Ruelle

Functionalization of carbon nanotubes
via plasma post-discharge surface treatment:

Implication as nanofiller in polymeric matrices
Benoit RUELLE
Dissertation prsente en vue de
lobtention du grade acadmique
de Docteur en Sciences
Directeurs de thse : Prof. M. HECQ
Prof. Ph. DUBOIS
Laboratoire de Chimie Inorganique et

Analytique
Anne acadmique 2009-2010
A ma famille et Emilie
Remerciements
A l'issue de la rdaction de cette recherche, je suis convaincu que la thse est loin d'tre un
travail solitaire. En effet, je n'aurais jamais pu raliser ce travail sans le soutien d'un grand
nombre de personnes dont la gnrosit, la bonne humeur et l'intrt manifests l'gard de
ma recherche m'ont beaucoup aids et permis de progresser.
En premier lieu, je tiens remercier mes directeurs de thse, les Professeurs Michel Hecq et
Philippe Dubois qui mont offert leur confiance et lopportunit de pouvoir raliser cette thse
dans dexcellentes conditions au sein de leurs laboratoires respectifs. Le fait de savoir que
leurs portes taient toujours grandes ouvertes ma t dune grande aide.
Jaimerais remercier plus particulirement trois personnes dont lapport et lcoute ont t
plus que primordiaux ce travail. Merci Carla Bittencourt pour toutes ses discussions
concernant les nanotubes de carbone, ses relectures et les magnifiques analyses XPS et TEM
haute rsolution. Un grand merci galement Thomas Godfroid pour les transformations de la
chambre plasma micro-onde, lide de la Nanotube Dance Party, sa science du plasma microonde, ses conseils et encouragements dans les moments difficiles. Enfin, dzikuj Sophie
Peeterbroeck pour son clairage concernant les composites polymres base de nanotubes, sa
patience, son soutien sans failles, ses relectures et, surtout, son amiti.
I would like to thank Professor Vittoria Vittoria for having welcomed me in the Department
of Chemical and Food Engineering in Salerno University. I am very grateful to Doctor
Giuliana Gorrasi and Professor Salvatore De Pasquale for their welcome, their kindness and
for having shared their enthusiasm. Special thanks to Giovanni, Salvatore, Dante and Luigi
for their help, their jokes and, of course, the Italian coffee breaks which are become a tradition
for me. Finally, a big thank to my friend Stefano for his welcome and the enjoyable time that I
spent with him and his friends. Grazie mille Salerno.
I would like to thank Doctor Hans Miltner, from the VUB, for his collaboration in the
crystallization studies and Professor Gustaaf van Tendeloo for the high resolution TEM
images performed in the EMAT in University of Antwerp.
Merci au Professeur Giovanni Camino, au Docteur Frdric Luizi, au Docteur Rony Snyders
et au Docteur Didier Villers davoir accept de faire partie de mon jury de thse.
Merci au F.R.I.A. (Fonds pour la Formation la Recherche dans lIndustrie et dans

lAgriculture) pour le financement de cette thse.
Je tiens aussi remercier les membres du SMPC pour leur sympathie et les moments partags,
et plus spcialement Ben, Nathalie, Deborah, Fabian, Fouad, Samira, Leila, Manu, Mounch,
Magali, Anne-Lise et Stphane.
Jaimerais galement remercier toute lquipe du LCIA, Rony, Stephanos, Monsieur Dauchot,
Corinne, Fabian, Yoann, Damien, Denis, Matthieu, David, Greg, Fabien, Dany, Adil,
Philippe, Daphn, Axel, Alisson, Alice, Damien, Xavier, Cristina ainsi que Mickal, Rachel,
Anne-so et Christine pour lambiance chaleureuse et joyeuse quils ont fait rgner au sein du
service ainsi que les coups de main ou conseils quils ont pu me prodiguer.
Jaimerais aussi remercier plus particulirement Laurent, mon compre depuis sept ans dj,
pour son soutien (surtout lors du retour du FRIA), son coute et nos dlires. Une pense
galement mes amis, Julien, Delphine, Melissa, Nathalie, Philippe, Renaud, Ahmed et Maria
pour tous ces moments partags et assez souvent folkloriques
Enfin, je tiens remercier du plus profond de mon cur ma famille qui ma soutenu et
encourag tout au long de mes tudes en tant tout moment mes cts.
Pour finir, je te remercie Milie, pour ton soutien quotidien, tout ce que tu mapportes, tes
relectures et de me supporter, surtout ces derniers mois
TABLE OF CONTENT
I. INTRODUCTION ....................................................................................................................1
II. STATE OF THE ART ............................................................................................................5
II.A. Carbon Nanotubes ............................................................................................................5
II.A.1. Structure of carbon nanotubes ....................................................................................6
II.A.2. Carbon nanotubes : Methods of production .............................................................9
II.A.3. Properties of carbon nanotubes ................................................................................11
II.A.4. A large range of applications ....................................................................................14
II.B.
Functionalization of carbon nanotubes ......................................................................18
II.B.1. Interests and types of functionalization...................................................................18

II.B.2. Non-covalent functionalization with surfactant or polymer.................................19
II.B.3. Covalent functionalization ........................................................................................21
II.B.3.a. End and defect-side chemistry.................................................................................21
II.B.3.b. Sidewall functionalization .......................................................................................24
II.C. Plasma treatment of carbon nanotubes ......................................................................28

II.C.1. Introduction to plasma : the fourth state of matter ................................................28
II.C.2. Characteristics and principal applications of plasma ............................................29
II.C.3. Applications of plasma into materials surface functionalization .........................32
II.D. Polymer nanocomposites-based on carbon nanotubes............................................34
II.D.1. Generalities on polymer nanocomposites ...............................................................34
II.D.2. Melt blending of polymer matrices with carbon nanotubes .................................36
II.D.3. Interfacial compatibilization: improvement of CNTs dispersion ........................37
II.D.4. Carbon nanotube-based nanocomposites: applications.........................................45
III. OBJECTIVES & PROPOSED RESEARCH STARTEGIES ............................54
TABLE OF CONTENT
IV. RESULTS & DISCUSSION ...........................................................................................58

IV.A. Microwave plasma post-discharge treatment ...........................................................58
IV.A.1. Microwave plasma experimental set-up and treatment parameters ....................59
IV.A.2. Atomic nitrogen treatment of carbon nanotubes ....................................................60
IV.A.2.a. Treatment of CNTs pressed pellet ........................................................................61
IV.A.2.b. Morphology of nitrogen-treated carbon nanotubes ...............................................65
IV.A.2.c. Treatment of CNT powder ....................................................................................67
IV.A.2.d. Origin of oxygen surface contamination ...............................................................69
IV.A.3. Quantification of primary amine functions grafted on CNTs ..............................72

IV.A.3.a. Limits of XPS N1s peak deconvolution ................................................................73
IV.A.3.b. Chemical derivatization of primary amines ..........................................................74
IV.A.3.c. Post-plasma reaction with hydrogen gas ...............................................................78
IV.A.3.d. Introduction of H2 in the -wave plasma discharge ..............................................78
IV.A.3.e. Introduction of H2 in the -wave plasma post-discharge ......................................83
IV.A.3.f. Stability of primary amines grafted on CNTs .......................................................86
IV.A.4. Influence of treatment time on nitrogen and NH2 functions grafting ...............88
IV.A.5. Conclusion ..................................................................................................................91
IV.B. Preparation and characterization of CNTs-g-PCL nanohybrids .........................92
IV.B.1. Preparation of CNTs-g-PCL nanohybrids ..............................................................92
IV.B.2. CNTs-g-PCL nanohybrids characterization ...........................................................96
IV.B.2.a. Content in PCL grafted on CNTs surface and free PCL chains ............................96
IV.B.2.b. Solvent dispersibility of CNTs-g-PCL nanohybrids .............................................98
IV.B.2.c. Morphology of CNTs-g-PCL nanohybrids ...........................................................99
IV.B.3. Conclusion ................................................................................................................104

IV.C. Carbon nanotubes-based nanocomposites .....................................................................105
IV.C.1. Thermal properties of CNTs-based nanocomposites ..........................................107
IV.C.1.a. Influence of CNTs on the thermal degradation of nanocomposites ....................107
IV.C.1.b. Thermal degradation of PCL nanocomposites ....................................................108
IV.C.1.c. Thermal degradation of HDPE nanocomposites .................................................111
IV.C.1.d. DSC analyses of PCL nanocomposites ...............................................................113
TABLE OF CONTENT
IV.C.1.e. DSC analyses of HDPE nanocomposites ............................................................116

IV.C.1.f. Conclusion ..........................................................................................................118
IV.C.2. Electrical properties of CNTs-filled nanocomposites .........................................119

IV.C.2.a. Electrical properties of nanocomposites and percolation theory ........................119
IV.C.2.b. Electrical resistivity measurements range ...........................................................123
IV.C.2.c. Electrical properties of PCL nanocomposites .....................................................124
IV.C.2.d. Electrical properties of HDPE nanocomposites ..................................................132
IV.C.2.e. Conclusions .........................................................................................................139
V. GENERAL CONCLUSIONS & PERSPECTIVES .............................................143

VI. EXPERIMENTAL PART & INSTRUMENTATION........................................147
VI.A. Materials ..........................................................................................................................147
VI.B. Preparation of samples .................................................................................................147
VI.C. Instrumentation .............................................................................................................150
TABLE OF CONTENT
1. INTRODUCTION
Since their first observation in 19911, carbon nanotubes (CNTs) have attracted a lot of
attention of many researchers and engineers around the world owing to their exceptional
properties. Their outstanding physical properties are a direct result of the near-perfect
structure of the CNTs, which at atomic scale can be thought of as a hexagonal sheet of carbon
atoms rolled into a seamless one-dimensional cylindrical shape. Due to this structure, no
previous material has displayed the combination of superlative properties attributed to them2.
Besides their extremely small (nanometric) size, their excellent electrical and thermal
conducting performances combined with their high toughness and transverse flexibility allow
for envisaging their use in a wide variety of applications like, for example, electronic
components, chemical and biological sensors, chemical and genetic probes, field emission
tips, mechanical memories, hydrogen and ion storage components, etc3. Carbon nanotubes are
thus expected to transmit their unique properties in multidisciplinary fields and should play a
key role in the nanotechnologies of the 21st century.
Offering in the same time a high aspect ratio (length-to-diameter) and a low density, this
allotropic variety of carbon is an ideal candidate as advanced filler materials in composites.
Indeed, CNTs can provide a three-dimensional conductive network through the polymer
matrix with exceedingly low percolation thresholds4. Furthermore, it has been suggested that
their high thermal conductivity can be exploited to make thermally conductive composites5.
In addition, the mechanical enhancement of polymer materials using CNTs as reinforcing
nanofillers should be exploited in the very near future6.
A key parameter for producing nanocomposite materials with largely improved
physical properties relies on the extent of dispersion of the individual CNTs. However, the
CNTs tend to form long bundles thermodynamically stabilized by numerous van der Waals
forces between the tubes and by entanglements, which occur during their synthesis. These
aggregates constitute a handicap in the majority of CNTs applications due to their low
solubility in water and organic solvents. In the case of their use as nanofillers, the
homogeneous dispersion of CNTs in polymer matrices is relatively difficult to achieve within
a large majority of polymer7.
I. INTRODUCTION
Figure 1- 1 : TEM micrograph of CNTs bundle/aggregate
The functionalization of the CNT sidewalls represents a solution in order to tune the
interactions between CNTs and the host polymer matrix and improve their dispersion ability.
This surface modification can be divided into two main approaches.
One is the non-covalent functionalization, used in different techniques such as the addition of
surfactants8, polymer wrapping9 or polymerization-filling technique (PFT)10, which allows the
unaltered CNTs for preserving their physical properties. However, the interaction between the
wrapping molecules and the nanotubes remain generally weak limiting the efficiency of the
properties transfer between the nanotubes and the host polymer, in particular a low load
transfer for mechanical properties.
The other approach relies on the covalent grafting of functional groups along the CNT
sidewalls. These chemical functions are interesting to improve the interactions with the
polymer matrix11 and can be used as anchoring sites for polymer chains8. The covalent
bonding between CNTs and the polymer allows obtaining an optimal interfacial strength and
thus a better load transfer. The covalent functionalization will influence the CNTs properties
and depends on the location of functionalization sites. Indeed, the functionalization can take
place on already existing CNT structural defects or defect sites can be created by the grafting
of chemical functions. In this last case, the treatment for grafting can degrade the CNTs
physical properties if the extent of functionalization is too important leading to damaged
CNTs structure12.
In the majority of so far reported functionalization processes such as chemical
functionalization performed in solvent13 or by ball milling14, only a very tiny fraction of
I. INTRODUCTION
carbon atoms along the nanotubes are functionalized. And some targeted reactive functions,
like amine groups, most often require time-consuming and costly multistep reactions that
rarely result in a controlled covalent functionalization process.
The aim of this research is to develop a non destructive and selective covalent
functionalization of CNTs and to disperse these treated CNTs homogenously within a
polymer matrix with covalent bonds at the interface polymer/nanotube with the preservation
of electrical properties of CNTs. It is thus important to control the quantity and the nature of
the grafted functions since it is necessary to have enough superficial chemical groups while
preserving inherent CNTs properties.
In the first part of the study, the treatment of CNTs via an original microwave plasma process
is developed to readily functionalize the CNTs surface with primary amine groups in one
single step. The use of plasma, which is an ionized gas consisting in ions, electrons and
neutral species, allows to treat the CNTs with a relative low time and environmentally
friendly process without any use of solvent. The quantity of grafted nitrogenated groups, and
particularly primary amines, was determined by x-ray photoelectron spectroscopy (XPS). The
covalent character of grafted nitrogen groups was studied by high-resolution ultraviolet
photoelectron spectroscopy (HR-UPS).
The so-grafted primary amine groups are used as initiation sites for promoting the ring
opening polymerization (ROP) of lactone monomers (i.e. -caprolactone) yielding polyestergrafted CNT nanohybrids. The morphology of the recovered nanohybrids and the quantity of
grafted polymer chains were studied by transmission electron microscopy (TEM) and
thermogravimetric analysis (TGA), respectively.
Finally, these nanohybrids are used as highly filled masterbatches to be dispersed in the
molten state within either poly(-caprolactone) (PCL) or high-density polyethylene (HDPE)
matrices to obtain high performance nanocomposites. The thermal behavior as well as their
electrical properties of the so-produced nanocomposites were characterized and correlated to
their morphology.
I. INTRODUCTION
References
1
Iijima S., Nature 1991, 354, 56-58

Thostenson E., Ren Z., Chou T.-W., Composites Sci. Tech. 2001, 61, 1899-1912
3
Collins P. G., Avouris P., Sci. Am. 2000, 12, 62-69
4
Kilbride B. E., Coleman J. N., Fraysse J., Fournet P., Cadek M., Drury A., J. Appl. Phys. 2002, 92, 4024-4030
5
Bagchi A., Nomura S., Compos. Sci. Technol. 2006, 66, 1703-1712
6
Wang W., Ciselli P., Kuznetsov E., Peijs T., Barber A. H., Philos. Trans. Roy. Soc. A 2008, 368, 1613-1626
7
Andrews R., Weisenberger M. C., Curr. Opin. Solid State Mater. Sci. 2004, 8, 31-37
8
Hirsch A., Angew. Chem. Int. Ed. 2002, 41, 1853-1859
9
Liu P., Eur. Polym. J. 2005, 41, 2693-2703
10
Bonduel D., Mainil M., Alexnadre M., Monteverde F., Dubois Ph., Chem. Commun. 2005, 781-783
11
Thostenson E. T., Li C., Chou T. W., Compos. Sci. Technol. 2005, 65, 491-516
12
Fu K., Huang W., Lin Y., Riddle L. A., Carroll D. L., Sun Y. P., Nano Lett. 2001, 1, 439-441
13
Banerjee S., Hemraj-Benny T., Wong S. S., Adv. Mater. 2005, 15, 3334-3339
14
Konya Z., Vesselenyi I., Niesz K., Kukovecz A., Demortier A., Fonseca A., Delhalle J., Mekhalif Z., Nagy J.
B., Koos A. A., Osvath Z., Kocsonya A., Biro L. P., Kiricsi I., Chem. Phys. Lett. 2002, 360, 429-435
2
I. INTRODUCTION
II. STATE OF THE ART

