ bands) sometimes 5
peaks (D, G, D’, G* and G’).
The numerical information derived from the fitting was used to identify frequency
positions at maximum intensity of the peaks and to calculate ratios of the
integrated areas as will be discussed below (See Figure 3.12).
When a study at different Elasers is carried out, the dispersion (change) of the peak
positions can be quantified as a function of laser energy. Of particular interest is
the shifting of the D- and G’-band peaks with different Elaser. Both peaks shift to
higher frequency values as Elaser is increased, in accordance with double
resonance theory30.

84
 3.3 Annealing treatments of carbon nanoribbons

Figure 3.13a summarizes the frequency behavior of the D-, G-, D’- and G’- bands
for the pristine sample (open squares), HT 1500 ºC (open circles), HT 2000 ºC
(open inverse-triangles) and HT 2800 ºC (open triangles) samples as a function of
Elaser. The peak position of the D- and G’-bands show a very strong dependence on
Elaser, up-shifting linearly as Elaser is increased.
These peaks show the same behavior as that reported for other sp2 carbon
materials.[31,32,33] On the other hand, the frequencies of the G and D’ bands hardly
change as Elaser is increased.

Figure 3.13 a) D-, G-, D’-, and G’- band positions as a function of Elaser for the pristine, HT 1500 ºC,
HT 2000ºC and HT 2800ºC samples, b) ID/IG ratios vs. Elaser, the black (violet) dashed line
represents a linear fit to the pristine (HT 2800 ºC) values

We can point out that the slopes of the dependence for the frequency of the D- and
G’- bands on Elaser take the highest values for the pristine sample and the lowest
for the HT 2800 ºC sample.
The pristine sample shows a frequency dependent slope for the D-band on Elaser of
∂ωD/∂Elaser ≈ 43 cm-1eV-1 and the slope of the G’-band is ∂ωG’/∂Elaser ≈ 96 cm-1eV-1.
For the HT 2800 ºC sample, the values are: ∂ωD/∂Elaser ≈ 36 cm-1eV-1 and
∂ωG’/∂Elaser ≈ 87 cm-1eV-1.
These numbers indicate a difference in dispersion directly related to the changes
induced by the heat treatments in the sample such as defect annealing, in-plane
crystallinizaton, and loop formation at the edges.

85
 3.3 Annealing treatments of carbon nanoribbons

There is a remarkable difference from the values reported by Vidano et al.30 and
Matthews et al.31 (∂ωD/∂Elaser ≈ 50 cm-1/eV and ∂ωG’/∂Elaser ≈ 100 cm-1/eV) where
sp2-hybridized carbon materials were used, such as disordered pyrolytic and
glassy carbons, crystalline graphite, PPP and HOPG.
The differences in these values might be attributed to the nature of our sample,
which could contain a starting 1:1 ratio of sp2-sp3 hybridization (e.g. pristine
sample), and a high density of edges10.

In 1984 Lespade and co-workers34 performed a Raman study of graphitization of

different carbon samples under heat treatments. They came up with four
“graphitization indices” derived from the Raman spectra.
In this discussion we will use two of them, the ID/IG ratios and FWHM of the G
band, in order to monitor the graphitization process of the heat treated
nanoribbons.
In Figure 11 we can observe an increase in the D-, D’- and G’-band intensities as
Elaser is decreased. In the case of the D-band, this tendency is quantified through
the ratio of the integrated intensities of the D- and G-bands (ID/IG) shown in Figure
3.13b as a function of Elaser.
For all of the samples, we observed lower ID/IG ratios when high Elaser energies are
used. For the lowest energy (Elaser=1.91 eV), the ID/IG ratios were the largest for all
four samples (unannealed and heat treated).
Looking at the set of ID/IG ratio values for each sample, the dependence on Elaser
for the pristine sample is the highest (making a rough linear fit to the values, the
slope of the line is -1.3 a.u./eV, as seen from the black dashed line in Figure
3.13b), while the dependence on Elaser decreases significantly for the sample heat
treated at 2800 ºC (fitting the values linearly, the slope of the line is -0.55 a.u./eV,
see violet dashed line in Figure 3.13b).
Once again, the differences on Elaser dependence of the ID/IG ratios may be
attributed to intrinsic changes triggered by the annealing treatments.

86
 3.3 Annealing treatments of carbon nanoribbons

The shifts to smaller values in the ID/IG ratios are noteworthy as we increase the
heat treatment temperatures from 1500 ºC to 2800 ºC, suggesting an increase in
crystalline order induced by the heat treatment temperature (THT). This latter
sample, as described previously23, consists of a highly crystalline graphite-like
material exhibiting multiple loops formed between adjacent graphene layers as a
result of the high temperature annealing treatment (see Figure 3.8).
The evolution of the FWHM (Full Width at Half Maximum) of the G band as a
function of heat treatment temperature for each laser line is plotted in Figure 3.14a.
The decrease in FWHM as THT is increased, confirms a sharpening of the G band
which is a direct evidence of enhanced crystallinization and the annealing of
structural defects.

Figure 3.14 a) FWHM values of the G band as a function of heat treatment temperature (THT) for
each of the seven excitation laser energies, and d) IG’/IG ratios vs. Elaser.

When looking at Figure 3.11, especially at low energies (1.91 eV - 2.17 eV), it is
evident that the intensity of the G’ band shows a progressive increase as the heat
treatment temperature is increased. Figure 3.14b shows a plot of the IG’/IG ratios as
a function of laser energy. Contrary to the behavior of the D band, the IG’/IG ratios
show their lowest values for the pristine sample and the highest ones for the HT
2800 ºC sample. The IG’/IG ratios reveal a dependence on Elaser, though it is not
linear and it does not follow the behavior shown by the ID/IG ratios.

87
 3.3 Annealing treatments of carbon nanoribbons

One remarkable feature of the ribbon samples is their inability to graphitize

completely. The double peak shape of the G’ band (characteristic of HOPG and
graphitized materials, explained by a second order double resonance Raman
process where we observe signal from the incident and scattered phonons28) is not
observed in our samples, not even for the material annealed at the highest
temperature (2800 ºC). This is probably due to the loop termination of the edges
which is itself symmetry-breaking and should contribute in intensity to disorder-
induced Raman features.
In summary, the bulk Raman measurements confirm that after thermal annealing,
the material enhanced its degree of crystallinity. The low residual D-band in the
HT 2800 ºC sample indicates a more crystalline structure when compared to its
unannealed counterpart, but showing still remaining symmetry-breaking elements
caused by the formation of loops.

In order to prepare isolated nanoribbon samples, a powder of the pristine sample

and samples heat treated at 1500 ºC and 2800 ºC were dispersed in methanol and
sonicated for 1 hour. After this time, one drop of the solution was deposited on a
Si/SiO2 substrate.

Due to the length dimensions of the nanoribbons (several microns), the

identification of the individual nanostructures was possible using an optical
microscope.
Once the isolated nanoribbons were detected, we performed scans of the
individual ribbons with Elaser = 2.33 eV at steps of 0.2 µm.
A computerized/motorized stage allowed us to precisely move the substrate in
order to scan the region of the substrate containing the nanoribbon of interest.
Each final measurement lasted 60 seconds and the optical power at the sample
position was maintained below 0.3 mW in order to avoid local heating effect.

88
 3.3 Annealing treatments of carbon nanoribbons

Figure 3.15 a) and c) SEM micrographs of a pristine nanoribbon and a nanoribbon annealed at
2800 ºC, respectively. The same pieces imaged by AFM after laser irradiation at 1.5 mV, b) pristine
and d) HT 2800 ºC. The inset in d) contains a higher resolution AFM image of the marked section.

Our first measurements, at relatively low power levels (∼1.5 mW) with Elaser = 2.33
eV and post-AFM-imaging (Atomic Force Microscopy) of the species, revealed
damage of the structure.
Figure 3.15 a and c shows two individual nanoribbons identified by SEM before
Raman spectroscopy , a pristine sample in a) and a heat treated one at 2800 ºC in
c).
After the material was scanned with the laser, at powers around 1.5 mW, the
structure was modified as can be observed in Figure 3.15 b) and d), at this point
what looked continuous in a) and c) became irregular with a segmented
morphology. We assume that some parts of the ribbons were not in complete
contact with the substrate, and therefore the heat induced by the incident laser
could not be efficiently dissipated and this led to the destruction of the material.

89
 3.3 Annealing treatments of carbon nanoribbons

This information suggests that the sample had an overall low thermal conductivity.
To prevent sample heating and temperature dependent shifts of the Raman
spectra, the power was decreased to ∼0.3 mW when probing the nanoribbons.

Figure 3.16 SEM micrographs of the individual nanoribbons, a) unannealed sample, b) sample
annealed at 1500 ºC and c) sample annealed at 2800 ºC. The green lines across the nanoribbons
indicate the areas where the measurements were recorded and the real magnitude of the laser
beam spot. Scale bars = 2 µm.

Once individual nanoribbons were identified using an optical microscope, scanning

electron microscopy was used to characterize them. Figure 3.16 shows the
isolated nanoribbons used in the measurements reported in Figure 3.17. The wide
green lines in Figure 16 represent the real area scanned by the incident laser
beam.

Raman spectroscopy line scans were performed across the nanoribbons, each
scan consisting of 4-6 individual measurements at steps of 0.2 µm, going from one
edge of the nanoribbon to the other (Figure 3.16 and 3.17d). Considering the spot
size of the laser beam, this procedure allowed us to hit the sample mainly on the
edges when the laser was far from the center of the nanoribbon.

The results of these measurements are plotted in Figures 3.17 a, b and c, and a
diagram of the positions of the measurements relative to the nanoribbon
morphology is shown in Figure 3.17d. It is clear that the spectra recorded far from
the center of the nanoribbons show more intense D band features for all three
nanoribbon samples.

90
 3.3 Annealing treatments of carbon nanoribbons

Figure 3.17 Raman spectroscopy line scans at Elaser = 2.33 eV in the range from -0.6 µm to 0.4 µm
in steps of 0.2 µm taken on individual nanoribbons: a) pristine sample, b) heat treated sample at
1500 ºC and c) heat treated sample at 2800 ºC. d) A diagram illustrating the position of the laser
spot (green circles) at each measurement across the nanoribbon. The numbers from 1 to 6 relate to
the actual laser spot positions labeled in Figure 18a.

We attribute this to a higher density of disorder-inducing elements present on the

ribbon edges, such as sp3 hybridized carbon atoms, defects10 and loops (whose
curvature is symmetry-breaking) in the case of the annealed samples23.
We believe that the spectra taken at the central part of the ribbons exhibit more
contributions coming from the hexagonal honeycomb planes which are
perpendicular to the laser beam.

Thus, the signal intensities differ along the scan as observed in Figure 3.18a,
where the intensities of the D-bands are plotted as a function of the distance
across individual nanoribbons.

91
 3.3 Annealing treatments of carbon nanoribbons

Figure 3.18 The intensities of the D- and D’- bands vs. distance in the line scan measurements on
the three individual nanoribbons are plotted in a); the circled regions represent the centers of the
isolated nanoribbons, and b) Raman measurements with lower D band intensities for each sample,
the spectrum of HOPG is also included for comparison.

Similar to the bulk observations, we note a decrease in the D-band intensity as we

go from the unannealed sample to the one annealed at 1500 ºC and finally to the
heat treated at 2800 ºC. Figure 3.18b shows the spectrum with the lowest D-band
intensity for each of the samples and the spectrum of HOPG is also shown for
comparison.
The increase in annealing temperature sharpens the G, G*, G’ and 2D’ features,
while it decreases the intensity of the disorder-induced features (D-, D’- and D+G-
bands).
As we have mentioned earlier, although high temperatures were used in the heat
treatments (up to 2800 ºC), we still observe a residual D-band, and the shape of
the G’-band indicates the inability of these samples to achieve long range AB
stacking order, characteristic of highly crystalline graphitic samples.

92
 3.4 Conclusions

3.4 Conclusions

We have found a new route to efficiently produce graphitic nanoribbons at

atmospheric pressure. We have demonstrated that heat treatments up to 2800 ºC
of as-produced carbon nanoribbons anneal point defects and interstitials within
individual graphene layers, thus leading to a material with a higher degree of
crystallinity, in which the reactive edges of the ribbons are transformed into loops in
a self-assembly process that can be strongly controlled by the variation of THT.

We have performed a detailed Resonance Raman study of heat treated

nanoribbons on both bulk and individual nanoribbon samples. At the bulk level, we
have confirmed a laser energy dependence of the D- and G’-bands frequency
positions; and in these dependence values we found differences within the
samples attributed to the annealing process.

The decrease of the ID/IG ratios and the FWHM of the G band as the heat treatment
temperature is increased suggests an increase in crystallinity of the hexagonal
honeycomb lattice, consequence of a partial graphitization process in which the
stacking order could not be achieved completely possibly caused by the loop
formation at the ribbon edges.

At the individual nanoribbon level, we have demonstrated that edge contributions

to the Raman spectra intensify the signal of the D-band, indicating that symmetry-
breaking elements such as sp3-like hybridized carbon atoms, structural defects and
loops are present particularly in the vicinity of the edges.
However, the large number of loops present at the edges of all the ribbons is
responsible for observing some residual D-band intensity.

The general up-shift of the decomposition temperature in the TGA analysis also
reveals a more stable and less reactive material as THT increases in the range
above which loop formation takes over.

93
 3.5 Related articles

3.5 Related Articles

1 . J. Campos-Delgado, H. Farhat, Y.A. Kim, A. Reina, J. Kong, M. Endo, H.

Muramatsu, T. Hayashi, H. Terrones, M. Terrones, M.S. Dresselhaus.
Resonant Raman study on bulk and isolated graphitic nanoribbons. Small
DOI: 10.1002/smll.200901059

2 . J. Campos-Delgado, Y.A. Kim, T. Hayashi, A. Morelos-Gómez, M.

Hofmann, H. Muramatsu, M. Endo, H. Terrones, R.D. Shull, M.S.
Dresselhaus, M. Terrones. Thermal stability studies of CVD-grown
graphene nanoribbons: defect annealing and loop formation. Chem. Phys.
Lett. 469, 177-182, 2009

3 . J. Campos-Delgado, J.M. Romo-Herrera, X. Jia, D.A. Cullen, H.

