Chemical modification and opticalspectroscopy of single-walled carbon

nanotubes

Ph.D. Dissertation

Katalin Nemeth

Doctoral School of Molecular- and NanotechnologiesUniversity of Pannonia

Head of School: Dr. Ferenc Vonderviszt

Supervisor: Prof. Katalin KamarasResearch Professor, Corresponding Member of the Hungarian Academy

of Sciences

Institute for Solid State Physics and Optics,Wigner Research Centre for Physics, Hungarian Academy of Sciences

2015

Chemical modification and optical spectroscopy of single-walled carbonnanotubes

Ertekezes doktori (PhD) fokozat elnyerese erdekeben

Irta:Nemeth Katalin

Keszult a Pannon Egyetem Molekularis- es NanotechnologiakDoktori Iskolajaban

Temavezeto: Prof. Kamaras Katalin

Az ertekezest temavezetokent elfogadasra javaslom:

Prof. Kamaras Katalin: igen/nem ...............................(alaıras)

A jelolt a doktori szigorlaton .......%-ot ert el.

Az ertekezest bıralokent elfogadasra javaslom:

Bıralo neve:...................................................... igen/nem

..................................(alaıras)

Bıralo neve:...................................................... igen/nem

..................................(alaıras)

A jelolt az ertekezes nyilvanos vitajan ........%-ot ert el.

Veszprem, 201..............................

A doktori (PhD) oklevel minosıtese:..........................

........................................a Bıralo Bizottsag elnoke

.........................................az EDHT elnoke

3

Contents

Kivonat 6

Abstract 8

Auszug 9

Foreword 10

1 Introduction 11

1.1 Carbon nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1.1.1 Allotropic forms of elemental carbon . . . . . . . . . . . . . . . 11

1.1.2 Basics of carbon nanotubes . . . . . . . . . . . . . . . . . . . . 12

1.1.3 Growth of carbon nanotubes . . . . . . . . . . . . . . . . . . . . 16

1.1.4 Electronic structure of single-walled carbon nanotubes . . . . . 18

1.1.5 Optical transitions of single-walled carbon nanotubes . . . . . . 19

1.1.6 Vibrational properties and Raman spectrum of single-walled car-

bon nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1.2 Chemistry of single-walled carbon nanotubes . . . . . . . . . . . . . . . 24

1.2.1 Structure and reactivity . . . . . . . . . . . . . . . . . . . . . . 24

1.2.2 Challenges in nanotube chemistry . . . . . . . . . . . . . . . . . 26

1.2.3 Reaction types and sites in nanotubes . . . . . . . . . . . . . . 26

1.2.4 Classical and modified Birch reduction . . . . . . . . . . . . . . 29

1.2.5 Alkali metal intercalation . . . . . . . . . . . . . . . . . . . . . 30

1.3 Optical spectroscopy of carbon nanotube thin films . . . . . . . . . . . 33

1.3.1 The Drude-Lorentz model . . . . . . . . . . . . . . . . . . . . . 33

1.3.2 Calculation of optical functions . . . . . . . . . . . . . . . . . . 35

2 Experimental 38

2.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.2 Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

2.2.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

2.2.2 Synthetic routes . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4

2.3 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

2.3.1 Raman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . 43

2.3.2 Thermogravimetry-mass spectrometry . . . . . . . . . . . . . . 44

2.3.3 1H-NMR spectrometry . . . . . . . . . . . . . . . . . . . . . . . 44

2.3.4 Wide range transmission spectroscopy . . . . . . . . . . . . . . 45

3 Results and discussion 52

3.1 Hydrogenation reactions on HiPco single-walled carbon nanotubes . . . 52

3.1.1 Raman spectroscopic and TG-MS results . . . . . . . . . . . . . 54

3.1.2 Optical spectroscopic results . . . . . . . . . . . . . . . . . . . . 56

3.2 Hydrogenation and n-butylation of HiPco single-walled carbon nanotubes 61

3.3 Hydrogenation of nanotube bundles by alkali metal intercalation . . . . 62

3.3.1 van der Waals interactions in nanotube bundles . . . . . . . . . 72

3.4 Conclusions and summary . . . . . . . . . . . . . . . . . . . . . . . . . 78

Acknowledgement 81

Theses 83

Tezispontok 85

List of publications 87

References 90

5

Kivonat

Egyfalu szen nanocsovek kemiai modosıtasa es optikai spektroszkopias vizsgalata

A disszertacio temaja egyfalu szen nanocsovek reduktıv addıcios reakcioinak vizs-

galata elsosorban termogravimetria-tomegspektrometria (TG-MS) es szeles tartomanyu

transzmisszios spektroszkopia segıtsegevel. Optikai spektroszkopiaval az addıcios reak-

ciok soran az elektronszerkezetben torteno valtozasok jol detektalhatok, melyekbol a

reakciok atmero-szelektivitasara kovetkeztethetunk.

Kereskedelmi forgalomban kaphato, nagy tisztasagu egyfalu szen nanocsoveket mo-

dosıtottam hidrogen- es n-butil-csoportokkal modosıtott Birch redukcioval es alkalifem-

interkalacioval. A modosıtott Birch redukcio egy kvazi-homogen, oldatfazisu reakcio-

nak tekintheto, melynek elsodlesges kiindulasi anyagai az egyedi nanocsovek, mıg az

interkalacios mechanizmus egy szilard fazisu, a nanocso kotegek belsejeben vegbemeno

reakcio, elsodleges kiindulasi anyagai tehat a nanocso kotegek.

HiPco nanocsovek hidrogenezese mindket reakcio eseten 2-4 H/100 C H-tartalmu

termeket eredmenyezett a TG-MS vizsgalatok szerint. Azonban, a ket reakcio atmero-

szelektivitasa elteronek bizonyult a spektroszkopiai meresek szerint: a modosıtott Birch

redukcio eseten normal (kisebb atmeroju csovek nagyobb reaktivitasa), az interkalacio

eseten fordıtott atmero-szelektivitast tapasztaltam. A kulonbseget a reakciok eltero

mechanizmusaval magyaraztam: az elso esetben a kis atmeroju csovek nagyobb reak-

tivitasaval, a masodik esetben az interkalacio energetikajaval. Az alkalifemek interkala-

cioja a nanocso kotegek belsejebe a koteg tagulasat igenyli. Ennek energiaigenye a szuk-

seges kotegtagulastol fugg, amely aranyos a kation meretevel es fugg a koteget felepıto

csovek atmerojetol.

6

Harom kulonbozo atlagos atmeroju kiindulasi nanocsovet (P2, HiPco, CoMoCat)

es ket eltero ionsugaru alkalifemet (K, Rb) alkalmazva, a nanocso kotegek

hidrogenezesenek atmero-szelektivitasa szeles atmero-tartomanyban es harom kulon-

bozo ionsugar/nanocso-atmero aranynal vizsgalhato. Fentieknek megfeleloen harom

kulonbozo atmero-szelektivitast mutattam ki optikai spektroszkopia segıtsegevel: K/P2

eseten semmilyen, K/HiPco eseten fordıtott, Rb/CoMoCat eseten normal atmero-

szelektivitast. A TG-MS meresek szerint a hidrogen-tartalom a K/P2 es K/HiPco

mintak eseten 2-4 H/100 C, a Rb/CoMoCat mintak eseten meglepoen kicsi,<1 H/100 C

volt.

Vizsgaltam HiPco nanocsovek reaktivitasat n-Bu csoportokkal szemben is, mindket

reakcioval. Mellekreakciokent mindket esetben hidrogenezest vartam. A varakozassal

ellentetben azonban a hidrogenezett nanocso volt a fo termek, a butilozas csak igen kis

szazalekban jatszodott le. Ebbol arra kovetkeztettem, hogy a hidrogenezes a butilozas-

sal szemben kedvezmenyezettebb es/vagy gyorsabb.

Minden esetben vizsgaltam azt is, hogy novelheto-e a kitermeles tobb (harom)

egymast koveto lepes alkalmazasaval, amikor egy reakcio kiindulasi anyaga az elozo

azonos reakciotıpus termeke. Azt talaltam, hogy kevesse bar, de minden esetben novel-

heto volt a kitermeles, meg a modosıtott Birch redukcio eseten is, ami egy kvazi-

homogen, oldatfazisu reakcio, es elsodleseges kiindulasi anyagai az oldoszerben jelenlevo

negatıvan toltott nanocsovek. A kotegek felbomlasa mar ekkor megkezdodik, az oldat-

ban egyedi nanocsovek is jelen vannak. A kitermeles novekedese azt igazolja, hogy

a megmarado kis kotegek az egymast koveto reakciok soran fokozatosan lazulnak es

bomlanak fel.

7

Abstract

Chemical modification and optical spectroscopy of single-walled carbon nanotubes

In this thesis, the reactivity and the diameter selectivity of reductive addition re-

actions of individual and bundled single-walled carbon nanotubes were investigated.

Two different reactions were applied: the well-known modified Birch reduction and a

solid-phase reaction, similar to those of used for the synthesis of some graphite in-

tercalation compounds. These routes were used to functionalize the nanotubes with

hydrogen and n-butyl groups. Thermogravimetry-mass spectrometry was used to de-

termine the functional group content and the thermal stability of the samples, and wide

range transmission spectroscopy on self-supporting thin films to determine the diam-

eter selectivity of the reactions inside one sample, since the energy of the electronic

transitions of the nanotubes has diameter-dependence.

8

Auszug

Chemische Modifikation und optische Spektroskopie der einwandigen Kohlenstoff-

nanorohren

Im Rahmen dieser Doktorarbeit wurden die Reaktionsfahigkeit und die Selektivitat

der Reaktionen an individuellen einwandigen Kohlenstoffnanorohren und Nanorohrbun-

deln untersucht. Zwei verschiedene reduzierende Additionsreaktionen sind hierbei

angewandt worden: die wohlbekannte modifizierte Birch Reduktion und eine Fest-

phasenreaktion, welche ahnlich zu denjenigen bei den Grafitinterkalationverbunden ge-

brauchlich ist. Diese Reaktionen wurden angewandt, um die Nanorohren mit Wasser-

stoff und n-Butyl Gruppen zu modifizieren. Termogravimetrie-Massenspektrometrie

wurde fur die Bestimmung des Funktionsgruppeninhalts und der termischen Festigkeit

der Proben benutzt. Breitbandige optische Spektroskopie wurde an selbsttragenden

Dunnschichten angewandt, um anhand von Anderungen in den Spektren Folgerungen

auf die Durchmesserselektivitat zu schließen.

9

Foreword

Carbon, this significantly important element has very unique properties arising from

the 1s22s22p2 electronic structure of the carbon atom. This electronic structure permits

wide variety of chemical bonding between two, or even more carbon atoms, building

up long chains, rings, networks, even structures with multiple carbon-carbon bonds.

Organic chemistry that deals with these structures is very vivid and complex, but

accordingly very interesting and challenging.

The same challenge exists when we consider the inorganic chemistry of carbon.

Even elemental carbon shows a huge diversity because of the hybridization and bonding

possibilities.

In this thesis I focus on one allotropic form of carbon, namely nanotubes. Covalent

hydrogenation and butylation reactions were done in the frame of this work to inves-

tigate selectivity and steric effects of these addition reactions. Optical spectroscopy

and thermogravimetry-mass spectrometry were the key tools to investigate the special

behaviour of the synthesized material systems.

In the Introduction part I present the basic physical and chemical properties of

nanotubes. A detailed description of the special optical spectroscopic technique and

data evaluation used for carbon nanotube samples is given. Then I present the syntheses

and the results obtained on the products, with special attention to optical spectroscopic

results.

10

1 | Introduction

1.1 Carbon nanotubes

Carbon nanotubes constitute a very special allotropic form of elemental carbon.

Although they can form individual particles in a system (giant molecules set up by

periodic units, like a polymer molecule), a carbon nanotube system is preferably de-

scribed as a solid. With the solid-state description most of the physical properties of

a carbon nanotube can be described and predicted. However, regarding the chemical

properties and reactivity, the solid-state picture is not always sufficient.

In this section I present the special structural, electronic and optical properties of

carbon nanotubes that are relevant and necessary for the topic of this thesis.

1.1.1 Allotropic forms of elemental carbon

A carbon atom (6C: 1s22s22p2) possesses four valence electrons. This means that

it can form four covalent bonds. Because of the collusion of spatial and energetic cir-

cumstances inside the atom, a carbon atom is really capable of forming up to four

stable covalent bonds even with other carbon atoms. This fact leads to the variety of

its allotropic forms (Figure 1.1).

These carbon structures are usually considered as three, two, one and zero dimen-

sional according to their symmetry properties. Diamond is a three-dimensional (3D)

structure because of the tetrahedral symmetry of sp3 hybridized carbon atoms which

build up the diamond lattice. sp2 hybridized carbon atoms build up the structure of

graphite (2D), nanotubes (1D) and fullerenes (quasi-0D) with a π-electron system de-

11

Figure 1.1: Allotropic forms of carbon [1].

localized over the whole σ-framework. For sp2 carbon atoms, the energetically favored

arrangement is trigonal planar, however in nanotubes and fullerenes the carbon atoms

are not in planar arrangement due to the curvature of these structures.

The strain in the σ-framework and π-electron system leads to the interesting physi-

cal and chemical properties of nanotubes, which is placed between that of graphite and

fullerenes.

1.1.2 Basics of carbon nanotubes

Carbon nanotubes are hollow cylindrical forms of graphitic sheets (Figure 1.1 c).

The multi-walled form of carbon nanotubes (multi-walled carbon nanotubes, MWNT)

was discovered by S. Iijima in 1991 (while using the arc-discharge method developed

for synthesis of fullerenes) [2], the single-walled form (single-walled carbon nanotubes,

SWNT) two years later [3]. The diameter of a SWNT is typically 0.5-2.5 nm, its length

is 50 µm-4 mm.

Theoretically SWNT can be derived from a single graphene layer rolled up seam-

lessly to form a cylinder [4]. We can choose a chiral vector, Ch on the graphene sheet

as in Figure 1.2. This will constitute the circumference of the nanotube (MWNT are

built up by coaxial SWNT). The lattice vectors of the graphene sheet, a1 and a2, build

up the Ch vector as Equation 1.1 shows, where n and m are integer numbers:

Ch = na1 +ma2 (1.1)

12

a1

a2

(n,0) zigzag

n m

T

(n,n) armchair

Ch=(7,4)θ

Figure 1.2: The honeycomb structure of the graphene sheet showing the chiral vectorCh, constructed from the lattice vectors a1 and a2, and the three main groups of single-walled nanotubes: (n, 0) zigzag, (n, n) armchair and (n,m) chiral. θ is the chiral angle,the angle between Ch and a1. The brown rectangle is the unit cell of the nanotube havingthe Ch vector as a circumference. T is the normal vector of Ch and the translationaldirection of the unit cell in order to construct a nanotube with an arbitrary length.

13

These (n,m) indices determine all geometrical-physical parameters of the nanotube,

like the diameter d :

d =|Ch|π

=a0π

√n2 + nm+m2 (1.2)

where a0 is the length of the base vector in graphene (0.2461 nm); the chiral angle θ;

the indices of the translational vector T (t1, t2):

t1 =n+ 2m

gcd(n+ 2m, 2n+m); t2 =

2n+m

gcd(n+ 2m, 2n+m)(1.3)

where gcd is the operator for greatest common divisor; the number of atoms in the unit

cell N :

N =4 · (n2 + nm+m2)

gcd(n+ 2m, 2n+m)(1.4)

the electronic structure and conduction properties (metallic or semiconducting), etc.

There is an infinite number of ways to roll up the graphene sheet into a cylinder

resulting in nanotubes with different properties. This vast number of nanotubes can

be divided into three main groups according to the shape of their unit cells: armchair,

zig-zag and chiral tubes (see Figure 1.3). Due to the six-fold symmetry of the graphene

sheet, structurally different nanotubes can be obtained only in the Θ = 0−30 range [5].

In reality, the length of a carbon nanotube is not infinite, and its end is usually closed

by a fullerene-like cap. A real single-walled carbon nanotube has three main structural

parts with different chemical and physical properties: the graphitic-like sidewall, the

fullerene-like cap (or an open end) and defects in the sidewall: absence or excess of

carbon atoms (pentagons or heptagons), contaminant atoms, distortion of the hexagons,

etc. These facts lead to a quite large variety of physical and chemical properties and

chemical reactions.

A real bulk nanotube sample is a black powder with very low density, mixtures of

bundled nanotubes with different chiralities and a given diameter and length distribu-

14

Figure 1.3: The three main groups of SWNTs. The translational period is indicated [6].

tion. Figure 1.4 shows an AFM image of an SWNT nanotube sample produced by the

electric arc-discharge method (see Section 1.1.3).

It can be derived from the band structure and the symmetry properties of the

graphene sheet by using tight-binding approximation, that the ratio of semiconducting

and metallic nanotubes is 2:1. If |n−m| = 3k, where k is an integer number, the nano-

tube is metallic, otherwise semiconducting. Due to curvature effects on the electronic

structure of the nanotubes, the smaller the tube diameter, the larger the deviation from

the 2:1 ratio. However, armchair nanotubes are always metallic. Even in thin armchair

nanotubes, the conjugated bond structure is parallel to the tube axis and extends to

the whole length of the tube. Thus, these electronic states are extended to the whole

molecule, they do not have any nodal planes with a component perpendicular to the

tube axis.