II.A. Carbon nanotubes
First TEM evidence for the tubular nature of nano-sized carbon filaments was
published in 1952 1 and many other reports have followed over the next forty years 2,3,4.
However, these fibers were not recognized as nanotubes and were not studied systematically.
It was only after the discovery of fullerenes in 1985 5 that carbon nanostructures were
explored further and attracted global scientific attention. Interest in carbon nanotube research
began in 1991 when the Japanese electron microscopist Sumio Iijima 6 observed tubular
features in carbon soot produced in an arc discharge. The structure of these tubules is
composed by concentric multiwalls forming one cylinder defining a multi-walled nanotube
(MWNT).
In 1993, two distinctive research groups demonstrated independently the ability to produce
single-sheet tubules, coined as single-walled carbon nanotubes (SWNTs) 7,8.
Promising physical, chemical and electronic properties of carbon nanotubes have been
highlighted in the next years, leading to a great expansion of the breadth and range of research
dedicated to these nanoparticles over the past several years (Figure 2-1). Once considered a
novelty, carbon nanotubes are today a key foundational material of the nanotechnology
revolution.
N u m b e r o f p u b lic a t ion s
8000
7000
6000
5000
4000
3000
2000
1000
0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Year
Figure 2- 1 : Number of publications on carbon nanotubes by year (www.isiknowledge.com)
II.A.1. Structure of carbon nanotubes

Carbon nanotubes (CNTs) are an allotropic form of carbon like diamond, graphite or
fullerenes. CNTs are composed entirely of sp bonds, with each carbon atom joined to three
neighbours forming a macromolecule of carbon, atoms.
An ideal nanotube can be thought of as a sheet of graphite (a hexagonal lattice of carbon)
rolled into a seamless cylinder (Figure 2-2). With a diameter varying from 1 to 5 nm8, the
cylinder can have typically length of about several microns and each end can be capped with a
hemisphere of fullerene-like structure. This nanostructure presents therefore a high aspect
ratio (length-to-diameter), which allows considering CNT as one-dimensional object.
Figure 2- 2 : Formation of a nanotube by a graphene rolling up 9
CNTs can be divided in two categories, single-walled nanotubes (SWNTs), which are
constituted by the fundamental cylinder structure, and multi-walled nanotubes (MWNTs)
consisting of multiple rolled graphite layers (Figure 2-3). The structure of MWNTs can be
described by the Russian Doll model where carbon sheets are arranged in different coaxial
tubes (number of tubes can vary from 2 to 50), or by the Parchment model, a single sheet is
rolled in around itself like a scroll of parchment 10. The interlayer distance in MWNTs is close
to 3.3, similar to graphene layers in graphite.
Figure 2- 3 : Schematic representation of a single-wall (a) and a multi-wall carbon nanotube (Russian Doll model) (b)
(from www.lps.u-psud.fr/IMG/pdf_coursUVSQ2008.pdf, 2009-05)
The structure of an individual carbon nanotube can be specified in terms of a vector, the chiral
vector Ch that joins two crystallographically equivalent sites on the original graphene sheet 11
(Figure 2-4). When this hexagonal lattice is rolled up, the two end-points of the vector are
superposed to form the cylinder. The vector Ch can be described by the following relation:
h = 1 + 2
(II.1)
where a1 and a2 are unit vectors of the hexagonal lattice and n and m are translational indices.
Each nanotube topology is usually characterized by these two integer numbers (n,m) leading
to a primary symmetry classification of nanotubes. A chiral nanotube is defined by a nanotube
whose mirror image cannot be superposed to the original one. On the contrary, the structure of
an achiral nanotube is undistinguishable from the mirror image one. There are only two cases
of achiral nanotubes: armchair (n,n) and zig-zag (n,0) nanotubes (Figure 2-5). The nanotube
structure can be also defined by the chiral angle between the vectors Ch and a1, with values
of in the range 0-30 due to the hexagonal symmetry of the lattice. The two limit cases
correspond to achiral nanotubes with values of 0 and 30 respectively for zig-zag and
armchair classes.
Figure 2- 4 : Schematic diagram showing how a hexagonal sheet of graphite is rolled to form a carbon nanotube 12
Figure 2- 5 : Illustrations of the atomic structure of armchair, zig-zag and chiral nanotubes
(from www.cobweb.ecn.purdue.edu, 2009-06)
In the multi-walled carbon nanotubes (Russian doll model), each individual concentric
tube can have different chirality.
This primary symmetric classification of CNTs is very important because variations in the
nanotube morphology can lead to changes in the properties of the nanotube. For example, the
structure of a CNT determines its electronic behavior. Indeed, the electronic properties of a
nanotube follow a general rule: if (n-m) is a multiple of 3, then the tube exhibits a metallic
behavior, otherwise it is an electrical semiconductor 13. So, armchair nanotubes present a
metallic behavior while zig-zag and chiral nanotubes can be either metallic or semiconducting.
Investigations of the influence of chirality on the mechanical properties of CNTs have also
been reported 14. Although the chirality has a relatively small influence on the elastic stiffness,
the Stone-Wales (Figure 2-6) transformation plays a key-role in the nanotube plastic
deformation under tension. This transformation corresponds to a reversible diatomic
interchange where the resulting structure is built upon two pentagons and two heptagons in
pairs.
Figure 2- 6 : The Stone-Wales transformation occurring in an armchair nanotube under axial tension11
The Stone-Wales transformation introduces a new defect in the nanotube structure, the
heptagon. Heptagons allow for concave areas within the nanotube. Thus, the heptagonal
defects in nanotubes can result in many possible equilibrium shapes. Indeed, most nanotubes
are not straight cylinders with hemispherical caps.
II.A.2. Carbon nanotubes: Methods of production

The morphology and properties of carbon nanotubes are closely related to their
method of production. The first developed production methods were the arc discharge and
laser vaporization processes. However, majority of works has been devoted to chemical vapor
deposition (CVD) techniques since they can be readily scaled up for industrial production.
The arc discharge method consists to generate an electric arc discharge between two
graphite electrodes under an inert gas atmosphere (Ar, He) 15. This method requires very high
temperature (>5000C) and produces a mixture of different structures, including fullerenes,
onion-like carbon particles and some graphite sheets. It is necessary to separate the carbon
nanotubes from the soot and catalytic metals present in the crude product. Although it is
possible to selectively grow SWNTs or MWNTs, carbon nanotubes produced by this method
are normally entangled with poor control over the length and diameter 16.
In the laser ablation method, a graphite target is vaporized by laser irradiation under
flowing inert gas atmosphere at high temperature (near 1200C)11. Laser ablation has become
the common method for producing SWNTs 17, the nanotubes produced are very pure.
However, as for the arc discharge technique, the quantities of the produced sample are limited
by the size of the used carbon source. In addition, subsequent purification steps are necessary
to separate the nanotubes from undesirable by-products.
These limitations have motivated the development of gas-phase techniques, such as
chemical vapor deposition (CVD) 18,19 (Figure 2-7), where CNTs are formed by the
decomposition of a hydrocarbon-containing gas onto a supported catalyst (e.g., Fe/Co on
alumina). The synthesis of CNTs is often thermally or plasma enhanced. This gas-phase
technique is a continuous process which allows to produce high purified nanotubes,
minimizing subsequent purification steps, with limited control over the structure and
diameter. The CVD technique is suitable for a large-scale industrial production of nanotubes
and can be also used to form vertically aligned carbon nanotubes. The MWNTs used in this
work are industrially produced by CVD and supplied by Nanocyl S.A.
Figure 2- 7 : Scheme of a CVD reactor: hydrocarbon gas is decomposed over a metal catalyst
in a quartz tube heated by a furnace at 600-1000C
(from http://www.ifw-dresden.de, 2009-06)
High pressure conversion of carbon monoxide (HiPCO) technique is considered as

an improved CVD process. This production method consists of the gas-phase growth of
SWNTs with carbon monoxide as a carbon source at high temperature and pression 20.
These different methods are still under development. There are numerous variations of these
techniques operating under different conditions, with different set-up, and process parameters.
For example, a group of researchers have recently developed an effective method for growing
SWNTs via a metal catalyst-free CVD process on a sputtering deposited SiO2 film 21. Each
technique provides diverse advantages and disadvantages over the quality and nature of the
so-recovered CNTs.
In the reported synthesis methods, by-products such as amorphous carbon, carbon
nanoparticles and catalytic metal particles are often present in the CNT powder. If the
10
quantity of these by-products is high, several purification steps can be necessary before an
acceptable purity level. Most common purification approaches are based on the difference in
resistance to thermal and/or chemical treatments between carbon nanotubes and by-products,
such as oxidation 22,23,24, acid treatment 25, annealing 26,27. Unfortunately, these purification
methods very often impair carbon nanotubes by the introduction of structural defects, the
nanotube tips opening and/or oxidation of the CNTs surface. A method to overcome this
drawback consists to physically separate the different nanostructures by ultrasonic vibrations,
however prolonged sonication has been reported to can introduce defects and can lead to the
destruction of CNT structure, in addition to the formation of amorphous carbon 28.
Alternative
non-destructive
chromatography
30,31
methods
such
as
microfiltration 29
or
size
exclusion
can be used on SWNTs but these techniques are prohibitively slow and
limit the quantity in produced CNTs.
II.A.3. Properties of carbon nanotubes

Carbon nanotubes have gained in interest as nanoscale materials due to their
exceptional physical properties such as their very high Youngs modulus, their ultimate
strength and their high electric and thermal conductivity. However, directly measuring the
properties of these nanoparticles is very difficult by conventional methods. Therefore, several
properties have been first evaluated using by mathematical modelling.
When CNTs were discovered, their structure immediately encouraged speculation
about their potential mechanical properties. Indeed, the sp2 carbon-carbon bond is one of the
strongest and, so, a fiber formed by these axially oriented covalent bonds would be an
extremely strong material. Theoretical investigations have led to estimated Youngs modulus
of CNTs in the range of 1 TPa 32. Different experimental methods have been developed to
measure the elastic properties of individual nanotubes. One method allows the determination
of the CNT stiffness by the observation of the amplitude of their intrinsic thermal vibrations
in a transmission electronic microscope (TEM). Average Youngs modulus values of 1.8
TPa 33 and 1.25 TPa 34 were obtained for MWNTs and SWNTs respectively. A second
technique consists in directly measuring the stiffness of MWNTs clamped at one extremity
using atomic force microscopy (AFM) 35 leading to an average value of 1.28 0.59 TPa. A
similar method by AFM, consisting now in clamping CNTs at both ends across a pore of an
11
alumina membrane, was used by Salvetat et al. 36. By this way, Youngs modulus values of
810 410 GPa for MWNTs have been measured.
The tensile strength of CNTs has also been estimated by different theoretical calculations, an
average value of 100 GPa was achieved 37,38,39. Experimental studies using AFM have allowed
to confirm this value ranging from 10 to 200 GPa as measured for the CNT tensile
strength 40,41, much higher them steel (~1 GPa). Taking into account a study carried out on 26
CVD-MWNTs, Barber et al have concluded that the CNT strengthening mechanism can be
associate to the interaction between the walls of the nanotubes 42. For illustration, a tensile
strength of 200 GPa for MWNT corresponds to the ability to endure a weight of 20 tons on a
cable with a cross-section of 1 mm. Carbon nanotubes thus represent the strongest and
stiffest materials in terms of tensile strength and elastic modulus. To highlight the nanotube
mechanical properties, Youngs modulus and tensile strength values of SWNTs, MWNTs and
other materials (stainless steel and Kevlar) are compared in Table 2-1.
Material
Youngs Modulus (TPa)
Tensile Strength (GPa)
Elongation at Break (%)
Stainless Steel 43
~ 0.2
0.3 - 1.5
10 - 30
Kevlar 44
~ 0.15
~ 2.8
~2
SWNT
~1
13 - 53
~ 16
MWNT
0.8 - 1.8
10 - 200
Table 2- 1 : Comparison of mechanical properties of different materials
The above mechanical properties of CNTs refer to axial properties of the nanotube.
Simple geometrical considerations suggest that CNTs should be much softer in the radial
direction than along the tube axis. Indeed, TEM observations of radial elasticity show that
even van der Waals forces can deform two adjacent nanotubes 45. Nanoindentation
experiments with AFM indicated Youngs modulus of the order of several GPa 46 in the radial
axis. In spite of these excellent mechanical properties, carbon nanotubes possess also a
remarkable flexibility. Besides their experimental observations, Iijima et al 47 simulated the
deformation properties of nanotubes bent to large angles. The bending is completely
reversible up to angles in excess of 110, despite the formation of complex kink shapes.
Figure 2-8 shows experimental and theoretical results, demonstrating the exceptional
resilience of CNTs at large strain.
12
Figure 2- 8 : HR-TEM micrograph and computer simulation of nanotube bend47
Carbon nanotubes are expected to be very good thermal conductors along the tube, but
good insulators laterally to the tube axis. It was demonstrated that thermal conductivity of
CNTs is determined primarily by phonons 48. Experimental thermal conductivity measured on
individual MWNTs reaches 3300 W/(m.K) 49, which is about twice as high as diamond.
Moreover, CNTs are thermally stable up to 2800C in vacuum and up to 700C under air11.
Other important physical properties of CNTs are their electronic properties, which are
dependent on the molecular structure (chirality) of CNTs as explained previously in paragraph
II.A.1. It was demonstrated that (n,m) carbon nanotubes are metallic when n-m is equal to a
multiple of 3, others are semi-conducting12.
The determination of electronic properties of MWNTs is a complicated task as the carbon
sheets in the MWNTs can have different electronic character and chirality. However,
Bachtold et al. 50 demonstrated that current flows only through the outer-shell of MWNTs
based on measurements of the Aharonove-Bohm oscillation. In studies of MWNTs with outer
shell side-bonded to metal electrodes, it was also concluded that the electrical transport is
dominated by outer-shell conduction 51.
The diameter being in the nanometer range gives rise to quantum effects in carbon nanotubes.
In fact, the unique electrical properties of nanotubes arise from the confinement of the
electrons in tubes, which allows motion only along the tube axis. The scattering processes are
therefore reduced in the tube and metallic CNTs can carry enormous current densities up to
13
109A/cm without being destroyed 52. This density is about 2-3 orders of magnitude higher
than the current density of metals such as copper. A high electrical conductivity value of 103
S/cm is often measured by both two- and four-probe processes on individual SWNTs 53 and
MWNTs 54,55,56,57.
II.A.4. A large range of applications

The exceptional properties of carbon nanotubes can be potentially used in many
applications ranging from nanodevices to macroscopic materials. First commercial
applications have very recently appeared, including, e.g. motor vehicle fuel system
components and specialized sport equipments. The world demand is expected to reach $200
million end of 2009 and by 2020 will likely approach $9 billion 58. Figure 2-9 depicts the
predicted world nanotube demand by market.
Figure 2- 9 : Scheme of world demand in nanotubes by market in 2009 (redrawn from ref. 57)
One of the first proposed applications for carbon nanotubes was the use of MWNTs as
reinforcement or as electrically conductive components in polymer composite materials 59.
The combination of their unique properties and their light mass allows to consider nanotubes
as ideal fillers to form composite materials with improved mechanical strength and electrical
conductivity. The use of CNTs in this field of application is discussed with more details in the
section II.D.
With the highest current density as compared to any metal, carbon nanotubes are ideal
candidates for the next generation of nanoelectronic applications. An electronic circuit
constituted of wires and intramolecular junctions fully based on CNTs can be suggested.
14
These components can not only connect different CNTs for integration, but can also act as
functional building blocks in the circuit, such as rectifiers, switches, amplifiers, etc 60. It was
shown that carbon nanotube-based field effect transistors (FETs) have excellent operating
characteristics that are as good as or better than currently used silicon devices 61. However, a
drawback for electronic applications is the SWNT band gap dependance on its chirality. For
an optimal device performance, it is necessary to be able to grow a specific type of tube at a
defined position and direction of alignment, with near 100% yield. Control of growth position
and alignment can be achieved now, however the control of chirality still an issue. It has been
reported that design of the catalyst in CVD process can be used in the control of the tube
chirality 62,63.
Carbon nanotubes have emerged as a promising class of electron field emitters,
because they possess a low threshold electric field for emission and they are stable at high
current density. A cathode made with aligned CNTs can be used to shoot electrons under
vacuum to the targeted anode. Depending on its nature, the electrode can produce visible light
or X-ray
64
. The small size of these electron guns enables miniaturisation of elements in
display technologies (Field Emission Display or FED) and instrumentation based on X-ray.
Carbon nanotubes with superior field emission properties can be used for the development of
the next generation flat-panel display TV based on field emission 65 (Figure 2-10).
Figure 2- 10 : CNTs-based nano-emissive display (NED) prototype