Muramatsu, Y.A. Kim, T. Hayashi, Z.F. Ren, D.J. Smith, Y. Okuno, T.
Ohba, H. Kanoh, K. Kaneko, M. Endo, H. Terrones, M.S. Dresselhaus, M.
Terrones. Bulk production of a new form of sp2 carbon: crystalline
graphene nanoribbons. Nanoletters 8, 2773-2778, 2008

94
 3.6 References

3.6 References

1
K. Nakada, M. Fujita, G. Dresselhaus, M.S. Dresselhaus. Edge state in
graphene ribbons: nanometer size effect and edge shape dependence. Phys.
Rev. B. Vol. 54, 24, 1996
2
K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos,
I.V. Grigorieva, A.A. Firsov. Electric field effect in atomically thin carbon films.
Science. 306, 666, 2004 and K.S. Novoselov, D. Jiang, F. Schedin, T.J. Booth,
V.V.Khotkevich, S.V. Morozov and A.K. Geim. Two-dimensional atomic
crystals. PNAS. Vol. 102, No. 30, 2005
3
A.M. Affoune, B.L.V. Prasad, AH.Sato, T. Enoki, Y. Kaburagi, and Y.
Hishoyama. Experimental evidence of a single nano-graphene. Chem. Phys.
Lett. 348, 17, 2001
4
C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud, D. Mayou, T. Li, J. Hass,
A.N.Marchenkov, E.H. Conrad, P.N. First, and W.A. de Heer. Electronic
confinement and coherence in patterned epitaxial graphene. Science. 312,
1191, 2006
5
X. Li, X. Wang, L. Zhang, S. Lee, H. Dai. Chemically derived, ultrasmooth
graphene nanoribbon semiconductors. Science 319, 1229, 2008
6
M. Zhang, D.H.Wu, C.L. Xu, Y.F. Xu, and W.K. Wang. Ribbon-like
nanostructures transformed from carbon nanotubes at high temperature and
pressure. Nanostructured Materials. 10, 1145, 1998
7
A. G. Cano-Márquez, F. J. Rodríguez-Macías, J. Campos-Delgado, C.G.
Espinosa-González, F. Tristán-López, D. Ramírez-González, D.A. Cullen, D.J.
Smith, M. Terrones, Y.I. Vega-Cantú. ExMWNTs: graphene sheets and ribbons
produced by lithium intercalation and exfoliation of carbon nanotubes.
Nanoletters 9, 4, 1527-1533, 2009

95
 3.6 References

8
D.V. Kosynkin, A.L. Higginbotham, A. Sinitskii, J.R. Lomeda, A. Dimiev, B.K.
Price, J.M. Tour. Longitudinal unzipping of carbon nanotubes to form graphene
nanoribbons. Nature 458, 2009
9
L. Jiao, L. Zhang, X. Wang, G. Diankov, H. Dai. Narrow graphene nanoribbons
from carbon nanotubes. Nature 458, 877-880, 2009
10
J. Campos-Delgado, J.M. Romo-Herrera, X. Jia, D.A. Cullen, H. Muramatsu,
Y.A. Kim, T. Hayashi, Z.F. Ren, D.J. Smith, Y. Okuno, T. Ohba, H. Kanoh, K.
Kaneko, M. Endo, H. Terrones, M.S. Dresselhaus, M. Terrones. Bulk production
of a new form of sp2 carbon: crystalline graphene nanoribbons. Nanoletters 8,
2773-2778, 2008
11
X. Jia, J. Campos-Delgado, E.E. Gracia-Espino, M. Hofmann, H. Muramatsu,
Y.A. Kim, T. Hayashi, M. Endo, J. Kong, M. Terrones, M.S. Dresselhaus. Loop
formation in graphitic nanoribbon edges using furnace heating or Joule heating.
J. Vac. Sci. Technol. B 27, 4, Jul/Aug, 1996-2002, 2009
12
J. Goma and M. Oberlin. Graphitization of thin carbon films. Thin Solid Films 65,
221, 1980
13
Y.S. Park, Y.C. Choi, K. S. Kim, D-C Chung, D.J. Bae, K.H. An, S.C. Lim, X.Y.
Zhu, Y.H. Lee. High yield purification of multiwalled carbon nanotubes by
selective oxidation during thermal annealing. Carbon 39, 655-661, 2001
14
I.W. Chiang, B.E. Brinson, R.E. Smalley, J.L. Margrave, R.H. Hauge.
Purification and characterization of single-wall carbon nanotubes. J. Phys.
Chem. B 105, 1157, 2001
15
H.R. Gutiérrez, U.J. Kim, J.P. Kim, P.C. Eklund. Thermal conversion of bundled
carbon nanotubes into graphitic ribbons. Nanoletters 5, 2195, 2005
16
A.C. Dillon, K.M Jones, T.A. Bekkedahl, C.H. Kiang, D.S. Bethune, M.J. Heben.
Storage of hydrogen in single-walled carbon nanotubes. Nature 386, 377, 1997

96
 3.6 References

17
J.L. Zimmerman, R.K. Bradley, C.B. Huffman, R.H. Hauge, J.L. Margrave. Gas-
phase purification of single-wall carbon nanotubes. Chem. Mater. 12, 1361,
2000
18
H. Murayama, T. Maeda. A novel form of filamentous graphite. Nature 345,
791, 1990
19
Y. Gogotsi, J.A. Libera, N. Kalashnikov, M. Yoshimura. Graphite polyhedral
crystals. Science 290, 317, 2000
20
M. Endo, B.J. Lee, Y.A. Kim, Y.J. Kim, H. Muramatsu, T. Yanagisawa, T.
Hayashi, M. Terrones, M.S. Dresselhaus. Transitional behaviour in the
transformation from active end planes to stable loops caused by annealing.
New Journal of Physics 5, 121.1, 2003
21
Z. Liu, K. Suenaga, P.J.F. Harris, S. Iijima. Open and closed edges of graphene
layers. Phys. Rev. Lett. 102, 015501, 2009
22
M. José-Yacamán, H. Terrones, L. Rendón, J.M. Domínguez. Carbon structures
grown from decomposition of a phenylacetylene and thiophene mixture on Ni
nanoparticles. Carbon 33, 669, 1995
23
J. Campos-Delgado, Y.A. Kim, T. Hayashi, A. Morelos-Gómez, M. Hofmann, H.
Muramatsu, M. Endo, H. Terrones, R.D. Shull, M.S. Dresselhaus, M. Terrones.
Thermal stability studies of CVD-grown graphene nanoribbons: defect
annealing and loop formation. Chem. Phys. Lett. 469, 177-182, 2009
24
T.C. Chieu, M.S. Dresselhaus. Raman studies of benzene-derived graphite
fibers. Phys. Rev. B 26, 10, 1982
25
A.M. Rao, A.W.P. Fung, M.S. Dresselhaus, M. Endo. Structural characterization
of heat-treated activated carbon fibers. J. Mater. Res. 7, 7, 1788-1794, 1992
26
M. Terrones, PhD. Thesis, University of Sussex, Sussex, UK, 1997
27
C. Jin, K Suenaga, S. Iijima. Vacancy migrations in carbon nanotubes.
Nanoletters 8, 1127, 2008

97
 3.6 References

28
S.V. Rotkin. SWNT nucleation: energetics of zipping-edge mechanism. Mat.
Res. Soc. Symp. Proc. 675, W2.9.1, 2001
29
M.S. Dresselhaus. G. Dresselhaus. K. Sugihara, I.L. Spain, H.A. Goldberg.
Springer Series in Materials Science Vol. 5. Graphite Fibers and Filaments.
Springer-Verlag, New York/Berlin/Heidelberg, 1988.
30
R. Saito, A. Grüneis, G.G. Samsonidze, V.W.Brar, G. Dresselhaus, M.S.
Dresselhaus, A. Jorio, L.G. Cançado, C. Fantini, M.A. Pimenta, A.G. Souza-
Filho. Double resonance Raman spectroscopy of single-wall carbon nanotubes.
New Journal of Physics 5, 157.1-157.15, 2003
31
L.G. Cançado, K. Takai, T. Enoki, M. Endo, Y.A. Kim, H. Mizusaki, A. Jorio,
L.N. Coelho, R. Magalhães-Paniago, M.A. Pimenta. General equation for the
determination of the crystallite size La of nanographite by Raman spectroscopy.
Appl. Phys. Lett. 88, 163106, 2006
32
R. P. Vidano, D. B. Fischbach, L. J. Willis, T. M. Loehr. Observation of Raman
band shifting with excitation wavelength for carbons and graphites. Solid State
Communications 39, 341-344, 1981
33
M. J. Matthews, M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus, M. Endo.
Origin of the dispersive effects of the Raman D band in carbon materials. Phys.
Rev. B 59, 10, 6585, 1999
34
P. Lespade, A. Marchand, M. Couzi, F. Cruege. Caracterisation de matériaux
carbonés par microspectrometrie Raman. Carbon 22, 4/5, 375-385, 1984

98
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

4. In situ reconstruction of graphitic nanoribbons by

Joule heating

The process by which electric current is passed through a conductor releasing heat
is called Joule heating. This phenomenon was first studied by James Prescott
Joule in 1841. He deduced that the heat produced was proportional to the square
of the current, multiplied by the electrical resistance of the conductor.
Q ∝ I2 ⋅R
This relationship is known as Joule’s law1.

Joule heating has provided an efficient way to achieve high temperatures in many
carbon materials, and it has been demonstrated that carbon nanostructures could
be modified substantially by Joule heating.
In 2001 Avouris and co-workers2, reported the breakdown of single shells of
multiwall carbon nanotubes (MWNTs) under a constant high bias regime. The
removal of carbon shells from these conductors was observed both electrically and
with the use of atomic force microscopy after the experiment (see Figure 4.1).

Figure 4.1 (A) The partial electrical breakdown of

a MWNT at constant voltage stress proceeds in a
series of discrete steps corresponding to the loss
of individual carbon shells from the MWNT.
Equally spaced dotted lines emphasize the
surprisingly regular spacing. (B) Images of
partially broken MWNTs show clear thinning, with
a decrease in radius equal to the intershell
spacing (0.34 nm) times the number of
completed breakdown steps. The two segments
of this sample were independently thinned by 3
and 10 shells, as depicted by the color overlays
(Extracted from Reference 2).

99
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

Electrically, the breakdown of a single carbon

shell results in a partial conductance drop that
typically occurs within a few milliseconds.
When stressed at sufficiently high bias,
multiple independent drops occur as one
carbon shell after another is broken. The
resulting staircase of current as a function of
time is shown in Figure 4.1A, as eight shells
of a MWNT are sequentially removed. Atomic
force microscopy (AFM) was used to resolve
the gradual thinning of the studied MWNT
(see Figure 4.1B). Although in this work the
primary factor of the breakdown was
attributed to current-induced defect formation
and not to the Joule heating effect, a latter
report by Huang et al.3 proved that the
breakdown of the tubes originated from heat-
stimulated imperfections.
The Joule heating experiments by Huang and
co-workers were conducted inside a
transmission electron microscope, which

Figure 4.2 (a) The current time (I-t) curve allowed a real time observation of the
of the breakdown of the MWNT shown in structural transformations occurring to the
panels b-e. The numbers below the MWNT. For this experiment, a multiwall
plateau indicate the total wall number. (b)-
carbon nanotube was end contacted and kept
(e) Sequential HRTEM images showing
free standing in high vacuum (10-8 Torr),
that the walls of the six-wall nanotube are
removed wall-by-wall from the outermost inside a HRTEM equipped with a scanning
wall (b) to the innermost wall (e). The tunneling microscopy (STM) probe. In such
numbers indicate the total number of an arrangement, atomic-scale imaging and I-
walls. The arrows mark kinks. The
th
V measurements were carried out
arrowheads denote the residue of the 4
rd 3 concurrently for the first time.
and the 3 walls after breakdown.

100
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

When passing a current of 240 µA, breakdown occurs at the midpoint of the
nanotube, resulting in the formation of a clean six-wall nanotube segment, which
was eliminated wall-by-wall by electric breakdown. The loss of one wall under a
constant bias of 3 V results in an instant current drop, as shown in Fig. 4.2a. The
current drops are approximately 13, 17, 25, and 31 µA for the 6th, 5th, 4th, and 3rd
wall breakdown (the innermost wall is labeled as the first wall), respectively.

The sequential wall-by-wall breakdown was recorded by HRTEM. Figure 4.2b-e

shows clearly that one wall is removed after each current drop, and each wall is
removed sequentially from the outermost wall (5th) to the innermost wall (2nd).
HRTEM indicates that the breakdown eventually results in the formation of a
double-wall nanotube. Their observation (Figure 1) that the breakdown is initiated
in the middle of the nanotube (not in the contacts) indicates that it is caused by
resistive heating. This is also supported by the defect generation and migration
along the nanotubes.
Typical defects observed are kinks, holes, and sliding between different walls.
These are apparently heat-stimulated imperfections.

Figure 4.3 a-d Tensile elongation of a single-wall carbon nanotube under a constant bias of 2.3 V
(images are scaled to the same magnification). Arrowheads mark kinks; arrows indicate features at
the ends of the nanotube that are almost unchanged during elongation. (Extracted from Ref. 4)

101
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

From the lattice spacing of 4 Å, they estimate that the temperature in the middle of
nanotube is between 2000 and 3000 ºC at the breakdown voltage of 3 V.
From the Fourier law, the middle section of the nanotube is the hottest spot;
therefore it is not difficult to understand that the breakdown occurs in the middle of
the nanotube, starting either from the outermost wall or from the innermost wall.3

The same set-up was used by Huang et al. in 2006 to prove the superplastic
properties of carbon nanotubes4 and to achieve the formation of nanotubes from
amorphous carbon nanowires5.

Huang and coworkers showed that, at high temperatures, individual SWNTs can
undergo superplastic deformation, becoming nearly 280 % longer and 15 times
narrower before breaking.
In their experiment, a SWNT with an initial length of 24 nm (Figure 4.3a) was
formed in situ by the electrical breakdown of a multwalled carbon nanotube inside
a high resolution transmission electron microscope equipped with a piezo
manipulator (STM tip).

The piezo manipulator was used to pull the SWNT to increase the strain (Figure
4.3 b-d), at a constant bias of 2.3 volts. At tensile failure, the SWNT was 91 nm
long, showing a tensile elongation of 280 %; its diameter was reduced 15-fold, from
12 to 0.8 nm.

During deformation at a bias of 2.3 V, the temperature is estimated to be more than

2000 ºC. At such high temperatures, the nanotube seems to be completely ductile.
Kinks and point defects are fully activated, resulting in superplastic deformation
impossible at low temperatures.

Kinks form frequently during tensile straining (Figure 4.3 b-d), and these kinks
propagate along the tube and then pile up (Figure 4.3d) or disappear at the ends.

102
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

The kink motion is evidence of kink mediated plasticity at high temperatures. The
nanotube narrows immediately after the kink passes.

Figure 4.4 Crystallization of a disordered nanotube caused by an I-V measurement. a) a nanotube

with disordered wall structures. b) crystallization of a) caused by making an I-V measurement, c) the
graphitization of b) was improved after holding at a high power (350 µW) for 3 min. (Extracted from
Reference 5).

In situ formation of nanotubes was accomplished when nanowires of amorphous

carbon were deposited by the electron-beam on the STM tip.

After growth, the nanowire was bonded to another electrode by deposition of

amorphous carbon between the nanowire tip and the electrode and then resistively
heated to temperatures higher than 2000 ºC by applying a high-bias voltage.