15

Figure 1.4: AFM image of a typical nanotube sample: the spaghetti-like bundles shownin the image are built up by ropes of individual nanotubes (own measurement).

1.1.3 Growth of carbon nanotubes

Carbon nanotubes are produced by decomposition of carbonaceous materials. The

so-formed carbon atoms will build up the nanotubes. By varying the parameters

(method, reactor geometry, starting materials, atmosphere, temperature, catalysts,

etc.) different samples are produced. The resulting material contains several types of

nanotubes, so it has a narrower or broader diameter and length distribution. More-

over, the individual nanotubes are not grown in a well-ordered structure, but they

form bundles randomly. In the bundles, different nanotubes stick together very tightly.

Besides, the sample always contains contaminants like amorphous carbon or residue of

catalysts.

In the following, a brief summary of nanotube growing methods is given:

16

Electric-arc discharge method

Striking an arc between graphite electrodes in an inert atmosphere (usually He,

He/Ar) is the traditional technique. Pyrolitic graphite and graphitic nanoparticles are

the contaminants. The process is quite simple and the product has a high structural

quality, even after purification [7, 8]. For synthesizing MWNT there is no need for

catalysts, but it is necessary for growing SWNT. Typical catalysts are Fe, Co, Ni, Y,

Gd, Fe/Ni, Co/Ni. The average diameter of so-produced SWNT is 1.2-2.0 nm [3,9–11].

Laser ablation

A piece of graphite (or graphite-metal/metal-oxide mixture) is vaporized by laser

irradiation at high temperature and in inert atmosphere. The carbon particles are swept

by a gas flow and deposited on a cooled collector. Default product without transition

metal catalyst is MWNT, catalyst is needed to produce SWNT [12–15].

Using catalysts, high quality but highly bundled SWNT can be produced. Their

average diameter is around 1.2-1.4 nm [15].

Chemical vapor deposition

While arc-discharge and laser ablation are called ”high temperature” (>3000 K)

and ”short time reaction” (10−3 − 10−6 s) methods, chemical vapor deposition (CVD)

is a medium temperature (700 − 1500 K) and long time (minutes to hours) reaction.

Advantages of CVD or CCVD (catalytic chemical vapor deposition) methods compared

to the former ones are that they contain less carbonaceous impurities, so only the

catalytic particles have to be removed.

Technically, a simple equipment consists of a quartz tube in a furnace. Hydrocarbon

(or other carbon-containing) gas is led into the quartz tube. In the presence of a metal

catalyst, thermal decomposition of the gas takes place. Carbon nanotubes grow over and

around the catalyst particles. Substrate materials are usually Si, mica, silica, quartz,

alumina. By varying the parameters (temperature, composition, catalyst, time, gas flow

etc.) physical and chemical properties can be varied [16].

17

One of the most popular CVD methods is the high pressure carbon monoxide, HiPco

method, where the thermal decomposition of Fe(CO)5 provides in situ the carbon

source CO and the catalytic Fe nanoparticles [17]. Average diameter of HiPco nanotubes

is 0.7-1.4 nm [16,17].

CoMoCat nanotubes are also CVD tubes, but using a Co-Mo mixture as a catalyst.

Carbon sources are usually CO, CH4 or C2H2 [18].

Finally, it must be emphasized that by varying the reaction conditions the consti-

tution of the products (diameter and length distributions, size of the bundles, func-

tionality) can be designed within certain limits, but an overall control has not yet been

achieved.

1.1.4 Electronic structure of single-walled carbon nanotubes

In graphite the fourth valence electron of each carbon atom is delocalized over

the whole honeycomb-like plane. It is demonstrated by the fact that the electrical

conductivity of graphite bulk material is 100 times greater along the planes than in the

perpendicular direction [19].

Let us consider a tubularly wrapped graphene sheet. This is a quasi-one dimen-

sional, ideal nanotube. The z direction, which is the direction of the nanotube axis,

will be infinite, and the electronic states in this direction will be continuous. Along the

circumference of the tube, electronic states are quantized by |Ch|.

This kind of quantization is a peculiarity of all one-dimensional electronic systems,

and has an important consequence: there appear narrow energy ranges in the solid’s

electronic band structure, which are allowed for a huge number of electrons. It leads to

the appearance of sharp peaks at these specific energies in their electronic density of

states (DOS), called Van Hove singularities [20–22].

Practically, from a chemist’s point of view, Van Hove singularities can be considered

as discrete molecular energy levels superimposed on the continuous solid electron states

(bands). In semiconducting nanotubes, there are forbidden states between the highest

energy level, where electrons are in the ground state (analogy with HOMO), and the

18

Figure 1.5: Electronic density of states (DOS) of (9,9) metallic and (10,8) semicon-ducting tubes (diameter: 1.22 nm). Optically allowed symmetric transitions betweenVan Hove singularities are indicated: M11 - 1st metallic transition; M22 - 2nd metallictransition. S11 - 1st semiconducting transition; S22 - 2nd semiconducting transition; S33

- 3rd semiconducting transition. M00 notation is used for the contribution of free chargecarriers (at zero energy) [23].

next allowed unoccupied state (analogy with LUMO). This energy range is called band

gap. In metallic nanotubes the density of states is finite (non-zero) and is populated

up to the 0-level (Fermi level) in the ground state (Figure 1.5).

1.1.5 Optical transitions of single-walled carbon nanotubes

Using optical spectroscopy to investigate the electronic structure and chemical

bonds of carbon nanotubes is plausible. Nanotube spectra have unique properties

originating from the one-dimensionality, diameter distribution and bundling. One-

dimensionality causes the above mentioned sharp peaks, the Van Hove singularities

in the DOS. Electronic transitions between these Van Hove singularities cause intense

peaks in the optical spectrum usually in the mid-infrared–visible range, superimposed

on the continuous electron states. Since electronic transition energies of nanotubes

with different diameters are very close to each other, diameter distribution of the sam-

19

ple causes broadening of the peaks related to the Van Hove singularities. Bundling of

the nanotubes changes the band structure through intertube interactions and lowering

symmetry properties.

The impurities, like amorphous carbon and remaining catalyst particles also con-

tribute to the optical spectrum. They increase the background and complicate the

evaluation of the measured spectra.

The electronic transitions between Van Hove singularities are demonstrated in Fi-

gure 1.5. These transitions can be easily induced optically and dominate the nanotube

spectrum shown in Figure 1.6.

Because of the one-dimensional structure of nanotubes, selection rules of optical

transitions are dependent on the polarization of the exciting light. Only symmetrical

transitions between Van Hove singularities are allowed, when light polarization is paral-

lel to the nanotube axis. For light polarized perpendicular the tube axis, selection rules

are different, but the intensity of these transitions is negligible beside the symmetrical

ones.

Plotting electronic transition energies versus tube diameter, the widely used

Kataura plot is obtained. It was developed by H. Kataura in 1999 using measured data

and calculations by the zone-folding method (Figure 1.7) [25,26]. The Kataura plot has

been subject to continuous improvement in experimental and theoretical techniques

over the years, and as a result, it has been refined but not changed substantially.

With increasing tube diameter the points in the Kataura plot get more dense, which

involves the densifying of the single peaks in the optical spectrum at high energies and

hardening the assignment to a given chirality. In the Raman spectrum, tubes having

transition energies that match the laser energy will be in resonance and provide an

enhanced Raman intensity.

The vibrational modes in the mid-infrared region are missing from the absorption

spectrum. The main reason for this absence is that in carbon nanotubes the transition

dipole moments connected to vibrational transitions are small [27].

20

0 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 5 0 0 0 0

π−π*

M 1 1S 2 2

S 1 1

M 0 0

Abso

rbanc

e (a.

u.)

W a v e n u m b e r ( c m - 1 )

S 3 3 + M 2 2

7 8 5 n ma )

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0

M 1 1S 2 2

S 1 1

M 0 0

Abso

rbanc

e (a.

u.)

W a v e n u m b e r ( c m - 1 )

S 3 3 + M 2 2

b )

Figure 1.6: (a) Absorbance spectrum of a real SWNT sample consisting of many differ-ent (n,m) tubes. The laser energy used in this work (785 nm) is indicated; (b) Zoomedin to the range of relevant electronic transitions. Each symmetric transitions has adistinct peak consisting of tubes with different diameters [24].

21

0 1 2 30

1

2

3 ..M 2 2

M 1 1

S 4 4

S 3 3

S 2 2

s e m i c o n d u c t i n g m e t a l l i c

Energ

y sep

aratio

n (eV

)

N a n o t u b e d i a m e t e r ( n m )

γ0 = 2 . 9 e V , a C C = 0 . 1 4 4 n m

7 8 5 n m

S 1 1

.

Figure 1.7: By plotting the energy of electronic transition vs. tube diameter, the socalled Kataura plot can be obtained [23]. Raman laser line used in this work (785 nm)is indicated. Tubes on this line are in resonance and have enhanced intensity in theRaman spectrum obtained by using this laser. With gray rectangle is indicated thediameter distribution of the nanotube sample, which absorption spectrum is shown inFigure 1.6.

1.1.6 Vibrational properties and Raman spectrum of single-

walled carbon nanotubes

The one-dimensional nature of carbon nanotubes has pronounced effects also on

their vibrational properties. As it was mentioned above, absorption spectroscopy is

unsuitable for investigation of the vibrational modes of pure nanotubes since they

have a very weak contribution. In contrast, Raman spectroscopy has proven to be a

very effective technique [28,29]. Raman active vibrational modes can have much larger

intensity due to resonance effects compared to IR active vibrational modes.

In the case of carbon nanotubes, a resonance effect is present, when the laser energy

matches that of a real electronic transition in the sample. These electronic transitions

appear in the absorption spectrum shown in Figure 1.6 as transitions between the Van

Hove singularities. Resonance causes enhanced intensity compared other excitations.

This, on the one hand, improves the quality of the Raman spectra, on the other hand,

22

0 2 5 0 5 0 0 7 5 0 1 0 0 0 1 2 5 0 1 5 0 0 1 7 5 0 2 0 0 0

Rama

n inte

nsity

(a. u.

)

R a m a n s h i f t ( c m - 1 )

R B M

D - b a n d

G - b a n d

Figure 1.8: General Raman spectrum of an SWNT sample [6].

it can cause complications in the evaluation, since the detected intensity does not

correlate with the amount [30].

The significance and popularity of Raman spectroscopy in nanotube characteriza-

tion is in its fast, simple and non-destructive features. Many processes can be followed

in situ by Raman spectroscopy. There is no need for special sample preparation, and

in the past few years there were important developments in the methodology. With

Raman spectroscopy, the electronic, vibrational and (under external pressure) elastic

properties of nanotubes can be investigated.

For investigating the electronic properties of carbon nanotube samples, absorption

spectroscopy is much more effective than Raman spectroscopy because there occurs no

complication of resonance effects.

The Raman spectrum of a solid, pure SWNT sample is shown in Figure 1.8. Three

main parts of the spectrum can be distinguished: at low wavenumbers the RBM (Ra-

dial Breathing Modes) region can be found, which appears only in nanotubes out of all

carbon allotropes. This is a collective expansion-contraction motion (like breathing) of

the carbon atoms building up the cylinder of the nanotube. The frequency of this mode

is inversely proportional to the tube diameter [29]. The RBM mode is usually used for

23

determining the diameter distribution of the sample by using several excitation wave-

lengths. For a precise characterization, more lasers are needed because of resonance: in

a bundled nanotube sample with 30-100 different diameters, there must be tubes with

an electronic excitation at the laser energy (resonance effect). RBM peaks of these

tubes will have anomalously large intensity. In some cases, this becomes a drawback:

from a simple Raman experiment, the quantitative determination of the amount of a

nanotube with a specific diameter is not possible.

Between 1200-1400 cm−1 the D-band (Disorder Induced Mode) can be found. This

mode can also be found in most forms of carbon. The origin of the D-band in nanotubes

lies in defects in the tube structure, like heptagons, pentagons, sp3 C atoms, ends and

caps of the tubes. All these defects increase the D-band intensity. In functionalized

nanotubes prepared from a good quality starting material, the increase of the intensity

of the D-band can be used for the quick verification of the functionalization.

The drawback of Raman measurements over absorption measurements in the case

of investigating chemical reactions of carbon nanotubes is that the change of the D-

band does not provide information about the selectivity of a reaction, only about its

occurrence.

The G-band around 1500-1700 cm−1 is also characteristic of graphite and nanotubes.

The origin of the G-band is the superposition of the tangential shifts of the C atoms

on the surface of the tubes.

1.2 Chemistry of single-walled carbon nanotubes

1.2.1 Structure and reactivity

Special properties of nanotubes compared to a simple graphene sheet play a very

important, moreover, determinative role in physical and chemical behaviour of real

nanotube samples. The differences between the ideal and real world become more and

more dominant as the diameter of the nanotube decreases, since the effect of finite

curvature and intertube interactions gets more significant [31].

24

Figure 1.9: Diagrams of (a) metallic (5,5) SWNT, (b) pyramidalization angle (θP ), and(c) the π-orbital misalignment angles (φ) along the C1-C4 bond in the (5,5) SWNTand its capping fullerene, C60 [32].

The first, chemically interesting statement is that a nanotube can be considered as

a cylindrical macromolecule expected to be chemically fairly inert. However, like other

non-planar conjugated systems, there are two curvature-induced effects that must be

considered: pyramidalization of conjugated carbon atoms and π-orbital misalignment

between adjacent carbon atoms (Figure 1.9) [32]. Both cause local strain, and therefore

enhance reactivity towards addition reactions. During addition, sp3 carbon atoms are

formed with tetrahedral symmetry which lowers local strain, and somewhat relaxes the

structure. Compared with fullerenes, which are curved in three dimensions, and with

graphene, which is not curved, the reactivity correlates with structural strain, therefore

expected to be placed between that of fullerenes and graphene.

In nanotubes - contrary to fullerenes - π-orbital misalignment is the main source of

strain. This effect is the most relevant in small-diameter tubes at C-C bonds enclosing

a higher angle with the nanotube axis [33].

Also, bond curvature determines the reactivity, especially towards [2+1] cycload-

ditions. There are three nonequivalent C-C bond types in nanotubes, defined by the

angle θ. θ is an acute angle between the tubular axis and the C-C bond in the plane of

25

the unfolded SWNT. θ is just equal to the chiral angle, the other two angles are 60−θ

and 60 + θ [34].

In extremely thin nanotubes, the curved surface causes rehybridization. Originally,

σ and pz orbitals are orthogonal to each other. Curvature induces a finite overlap

between them. The energy of the σ∗ state shifts upward a little, while the energy

of π∗ state shifts downward. These shifts increase with decreasing tube diameter. In

semiconducting tubes with <1 nm diameter, the π∗ band can slide into the bandgap,

and make them metallic [35,36].

1.2.2 Challenges in nanotube chemistry

Comparing a flask of nanotubes with an ordinary chemical system, a chemist faces

the following challenges. First, nanotubes are insoluble in any solvent. Dispersions by

using surfactants can be made with different stability, but sedimentation starts in a

short time. To get a homogeneous dispersion, nanotubes must be covered strongly

with surfactants (then usually it is impossible to get rid of them without re-bundling

of the nanotubes) [37], functionalized with solubility-increasing side groups (this may

completely change the original properties) [38,39], or modified by adsorption of charged

species [40].

Second, a nanotube sample is a mixture of 10-100 different (n,m) nanotube types.

This varies with growing techniques and circumstances, even from batch to batch pre-

pared by the same technique in the same lab. The nanotube composition can be more

or less well determined by Raman spectroscopy using several lasers.

Third, nanotubes form close-packed bundles held together by van der Waals forces.

Tubes inside a bundle can be more difficult to reach by a reactant species. A particular

tube could be highly reactive, but useless in a reaction which cannot unfold the bundle:

that tube will remain intact, and shows itself less reactive.

1.2.3 Reaction types and sites in nanotubes

A real nanotube has three different parts regarding the chemical reactivity:

26

• perfect sidewall: sp2 carbon atoms in a honeycomb lattice

• defects: sp3 carbon atoms, lack of atoms, Stone-Wales defects

• fullerene-like cap or an open end

Nanotube modifications can be divided into non-covalent, covalent and encapsula-

tion methods. Non-covalent methods are usually used for increasing the solubility of

the nanotubes, or further the adsorption of other functional molecules or particles.

Functionalization means collectively those reactions that provide a special func-

tionality. Most of the chemical reactions on nanotubes are performed on buckypapers.

Buckypapers are thin sheets made of nanotubes by filtering nanotube dispersions. Na-

notube bundles in the buckypaper form are oriented in 2D (because of the filtration),

while in the powder form, they are oriented in 3D.

In the following a short summary of chemistry on carbon nanotubes is given:

Covalent chemistry

Covalent chemistry of nanotubes is very complex. The most evident is addition

reactions to sidewalls (sidewall functionalization; addition reactions only) as it is shown

in Figure 1.10, and adding –COOH or –[F] n groups to the ends and defects of the tube

(defect functionalization; either addition or substitution reactions) as it is shown in

Figure 1.11.

Noncovalent chemistry

Noncovalent functionalization is the most widespread modification of nanotubes

used in applications. This covers usually wrapping by polymer- and biomolecules for

special functionality, attachment and covering by surfactants to enhance solubility,

attachment of molecules with π-systems by van der Waals and π − π interactions,

inorganic nanoparticles (usually oxides of transition metals) via polymer wrapping.

The advantage of this type of modifications is that the original electronic structure of

the tubes can be retained [43].