(Image: Motorola Labs)
Carbon nanotubes present also some advantages for sensing applications. Their small
size with very large surface, high sensitivity via defects, fast response and good reversibility
at room temperature enable them to be used as a resistive chemical sensor 66. Actually, the
conductivity of CNTs changes when they are exposed to certain gaseous chemical species
15
(NO2, NH3) 67. It has been reported that the sensor selectivity can be tuned by grafting onto the
CNTs chemical species that will selectivety interact. One promising example is DNAdecorated carbon nanotube transistors acting as chemical sensors 68, wherein attached singlestrand DNA (ss-DNA) acts as a chemical recognition site and nanotubes as electronic read-out
components. Response of these systems differs in sign and magnitude for different gases and
odours and can be tuned by choosing the base sequence of the ss-DNA. For example, SWNTs
coated with ss-DNA have been used for sensing of a low concentration of dopamine in the
presence
of
excess
ascorbic
acid 69.
CO2
detection
using
SWNTs
and
microelectromechanical system has been demonstrated 70 and H2 sensing has been carried out
at room temperature using SWNTs films 71.
Techniques for the attachment of MWNTs at the tips of atomic force microscopy (AFM)
cantilevers have been developed for their use as scanning probes 72. These modified tips
present a number of advantages, including crash-proof operation and the capability of imaging
deep structures inaccessible to conventional tips. Moreover, the extremity of nanotube
modified tips can be functionalized with distinct chemical groups and used to demonstrate
nanoscale imaging with the ability to discriminate local chemistry of the surface being
probed 73.
Carbon nanotubes are now considered for energy storage and production because of
their small dimensions, smooth surface topology and perfect surface specificity since only the
graphite planes are exposed in their structure 74. It has been claimed that CNTs could represent
ideal candidates for the storage of hydrogen due to their large surface coupled with tubular
structure suggesting important capillarity effects. A storage capacity near 6.5% by weight, if
readily cyclable, is required for automotive applications. Theoretical estimates suggest a
storage capacity of 8-10wt%, but it turns out that experimental values are less than 1wt% 75.
Thus, it seems that a consensus has been reached now to say that CNTs are not a useful
hydrogen storage medium 76. However, some studies demonstrate that Ti coating could
increase the capacity storage of nanotubes up to 10wt% 77.
Another major area of research on CNTs concerns biomedical materials and devices.
Many applications for CNTs have been proposed including biosensors (examples of DNA
decorated CNTs), drug and vaccine delivery vehicles and novel biomaterials 78. For example,
SWNTs can be used as DNA transporters and near-infrared (NIR) photothermal agents 79.
Actually, folic acid functionalized carbon nanotubes can selectively kill cancer cells because
16
of the NIR light heating effect of SWNTs. The transporting capabilities of CNTs combined
with suitable functionalization chemistry and their intrinsic optical properties can lead to
novel nanomaterials for drug delivery and cancer therapy.
However, many uncertain factors may limit the biomedical applications of CNTs like their
aggregation and their questionable biocompatibility. It has been reported 80 that CNTs possess
some degree of toxicity (in vitro and in vivo), predominately due to the presence of residual
transition metal catalysts. Pristine CNTs has been shown to have minimal cytotoxicity at high
concentration, while functionalized CNTs are not toxic 81. These initial results urge caution
when handling CNTs and the introduction of safety measures in manufacturing facilities and
laboratories should be seriously considered.
As discussed before, carbon nanotubes could be used in a large range of diversified
applications, however it remains several drawbacks to be overcome such as their tendency to
aggregate. Indeed, CNTs form long bundles thermodynamically stabilized by numerous -
interactions between sidewalls. One solution consists to directly design carbon nanotubes
during their synthesis depending on the targeted application. For instance, the development of
techniques for the controlled growth of CNTs aligned on suitable substrates has received
special attention as well-defined CNTs architectures are important for applications such as
electronic displays or electron field emitters 82. Figure 2-11 shows an example of very well
defined vertically aligned CNTs using nanolithography. This example demonstrates the high
control reached on the growth of CNTs aligned on a flat surface.
Figure 2- 11 : Each Nanobama face is made of 150 million vertically aligned CNTs and measures 0.5mm 83
17
For other applications requiring non-aligned carbon nanotubes, the functionalization of

CNTs represents a solution to disaggregate them. The functionalization consists to graft
functional groups or polymer chains to CNTs surface. The different ways approached so far to
functionalize carbon nanotubes are presented in the next section.
II.B. Functionalization of carbon nanotubes

II.B.1. Interests and types of functionalization
As previously discussed, the CNTs unique properties make them desirable for many
different applications. However, to exploit as much as possible these properties, most of the
applications require the functionalization of carbon nanotubes, such as changing the surface
properties to make nanotubes soluble in different media, or attaching functional groups or
polymer chains for specific utilizations of modified nanotubes.
The CNT functionalization methods can be divided into two major groups. The first
one is the functionalization from inside (endohedral functionalization) meaning that CNTs are
treated by filling their inner empty cavity with different molecules 84 or nanoparticles 85
(Figure 2-12).
a
Figure 2- 12 : Schematic representation (a) and HRTEM image (b) of a SWNT filled with C60 fullerenes
(from www.fz-juelich.de/iff/datapool/iee/TEM.jpg)
The introduction of these particles, for instance Buckminsterfullerene C60, can be achieved by
spontaneous penetration of nanoparticles or by wet chemistry, where nanotubes are filled with
some compounds, which react under particular thermal or chemical conditions and produce
inside the nanoparticles.
18
The second group is the exohedral functionalization, which consists to graft molecules
on the outer surface of nanotubes. Several approaches have been developed and include defect
functionalization (Figure 2-13A), covalent functionalization (Figure 2-13B) and noncovalent
functionalization with surfactants (Figure 2-13C) or polymers (Figure 2-13D) 86.
CNT
Figure 2- 13 : Functionalization possibilities for CNTs: defect functionalization (A),

covalent sidewall functionalization (B), noncovalent functionalization with surfactants (C)
and polymer wrapping (D)86
The different types of exohedral functionalization can be classified via the nature of the
interactions between the surface of carbon nanotubes and the functional groups or polymer
chains. These interactions can rely upon covalent or non-covalent bonds.
II.B.2. Non-covalent functionalization with surfactant or polymer

In comparison to a covalent bond, the energy released in the formation of a noncovalent bond is lower. However, in supramolecular systems like carbon nanotubes, the
number of these bonds can be high and become dominant. The noncovalent interaction is
based on van der Waals forces or - stacking and it is controlled by thermodynamics.
The great advantage of this type of functionalization relies upon the possibility of attaching
various groups without disturbing the electronic system of the rolled graphene sheets of
CNTs 87.
19
The formation of non-covalent aggregates with surfactants is a suitable method for dispersing
individual nanotubes in aqueous 88 or organic solvents 89. To solubilize CNTs in aqueous
media, the used surfactant is an amphiphilic molecule whose the hydrophobic part is oriented
toward the surface of CNTs, whereas the polar moiety interacts with solvent molecules 90.
Moore et al. 91 studied various aqueous dispersions of CNTs with anionic, cationic and nonionic surfactants. They show that the ability of surfactant to disperse individual CNTs is
linked to the size of the hydrophilic groups due to an enhanced steric stabilization.
Another example of non-covalent functionalization is the anchoring through strong -
stacking interactions of polynuclear aromatic compounds to the side-walls of CNTs, such as
planar pyrene moiety 92. (Figure 2-14).
Figure 2- 14 : Interaction of nanotubes with pyrene derivatives 93
Carbon nanotubes can be also wrapped with polymer chains to form supramolecular
complexes of CNTs. These structures can be formed by solution mixing of nanotubes with
polymer chains. For example, the wrapping of SWNTs with polymer bearing polar sidechains, such as polyvinylpyrrolidone (PVP) or polystyrenesulfonate (PSS), leads to the
suspension of SWNT/polymer complexes in water 94. In 2003, this method was used to wrap
SWNTs with DNA oligonucleotide sequences to improve separation of metallic versus
semiconducting nanotubes 95.
Another way to get these nanotubes/polymer complexes is the in situ polymerization directly
in presence of CNTs. The polymerization-filling technique (PFT) allows to produce carbon
nanotubes coated with polyethylene 96. First, methylaluminoxane (MAO), a commonly used
co-catalyst in metallocene-based olefin polymerization, is anchored on the nanotubes surface.
Then, the metallocene catalyst (e.g., Cp2*ZrCl2) is added. Finally, the polymerization of
ethylene is started, polyethylene is formed near the CNT surface and, with increasing
molecular mass, precipitates onto nanotubes. The bundles of CNTs are destructed along with
20
the formation of the polyethylene coating on nanotube surface. The different steps of this
technique are represented in figure 2-15.
Figure 2- 15 : Schematic representation of the polymerization-filling technique (PFT) applied to CNTs96
Although the non-covalent functionalization allows the preservation of the electronic structure
of CNTs, the weakness of interface between nanotubes and physisorbed polymer chains is a
limitation for applications involving mechanical reinforcement 97.
II.B.3. Covalent functionalization

Two major groups of chemical functionalization of CNTs via covalent attachment can
be distinguished, the end and defect-group chemistry and the sidewall functionalization.
II.B.3.a. End and defect-side chemistry
The functionalization via end and defect-side chemistry consists to graft functional
group directly on the already existing defects in the structure of CNTs. Indeed, carbon
nanotubes are generally described as perfect graphite sheets rolled into nanocylinders. In
reality, all CNTs present defects and can be curved. Typically, around 1-3% of carbon atoms
of a nanotube are located at a defect site 98.
21
The end caps of nanotubes can be composed of highly curved fullerene-like hemispheres,
which are highly reactive as compared to the sidewalls 99. The sidewalls themselves contain
defect sites such pentagon-heptagon pairs called Stone-Wales defects (see previously
discussed Figure 2-6), sp3-hybridized defects and vacancies in the nanotube lattice 100. The
most frequently encountered type of defect is the so-called Stone-Wales (or 7-5-5-7) defect
which leads to a local deformation of the nanotube curvature. Addition reactions are most
favored at the carbon-carbon double bonds in these positions. The different typical defects are
showed in the Figure 2-1686.
Figure 2- 16 : Typical defects in a SWNT
Five- or seven-membered rings in the carbon framework, instead of the normal sixmembered rings, leading to a bend in the tube (Figure 2-16, arrow A). The sp3-hybridized
defects is indicated with the letter B (R=H and OH). Carbon framework can be damaged by
oxidative conditions, which leaves COOH groups on the edge of a hole (Figure 2-16,C).
Figure 2-16D shows open end of the SWNT terminated with COOH groups. Other
terminal groups, such as OH, -H and =O, are also possible.
Frequently, these intrinsic defects are supplemented by oxidative damages to the
nanotube framework by strong acids during the purification step. These acids leave vacancies
functionalized with oxygenated functional groups such as carboxylic acid, ketone, alcohol and
ester groups 101. The oxidative introduced carboxyl groups represent useful sites for further
modifications in organic solvents such as the coupling of molecules through the creation of
22
amide and ester bonds. Figure 2-17 represents common functionalization routes used to
derivatize SWNTs at ends and defect sites through solution-based chemistry.
By this way, nanotubes can be flanked with a wide range of functional moieties, for which
bifunctional molecules are often utilized as linkers. This method was used to graft amine
moieties onto carbon nanotubes via the reaction with diamines such as triethylenetetramine 102,
ethylenediamine 103 or 1,6-hexamethylenediamine 104. Another approach involves reduction of
the carboxyl groups to hydroxyls, followed by transformation into amino groups via
phthalimide coupling and hydrolysis 105. Gromov et al. have developed two other approaches,
using amino-decarboxylation substitution, to get primary amino groups directly attached on
the carbon nanotube surface 106. The carboxylic groups, in majority at CNTs open ends, are
replaced with amino groups via Hofmann rearrangement of carboxylic acid amide or via
Curtius reaction of carboxylic acid chloride with sodium azide.
Figure 2- 17 : Schematic of common functionalization ways at ends and defect sites 107
23
One advantage of the defect-side chemistry is that functionalized carbon nanotubes retain
their pristine electronic and mechanical properties 108. However, the repartition of grafted
functions is not homogenous along CNTs. Indeed, the carboxylic groups, used as anchoring
sites, are placed in majority at the extremities of carbon nanotubes.
II.B.3.b. Sidewall functionalization
The second group of covalent grafting reactions is the sidewall functionalization, which
consists to graft chemical groups through reactions onto the -conjugated skeleton of CNTs.
First covalent sidewall functionalization was carried out on the basis of well-developed
addition chemistry on fullerenes. However, it is predicted that the sidewall addition chemistry
of CNTs differs from that of fullerenes, even though both are curved, conjugated carbon
systems 109. Indeed, the chemical reactivity in strained carbon systems arises from two factors,
the pyramidalization at the carbon atoms and the -orbital misalignment between adjacent
carbon atoms 110. In fullerenes, the former is the most influent factor due to their pronounced
curvature. The relief of pyramidalization strain energy results in energetically favorable
addition reactions onto fullerene structures. In carbon nanotubes, the pyramidalization strain
is not as acute and -orbital misalignment is expected to have a greater influence 111. This
misalignment, associated with bonds at an angle to the tube circumference (bonds that are not
perpendicular or parallel to the tube axis), is the origin of torsional strain in CNTs. The relief
of this strain controls the extent of reactivity of carbon nanotubes. These two factors
influencing the reactivity of CNTs are represented in Figure 2-18. We can observe that the CC bond in CNTs structure will be more or less reactive in function of its angle in to the tube
circumference.
The strain in CNTs arises thus from these factors which scale inversely with tube
diameter86, thinner CNTs are more reactive than larger tubes. However, the reactivity of CNT
sidewalls remains low and sidewall-functionalization is only successful if a highly reactive
reagent is used, whereas the nanotube caps are quite reactive due to their fullerene-like
structure. Another constraint for sidewall functionalization is the tendency of CNTs to form
bundles and to limit the available nanotube surface for the grafting of chemical reagents.
24
Figure 2- 18 : Pyramidalization angles (p) and the -orbital misalignment angles () along C1-C4 in the (5,5) SWNT
and its capping fullerene C60 112
A large majority of covalent sidewall functionalizations is carried out in organic

solvent, which allows the utilization of sonication process to improve the dispersion of CNTs
and, thus, the available surface of carbon nanotubes. However, precipitation immediately
occurs when this process is interrupted 113. The required reactive species such as carbenes,
nitrenes or radicals, are in general made available through thermally activated reactions 114.
However, the addition mechanisms are not understood completely yet. Normally, the addition
reaction could be initiated exclusively on the intact sidewall or in parallel at defect sites. Most
common sidewall functionalizations carried out in organic solvents, such as carbene 115 or
nitrene 116 [2+1] cycloaddition reactions or radical additions via diazonium salts 117 for
instance, are represented in Figure 2-19.
25
Pristine
CNTs
a'
Figure 2- 19 : Schematic describing various common covalent sidewall functionalization reactions of CNTs
using organic solvents107
The first studied sidewall functionalization is the fluorination of CNTs. Carbon

nanotubes have been fluorinated by fluorine in the range between room temperature and
600C 118,119,120. This reaction is very useful because further nucleophilic substitutions can be
easily accomplished and open a flexible approach for providing the CNT sidewalls with
26
various types of functional groups 121. Some examples of these reactions are illustrated in the
lower part of Figure 2-19a.
Among these reactions, several diamines are reported to react with fluoronanotubes via
nucleophilic substitution reactions 122, leading to the formation of amino-functionalized CNTs
(Figure 2-19a). However, by using a bifunctional reagent such as diamines with sufficiently
long carbon chain, the nanotubes can be covalently cross-linked with each other. Modification
with amino-containing substituents was developed by the use of radical process of photolysis
of acetonitrile 123. The 1,3-dipolar cycloaddition reaction can be also used to graft linkers, with
amino groups at their ends, uniformly distributed around sidewalls 124.
However, carbon nanotube functionalizations carried out in organic solvent present
some inconvenients. The first one is the necessity to agitate carbon nanotubes by sonication
process to get a good dispersion of them in organic solvent. Indeed, intensive or too long
sonication causes severe damages to the tube walls and the average length of CNTs can be
reduced 125. Secondly, this functionalization inorganic solvent requires generally multistep
time-consuming reactions. Finally, the use of organic solvent is polluting and difficult to
upgrade for industrial processes.
To overcome the drawbacks of the wet chemistry functionalization, dry processes
are being developed. The ball-milling of MWNTs in reactive atmospheres was shown to
functionalize nanotubes with different chemical groups, such as amides, carbonyls, thiols and
mercaptans for instance, depending on the used atmosphere 126. Using this method, CNTs are
broken and radicals formed on their newly created extremities can then react with the
introduced gas. A relatively long time (min. 24h) is necessary to get an enough quantity of
functions onto CNTs. However, the average length of CNTs decreases for increasing
treatment and amorphous carbon is observed after the breaking process 127. Another drawback
is the presence of functional groups exclusively on the ends of cup-stacked carbon nanotubes.
An alternative dry approach to chemical modification of CNTs uses plasma
discharges. Generalities on plasma processes and their applications in CNT functionalization
are presented in the next chapter.
27
II.C. Plasma treatment of carbon nanotubes