The high temperatures promoted the graphitization of the structure, leading to the
formation of concentric nanotube walls (see Figure 4.4). These experiments helped
to understand the dynamics of the non-catalytic growth of carbon nanotubes.

103
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

Figure 4.5 illustrates the experiment they

performed, where a double wall carbon
nanotube was first damaged by electron
beam irradiation (Figure 4.5a, no voltage
is applied), creating adatoms (black
spots) and monovacancies (white spots).
When a voltage up to 0.7 V was
gradually applied, most of the dark spots
are wiped out although the bright spots
remain there, as shown in Figure 4.5b
(also shown schematically in Figure
4.5bs).
In addition to the disappearance of
carbon adatoms and clusters, a few
vacancies with an enlarged size were
newly formed on the nanotube wall
(Figure 4.5b). They might correspond to
multivacancies such as di-, triple-, or
octavacancies created either by the
coalescence of nearby monovacancies or
through the further carbon loss around
the monovacancies. When they further
Figure 4.5 Time sequential HR-TEM images for
raised the temperature by increasing the
the formation, migration, and coalescence of
large vacancies in a DWNT (a–h) and their input voltage to 1.1 V (a current about 78
respectively schematic images (as-hs). Scale µA), most of the smaller vacancies
bar = 5 nm. disappeared and no more obvious
carbon adatoms or clusters (dark
6
The japanese group led by Iijima , contrast spots) could be observed, as
reported in 2008 the formation and shown in Figure 4.5c (also shown
dynamics of vacancies in carbon schematically in Figure 4.5cs).
nanotubes in a Joule heating regime.

104
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

Instead, a few even larger vacancies with a diameter of about

1–2 nm were created, corresponding to tens of carbon atoms loss. They were
mostly formed due to the coalescence of those mobile smaller vacancies in
neighboring hexagons.
Examples of even larger vacancies are given in Figure 4.5d-h (also shown
schematically in Figure 4.5ds-hs), where particular attention is paid on three large
vacancies marked as A, B, and C respectively in Figure 4.5d. Vacancies A and B
are separated by a center-to-center distance of about 7 nm at t = 170 s (Figure
4.5d). Vacancy B migrates toward A, and vacancy C migrates downward to the
sidewall, while vacancy A strolls around its original position. The migration of
vacancies B and C is quite smooth, and they do not cause any massive structural
distortions on the nanotube walls. Although each of them is trying to keep an
energy-favored round shape and their edge-shapes are slightly and frequently
changing, suggesting the diffusion of carbon atoms through the edge would
become easier at this high temperature. It takes about 18 s for vacancy B to arrive
at vacancy A.
Once they get in touch, coalescence starts gradually and a narrow neck forms
between them. Within 0.5 s, the coalescence process is completed and a much
larger hole is formed. Soon after, the edge of the hole is annealed again into a
round-shaped one with a diameter of about 4 nm, and its surface area is roughly
equal to the sum of vacancies A and B.
More importantly and surprisingly, this large hole is still mobile. It moves upward
and finally migrates to the left end of this DWNT (images not shown here). When it
reaches the sidewall, we could recognize that the large hole lies in the outer shell
of the DWNT because a portion of the outer wall disappears, not the inner wall,
and consequently so do vacancies A, B, and C.

Note that this newly formed large hole could survive without any splitting, even at
room temperature after removing the applied voltage. In the above experiments,
we find the vacancies in the outer wall are definitely more likely to migrate. This is
reasonable because most of the current is supposed to flow through the outmost

105
 4. In situ reconstruction of graphitic nanoribbons by Joule heating

shells, therefore the “hotter” outer wall is more favorable for the formation and
migration of vacancy, and a CNT with a larger diameter is more likely to
accommodate a large vacancy due to the weakened curvature effect. The
presence of inner wall is also playing a key factor here. When a large hole is
formed on the outer wall by losing tens of carbon atoms or even more, the outer
shell is hardly able to reconstruct by diameter shrinkage because of the existence
of inner walls.

By adopting the previously proposed thermal conductively for a suspended CNT,

the temperature on the CNT is estimated to range from ∼1000 to ∼1800 K. For the
migration of the vacancy B in Figure 4.5, we know the traveling distance within a
given time (about 7.3 nm for 18 s). The calculated vacancy migration speed is 0.4
nm/s.

The Joule heating experiments on our graphitic nanoribbons that we will describe
in this chapter were conducted using the facilities of the Department of Physics in
Boston College in Boston, Massachusetts, USA.

The experiments that we have carried out in this system have provided two
different and interesting results. On one hand, we were able to produce for the first
time atomically smooth zigzag or armchair edges from defective rough edges
present in graphite nanoribbons, by applying a controlled voltage7.

On the other hand, the increment in temperature induced loop formation between
adjacent sheets8 (single-, double- and multiple-loops), as described previously for
the furnace heating experiments (see Chapter 3 section 3). We will describe in
detail both processes and list the differences in sample preparation that lead to
each phenomenon.

106
 4.1 Conditions of the experiment

4.1 Conditions of the experiment

Resistive Joule heating experiments on the graphitic nanoribbons were carried out
using an integrated STM-TEM system (a JEOL 2010F HRTEM integrated with a
Nanofactory STM holder) which is further attached to a piezoelectric stage that
allows us to move the tip in all 3 (x, y, z) directions (see Figure 4.6).

The STM tip and sample holder also serve as two electrodes.
The as-prepared nanoribbon sample is placed in between the STM tip and the
sample holder, and a controlled bias voltage was applied across the ribbon.8

Figure 4.6 A schematic diagram of the TEM-STM system. The inset on the right shows a low
magnification TEM image of a nanoribbon piece placed in between two electrodes (the sample
8
holder and the STM tip).

107
 4.2 Crystallization and sharp edge formation

4.2 Crystallization and sharp edge formation

A short and narrow graphitic nanoribbon (315 nm long and 66 nm wide) was
placed inside the Joule heating set-up described above (inset of Figure 4.8A),
before a controlled voltage was applied, the structure of the nanoribbon was
intentionally damaged through irradiation of the electron beam for a long-time
period (~15 min). The effects of this irradiation are evident in Figure 4.8A in the
left-hand panel and Figure 4.10A, where the structure of the nanoribbon before the
Joule Heating experiment is shown.
Comparing this image with the pristine nanoribbon morphology (Figure 3.3 b and c,
Figure 3.8 Pristine) we can not help noticing that the sheets edges became
irregular and unparallel probably due to the evaporation of C species, triggered by
e-beam irradiation.
Once a controlled bias voltage was applied through the nanoribbon, the effects
induced by Joule heating were monitored by HRTEM and by changes in the
current vs. voltage (I-V) curves. The material behaved like a metal (straight line for
the I-V curves) following Ohm’s law.
As we applied a voltage (up to 1.6 V) over the length of the ribbon, the I-V curve,
depicted in Figure 4.7A, showed three distinct regimes: i) a linear regime from 1 to
1.25 V (which lasted 7 min and 30 sec), ii) a slowly increasing slope regime from
1.25 to 1.6 V (for ~ 2 min), and iii) a rapidly increasing slope regime at 1.6 V (for a
lapse of ~ 27 minutes). In the nonlinear regime at 1.6 V, the resistance decreases
with increasing input energy, this tendency is observed in Figure 4.7B, where
resistance versus input energy is plotted. As current flows at an applied voltage of
1.6 V, a restructuring process takes place, where the degree of crystallinity of the
ribbon improves rapidly causing the resistance values to drop (Figures 4.8D and
4.9), and the sample thickness decreases, until all the graphene layers evaporate
and the sample breaks from the middle (Figure 4.9C).
During the experiment, linear I-V sweeps were performed. Figure 4.8 shows
representative images at different stages of the experiment (left-hand panels) and
their corresponding I-V curves (right-hand panels).

108
 4.2 Crystallization and sharp edge formation

Figure 4.7 A) I vs. V curve during Joule heating, indicating three regimes: i) a linear regime from 1
to 1.25 V, ii) a slowly increasing slope regime from 1.25 to 1.6 V, and iii) a rapidly increasing slope
regime at 1.6 V. B) Resistance versus input energy at 1.6 V applied bias (Adapted from Ref. 6).

Before Joule Heating (Figure 4.8A) the structure, as described before, was
damaged by the electron beam and consisted of graphene sheets with irregular
and unparallel edges. The resistance value at this point corresponds to 20 kΩ.
When 1 V of bias voltage was applied (Figure 4.8 B and C), the structure of the
nanoribbon suffered major transformations, a progressive crystallization process
started taking place at this voltage, the irregular lines representing the sheets
edges seen in Figure 4.8B commence to straighten up and align in Figure 4.8C;
the transformations are also reflected in the change of resistance values, dropping
from 11.4 kΩ to 10.5 kΩ. During the highest voltage regime (1.6 V, Figure 4.8D)
the sample decreases its resistance value to 9.1 kΩ and below, graphitization is
fully achieved and we observe a defect-free honey-comb structure and the
formation of sharp edges (Figure 4.9b). Arrays of zig-zag (armchair) edges are
formed and the evolution of AA- to AB- stacking order is also achieved.7
We believe that the high temperatures induced by Joule Heating are responsible
for the crystallization process. The e-beam irradiation played also a crucial role in
the sharp edge formation as described above. Both effects combined, eventually
led to the rupture of the nanoribbon (Figure 4.9C).
At high applied voltages (1.6 V), the preferred reconstruction-crystallization effects
induced by the high temperature caused by Joule heating are seen in going from
Figure 4.10 A to B.

109
 4.2 Crystallization and sharp edge formation

Figure 4.8 Representative HRTEM images at different stages of the experiment (left-hand panels)
and their corresponding I-V curves (right-hand panels). A) before Joule Heating, the inset in the left
panel shows the nanoribbon attached to both electrodes, B) and C) at an applied voltage of 1 V, D)
at an applied voltage of 1.6 V. Resistance values are calculated by fitting linearly the data and
taking the inverse of the slope values.

110
 4.2 Crystallization and sharp edge formation

The well-resolved HRTEM images allowed us to determine the edge configuration

by looking at the hexagons orientation (see Figures 4.9B and 4.10D). Before Joule
heating the edges seen in the material are irregular and very few correspond to
zigzag (pink lines) or armchair segments (green lines). After 10 min of annealing at
1.6 V, the reconstruction process induced the formation of straight edges in either
armchair or zigzag configurations, indicating that the activation energy of atoms
forming zigzag or armchair edges is lower than for other edge configurations. Other
types of edges are seen infrequently because a mixture of zigzag and armchair
edges is metastable due to an energy penalty at the edge junctions.

Figure 4.9 TEM images of the Joule Heating experiment at an applied bias voltage of 1.6 V. A)
Image of the nanoribbon before rupture, B) high resolution image of the selected area in A; the inset
shows a detail of the marked region where the highly crystalline honey-comb lattice of carbon atoms
is evident. C) Low resolution image of the nanoribbon after rupture.

Figure 4.10D depicts the reconstructed graphitic material shown in Figure 4.10B.
The measured in-plane lattice spacing for our ribbons is 0.24 ± 0.02 nm, consistent
(within the accuracy of our TEM measurements) with literature values for graphite
[which is 3 ac–c, where ac–c is the nearest-neighbor carbon-to-carbon distance9].
As a result of carbon atom vaporization and Joule heating, the defective graphitic
edges in the as-grown nanoribbon sample crystallize and finally become atomically
sharp and highly crystalline. The maximum length of the smooth edges observed
after the process in Fig. 4.10C is ~29 nm. The mechanism of reconstruction or
crystallization for the nanoribbons and edges is attributed primarily to the initial
edge configuration of the material (irregular and unparallel), carbon atom

111
 4.2 Crystallization and sharp edge formation

vaporization, the current flow along the ribbon and edges, and the high
temperature associated with the resistive Joule heating.

Figure 4.10 Crystallization and edge formation in graphitic nanoribbons. A) Nanoribbon sample
before Joule heating, showing very few zigzag (pink lines) and armchair edges (green lines). B) The
same ribbon sample after Joule heating (for 10 min at 1.6 V), in which most of the edges seen are
either zigzag or armchair edges, as indicated in C). The inset hexagons indicate the zigzag- or
armchair-edge orientations associated with the lattice patterns in A) and C). D) High- magnification
image of the annealed sample showing that well-defined zigzag-armchair and zigzag-zigzag edges
are formed. Scale bars in A), B) and C), 4 nm; in D), 1 nm (Adapted from Reference 7).

The graphitization steps of carbon at high temperatures have been described by

Goma and Oberlin10. According to their observations, the graphitization process
happens in two phases, 1) a pregraphitization stage, below 2000 ºC, where small
mosaic elements about 10 Å in size become associated in an almost parallel
alignment with tilt and twist boundaries to form distorted layers; 2) above 2000 ºC
dewrinkling of carbon layers occurs and La (the in-plane crystallite size) increases
markedly, triperiodic order is achieved. For our ribbon samples, we observed the
transformation of AA stacking into ABAB stacking, and we attribute these

112
 4.2 Crystallization and sharp edge formation

transformations mainly to the high temperatures reached (above 2000°C) in the

nonlinear voltage regime. The dynamics of the edge reconstruction are shown in
Fig 4.11, A to D. In this figure we see a time resolved sequence of a three layered
graphene edge structure placed on top of another few-layered graphene sheet with
a zigzag-armchair-zigzag-armchair edge configuration forming 150° angles (note
that the angles between zigzag and armchair edges can be 30°, 90°, or 150°). As
we apply a bias voltage of 1.6 V (reaching high annealing temperatures), the
armchair edge above the zigzag edge starts to evaporate, resulting in the upward
movement of the zigzag edge (red arrows) with a speed of 2.3 nm/min. This low
speed at the investigated high temperatures indicates that the necessary activation
energy is much larger than the activation energy of defect migration observed by
Jin et al.6 in carbon nanotubes.

Figure 4.11 Edge motion under Joule heating. A) Three-layer zigzag-armchair-zigzag-armchair

edge array. The red arrows indicate the position of the zigzag-armchair edge junction. After some
time of Joule heating, the junction moves up [B and C], keeping the short zigzag-edge length almost
unchanged. Eventually, the zigzag edge joins with the upper zigzag edge, forming a stable zigzag-
zigzag-armchair edge array (D). The sketches on the left of A) and on the right of D) are simulated
7
structures of A) and D), respectively. Scale bar in A, 2 nm.

Eventually, as shown in Figure 4.11D, the armchair edge is eliminated, and the
lower zigzag edge joins with the upper zigzag edge and forms a stable zigzag-
zigzag junction.

113
 4.2 Crystallization and sharp edge formation

Joule heating involves both current flow and atomic vibrations. Point (localized)
defects are associated with large amplitude vibrations, and these are likely to be
annealed first. In addition, zigzag edged graphene ribbons are the only graphene
structures which have electronic states that are localized along their edges11.