27

Figure 1.10: Summary of sidewall covalent chemistry of nanotubes [41].

28

Figure 1.11: Summary of defect functionalization through –COOH and –[F]n groups[42].

Encapsulation

Filling different species into carbon nanotubes (X@CNT) was reported first by

Smith et. al. in 1998. They filled C60 into nanotubes [44]. Since then, several organic

molecules, metal organic complexes [45, 46], metal clusters [47], endohedral fullerenes

etc. [43] were also used for encapsulation. Nanotubes can be used as nanoreactors for

oriented reactions, like polymerization, fabrication of nanoribbons [48,49], and double-

walled carbon nanotubes [50].

In the following I will focus on those two reductive addition reactions which were

used in the frame of this work.

1.2.4 Classical and modified Birch reduction

Birch reduction was developed by A. Birch in 1946-1949 for the hydrogenation of

aromatic compounds, especially benzene [51,52].

The base of the reaction is the dissolution of alkali metals in liquid ammonia (thus

they are also referred to as dissolving metal reductions) by producing ”solvated elec-

29

trons”: these are the blue colored complexes of [M(NH3)6]+e−, where M=Li, Na or

K [53]. These complexes act as very strong and selective reducing agents.

The mechanism of the classical Birch reduction is shown in Figure 1.12. The main

steps are the addition of an electron to the aromatic system resulting in a carbanion

and a radical in para position. The carbanion part reacts with electrophilic agents, like

protons, alcohols, carbocations resulting in a saturated C atom and a newly formed

carbanion also in para position. This carbanion can also react with electrophilic agents.

Figure 1.12: The mechanism of the classical Birch reduction [6].

In modified Birch reductions, the primary reducing agents can be other complexes,

for example radical anions formed from aromatic compounds like naphthalene. By using

different solvents and circumstances, these reactions can be directed well and used for

reductive addition reactions of aromatic and conjugated electronic systems.

1.2.5 Alkali metal intercalation

The existence of graphite intercalation compounds (GICs) and the need of tuning

electronic properties of carbon nanotubes have given the motivation for n-doping of

nanotubes: charge transfer reactions with electron donors and synthesizing alkali metal

intercalated carbon nanotubes [54–58].

Intercalation is different from the above mentioned encapsulation. Encapsulation

happens inside the hollow of an individual nanotube, but intercalation happens in a

nanotube bundle, the intercalating species are located between the individual tubes in

the interstitial channels.

30

Alkali metal intercalated nanotubes are at least as air- and water sensitive as al-

kali metals themselves. Intercalation can be achieved by electrochemical and simple

chemical means [59, 60]. One of the first observations was the work of Nalimova et

al. [61]. They reacted Li with carbon nanotubes in large excess at room temperature in

ultrahigh vacuum. Li intercalation was proved by XRD and IR studies. They pointed

out that using Li in excess may lead also to the insertion of Li into the hollow of the

individual tubes (encapsulation).

Suzuki et al. studied K and Cs intercalation at room temperature [62]. They found

that intercalation into the bundles is reversible and has an equilibrium value of KC24,

CsC24 and CsC8.

Bower et al. showed the reversibility of intercalation and structural disorder by

TEM and EELS measurements [63].

After investigating the structure and conduction properties of alkali metal interca-

lated nanotubes, and the reversibility of the process, Pichler et al. determined more

precisely the electronic spectra of K-intercalated materials paying attention to the origi-

nally unoccupied electronic states. They found that the electronic and optical properties

can be well described within the frame of a simple Drude-Lorentz model [64,65].

The first review of nanotube intercalation was provided by J. E. Fischer [66]. He

suggested a model with chain-like structure in the triangular and hexagonal channels

between the tubes. The structural parameters are determined by the size of interstitial

channels and the Coulomb repulsion between the alkali ions. Alkali metal intercalation

destroys crystallinity, although the process is almost completely reversible.

Bendiab et al. studied possible Rb positions and Rb-doped nanotube bundles by

X-ray and neutron diffraction and used ab initio calculations [67]. They found that

the most probable and energetically favorable positions for a single Rb ion are the

interstitial channels between three tubes in the bundle. Reaching saturation (RbC8),

these sites will be no longer available, and the outer surface of the bundles becomes

preferential for further Rb uptake.

31

Diameter-selective doping was presented by Kukovecz et al. [68]. They showed by

resonant Raman spectroscopy and conductivity measurements that the intercalation

of potassium depends on the diameter of nanotubes. They studied pulsed laser vapor-

ization (PLV) and HiPco tubes (1.2-1.4 and 0.8-1.2 nm diameter, respectively). From

radial breathing modes (RBM) of the Raman spectra, they found that the doping level

showed reversed dependence on the tube diameter below a certain limit. They calcu-

lated the required relative lattice expansion as a function of tube diameter (the size of

intertube channels naturally increases with increasing tube diameter). From these data

they obtained that above 1.4 nm tube diameter there is no need for lattice expansion.

Between 1.4 and 0.7 nm diameter larger and larger lattice expansion is required. Be-

low 0.8 nm, the energy gain of charge transfer is not enough for covering the energy

requisite, and hence, the intercalation becomes hindered and unfavorable.

Kukovecz et. al. also observed an unexpected intensity loss in the RBM region of

small diameters. They explained this by the higher dopant/C ratio (decreasing number

of C atoms per unit length in smaller diameter tubes), higher quenching of resonance of

nanotube dopants and the increasing stacking of smaller diameter tubes. Because of the

smaller number of C atoms per unit length in thinner tubes, the secondary interactions

are weaker, thus they can be more easily pushed apart by the dopants. Because of these

competing effects, they expected a minimum in doping level for K between 0.9-1.2 nm

tube diameter.

Vigolo et al. investigated K, Rb and Cs-intercalated nanotube bundles and their

dispersibility by Raman spectroscopy and transmission electron microscopy [69]. By

alkali intercalation, SWNT dispersibility in dimethyl-sulfoxide (DMSO) increased in-

dicating debundling process and charge transfer. Debundling was confirmed by TEM

images.

Debundling and resulting increase in the volume of a buckypaper sample was also

described by Tanaike et al. [70]. Swelling, like in GICs, and highly increased solubility

in DMSO was detected caused by electrochemical intercalation of Li ions.

32

In summary, intercalation of nanotubes by alkali metals is extensively studied and

also a practically interesting field of nanotube science. However, covalent chemical

reactions of alkali metal intercalated nanotubes have not yet been published.

1.3 Optical spectroscopy of carbon nanotube thin

films

Optical spectroscopy is beyond doubt one of the most important and useful tech-

niques to investigate material systems. It is based on the interaction of electromagnetic

radiation with matter. Electromagnetic radiation transmits energy to the system un-

der consideration. The system absorbs an exact amount of energy depending on the

relation between its intrinsic properties and the electromagnetic radiation. Here we con-

sider optical spectroscopy in the far infrared-ultraviolet range of the electromagnetic

spectrum (λ=400 µm–190 nm; E=3.1 meV–6.5 eV). Absorbing these energies changes

the rotational, vibrational and/or electronic states of the system.

The response to electromagnetic radiation is determined by the system’s micro-

scopic (molecular, electronic and crystal structure) and macroscopic properties (shape,

thickness, heterogeneity etc.).

1.3.1 The Drude-Lorentz model [24,71]

The Drude-Lorentz model is a very simple, commonly used model to describe optical

functions of solids, and based on the damped harmonic oscillator model. Due to the

scales (optical wavelengths are much larger than atomic distances), both light and

matter can be treated classically.

When describing electronic transitions, we start from the band structure. Optical

transitions inside a band (like behavior of free electrons in metals) can be described

by the classical Drude model. Excitations between the bands are modeled by a sum of

Lorentz oscillators. Both vibrational and electronic excitations can be discussed within

the framework of the dielectric formalism.

33

When light interacts with matter, it can be reflected and absorbed, and the residual

amount will be transmitted. We can measure the reflected and transmitted light in-

tensity relative to the incident intensity (IR/I0 and IT/I0, respectively). The absorbed

intensity, IA/I0 can be calculated by using the relations between the intensities given

by Equation 1.5:

I0 = IR + IT + IA (1.5)

To describe light absorption in non-magnetic insulators, bound electrons can be

modeled as damped harmonic oscillators excited by the E = E0eiωt electric field

(Lorentz-model):

med2r

dt2+meγ

dr

dt+meω

20r = eE (1.6)

where e is the charge, me is the mass of electron, r is the momentary deflection, γ is

the damping, ω0 is the excitation frequency.

The solution of Equation 1.6 is:

r(ω) =eE

me(ω20 − ω2 − iωγ)

(1.7)

The dielectric function of the Lorentz oscillator, which describes its response to the

exciting electric field:

εr(ω) = 1 +e2

ε0me

1

ω20 − ω2 − iωγ

(1.8)

where ε0 is the dielectric constant of vacuum.

In real systems, high frequency excitations also contribute to the dielectric function,

that can be considered as a constant (ε∞):

εr(ω) = ε∞ +e2

ε0me

1

ω20 − ω2 − iωγ

(1.9)

34

Metallic electrons are considered in this so-called Drude-Lorentz dielectric function

as oscillators with zero excitation frequency. Further generalization to the case of many

oscillators (N oscillators in unit volume taking part in the j-th excitation):

εr(ω) = ε∞ +e2

ε0me

∑j

Nj

ω20j − ω2 − iωγj

(1.10)

1.3.2 Calculation of optical functions

The dielectric function, and therefore all optical functions (refractive index, optical

conductivity etc.) are complex quantities. The real and imaginary parts of the Drude-

Lorentz dielectric function are:

ε = ε′ + iε′′ (1.11)

ε′r = ε∞ +e2

ε0me

ω20 − ω2

(ω20 − ω2)2 + ω2γ2

(1.12)

ε′′r =e2

ε0me

ωγ

(ω20 − ω2)2 + ω2γ2

(1.13)

The complex refractive index N and the complex optical conductivity σ are defined

by Equations 1.14 and 1.15, respectively:

N =√εr = n− iκ (1.14)

σ(ω) =[εr(ω)− ε0]ω

i(1.15)

In Equation 1.14, the real part of the complex refractive index (n) is in connection

with reflectance (change in the phase), and imaginary part, κ is in connection with the

exponential decrease of the amplitude.

35

The well-known and widely used Beer’s law makes connection between the optical

density D calculated from the transmittance T (T = IT/I0), the concentration c and

the thickness d of the sample:

D = − log T = εcd (1.16)

where ε is the extinction coefficient.

Beer’s law is valid only for sufficiently low concentrations and when reflection can

be neglected. In this case D = A, where A is the absorbance.

In our case, the simple Beer’s law is not sufficient. On the one hand, nanotube sam-

ples always contain metallic tubes. Their reflectance is not negligible in the far-infrared

region, and has an effect even in the near-infrared region. The relevant information

from our point of view is mostly in the near-infrared part of the spectra. On the other

hand, the Drude-Lorentz fit of the complex dielectric function (or any other optical

functions) will be proper, relevant and contain every information if both the real and

the imaginary parts are known.

In order to get both components of the complex optical functions of our samples,

we apply the single layer model, which works well for self-supporting nanotube thin

films, the form of samples measured in our experiments.

The single layer model describes the transmittance of a layer with finite thick-

ness d, and with parallel surfaces surrounded by vacuum. Inside the layer multiple

reflection-transmission events happen. The sum of these events appear in the transmit-

ted intensity, IT .

The transmittance of a single layer:

T =ITI0

= |t|2 (1.17)

where t is the transmission coefficient.

t = |t|eiφ =√Teiφ =

4N

(1 + N)2e−iδ − (1− N)2e−iδ(1.18)

36

where δ = 2πωdN , ω is the wavenumber of the exciting light, d is the thickness, N is

the complex refractive index of the layer [21,24]

Thus, for obtaining the total optical function of the system, we need to know the

thickness of the sample d and the phase φ. d is available experimentally: thickness of

nanotube thin films can be measured by atomic force microscopy (see Experimental).

φ is experimentally unavailable, but under specific conditions (which luckily hold in

our case) it can be calculated by using the Kramers-Kronig relations.

37

2 | Experimental

2.1 Objectives

As it was shown in the Introduction, carbon nanotubes form bundles of different

tubes held together by van der Waals forces. The first step, when one applies chemical

reaction to nanotubes is to bring reactants close to the tube surface, preferably to each

and every one with equal chance.

This is quite well solved by forming negatively charged nanotubes through reduction

reactions. When each tube becomes negatively charged, Coulomb repulsion will exfoli-

ate the bundles. This phenomenon is well represented by Birch reductions [72–75]. In

these reactions, ammonia, tetrahydrofuran and ethylenediamine are used as solvents,

and lithium or potassium as reductive agents. After applying this step, electrophilic

species can attack the so-formed carbanions and result in different functionalized ma-

terials. A similar process happens in alkali metal intercalated graphite resulting in

partially hydrogenated graphite [54,76].

Following the example of graphite intercalation compounds, direct reduction by in-

tercalating alkali metals into carbon nanotube bundles were applied [54,70,77]. Based

on these results and the similar reactivity of graphite and carbon nanotubes, alkali

metal intercalation is expected to exfoliate and reduce nanotube bundles. By this

method, the step of carbanion formation during the reductive modification could be

necessarily separated in space and time from the step of attachment of the functional

group (unlike other Birch-type reactions) [78].

38

Nanotube Company TypeDiameter range Mean diameter

(nm) (nm)

P2 Carbon SolutionsO2-purified

1.2-1.7 1.60arc-discharge

HiPco CNI Nanotechnologies CVD 0.8-1.3 1.08

CoMoCat CGSouthWest

CVD 0.57-1.17 0.90NanoTechnologies

Table 2.1: Specific parameters of investigated single-walled carbon nanotubes.

When adding alkali metals to carbon nanotubes in excess, a stable phase (with

composition KC27 in case of potassium determined by XPS) is formed in a few minutes

at 180 C [79]. Over longer time a saturation concentration (about KC9 excluding

encapsulation) is reached. Sidewall functionalization of the nanotubes may be able to

proceed from phase KC27 by electrophilic addition.

The general selectivity of sidewall reactions in nanotubes is believed to be deter-

mined principally by structural strain caused by π-orbital misalignment on the curved

surface, resulting in higher reactivity of smaller diameter tubes [32]. Taking into ac-

count, however, that realistic nanotube samples consist of bundles, there are other

important phenomena that must be considered, such as kinetics, steric effects and en-

ergetics of all processes and intermediate products in a reaction.

In my thesis I studied reductive addition reactions of single-walled carbon nanotubes

and SWNT bundles investigating the diameter and other selectivity of these reactions.

Several types of commercial single-walled carbon nanotubes were used as starting

materials representing a wide diameter range, overall between 0.57-1.7 nm. Parameters

of the starting nanotubes are summarized in Table 2.1.

The products were investigated by thermogravimetry-mass spectrometry (TG-MS),

1H-NMR spectrometry, Raman spectroscopy and wide range optical transmission spec-

troscopy to obtain detailed information about the reactivity through composition, ther-

mal and optical properties of the samples, with special attention to diameter selectivity.

39

For a more detailed comparison to hydrogenation by alkali metal intercalation, hy-

drogenation of HiPco by modified Birch reduction, using K-naphthalenide in tetrahy-

drofuran was also performed [74,75]. HiPco nanotubes were also modified with n-butyl

groups both by alkali metal intercalation and modified Birch reduction. All the control

samples were studied by TG-MS and Raman spectroscopy.

2.2 Preparation

2.2.1 Materials

The following materials were used:

• SWNT: P2, HiPco, CoMoCat (Table 2.1, as received, without further purification)

• Solvents: toluene (VWR , 99.8+%, redistilled), tetrahydrofuran (VWR, 99.7+%,

redistilled). Anhydrous solvents must be used during the syntheses, because both

the alkali metals and the carbanions are air and water sensitive. Special care has to

be taken all along the reactions. Air sensitive steps have to be performed in inert

atmosphere. In our case we used an Ar-filled dry box. Toluene was cryo-distilled

from Na-K alloy in a vacuum line. THF was redistilled from K-benzophenone.

• Reducing agents: K, Rb (Sigma-Aldrich, 98+%), naphthalene (Sigma-Aldrich,

99+%, as received)

• Reactants: methanol (VWR, 99.8+%, as received), 1-iodobutane (Sigma-Aldrich,

99%, as received)

• Filtering and washing: tetrahydrofuran, ethanol (VWR, 99.7+%, as received), 1:3

HCl:H2O, distilled water, acetone (VWR, 99.8+%, as received)

2.2.2 Synthetic routes

The functionalization method, which was inspired by exfoliated graphite intercala-

tion compounds, is described below. The reaction scheme is shown in Figure 2.1.

40

Figure 2.1: Reaction scheme of reduction of nanotubes by alkali metal and subsequentaddition of hydrogen (n-butyl) groups.

About 100 mg of as-received SWNT was first annealed in dynamic vacuum

(10−6 mbar) at 250 C for 12 hours, followed by transfer into an Ar dry box. In the

dry box, alkali metal (potassium in case of P2 and HiPco and rubidium in case of

CoMoCat) was added in a glass vial, keeping the carbon:alkali metal molar ratio 4:1.

The glass vial was sealed on a vacuum line. Annealing at 200 C for 12 hours was

enough for the alkali metal to intercalate into the nanotube bundles. Intercalation was

indicated by the copper/gold color of the sample [80].