II.C.1. Introduction to plasma: the fourth state of matter
Plasma is physically defined as an ionized gas with an equal number of positively
and negatively charged particles. It consists of free electrons, ions, radicals, UV-radiation and
various highly excited neutral and charged species 128. The entire plasma is electrically neutral
but its behavior is complex because the particle motions are controlled by electric and
magnetic fields. In 1879, Sir William Crookes firstly defined this state of ionized gas as the
fourth state of matter after solid, liquid and gas (in order of increasing energy). The nature of
the cathode ray matter was in the Crookes tube subsequently identified by Sir J. J. Thomson
in 1897 129 and called plasma by Irving Langmuir when he studied the electrified gases in
vacuum tubes in 1928 130. It is often considered that 99% of the matter in the universe is in the
plasma state. This estimate may not be very accurate, but it is certainly a reasonable one in
view of the fact that stellar interiors and atmospheres, gaseous nebulae and much of the
interstellar medium are plasmas.
According to gas temperature, plasma can be classified into thermal (or equilibrium)
plasma, which is characterized by being fully ionized (gas temperature Tg = electron
temperature Te) or cold or (non-equilibrium) plasma with the gas only partially ionized
(TgTe). Thermal plasma implies that the temperature of all active species (electrons, ions
and neutrals) is the same. This is, for example, true for stars as well as for fusion plasma.
High temperatures are required to form these equilibrium plasmas, typically ranging from
4000K to 20000K 131.
In the non-thermal or cold plasmas, the temperature of neutral and positively charged
species is low, while the electrons are in much higher temperature because they are light and
easily accelerated by the applied electromagnetic fields. As a result the generated plasma is in
the state of non-equilibrium and the reaction may proceed at low temperature. Indeed, the
electrons can reach temperatures of 104 105 K (1-10 eV) while the temperature of the gas
can be as low as room temperature 132. Figure 2-20 represents the wide range of plasma types
in function of their electron densities and their temperatures.
28
Figure 2- 20 : Diverse types of plasma represented in function of electron density and temperature 133
II.C.2. Characteristics and principal applications of plasma

The electrically induced discharge in gas is the most common method for plasma
ignition. Direct current (DC) discharges, pulsed DC discharges, radio frequency (RF)
discharges (13.56 MHz) and microwave discharges (2.45 GHz) represent the plasma
categorization based on electric apparatus 134. These common standard frequencies were
chosen in order to avoid interfering with telecommunication. The basic feature of a variety of
electrical discharges is that they produce plasmas in which the majority of the electrical
energy primarily goes into production of energetic electrons, instead of heating the entire gas
stream 135. These energetic electrons induce ionization, excitation and molecular fragmentation
process of the background gas molecules to produce excited species which create a
chemically-richenvironment. Due to their essential role, the electrons are therefore
considered to be the primary agents in the plasma.
29
Plasma processes provide a cost effective and environmentally friendly alternative to

many important industrial processes because the method produces no waste products and in
most cases exposes operators to no significant hazards. Plasmas find thus well-established use
in industrial applications, however they are also gaining more interest in the field of life
sciences, related to environmental issues 136 and biomedical applications 137,138. Display
technologies, which are based on microdischarges, have been developed and offer the
possibility of large lightweight, flat TV monitors 139. Another application field of plasmas is in
analytical spectrochemistry for the trace analysis of solids, liquids and gases 140. Atomic
lasers 141, ozone generation 142, light systems (by example fluorescence lamps) 143 and particle
sources 144 are also developed.
Gas discharge plasmas can be used in a large variety of applications requiring surface
modification. Plasma processing is generally used for film deposition 145,146, etching 147 and
may also be used for resist development and removal 148. Surface modification by plasmas
plays a crucial role in the microelectronics industry, for the microfabrication of integrated
circuits130, and in materials technology 149.
All plasmas that are created by the injection of microwave power, electromagnetic
radiation in the frequency range of 300 MHz to 10 GHz, can in principle be called
microwave induced plasmas (MIPs) 150. This is, however, a general term that comprises
several different plasma types, such as cavity induced plasma, atmospheric plasma torches,
etc. A striking feature of MIPs is the wide range of operational conditions that can be applied:
dependent on the plasma source, power levels can range from a few Watts up to several
hundreds of kiloWatts, the discharge pressure might range from less than 10-2 Pa up to several
times atmospheric pressure, whereas many different discharge gases might be used (both
noble and molecular gases). A result of this broad range of operational conditions is that
plasmas with widely varying plasma parameters (such as electron density, electron
temperature, gas temperature, ionization degree and chemical composition) can be created.
Moreover, the geometry of the plasmas can be largely influenced. Because practical
applications as well as laboratory research usually need plasmas with specific characteristics
and a controllable geometry, MIPs are in many cases preferable to other plasmas.
Among these plasmas, surface wave discharges (SWDs) can be used when highdensity plasmas have to be created. SWDs are sustained by waves that are conducted parallel
to the surface of the dielectric walls surrounding the plasma. The most common SWD
configuration is the surfatron, but the waveguide used in this work is a surfaguide 151. The
microwave induced plasmas present a higher density of electron in comparison of other type
30
of plasmas such as RF plasmas by instance, because electron are more easily created with
microwaves (at same power) and they present a more important density.
Figure 2- 21 : Ar + N2 surface microwave discharge
The high electron density of the MIPs is a key parameter for the creation of atomic
nitrogen in Ar+N2 microwave plasma Godfroid et al. have demonstrated that production of
atomic nitrogen in Ar+N2 SWD is achieved by two electronic mechanisms 152. The first is the
dissociation by direct electronic impact, which is achieved by collision between a nitrogen
molecule and an electron:
4
4
+ 2 +
g , n=0 + ( ) + ( )
(II.2)
this reaction leads to the production of atomic nitrogen in the fundamental state N(4S).
The second one is called dissociative recombination and is achieved by the recombination
between an electron and an ionized nitrogen molecule:
+ 2+ (4) + (4)
(II.3)
Finally, the set-up and parameters of the microwave discharge have been studied to get the
higher nitrogen dissociation rate 153. The most influential parameters proved to be the power
supply pulse, the dilution of N2 in Ar and the total gas flow.
31
II.C.3. Applications of plasma into materials surface functionalization

Compared with other chemical modification methods, the plasma induced
functionalization present interesting properties and is known as is an environmentally
friendly, solvent-free and time efficient process. Moreover, this treatment allows the grafting
of a wide range of different functional groups depending on plasma parameters such as power,
gas used, duration of treatment and pressure. Moreover, this method provides the possibility
of scaling up to produce large quantities necessary for commercial use. Plasma treatment has
been widely used for surface activation of various materials, ranging from organic polymers
to ceramics and metals. In this domain, we can cite the polymer surface functionalization in
pulsed and continuous nitrogen microwave plasma, under admixture of hydrogen 154. This
process allows grafting nitrogen functional groups on polystyrene with a high selectivity in
primary amine groups.
The surface modification of carbon nanotubes can be also carried out through a wide
range of plasma processes. In addition to the already cited advantages of plasma treatment,
the amount of functional groups can also be tailored. This is important since having saturation
of these groups on the surface can alter the electronic conductivity of nanotube material. For
instance, the resistivity of CNTs increases with the degree of fluorination 155. The most
common plasma treatment of CNTs is the low pressure RF cold plasma which is successfully
used to bind oxygen155,156,157, hydrogen 158 and fluorine groups 159,160.
It has been observed that a complete purification of carbon nanotubes powder can be
reached out after CNTs treatment in glow discharges (RF or MW). Indeed, amorphous carbon
domains are eliminated and the impurities are removed by ion bombardment and irradiation in
an O2 RF plasma 161. However, it was also found that the average diameter of MWNTs
decreases with treatment duration in three steps. At first, ion bombardment causes the creation
of vacancies and interstitials in MWNTs, the superficial structure, including produced
amorphous carbon, are thus lost. Secondly, the ion beams continue to react with amorphous
carbon until they are peeled totally from the CNTs. The last step is the oxidation of
amorphous carbon under O2 plasma irradiation in CO2, inducing the decrease of the MWNTs
average diameter. The nature of the plasma gas is important because it is known that
oxygenated ions or radicals are very reactive species in etching processes. However,
destruction of carbon nanotube sidewalls is also observed for other less reactive plasma gas
such as CF4 162 or Ar 163. Moreover, it was shown that UV photons promote the
defunctionalization of moieties grafted on CNTs 164. Figure 2-22 represents high resolution
32
transmission electronic microscopy images of MWNTs before and after Ar plasma discharge.
We can observe that ion bombardment and irradiation in Ar plasma deforms the graphite
layers and destructs the MWNTs. A solution consists to reduce the power supplied to the
plasma or the duration of the treatment to limit destruction of sidewalls. However, these limits
are very restricting to get a suitable and controlled functionalization of carbon nanotubes.
Figure 2- 22 : HRTEM images of MWNTs (a) before and (b) after Ar microwave plasma treatment for 3 minutes162
Another solution is to avoid the functionalization of CNTs directly inside the plasma
where the density of high energetic ions is very high and, instead place the CNTs outside the
discharge production zone to reduce the detrimental effects associated with ion bombardment
and irradiation. In this approach, radicals become most important reactive species to graft
functional moieties onto CNTs surface. As aforementioned, the application of microwave
discharge sources for the production of intense beams of atomic, radical and metastable
species is well established. For example, carbon nanotubes have been hydrogenated by atomic
hydrogen generated in a H2 microwave plasma163. In this type of functionalization, the exact
location of the sample and its distance from the plasma discharge zone become a key-factor in
determining the extent of functionalization. Khare et al. have studied the exposition of
SWNTs to microwave-generated N2 plasma, by placing CNTs at different distances from
discharge. 165 At the shortest distance (1cm), they observed the highest concentration in
nitrogen groups but, also, a loss of integrity of SWNTs due to highly reactive species, like
N2+. At intermediate distances, the incorporation of nitrogen, but also of high quantity of
oxygen, onto the CNTs is obtained while functionalization was not observed for the maximal
distance of 7 cm due to total recombination of atomic nitrogen. It is thus important to place
33
CNTs outside the glow discharge zone at a minimal distance to avoid interaction with highly
reactive ions and UV photons. With this type of treatment, radicals become the most
important reactive species to graft functional groups onto carbon nanotubes. The efficiency of
this approach depends on the extent of radicals production, in our case of atomic nitrogen.
Within the scope of this work, the functionalization of carbon nanotubes will be studied by
this approach with microwave plasma allowing a high efficiency of atomic nitrogen
production139 .
II.D. Polymer nanocomposites-based on carbon nanotubes

II.D.1. Generalities on polymer nanocomposites
A nanocomposite is defined as a composite material where at least one of the
dimensions of one of its constituents is on the nanometer size scale 166. This term implies also
the combination of two (or more) distinct materials, such as a ceramic and a polymer. The
challenge and interest in the development of nanocomposite is to find ways to create
macroscopic components that benefit from the unique properties of the very small objects
dispersed within them. Natural materials, such as bone or wood, are good examples of the
successful implementation of this concept, offering excellent mechanical properties compared
to those of their separated constituents.
The particularity of nanofillers is that they have at least one of their dimensions in
the same range than the radius of gyration of typical polymer chains. Moreover, when particle
size decreases to nanometer size, the specific surface area of fillers drastically increases and,
so, significantly increases the resulting interfacial volumes. These regions can have distinctly
different properties from the bulk polymer and can represent substantial volume fraction of
the matrix for nanoparticles with surface areas in the order of hundreds of m/g, like carbon
nanotubes (Figure 2-23). The essential of the entire polymer matrix in a nanocomposite can
behave as if it was part of interface with specific properties. Another key-advantage is that
the improvement of the matrix properties can be reached at low nanofiller contents in
comparison to more conventional microfiller-based composites. These considerations on the
nanoparticle specificities explain why these fillers can have remarkably high impact on the
final properties of nanocomposites 167.
34
Particle size
Volume content
Number of particles
Interface
Aspect ratio
10m
30%
~106
~0.1m/g
~20
1m
30%
~1010
~1m/g
~100
10nm
3%
~1015
~100m/g
~1000
Figure 2- 23 : Influence of particle characteristics on various geometric factors 168
Carbon nanotubes (CNTs) are considered as ideal candidates to prepare polymer

nanocomposites presenting high-performance properties due to the combination of their
remarkable physical properties with their low density. The resulting CNTs-based materials
can be substantially stronger and lighter than conventional carbon fiber-reinforced polymer
composites currently used in many applications 169. The incorporation of CNTs also provide
electrical conductivity to insulating polymer matrix 170, delay thermal degradation of the
polymer matrix and act as efficient flame retardant additives 171.
Nanocomposite properties depend on several factors: the polymer nature, the
synthetic process of CNTs, the nanotube purification process (if any), the amount and type of
impurities on nanotubes or their aspect ratio. However, the key-factor remains the quality of
CNTs dispersion within the host polymer matrix 172,173,174. The variations in nanotube and
nanotube/polymer properties account for many of the apparent inconsistencies in the
literature. However, these variations are difficult to quantify. In this work, we have performed
our studies using CNTs from the same batch, thereby reducing the variability between
samples and clarifying overall trends.
Several methods have been used to disperse carbon nanotubes in polymer
matrices
175
. The most straightforward method is the direct melt blending. Two other methods
are the solution-mediated processes and in-situ polymerization however these techniques are
less versatile and more environmentally contentious.
35
II.D.2. Melt blending of polymer matrices with carbon nanotubes

The melt blending approach is a very suitable technique for industrial scale processes
because of its high speed and simplicity 176. The strategy consists in blending a molten
thermoplastic with carbon nanotubes in order to optimize the polymer-nanotube interactions
and to form a nanocomposite. Amorphous polymers, such as polystyrene (PS), can be
processed above their glass transition temperature while semi-crystalline polymers, such as
polyethylene (PE), need to be heated above their melt temperature to induce sufficient
softening. In general, melt processing involves the melting of polymer to form a viscous
liquid. Any additives, such as carbon nanotubes, can be mixed into the melt by shear mixing.
Bulk samples can then be fabricated by techniques such as compression molding, injection
molding or extrusion.
The optimization of processing conditions is very important because CNTs can affect melt
properties such as viscosity, resulting in unexpected polymer degradation because of
excessive shear rates 177. Successful examples of melt blended compositions include
poly(methyl methacrylate)/MWNTs 178, polycarbonate/MWNTs 179, polypropylene/SWNTs 180,
nylon-6/MWNTs 181,182, polypropylene/SWNTs 183 or polyimide/SWNTs composites 184. All
these examples show some interesting effects in terms of mechanical reinforcement,
crystallization behavior or electrical and thermal percolation thresholds but these effects are
always dependent on the studied matrix. Most of time the melt blending method does not give
very good dispersion of the nanofillers in the majority of polymer matrices because the
aggregates of CNTs are highly stabilized and the compatibility between the polymer chains
and carbon nanotubes is low.
As discussed before, carbon nanotubes have important surface properties, which
affect their dispersion within a polymer matrix 185. Indeed, many van der Waals interactions
(essentially - interactions) between CNTs hold them together in bundles of 50 to few
hundred individual nanotubes packed in a hexagonal array (Figure 2-24a). The average
diameter of a bundle is typically between 10 and 100 nm. Unfortunately, CNTs keep their
tendency to aggregate in large majority of polymer host matrices (Figure 2-24b). These
entanglements and bundles lead to many defect sites in the composites and limit the efficiency
of properties transfer from nanotubes to polymer composites 186. Baughman et al. proved that
bundling results in diminished mechanical and electrical properties as compared to theoretical
predictions based on the dispersions of individual SWNTs58.
36
200nm
Figure 2- 24 : Transmission electron microscopy (TEM) images of (a) a SWNTs rope (axial view) 187 and
(b) bundle-like aggregate of MWNTs subsisting in PE matrix after melt blending
The challenge is thus to incorporate individual CNTs, or at least relatively thin

nanotubes bundles, inside the polymer matrix. Different techniques are proposed in the
literature to overcome this drawback. One of them is the modification of CNTs sidewalls via a
pre-treatment of CNTs.
II.D.3. Interfacial compatibilization: improvement of CNTs dispersion

This second approach implies a modification of the CNTs by sidewall
functionalization. In addition to pre-disaggregation of CNTs bundles, the interfacial
compatibilization also improves the interactions polymer-filler107. Indeed, due to the relatively
smooth grapheme-like surface of nanotubes, there is a lack of interfacial adhesion between the
polymer matrix and CNTs. A crucial parameter for the production of this type of carbon
nanotube-based nanocomposites is the control of the degree of grafting/functionalization on
the CNTs walls. For example, it was shown by molecular dynamics that a very high number
of functional groups covalently attached onto the CNT surface can decrease the maximum
buckling force of nanotubes by about 15% and therefore reduce the mechanical properties of
the final nanotube-based composites 188. Moreover, the degree of covalent functionalization
should not be too high in order not to significantly disturb the electron system of nanotubes,
and thus affect the electrical properties, but it should be high enough to provide good
compatibility between the filler and polymer matrix. However, several researchers have
37
reported that functionalization can improve the electrical properties of nanocomposites 189. It
appears that the disadvantages of functionalization with respect to CNTs conductivity can be
outweighed by the improved dispersion 190 and some grafted functional groups can have a
doping effect if the linked heteroatom is an electron donor as nitrogen for instance 191. It is
thus important to evaluate the content and to determine the exact location of grafted
functional groups on the CNTs surface in order to control the carbon nanotube integrity and to
avoid intensive lost of properties.
Covalent functionalization can be performed by either defect-side chemistry, modification of
surface-bound
reactive
groups,
e.g.,
carboxylic
acid
groups
or
direct
sidewall
functionalization. Some pertinent examples are exposed hereafter.