Figure 4.12 Transport properties of graphene

nanoribbon heterojunctions based on a single
π-orbital tight-binding model (the Fermi energy
is located at E=0 eV). A) to C) Three models
considered here, where the details of the
edges are highlighted in red. In each case, the
two electrodes consist of two zigzag-edge
nanoribbons. The conductance corresponding
to systems A) to C) is presented in D.
(Adapted from Ref. 7).

Because the electronic dispersion is

quite large, the electronic flow in
zigzag-edged nanoribbons occurs
mainly along these zigzag edges. When a zigzag edge meets a non-zigzag edge,
the electronic flow is reduced, and the system acts as if a large resistance were
introduced at the junction.
Therefore, heating will result and will cause electron flow away from the edge
junction or, if enough energy is dissipated at the junction, will result in a
modification of the structure. In that case, provided that sufficient energy is
dissipated in this manner, the atomic structure will rearrange locally until electronic
flow is reestablished. Starting from a zigzag edge, this can take place if the
structure is annealed into a zigzag edge. For this reason, zigzag-zigzag junctions
are the favored junction formation (Figure 4.11). This zigzag edge formation
mechanism is therefore local, because it is due to a high local resistance at the
edge intersection. The conductance of the heterojunctions between zigzag and
armchair edges is considerably reduced because of the fact that transmission is
hindered on those non-zigzag edges where no electronic state is present.

114
 4.2 Crystallization and sharp edge formation

The Joule annealing process and its associated current flow create stable “edge
arrays” (see Figures 4.10C and 4.11 A-D). These edge arrays indicate that an
offset is present between the edges of two adjacent layers. The results of a careful
measurement of the edge-to-edge distances in these edge arrays is plotted in Fig.
4.13a, which shows clear peaks at 0.34-, 0.38-, and 0.43-nm edge-to-edge
distances.

Figure 4.13 Edge arrays and their time evolution. a) The experimental edge-to-edge distances in
the edge arrays show three main peaks at 0.34. 0.38 and 0.43 nm. These peaks correspond to the
separation (off-set) of the two adjacent AB stacked zigzag edges, AB stacked armchair edges, and
AA stacked zigzag edges, respectively, as indicated in the inset of a). b) Time evolution of the
7
stacking of all the edges obtained from analysis of Figure 4.9. Error of points is 2%.

115
 4.2 Crystallization and sharp edge formation

These distances correspond to the offset of adjacent edges of differently stacked

graphene layers (AB zigzag, AB armchair, AA zigzag). Of particular interest, the
evolution of the edge arrays (Fig. 413b) after 10 min of irradiation and Joule
annealing shows a clear increase for the AB stacked zigzag edges and a decrease of
the AA stacked zigzag edges.
The percentage of AB stacked zigzag edges increases from 17 to 32 %, whereas
the percentage of AA stacked zigzag edges decreases from 17 to 13% after a 10-
min annealing treatment, indicating that the AB stacked layer configuration is more
stable than AA stacking. This observation is attributed to the fact that the ABAB
stacked configuration is thermodynamically more stable than AA stacking12 and is
consistent with the Goma-Oberlin mechanism9.

4.3 Loop formation at the edges

Resistive Joule heating experiments on the as-produced graphitic nanoribbons

(consciously minimizing the electron beam exposure before and after the constant
bias voltage was applied) were carried out using the integrated scanning tunnel
microscope (STM)-TEM system described in Section 4.1. Figure 4.15 a-d depicts
TEM images of the nanoribbon edges after Joule heating. As we can see from
these images, in the interior region of the ribbon material, the originally wavy
fringes of the pristine material (see Figures 3.8 Pristine and 4.14a) are developed
into close packed straight lines, indicating an increased crystallinity of the graphitic
nanoribbon material, which is consistent with our observation of furnace heated
samples at 1500 ºC and 2800 ºC (Figures 3.8 HT 1500 ºC and 4.14 b and c).
Along the edges of the ribbon material, the open ends disappear, thus forming
loops of different shapes (Figure 4.15a-d). The loops formed near the electrodes
(Figure 4.15a) reveal a larger degree of curvature with a minimum of 2-3 layers
contained within the loops, a result which is consistent with the furnace-heated
samples13 in the temperature range between 2000 ºC - 2500 ºC.
The loops formed further away from the electrodes (Fig. 4.15 b-c) display a smaller
curvature with more than 10 layers contained on average within a single loop.

116
 4.3 Loop formation at the edges

Figure 4.14 TEM micrographs of the edges of graphitic nanoribbons annealed in furnace: a) pristine
sample, b) heat treated sample at 1500 ºC, and c) heat treated sample at 2800 ºC. The scale bars
8
for the three images correspond to 5 nm.

Finally, the loops formed near the central region of the ribbon material (Figure
4.15d) reveal larger loops and these even contain some facets (graphitization
effect). Compared with the furnace heat treatment at 2800 ºC (Figure 4.14c) where
multiple loops (~6 layers) are formed, Joule heating (Figure 4.15c-d) results in loop
formation with more layers (>10 layers), which might indicate that a higher local
temperature is reached.

The effect of loop formation by Joule heating is primarily attributed to the high
temperature achieved by resistive Joule heating.
We attribute the different shapes of the loops, that is, smaller loops (double or
triple) near the electrodes (Figure 4.15a), and larger multiple loops and faceted
formations near the central region of the ribbon (Figure 4.15d), to the larger
temperature difference across the ribbon.

In our Joule heating experiments, the two electrodes serve as heat sinks.
Due to the achievement of good thermal conductivity near the electrodes, the
temperature near the electrodes is much lower than that in the central region of the
ribbon.
The formation of different loop structures along the nanoribbon is a direct evidence
of temperature gradient.

117
 4.3 Loop formation at the edges

Figure 4.15 TEM images showing loop formation by Joule heating: a) near the electrode, b) and c)
further away from the electrodes than in (a), and d) near the central region of the suspended
graphitic nanoribbon material. (Scale bar is 5 nm)8.

This result is also consistent with the furnace heating experiments, which showed
that different shapes of loops are formed at different annealing temperatures
(Figure 3.8 and 4.14).

4.4 Estimation of the achieved temperatures

In order to estimate the temperature of the graphitic nanoribbons during the Joule
heating experiment, Pt nanoparticles were deposited chemically on the as-

118
 4.4 Estimation of the achieved temperatures

prepared nanoribbon surface (Figure 4.16a), and the structural changes in the Pt
nanoparticles were monitored in-situ.
The Pt anchoring process consisted of sonicating for 15 minutes a mixture of
graphitic nanoribbons (10 mg), plus 10 ml of N,N-dimethylformamide (Sigma-
Aldrich®, 99% ), (1,5-Cyclooctadiene) dimethylplatinum (II) (Sigma-Aldrich ®,
97%) as a platinum source and polyvinylpyrrolidone (Sigma-Aldrich®, average mol
wt. 10,000) as a passivating agent.
After sonication, the suspension was maintained under an Ar-H2 (5% H2)
atmosphere to increase the reduction rate, and the suspension was subsequently
placed in a glycerin bath at 110 ºC for 40 minutes.
Next, the suspension was allowed to cool down to room temperature and the
composite material (graphitic nanoribbons with Pt particles) was recovered by
filtration. These graphitic nanoribbons exhibited platinum nanoparticles (with an
average size of ∼6 nm) anchored to their surface (see Figure 4.16a). Finally a
thermal treatment was carried out at 350 ºC under an Ar atmosphere for 15
minutes in order to eliminate any residues of organic material that could remain on
the surface of the composite material.

Figure 4.16 A sequence of TEM images showing Pt nanoparticles on the ribbon surface (a) before
Joule heating, (b) after Joule heating for 11 minutes under a constant bias of ~2V, and (c) after
Joule heating for another 4 minutes under a constant bias of ~2V. Here we see that the Pt particles
melt and merge into bigger clusters (b), and start to evaporate from the central region of the ribbon
8
(b), and eventually evaporate across almost the entire ribbon sample (c). (Scale bar is 100 nm).

119
 4.4 Estimation of the achieved temperatures

The modified nanoribbon material was then mounted on the Joule heating set-up.
As we increased the applied voltage across the nanoribbons, the Pt nanoparticles
near the central region of the ribbon started to melt and merge with other small
neighboring Pt nanoparticles (some particles finally reached a diameter of 13 nm).
Subsequently, and starting from the central region, the Pt nanoparticles
evaporated, resulting in a clean surface (devoid of Pt nanoparticles) near the
center of the ribbon sample (Figure 4.16b).
When a higher voltage is applied, additional Pt nanoparticles evaporate and
eventually almost the entire ribbon is free of Pt nanoparticles (Figure 4.16c).
From these experiments, we confirmed that good thermal contacts are made near
the electrodes, and that the two electrodes serve as heat sinks. The central region
of the ribbon (Figure 4.14c) exhibits the highest temperature at a given applied
voltage. Given the bulk Pt boiling point of 3827 ºC, and that the boiling point of the
Pt nanoparticles will have a lower boiling point than their bulk counterpart material
due to size effects14, we estimate the temperature of the suspended ribbon sample
under Joule heating to be above 2800 ºC based on the loop formation morphology
(comparison of Figures 3.8 HT 2800 ºC and 4.14d).

4.5 Conclusions

As we have carefully described in this chapter, Joule heating experiments of

graphitic nanoribbons lead either to sharp edge reconstruction or to loop
termination of the edges, depending on the treatment given to the starting material.

When an as-produced nanoribbon is submitted to Joule heating (minimizing e-

beam exposure), there is a significant rise in temperature.
This increase in temperature results in the annealing of defects, in-plane
crystallization, and for the edges, the adjacent sheets (with parallel edges) find a
more stable configuration forming loops, as seen in furnace heating13, and Iijima's
recent work15 (see Figure 3.5d).

120
 4.5 Conclusions

The configuration of the loops formed along the nanoribbon in this Joule heating
experiment, gave irrefutable evidence of a temperature gradient, where the highest
temperature is found in the center of the nanostructure.

When the same experiment is conducted in a nanoribbon specie that has been
irradiated with the electron beam (for the purpose of imaging or to intentionally
damage the structure), the results are completely different

Our experiments show that e-beam irradiation of the sample prior to Joule heating
(between 10 min - 15 min) plays a decisive role in the reconstruction process.
This irradiation damages the structure, leading to the creation of vacancies and,
most importantly, the sheets edges become irregular and unparallel probably due
to the indistinctive evaporation of carbon atoms.

The irregularity of the edges does not favor the loop formation but the sharp edge
reconstruction (primarily zigzag or armchair configurations) in the Joule heating
process.

We have also carried out experiments of indirect temperature estimation during the
Joule Heating treatments by depositing platinum nanoparticles (~6 nm in diameter)
on the nanoribbons.

Our results showed that the temperatures achieved are high enough to melt and
evaporate the Pt nanoparticles located in the middle of the nanoribbons.
We are certain that the achieved temperatures surpass the 2800 ºC at the center of
the studied nanostructure, consistent with Huang’s stimations3-5 on experiments
conducted in the same set-up using carbon nanotubes.

121
 4.6 Related articles

4.6 Related articles

1 . X. Jia, J. Campos-Delgado, E.E. Gracia-Espino, M. Hofmann, H.

Muramatsu, Y.A. Kim, T. Hayashi, M. Endo, J. Kong, M. Terrones, M.S.
Dresselhaus. Loop formation in graphitic nanoribbon edges using furnace heating
or Joule heating. J. Vac. Sci. Technol. B 27, 4, Jul/Aug, 1996-2002, 2009

2 . X. Jia, M. Hofmann, V. Meunier, B.G. Sumpter, J. Campos-Delgado, J.M.

Romo-Herrera, H. Son, Y.-P. Hsieh, A. Reina, J. Kong, M. Terrones, M.S.
Dresselhaus. Controlled formation of Sharp zigzag and armchair edges in
graphitic nanoribbons. Science 323, 1701-1705, 2009

122
 4.7 References

4.7 References

1
Yves Quéré. Physics of Materials. Gordon and Breach Science Publishers.
Amsterdam, 1998
2
P.G. Collins, M.S. Arnold, Ph. Avouris. Engineering carbon nanotubes and
nanotube circuits using electrical breakdown. Science 292, 706-709, 2001
3
J.Y. Huang, S. Chen, S.H. Jo, Z. Wang, D.X. Han, G. Chen, M.S. Dresselhaus,
Z.F. Ren. Atomic-scale imaging of wall-by-wall breakdown and concurrent
transport measurements in multiwall carbon nanotubes. Phys. Rev. Lett. 94,
236802, 2005
4
J.Y. Huang, S. Chen, Z.Q. Wang, K. Kempa, Y.M. Wang, S.H. Jo, G. Chen,
M.S. Dresselhaus, Z.F. Ren. Superplastic carbon nanotubes. Nature 439, 281
2006
5
J.Y. Huang, S. Chen, Z.F. Ren, G. Chen, M.S. Dresselhaus. Real-time
observation of tubule formation from amorphous carbon nanowires under high-
bias Joule Heating. Nano Letters 6, 8, 1699-1705, 2006
6
C. Jin, K. Suenaga, S. Iijima. Vacancy migrations in carbon nanotubes.
Nanoletters 8, 4, 1127-1130, 2008.
7
X. Jia, M. Hofmann, V. Meunier, B.G. Sumpter, J. Campos-Delgado, J.M.
Romo-Herrera, H. Son, Y.-P. Hsieh, A. Reina, J. Kong, M. Terrones, M.S.
Dresselhaus. Controlled formation of Sharp zigzag and armchair edges in
graphitic nanoribbons. Science 323, 1701-1705, 2009
8
X. Jia, J. Campos-Delgado, E.E. Gracia-Espino, M. Hofmann, H. Muramatsu,
Y.A. Kim, T. Hayashi, M. Endo, J. Kong. M. Terrones, M.S. Dresselhaus. Loop
formation in graphitic nanoribbon edges using furnace heating or Joule heating.
J. Vac. Sci. Technol. B 27, 4, 1996-2002, 2009
9
B.T. Kelly. Physics of Graphite. Applied Science Publishers. London, 1981

123
 4.7 References

10
J. Goma, M. Oberlin. Graphitization of thin carbon films. Thin Solid Films 65,
221-232, 1980
11
K. Nakada, M. Fujita, G. Dresselhaus, M.S. Dresselhaus. Edge state in
graphene ribbons: nanometer size effect and edge shape dependence. Phys.
Rev. B 54, 24, 17954-17961, 1996
12
S. Latil, V. Meunier, L. Henrard. Massless fermions in multilayer graphitic
systems with misoriented layers: Ab initio calculations and experimental
fingerprints. Phys. Rev. B 76, 201402, 2007
13
J. Campos-Delgado, Y.A. Kim, T. Hayashi, A. Morelos-Gómez, M. Hofmann, H.
Muramatsu, M. Endo, H. Terrones, R.D. Shull, M.S. Dresselhaus, M. Terrones.
Thermal stability studies of CVD-grown graphene nanoribbons: defect
annealing and loop formation. Chem. Phys. Lett. 469, 177-182, 2009
14
M. Attarian Shandiz. Effective coordination number model for the size
dependency of physical properties of nanocrystals. J. Phys.: Condens. Matter
20, 325237, (2008).
15
Z. Liu, K. Suenaga, P.J.F. Harris, S. Iijima. Open and closed edges of graphene
layers. Phys. Rev. Lett. 102, 015501, 2009

124
 5. Conclusions and Perspectives

5. Conclusions and Perspectives

The synthesis of doped SWNTs and graphitic nanoribbons has been addressed
using an aerosol-assisted CVD method.
Using the floating catalyst technique, ethanol-ferrocene solutions were prepared
with precursor compounds of the doping element.
The synthesis of single-wall carbon nanotubes in sulphur, nitrogen, phosphorous
and silicon environments was achieved via aerosol-assisted-CVD.
Precursors containing the target doping element were mixed in ethanol-ferrocene
solutions at different concentrations, thiophene for sulphur incorporation,
triphenylphosphine was used in the case of phosphorous, methoxytrimethylsilane
was used as silicon precursor, and benzylamine and pyrazine were used as
nitrogen precursors.