Subsequently, the intercalated sample was taken back to the dry box. The vial

was opened and the intercalated nanotubes were put into a Schlenk-type flask with

a funnel (Figure 2.2). 40 ml anhydrous toluene was added to the flask and 20 ml to

the funnel. Toluene was used as an aprotic solvent to avoid side reactions with any

other H source. Outside of the dry box, sonication was applied for 15 minutes to

enhance the intercalation process. Next, 5 ml methanol was filled fast and carefully

to the funnel. Methanol/toluene was added dropwise into the flask during sonication.

Sonication was continued for 2 more hours, and the mixture was left overnight. The

sample was filtered on a Millipore nylon membrane filter (0.1 µm pore size), washed

with ethanol, 1:3 HCl:H2O, distilled water, ethanol and acetone. Finally, it was dried

in dynamic vacuum at 200 C for 12 hours.

The product obtained this way was transferred back into the dry box. The whole

process described above, except the initial annealing, was repeated two more times in

order to investigate whether it is possible to improve the degree of hydrogenation by

applying successive steps.

The main products of reactions with methanol are hydrogenated nanotubes, but

there are side reactions, such as hydrogen evolution, when attachment of H to the

41

Figure 2.2: The Schlenk-type reaction flask. In the flask, there are K-intercalated nano-tubes. It is clearly seen that they are not dispersed in the toluene, even after sonication,since they are negatively charged in a totally apolar solvent.

nanotube is kinetically hindered, or when the unreacted alkali metal reduces methanol

directly. In case of reactions with 1-iodobutane, n-butylated nanotubes are the product.

Reference samples were made of pristine nanotubes by performing the same steps

as at the hydrogenation reactions (initial annealing, annealing in sealed glass tube,

addition of methanol, washing, annealing in dynamic vacuum), except for the addition

of alkali metal.

The preparation using modified Birch reduction (the reaction scheme is shown in

Figure 2.3) was started by a prior annealing of 100 mg of the as-received HiPco for

12 hours at 250 C in dynamic vacuum (at 10−6 mbar). Subsequently, the nanotubes

were transferred into the dry box. The same Schlenk-type flask as for alkali metal in-

tercalation was used for the reaction. Naphthalene and potassium were dosed in excess.

Then THF was added both to the flask and to the funnel (100 ml and 20 ml, respec-

tively). The sample was left for 15 min. Meanwhile potassium reacted with naphthalene

and composed a dark green complex. This complex reduced the nanotubes to carban-

ions in an equilibrium process. Then the flask was sonicated for 15 min to loosen the

nanotube bundles and to promote the carbanion formation. 5 ml methanol was added

to the funnel. The reactant was added slowly, dropwise during continuous sonication.

The reaction is quite fast, indicated by the almost instant disappearance of the green

42

Figure 2.3: Reaction scheme of hydrogenation (n-butylation) of nanotubes by modifiedBirch reduction.

color of the complex. Small bubbles were also observed, indicating the hydrogen evo-

lution from the side reaction of potassium and methanol, and their excess. Sonication

was kept going for 1 more hour. The sample was left overnight. The same filtering

and washing procedure was applied as after the alkali metal intercalated reaction, but

completed by a step of washing with THF.

Using HiPco, two more series of samples were prepared. By potassium intercala-

tion and by modified Birch reduction, following the processes described above, n-butyl

groups were attached to the HiPco nanotubes. The reagent was 1-iodobutane instead

of methanol. Three successive steps were also performed.

2.3 Characterization

In this section I present the methods that were used for sample characterization.

The conclusions are based on the results of thermogravimetry-mass spectrometric and

wide range optical measurements, while Raman spectroscopic and 1H-NMR results

support well the main results of the former techniques.

2.3.1 Raman spectroscopy

Raman studies were carried out to detect changes in the D/G mode intensity ratio

of the samples. This gives information about the formation of defects (like sp3 carbon

atoms in the sidewall), thus indirectly about the success of the reactions (Section 1.1.6).

Raman spectra were taken by a Renishaw 1000B spectrometer using 785 nm excita-

tion wavelength, with 4 cm−1 spectral resolution. The laser power was kept sufficiently

low in order to exclude heat damage (2.5 mW/µm2).

43

2.3.2 Thermogravimetry-mass spectrometry

Thermogravimetry-mass spectrometry is widely used to determine the composition

and thermal stability of samples. The main point of the measurement is to determine

the mass loss of a sample as a function of temperature. The thermal decomposition

products are introduced into a mass spectrometer to determine their composition and

their contribution to the mass loss. On the basis of these data conclusions can be drawn

on the structure of the sample from the TG (thermogravimetric) and DTG (derivated

thermogravimetric) curves by obtaining the contribution of the single fragments.

In the case of our samples, TG-MS data give information about the side group

content and thermal stability, which is determined by the diameter of the tube. From

these data we could draw conclusions about the efficiency and the diameter selectivity

of the reactions.

Measurements and data evaluation were done by Emma Jakab in the Institute of

Materials and Environmental Chemistry, Research Centre for Natural Sciences.

Mass change with temperature is directly measured by a Perkin-Elmer TGS-2 ther-

mobalance and a HIDEN HAL 2/301 PIC quadrupole mass spectrometer. 2-4 mg sam-

ple in a Pt vessel was heated up to 800 C with 20 C/min rate in Ar atmosphere. A

portion of the volatile products was introduced into the mass spectrometer (operated

at 70 eV in electron impact ionization mode) through a heated glass-lined steel cap-

illary. Ion intensities were normalized to 38Ar isotope of the carrier gas to eliminate

errors resulting from the shift in MS intensities. To measure hydrogen (m/z = 2), a

calibration with TiH2 is necessary. During the measurements, signals of 16 ions can be

followed.

2.3.3 1H-NMR spectrometry

For quantitative determination of H-content, 1H-NMR spectrometry can be well

used. Unlike TG-MS, NMR measures together all the 1H atoms, which are also present

in the residual solvents (toluene, water, ethanol, acetone etc.) and water condensed

44

from the air, and a very little amount can come also from the sample holder. Therefore

it is very important to correct with the H-content of a carefully produced reference

sample.

Wide line 1H-NMR measurements and data evaluation were performed by Monika

Bokor, Tamas Verebelyi and Kalman Tompa in the Institute for Solid State Physics

and Optics, Wigner Research Centre for Physics.

Measurements and data acquisition were accomplished by a Bruker AVANCE III

NMR spectrometer at the frequency of 82.4 MHz with a stability better than ±10−6.

The inhomogeneity of the magnetic field was 2 ppm. Free induction decays (FIDs) were

measured at room temperature. Known amounts (weight) of the nanotubes (typically

7-15 mg of P2 and CoMoCat, 2-3 mg of HiPco samples) or adamantane (99+%, Sigma-

Aldrich) were put in Teflon capsules. The FID measured on the empty capsule was

subtracted from the FID of the actual capsuled sample to correct for background. The

amplitude of the FID at zero time is proportional to the number of 1H nuclei in the

sample [81]. The first 9-10 µs of the FID was lost in the dead time of the spectrometer.

The observed FIDs were extrapolated back to zero time by fitting Gaussian functions

to obtain its zero-time amplitude. The FID of adamantane was used for calibration in

calculating the hydrogen concentrations. There were residual magnetic catalyst parti-

cles in the samples (typically 2-5 w/w%), but they did not disturb the measurement

significantly [74].

2.3.4 Wide range transmission spectroscopy

Wide range optical transmission measurements on carbon nanotube self-supporting

thin films combined with Kramers-Kronig transformation and fitting with the Drude-

Lorentz model give detailed information about the optical transitions in the sample.

From these results, quantitative information can be obtained of the contributions of

tubes with different diameters. Since addition reactions remove electrons from the nano-

tube’s sp2 electronic system, the peak intensity of transitions related to those electrons

will decrease. With proper data evaluation this decrease can be quantitatively deter-

45

mined. From the different degree of decrease of different diameters, we can judge the

diameter selectivity of addition reactions.

Thin film preparation

For wide range optical measurements, self-supporting nanotube thin films were pre-

pared by vacuum filtration (see Figure 2.4).

Typically 5-10 mg nanotube sample was dispersed in a Triton X-100-water solu-

tion (Triton X-100 from VWR, 98+%) to get individually dispersed nanotubes, and

was sonicated for at least 2 hours to obtain a homogeneous dispersion. Sonication

time is dependent on the sample and can vary in a wide range. The dispersion was

left overnight (or several days) to let the bigger bundles precipitate. The top of the

dispersion remained homogeneous and contained more or less individually dispersed

nanotubes. From the top of the dispersion 10-50 ml was added to the funnel of the fil-

ter filled with distilled water and filtered on a nitrocellulose, acetone-soluble membrane

filter (Millipore, 0.1 µm pore size). Special attention was paid to avoid any movement

of the filter. The sample was washed with distilled water, then the membrane filter was

carefully dissolved in acetone. The nanotube thin films were placed over a 2 mm diame-

ter hole on a graphite disc. Graphite is the only suitable frame material, because it has

similar thermal dilatation to nanotube thin films. This is very important, because the

so-made nanotube thin films were annealed in dynamic vacuum at 200 C for 12 hours

in order to get rid of any volatile and dopant species.

For measuring the thickness of the film by AFM, the same filter was used, but

another piece of film was moved on a Si piece and was cut in the middle. The height

of this stage was used considering the difference between the value of Si and smooth

nanotube surface after flattening the raw AFM image (the fold and crease of the stage

was not considered). For precise determination of the thickness, a histogram excluding

the stage values was used. Areas used for thickness determination were 10x10 µm, and

at least 10 areas were used at one sample. The samples were usually 160-290 nm thick.

46

Figure 2.4: Self-supporting thin film preparation. a) filtering; b) filtrate on solublemembrane filter; c) transparent self-supporting thin film [24].

Measurement parameters

Wide range optical transmission measurements were carried out on self-supporting

thin films made of the samples [82]. Transmission data between 25-52500 cm−1 were

recorded, by a Bruker IFS 66v/s FT-IR instrument in the far (25-1000 cm−1) and

mid-infrared (400-7500 cm−1) region, a Bruker Tensor37 in the near infrared (4000-

15000 cm−1), and a Jasco v550 spectrometer in the visible and UV (11100-52500 cm−1).

Measuring in wide spectral range requires using more than one instrument. In dif-

ferent spectral ranges different light sources, optical elements, detectors etc. have to

be used, and also different substrates are needed to support the sample. Evaluation

of the spectra is made difficult by the absorption of the substrate, because there is

no substrate transparent in the whole spectral range. To solve the problem, the best

solution would be to get rid of the substrates. The special mechanical properties of

carbon nanotube thin films permit this [82].

Data evaluation

By using a Matlab-based program written by Aron Pekker, Kramers-Kronig trans-

formations were performed on the transmittance data to calculate phase φ:

φ(ω0) = 2πdω0 −∫ +∞

0

ln(√T (ω)/

√T (ω0))

ω2 − ω20

dω (2.1)

47

where d thickness is determined experimentally by AFM measurements. Standard ex-

trapolations were used to zero and infinite frequencies: at low wavenumbers frequency-

independent, at high wavenumbers power law decrease.

The optical conductivity data were calculated and fitted by Drude-Lorentz oscil-

lators. After subtracting the background and all other peaks related to transitions

between different Van Hove singularities, we obtained the contribution of the specific

peaks [83]. However, these peaks cannot be assigned to single nanotubes, because the

transition energies are closer to each other than the width of the peaks.

Out of the available optical functions, optical conductivity σ was chosen since it

is additive when several independent processes are involved, like light absorption by

different nanotubes [84]. Each process can be modeled by a separate harmonic oscillator:

σ = σ′ + iσ′′ =∑j

σj =∑j

f(ω0,j, ωj, γj) (2.2)

From Equations 1.10 and 2.2 the real part of the optical conductivity can be written

as:

σ′ =e2ω2

meV

∑j

Njγj(ω2

0,j − ω2)2 + γ2jω2j

(2.3)

From the σ′(ω) curve, the parameters of the individual oscillators can be determined

by standard numerical processes. The fitted Drude-Lorentz oscillators represent the

contribution of the transitions between Van Hove singularities of 1D electronic systems

(nanotubes), the excitation of the full π-electron system (π−π∗ transitions in nanotubes

and other carbonaceous materials as contaminants). The π−π∗ transitions (peaks above

∼40000 cm−1) and metallic contaminants, like remaining catalysts are considered as

a constant (broad and weak Drude oscillator), other contaminants are considered as a

few Lorentzians.

In Figure 2.5 the Drude-Lorentz fit of the optical conductivity spectrum (σ′ vs. ω)

is shown, together with the process to obtain the contribution of the specific peaks.

The optical conductivity spectrum has very similar shape as the absorbance spectrum

48

0 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 5 0 0 0 00

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

o p t i c a l c o n d u c t i v i t y b a c k g r o u n d b a c k g r o u n d c o m p o n e n t s b a c k g r o u n d c o r r e c t e d

a )

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 00

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

S 3 3 + M 2 2

M 1 1

S 2 2

S 1 1

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

b a c k g r o u n d c o r r e c t e d s i n g l e t r a n s i t i o n s e x t r a c t e d f i t t e d L o r e n t z i a n s

b )M 0 0

Figure 2.5: a) Fitting and background correction of nanotube sample’s optical con-ductivity spectrum; b) extraction of single transitions from the background-correctedoptical conductivity spectrum and the fitted Drude-Lorentz oscillators. Own measure-ment of P2 reference sample is used to show the evaluation process.

49

shown in Figure 1.6. This was the other reason why we have chosen optical conductivity

representation out of the available optical functions.

Optical conductivity spectra of the starting materials

Figure 2.6 shows the diameter dependence of S11 and S22 transitions of the start-

ing P2, HiPco and CoMoCat nanotubes. It can be clearly seen how the transition

wavenumbers change with tube diameter: the smaller the mean diameter, the larger

the wavenumber. Thus, the lower wavenumber part of the peak is related to the larger

diameter tubes within one sample, the higher wavenumber part to the smaller diameter

tubes. When detecting the different changes in the intensities of the lower and higher

wavenumber parts within one sample, we can draw conclusions about the diameter

selectivity of an addition reaction.

The width of the peaks is due to the diameter distribution of the samples. P2

nanotubes have a more or less narrow, featureless S11 peak indicating that the single

components are very close to each other. The S22 peak of P2 shows a more complicated

structure, since the energy of these transitions is separated more from each other (see

Figure 1.5).

HiPco nanotubes have very wide and much more structured S11 and S22 peaks. This

indicates that HiPco has a wide diameter distribution but contains quite less types of

nanotubes than P2.

CoMoCat nanotubes show also wide peaks, but it is because of their small mean

diameter (large difference in transition energies between tubes with close diameters).

Accordingly, these wide peaks are very much structured due to the small number of

constituting (n,m) tubes.

50

2 0 0 0 4 0 0 0 6 0 0 0 8 0 0 0 1 0 0 0 0 1 2 0 0 0 1 4 0 0 00

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0 P 2 ( 1 . 6 0 n m ) H i P c o , x 0 . 2 ( 1 . 0 8 n m ) C o M o C a t ( 0 . 9 0 n m )

Op

tical co

nduc

tivity

(Ω-1 cm

-1 )

W a v e n u m b e r ( c m - 1 )

a )

5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 00

5 0

1 0 0

1 5 0

2 0 0

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

P 2 ( 1 . 6 0 n m ) H i P c o , x 0 . 2 ( 1 . 0 8 n m ) C o M o C a t ( 0 . 9 0 n m )

b )

Figure 2.6: Diameter dependence of a) S11; and b) S22 transitions. The tendency towardslower wavenumbers with increasing mean diameters can be clearly seen.

51

3 | Results and discussion

In this section I present the results obtained by Raman spectroscopy,

thermogravimetry-mass spectrometry (TG-MS), wide range transmission spectroscopy

and 1H-NMR spectrometry used for characterizing the samples. The samples, reference

samples and starting materials were investigated by Raman spectroscopy to determine

the D/G mode intensity ratio. This quick measurement gives the first evidence that

sidewall functionalization has taken place. TG-MS and 1H-NMR measurements give

quantitative information about the H and n-Bu content of the samples, and TG-MS in

addition about the diameter dependence by the evolution temperature. Quantitative

evaluation of the results obtained by wide range optical spectroscopic measurements

yields information about diameter selectivity within one sample.

3.1 Hydrogenation reactions on HiPco single-

walled carbon nanotubes

On HiPco nanotubes both alkali metal intercalation (by using potassium as inter-

calating agent) and modified Birch reduction were performed in three successive steps

in order to study the difference in reactivity and diameter selectivity.

The as-made samples were characterized by Raman spectroscopy, TG-MS and wide

range optical spectroscopy.

52

1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0

Norm

alized

Rama

n inte

nsity

(a. u.

)

R a m a n s h i f t ( c m - 1 )

D b a n d

G b a n d H i P c o s t a r t i n g m a t e r i a l 1 x 2 x 3 x h y d r o g e n a t e d

Figure 3.1: Raman spectra of HiPco samples hydrogenated by alkali metal (potassium)intercalation. D and G bands are indicated.

1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0

Norm

alized

Rama

n inte

nsity

(a. u.

)

R a m a n s h i f t ( c m - 1 )

D b a n d

G b a n d H i P c o s t a r t i n g m a t e r i a l 1 x 2 x 3 x h y d r o g e n a t e d

Figure 3.2: Raman spectra of HiPco samples hydrogenated by modified Birch reduction.D and G bands are indicated.