Hydroxyl-functionalized carbon nanotubes have been used to reinforce poly(vinyl alcohol)
(PVA) 192. The idea is that the OH groups from SWNTs form hydrogen bonds with the OH
of the PVA allowing a twofold increase of the modulus by the addition of a tiny amount of
nanotubes (as low as 0.8 wt%) and a strength increase from 74 to 107 MPa is also observed.
Another example is the fluorinated SWNTs (F-SWNTs) dispersed in a PE matrix 193. It was
demonstrated that, in comparison with polyethylene composites filled with purified
nanotubes, improved dispersion and enhanced interfacial / mechanical properties are achieved
for F-SWNT-based composites. A very good dispersion of amino-functionalized doublewalled carbon nanotubes (DWNTs) in an epoxy resin by a standard calendaring technique has
also been observed and compared to theoretical predictions 194. Investigation of the
mechanical properties resulted in an increase of strength, Youngs modulus and strain to
failure at a nanotube content of only 0.1wt%. On Figure 2-25, the difference in term of
nanofillers dispersion quality between untreated DWNTs-based and amino-functionalized
DWNTs nanocomposites appears clearly.
38
Figure 2- 25 : TEM photographs of the dispersion of (a) DWNTs and (b) amino-functionalized DWNTs
in an epoxy matrix 195
Several authors have also studied the effect of the chemical functionalization on
electrical properties of polymer nanocomposites-based on carbon nanotubes.
Valentini et al. 196 pointed out that the amino-functionalized SWNT in epoxy matrix allows
migration of intrinsic charges, which contribute to the overall conductivity. Tamburri et al.184
found that extensive functionalization of CNTs with OH and COOH groups enhances the
current in a conducting polymer (1,8-diaminophtalene-based) by factors of 90 and 140,
respectively, whereas the untreated nanotubes showed an enhancement by only 20.
In 2005, Kim et al. 197 measured the electrical conductivity of epoxy composites filled with
MWNTs oxidized under both strong and mild conditions. The authors showed that the amount
of damage has much influence on the electrical properties of the MWNTs, causing the
electrical conductivity to be lowered and the percolation threshold to be raised. On the other
hand, the oxidation of MWNTs under mild conditions endows the MWNTs composite with
high conductivity and a relatively low percolation threshold. These results show the
importance to control the oxidative conditions, in which no damage is caused to the nanotubes
structure and no impurities are generated, allowing the production of CNTs-based
nanocomposites with better electrical properties.
Polyimide nanocomposites filled with unmodified, acid-modified and amino-modified (acid
treatment followed by reaction with ethylenediamine) MWNTs were also prepared and
characterized 198. TEM analysis revealed that both types of functionalized carbon nanotubes
were dispersed uniformly in the polyimide matrix. The measure of volume electrical
39
resistivity of each polyimide nanocomposite is represented in Figure 2-26. The

polyimide\MWNTs nanocomposite present much better electrical properties (conductivity and
percolation threshold) compared to other (nano)composites. This observation can be
explained by the acid treatment used in both functionalization approaches, which creates
damages on the CNTs surface and reduce the average length of carbon nanotubes. These
nanotubes are easier to disperse in the polyimide matrix but their electrical conductivity is
reduced. The percolation threshold of the functionalized MWNTs is increased due to decrease
of nanotube length.
Figure 2- 26 : Influence of acid and amine modification of MWNTs on volume electrical resistivity
of polyimide nanocomposites198
The above examples highlight the importance of the control over the covalent
functionalization of carbon nanotubes to obtain CNTs nanocomposites with higher best
electrical properties.
The majority of functionalized CNTs used as nanofillers are polymer-grafted
nanotubes because the use of long polymer chains, in comparison to small functional groups,
allows for a better pre-disaggregation of CNTs ropes. Moreover, an optimal interfacial
interaction can be created when the grafted polymer chains are miscible (or display good
compatibility) into the host polymer matrix. Two main strategies for the polymer covalent
40
grafting to the nanotube surface exist: the grafting from and the grafting onto methods
(Figure 2-27).
Grafted
functional
groups
Anchored
polymerization
initiators
Grafting onto
Grafting from
End-functionalized
polymer chains
Monomer solution
Polymer-grafted carbon nanotube

Figure 2- 27 : Sketch representing the grafting onto and grafting from methods
The grafting from method is based on the initial immobilization of initiators onto
the nanotube surface, followed by the in-situ polymerization of appropriate monomers with
the formation of the polymer chains chemically attached bound to the nanotube 199. The
advantage of this technique is that the nanotubes with high grafting density can be prepared.
This technique has been widely used for the preparation of PMMA / CNTs nanocomposites /
nanohybrids (Figure 2-28).
41
Figure 2- 28 : General strategy for grafting polymers onto carbon nanotubes via ATRP initiated
from the MWNT surface 200
More precisely, methyl methacrylate (MMA) monomer is polymerized on the

functionalized MWNTs surface via a (latent) free- radical initiator. Authors observed that
MWNTs were more uniformly dispersed in the PMMA matrix than in a composite simply
prepared by melt blending and reported that successful dispersion of MWNTs resulted from
the CNTs surface grafting of PMMA 201. The preparation of poly(n-butyl methacrylate)
grafted SWNTs by atom transfer radical polymerization (ATRP) of n-butyl methacrylate
using 2-bromopropionate groups covalently immobilized SWNT as initiator has been also
reported 202. The authors of this paper also synthesized polystyrene grafted SWNTs with the
same anchored initiator approach 203 and observed in both cases an improvement of nanotube
dispersion
whatever
the
polymer
matrix.
Nanotube
functionalization
with
4-hydroxymethylaniline (HMA) via the diazonium salt method has been carried out on
SWNTs. The-OH groups that are generated from the functionalization were used to initiate
the ring-opening polymerization (ROP) of -caprolactone (-CL) to form PCL grafted CNTs
nanohybrids 204 (Figure 2-29). The PCL grafting proved to allow a dramatic increase of carbon
nanotubes suspension in chloroform.
42
Figure 2- 29 : Two-step strategy to yield PCL-g-SWNTs nanohybrids202
Saeed et al. 205 characterized PCL/MWNTs nanocomposites yielded by in-situ ring

opening polymerization (ROP) of -caprolactone (-CL) in presence of unmodified MWNTs
(P-MWNTs) and acid-modified MWNTs (A-MWNTs). The preparation of these
nanocomposites is schematized in Figure 2-30. SEM micrographs of fractured surface of both
nanocomposites show that the A-MWNTs were better dispersed than P-MWNTs in PCL
matrix. Percolation thresholds of both nanocomposites properties were found at 2 wt% of
MWNTs as studied by rheological investigations. However, the electrical conductivity of the
PCL/P-MWNTs nanocomposite was between 10-1 and 10-2 S/cm in presence of 2 wt% of
MWNTs although the PCL/A-MWNTs conductivity reached 10-2 S/cm at a filling level as
high as 7 wt% of A-MWNTs. The electrical conductivity of PCL nanocomposites filled with
unmodified MWNTs was higher than that of PCL/A-MWNTs over the entire range of
nanotube loading. Such an observation has been explained by the destruction of -electron
pathway network of MWNTs by the acid treatment.
43
Figure 2- 30 : Preparation of PCL/MWNTs nanocomposites by ROP of -CL in the presence of either

unmodified MWNTs (P-MWNTs) or from acid treated MWNTs (A-MWNTs)205
The grafting onto approach is based on the attachment of already preformed endfunctionalized polymer chains onto functional groups available at the nanotube surface. An
advantage of this method is that preformed (commercial) polymers of controlled mass (and
mass distribution) can be used. However, the main limitation of this technique is that the
fixation of the first polymer chains can sterically prevent diffusion of additional
macromolecules onto the surface usually leading to low grafting density. Also only polymers
containing reactive functional group(s) can be considered.
One of the first examples of grafting to approach was published by Fu et al. 206, where
carboxylic acid groups on the nanotube surface were converted to acyl chlorides by refluxing
the samples in thionyl chloride. Then, the acyl chloride functions were reacted with hydroxyl
groups of PEG chains via esterification reaction.
Another original polymer grafting relies upon the use of either polystyrene, PCL or their
copolymers, end-capped by a 2,2,6,6,-tetramethylpiperidine-1-oxyle (TEMPO) group, which
can react with MWNTs according to a free-radical mechanism (Figure 2-31)
207
. Other
alkosyamine end-groups can be used as well like PS-NO.
44
Figure 2- 31 : Grafting of PS, PCL and copolymer chains on MWNTs via a controlled radical mechanism 208
It comes out from all these studies that covalent functionalization by different
molecules or polymer chains onto carbon nanotube surface can provide an efficient approach
for yielding homogeneous dispersion of CNTs in the polymer matrix and excellent interfacial
stress transfer between nanotube and polymer. All discussed examples show that a good
compatibility between CNTs and polymer matrix is needed to substantially enhance the
physical properties of the final nanocomposites. Furthermore, the covalent CNTs surface
treatment / functionalization has to been under control to preserve the nanotubes properties
and allow their transfer to polymer matrix.
II.D.4. Carbon nanotube-based nanocomposites: applications