Electron microscopy of the samples as well as RBM Raman signal confirmed the
presence of carbon nanotubes in the synthesized materials.

Electron microscopy analysis revealed that most of the materials consisted of

doped SWNTs entangled with metallic nanoparticles, in the case of S and Si,
interesting morphologies were formed, such as short bamboo-like MWNTs and Si
nanorods with metallic tips, respectively.

In particular, the incorporation of sulphur promoted a very dense formation of by-

products. The enhancement of the catalytic activity of iron by sulphur addition was
confirmed through the copious presence of short, twisted bamboo-like MWNTs
along with SWNTs. We believe that the modification of the experimental
parameters (e.g. reduction of ferrocene concentration) could yield to the
optimization of S-doped SWNTs.

The RBM analysis of our N- and P- doped materials showed that as the doping
precursor was increased in the sprayer solution, narrower diameter tubes were

125
 5. Conclusions and Perspectives

favored. The latter is consistent with theoretical calculations indicating that dopants
of heavier elements embedded in the hexagonal carbon lattice are more
energetically favored in narrower tubes exhibiting higher radii of curvature.

We have used the IG'Def /IG'Pris relative intensities as a direct doping index for N, P

and Si.
Our samples synthesized with sulphur could not be included in this analysis due to
the poor quality of the recorded Raman spectra. Hence we do not have strong
evidence for the effect of sulphur doping of SWNTs.

Si-doped samples showed low IG'Def /IG'Pris relative intensity values which are

directly related to the small amount of silicon atoms available per carbon atoms
during the synthesis. Nitrogen doping is more effective when pyrazine is used
(compared to benzylamine), and phosphorous doping is very effective even at low
TPP concentrations.
Our Raman records show that increasing the precursor concentration leads to

higher doping levels, increasing the I G'Def / I G'Pris relative intensities, and that the

frequency splitting of the G’ band depends more on the doping element than the
doping amount.
We are carrying out more synthesis experiments in order to have a higher number
of samples per doping element, to be able to run Raman spectroscopy
measurements and to elucidate the nature of the G’ band splitting peculiarities
observed in this work.

Our experiments for sulphur doping of SWNTs, led us to find a new route to
efficiently produce graphitic nanoribbons at atmospheric pressure. The process is
based on the pyrolytic decomposition at 950 ºC of solutions of ferrocene-ethanol-
thiophene at concentrations of 1.25 - 98.63 - 0.12 wt. %, respectively.

The morphology of the initial black powder that was collected consisted of ribbon-
like structures of graphitic layers aligned parallel to the nanoribbon main axis with

126
 5. Conclusions and Perspectives

average dimensions of several microns in length, widths ranging from 20-300 nm

and thicknesses of 10-20 nm. It is interesting to note that the initial ribbons
revealed both flat regions as well as rippled areas. The structure provides the
material with a high proportion of edges, which are passivated by the presence of
oxygen groups and sp3 hybridization.

We have demonstrated that heat treatments up to 2800 ºC of as-produced carbon

nanoribbons anneal point defects and interstitials within individual graphene layers,
thus leading to a material with a higher degree of crystallinity, in which the reactive
edges of the ribbons are transformed into loops in a self-assembly process that
can be strongly controlled by the variation of THT.

We have performed a detailed Resonance Raman study of heat treated

nanoribbons on both bulk and individual nanoribbon samples. At the bulk level, we
have confirmed a laser energy dependence of the D- and G’-bands frequency
positions; and in these dependence values we found differences within the HT
samples attributed to the annealing process.
The decrease of the ID/IG ratios and of the FWHM of the G band as the heat
treatment temperature is increased suggests an increase in crystallinity of the
hexagonal honeycomb lattice, a consequence of a partial graphitization process in
which the stacking order could not be achieved completely possibly caused by the
loop formation at the ribbon edges.

At the individual nanoribbon level, we have demonstrated that edge contributions

to the Raman spectra intensify the signal of the D-band, indicating that symmetry-
breaking elements such as sp3-like hybridized carbon atoms, structural defects and
loops are present particularly in the vicinity of the edges.
However, the large number of loops present at the edges of all the ribbons is
responsible for observing some residual D-band intensity. The general up-shift of
the decomposition temperature in the TGA analysis also reveals a more stable and

127
 5. Conclusions and Perspectives

less reactive material as THT increases in the range above which loop formation
takes over.

The annealing processes that resulted from the Joule Heating experiments, proved
to produce both, loops at the edges between parallel adjacent graphene sheets,
and sharp zigzag or armchair edges when the sheets have been previously
damaged by irradiation with an electron beam.
We have also carried out experiments of temperature estimation during the Joule
Heating treatments. Our results showed that the temperatures achieved surpass
the 2800 ºC.

The work reported in this thesis has opened new lines of research related to the
CVD synthesis of doped-SWNTs and graphitic nanoribbons.

Although a lot of experiments and characterization have been carried out, as this
thesis proves, the scientific curiosity keeps motivating us to imagine, conceive and
design new experiments in the search for new properties and new materials.

5. 1 Future work on doped-SWNTs

The optimization of sulphur-doping experiments is pending.

Although we were able to produce SWNTs in the presence of sulphur, the
generation of by-products was greatly enhanced, and the SWNT Raman signal
was screened in our measurements.

We believe that a lower concentration of ferrocene (~0.5 wt. % in contrast to the

1.25 wt. % used in our experiments reported in Chapter II), combined with the
sulphur effect will provide enough catalytic action to synthesize SWNTs, while
bigger Fe-S particles that promote the formation of short-twisted bamboo-like
MWTNs will not be as numerous.

128
 5.1 Future work on doped SWNTs

Also the use of Ar-H2 instead of pure Ar as a carrier gas could inhibit the formation
of amorphous carbon residues and extend the life-time of the catalyst particles.

The introduction of MTMS in the ethanol-ferrocene solutions to produce Si-doped

SWNTs at concentrations above 0.1 wt. % triggered the formation of interesting by-
products and decreased the SWNT formation. We believe that the modification of
experimental parameters (increase of temperature and flow rate) will increase the
velocity of the reaction and thus, promote the formation of SWNTs before the
catalytic particles become de-activated by the formation of Fe-Si alloys.

Our results proved that nitrogen doping is more efficient when pyrazine is used as
precursor (compared to benzylamine), evaluated through the IG'Def /IG'Pris relative

intensities. However, XPS measurements need to be performed in order to confirm

this idea. Quantification of P-doping should also be addressed using this
technique.

Testing the biocompatibility of cells with the different doped SWNTs is also
pending.

5.2 Future work on graphitic nanoribbons

We have performed a lot of characterization on our graphitic nanoribbons. XPS

measurements of pristine nanoribbons revealed a striking 1:1 ratio of sp2/sp3
hybridization.
Our observations lead us to think that this high proportion of sp3 hybridization can
be attributed to the elevated proportion of edges present in the sample.
This possible explanation needs to be confirmed by XPS measurements of the
heat treated nanoribbons. We expect to find a decreased signal of the sp3
hybridization due to the presence of loops which have passivated the edges in the
annealed samples.

129
 5.2 Future work on graphitic nanoribbons

So far we have worked with pristine and heat treated nanoribbons but we have not
done chemistry-related experiments. We believe that due to the nature of our
sample, we could chemically treat or exfoliate the sheets to end up with single-
layer or double layer graphene nanoribbons.

Another line of research could be to try to reinforce materials with our nanoribbons,
such as polymers and ceramics.

130
 APPENDIX A

APPENDIX A
Carbon nanotubes synthesis methods and characterization tools

Synthesis methods
Arc discharge
An arc discharge is generated when a bias voltage is applied between two metallic
electrodes, if the bias is sufficiently high to produce a discharge; the result is a
small current flowing through the circuit (both electrodes and the existing gas in
between).1

2
Figure A1 Diagram of an arc discharge chamber used to produce carbon nanotubes.

The carbon arc provides a simple and traditional tool for generating the high
temperatures needed for the vaporization of carbon atoms into a plasma (>3000
°C). Typical conditions for operating a carbon arc for the synthesis of carbon
nanotubes include the use of carbon rod electrodes of 5-20 mm diameter,
separated by ~1 mm with a voltage of 20-25 V across the electrodes and a DC
electric current of 50-120 A flowing between the electrodes. The arc is typically
operated in ~500 Torr He with a flow rate of 5-15 ml/s for cooling purposes. As the
carbon nanotubes form, the length of the positive electrode (anode) decreases,
and carbon material is deposited in the negative electrode (See Figure A1)3.
The first reports of single-wall carbon nanotube synthesis go back to 1993.
Independently, two groups achieved their synthesis and both reports appeared one
after the other in Nature magazine (Iijima, et al.4 and Bethune, et al.5, see Figure
A2). Both experiments were carried out in an arc discharge chamber with graphite

131
 APPENDIX A

electrodes (set-up used to produce C60), with some differences in the experimental
parameters.

Figure A2 TEM images of arc-discharge produced SWNTs, a) extracted from Reference 4, b)

extracted from reference 5

Catalysts used to prepare isolated single-wall carbon nanotubes (SWNTs) include

transition metals such as Co, Ni, Fe and rare earths such as Y and Gd, while
mixed catalysts such as Fe/Mi, Co/Ni and Co/Pt have been used to synthesize
ropes of SWNTs. The iron or cobalt catalyst in the arc process forms both SWNTs
and nanometer/size carbide particles surrounded by graphene layers, as well as
metal clusters encapsulated within graphene layers. Thus purification is necessary
to separate out a pure single-wall nanotube sample.4

Laser ablation
Using this technique the first report of SWNTs was published by Thess and co-
workers6 in 1996. A Co-Ni/graphite composite laser vaporization target was used,
consisting of 1.2 atom % Co-Ni alloy with equal amounts of Co and Ni added to the
graphite. Two sequenced laser pulses were used to evaporate the target placed
inside a furnace operated at 1200 ºC. Flowing argon gas sweeps the entrained
nanotubes from the high temperature zone to the water-cooled Cu collector
downstream, just outside the furnace, Figure A3 shows a schematic representation
of a typical laser ablation set-up.

132
 APPENDIX A

7
Figure A3 Diagram of the set-up used in the laser vaporization method to produce SWNTs

The material thus produced appears as a mat of ropes 10-20 nm in diameter and
up to 100 µm or more in length.

Figure A4 a) SEM micrograph of SWNT material showing a mat of tangled carbon fibers, b) TEM
image of a single rope of SWNTs made up of ~100 SWNTs as it bends through the image plane of
the microscope, showing uniform diameter and triangular packing of the tubes within the rope.
(Extracted from Ref. 6)

Under TEM examination, each “rope” consisted primarily of a bundle of SWNTs

aligned along a common axis, see Figure A4. The single-wall nanotubes are held
together by weak van der Waals inter-nanotube bonds to form a two-dimensional
triangular lattice with a lattice constant of 1.7 nm, and an inter-tube separation of
0.315 nm at closest approach within a rope.6

Chemical Vapor Deposition

Chemical vapor deposition (CVD) is a chemical process used to produce high-
purity, high-performance solid materials. The process is often used in the

133
 APPENDIX A

semiconductor industry to produce thin films. In a typical CVD process, the

substrate is exposed to one or more volatile precursors, which react and/or
decompose on the substrate surface to produce the desired deposit. Frequently,
volatile by-products are also produced, which are removed by gas flow through the
reaction chamber.
A number of forms of CVD are in wide use and are frequently referenced in the
literature (atmospheric pressure CVD, low-pressure CVD, ultrahigh vacuum CVD,
aerosol assisted CVD, direct liquid injection CVD, microwave plasma-assisted
CVD, plasma-enhanced CVD, remote plasma-enhanced CVD). 8

Figure A5 A) and B) extracted from reference 13, C) - G) extracted from reference 14. A) Optical
image showing a human hair and two as-grown SWNT strands (indicated by black arrows), B)
HRTEM image of a top view of a SWNT rope, white arrows indicate the arrangement of the triangle
lattice of a large area in a SWNT strand. The inset shows a cross-sectional view of a polycrystalline
bundle. C) SWNT forest grown with water-assisted CVD. Picture of a 2.5-mm-tall SWNTs forest on
a silicon wafer. A matchstick on the left and a ruler with millimeter markings on the right is for size
reference. D) SEM image of the same SWNT forest, E) SEM image of the SWNT forest ledge.
Scale bar = 1 µm. F) Low resolution TEM image of the nanotubes. Scale bar = 100 NM. G) High-
resolution TEM image of the SWNTs. Scale bar = 5 nm.

The synthesis of carbon nanotubes from the vapor phase utilizes equipment similar
to that used for the preparation of vapor-grown carbon fibers, with the furnace
temperature held at 1100 ºC and using Fe catalyst particles, but using a low
benzene gas pressure. A variety of other hydrocarbons, catalysts and catalyst

134
 APPENDIX A

supports have been used successfully by various groups worldwide to synthesize

carbon nanotubes.5
Other metallic catalysts used are Fe, Ni, Co and Pt. These are, either seeded as
nanoparticles in substrates previous to the CVD reaction or as “floating” catalysts,
where compounds containing the metals are mixed with the hydrocarbon at the
time of reaction. The most common hydrocarbons used are methane, benzene,
acetylene, toluene and ethylene. A typical CVD set-up contains a furnace, a
reaction chamber, Ar or He flow to keep an inert atmosphere and carry the gases
outside the furnace and the reaction zone.
The first report of CVD produced SWNTs was by H. Dai in 19969 decomposing CO
and using Mo as catalyst, and later works reported the use of different
hydrocarbons (methane10, ethylene11).
In 2002 Maruyama and co-workers12 reported the synthesis of high purity SWNTs
with this method. Using alcohol as the carbon source at low temperatures (700 ºC -
800 ºC, it is worth mentioning that low pressures are involved in the process) a
product free of impurities as amorphous carbon, multi-walled carbon nanotubes,
metal nanoparticles and carbon nanoparticles were completely suppressed by the
etching effect of the –OH radical.
In the same year, Zhu et al13 achieved the synthesis of long strands of SWNTs
(Figure 3A and B). Ferrocene and small concentrations of thiophene were mixed
with n-hexane and the pyrolysis temperature was 1150 ºC.
The so-called super-growth of SWNTs was achieved by Hata in 200414, when
SWNT forests of 2.5 mm were synthesized during 10 minutes of reaction (see
Figure 3 C-G). This group added minute amounts of water vapor to the Ar or He
flow. Catalysts were seeded in different substrates by sputtering and ethylene was
used as a carbon source.