53

Figure 3.3: Typical TG-MS curve of HiPco hydrogenated by intercalating potassiuminto the bundles. Black solid lines represent the mass loss (TG curve) and its derivative(DTG curve). Notations: (red circle) hydrogen (m/z 2); (blue triangles) methane (m/z16); (green diamonds) methanol (m/z 31); (magenta squares) m/z 92 toluene.

3.1.1 Raman spectroscopic and TG-MS results

On the Raman spectra of both series of samples (shown in Figures 3.1 and 3.2),

it can be seen that the D/G mode intensity ratio has increased in all hydrogenated

products (0.10-0.13) compared to the starting material (0.08), although the change is

very small, due to the low degree of hydrogenation as detected by TG-MS. However, it

must be emphasized that D/G ratio cannot be used as quantitative estimation of the

sidegroup content due to the complicated origin of the D band [29].

After verifying by Raman spectroscopy that sp3 carbon atoms were formed, the

samples were measured by TG-MS. Typical TG-MS curves for each series of samples

are shown. According to TG-MS results, in both cases, hydrogen evolution starts at

around 400-450 C with a maximum at 500 C (Figures 3.3 and 3.4). The starting

and maximum temperatures show a strict correlation with the tube diameter [74].

Due to increasing stability of the C–H bond on smaller diameter tubes, the increasing

bonding energy leads to higher evolution temperature. The second maximum at 700 C

originates from secondary decomposition products [74].

54

Figure 3.4: Typical TG-MS curve of HiPco hydrogenated by modified Birch reduction.Black solid lines represent the mass loss (TG curve) and its derivative (DTG curve).Notations: (red circle) m/z 2 hydrogen; (blue triangles) m/z 16 methane; (green di-amonds) m/z 29 –CHO; (magenta squares) m/z 31 methanol; (grey circles) m/z 72tetrahydrofuran; (brown triangles) m/z 128 naphthalene.

In Figure 3.4, a shoulder can be observed in the hydrogen peak at 500-550 C as a

maximum of hydrogen evolution from the C–H bonds on the nanotube walls. Then at

650-700 C comes the peak of the secondary decomposition products.

Both synthetic methods provide hydrogenated carbon nanotubes. The relatively

low evolution temperature of hydrogen confirms that they are from C–H bonds on the

sidewalls. The quantity of hydrogen was determined and was found to be 2-4 H/100 C,

so the yield is very small.

No difference in the H content of the samples prepared by the two methods

was found. This is quite surprising, since the modified Birch reduction is a quasi-

homogeneous liquid-phase reaction, while the intercalation-driven reaction is a solid-

phase reaction on nanotube bundles. When nanotubes, potassium and naphthalene are

present in THF at room temperature, the first and very quick process taking place is

the electron transfer from potassium to naphthalene. We believe that there is no (or

very little) direct electron transfer from K to nanotubes, since we did not observe any

color change from black to gold. The electron transfer leads to the formation of naph-

thalenide radical anions indicated by the changing color of the solution from colorless

55

to green. This step corresponds to the dissolution of alkali metals in liquid ammonia

at -78 C in the classical Birch reduction. The next step in the classical route is the

formation of carbanion from the aromatic molecule desired to react. In our case, the

nanotubes are reduced in a quasi-equilibrium process by naphthalenide radical anions

and form negatively charged nanotubes (nanotubide anions) [85, 86]. This step results

in more or less homogeneously dispersed nanotubes due to Coulomb repulsion between

the negatively charged nanotubides. These dispersed species react with methanol to

produce hydrogenated nanotubes.

The quasi-equilibrium process here means that naphthalene is added in excess to

nanotubes, and so potassium in excess to naphthalene. So as naphthalenide radical

anions transfer their electrons to nanotubes, they re-form by involving oxidation of

new potassium atoms. With time, the whole process leads to dispersed nanotubide

anions, more or less individual; at least, bundles get loosened. Thus, it is expected that

many more individual nanotubes are freely available for methanol molecules to become

hydrogenated. So in this case we can expect that simple energetic considerations are

enough to explain diameter selectivity.

3.1.2 Optical spectroscopic results

In the investigation of diameter selectivity of addition reactions on carbon nano-

tubes, the key tool is wide range optical spectroscopy, since exact, quasi-quantitative

conclusions can be drawn from the special features of the spectra [87–89].

The most precise procedure to extract changes from solid-state spectra is to compare

the optical conductivity. This quantity is additive when several independent processes

are involved, like light absorption by different nanotubes, and its calculation takes

into account the reflectance at the interfaces, which can heavily influence the optical

density calculated from transmittance [84]. Contrary to Raman spectroscopy, there is

no resonance process which prefers certain nanotubes over others with selective increase

in the scattering intensity.

56

The energy of interband transitions scales inversely with tube diameter. These ener-

gies are very close to each other if the tube diameters are so, which causes an enlarged

width of the peaks. Upon addition reactions, sp3 C atoms are formed in the nano-

tube sidewall, which decreases the intensity of interband transitions, because a smaller

number of electrons will be involved in the excitation process. The change in the peak

shape of a functionalized sample compared to its starting material and reference sample

reflects the relative amount of electrons localized on sp3 orbitals in the nanotubes with

different diameters.

The results of wide range optical transmission measurements on samples prepared

by modified Birch reduction are shown in Figure 3.5. They confirm the above sketched

picture of quasi-homogeneous reaction on loosened bundles (Section 3.1.1), since they

show the expected diameter selectivity (described in Section 1.2.1): higher reactivity of

smaller diameter tubes (which have larger curvature, so larger structural strain, higher

reactivity towards addition reactions on through sp3 carbon atoms can be formed).

Figure 3.5a represents the Drude-Lorentz fit of calculated optical conductivity spec-

tra from wide range transmission data. The single groups of peaks are related to the

single types of transitions as noted in the figure. They have been obtained by subtract-

ing all other groups of peaks as background (as shown in Figure 2.5). By observing the

S11 peak shown in Figure 3.5b, it can be clearly seen that the high-wavenumber part

of the peaks shows larger decrease in intensity than the low-wavenumber part. This

means that nanotubes with higher transition energies (these are the smaller diameter

tubes) have more electrons removed from the sp2 electronic system and are localized

in a sidewall C–H bond on an sp3 orbital.

As a conclusion, it can be stated that there is practically no difference between the

efficiency of hydrogenation of HiPco single-walled carbon nanotubes either by modified

Birch reduction in liquid phase, or by intercalation of potassium into the bundles in

solid phase, although the degree is very small (as detected by TG-MS), 2-4 H/100 C.

57

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 00

2 0 0

4 0 0

6 0 0

8 0 0

1 0 0 0

1 2 0 0

1 4 0 0

Optica

l cond

uctivt

iy (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

M 0 0

S 1 1

S 2 2

S 3 3 + M 1 1

a )

6 0 0 0 6 5 0 0 7 0 0 0 7 5 0 0 8 0 0 0 8 5 0 07 0 0

8 0 0

9 0 0

1 0 0 0

1 1 0 0

Optica

l cond

uctivt

iy (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

b )

Figure 3.5: a) Drude-Lorentz fit of calculated optical conductivity spectra from widerange transmission measurements of HiPco nanotubes hydrogenated by modified Birchreduction. The groups of peaks related to the single transitions are indicated. Notations:(black line) 1x; (red line) 2x; (green line) 3x hydrogenated sample. b) zoom in to S11

transitions.

58

On the other hand, by looking at the optical conductivity spectra of samples hydro-

genated by intercalating potassium shown in Figure 3.6, one can observe the difference

between the S11 peaks compared to those in Figure 3.5.

In this case, the lower wavenumber part of the S11 peaks decreases more with

the degree of hydrogenation. This surprising dissimilar behaviour of S11 peaks can be

explained by the different reaction mechanism and driving forces. According to the

detailed work of Kukovecz et. al [68], the rate of potassium intercalation into nanotube

bundles strongly depends on the size of the intertube channels. Since a tight bundle

of nanotubes consists of tubes with similar diameters nestled tightly parallel with each

other, almost as much ordered as in a crystal, the size of these interstitial channels

can be well estimated from the tube diameter (as described in Section 3.3.1), and they

become larger with increasing tube diameter. The larger interstitial channels infer less

tight bundles, which eases the intercalation of potassium cations.

The decreasing trend of S22 transition peaks at both series of samples is apparent.

In the case of S33 and M11 peaks we cannot make such a clear statement, since their

energies are very close to each other and mixed very much. However, in the case of M00

(Drude) peaks at zero wavenumber, the changes in intensity seem to be independent

of functionalization rate. This is because the intensity of this peak is connected to

(beside the rate of covalent functionalization of metallic tubes) the amount of free

charge carriers [58, 90]. Exposing carbon nanotube thin films to air (specifically to

oxygen), weak p-doping occurs. This increases the concentration of free charge carriers

as a function of exposure time and causes also an increase in M00 peak intensity at

wavenumbers below 400 cm−1. This increase, and the intensity decrease from covalent

functionalization of metallic tubes are added, and after all, since we did not take care

of exposure time in our experiments, they make this peak not representative to either

case.

59

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 00

2 0 0

4 0 0

6 0 0

8 0 0

1 0 0 0

1 2 0 0

1 4 0 0

1 6 0 0

S 2 2

S 3 3 + M 1 1

M 0 0

Op

tical co

nduc

tivity

(Ω-1 cm

-1 )

W a v e n u m b e r ( c m - 1 )

S 1 1 a )

5 0 0 0 5 5 0 0 6 0 0 0 6 5 0 0 7 0 0 0 7 5 0 0 8 0 0 02 0 0

4 0 0

6 0 0

8 0 0

1 0 0 0

1 2 0 0

1 4 0 0

1 6 0 0

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

b )

Figure 3.6: a) Drude-Lorentz fit of calculated optical conductivity spectra from widerange transmission measurements of HiPco nanotubes hydrogenated by alkali metal(potassium) intercalation. The groups of peaks related to the single transitions areindicated. Notations: (black line) 1x; (red line) 2x; (green line) 3x hydrogenated sample.b) zoom in to S11 transitions.

60

Figure 3.7: Typical TG-MS curve of HiPco n-butylated by intercalating potassium intothe bundles. Black solid lines represent the mass loss (TG curve) and its derivative(DTG curve). Notations: (red circle) m/z 2 hydrogen; (blue triangles) m/z 18 water;(green squares) m/z 43 C3H7; (magenta triangles) m/z 92 toluene.

3.2 Hydrogenation and n-butylation of HiPco

single-walled carbon nanotubes

Both types of reductive addition reactions were applied to HiPco nanotubes to

investigate the selectivity towards addition of n-butyl groups (using 1-iodobutane as

reagent).

TG-MS measurements shown in Figures 3.7 and 3.8 demonstrate that hydrogenation

takes place in both types of reactions beside n-butylation. The calculated amount of

n-butyl groups in both types of reactions was less than 1 n-Bu/100 C, the hydrogen

content was 1-3 H/100 C. The detachment of n-Bu groups starts around 300 C, at

lower temperature than hydrogen (400-450 C) in agreement with the previous work of

Borondics et. al [74]. This means that the C–Bu bonds are less stable than C–H bonds

in the sidewall.

61

Figure 3.8: Typical TG-MS curve of HiPco n-butylated by modified Birch reduction.Black solid lines represent the mass loss (TG curve) and its derivative (DTG curve). No-tations: (red circle) m/z 2 hydrogen; (blue triangles) m/z 18 water; (magenta squares)m/z 43 C3H7; (turquoise triangles) m/z 44 CO2; (green diamonds) m/z 128 naphtha-lene.

Since pure n-butylated samples could not be synthesized, furthermore, n-butylation

seemed to be only a side reaction under these circumstances, further attention was paid

on studying hydrogenation reactions on nanotube bundles.

3.3 Hydrogenation of nanotube bundles by alkali

metal intercalation

As described in Section 3.1, the diameter distribution of HiPco nanotubes is exactly

in the region where a small lattice expansion is needed to intercalate potassium cations.

The tendency is so well-seen in the optical spectra of hydrogenated HiPco in Figure 3.6

that we think, in this diameter range, the process that determines the selectivity is

purely the intercalation. Its rate is roughly determined by the size of the interstitial

channels.

To investigate the diameter selectivity of intercalation processes, besides K/HiPco

combination with 0.256 diameter ratio, two other combinations with extreme diameter

62

Nanotube Mean diameter (nm) Dopant Ionic radius (nm) dalkali metal/dSWNT

P2 1.60 K 0.138 0.173

HiPco 1.08 K 0.138 0.256

CoMoCat 0.90 Rb 0.152 0.338

Table 3.1: Alkali metal cation/nanotube diameter ratios used for investigating the effectof bundling.

1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0

Norm

alized

Rama

n inte

nsity

(a. u.

)

R a m a n s h i f t ( c m - 1 )

P 2 s t a r t i n g m a t e r i a l 1 x 2 x 3 x h y d r o g e n a t e d

G b a n d

D b a n d

Figure 3.9: Raman spectra of P2 samples hydrogenated by potassium intercalation. Dand G bands are indicated.

ratios: a small one (K/P2 with 0.173) and a large one (Rb/CoMoCat with 0.338) (as

given in Table 3.1) were used.

The Raman spectra recorded on both K/P2 and Rb/CoMoCat hydrogenated

SWNT samples and shown in Figures 3.9 and 3.10, represent the same trend as all

the other functionalized samples: an increase in the D/G mode intensity ratio upon

functionalization indicating the formation of new sp3 C atoms in the sidewalls.

TG-MS measurements on the K/P2 system (a typical curve is shown in Figure 3.11)

also verify the presence of hydrogen bound to the nanotube sidewalls. The difference

compared to the HiPco samples is the temperature of hydrogen evolution. It starts at

the temperature about 50 C lower (starting at 300 C with peak at 350 C) than for

63

1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0

Norm

alized

Rama

n inte

nsity

(a. u.

)

C o M o C a t s t a r t i n g m a t e r i a l 1 x 2 x 3 x h y d r o g e n a t e d

R a m a n s h i f t ( c m - 1 )

D b a n d

G b a n d

Figure 3.10: Raman spectra of CoMoCat samples hydrogenated by rubidium interca-lation. D and G bands are indicated.

HiPco. It is explained by the larger mean diameter of P2 tubes, which decreases the

stability of sidewall C-H bonds.

TG-MS curves of CoMoCat samples (Figure 3.12), however, do not show a distinct

peak of hydrogen evolution. Since CoMoCat tubes have the smallest diameter, so the

largest curvature among the studied three types of nanotubes, the simple energetic

expectations would dictate the highest degree of hydrogenation.

Although Raman spectra of CoMoCat samples show an obvious increase in the D/G

mode intensity ratio (Figure 3.10), TG-MS and 1H-NMR results prove only the presence

of not more than 1 H/100 C, which is the detection limit of TG-MS. Thus, on the TG-

MS curves of CoMoCat samples, there is no distinct hydrogen evolution peak above

400-450 C, as would be expected from the hydrogen evolution temperature measured

in HiPco. The first distinct peak is far above 700 C, which is unambiguously related

to secondary decomposition products, since even at so small diameters that CoMoCat

samples represent, the detachment of functional groups ends at about 500 C [91].

The quantitative evaluation of TG-MS and 1H-NMR results revealed that the overall

H content of the P2 samples is a little bit higher than HiPco’s. Results on all samples

64

Figure 3.11: Typical TG-MS curve of P2 hydrogenated by potassium intercalation.Black solid lines represent the mass loss (TG curve) and its derivative (DTG curve).Notations: (red circle) m/z 2 hydrogen; (blue triangles) m/z 16 methane; (green dia-monds) m/z 31 methanol; (magenta squares) m/z 92 toluene.

Figure 3.12: Typical TG-MS curve of CoMoCat hydrogenated by rubidium interca-lation. Black solid lines represent the mass loss (TG curve) and its derivative (DTGcurve). Notations: (red circle) m/z 2 hydrogen; (blue triangles) m/z 28 CO; (greendiamonds) m/z 44 CO2; (magenta squares) m/z 92 toluene.

65

Sample Tmax [H] (C) H /100C (TG-MS) [H] (mmol/g) H /100C

(NMR) (NMR)

Sample number

1 2 3

HiPco 480 1.90 4.31 3.21 – –

P2 350 2.17 2.41 3.61 4.63±0.27 5.6

CoMoCat – <1 <1 <1 2.39±0.15 2.9

Table 3.2: H content and evolution temperature of the samples. 1H-NMR data arerelated to the samples with the highest degree of hydrogenation measured by TG-MS. H content determined by NMR are corrected with the H content of the referencesamples.

hydrogenated by alkali metal intercalation are summarized in Table 3.2. 1H-NMR data

on HiPco samples are not trustworthy because of the small amount of the measured

samples, therefore they are not included in Table 3.2. The general shape and data

evaluation of the NMR spectra are shown in Figure 3.13.

For the evaluation of the NMR spectra we used the Curie-Langevin law for the

nuclear magnetism, which says that the net magnetism of a macroscopic sample (the

NMR intensity extrapolated to t=0) is proportional to the amount of that type of

nuclei in the sample.

Optical conductivity spectra of K/P2 and Rb/CoMoCat samples are shown in Fig-

ures 3.14 and 3.15, respectively. Differences can be observed comparing the spectra

either with HiPco or with each other.

In the K/P2 system, the intensity loss is independent of wavenumber, at

Rb/CoMoCat it shows an increase with increasing wavenumber. Comparing the three

systems’ optical conductivity spectra (K/HiPco, K/P2 and Rb/CoMoCat), three differ-

ent behaviors can be observed. A simple possible explanation of the results is discussed

in 3.3.1.