Even if the market is still in the early stage of product and process development, the
current production of CNT-filled polymer composites is estimated at ca. 80.000 tons 209. A
part of CNT-based polymer composites exploit carbon nanotubes as a conductive dispersed
45
phase in an insulating matrix for applications ranging from electronics to automotive and
aerospace sectors.
The first use of conductive CNT-filled composite was by Hyperion for electrostatic
dissipation applying paints onto car components 210. Another use of these nanocomposites is as
antistatic shielding on airplane wings and fuselages for example. A future market targets the
electromagnetic interference (EMI) shielding 211, a critical application for many industrial
sectors. The required composite conductivity of 1 S/cm can be achieved. Nanocomposites of
PCL filled with MWNTs were foamed by supercritical CO2 in order to prepare materials with
reduced electromagnetic interference 212. Interestingly enough, such carbon nanotube-filled
nanocomposites represent very promising EMI shielding materials because their performances
result from absorption at low filler content and not from reflection.
A fourth type of application of conducting composites is the production of transparent
conductors. There is a huge market for transparent electronic conductors such as indium tin
oxide (ITO) in displays. In this field, there is a drive towards flexible displays on plastic
substrates. However ITO is brittle and has poor adhesion to plastic. Conducting composites of
SWNTs can be transparent if thin enough and the quality of nanotube dispersion high enough
to reduced the required content in CNTs 213. Carbon nanotube-based nanocomposites can also
be used in multilayer printed circuits 214. The mechanical properties of CNTfillers/nanocomposites can be also exploited fibers, by aligning CNTs. One most promising
example consists to draw and spin a nanotube fiber directly from a CVD reaction chamber
(Figure 2-32) 215. The resulting fiber can then be post-impregnated with epoxy to produce the
expected composite. These fibers presently have remarkable failure strengths as high as 1
GPa.
Figure 2- 32 : Scanning electron micrograph of a spun CNTs fiber215
46
Carbon nanotubes are effective fire-retardant additives for polymers at quite low loading level
of a few weight percents in CNTs. This specific property has been recently used by
Kabelwerk Eupen, in collaboration with Nanocyl S.A., to produce the first electrical cable
prototype shielded by a fire-proofed nanocomposite layer based on carbon nanotubes 216.
During the combustion of the nanocomposite layer, an insulating char layer is formed acting
as a thermal shield and therefore protecting the inner material against melting or electrical
contacts. This last application appears very promising since the wire and cable industry will
need to replace the use of harmful halogen flame retardants soon.
CNTs can be also incorporated in antifouling paints based on silicone resins 217. In contrast
with the current coatings used for marine applications, the anti-fouling properties are not
resulting from the continuous release of biocides, e.g. copper or tin-based biocides, but are
explained by the nanostructured surface of the CNTs / PMDS coatings of the CNTs.
Even if carbon nanotubes present exceptional physical properties, the number of
applications based on CNT-filled polymer nanocomposites remains low. This is largely
explained by the difficulty to transfer CNTs properties to the host polymer matrix due to poor
filler dispersion and lack of interfacial compatibility in the majority of materials. Sidewall
modification of CNTs is a good solution to overcome these drawbacks. However up to now, it
is difficult to get optimal nanocomposite properties without damaging their structure resulting
in some lost of their intrinsic properties and those of the final material. It is thus important to
develop a suitable technique allowing for the controlled functionalization of carbon nanotube
surface and preserving their intrinsic properties. The grafting of polymer chains from these
functional groups can then lead to pre-disaggregated CNT bundles with improved interactions
between carbon nanotubes and the host polymer matrix.
47
References
1
Radushkevish L.V., Lukyanovich V.M., Zurn. Fisic. Chim. 1952, 26, 88-95
Hillert M., Lange L., Z Kristallogr 1958, 111, 24-34
3
Baker R.T.K., Harris P.S., Thomas R.B., Waite R.J., J Catal 1973, 30, 86-95
4
Oberlin A., Endo M., Koyama T., J Cryst Growth 1976, 32, 335-349
5
Kroto H.W., Heath R. J., OBrien S. C., Curl R.F., Smalley R.E., Nature 1985, 318, 162
6
Iijima S., Nature 1991, 354, 56-58
7
Iijima S., Ichihashi T., Nature 1993, 363, 603-605
8
Bethune D.S., Klang C.H., de Vries M.S., Gorman G., Savoy R., Vazquez J., Beyers R., Nature 1993, 363,
605-607
9
http://hal.archives-ouvertes.fr/docs/00/34/77/13/PDF/These_dan_elgrabli_2008.pdf
10
Amelinckx S., Lucas A., Lambin P., Rep. Prog. Phys. 1999, 62, 1471
11
Belin T., Epron F., Mater. Sci. Engng 2005, B119, 105-118
12
Thostenson E.T., Ren Z., Chou T.-W., Compos. Sci. Technol. 2001, 61, 1899-1912
13
Saito R., Fujita M., Dresselhaus G., Dresselhaus M., Appl. Phys. Lett. 1992, 60, 2204-2206
14
Yakobson B.I., Brabec C.J., Bernholc J., Phys. Rev. Lett. 1996, 76(14), 25112514
15
Journet C., Bernier P., Appl. Phys. A-Mater. Sci. & Proces. 1998, 67(1), 1-9
16
Journet C., Maser W.K., Bernier P., Loiseau A., delaChapelle M .L., Lefrant S., Deniard P., Lee R., Fischer
J.E., Nature 1997, 388 (6644), 756-758
17
Puretzky A.A., Geohegan D.B., Fan X., Pennycook S.J., Appl. Phys. Lett. 2000, 76(2), 182-184
18
Endo M., Takeuchi K., Kobori K., Takahashi K., Kroto H.W., Sarkar A., Carbon 1995, 33, 873-881
19
Ivanov V., Fonseca A., Nagy J.B., Lucas A., Lambin P., Bernaerts D., Zhang X.B., Carbon 1995, 33, 17271738
20
Nikolaev P., Bronikowski M.J., Bradley R.K., Rohmund F., Colbert D.T., Smith K.A., Smalley R.E., Chem.
Phys. Lett. 1999, 313, 91-97
21
Liu B., Ren W., Gao L., Li S., Pei S., Liu C., Jiang C., Cheng H.-M., J. Am. Chem. Soc. 2009, 131(6), 20822083
22
Borowiak-Palen E., Pichler T., Liu X., Knupfer M., Graff A., Jost O., Pompe W., Kalenczuk R.J., Fink J.,
Chem. Phys. Lett. 2002, 363, 567-572
23
Huang S.M., Dai L.M., J. Phys. Chem. B 2002, 106, 3543-3545
24
Hou P.X., Liu C., Tong Y., Liu M., Cheng H.M., J. Mater. Res. 2001, 16, 2526-2529
25
Farkas E., Anderson M.E., Chen Z.H., Rinzler A.G., Chem. Phys. Lett. 2002, 363, 111-116
26
Chiang I.W., Brinson B.E., Smalley R.E., Margrave J.L., Hauge R.H., J. Phys. Chem.B 2001, 105, 1157-1161
27
Georgakilas V., Voulgaris D., Vazquez E., Prato M., Guldi D.M., Kukovecz A., Kuzmany H., J. Am. Chem.
Soc. 2002, 124 (48), 14318-14319
28
Girifalco L.A., Hodak M., Lee R.S., Phys. Rev. B 2000, 62, 13104-13110
29
Bandow S., Rao A.M., Williams K.A., Thess A., Smalley R.E., Eklund P.C., J. Phys. Chem. B 1997, 101,
8839-8842
30
Niyogi S., Hu H., Hamon M.A., Bhowmik P., Zhao B., Rozenzhak S.M., Chen J., Itkis M.E., Meier M.S.,
Haddon R.C., J. Am. Chem. Soc. 2001, 123 (4), 733-734
31
Zhao B., Hu H., Niyogi S., Itkis M.E. Hamon M.A., Bhowmik P., Meier M.S., Haddon R.C., J. Am. Chem.
Soc. 2001, 123 (47), 11673-11677
32
Lu J.P., J Phys Chem Solids 1997, 58(11), 1649-1652
33
Treacy M.M.J., Ebbesen T.W., Gibson J.M., Nature 1996, 381 (6584), 678-680
34
Krishnan A., Dujardin E., Ebbesen T.W., Yianilos P.N., Treacy M.M.J., Phys. Rev. B 1998, 58(20), 1401314019
35
Wong E.W., Sheehan P.E., Lieber C.M., Science 1997, 277, 1971-1975
36
Salvetat J.-P., Kulik A.J., Bonard J.-M., Briggs G.A.D., Stockli T., Metenier K., Bonnamy S., Beguin F.,
Burnham N.A., Forro L., Adv. Mater. 1999, 11, 161-165
37
Xiao J.R., Gama B.A., Gillepsie J.W., Int. J. Solid. Str. 2005, 42, 3075-3092
2
48
38
Sun X., Zhao W., Mater Sci. Engng. A. 2005, 390, 366-371
Belytschko T., Xiao S.P., Schatz G.C., Ruoff R., Phys. Rev. B 2002, 65 (23), p. 235430
40
Popov V.N., Mater. Sci. Engng. 2004, R 43, 61-102
41
Yu M.F., Files B.S., Arepalli S., Ruoff R.S., Phys. Rev. Lett. 2000, 84, 5552-5555
42
Barber A., Andrews R., Schadler L.S., Wagner H.D., Appl. Phys. Lett. 2005, 87, p.203106
43
http://www.engineershandbook.com/Tables/steelprop.htm (February 2009)
44
Wagner H.D., Encycl. Polym. Sci. Technol. 2002, DOI: 10.1002/0471440264.pst317
45
Ruoff R.S., Tersoff J., Lorents D.C., Subramoney S., Chan B., Nature 1993, 364, 514 - 516
46
Yu M.-F., Kowalewski T., and Ruoff R.S., Phys. Rev. Lett. 2000, 85 (7), 1456-1459
47
Iijima S., Brabec C., Maiti A., Bernholc J., J. Chem. Phys. 1996, 104, 2089-2092
48
Hone J., Whitney M., Zettl A., Synth. Metals 1999, 103, 2498-2499
49
Kim P., Shi L., Majumbar A., McEuen P.L., Phys. Rev. Lett. 2001, 87 (21), p.215502
50
Bachtold A., Strunk C., Salvetat J.-P., Bonard J.-M., Forro L., Nussbaumer T., Schnenberger C., Nature
1999, 397, 673-675
51
Schenenberger C., Bachtold A., Strunk C., Salvetat J.-P., Forro L., Appl. Phys. A 1999, 69, 283-295
52
Frank S., Poncharal P., Wang Z.L., deHeer W.A., Science 1998, 280, 1744-1746
53
Tans S.J., Devoret M.H., Dai H., Thess A., Smalley R., Geerlings L.J., Dekker C., Nature 1997 , 386, 474-477
54
Dai H., Wong E.W., Lieber C.M., Science 1996, 272, 523
55
Ebbesen T.W., Lezec H.J., Hiura H., Bennett J.W., Ghaemi H.F., Thio T., Nature 1996, 382 , 54-56
56
Van Haesendonck C., Stockman L., Vullers R.J.M., Bruynseraede Y., Langer L., Bayot V., Grivei E., Issi J.P., Heremans J.P., Olk C.H., Surf. Sci. 1997, 386, 279-289
57
Ando Y., Zhao X., Shimoyama H., Sakai G., Kaneto K., Int. J. Inorg. Mater. 1999, 1, 77-82
58
http://www.freedoniagroup.com/brochure/20xx/2019smwe.pdf
59
Baughman R.H., Zakhidov A.A., de Heer W.A., Science 2002, 297, 787-792
60
Wei D.C., Liu Y.Q., Adv. Mater. 2008, 20, 2815-2841
61
Avouris P., Acc. Chem. Res. 2002, 35, 1026-1034
62
Bachilo S.M., Balzano L., Herrerra J.E., Pompeo F., Resasco D.E., Weisman R.B., J. Am. Chem. Soc. 2003,
125, 11186-11187
63
Li Y., Mann D., Rolandi M., Kim W., Ural A., Hung S., Javey A., Cao J., Wang D., Yenilmez E., Wang Q.,
Gibbons J.F., Nishi Y., Dai H., Nano Lett. 2004, 4, 317-321
64
Cheng Y., Zhou O., C.R. Physique 2003, 4, 1021-1033
65
Saito Y., Uemura S., Carbon 2000, 38, 169-182
66
Tang Z.K., Zhang L., Wang N., Zhang X.X., Wen G.H., Li G.D., Wang J.N., Chan C.T., Sheng P., Science
2001, 292, 2462-2465
67
Kong J., Franklin N.R., Zhou C.W., Chapline M.G., Peng S., Cho K.J., Dai H.J., Science 2000, 287, 622-625
68
Staii C., Johnson A.T.Jr, Chen M., Gelperin A., Nano Lett. 2005, 5, 1774-1778
69
Hu C.G., Zhang Y.Y., Bao G., Zhang Y.L., Liu M.L., Wang Z.L., J. Phys. Chem. B 2005, 109, 20072-20076
70
Zribi A., Knobloch A., Rao R., Appl. Phys. Lett. 2005, 86, 203112
71
Oakley J.S., Wang H.-T., Kang B.S., Wu S., Ren F., Rinzler A.G., Pearton S.J., Nanotechnology 2005, 16,
2218-2221
72
Dai H., Hafner J.H., Rinzler A.G., Colbert D.T., Smalley R.E., Nature 1996, 384, 147-149
73
Wong S.S., Joselevich E., Woolley A.T., Cheung C.L., Lieber C.M., Nature 1998, 394, 52-54
74
Ajayan P.M., Zhou O.Z., Carbon nanotubes 2001, 80, 391-425
75
Kajiura H., Tsutsui S., Kadono K., Kakuta M., Ata M., Murakami Y., Appl. Phys. Lett. 2003, 82, 1105-1107
76
Hirscher M., Becher M., Haluska M., von Zeppelin F., Chen X., Dettlaff-Weglikowska U., Roth S., J. Alloys
Compd. 2003, 356-357, 433-437
77
Yildirim T., Ciraci S., Phys. Rev. Lett. 2005, 94, 175501
78
Lin Y., Taylor S., Li H.P., Fernando K.A.S., Qu L.W., Wang W., Gu L.R., Zhou B., Sun Y.P., J. Mater.
Chem. 2004, 14, 527-541
79
Kam N.W.S., OConnell M., Wisdom J.A., Dai H.J., Proc. Natl. Acad. Sci. USA 2005, 102, 11600-11605
80
Smart S.K., Cassady A.I., Lu G.Q., Martin D.J., Carbon 2006, 44, 1034-1047
81
Gao J., Xu B., Nano Today 2009, 4, 37-51
39
49
82
Talapatra S., Kar S., Pal S.K., Vajtai R., Ci L., Victor P., Shaijumon M.M., Kaur S., Nalamasu O., Ajayan
P.M., Nature Nanotech. 2006, 1, 112-116
83
Hart J., Tawfick S., De Volder M., Walker W., www.nanobama.com, 2008
84
Dujardin E., Ebbesen T.W., Hiura H., Tanigaki K., Science 1994, 265, 1850-1852
85
Monthioux M., Carbon 2002, 40, 1809-1823
86
Hirsch A., Angew. Chem. Int. Ed. 2002, 41 (11), 1853-1859
87
Tasis D., Tagmatarchis N., Georgakilas V., Prato M., Chem. Eur. J. 2003, 4000-4008
88
Islam M.F., Rojas E., Bergey D.M., Johnson A.T., Yodh A.G., Nanolett. 2003, 3, 269-273
89
Ausman K. D., Piner R., Lourie O., Ruoff R. S., Korobov M., J. Phys. Chem. B 2000, 104 (38), 8911-8915
90
Bandow S., Rao A. M., Williams K. A., Thess A., Smalley R. E., Eklund P. C., J. Phys. Chem. B 1997, 101,
8839-8842
91
Moore V.C., Strano M. S., Haroz E. H., Hauge R. H., Smalley R. E., Schmidt J., Talmon Y., Nanolett. 2003, 3,
1379-1382
92
Chen R. J., Zhang Y., Wang D., Dai H., J. Am. Chem. Soc. 2001, 123, 3838-3839
93
Besteman K., Lee J.-O., Wiertz F. G. M., Heering H. A., Dekker C., Nanolett. 2003, 3, 727-730
94
OConnell M. J., Boul P., Ericson L. M., Huffman C., Wang Y., Haroz E., Kuper C., Tour J., Ausman K. D.,
Smalley R. E., Chem. Phys. Lett. 2001, 342, 265-271
95
Zheng M., Jagota A., Strano M. S., Santos A. P., Barone P., Chou S. G., Diner B. A., Dresselhaus M. S.,
Mclean R. S., Onoa G. B., Samsonidzc G. G., Semke E. D., Usrcy M., Walls D. J., Science 2003, 302, 15451548
96
Bonduel D., Mainil M., Alexandre M., Monteverde F., Dubois P., Chem. Commun. 2005, 781-783
97
Manias E., Polizos G., Heidecker M. J., Polymer Nanocomposites, TMS Editions Warrendale 2006
98
Hu H., Bhowmik P., Zhao B., Hamon M. A., Itkis M. E., Haddon R. C., Chem. Phys. Lett. 2001, 345, 25-28
99
Sinnott S. B., J. Nanosci. Nanotechnol. 2002, 2, 113-123
100
Charlier J.-C., Acc. Chem. Res. 2002, 35, 1063-1069
101
Chen J., Hamon M. A., Hu H., Chen Y., Rao A. M., Eklund P. C., Haddon R. C., Science 1998, 282, 95-98
102
Gojny F.H., Nastalczyk J., Roslaniec Z., Schulte K., Chem. Phys. Lett. 2003, 370, 820-824
103
Meng H., Sui G. X., Fang P. F., Yang R., Polym. 2008, 49, 610-620
104
Li J., Fang Z. P., Tong L. F., Gu A. J., Liu F., J. Appl. Polym. Sci. 2007, 106, 2898-2906
105
Ramanathan T., Fischer F. T., Ruoff R. S., Brinson L. C., Chem. Mater. 2005, 17, 1290-1295
106
Gromov A., Dittmer S., Svensson J., Nerushev O. A., Perez-Garcia S. A., Licea-Jimnez L., Rychwalski R.,
Campbell E. E. B., J. Mater. Chem. 2005, 15, 3334-3339
107
108
Zhang J., Zou H., Qing Q., Yang Y., Li Q., Liu Z., Guo X., Du Z., J. Phys. Chem. B 2003, 107, 3712-3718
109
Chen Z., Thiel W., Hirsch A., ChemPhysChem. 2003, 4, 93-97
110
Haddon R. C., Acc. Chem. Res. 1998, 21, 243-249
111
Hamon M. A., Itkis M. E., Niyogi S., Alvaraez T., Kuper C., Menon M., Haddon R. C., J. Am. Chem. Soc.
2001, 123, 11292-11293
112
Niyogi S., Hamon M. A., Hu H., Zhao B., Bhowmik P., Sen R., Itkis M. E., Haddon R. C., Acc. Chem. Res.
2002, 35, 1105-1113
113
Tasis D., Tagmatarchis N., Bianco A., Prato M., Chem. Rev. 2006, 106, 1105-1136
114
Balasubramanian K., Burghard M., Small 2005, 1, 180-192
115
Lee W. H., Kim S. J., Lee W. J., Lee J. G., Haddon R. C., Reucroft P. J., Appl. Surf. Sci. 2001, 181, 121-127
116
Holzinger M., Steinmetz J., Samaille D., Glerup M., Paillet M., Bernier P., Ley L., Graupner R., Carbon
2004, 42, 941-947
117
Bahr J. L., Yang J., Kosynkin D. V., Bronikowski M. J., Smalley R. E., Tour J. M., J. Am. Chem. Soc. 2001,
123, 6536-6542
118
Nakajima T., Kasamatsu S., Matsuo Y., Eur. J. Solid State Inorg. Chem. 1996, 33, 831-840
119
Hamwi A., Alvergnat H., Bonnamy S., Beguin F., Carbon 1997, 35, 723-728
120
Mickelson E. T., Huffman C. B., Rinzler A. G., Smalley R. E., Hauge R. H., Margrave J. L., Chem. Phys.
Lett. 1998, 296, 188-194
121
Khabashesku V. N., Billups W. E., Margrave J. L., Acc. Chem. Res. 2002, 35, 1087-1095
50
122
Stevens J. L., Huang A. Y., Peng H., Chiang I. W., Khabashesku V. N., Margrave J. L., Nano Lett. 2003, 3,
331-336
123
Nakamura T., Ohana T., Ishihara M., Hasegawa M., Koga Y., Diam. Rel. Mater. 2008, 17, 559-562
124
Pantarotto D., Charalambos D. P., Graff R., Hoebeke J., Briand J.-P., Prato M., Bianco A., J. Am. Chem. Soc.
2003, 125, 6160-6164
125
Monthioux M., Smith B. W., Burteaux B., Claye A., Fischer J. E., Luzzi D. E., Carbon 2001, 39, 1251-1272
126
Konya Z., Vesselenyi I., Niesz K., Kukovecz A., Demortier A., Fonseca A., Delhalle J., Mekhalif Z., Nagy J.
B., Koos A. A., Osvath Z., Kocsonya A., Biro L. P., Kiricsi I., Chem. Phys. Lett. 2002, 360, 429-435
127
Ma P. C., Tang B. Z., Kim J.-K., Chem. Phys. Lett. 2008, 458, 166-169
128
Chapman B., Glow Discharge Processes, Wiley Interscience Publication 1980
129
Thomson J. J., Philosophical Magazine 1897, 44, 293- http://web.lemoyne.edu/~GIUNTA/thomson1897.html
130
Langmuir I., Proc. Nat. Acad. Sci. US 1928, 14, 627-637
131
Bogaerts A., Neyts E., Gijbels R., van der Mullen, Spectrochimica Acta B 2002, 57, 609-658
132
Grill A., Cold Plasma in Materials Fabrication, IEEE Press 1994
133
www.bmbf.de/pub/plasma_technology.pdf
134
Denes F. S., Manolache S., Prog. Polym. Sci. 2004, 29, 815-885
135
Fridman A., Chirokov A., Gutsol A., J. Phys. D :Appl. Phys. 2005, 38, R1-R24
136
Hammer T., Contrib. Plasma Phys. 1999, 39, 441-462
137
Frauchiger V. M., Schlottig F., Gasser B., Textor M., Biomaterials 2004, 25, 593-606
138
Favia P., Sardella E., Gristina R., dAgostino R., Surf. Coat. Techn. 2003, 169-170, 707-711
139
Punset C., Boeuf J.-P., Pitchford L. C., J. Appl. Phys. 1998, 83, 1884-1897
140
Marcus R. K., Evans E. H., Caruso J. A., J. Anal. At. Spectrom. 2000, 15, 1-5
141
Thiem T. L., Microchemical J. 1994, 50, 214-219
142
Balakrishnan P. A., Arunagiri A., Rao P. G., J. Electrostatics 2002, 56, 77-86
143
Ravi J., Maya J., J. Appl. Phys. 2000, 87, 4107-4112
144
Nakashima Y., Kato T., Ishimoto Y., Orito K., Natori T., Fukasawa T., Watanabe K., Kobayashi S., Shoji M.,
Kubota Y., J. Nucl. Mater. 2003, 553, 313-316
145
Werbowy A., Zdunek K., Dusinski E., Szmidt J., Elert M., Surf. Coat. Techn. 2003, 174-175, 170-175
146
Valentini L., Armentano I., Kenny J. M., Lozzi L., Santucci S., Diam. Relat. Mater. 2003, 12, 821-827
147
Inoue S., Kajikawa K., Mat. Sci. Eng. B 2003, 103, 170-178
148
McOmber J. I., Nair R.S., Nucl. Instrum. Meth. Phys. Resear. B 1991, 55, 281-286
149
Hino T., Akiba M., Fus. Eng. Des. 2000, 49-50, 97-105
150
Timmermans E., Atomic and Molecular Excitation Processes in Microwave Induced Plasma, Ph. D. Thesis,
Eindhoven University of Technology, 1999
151
Moisan M., Zakrzewski, J. Phys. D : Appl. Phys. 1991, 24, 1025-1048
152
Godfroid T., Dauchot J.-P., Hecq M., Surf. Coat. Technol. 2005, 200, 649-654
153
Godfroid T., Dauchot J.-P., Hecq M., Surf. Coat. Technol. 2003, 174-175, 1276-1281
154
Meyer-Plath A. A., Finke B., Schrder K., Ohl A., Surf. Coat. Technol. 2003, 174-175, 877-881
155
Felten A., Bittencourt C., Pireaux J.-J., Van Lier G., Charlier J.-C., J. Appl. Phys. 2005, 98, 074308
156
Bubert H., Haiber S., Brandl W., Marginean G., Heintze M., Bruser V., Diam. Rel. Mater. 2003, 12, 811-815
157
Ionescu R., Espinosa E.H., Sotter E., Llobet E., Vilanova X., Correig X., Felten A., Bittencourt C., Van Lier
G., Charlier J.-C., Pireaux J. J., Sensors and Actuators B 2006, 113, 36-46
158
Zhang J., Feng T., Yu W., Liu X., Wang X., Li Q., Diam. Rel. Mater. 2004, 13, 54-59
159
Plank N. O. V., Jiang L., Cheung R., Appl. Phys. Lett. 2003, 83, 2426-2428
160
Valentini L., Puglia D., Armentano, Kenny J. M., Chem. Phys. Lett. 2005, 403, 385-389
161
Xu T., Yang J., Liu J., Fu Q., Appl. Surf. Sc. 2007, 253, 8945-8951
162
Kalita G., Adhikari S., Aryal H. R., Ghimre D. C., Afre R., Soga T., Sharon M., Umeno M., Physica E 2008,
41, 299-303
163
Qin Y., Hu M., Appl. Surf. Sc. 2008, 254, 1757-1762
164
Khare B. N., Meyyappan M., Cassell A. M., Nguyen C. V., Han J., Nano Lett. 2002, 2, 73-77
165
Khare B., Wilhite P., Tran B., Teixeira E., Fresquez K., Mvondo D. N., Bauschlicher C., Meyyappan M., J.
Phys. Chem. B 2005, 109, 23466-23472
51
166
Ajayan P. M., Redlich P., Rhle M., J. Microsc. 1997, 185, 275-282
Advani S. G., Processing and properties of nanocomposites, World Scientific Publishing Company 2007
168
Schulte K., Gojny F. H., Fiedler B., Sandler J. K. W., Bauhofer W., Polymer Composites From Nano- to
Macroscale, Springer US, 2005
169
Bal S., Samal S. S., Bull. Mater. Sci. 2007, 30, 379-386
170
Du F., Scogna R. C., Zhou W., Brand S., Fischer J. E., Winey K. I., Macromolecules 2004, 37, 9048-9055
171
Laoutid F., Bonnaud L., Alexandre M., Lopez-Cuesta J.-M., Dubois Ph., Mater. Sci. Eng. R 2009, 63, 100125
172
Grossiord N., Loos J., Regev O., Koning C. E., Chem. Mater. 2006, 18, 1089-1099
173
Xie X.-L., Mai Y.-W., Zhou X.-P., Mater. Sci. Eng. :R 2005, 49, 89-112
174
Szleifer I., Yerushalmi-Rozen R., Polymer 2005, 46, 7803-7818
175
Moniruzzaman M., Winey K. I., Macromolecules 2006, 39, 5194-5205
176
Andrews R., Jacques D., Minot M., Rantell T., Macromol. Mater. Eng. 2002, 287, 395-403
177
Potschke P., Bhattacharyya A. R., Janke A., Goering H., Compos. Interf. 2003, 10, 389-404
178
Jin Z., Pramoda K. P., Xu G., Goh S. H., Chem. Phys. Lett. 2001, 337, 43-47
179
Poetschke P., Bhattacharyya A. R., Janke A., Goering H., Compos. Interfaces 2003, 10, 389-404
180
Bhattacharyya, A. R.; Sreekumar, T. V.; Liu, T.; Kumar, S.; Ericson, L. M.; Hauge, R. H.; Smalley, R. E.
Polymer 2003, 44, 2373-2377
181
Liu T., Phang I. Y., Shen L., Chow S. Y., Zhang W. D., Macromolecules 2004, 37, 7214-7222
182
Zhang W. D., Shen L., Phang I. Y., Liu T., Macromolecules 2004, 37, 256-259
183
Bhattacharyya, A. R.; Sreekumar, T. V.; Liu, T.; Kumar, S.; Ericson, L. M.; Hauge, R. H.; Smalley, R. E.
Polymer 2003, 44, 2373-2377
184
Siochi, E. J.; Working, D. C.; Park, C.; Lillehei, P. T.; Rouse, J. H.; Topping, C. C.; Bhattacharyya, A. R.;
Kumar, S. Composites, Part B 2004, 35B, 439-446
185
Peeterbroeck S., Breugelmans L., Alexandre M., BNagy J., Viville P., Lazzaroni R., Dubois Ph., Compos. Sci.
Technol. 2007, 67, 1659-1665
186
Allaoui A., Bai S., Cheng H. M., Bai J. B., Comp. Sci. Technol. 2002, 62, 1993-1998
187
Thess A., Lee R., Nikolaev P., Dai H., Petit P., Robert J., Xu C., Lee Y. H., Kim S. G., Rinzler A. G., Colbert
D. T., Scuseria G. E., Tomanek D., Fischer J. E., Smalley R. E., Science 1996, 273, 483-488
188
Garg A., Sinnott S. B., Chem. Lett. Phys. 1998, 295, 273-278
189
Lau K. T., Hui D., Compos. Part B : Eng. 2002, 33, 263-277
190
Tamburri E., Orlanducci S., Terranova M. L., Valentini F., Palleschi G., Curulli A., Brunetti F., Passeri D.,
Alippi A., Rossi M., Carbon 2005, 43, 1213-1221
191
Terrones M., Ajayan P. M., Banhart F., Blase X., Carroll D. L., Charlier J. C., Czerw R., Foley B., Grobert
N., Kamalakaran R., Kohler-Redlich P., Rhle M., Seeger T., Terrones H., Appl. Phys. A 2002, 74, 355-361
192
Liu L., Barber A. H., Nuriel S., Wagner H. D., Adv. Funct. Mater. 2005, 15, 975-980
193
Schofner M.L., Khabashesku V.N., Barrera E.V., Chem Mater. 2006, 18, 906-913
194
Gojny F.H., Wichmann M.H.G., Kpke U., Fiedler B., Schulte K., Compos. Sci. Technol. 2004, 64, 23632371
195
Shaffer Milo S.P., Windle A.H., Adv. Materials 1999, 11 (11), 937-941
196
Valentini L., Armentano I., Puglia D., Kenny J. M., Carbon 2004, 42, 323-329
197
Kim Y.J., Shin T.S., Choi H.D., Kwon J.H., Chung Y.C., Yoon H.G., Carbon 2005,43 (1), 23-30
198
Yuen S.-M., Ma C.-C. M., Lin Y.-Y., Kuan H.-C., Compos. Sci. Technol. 2007, 67, 2564-2573
199
Tsubokawa N., Polym. J. 2005, 37, 637-655
200
Kong H., Gao C., Yan D., J. Am. Chem. Soc. 2004, 126, 412-413
201
Park S. J., Cho M. S., lim S. T., Choi H. J., Jhon M. S., Macromol. Rapid Commun. 2003, 24, 1070-1073
202
Qin S., Qin D., Ford W. T., Resasco D. E., Herrera J. E., J. Am. Chem. Soc. 2004, 126, 170-176
203
Qin S., Qin D., Ford W. T., Resasco D. E., Herrera J. E., Macromolecules 2004, 37, 752-757
204
Buffa F., Hu H., Resasco D. E., Macromolecules 2005, 38, 8258-8263
205
Saeed K., Park S.-Y., J. Appl. Polym. Sci. 2007, 104, 1957-1963
206
Fu K., Huang W., Lin Y., Riddle L. A., Carroll D. L., Sun Y. P., Nanolett. 2001, 1, 439-441
207
Lou X., Detrembleur C., Sciannamea V., Pagnoulle C., Jerome R., Polymer 2004, 45, 6097-6102
167
52
208
Liu P., Eur. Polym. J. 2005, 41, 2693-2703