Characterization tools
The characterization of carbon nanotubes and carbon nanostructures aims to
identify the morphology, dimensions, structure, chemical composition, physical,

135
 APPENDIX A

chemical and electronic properties. Researchers in nanosciences use several

techniques as standard characterization tools, some of which are described below.

Electron microscopy
Electron microscopy is a type of microscopy that uses a particle beam of highly
energetic electrons to illuminate a specimen and create a highly-magnified image.
Electron microscopes have much greater resolving power than light microscopes
that use electromagnetic radiation and can obtain much higher magnifications of up
to 2 million times. The greater resolution and magnification of the electron
microscope is because the wavelength of an electron, its de Broglie wavelength, is
much smaller than that of a photon of visible light.15
The basic steps involved in all electron microscopes are:
1. A stream of electrons is formed (by the electron source) and accelerated
toward the specimen using a positive electrical potential
2. This stream is confined and focused using metal apertures and magnetic
lenses into a thin, focused, monochromatic beam
3. This beam is focused onto the sample using a magnetic lens
4. Interactions occur inside the irradiated sample, affecting the electron beam
These interactions and effects are detected and transformed into an image.
The electron microscope uses electrostatic and electromagnetic lenses in forming
the image by controlling the electron beam to focus it at a specific plane relative to
the specimen.16

The Scanning Electron Microscope (SEM) is a type of electron microscope that

images the sample by scanning it with a high-energy beam of electrons in a raster
scan pattern. The electrons interact with the atoms that make up the sample
producing signals that contain information about the sample’s surface topography,
composition and other properties such as electrical conductivity.
The types of signals produced by an SEM include secondary electrons, back
scattered electrons, characteristic x-rays, light, specimen current and transmitted
electrons. These types of signal all require specialized detectors that are not

136
 APPENDIX A

usually all present on a single machine. For conventional imaging in the SEM,
specimens must be electrically conductive, at least at the surface, and electrically
grounded to prevent the accumulation of electrostatic charge at the surface.12
Transmission electron microscopy (TEM) is a microscopy technique whereby a
beam of electrons is transmitted through an ultra thin specimen, interacting with the
specimen as it passes through. An image is formed from the interaction of the
electrons transmitted through the specimen; the image is magnified and focused
onto an imaging device, such as a fluorescent screen, on a layer of photographic
film, or to be detected by a sensor such as a CCD camera.12

X-ray diffraction
X-ray crystallography is a method of determining the arrangement of atoms with a
crystal, in which a beam of X-rays strikes a crystal and scatters into many different
directions.17
Bragg’s law states that when X-rays hit an atom, they make the electronic cloud
move as does any electromagnetic wave. The movement of these charges re-
radiates waves with the same frequency; this phenomenon is known as Rayleigh
scattering (or elastic scattering). These re-emitted wave fields interfere with each
other either constructively or destructively (overlapping waves either add together
to produce stronger peaks or subtract from each other to some degree), producing
a diffraction pattern on a detector or film. The resulting wave interference pattern is
the basis of diffraction analysis.
The interference is constructive when the phase shift is a multiple of 2π; this
condition can be expressed by Bragg’s law
nλ = 2d sin θ
where n is an integer determined by the order given, λ is the wavelength of the X-
rays, d is the spacing between the planes in the atomic lattice and θ is the angle
between the incident ray and the scattering planes.
This technique has been used widely because it constitutes a non-destructive,
accurate method to investigate the atoms arrangement of a particular sample
through the diffraction pattern.

137
 APPENDIX A

Figure A6 X-ray diffraction profiles of a) a graphite rod, and b) a core of a deposit grown after arc
18
discharge. The core consists of bundles of carbon tubules and hollow nanoparticles

Carbon materials are often characterized by X-ray diffraction since it is a sensitive

tool to identify sp2 and sp3 hybridization through the (002) and (111) reflections,
see Figure A6. Impurities and by-products like metal nanoparticles produced in the
CVD synthesis of multi-wall carbon nanotubes are easily detected when strong iron
carbides reflections are present in the spectra.

XPS
X-ray photoelectron spectroscopy (XPS) is a surface analytical technique, which is
based upon the photoelectric effect. Each atom in the surface has core electrons
with the characteristic binding energy, that is conceptually, not strictly, equal to the
ionization energy of that electron. When an X-ray beam is directed to the sample
surface, the energy of the X-ray photon is adsorbed completely by the core
electron of an atom. If the photon energy, hν, is large enough, the core electron will
then escape from the atom and be emitted from of the surface. The emitted
electron with the kinetic energy of Ek is referred to as the photoelectron.
The binding energy is given by the relation
Eb = hν - Ek
The core electron of an element has a unique binding energy, which seems like a
“fingerprint”. Thus almost all elements except for hydrogen and helium can be
identified via measuring the binding energy of its core electron. Furthermore, the

138
 APPENDIX A

binding energy of the core electron is very sensitive to the chemical environment of
the element. The same atom when bonded to the different chemical species, leads
to a change in the binding energy of its core electron. Since the number of
photoelectrons of an element is dependent upon the atomic concentration of that
element in the sample, XPS is used to not only identify the elements but also
quantify the chemical composition.19

20
Figure A7 An example of XPS C1s peak deconvolution of MWNTs and the assigned bonds

Figure A7 is included to illustrate a representative XPS measurement from carbon

nanostructures, it is extracted from Reference 19 where an XPS study on
functionalized MWNTs was carried out. In the figure, the C1s signal is shown and a
deconvolution depicts the signal contribution from the different types of bonds (i.e.
C=C, C-C, C-O, C=O, COO).

EDX
EDX analysis stands for Energy Dispersive X-ray analysis. It is sometimes referred
to also as EDS or EDAX analysis. It is a technique used for identifying the
elemental composition of the specimen, or an area of interest thereof. The EDX
analysis system works as an integrated feature of an electron microscope, and can
not operate on its own without the latter.
During EDX analysis, the specimen is bombarded with an electron beam inside the
scanning electron microscope. The bombarding electrons collide with the specimen

139
 APPENDIX A

atoms’ own electrons, knocking some of them off in the process. A position
emptied by an ejected inner shell electron is eventually occupied by a higher-
energy electron from an outer shell. To be able to do so, however, the transferring
outer electron must give up some of its energy by emitting an X-ray.
The amount of energy released by the transferring electron depends on which shell
it is transferring from, as well as which shell it is transferring to. Furthermore, the
atom of every element releases X-rays with unique amounts of energy during the
transferring process. Thus, by measuring the amounts of energy present in the X-
rays being released by a specimen during electron beam bombardment, the
identity of the atom from which the X-ray was emmited can be established.21

EELS
When an electron beam is incident into specimen,
a part of the electrons is inelastically scattered and
loses a part of its energy. Elemental composition
and the atomic bonding state can be determined by
analyzing the energy with the spectroscope
attached under the electron microscope (Electron
Energy Loss Spectroscopy). Because the
analyzing region can be selected from a part of the
enlarged electron microscopic image, one can
analyze a very small region. Moreover, by selecting
electrons with a specific loss energy by a slit so as
Figure A8 EELS spectra of
carbon allotropes22
to image them, the element distribution in a
specimen can be visualized (Elemental Mapping).

Diamond, graphite and fullerene are the materials that consist of only carbon, so
that, all of these specimens have absorption peaks around 284 eV in EELS
corresponding to the presence of carbon atoms. From the fine structure of the
absorption peak, the difference in bonding state and local electronic state can be
detected (see Figure A8). The sharp peak at the absorption edge corresponds to

140
 APPENDIX A

the excitation of a carbon K-shell electron (1s electron) to an empty anti-bonding π-

orbital. In the case of diamond this sharp peak is absent.22

Raman spectroscopy
This technique is based on the Raman effect, which consists of the inelastic
scattering of light by matter.

Figure A9 Representative Raman spectra of carbon materials. a) First- and second-order Raman
23
spectra of highly oriented pyrolytic graphite (HOPG) and “glassy” carbon, extracted from Ref. , b)
comparison of Raman spectra at 514 nm for bulk graphite and graphene. They are scaled to have
similar height of the G’ peak at ~2700 cm-1, extracted from Ref. 24
, c) Raman spectra for purified
25
SWNTs excited at five different laser frequencies, extracted from Ref. .

141
 APPENDIX A

Raman analysis has proven to be a powerful, non-destructive technique, very

helpful in the study of carbons. Its sensitivity to detect disorder, graphitization and
stacking order of graphitic samples (see Figure A9a), plus the characteristic
fingerprints of single-wall carbon nanotubes and graphene (see Figure A9 b and c),
make of this technique a basic tool in the analysis of carbon materials.
Due to the importance of this characterization tool to analyze carbon structures, we
have devoted a complete appendix to address the principles behind Raman
scattering, focusing on carbon materials (see Appendix B).

142
 APPENDIX A

References

1
Translated from:

http://bacterio.uc3m.es/docencia/profesores/herreros/itts/ficheros/Arco.pdf
2
http://mrsec.wisc.edu/Edetc/SlideShow/slides/nanotubes/growing_arc.html
3
R. Saito, G. Dresselhaus, M.S. Dresselhaus. Physical Properties of Carbon
Nanotubes. Chapter 5: Synthesis of Carbon Nanotubes. Imperial College
Press. London, 1998
4
S. Iijima, T. Ichihashi. Single-shell carbon nanotubes of 1-nm diameter. Nature
363, 603-605, 1993
5
D.S. Bethune, C.H. Klang, M.S. de Vries, G. Gorman, R. Savoy, J. Vazquez, R.
Beyers. Cobalt-catalysed growth of carbon nanotubes with single-atomic-layer
walls. Nature 363, 605-607, 1993
6
A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y.H. Lee, S.G.
Kim, A.G. Rinsler, D.T. Colbert, G.E. Scuseria, D. Tománek, J.E. Fischer, R.E.
Smalley. Crystalline ropes of metallic carbon nanotubes. Science 273, 483-487,
1996
7
http://nano.mtu.edu/PulsedLaserVaporization_start.html
8
http://en.wikipedia.org/wiki/Chemical_vapor_deposition
9
H. Dai, A. G. Rinzler, P. Nikolaev, A. Thess, D. T. Colbert, R.E. Smalley. Single-
wall nanotubes produced by metal-catalyzed disproportionation of carbon
monoxide. Chem. Phys. Lett. 260, 471-475, 1996
10
J. Kong, A. M. Cassel, H. Dai. Chemical vapor deposition of methane for single-
walled carbon nanotubes. Chem. Phys. Lett. 292, 567-574, 1998
11
J. H. Hafner, M. J. Bronikowski, B.R. Azamian, P. Nikolaev, A. G. Rinzler, D. T.
Colbert, K.A. Smith, R.E. Smalley. Catalytic growth of single-wall carbon
nanotubes from metal particles. Chem. Phys. Lett. 296, 195-202, 1998

143
 APPENDIX A

12
S. Maruyama, R. Kojima, Y. Miyauchi, S. Chiashi, M. Kohno. Low-temperature
synthesis of high-purity single-walled carbon nanotubes from alcohol. Chem.
Phys. Lett. 360, 229-34, 2002
13
H.W. Zhu, C.L. Xu, D.H. Wu, B.Q. Wei, R. Vajtai, P.M. Ajayan. Direct synthesis
of long single-walled carbon nanotube strands. Science 296, 884-886, 2002
14
K. Hata, D. N. Futaba, K. Mizuno, T. Namai, M. Yumura, S. Iijima. Water-
assisted highly efficient synthesis of impurity-free single-walled carbon
nanotubes. Science 306, 1362-1364, 2004
15
http://en.wikipedia.org/wiki/Electron_microscope
16
http://www.unl.edu/CMRAcfem/em.htm
17
http://en.wikipedia.org/wiki/X-ray_crystallography
18
Y. Saito, T. Yoshikawa, S. Bandow, M. Tomita, T. Hayashi. Interlayer spacings in
carbon nanotubes. Phys. Rev. B 48, 3, 1907-1909, 1993
19
http://www.nuance.northwestern.edu/KeckII/xps1.asp
20
T.I.T. Okpalugo, P. Papakonstantinou, H. Murphy, J. Mc Laughlin, N.M.D. Brown. High
resolution XPS characterization of chemical functionalized MWCNTs and SWCNTs.
Carbon 43, 153-161, 2005
21
http://www.siliconfereast.com/edxwdx.htm
22
http://eels.kuicr.kyoto-u.ac.jp/eels.en.html
23
R. J. Nemanich, S.A. Solin. First- and second-order Raman scattering from finite-size
crystals of graphite. Phys. Rev. B 20, 2, 1979
24
A. C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec,
D. Jiang, K.S. Novoselov, S. Roth, A.K. Geim. Raman spectrum of graphene and
graphene layers. Phys. Rev. Lett. 97, 187401, 2006
25
A. M. Rao, E. Richter, S. Bandow, B. Chase, P.C. Eklund, K.A. Williams, S. Fang, K.R.
Subbaswamy, M. Menon, A. Thess, R.E. Smalley, G. Dresselhaus, M.S. Dresselhaus.
Diameter-selective Raman scattering from vibrational modes in carbon nanotubes.
Science 275, 187-191, 1997