TG-MS measurements, namely, the temperature of hydrogen evolution also shows

the difference in mean diameters between P2 and HiPco. As the mean diameter of a

sample increases, the temperature of hydrogen evolution decreases.

66

0 5 0 1 0 0 1 5 0 2 0 0 2 5 0- 1 0 0

0

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

7 0 0

1 H NMR

inten

sity (a

. u.)

T i m e ( s )

e m p t y T e f l o n c a p s u l e a d a m a n t a n e

a )

0 5 0 1 0 0 1 5 0 2 0 0 2 5 0- 1 0 0

0

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

7 0 0

1 H NMR

inten

sity (a

. u.)

T i m e ( s )

c o r r e c t e d i n t e n s i t y o f a d a m a n t a n e f i t t e d a n a l y t i c f u n c t i o n

b )

0 1 0 2 0 3 0 4 0 5 0

0

1

2

3

4

5

6

1 H NMR

inten

sity (a

. u.)

T i m e ( s )

C o M o C a t r e f e r e n c e s a m p l e e m p t y T e f l o n c a p s u l e c o r r e c t e d s a m p l e i n t e n s i t y f i t t e d a n a l y t i c f u n c t i o n

c )

Figure 3.13: a) the signal of the empty Teflon capsule as sample holder (highlightedwith green) and the adamantane. b) the signal of the adamantane corrected with thesignal of the sample holder and the fitted analytic function. c) signals and fit relatedto the CoMoCat reference sample.

67

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 00

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

S 3 3 + M 2 2

M 1 1

S 1 1

S 2 2

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

M 0 0 a )

8 0 0 0 9 0 0 0 1 0 0 0 0 1 1 0 0 0 1 2 0 0 05 0

1 0 0

1 5 0

2 0 0

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

b )

Figure 3.14: a) Drude-Lorentz fit of calculated optical conductivity spectra from widerange transmission measurements of P2 nanotubes hydrogenated by alkali metal (potas-sium) intercalation. The groups of peaks related to the single transitions are indicated.Notations: (gray line) reference; (black line) 1x; (red line) 2x; (green line) 3x hydro-genated sample. b) zoom in to S22 transitions.

68

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0 3 0 0 0 0 3 5 0 0 00

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0

M 1 1 + S 3 3

S 2 2

S 1 1

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

M 0 0 a )

1 2 0 0 0 1 4 0 0 0 1 6 0 0 0 1 8 0 0 02 5

5 0

7 5

1 0 0

1 2 5

1 5 0

Optica

l cond

uctivi

ty (Ω

-1 cm-1 )

W a v e n u m b e r ( c m - 1 )

b )

Figure 3.15: a) Drude-Lorentz fit of calculated optical conductivity spectra from widerange transmission measurements of CoMoCat nanotubes hydrogenated by alkali metal(rubidium) intercalation. The groups of peaks related to the single transitions are in-dicated. Notations: (gray line) reference; (black line) 1x; (red line) 2x; (green line) 3xhydrogenated sample. b) zoom in to S22 transitions.

69

In summary, by using three different single-walled carbon nanotubes (P2, HiPco,

CoMoCat) and two alkali metals (K, Rb), the selectivity of hydrogenation by inter-

calation could be investigated in a wide diameter range and at three different ionic

radius/nanotube diameter ratios by optical spectroscopy.

Besides, for all reactions, the effect of using three successive steps were studied,

whether the efficiency could be increased.

The degree of hydrogenation of HiPco nanotubes by both reactions was measured

to be 2-4 H/100 C. However, the changes in the characteristic S11 transitions between

the Van Hove singularities located at the near-infrared–visible spectral range showed

different diameter selectivity. For modified Birch reduction, according to spectroscopic

results, the diameter selectivity was ”normal”, curvature-determined (Figure 3.5): the

smaller the diameter, the higher the H content. For hydrogenation by potassium inter-

calation, diameter selectivity was perfectly reversed (Figure 3.6). It is explained by the

mechanism of alkali metal intercalation into nanotube bundles. The diameter range rep-

resented by HiPco is exactly at the range where the intercalation of potassium cations

requires no or only a small lattice expansion [68]. The size of the channels inside the

bundles is proportional to the diameter of the tubes building up the bundle.

At smaller ionic radius/nanotube diameter ratio than that of K/HiPco (K/P2), the

degree of hydrogenation was also proven to be 2-4 H/100 C, but on the average a little

bit higher than at HiPco. However, no diameter selectivity could be detected by optical

spectroscopy (Figure 3.14). In this case, for this large tube diameters, there is no need

for lattice expansion, but it is required for the smaller diameter tubes of P2. Here, the

selectivity determining factors are the lattice expansion together with the effect of the

curvature. These two opposite effects seem to equalize each other in the diameter range

represented by P2.

At larger ionic radius/nanotube diameter ratio (Rb/CoMoCat), the degree of hy-

drogenation is drastically reduced (∼1 H/100 C), and ”normal”diameter selectivity was

detected by optical spectroscopy (Figure 3.15). It can be explained by the block of Rb

intercalation into the narrow channels inside the bundles [67]. Since electron transfer

70

Figure 3.16: The cutaway view of a bundle built up by 19 tubes placed perfectly parallelto each other in the x−y plane (z direction is parallel to the tube axis) with intercalatedalkali cations in the triangular channels (1 chain/channel).

is only possible at the surface of the bundles, the degree will be very small, and the

determining factor will be only the curvature.

The S11 peak of HiPco and the S22 peak of P2 and CoMoCat samples were used

for analyzing the diameter selectivity. When sorting out the appropriate peak to draw

conclusions by, there are various factors and peculiarities to consider: mean diameter,

diameter distribution and the number and of (n,m) tubes. The mean diameter deter-

mines the wavenumbers (so the range of the spectrum) where the S11 and S22 peaks will

be located. If the mean diameter is large, it may infer that the single peaks are close

to each other. The latter depends also on the diameter distribution and the number of

nanotube types. If the former is large and the latter is small, the peak will be quite

wide, but very much structured. In other cases, like P2, where the mean diameter is

large and the diameter range is small, the S11 peak will be narrow and featureless. In

such cases, using S22 peak may become reasonable. In the case of CoMoCat, where

the mean diameter is small, the energy separation of the single transitions is larger.

Here, both S22 and S11 peaks show a rich structure. It also increases the importance

of choosing the right peak that that range of the spectrum will be the reference when

merging the different spectral ranges measured by different instruments on occasion of

bad overlap.

71

When investigating the diameter selectivity of alkali metal intercalation within one

sample, the activation energy of intercalation, the required lattice expansion and other

processes must be considered. In Figure 3.16 the simplest arrangement is shown: alkali

metal cations intercalated in single chains into nanotube bundles with triangular ge-

ometry. When alkali cations are ordered like this, according to [68], there is no need for

lattice expansion above 1.5 nm tube diameter. This case is represented by our K/P2

samples. K/HiPco samples represent the case with relative lattice expansion of 1-7 %.

For Rb intercalation, a lattice expansion over 13 % would be expected even for such

large diameter tubes like 1.4 nm [67].

A simple discussion of the energetics of intercalation is given in Section 3.3.1.

3.3.1 van der Waals interactions in nanotube bundles

Introduction

In the investigation of alkali metal intercalation, it is essential to study the structure

and interactions inside nanotube bundles. The simplest case is to consider a triangular

geometry built up by infinitely long nanotubes with the same chirality (preferably

armchair for simplifying the calculations).

Girifalco et al. studied the potential energies of interaction between graphitic ma-

terials [92]. They used Lennard-Jones carbon-carbon potential (LJ) and assumed con-

tinuous distribution of atoms in tube surface. Since this is very much similar in na-

notubes and graphene, in the following they performed the calculations considering

graphene. They found that by using certain reduced parameters, the potential energies

of graphene-graphene, tube-tube, C60-C60 and tube-C60 in various arrangements can

be plotted on the same curve.

In the following we restrict our considerations to pairs of nanotubes. In the contin-

uum model, the φ(R) potential per unit area of interacting tubes between two identical,

parallel and infinitely long tubes is:

72

φ(R) = n2σ

∫u(x) dΣ1 dΣ2 (3.1)

where nσ is the mean surface density of carbon atoms, x is the distance between two

surface elements dΣ1 and dΣ2 on the different tubes and R is the perpendicular distance

between tube centers. The integrals are independent of tube radius [92].

Girifalco et al. computed the potentials for tubes with diameters between 0.54 and

3.80 nm (armchair tubes in the n range from 4 to 28). As expected, the potential has

a longer range, and the minimum occurs at a higher reduced distance (the distance of

the tubes referred to their diameters) for smaller diameter tubes.

Sun et al. generalized Girifalco’s equation for two not identical tubes [93].

The simplified formula given by Equation 3.3 was used for calculating φ(R) by

introducing the reduced parameter R from [92]:

R =R− ρR0 − ρ

(3.2)

where R0 = 2r+3.13 [A] is the equilibrium distance of the tubes. In our case, defining ρ

as the sum of the radii, ρ = 2r [A], where r is the tube radius, is a good approximation

for ρ.

φ(R) =φ(R)

|φ(R0)|= − 1

0.6

[(3.41

3.13R + 0.28)4 − 0.4(

3.41

3.13R + 0.28

)10]

(3.3)

For calculating |φ(R0)|, Equation 3.4 is a good approximation [92]:

φ0(r) = −0.1135√r + 9.39× 10−3 [eV/A] (3.4)

Results

The potential energy as a function of tube-tube distance was calculated for the

tubes listed in Table 3.1. The calculation was performed for all tube types for the

mean diameter and for the lower and upper limits of diameter distribution. The min-

imum required tube-tube distance was calculated for hosting one K+ or one Rb+ in

73

1 . 0 1 . 2 1 . 4 1 . 6 1 . 8 2 . 0 2 . 2 2 . 4 2 . 6 2 . 8 3 . 0- 4

- 3

- 2

- 1

0

1

2

3

4

5 K +

Poten

tial e

nergy

(eV n

m-1 )

T u b e - t u b e d i s t a n c e ( n m )

P 2 H i P c o C o M o C a t

R b +

Figure 3.17: van der Waals potentials between two identical nanotubes with meandiameters from Table 2.1. The minimum tube-tube distance for hosting one K+ andone Rb+ between three tubes is indicated.

the triangular lattice by using simple geometric considerations. For K+ the minimum

distance is 1.784 nm, for Rb+ 1.965 nm.

Figure 3.17 shows the van der Waals potential curves according to Equation 3.3

together with the use of Equation 3.4 for tubes with mean diameter from Table 2.1.

It can be seen that only mean diameter P2 tubes, placed as in Figure 3.16, require no

lattice expansion for hosting one K+ between three tubes.

Figures 3.18, 3.19 and 3.20 show the van der Waals potential energies for each

nanotube samples for their mean diameters and their upper and lower diameter limits.

Presumed states of alkali metal intercalation is shown in Figure 3.21. In the first

step, alkali metal atoms condensate onto the surface of a bundle. Then they reduce

the nanotubes on the surface. The so-formed alkali cations migrate into the interior of

the bundle at the ends of the tubes into the triangular channels. In this model, the

activation energy that must be invested is treble of the lattice expansion energy. Latter

is the potential energy difference between the intercalated and initial states.

Figures 3.22 and 3.23 show the required lattice expansion and the lattice expansion

energy for accomodation a cation in the triangular channel, respectively.

74

1 . 4 1 . 6 1 . 8 2 . 0 2 . 2 2 . 4 2 . 6 2 . 8 3 . 0- 4

- 3

- 2

- 1

0

1

2

3

4

5 K +

Poten

tial e

nergy

(eV n

m-1 )

T u b e - t u b e d i s t a n c e ( n m )

1 . 7 1 . 6 1 . 2

Figure 3.18: van der Waals potentials between two identical nanotubes for P2 tubeswith diameters from Table 2.1. The minimum tube-tube distance for hosting one K+

between three tubes is indicated.

1 . 0 1 . 2 1 . 4 1 . 6 1 . 8 2 . 0 2 . 2

- 3

- 2

- 1

0

1

2

3

4K +

Poten

tial e

nergy

(eV n

m-1 )

T u b e - t u b e d i s t a n c e ( n m )

1 . 3 1 . 0 8 0 . 8

Figure 3.19: van der Waals potentials between two identical nanotubes for HiPco tubeswith diameters from Table 2.1. The minimum tube-tube distance for hosting one K+

between three tubes is indicated.

75

0 . 8 1 . 0 1 . 2 1 . 4 1 . 6 1 . 8 2 . 0 2 . 2 2 . 4

- 3

- 2

- 1

0

1

2

3

4 R b +

Poten

tial e

nergy

(eV n

m-1 )

T u b e - t u b e d i s t a n c e ( n m )

1 . 1 7 0 . 9 0 0 . 5 7

Figure 3.20: van der Waals potentials between two identical nanotubes for CoMoCattubes with diameters from Table 2.1. The minimum tube-tube distance for hosting oneRb+ between three tubes is indicated.

Figure 3.21: Presumed initial, intermediate and final states of alkali metal intercalation.

76

0 . 5 1 . 0 1 . 5 2 . 0 2 . 5

0 . 0 00 . 0 20 . 0 40 . 0 60 . 0 80 . 1 00 . 1 20 . 1 40 . 1 60 . 1 8

Requ

ired l

attice

expa

nsion

(nm)

T u b e d i a m e t e r ( n m )

K +

R b +

P 2 H i P c o C o M o C a t

Figure 3.22: Lattice expansion values for all tube diameters from Table 2.1 to accomo-date the respective alkali cation.

0 . 5 1 . 0 1 . 5 2 . 0 2 . 5

0 . 0

0 . 2

0 . 4

0 . 6

0 . 8

1 . 0

1 . 2

Lattic

e exp

ansio

n ene

rgy (e

V nm-1 )

T u b e d i a m e t e r ( n m )

K +

R b +

P 2 H i P c o C o M o C a t

Figure 3.23: Lattice expansion energies for all tube diameters from Table 2.1 for acco-modating K+ and for CoMoCat/Rb+.

77

From these results we can conclude that the required lattice expansion energy for

K+ has a maximum at around 0.6 nm tube diameter. This value is a bit lower than

Kukovecz et al. suggested (0.9-1.2 nm) [68].

Below a certain limit (which is the lattice parameter required to accomodate one

K+ or one Rb+) the required lattice expansion is zero, above this limit it varies linearly

with tube diameter. The lattice expansion energies, however, has a maximum value.

This means that K+ intercalation becomes more and more energetically favorable for

smaller tube diameters than ∼0.6 nm.

This simple model can explain the observed anomaly in diameter selectivity for all

tube samples. However, in the case of CoMoCat, this model cannot explain the little

degree of hydrogenation compared to HiPco. The stability and the lattice expansion

energy considerations would lead to a higher degree of hydrogenation compared to

HiPco. Since we observed anomalously little hydrogen content in all three parallel

experiments, the explanation must be founded on another factor then energetics.

Possible explanations can be the kinetical hindrance of Rb intercalation caused by

its larger size compared to K, thus, electron transfer is also only possible on the surface,

which means that only a limited number of tubes are involved in the reaction. Another

possible reason can be that considering an average size bundle built up by smaller

diameter tubes has smaller surface, which can be coated only smaller number of the

larger Rb atoms, or the change in the numerical value of energetics of the processes

(heat of evaporation, heat of adsorption, ionization energy etc.). Anyhow, this anomaly

requires further investigations.

3.4 Conclusions and summary

Reductive addition reactions of single-walled carbon nanotubes representing

a wide diameter range were applied and investigated by Raman spectroscopy,

thermogravimetry-mass spectrometry, 1H-NMR spectrometry and wide range optical

spectroscopy.

78

To compare efficiency, diameter and other selectivity and reactivity of the nanotu-

bes, two different reaction mechanisms and functional groups (modified Birch reduc-

tion and alkali metal intercalation; H and n-Bu) were used. The main difference in the

applied mechanisms is that modified Birch reduction is a quasi-homogeneous, solution-

phased reaction, while alkali metal intercalation followed by addition of electrophilic

reagents is a solid-phase reaction. In the case of modified Birch reduction, the real

reactants are the more or less individual nanotubide anions, in the case of alkali metal

intercalation, the intercalated nanotube bundles.

At the n-butylation reaction of HiPco nanotubes, hydrogenation as a side reaction

was expected. On the contrary, at both mechanism, hydrogenated nanotubes were the

main products, and n-butylation seemed to be the side reaction. This lets us conclude

that hydrogenation is far more fast and/or favorable than n-butylation.

The degree of functionalization could be increased in all reactions by successive

steps. This can be explained by the bundling of the nanotubes. Even in the case of

modified Birch reduction we cannot speak about a homogeneous reaction with individ-

ually dispersed nanotubes, albeit the size of the bundles is much smaller and they are

much loosened due to the repulsion of the negatively charged nanotubes, and due to

the possibility for these smaller bundles to repel from each other in the liquid phase.

Alkali metal intercalation followed by addition of electrophilic reagents is a solid-

phase reaction, and even when toluene is added as a ”medium” for the reaction, it

remains a solid-phase one, since the negatively charged nanotube bundles with alkali

metal cations in their interstitial channels (or on their surfaces) will not be dispersed

in the apolar toluene.

As proven by optical spectroscopy, the diameter selectivity of hydrogenation by

using two different reaction mechanisms is different. The difference is explained by the

tendency of nanotubes towards bundling. Since bundling can limit the accessibility of

the individual tubes inside, this effect also must be taken into account besides the

simple increasing diameter-decreasing reactivity considerations.