http://www.principiaconsulting.com/publishing/news.cfm?article=64
210
Hyperion Catalysis International, Plast. Add. Comp. 2001, 3, 20-22
211
Kim W.-S., Song H. S., Lee B. O., Kwon K.-H., Lim Y.-S., Kim M.-S., Macromol. Res. 2002, 10, 253-258
212
Thomassin J.-M., Pagnoulle C., Bednarz L., Huynen I., Jerome R., Detrembleur C., J. Mater. Chem. 2008, 18,
792-796
213
Park C., Ounaies Z., Watson K., Crooks R. E., Connell J., Lowther S. E., Siochi E. J., Harrison J. S., St. Clair
T. L., Chem. Phys. Lett. 2002, 364, 303-307
214
Shibayama K., Nakasuga A., Japan Patent number JP 2004075706, 2004
215
Li Y.-L., Kinloch I. A., Windle A. H., Science 2004, 304, 276-278
216
www.nanocyl.com
217
Beigbeder A., Degee P., Conlan S. L., Mutton R. J., Clare A. S., Pettitt M. E., Callow M. E., Callow J. A.,
Dubois Ph., Biofouling 2008, 24, 291-302
209
53
III. OBJECTIVES &

PROPOSED RESEARCH STRATEGIES
Polymer nanocomposites represent a new class of composite materials and are defined
as multiphase materials where fillers have at least one dimension in the nanometer range. Due
to their wonderful combination of physical and chemical properties, the carbon nanotubes
(CNTs) should assume an important role in the field of nanomaterials 1,2. However, an
important parameter for preparing polymer/CNTs nanocomposites with enhanced electrical,
mechanical and thermal properties is the extent of dispersion of the individual CNTs
throughout the polymer matrix 3,4,5. Due to numerous intermolecular van der Waals forces and
their entanglements, the CNTs tend to form aggregates limiting their dispersion ability in
solvents and polymer matrices. To overcome the CNTs bundle-like agglomerations, different
techniques are presented in the literature such as melt blending under very high shear
conditions or surfactant-assisted processing of composites materials for example.
One of the most promising techniques is to covalently attach chemical species onto the
CNTs surface improving the CNTs dispersion 6. Moreover, the covalent functionalization can
provide a means for engineering the nanotube/polymer interface for optimal nanocomposite
properties. With respect to mechanical properties, for example, the interfacial adhesion can be
modified through the covalent link between the CNT and the polymer matrix to maximize
load transfer. However, a noticeable drawback of the covalent functionalization is the
disruption of the extended conjugation creating defect sites on the CNTs surface. As a
consequence, their mechanical, thermal and more particularly electrical properties can be
damaged and sharply decrease. A high control over the functionalization process is thus
necessary to graft chemical groups on already existing defect sites and to limit the damaging
of CNTs. On the other hand, in the majority of so far reported processes such as chemical
functionalization performed in solvents 7,8 or by ball milling 9, only a tiny fraction of the
carbon atoms on the nanotubes surface get functionalized.
Accordingly, our main objectives are to develop a non destructive and selective
covalent functionalization process and to disperse these treated CNTs within a polymer
matrix. The covalent bonds at the interface polymer/nanotube as well as the preservation of
III. PROPOSED REASEARCH STRATEGIES
54
electrical properties of CNTs are needed to obtain nanocomposites with enhanced electrical
conductivity.
To reach our objectives, it has been decided to develop a three-step strategy combining a
highly efficient -waves post-discharge plasma functionalization, the subsequent CNTs-gpolyester nanohybrid formation and the melt blending within the host polymer matrices
(Figure 3-1).
Figure 3- 1 : Schematic representation of the three-step thesis strategy
In the first part of the study, the treatment of CNTs via an original microwave plasma
process will be developed to readily functionalize the CNTs surface with primary amine
groups in one single step. The use of plasma, which is an ionized gas consisting in ions,
electrons and neutral species, allows to treat the CNTs with a relative low time and
environmentally friendly process without any use of solvent. Another advantage of plasma is
55
the versatility of the technique which is indispensable to develop a highly controlled CNTs
functionalization process. The process consists to submit the CNTs under a reactive flow of
atomic nitrogen formed in an Ar+N2 -wave plasma, which are known to easily yield radicals
under relatively high gas pressure allowing the production of a high density of reactive
species.
The first part of the functionalization process will be carried out on pressed CNTs pellet to
verify the efficiency of the nitrogenated functions grafting via X-ray photoelectron
spectroscopy (XPS). The covalent character of bonds between CNTs and nitrogenated groups
and the morphology of treated CNTs will be characterized by high-resolution ultraviolet
photoemission spectroscopy (HR-UPS) and high-resolution transmission electron microscopy
(HR-TEM) respectively. Therefore, the addition of hydrogen in the set-up will be envisaged
to increase the primary amine selectivity and grafting efficiency on the surface of CNTs
powder.
In the second part, the so-grafted primary amines will be further used as initiation sites for
promoting the ring opening polymerization (ROP) of lactone monomers (i.e., -caprolactone)
yielding polyester-grafted CNT nanohybrids. The morphology of the recovered nanohybrids
will be characterized by TEM and the quantity of grafted polymer chains by
thermogravimetric analysis (TGA). Interestingly enough, the absolute quantification of
primary amines will be performed via both TGA and XPS analyses.
Finally, the so-formed CNTs-g-PCL nanohybrids will be used as highly filled masterbatches
to be dispersed in the molten state within either poly(-caprolactone) (PCL) or high-density
polyethylene (HDPE) matrices. The thermal behavior, the electrical properties and the
morphology of the so-produced nanocomposites will be characterized. The use of two
polymer matrices, having different miscibility behavior with PCL chains, will determine the
influence of the CNTs coating on the resulting properties of the nanocomposites.
56
References
1
Thostenson E. T., Ren Z., Chou T.-W., Compos. Sci. Technol. 2001, 61, 1899-1912
Ajayan P. M., Schadler L. S., Giannaris C., Rubio A., Adv. Mater. 2000, 12, 750-753
3
Peeterbroeck S., Alexandre M., Jrme R., Dubois Ph., Polym. Deg. Stab. 2005, 221, 115-126
4
Biercuk M. J., Llaguno M. C., Radosavljevic M., Hyun J. K., Johnson A. T., Appl. Phys. Lett. 2002, 80, 27672769
5
Coleman J. N., Khan U., Blau W. J., Gunko Y. K., Carbon 2006, 44, 1624-1652
6
Dyke C. A., Tour J. M., J. Phys. Chem. A 2004, 108, 11151-11159
2
Balasubramanian K., Burghard M., Small 2005, 1, 180-192