144
 APPENDIX B

APPENDIX B
Raman spectroscopy of sp2 carbon materials

Raman scattering refers to the inelastic scattering of light. During a scattering

event (1) an electron is excited from the valence energy band to the conduction
energy band by absorbing a photon, (2) the excited electron is scattered by
emitting (or absorbing) phonons, and (3) the electron relaxes to the valence band
by emitting a photon. We generally observe Raman spectra for the scattered
photon (light) whose energy is smaller by the phonon energy than that of the
incident photon in accordance with the so-called Stokes process, discussed below.
By measuring the intensity of the scattered light as a function of frequency
downshift (losing energy) of the scattered light, which is what is plotted in Raman
spectra, we obtain an accurate measure of the phonon frequencies of the material.
Raman scattering can occur for phonon emission or by phonon absorption, and
these two processes are called the Stokes process and anti-Stokes process,
respectively1.
The number of emitted phonons before relaxation of the lattice can be one, two,
and so on, which we call, respectively, one-phonon, two-phonon and multi-phonon
Raman processes. The order of a scattering event is defined as its number in the
sequence of the total scattering events, including elastic scattering by an
imperfection (such as a defect or edge) of the crystal. The lowest order process is
the first-order Raman scattering process which gives Raman spectra involving one-
phonon emission.1
To run Raman spectroscopy experiments, lasers are used as a source of light, the
excitation energy can be tuned by changing the laser wavelength in the visible
range.
In order to interpret the Raman spectra of carbon nanostructures it is necessary to
understand the phonon dispersion of graphene or 2D graphite.
The phonon dispersion relation shows interesting features in crystals with two or
more atoms per primitive basis. For each polarization mode in a given propagation
direction the dispersion relation ω versus Κ develops two kinds of branches, known

145
 APPENDIX B

as the acoustical and optical branches. We have longitudinal LA and transverse

acoustical TA modes, and longitudinal LO and transverse optical TO modes.
If there are p atoms in the primitive cell, there are 3p branches to the dispersion
relation: 3 acoustical branches and 3p - 3 optical branches.2
Since the unit cell of monolayer graphene contains two carbon atoms, A and B,
there are six phonon dispersion bands, in which three are acoustic branches (A)
and the other three are optic (O) phonon branches.3 For one acoustic branch (A)
and one optic (O) phonon branch, the atomic vibrations are perpendicular to the
graphene plane, and they correspond to the out-of-plane (o) phonon modes. For
two acoustic and two optic phonon branches, the vibrations are in-plane (i).
Traditionally, the directions of the vibrations are considered with respect to the
direction of the nearest carbon-carbon atoms and, therefore, the phonon modes
are classified as longitudinal (L) or transverse (T) according to vibrations parallel
with or perpendicular to, respectively, the A-B carbon-carbon directions. Therefore,
along the high symmetry ΓΜ and ΓΚ directions, the six phonon dispersion curves
are assigned to LO, iTO, oTO, LA, iTA and oTA phonon modes, (see Figure B1).
Near the zone center (Γ point), the in-plane iTO and LO optic modes correspond to
the vibrations of the sublattice A against the sublattice B, and these modes are
degenerate at the Γ point. The phonon modes around the K point are especially
important, since the D-band and the G’-band are related to phonon modes in the
vicinity of the K point.

Figure B1 Calculated phonon dispersion relation of graphene showing the iLO, iTO, oTO, iLA, iTA
3
and oTA phonon brances.

146
 APPENDIX B

To correctly describe the dispersion of the LO and iTO phonon branches near the Γ
and K points, it is important to consider the renormalization of the phonon energies,
associated with a process in which a phonon can create an electron-hole pair. This
important electron-phonon coupling cannot be understood within the framework of
the Born-Oppenheimer approximations, and gives rise to an interesting effect
known as the Kohn anomaly.3 The Kohn anomaly is specially important for metallic
SWNTs and for graphene and is responsible for a softening of certain Γ and K
points phonons which increases the dispersive behavior of the D and G’ bands,
which will be explained in detail in future pages.

Resonance Raman spectroscopy

For a photo-excited electron, it will experience some processes before combining
with a hole in the valence energy band. One of these processes is a scattering
process by a quantized lattice vibration, known as a phonon with a phonon wave
vector q. The electronic wave vector k is scattered to k – q (phonon emission) or by
k + q (absorption), respectively, and the electron energy is shifted downward or
upward by hω q , which is given by the observation of the scattered light and is

known as the Stokes or anti-Stokes Raman shift. If a one-phonon scattering event

occurs it is called a first-order Raman process. If either the initial k or the scattered
k ± q electronic states is a real electronic state, the scattering amplitude is
enhanced, and this is known as the incident or scattered resonance Raman
process, respectively (see Figure B2, a1 and a2).
One important aspect of the Raman process in a crystal is that the photo-excited
electron has to go back to its original k state to recombine with the hole at the k
state. Because of this restriction, only a q ~ 0 phonon wave vector is selected in the
first-order Raman process. Thus, in standard solid state textbooks, we see the
statement that only the phonon energy at the center (Γ point) of the Brillouin zone,
which is called the zone-center phonon mode, is observed by the Raman
spectroscopy for solids.
We also expect second-order Raman processes to occur (Figure B2, c1 and c2),
by emitting (1) two phonons with the same frequency or (2) two phonons with

147
 APPENDIX B

different frequencies, which are called overtone or combination Raman modes,

respectively. Another second-order process is given by one-phonon scattering and
one elastic scattering process (Figure B2, b1-b4). Elastic scattering occurs in the
presence of a defect in the crystal. In the latter case, the one phonon energy is
observed in the second-order process. We refer to it as a one-phonon, second-
order process.

Figure B2 (a1, a2) First-order, (b1-b4) one-phonon second-order and (c1, c2) two-phonon second-
order, resonance Raman spectral processes. Top panels: incident photon resonance; bottom
panels: scattered resonance conditions. For one-phonon, second-order transitions, one of the two
4
scattering events is elastic scattering (dashed line). Resonance points are shown as solid circles.

In a second-order process, since an electron is represented by wave vectors q and

–q, respectively, by scattering from and then by back-scattering to the original k
state, non-zone-center phonon modes (q ≠ 0) can be observed in the second-order
Raman spectra. That is, (1) the electron is excited at k, (2) scattered to k + q, (3)
scattered back to k and (4) recombined with a hole at k (see figure B2, b1-b4 and
c1,c2). However, the probability for an electron to be scattered by a phonon is
generally small, and thus the probability for the event occurring twice in a second-
order process is much smaller. Thus, higher-order Raman processes in solids have
not been considered much except for double resonance Raman spectroscopy.4
The double resonance effects occur when both the scattered intermediate state
and the initial (or final) state are actual electronic states,5 and in this situation the
Raman intensity can be increased by orders of magnitude.

148
 APPENDIX B

In a double resonance process, the resonance enhancement results in Raman

intensities that are almost of the same order as in the first-order resonance Raman
spectra, and the corresponding q is selected by the resonance condition.4

The most outstanding features in the Raman spectra of carbon nanotubes are: the
radial breathing mode (RBM) appearing in the range 100 cm-1 – 400 cm-1, that is a
totally symmetric vibrational mode associated with the vibration of carbon atoms in
a radial direction in relation to the nanotube axis; the G band appearing at ~1580
cm-1 is associated with the highest-frequency optical phonon modes at the Γ-point
in the Brillouin zone;4 and the G’ band that appears in the high frequency region
(~2700 cm-1), this band originates from a second-order process involving two iTO
phonons near the K point3. In addition to these bands we see the defect-induced
band (D band) at 1350 cm-1 when a certain amount of disorder or symmetry-
breaking defects are present in the nanotube or in graphitic structures especially
for spectra taken at the graphene edges. In some graphitic structures another
defect-induced band appears at 1620 cm-1. These bands originate from an
intervalley and intravalley double resonance process, for the D and D’ bands,
respectively (see Figure B4).

First-order Raman scattering

In sp2-carbon materials the only first-order Raman feature that we observe is the G
band; however, in the Raman spectra of single-wall carbon nanotubes we observe
another first-order Raman mode at low frequencies, the so-called radial breathing
mode. This feature is inherent to SWNTs and its observation in a Raman spectrum
provides irrefutable evidence of the presence of SWNTs in the sample.

Radial Breathing Mode (RBM)

The RBM Raman features (appearing between 120 cm-1 <ωRBM < 250 cm-1 for
SWNTs within 1 nm< dt < 2 nn) correspond to the atomic vibration of the C atoms in
the radial direction, as if the tube were breathing (see Figure B3 a and c). These
features are very useful for characterizing nanotube diameters through the relation

149
 APPENDIX B

ωRBM= 234 cm-1/dt + 10 cm-1 for samples in bundle. For dt < 1 nm, this relation is not
expected to hold due to nanotube lattice distortions leading to a chirality
dependence of ωRBM. For large diameter tubes (dt > 2 nm) the intensity of the RBM
feature is weak and is hardly observable6.
Because of its dependence on the nanotube diameter, the spectra of this mode are
largely used for the characterization of the diameter distribution in a carbon-
nanotube sample. The Raman intensity for the RBM is strongly enhanced when the
incident or scattered light is in resonance with an excitonic transition. Thus, the
Raman spectra of the RBMs for a sample composed of an ensemble of different
nanotube chiralities is strongly dependent on Elaser because for each value of Elaser,
the optical transition energies (Eii) for different SWNTs are in resonance, and thus,
the intensity of the RBM for these nanotubes is resonantly enhanced. 7

Figure B3 Schematic of the nanotube vibrations, a) RBM mode, and b) the G band atomic
vibrations along the nanotube circumference and along the nanotube axis (extracted from reference
8
6) . c) Black line, Raman spectrum of a bundle of SWNTs synthesized by CVD with Elaser = 2.41
eV, the labels name the first order, second order and combination Raman features; the light-blue
line depicts the spectrum of HOPG for comparison.

At a fixed Elaser value, the RBM linewidth is found to increase with increasing
nanotube diameter and as the energy difference |Elaser - Eii| between Elaser and the
resonant interband energy Eii increases. The smallest linewidths are, as expected,
obtained under fully resonant conditions where Elaser = Eii.9 If one interband energy
Eii and a nanotube diameter are specified, then its corresponding unique (n,m)
indices can be identified from the Kataura plot (Figure 1.5).
150
 APPENDIX B

It is clear that a single Raman measurement gives an idea of the tubes that are in
resonance with that laser line, but does not give a complete characterization of the
diameter distribution of the sample. However, by taking Raman spectra using many
laser lines, a good characterization of the diameter distribution in the sample can
be obtained.6

Tangential modes - G band

The graphite-like band (G-band) in carbon nanotubes is directly related to the G-
band in graphite that is identified with an in-plane tangential optical phonon
involving the stretching of the bond between the two atoms in the graphene unit
cell, see Figure B3b. While the G-band in graphite exhibits a single Lorentzian
feature at ~1582 cm-1, the G-band of carbon nanotubes is composed of several
peaks originating from the quantum confinement of the wavevector along the
SWNT circumferential direction, and the folding of the graphite Brillouin zone into
the SWNT Brillouin zone.7 The observation of these characteristic multi-peak
features around ~1580 cm-1 also provides a signature of carbon nanotubes. This
multi-peak feature gives information about the metallic (or semiconducting)
character of the SWNTs in resonance with a given laser line. The two most
intense G peaks are labeled G+, for atomic displacements along the tube axis, and
G-, for modes with atomic displacement along the circumferential direction (see
Figure B3 b and c), and the lowering of the frequency for the G- mode is caused by
the curvature of the nanotube which softens the tangential vibration in the
circumferential direction.
The difference between the G band lineshape for semiconducting and metallic
SWNTs is evident in the lineshape of the G- feature, which is broadened for
metallic SWNTs in comparison with the Lorentzian lineshape for semiconducting
tubes, and this broadening is related to the presence of free electrons in nanotubes
with metallic character.10
However, when multi-wall carbon nanotubes, HOPG, graphene or any other sp2
carbon system is being analyzed, the G band shape is consistent with a single
sharp Lorentzian peak, centered around 1580 cm-1. 11,12,13

151
 APPENDIX B

Second-order Raman scattering

In second-order double-resonance (DR) Raman processes for carbon materials,
the electron (1) absorbs a photon at a k state, (2) scatters to k + q states, (3)
scatters back to a k state, and (4) emits a photon by recombining with a hole at a k
state. The two scattering processes consist of either elastic scattering by defects of
the crystal or inelastic scattering by emitting a phonon, as shown in Figure B2.
Thus (1) one-elastic and one-inelastic scattering event and (2) two-inelastic
scattering events are relevant to second-order Raman spectroscopy.1
The most important second-order Raman features are the D band and the G’ band,
they originate from a second-order process involving two iTO phonons near the K
point for the G’ band and one iTO phonon and one defect in the case of the D
band. Since the G’ band is approximately twice the D band frequency (ωG’ ~ 2ωD),
some authors prefer to call it the 2D band. However this two-phonon band is
allowed in the second-order Raman spectra of sp2-materials without any kind of
disorder or defects (see spectrum of HOPG in Figure B3c). In order to prevent any
misleading connection of this feature with disorder or defects, and to avoid
confusion between the designation of “2D” to denote two dimensionality, we will
use here the conventional notation of “G’ band” as is used in the graphite and
nanotube literature.3
In a DR Raman process, two resonance conditions for three intermediate states
should be satisfied, in which the intermediate k + q state is always a real electronic
state and either the initial or the final k state is a real electronic state.1 This DR
mechanism is called an intervalley process because it connects points in circles
around inequivalent K and K’ points in the first Brillouin zone of graphene. On the
other hand, the double resonance process responsible for the D’ band is an
intravalley processs, since it connects two points belonging to the same circle
around the K point (or the K’ point), see Figure B4.3
Through history, the D and the G’ bands have been very useful in the
characterization of graphitic materials. For years the origin of the D band and its
dispersion was not fully understood. In 1970 Tuinstra and Koening11 attributed the

152
 APPENDIX B

presence of the D band in the Raman spectrum of graphite to particle size effects
in which due to finite crystal size an A1g mode of the lattice became active, they
also correlated the intensity of this mode to the in-plane crystallite size La. The
following experiments showed that the presence of the D band in the Raman
spectra of sp2-carbon materials was also related to disorder features present in the
material. In present days the ratio of the integrated areas (or the intensities) of the
D band over the G band (ID/IG ratio) is a direct quantification of the degree of
disorder or defects present in the material and is helpful in the determination in-
plane crystallite sizes.14 Edges and crystal boundaries within a sample also
contribute to the D band intensity.
In the Raman spectra of carbon materials, the presence of the G’ band and most
importantly of its doublet morphology represent a strong proof of 3D order or ABAB
stacking achievement in a graphitic structure (see Figure 3.18b).
Recently the G’ band has been used as a characterization tool of single-layer and
few-layer graphene to determine the number of sheets, with ABAB… stacking,
present in the sample.13

Disorder induced modes

In contrast to the first-order process, where only phonons near the Γ point of the
Brillouin zone are probed, in second-order Raman spectra it is possible to probe
features identified with phonons associated with interior points in the Brillouin zone,
see Figure B4. This becomes possible because the selection rules for the second-
order features are relaxed with respect to the first-order features.