79

Anyhow, by applying successive steps, each step seems to increase the loosening of

the bundles making new, formerly intact tubes available for the reactants, proven by

the increasing degree of functionalization in each case.

Investigating further the hydrogenation by alkali metal intercalation by using differ-

ent starting SWNT, with three different ratios of ionic radii of alkali metal cations and

mean diameters of nanotubes, three different diameter selectivities could be detected.

It was demonstrated that the selectivity of this type of reductive hydrogenation is very

sensitive to energetics of the competing processes. With these three starting nanotubes

all the possible cases could be represented.

80

Acknowledgement

First of all, I would like to thank my supervisor, Katalin Kamaras for supporting me

and assuming the guidance of my PhD work. She showed great patience to me all the

time from my first steps. She involved me in lots of projects and collaborations, which

I found really great and interesting. I am also grateful for the useful and important

scientific and social skills I learnt from her.

I am truly grateful to Sandor Pekker for his extensive help with my work, and for

the inspiring discussions.

I am also grateful to Eva Kovats and Ferenc Borondics for their useful advice and

help with the laboratory work.

I would like to thank Emma Jakab for the essential thermogravimetry-mass spec-

trometric measurements on my samples.

Warm thanks are due to Aron Pekker, Peter Nemes-Incze and Hajnalka Maria

Tohati for their help with the transmission spectroscopic and AFM measurements.

I thank Monika Bokor, Tamas Verebelyi and Kalman Tompa for the NMR mea-

surements on my samples, and for the useful discussions.

Thanks to all my colleagues: Bea Botka, Balint Korbuly, Eva Kovats, Peter Matus,

Aron Pekker, Gyongyi Pergerne Klupp, Laszlo Ratkai, Zsolt Szekrenyes and Hajnalka

Maria Tohati for creating a friendly, cheerful and supporting milieu in the lab and in

the office. The amount of chocolate and coffee had together speaks for itself.

I am grateful to Laszlo Kovacs and Miklos Veres for allowing me to use the UV-Vis

and Raman spectrometers, respectively.

81

I am grateful to the Institute for Solid State Physics and Optics, Wigner Research

Centre for Physics for the opportunity to work among its walls.

Last but not least, I am truly grateful to my family and friends for their love,

support and encouragement all along the years. Special thanks are concerned to my

husband, Laszlo for his mental support, for the solid background of which this work

would not have been possible without and for his incontestable care and love.

82

Theses

1. I synthesized hydrogenated HiPco nanotubes in three successive steps by modi-

fied Birch reduction and potassium intercalation followed by addition of methanol.

According to thermogravimetry-mass spectrometric measurements, the H-content

was 2-4 H/100 C in both cases. However, I detected by optical spectroscopy that

the diameter selectivity of these reactions was different: in the case of modified

Birch reduction the smaller diameter tubes, in the case of potassium intercalation

the larger diameter tubes were more reactive. I explained this discrepancy by the

different reaction mechanisms: at modified Birch reduction the individual nano-

tubes, at potassium intercalation the nanotube bundles are the primer reactants.

In the latter case, the determinative factor is the accessibility of the individual

nanotubes inside the bundles. The incorporation of potassium cations is easier

into bundles of larger diameter tubes.

2. I prepared hydrogenated and n-butylated HiPco single-walled carbon nanotubes

in three successive steps by both modified Birch reduction and potassium inter-

calation. According to thermogravimetry-mass spectrometry measurements the

n-butyl content of the samples was ∼1 n-Bu/100 C, the hydrogen content was

2-3 H/100 C in spite of lack of addition of methanol to the system. I explained

this anomaly by the presence of other protic H sources (water, oxide, hydroxide)

and that the hydrogenation can be faster and/or favorable than n-butylation.

3. I investigated the diameter selectivity of hydrogenation by alkali metal intercala-

tion on single-walled carbon nanotubes by using three different mean diameters

and diameter distributions (P2, HiPco, CoMoCat) and two alkali metals with

different ionic radii (K, Rb) on a wide diameter range and at three different ionic

radius/nanotube diameter ratios. Thermogravimetry-mass spectrometric and 1H-

83

NMR results revealed that the degree of hydrogenation is 2-4 H/100 C in the case

of small (K/P2) and medium (K/HiPco) ratios, and at around 1 H/100 C at large

(Rb/CoMoCat) ratio.

4. I investigated the diameter selectivity of hydrogenation of single-walled carbon

nanotubes by alkali metal intercalation followed by addition of methanol on a

wide diameter range. By optical spectroscopy I detected no selectivity at small

ionic radius/nanotube diameter ratio, reversed selectivity at medium ratio and

normal selectivity at large ratio. I explained these results by the activation energy

of alkali metal intercalation into nanotube bundles, which depends on nanotube

diameter and ionic radii.

5. I prepared hydrogenated and n-butylated HiPco single-walled carbon nanotubes

by modified Birch reduction and potassium intercalation in three successive steps,

respectively. It was demonstrated by thermogravimetry-mass spectrometry that

the degree of functionalization could be increased at both types of reactions by

applying successive steps. I explained this by the step-by-step loosening of the

nanotube bundles. This even affects in the homogeneous phase modified Birch

reduction.

84

Tezispontok

1. Hidrogenezett

HiPco nanocsoveket allıtottam elo harom lepesben modosıtott Birch-redukcioval

es kalium interkalacioval. Termogravimetria-tomegspektometriai meresek szerint

a H-tartalom 2-4 H/100 C-nek adodott mindket reakcio eseten. Optikai spek-

troszkopia segıtsegevel azonban kimutattam, hogy az atmero-szelektivitas kulon-

bozo: modosıtott Birch-redukcio eseten a kisebb atmeroju csovek, a kalium in-

terkalacio eseten a nagyobb atmerojuek bizonyultak reaktıvabbaknak. Az elterest

a kulonbozo reakcio-mechanizmussal magyaraztam: modosıtott Birch-redukcio es-

eten az egyedi nanocsovek reaktivitasa a donto, kalium interkalacional az egyedi

nanocsovek hozzaferhetosege a kotegeken belul a beepulo K-ionok szamara. A

K-ionok beepulese konnyebb a nagyobb atmeroju csovek kotegeibe.

2. Hidrogenezett es n-butilozott HiPco nanocsoveket allıtottam elo harom lepes-

ben modosıtott Birch-redukcioval es kalium interkalacioval. Termogravimetria-

tomegspektrometriai meresek szerint a mintakban ∼1 n-Bu/100 C n-butil-

tartalom, es 2-3 H/100 C hidrogen volt annak ellenere, hogy hidrogenezo szert

nem adtam a rendszerhez. Ezt azzal magyaraztam, hogy a hidrogenezodest a je-

lenlevo protikus H-tartalmu szennyezok okozzak (vız, oxidok, hidroxidok), es a

hidrogenezodes gyorsabb es/vagy kedvezmenyezettebb lehet, mint a butilozas.

3. Alkalifem interkalacios hidrogenezes kitermeleset tanulmanyoztam egyfalu szen

nanocsoveken szeles atmero-tartomanyban, harom kulonbozo atlagos atmeroju

nanocso (P2, HiPco, CoMoCat) es ket kulonbozo ionatmeroju alkalifem (K,

Rb) felhasznalasaval. Termogravimetria-tomegspektrometria es 1H-NMR meresek

segıtsegevel kimutattam, hogy a hidrogen-tartalom 2-4 H/100 C kis es kozepes

85

ionatmero/nanocso-atmero aranyonal (K/P2 es K/HiPco), es <1 H/100 C nagy

ionatmero/nanocso-atmero aranynal (Rb/CoMoCat).

4. Alkalifem interkalacios hidrogenezes atmero-szelektivitasat tanulmanyoztam egy-

falu szen nanocsoveken szeles atmero-tartomanyban, harom kulonbozo atlagos at-

meroju nanocso (P2, HiPco, CoMoCat) es ket kulonbozo ionatmeroju alkalifem

(K, Rb) felhasznalasaval. Optikai spektroszkopia segıtsegevel kimutattam, hogy

kis ionatmero/nanocso-atmero aranynal nem mutathato ki atmero-szelektivitas,

kozepes aranynal a nagyobb atmeroju, nagy aranynal a kis atmeroju nanocso-

vek a reaktıvabbak. Ezt az interkalacio nanocso- es ionatmero-fuggo aktivalasi

energiajaval magyaraztam.

5. Modosıtott Birch-redukcioval, illetve kalium interkalacioval hidrogenezett,

valamint n-butilozott HiPco nanocsoveket allıtottam elo. Harom egymast

koveto lepesben elvegezve a reakciokat, termograviemtria-tomegspektromertias

meresekkel kimutattam, hogy a kitermeles mindket reakciotıpusnal es mind-

ket funcios csoportnal novelheto egymast koveto lepesek alkalmazasaval. Ezt a

nanocso-kotegek lepesenkenti fokozatos fellazulasaval magyaraztam, amely sz-

erepet jatszik meg a homogen fazisu modosıtott Birch-redukcional is.

86

List of publications

Publications related to the thesis:

1. K. Nemeth, A. Pekker, F. Borondics, E. Jakab, N. M. Nemes, K. Kamaras andS. Pekker. Investigation of hydrogenated HiPCo nanotubes by infrared spec-troscopy. Phys. Status Solidi B 247:2855, 2010.

2. K. Nemeth, E. Jakab, F. Borondics, H. M. Tohati, A. Pekker, M. Bokor, T. Vere-belyi, K. Tompa, S. Pekker and K. Kamaras. Breakdown of diameter selectivityin a reductive hydrogenation reaction of single-walled carbon nanotubes. Chem.Phys. Lett., 618:214, 2015.

Other publications:

1. G. Inzelt, Z. Puskas, K. Nemeth and I. Varga. Electrochemically induced trans-formation of ruthenium(III) trichloride microcrystals in salt solutions. J SolidState Electrochem., 9:823, 2005.

2. G. Inzelt, K. Nemeth and A. Roka. Electrochemical quartz crystal microbalancestudy of redox transformations of TCNQ microcrystals in concentrated LiCl so-lutions. Electrochimica Acta, 52:4015, 2007.

3. H.-M. Tohati, B. Botka, K. Nemeth, A. Pekker, R. Hackl and K. Kamaras. In-frared and Raman investigation of carbon nanotube-polyallyamine hybrid sys-tems. Phys. Status Solidi B, 247:2884, 2010.

4. E. A. Francis, S. Scharinger, K. Nemeth, K. Kamaras and C. A. Kuntscher. In-vestigation of the Jahn–Teller effect in the C−60 monoanion under high pressure.Phys. Status Solidi B, 247:3047, 2010.

5. E. A. Francis, S. Scharinger, K. Nemeth, K. Kamaras and C. A. Kuntscher.Pressure-induced transition from the dynamic to static Jahn–Teller effect in(Ph4P)2IC60. Phys. Rev. B, 85:195428, 2012.

6. B. Botka, M. E. Fustos, H. M. Tohati, K. Nemeth, G. Klupp, Z. Szekrenyes,D. Kocsis, M. Utczas, E. Szekely, T. Vaczi, G. Tarczay, R. Hackl, T. W. Cham-berlain, A. N. Khlobystov and K. Kamaras. Interactions and chemical transfor-mations of coronene inside and outside carbon nanotubes. Small, 10:1369, 2014.

7. C. Muller, K. Nemeth, S. Vesztergom, T. Pajkossy and T. Jacob. The in-terface between HOPG and 1-butyl-3-methyl-imidazolium hexafluorophosphate.Phys. Chem. Chem. Phys., submitted.

87

Posters:

1. K. Nemeth, F. Borondics, E. Jakab, A. Pekker, K. Kamaras, S. Pekker:Sidewall functionalization of HiPCo nanotubes in tolueneIWEPNM, Kirchberg in Tirol, Austria (2008)

2. K. Nemeth, F. Borondics, E. Jakab, A. Pekker, K. Kamaras, S. Pekker:Reductive functionalization of HiPCo nanotubesSIWAN, Szeged, Hungary (2008)

3. K. Nemeth, A. Pekker, F. Borondics, K. Kamaras, S. Pekker:Infrared and Raman spectra of hydrogenated HiPCo nanotubesIWEPNM, Kirchberg in Tirol, Austria (2010)

4. K. Nemeth, A. Pekker, F. Borondics, K. Kamaras, S. Pekker:Sidewall functionalization of HiPCo single-walled carbon nanotubesFISS, Krutyn, Poland (2010)

5. K. Nemeth, A. Pekker, F. Borondics, K. Kamaras, S. Pekker:Investigation of diameter selectivity of reductive hydrogenation on single-walledcarbon nanotubesIWEPNM, Kirchberg in Tirol, Austria (2012)

6. H.-M. Tohati, B. Botka, K. Nemeth, A. Pekker, K. Kamaras:Infrared and Raman investigation of carbon nanotube-based hybrid systemsIWEPNM, Kirchberg in Tirol, Austria (2010)

7. E. A. Francis, S. Scharinger, K. Nemeth, K. Kamaras, C. A. Kuntscher:Investigation of the Jahn–Teller effect in C−60 monoanion under high pressure byinfrared spectroscopyIWEPNM, Kirchberg in Tirol, Austria (2010)

8. H.-M. Tohati, K. Nemeth, A. Pekker, K. Kamaras:Infrared measurements on carbon nanotube-poly(allylamine hydrochloride) hy-brid systemsFISS, Krutyn, Poland (2010)

9. H.-M. Tohati, K. Nemeth, K. Kamaras, S. Ben-Valid, A. Zeng, L. Reiss, S.Yitzchaik, M. Pietraszkiewicz, O. Pieraszkiewicz, L. Maggini, D. Bonifazi:Infrared spectroscopic investigation on non-covalently functionalized single walledcarbon nanotubesACN’2011, St. Petersburg, Russia (2011)

10. H.-M. Tohati, K. Nemeth, K. Kamaras:Wide range optical study on double-walled carbon nanotubes prepared from sep-arated outer tubesIWEPNM, Kirchberg in Tirol, Austria (2012)

88

Oral presentations:

1. K. Nemeth, A. Francis:Synthesis of (Ph4P)2C60IMini-Workshop on ”Synthesis and spectroscopic characterization of carbon nanos-tructures”, Augsburg, Germany (2009)

2. K. Nemeth, A. Francis, M. Gyori, P. Matus, K. Kamaras:Synthesis and infrared measurements of (Ph4P)2C60I and TDAE-C60

Mini-Workshop on ”Synthesis and spectroscopic characterization of carbon nanos-tructures”, Augsburg, Germany (2010)

89

References

[1] http://teachers.yale.edu/curriculum/extra/images/2010/10.05.07.02.jpg.

[2] S. Iijima. Helical microtubules of graphitic carbon. Nature, 354:56, 1991.

[3] S. Iijima and T. Ichihashi. Single-shell carbon nanotubes of 1-nm diameter. Nature,363:603, 1993.

[4] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos,I. V. Grigorieva, and A. A. Firsov. Electric field effect in atomically thin carbonfilms. Science, 306:666, 2004.

[5] M. Damnjanovic, I. Milosevic, T. Vukovic, and R. Sredanovic. Full symmetry,optical activity, and potentials of single-wall and multiwall nanotubes. Phys. Rev.B, 60:2728, 1999.

[6] F. Borondics. Egyfalu szen nanocsovek optikai vizsgalata es funkcionalizalasa (PhDthesis). Eotvos Lorand University, 2007.

[7] E. G. Gamaly and T. W. Ebbesen. Mechanism of carbon nanotube formation inthe arc discharge. Phys. Rev. B, 52:2083, 1995.

[8] M. E. Itkis, D. E. Perea, S. Niyogi, S. M. Rickard, M. A. Hamon, H. Hu, B. Zhao,and R. C. Haddon. Purity evaluation of as-repared single-walled carbon nanotubesoot by use of solution-phase near-IR spectroscopy. Nano Lett., 3:309, 2003.

[9] D. S. Bethune, C. H. Kiang, M. S. de Vries, G. Gorman, R. Savoy, J. Vazquez, andR. Beyers. Cobalt-catalyzed growth of carbon nanotubes with single-atomic-layerwalls. Nature, 363:605, 1993.

[10] D. Zhou, S. Seraphin, and S. Wang. Single-walled carbon nanotubes growing fromYC2 particles. Appl. Phys. Lett., 65:1593, 1994.

[11] P. M. Ajayan, J. M. Lambert, P. Bernier, L. Barbedette, C. Colliex, and J. M.Planeix. Growth morphologies during cobalt-catalyzed single-shell carbon nano-tube synthesis. Chem. Phys. Lett., 215:509, 1993.

[12] 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. Tomanek, J. E. Fischer, andR. E. Smalley. Crystalline ropes of metallic carbon nanotubes. Science, 273:483,1996.

[13] T. Guo, P. Nikolaev, A. Thess, D. T. Colbert, and R. E. Smalley. Catalytic growthof single-walled nanotubes by laser vaporization. Chem. Phys. Lett., 243:49, 1995.

90

[14] W. K. Maser, E. Munoz, A. M. Benito, M. T. Martınez, G. F. de la Fuente,Y. Maniette, E. Anglaret, and J.-L. Sauvajol. Production of high density single-walled nanotube material by a simple laser-ablation method. Chem. Phys. Lett.,292:587, 1998.