Figure B4 Top panel: one-phonon second-

order double resonance process for the D
band (intervalley process), and bottom panel:
one-phonon second-order double resonance
process for the D’ band (intravalley process).
(Adapted from reference 3)

153
 APPENDIX B

The most common example of a second-order feature in the Raman spectra of

carbon nanotubes is the disorder-induced band, observed in the ~1300-1400 cm-1
range (D band). The D band common to all sp2-hybridizaed disordered carbon
materials, originates from phonons close to the K point of the graphite Brillouin
zone, and is activated by the presence of defects, such as impurities or missing
atoms, finite-size effects and in the case of carbon nanotubes, molecules linked to
the nanotube sidewalls.7 Around
1620 cm-1 another disorder-
induced feature appears and is
called D’ band and it originates
from an intravalley second-order
double-resonance process.
In carbon materials, there are
several weak Raman signals
whose phonon frequencies change
with changing laser excitation
energy which is called “dispersive”
behavior.1

Figure B5 Electronic energy bands of 2D

graphite (top). Phonon dispersion curves
of 2D graphite (bottom). Both the phonon
branch that is strongly coupled to the
electronic bands in the optical excitation,
and the electronic bands near the Fermi
level (E = 0) that are linear in k are
indicated by heavy lines. (Extracted from
15
reference )

A well-known dispersive spectral feature is the disorder-induced D band which

appears at around 1350 cm-1 for laser excitation at about 2.41 eV. The D band
frequency is highly dispersive and increases with increasing laser energy (Elaser) by

154
 APPENDIX B

44-51 cm-1/eV for various graphite and sp2 materials and by 53 cm-1/eV for
SWNTs, respectively.5
Double resonance Raman theory, as discussed in this section, works well for
explaining the dispersive phonon modes in which a non-zone-center (q ≠ 0) phonon
mode and a second-order Raman process are relevant to these weak spectral
features.
As shown in Figure B5, the electronic structure of 2D graphite near the Fermi
energy is linear in wave vector k, which is expressed by the crossed solid lines in
Figure B2. The crossing point corresponds to the Fermi energy located at the K
point of the Brillouin zone, this crossing point in graphene is called the Dirac point.
When the laser energy Elaser increases, the resonance k vector for the electron
moves away from the K point. In the DR process, the corresponding q vector for
the phonon increases with increasing k, measured from the K point. Thus by
changing the laser energy, we can observe the phonon energy hω (q ) along the
phonon dispersion relations (Figure B5). This effect is observed experimentally as
a dispersion of the phonon energy as a function of excitation laser energy.1

G’ band
The most intense feature in the second-order Raman spectra of SWNTs (and of
sp2-carbon materials) is the overtone of the D band feature discussed above, the
so-called G’ band, whose frequency is 2ωD (See Figure B3c) and its frequency
dispersion is twice the D band frequency dispersion. While the D band originates
from a double-resonance process involving a phonon and a defect, in the G’ band,
instead of a defect, another phonon is responsible for the momentum conservation
in the double-resonance process, and the G’ band is symmetry-allowed.
In systematic studies of heat treatments of carbonaceous materials16,17 it has been
shown that in the graphitization process, when the ABAB stacking of graphitic
layers is achieved, the G’ peak evolves to a double-peak morphology. The
explanation underlying the splitting observed on Figure 3.18b can be extracted
from the article by Al-Jishi and Dresselhaus.18 These authors calculated the
phonon density of states for hexagonal graphite. They found a number of peaks

155
 APPENDIX B

relevant to this discussion, in particular, a strong one at 1345 cm-1 due to the
extremum in the dispersion curves along the ΓΚ direction and a second line at
1365 cm-1 due to Μ point zone edge phonons. Thus, when the graphite is
structurally well stacked (ordered), overtones could be expected in the second-
order Raman spectrum at twice these frequencies. On the other hand, for the most
poorly organized materials, one would expect these features to be less well defined
and only an overtone corresponding to the average is observed experimentally.
The appearance of the doublet in the ordered graphites is simply a manifestation of
the sharpening of the features in the phonon DOS curves upon establishing good
ABAB stacking.18

In 2008 Maciel and co-workers19 proved that the frequency of the G’ band in the
Raman scattering of SWNTs is dependent of the diameter distribution in the
sample and that the observation of a second peak at lower frequencies is indeed
not related to the diameter distribution of the samples, but there is actually a
second feature at lower frequencies due to the presence of defects, which
disappears under heat treatments. The observation of a two peaks G’ band is due
to the information coming from defect sites (G’Def) and the higher frequency peak is
associated with non-defective tube segments (G’Pris).
The defect-induced nature of the G’Def peak was demonstrated when the G’ band
spectra of pristine CVD-produced SWNTs was analyzed in the same work. The
authors found that in the as-produced SWNTs, a weak G’Def peak was observed,
indicating the presence of defects in the structure (vacancies, structural
heterogeneity). However, the G’Def peak disappears after annealing of the sample
at 200 ºC, 400 ºC and 600 ºC in argon flow (see Figure B6). The defects were
present only before the annealing step. These results are congruent with
previously reported works on the low crystallinity (quality) of CVD-grown SWNTs
when compared to their arc-discharge-produced counter part. Properly heat treated
CVD-grown SWNTs do not have disorder induced features.

156
 APPENDIX B

Figure B6 The G’ features of CVD-SWNT samples before and after heat treatment at 400 ºC and at
600 ºC (Elaser=2.41 eV). The spectra were deconvoluted into four Lorentzian (dashed lines for G’Pris
and dotted lines for G’Def) the corresponding RBM spectra were superimposed to the G’ band
(extracted from Reference 19).

Later in 2008, the same group proved that the G’ band can be used to probe
defects in the lattice of single-wall carbon nanotubes20. These defects can be
vacancies, Stone-Wales transformations or even foreign atoms inserted in the
honeycomb structure (i.e. doping). Figure B7a is extracted from that report, where
we see a plot of the G’ band for different sp2 carbon materials. The three upper
curves are for single-wall carbon nanotubes prepared by different methods. The
upper plot is for heat-treated undoped SWNTs and shows a single G’P peak (P for
pristine) centered at ωG ' = 2,676 cm-1.
P

On doping, a new G’ peak is observed at a lower/higher frequency for n/p doping.

The spectra of graphene, highly oriented pyrolytic graphite (HOPG) and
amorphous carbon are also presented in Figure B7a to clearly show that the new
peak cannot be related to other forms of sp2 carbon. This new peak, from now on is
called the G’Def peak (Def for ‘defect’), and appears in the G’ spectra of doped
SWNTs.
The presence of a negatively charged defect provokes the renormalization effect
on the local electronic and vibration structure, as discussed below.

157
 APPENDIX B

The second-order symmetry-allowed G’ Raman scattering process occurs in all sp2

carbon materials, involving in-plane transversal optical (iTO) phonons near the K
point in the hexagonal Brillouin zone, selected by the double-resonance process.

a b

2
Figure B7 a) The G’ band in different sp carbon materials measured at room temperature with
Elaser = 2.41 eV (514 nm). The arrows point to defect-induced peaks in the G’ for doped SWNTs.
The p/n doping comes from substitutional boron/nitrogen atoms, the nearest neighbours of carbon
in the periodic table. The spectra of graphene, HOPG and amorphous carbon are shown for
comparison. b) Schematic model showing the renormalization of electron and phonon energies near
a negatively charged defect. The upper panel shows the dispersion of the π and π* electrons along
the KMK’ direction in the hexagonal Brillouin zone, near the Fermi level (EF). The lower panel shows
the iTO phonon branch along the ΓK direction. The arrows in the upper panel indicate electron
transitions by photon absorption (vertical arrows) and phonon scattering (nearly horizaontal arrows).
The arrows in the lower panel indicate the energies for the same phonons when they are scattering
electrons in the upper panel. Dashed lines represent the renormalized structure near the defect
sites. (Adapted from Reference 20)

An excited electron with wave vector k resonantly selects a phonon with wave
vector q ~ -2k. This relation between k and q, together with the Dirac-Fermion
behavior for high-speed electrons and the Kohn anomaly for phonons, makes the
G’ band strongly dependent on the particular electronic structure, thereby providing
information about the electronic structure.

158
 APPENDIX B

The excited electron with wave vector of modulus k is generated by resonance with
the excitation laser
Elaser = Ec – Ev = 2 hvF k ≈ hvF q
where the subscripts v and c stand for the valence and conduction bands,
respectively. Changing Elaser changes k linearly, and consequently q, so that the
electronic structure can be probed using phonon measurements.
A schematic model showing the defect-induced renormalization of the electron and
phonon energies and its influence on the double-resonance G’ scattering process
is shown in Figure B7b.
The two resonant processes, the optical absorption and the scattering of the
excited electron by a phonon, are presented by the vertical and nearly horizontal
arrows, respectively. The G’Pris follows the unperturbed process (solid lines),
whereas the G’Def follows the process where electron and phonon energies are
renormalized (dashed lines).20

In the second-order Raman spectra of graphite and SWNTs there are other
features worth mentioning. For example the M and iTOLA features21 located at
1700 cm-1 and 1950 cm-1, respectively (iTOLA is the combination mode of the iTO
+ LA processes); the G* band22 at 2450 cm-1; the D+G band (combination mode of
the phonons that originate the D and G bands) located at 2900 cm-1; the 2G band
(overtone of the G band) at 3180 cm-1, and the 2D’ band12,23 (overtone of the D’
band) found at 3250 cm-1.

159
 APPENDIX B

References

1
M.S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio. Raman spectroscopy of
carbon nanotubes. Physics Reports 409, 47, 2005
2
C. Kittel. Introduction to Solid State Physics. 7th edition. Wiley. USA, 1996
3
L.M. Malard, M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus. Raman
spectroscopy in graphene. Physics Reports 473, 51-87, 2009
4
R. Saito, A. Grüneis, G.G. Samsonidze, V.W. Brar, G. Dresselhaus, M.S.
Dresselhaus, A. Jorio, L.G. Cançado, C. Fantini, M.A. Pimenta, A.G. Souza
Filho. Double resonance Raman spectroscopy of single-wall carbon
nanotubes. New Journal of Physics 5, 157.1-157.15, 2003
5
R. Saito, A. Jorio, A.G. Souza Filho, G. Dresselhaus, M.S. Dresselhaus, M.A.
Pimenta. Probing phonon dispersion relations of graphite by double resonance
Raman scattering. Phys. Rev. Lett. 88, 027401, 2002
6
A. Jorio, M.A. Pimenta, A.G. Souza-Filho, R. Saito, G. Dresselhaus, M.S.
Dresselhaus. Characterizing carbon nanotube samples with resonance Raman
scattering. New Journal of Physics 5,139.1-139.17, 2003
7
A. Jorio, M.S. Dresselhaus, G. Dresselhaus, Editors. Carbon nanotubes.
Advanced topics in the synthesis, structure, properties and applications.
Springer. Berlin, Heidelberg, 2008
8
F. Lupo, J.A. Rodríguez-Manzo, A. Zamudio, A.E. Elías, Y.A. Kim, T. Hayashi,
M. Muramatsu, R. Kamalakaran, H. Terrones, M. Endo, M. Rühle, M. Terrones.
Pyrolytic synthesis of long strands of large diameter single-walled carbon
nanotubes at atmospheric pressure in the absence of sulphur and hydrogen.
Chem. Phys. Lett. 410, 384-390, 2005
9
M.S. Dresselhaus, G. Dresselhaus, A. Jorio, A.G. Souza-Filho, R. Saito. Raman
spectroscopy on isolated single wall carbon nanotubes. Carbon 40, 2043-2061,
2002

160
 APPENDIX B

10
A. Jorio, R. Saito, G. Dresselhaus, M.S. Dresselhaus. Determination of
nanotubes properties by Raman spectroscopy. Phil. Trans. R. Soc. Lond. A
362, 2311-2336, 2004
11
F. Tuinstra, J.L. Koenig. Raman spectrum of graphite. The Journal of Chemical
Physics 53, 3, 1970. R. J. Nemanich, S.A. Solin. First- and second-order
Raman scattering from finite-size crystals of graphite. Phys. Rev. B 20, 2, 1979
12
Y. Kawashima, G. Katagiri. Fundamentals, overtones, and combinations in the
Raman spectrum of graphite. Phys. Rev. B 52, 14, 1995
13
A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S.
Piscanec, D. Jiang, K.S. Novoselov, S. Roth, A.K. Geim. Raman spectrum of
graphene and graphene layers. Phys. Rev. Lett. 97, 187401, 2006
14
L.G. Cançado, K. Takai, T. Enoki, M. Endo, Y.A. Kim, H. Mizusaki, A. Jorio,
L.N. Coelho, R. Magalhães-Paniago, M.A. Pimenta. General equation for the
determination of the crystallite size La of nanographite by Raman spectroscopy.
Appl. Phys. Lett. 88, 163106, 2006
15
M.J. Matthews, M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus, M. Endo.
Origin of dispersive effects of the Raman D band in carbon materials. Phys.
Rev. B 59, 10, R6585-R6588, 1999
16
P. Lespade, A. Marchand, M. Couzi, F. Cruege. Caracterisation de matériaux
carbonés par microspectrométrie Raman. Carbon 22, 4/5, 375-385, 1984
17
H. Wilhelm, M. Lelaurain, E. McRae, B. Humbert. Raman spectroscopic studies
on well-defined carbonaceous materials of strong two-dimensional character.
Journal of Applied Physics 84, 12, 6552-6558, 1998
18
R. Al-Jishi, G. Dresselhaus. Lattice-dynamical model for graphite. Phys. Rev. B
26, 8, 4514-4522, 1982
19
I.O. Maciel, M.A. Pimenta, M. Terrones, H. Terrones, J. Campos-Delgado, A.
Jorio. The two peaks G’ band in carbon nanotubes. Phys. Stat. Sol. (b) 245,
10, 2197-2200, 2008

161
 APPENDIX B

20
I.O. Maciel, N. Anderson, M.A. Pimenta, A. Hartschuh, H. Quian, M. Terrones,
H. Terrones, J. Campos-Delgado, A.M. Rao, L. Novotny, A. Jorio. Electron and
phonon renormalization near charged defects in carbon nanotubes. Nature
Materials 7, 878-883, 2008
21
V.W. Brar, G.G. Samsonidze, M.S. Dresselhaus, G. Dresselhaus, R. Saito, A.K.
Swan, M.S. Ünlü, B.B. Goldberg, A.G. Souza-Filho, A. Jorio. Second-order
harmonic and combination modes in graphite, single-wall carbón nanotube
bundles, and isolated single-wall carbon nanotubes. Phys. Rev. B 66, 155418,
2002
22
T. Shimada. T. Sugai, C. Fantini, M. Souza, L.G. Cançado, A. Jorio, M.A.
Pimenta, R. Saito, A. Grüneis, G. Dresselhaus, M.S. Dresselhaus, Y. Ohno, T.
Mizutani, H. Shinohara. Origin of the 2450 cm-1 Raman bands in HOPG,
single-wall and double-wall carbon nanotubes. Carbon 43, 1049-1054, 2005
23
S.D.M. Brown, P.Corio, A. Marucci, M.A. Pimenta, M.S. Dresselhaus. Second-
order resonant Raman spectra of single-walled carbon nanotubes. Phys. Rev.
B 61, 11, 7734, 2000

162