[15] A. G. Rinzler, J. Liu, H. Dai, P. Nikolaev, C. B. Huffman, F. J. Rodriguez-Macıas,P. J. Boul, A. H. Lu, D. Heymann, D. T. Colbert, R. S. Lee, J. E. Fischer, A. M.Rao, P. C. Eklund, and R. E. Smalley. Large-scale purification of single-walledcarbon nanotubes: process, product, and characterization. Appl. Phys. A, 67:29,1998.

[16] H. Dai. Carbon nanotubes: opportunities and challenges. Surf. Science, 500:218,2002.

[17] P. Nikolaev, M. J. Bronikowski, R. K. Bradely, F. Rohmund, D. T. Colbert, K. A.Smith, and R. E. Smalley. Gas-phase catalytic growth of single-walled carbonnanotubes from carbon monoxide. Chem. Phys. Lett., 313:91, 1999.

[18] G. Lolli, L. Zhang, L. Balzano, N. Sakulchaicharoen, Y. Tan, and D. E. Resasco.Tailoring (n,m) structure of single-walled carbon nanotubes by modifying reactionconditions and the nature of the support of CoMo catalysts. J. Phys. Chem. B,110:2108, 2006.

[19] P. R. Wallace. The band theory of graphite. Phys. Rev., 71:622, 1947.

[20] L. Van Hove. The occurence of singularities in the elastic frequency distributionof a crystal. Phys. Rev., 89:1189, 1953.

[21] M. Dressel and G. Gruner. Electrodynamics of Solids. Cambridge University Press,2003.

[22] Solyom J. A modern szilardtest-fizika alapjai I., Szerkezet es dinamika. ELTEEotvos Kiado, 2009.

[23] http://www.photon.t.u-tokyo.ac.jp/∼maruyama/kataura/kataura.html.

[24] A. Pekker. Wide range optical studies on single-walled carbon nanotubes (PhDthesis). Budapest University of Technology and Economics, 2011.

[25] H. Kataura, J. Kumazava, Y. Maniwa, I. Umezu, S. Suzuki, Y. Ohtsuka, andY. Achia. Optical properties of single-walled carbon nanotubes. Synth. Met.,103:2555, 1999.

[26] R. Saito, M. Fujita, and M. S. Dresselhaus. Electronic structure of chiral graphenetubules. Appl. Phys. Lett., 60:2204, 1992.

[27] A. Pekker, A. Botos, A. Rusznyak, J. Koltai, J. Kurti, and K. Kamaras. Vibra-tional Signatures in the Infrared Spectra of Single- and Double-Walled CarbonNanotubes and Their Diameter Dependence. J. Phys. Chem. Lett., 2:2079, 2011.

[28] P. C. Eklund, J. M. Holden, and R. A. Jishi. Vibrational modes of carbon nano-tubes; spectroscopy and theory. Carbon, 33:959, 1995.

91

[29] S. Reich, C. Thomsen, and J. Maultzsch. Carbon Nanotubes: Basic Concepts andPhysical Principles. Wiley-VCH, Weinheim, 2004.

[30] A. Jorio, M. A. Pimenta, A. G. Souza Filho, R. Saito, G. Dresselhaus, and M. S.Dresselhaus. Characterizing carbon nanotube samples with resonance Raman scat-tering. New Journal of Physics, 5:139, 2003.

[31] I. Cabria, J. W. Mintmire, and C. T. White. Metallic and semiconducting narrowcarbon nanotubes. Phys. Rev. B, 67:121406, 2003.

[32] S. Niyogi, M. A. Hamon, H. Hu, B. Zhao, P. Bhowmik, R. Sen, M. E. Itkis, andR. C. Haddon. Chemistry of single-walled carbon nanotubes. Acc. Chem. Res.,35:1105, 2002.

[33] M. S. Dresselhaus, G. Dresselhaus, and R. Saito. Physics of carbon nanotubes.Carbon, 33:883, 1995.

[34] J. Li, G. Jia, Y. Zhang, and Y. Chen. Bond-curvature effect of sidewall [2+1]cycloadditions of single-walled carbon nanotubes: A new criterion to the adductstructures. Chem. Mater., 18:3579, 2006.

[35] X. Blase, L. X. Benedict, E. L. Shirley, and S. G. Louie. Hybridization effects andmetallicity in small radius carbon nanotubes. Phys. Rev. Lett., 72:1878, 1994.

[36] V. Zolyomi and J. Kurti. First-principles calculations for the electronic band struc-tures of small diameter single-wall carbon nanotubes. Phys. Rev. B, 70:085403,2004.

[37] M. J. O’Connell, S. M. Bachilo, C. B Huffman, V. C. Moore, M. S. Strano, E. HHaroz, K. L. Rialon, P. J. Boul, W. H. Noon, C. Kittrell, J. Ma, R. H Hauge, R. B.Weisman, and R. E. Smalley. Band gap fluorescence from individual single-walledcarbon nanotubes. Science, 297:593, 2002.

[38] M. N Tchoul, W. T. Ford, G. Lolli, D. E. Resasco, and S. Arepalli. Effect of mildnitric acid oxidation on dispersabiltiy, size and structure of single-walled carbonnanotubes. Chem. Mater., 19:5765, 2007.

[39] K. A. Worsley, I. Kalinina, E. Bekyarova, and R. C. Haddon. Functional-ization and dissolution of nitric acid treated single-walled carbon nanotubes.J. Am. Chem. Soc., 131:18153, 2009.

[40] J. Zhu, M. Yudasaka, M. Zhang, and S. Iijima. Dispersing carbon nanotubes inwater: A noncovalent and nonorganic way. J. Phys. Chem. B, 108:11317, 2004.

[41] S. Banerjee, T. Hemraj-Benny, and S. S. Wong. Covalent surface chemistry ofsingle-walled carbon nanotubes. Adv. Mater., 17:17, 2005.

[42] H. K. F. Cheng, T. Basu, N. G. Sahoo, L. Li, and S. H. Chan. Current Advancesin the Carbon Nanotube/Thermotropic Main-Chain Liquid Crystalline PolymerNanocomposites and Their Blends. Polymers, 4:889, 2012.

[43] D. Tasis, N. Tagmatarchis, A. Bianco, and M. Prato. Chemistry of carbon nano-tubes. Chem. Rev., 106:1105, 2006.

92

[44] B. W. Smith, M. Monthioux, and D. E. Luzzi. Encapsulated C60 in carbon nano-tubes. Nature, 396:323, 1998.

[45] L. Guan, Z. Shi, M. Li, and Z. Gu. Ferrocene-filled single-walled carbon nanotubes.Carbon, 43:2780, 2005.

[46] D. Kocsis, D. Kaptas, A. Botos, A. Pekker, and K. Kamaras. Ferrocene encapsu-lation in carbon nanotubes: Various methods for filling and investigation. phys.stat. sol. (b), 248:2512, 2011.

[47] P. M. F. J. Costa, J. Sloan, T. Rutherford, and M. L. H. Green. Encapsulation ofRexOy clusters within single-wall carbon nanotubes and their in tubulo reductionand sintering to Re metal. Chem. Mater., 17:6579, 2005.

[48] A. Chuvilin, E. Bichoutskaia, M. C. Gimenez-Lopez, T. W. Chamberlain, G. A.Rance, N. Kuganathan, J. Bisupek, U. Kaiser, and A. N. Khlobystov. Self-assemblyof a sulphur-terminated graphene nanoribbon within a single-walled carbon nano-tube. Nature Materials, 10:687, 2011.

[49] C. Pham-Huu, N. Keller, C. Estournes, G. Ehret, J. M. Greneche, and M. J.Ledoux. Microstructural investigation and magnetic properties of CoFe2O4

nanowires synthetized inside carbon nanotubes. Phys. Chem. Chem. Phys., 5:3716,2003.

[50] B. Botka, A. Pekker, A. Botos, K. Kamaras, and R. Hackl. A systematic study ofoptical and Raman spectra of peapod-based DWNTs. phys. stat. sol. (b), 247:2843,2010.

[51] A. J. Birch. The reduction of organic compounds by metal-ammonia solutions.Q. Rev. Chem. Soc., 4:69, 1950.

[52] A. J. Birch and H. Smith. Reduction by metal-amine solutions: applications insynthesis and determination of structure. Q. Rev. Chem. Soc., 12:17, 1958.

[53] H. E. Zimmerman. A Mechanistic Analysis of the Birch Reduction. Acc. Chem.Res., 45:164, 2012.

[54] M. S. Dresselhaus and G. Dresselhaus. Intercalation compounds of graphite. Ad-vances in Physics, 51:1, 2002.

[55] A. M. Rao, P. C. Eklund, S. Bandow, A. Thess, and R. E. Smalley. Evidence forcharge transfer in doped carbon nanotube bundles from Raman scattering. Nature,388:257, 1997.

[56] O. Zhou, R. M. Fleming, D. W. Murphy, C. H. Chen, R. C. Haddon, A. P. Ramirez,and S. H. Glarum. Defects in carbon nanostrucures. Science, 263:1744, 1994.

[57] R. S. Lee, H. J. Kim, J. E. Fischer, A. Thess, and R. E. Smalley. Conductivityenhancement in single-walled carbon nanotube bundles doped with K and Br.Nature, 388:255, 1997.

[58] R. Jacquemin, S. Kazaoui, D. Yu, A. Hassanien, N. Minami, H. Kataura, andY. Achiba. Doping mechanism in single-wall carbon nanotubes studied by opticalabsorption. Synth. Met., 115:283, 2000.

93

[59] G. Sumanasekera, J. Allen, and P. Eklund. Electrochemical doping of bundles ofundoped single wall carbon nanotubes. Mol. Cryst. Liq. Cryst., 340:535, 2000.

[60] S. Suzuki and M. Tomita. Observation of potassium-intercalated carbon nanotubesand their valence-band excitation spectra. J. Appl. Phys., 79:3739, 1996.

[61] V. A. Nalimova, D. E. Sklovsky, G. N. Bondarenko, H. Alvergnat-Gaucher, S. Bon-namy, and F. Beguin. Lithium interaction with carbon nanotubes. Synth. Met.,88:89, 1997.

[62] S. Suzuki, C. Bower, and O. Zhou. In-situ TEM and EELS studies of alkali-metalintercalation with single-walled carbon nanotubes. Chem. Phys. Lett., 285:230,1998.

[63] C. Bower, S. Suzuki, K. Taginaki, and O. Zhou. Synthesis and structure of pris-tine and alkali-metal-intercalated single-walled carbon nanotubes. Appl. Phys. A,67:47, 1998.

[64] T. Pichler, M. Knupfer, M. S. Golden, and J. Fink. Localized and delocalizedelectronic states in single-wall carbon nanotubes. Phys. Rev. Lett., 80:4729, 1998.

[65] T. Pichler, M. Sing, M. Knupfer, M. S. Golden, and J. Fink. Potassium inter-calated bundles of single-wall carbon nanotubes: electronic structure and opticalproperties. Solid State Communications, 109:721, 1999.

[66] J. E. Fischer. Chemical doping of single-wall carbon nanotubes. Acc. Chem. Res.,35:1079, 2002.

[67] N. Bendiab, A. M. Saita, R. Aznar, J. L. Sauvajol, R. Almairac, I. Mirebeau,and G. Andre. Rubidium localization in single-walled carbon nanotube bundles:structural study. Phys. Rev. B, 78:104108, 2008.

[68] A. Kukovecz, T. Pichler, R. Pfeiffer, C. Kramberger, and H. Kuzmany. Diameterselective doping of single wall carbon nanotubes. Phys. Chem. Chem. Phys., 5:582,2003.

[69] B. Vigolo, C. Herold, J.-F. Mareche, P. Bourson, S. Margueron, J. Ghanbaja, andE. McRae. Direct revealing of the occupation sites of heavy alkali metal atomsin single-walled carbon nanotube intercalation compounds. J. Phys. Chem C,113:7624, 2009.

[70] O. Tanaike, O. Kimizuka, N. Yoshizawa, K. Yamada, X. Wang, H. Hatori, andM. Toyoda. Debundling of swcnts through a simple intercalation techniqe. Elec-trochem. Commun., 11:1441, 2009.

[71] Bukki T., Kamaras K., and Kalman P. Bevezetes a modern optikaba V. Mue-gyetemi Kiado, 2000.

[72] Y. Chen, R. C. Haddon, S. Fang, A. M. Rao, P. C. Eklund, W. H. Lee, E. C.Dickey, E. A. Grulke, J. C. Pendergrass, A. Chavan, B. E. Haley, and R. E. Smal-ley. Chemical attachment of organic functional groups to single-walled carbonnanotube material. J. Mater. Res., 13:2423, 1998.

94

[73] S. Pekker, J.-P. Salvetat, E. Jakab, J.-M. Bonard, and L. Forro. Hydrogenation ofcarbon nanotubes and graphite in liquid ammonia. J. Phys. Chem. B, 105:7938,2001.

[74] F. Borondics, M. Bokor, P. Matus, K. Tompa, S. Pekker, and E. Jakab. Reduc-tive functionalization of carbon nanotubes. Fullerenes, Nanotubes and CarbonNanostructures, 13:375, 2005.

[75] F. Borondics, E. Jakab, and S. Pekker. Functionalization of carbon nanotubes viadissolving metal reductions. J. Nanosci. Nanotechnol., 7:1, 2007.

[76] D. E. Bergbreiter and J. M. Killough. Reactions of potassium-graphite. J. Am.Chem. Soc., 100:2126, 1978.

[77] A. Dailly, J. W. L. Yim, C. C. Ahn, E. Miura, R. Yazami, and B. Fultz. Purificationof carbon single-wall nanotubes by potassium intercalation and exfoliation. Appl.Phys. A, 80:717, 2005.

[78] D. Voiry, O. Roubeau, and A. Penicaud. Stoichiometric control of single walledcarbon nanotubes functionalization. J. Mater. Chem., 20:4385, 2010.

[79] Y. Iwasa, H. Fudo, Y. Yatsu, T. Mitani, H. Kataura, and Y. Achiba. Phase stabilityof doped carbon nanotubes. Synth. Met., 121:1203, 2001.

[80] N. M. Nemes, J. E. Fischer, K. Kamaras, D. B. Tanner, and A. G. Rinzler. Elec-tronic and structural properties of alkali doped SWNT. AIP Conference Proceed-ings, 633:259, 2002.

[81] K. Tompa, P. Banki, M. Bokor, G. Lasanda, and L. Vasaros. Diffusible and residualhydrogen in amorphous Ni(Cu)-Zr-H alloys. J. Alloys Comp., 350:52, 2003.

[82] Z. C. Wu, Z. H. Chen, X. Du, J. M. Logan, J. Sippel, M. Nikolou, K. Kamaras,J. R. Reynolds, D. B. Tanner, A. F. Hebard, and A. G. Rinzler. Transparent,conductive carbon nanotube films. Science, 305:1273, 2004.

[83] A. Pekker and K. Kamaras. Wide-range optical studies on various single-walledcarbon nanotubes: origin of the low-energy gap. Phys. Rev. B, 84:075475, 2011.

[84] A. Pekker, F. Borondics, K. Kamaras, A. G. Rinzler, and D. B. Tanner. Calculationof optical constants from carbon nanotube transmission spectra. phys. stat. sol.(b), 243:3485, 2006.

[85] S. Fogden, C. A. Howard, R. K. Heenan, N. T. Skipper, and M. S. P. Shaffer.Scalable method for the reductive dissolution, purification and separation of single-walled carbon nanotubes. ACS Nano, 6:54, 2012.

[86] B. Gebhardt, Z. Syrgiannis, C. Backes, R. Graupner, F. Hauke, and A. Hirsch.Carbon nanotube sidewall functionalization with carbonyl compounds - ModifiedBirch conditions vs the organometallic reduction approach. J. Am. Chem. Soc.,133:7985, 2011.

[87] M. Muller, R. Meinke, J. Maultzsch, Z. Syrgiannis, F. Hauke, A. Pekker, K. Ka-maras, A. Hirsch, and C. Thomsen. Electronic Properties of Propylamine-Functionalized Single-Walled Carbon Nanotubes. ChemPhysChem, 11:2444, 2010.

95

[88] A. Pekker, D. Wunderlich, K. Kamaras, and A. Hirsch. Diameter selectivity ofnanotube sidewall functionalization probed by optical spectroscopy. phys. stat.sol. (b), 245:1954, 2008.

[89] K. Nemeth, A. Pekker, F. Borondics, E. Jakab, N. M. Nemes, K. Kamaras, andS. Pekker. Investigation of hydrogenated HiPco nanotubes by infrared spec-troscopy. phys. stat. sol. (b), 247:2855, 2010.

[90] H. Hu, B. Zhao, M. A. Hamon, K. Kamaras, M. E. Itkis, and R. C. Haddon. Side-wall functionalization of single-walled carbon nanotubes by addition of dichloro-carbene. J. Am. Chem. Soc., 125:14893, 2003.

[91] J. Cabana, S. Lavoie, and R. Martel. Thermal Chemistry of Methylene- andPhenyl-Functionalized Caron Nanotubes. J. Am. Chem. Soc., 132:1389, 2010.

[92] L. A. Girifalco, M. Hodak, and R. S. Lee. Carbon nanotubes, buckyballs, ropesand a universal graphitic potential. Phys. Rev. B, 62:13104, 2000.

[93] C.-H. Sun, L.-C. Yin, F. Li, G.-Q. Lu, and H.-M. Cheng. Van derWaals interactions between two parallel infinitely long single-walled nanotubes.Chem. Phys. Lett., 403:343, 2005.

96