Gromov A., Dittmer S., Svensson J., Nerushev O. A., Perez-Garcia S. A., Licca-Jimenez L., Rychwalski R.,
Campbell E. E. B., J. Mater. Chem. 2005, 15, 3334-3339
9
Barthos J. R., Mhn D., Demortier A., Pierard N., Morciaux Y., Demortier G., Fonseca A., Nagy J. B., Carbon
2005, 43, 321-325
8
57
IV. RESULTS & DISCUSSION

The first challenge of this work relied upon the development of a strategy to graft
functional groups at the carbon nanotube (CNT) surface without reducing their intrinsic
properties. We have focused on the grafting of primary amines because these reactive
functions can be used as anchoring sites for polymer chains.
The second challenge consisted of the polymerization of lactone monomers from the reactive
primary amine functions grafted at the CNT surface. The aim was to synthesize nanohybrids
(CNTs-polymer) in order to pre-disaggregate the carbon nanotube bundles.
The third challenge was to use these nanohybrids as masterbatches highly filled in CNTs. The
objective relied upon the dispersion in the molten state of the aforementioned nanohybrids
within matrices such as poly(-caprolactone) (PCL) and high density polyethylene (HDPE),
yielding nanocomposites with improved properties. We will discuss the impact of
nanohybrids, and more specifically polyester chains grafted on the surface of CNT
nanofillers, on thermal and electrical properties of the recovered nanocomposites.
The CNTs used in this work were multi-walled carbon nanotubes (MWNTs) which are
industrially produced by Nanocyl S.A. (Nanocyl-7000), via a catalytic carbon vapour
deposition (CCVD) process 1. Carbon nanotubes have a purity of min. 90% carbon.
IV.A. Microwave plasma post-discharge treatment

In the first part of this work, a strategy to functionalize, in one step, the CNT surface
with nitrogenated functions is developed. As discussed before, within this framework, the use
of a plasma technique is a promising functionalization method. Plasma induced
functionalization presents some important characteristics such as a relative low time process
and the absence of solvents, thus being an environmentally friendly strategy. In plasma
induced functionalization, a large range of functional groups can be grafted on the material
surface via an unique step treatment. However, for the preservation of the carbon nanotube
structure, and thus of their physical properties, it is important to avoid the interaction of high
IV. RESULTS AND DISCUSSION
58
energetic particles and radiation with the CNT surface. This can be achieved by placing the
CNTs outside the glow discharge during the plasma treatment. It is thus necessary to create
reactive species with a sufficient mean lifetime to reach the nanotubes in the post-discharge
chamber. Microwave (-wave) plasma has been used because this type of plasma yields easily
radicals under high gas pressure allowing the production of a high density of reactive species.
IV.A.1. Microwave plasma experimental set-up and treatment parameters

The CNTs are treated in the post-discharge of an Ar + N2 microwave plasma sustained
by a surface wave launched in a quartz tube via a surfaguide supplied by a 2.45 GHz
microwave generator, which can be pulsed. The experimental set-up is schematically
represented in Figure 4-1.
Figure 4 - 1 : Schematic representation of the microwave plasma set-up with zoom on the post-discharge part
Plasma parameters, such as gas flow, pressure, pulse time of the power supply or dilution of
molecular nitrogen (N2) in argon (Ar), have influence in the creation of atomic nitrogen
species. Godfroid et al. 2 have studied the influence of these parameters on the dissociation
rate of N2 and a maximum value of 43% was achieved. In the first set of experiences of this
work, the plasma parameters were fixed to reach this optimal nitrogen dissociation. These
59
optimal plasma parameters are a discharge pressure of 36 Torrs and gas flow rate of 1750
sccm and 50 sccm for Ar and N2 respectively. The average power is maintained at 400 W but
the power in the pulse is the maximum power of 1600 W which can be delivered by the
microwave generator. The optimum modulation is a period of 900 s and a pulse time of 225
s.
It is important to note that the discharge tube and the post-discharge chamber are
separated by a diaphragm that consists of an aluminium ring with a circular aperture whose
diameter measures few millimetres. This diaphragm plays two-roles in the post-discharge
treatment. Firstly, the separation between the plasma set-up and the post-discharge protects
the samples placed in the post-discharge from irradiation of high-energy particles from wave plasma. Moreover, the average distance of 40 cm between the end of the discharge tube
and sample holder promotes the interaction of nanotubes with atomic nitrogen species that
have enough mean lifetime to reach them contrary to others species.
Secondly, the diaphragm allows to differentiate and to control the post-discharge pressure
from the discharge one. The discharge pressure needed to reach an optimal nitrogen
dissociation rate can be different from the required post-discharge pressure for optimal
nitrogenated groups grafting on the surface of CNTs. Being therefore independently adjusted
from the pressure of the discharge tube, the pressure in the post-discharge chamber becomes
an important parameter for the CNT treatment because the number of reactive species varies
in function of this pressure. Our preliminary results showed that a post-discharge pressure of 3
Torrs allows efficient nitrogen grafting on the CNT surface 3. The treatment time also
influences the amount of functional groups grafted on the CNT surface.
IV.A.2. Atomic nitrogen treatment of carbon nanotubes

The first set of samples to be investigated is composed of CNTs plasma treated under
atomic nitrogen flow generated by Ar + N2 microwave plasma. The influence of the
interaction of the atomic nitrogen with the CNTs on the surface chemistry, electronic states of
the valence band and morphology of CNTs has been investigated.
60
IV.A.2.a. Treatment of CNTs pressed pellet

First the microwave post-discharge treatment was carried out on a pellet of pressed
CNTs to determine the impact of atomic nitrogen on the chemical composition of the CNT
surface. To form the carbon nanotubes pellet, the CNTs were compacted in a mechanical
press under 4.103 Pa. The pellet is placed on a sample holder in the post-discharge chamber of
the microwave plasma. The post-discharge chamber is pumped down to a pressure lower than
10-4-10-3 Pa in order to remove O2 and H2O traces at the most. The pellet of carbon nanotubes
is then treated during 45 minutes. The applied microwave plasma parameters are summarized
in Table 4-1. After treatment, the discharge tube and the post-discharge chamber gradually
return to ambient atmosphere before analysis of the treated CNTs pellet in the X-ray
Photoelectron Spectroscopy (XPS) set-up.
Microwave plasma parameters

Power parameters
Power in the pulse
Average power
Period
Pulsation time
1600 W
400 W
900 s
225 s
Flow rate
Argon
Nitrogen
1750 sccm
50 sccm
Pressures
Discharge tube
Post-discharge chamber
36 Torrs
3 Torrs
Table 4 - 1 : Used parameters for CNTs pellet treatment
Figure 4-2 shows the comparison of XPS survey spectrum recorded on pristine carbon
nanotubes and on N-functionalized carbon nanotubes pellet. The peak at 284.3 eV binding
energy, observed in both spectra, arises from photoelectrons emitted from the C 1s core level.
The presence of a low intensity peak near 532 eV in the spectrum recorded on untreated
sample (highlighted by the black ring in Figure 4-2a), indicates that pristine CNTs are slightly
oxidized, 0.6 atomic % of oxygen determined by the area measurement of this O 1s peak.
The presence of the peak near 399 eV in the spectrum recorded on treated CNTs (Figure 42b), attests for the presence of nitrogen groups on the CNTs pellet 4. The relative atomic
61
concentration of nitrogen is 11.5 0.3 at.%. The increase of O 1s peak intensity at 532 eV is
also observed, the oxygen relative atomic concentration is 4.7 0.5 at.%, which means an
oxygen incorporation of about 4 at.% after CNTs treatment. The oxygen may originate either
from residual gases in the plasma set-up, such as O2 and H2O vapour, or during the exposition
of freshly treated CNTs to ambient atmosphere. It is worth pointing out that these atomic
percentages are an average value obtained after measurements on minimum three points of the
CNTs pellet surface.
Element
Binding
Energy (eV)
At. %
284.3
83.8
399.4
11.5
531.8
4.7
O 1s
Intensity (arb. Units)
C 1s
N 1s
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
Binding Energy (eV)

Figure 4 - 2 : XPS spectra of (a) pristine CNTs and (b) CNTs pellet exposed to atomic nitrogen
in the Ar + N2 microwave plasma post-discharge chamber
Figure 4-3 shows a comparison between the C 1s peaks recorded on pristine and
treated CNTs with high resolution XPS in the synchrotron of Hamburg by analysing the
kinetic energy of photoelectrons generated using 3300 eV photons. The use of HR-XPS
analysis allows to measure precisely the binding energy range of carbon components. For
both samples, we can observe the typical C 1s peak of CNTs with the contribution peak at 291
eV, commonly related to the -* transition levels 5. The energy position of this plasmon
loss peak is known to be related to the delocalization of 2p- electrons. Indeed, larger electron
62
delocalization is expected for higher plasmon energy while smaller plasmon energy can be
related to electronic localization6. In Figure 4-3, there are no changes in energy position of the
plasmon loss peak after nitrogen functionalization. This observation suggests that the degree
of electron delocalization is not modified by grafting nitrogenated groups onto the CNTs
surface. The chemical modification produced by the microwave plasma post-discharge
treatment is revealed by the new broad structure near 287.5 eV that is attributed to
photoelectrons emitted from carbon atoms belonging to functional groups containing nitrogen
and/or oxygen atoms such as amine, amide or nitrile groups7.
Figure 4 - 3 : HR-XPS C 1s spectra recorded on (a) pristine CNTs and (b) atomic nitrogen exposed CNTs
Thus XPS analysis results seem to indicate that the Ar + N2 microwave plasma postdischarge treatment allows the grafting of nitrogenated groups onto carbon nanotubes surface.
However, there are not absolute evidence of effective covalent bonds between N groups and
CNTs. In order to verify the covalent character of these bonds, high resolution ultraviolet
photoelectron spectroscopy (HR-UPS) was carried out, in the synchrotron of Hamburg, on
three different samples: pristine CNTs and CNTs pellet exposed to atomic nitrogen for 5 and
15 minutes, respectively, under plasma parameters presented in Table 4-1. The valence band
spectra were recorded at a photon energy of 55 eV. This analysis reveals changes on the
valence electronic states with nitrogen grafting time (Figure 4-4).
63
The valence band structure of the pristine carbon nanotubes (Figure 4-4 a) is essentially the
same as that of graphite. The characteristic features arising from the threefold coordination of
the carbon atoms are the 2p- band near 3.5 eV, the 2p- states near 5.5 and 8.0 eV and the
mixed 2s-2p hybridized states at 13.6 eV. The feature near 14 eV was reported to be related to
oxygen grafting on the sidewalls and at the edges of the CNTs 8.
In addition to the features observed in the valence band spectrum of pristine CNTs, spectra
recorded after 5 minutes (Figure 4-4 b) and 15 minutes (Figure 4-4 c) of treatment have a
band peaking near 4.2 eV that is attributed to photoelectrons emitted from nitrogen alone pair
states (sp2 hybridized nitrogen) 9, while the 2p- band of carbon atoms at 3.5 eV tends to
disappear. The dominant feature centred at 8.5 eV contains contributions from C and N 2p
electrons associated with bonds overlapping the -bond feature of the C 2p electrons located
at about 7.5 eV. A band located at 9.8 eV, which appears with treatment and contributes to
this large peak, is generated by photoelectrons emitted from C 2p N 2p -states 10. In brief,
the 2p- states can no longer be observed and the contribution of 2p- states is enhanced with
atomic nitrogen treatment giving evidence for the covalent character of the nitrogen grafting
on the carbon nanotube surface.
C-N states
Intensity (arb. units)
C-N states
c
b
C-C states
C-C states
20
15
10
Binding Energy (eV)

Figure 4 - 4 : Valence band spectra recorded by HR-UPS analysis on pristine CNTs (a)
and CNTs exposed to N for 5 minutes (b) and 15 minutes (c)
64
In conclusion, photoelectron spectroscopy analysis shows that, by exposing CNTs to

atomic nitrogen generated in the microwave plasma, nitrogenated chemical groups can be
covalently grafted onto the CNTs surface altering the density of electronic states. The quantity
of grafted nitrogenated groups can reach high values, such as 11.5 at.% of the CNTs pellet
surface, which shows the high grafting efficiency of the process.
Since the efficiency of the nitrogen functionalization has been evidenced, the next step
consists in the study of the influence of the treatment on the carbon nanotube morphology.
IV.A.2.b. Morphology of nitrogen-treated carbon nanotubes

Different examples in the literature showed that functionalization of carbon nanotubes
using plasma source can present a problem of surface etching leading to a decrease or even
loss of the CNT physical properties due to structural damages. This problem was already
observed in functionalization performed inside 11,12 and outside 13 of the discharge. It was thus
important to analyse the morphology of the Ar +N2 microwave plasma post-discharge treated
CNTs.
Some CNTs were deposited on a Transmission Electronic Microscopy (TEM) grid and the
supported samples were directly exposed under atomic nitrogen flow in the post-discharge of
the Ar + N2 microwave plasma for 15 minutes, with plasma parameters presented in the Table
4-1. Then, the samples were analyzed by High Resolution TEM (HRTEM), in the EMAT
laboratory at University of Antwerp, to evaluate the possible etching of CNTs sidewalls due
to exposure to atomic nitrogen. Figure 4-5 shows typical micrographs recorded on these
treated samples.
65
Figure 4 - 5 : HRTEM images of CNTs treated in the post-discharge chamber of Ar + N2 microwave plasma
The excellent morphology preservation of the CNT sidewalls indicates that the -wave
plasma post-discharge treatment does not damage the CNTs structure. Before treatment, the
outer surfaces of these CNTs probed to be covered by amorphous carbon traces which were
formed during their synthesis 14. After atomic nitrogen treatment, the relative quantity in
amorphous carbon seems to decrease indicating that the treatment process cleans the CNT
surface.
The absence of damages in the structure of carbon nanotubes after the plasma treatment
indicates also that nitrogen groups grafting results only from chemical reaction between CNTs
surface and atomic nitrogen species, without side-effects of irradiation or high energy particle
interaction at CNT surface.
We have seen that CNTs pellet treatment in the Ar + N2 -wave plasma post-discharge allows
for the covalent grafting of nitrogenated functions without damaging their structure. However,
before using these so-functionalized CNTs in different applications, it has been necessary to
scale up the treatment directly onto the whole content of CNTs in powder (and not pressed in
pellet anymore).
66
IV.A.2.c. Treatment of CNT powder

We have observed that the treatment of CNTs pellet in post-discharge of the Ar + N2
-wave plasma allows a very efficient covalent functionalization reaction on CNT surface.
However, the functionalization is not homogeneous because just the CNTs near the top
surface of the pellets are exposed to the plasma. Indeed, atomic nitrogen species could only
interact with the nanotubes present on the upper surface of the sample, which have the highest
amount of nitrogenated groups on their surface.
A mechanical stirring system of CNT powder has thus been developed with the aim of
treating a large amount of CNTs (Figure 4-6). This system is constituted by a crucible directly
mounted on a vibrating speaker membrane. In the crucible, the CNT powder is stirred during
the treatment allowing a permanent renewal of carbon nanotube accessible to the reactive
atomic nitrogen flow.
Figure 4 - 6 : Experimental -wave plasma set-up constituted of the discharge tube and the post-discharge chamber,
which contains the CNTs mechanical stirring system
To avoid the dispersion of the CNT powder in the post-discharge chamber, the Ar flow has
been fixed at 750sccm. This decrease of the Ar flow influences the N2 dissociation in the
plasma discharge, actually equal to 20%, the half value obtained with optimal plasma
67
parameters. The treatment time was set at 2 hours. While the quantity of treated CNTs powder
was fixed at 0.5g.
The aim of this experiment was to determine whether nitrogenated groups were
grafted homogeneously on the whole surface of the carbon nanotubes, involving an efficient
nitrogen functions grafting. After the functionalization process, the treated CNTs were
pressed in pellet to be analyzed by XPS.
XPS analysis measures the kinetic energy and the number of electrons that escape from the
top 10 nm of the material being analyzed. It is important to note that since carbon nanotubes
present an average diameter of ca. 10 nm, XPS results arise from the average analysis of the
volume of the first CNTs layer and not of the extreme surface of the nanotubes. Thus the
absolute quantification of functional groups grafted at the CNT surface can not be determined.
However, these values are representative of the functionalization efficiency and can be
compared to each other.
Figure 4-7 shows the comparison of XPS survey spectrum recorded on CNTs and on Nfunctionalized CNTs. A low intensity peak of nitrogen at 399.7 eV appears in the spectrum
recorded on treated CNTs (Figure 4-7b). The measured quantity of N element is near to 0.9
0.1 at.%. The increase of O 1s peak intensity at 532 eV is also observed, the oxygen relative
atomic concentration is equal to 1.1 0.3 at.%, which means an oxygen incorporation of
about 0.5 at.% after CNTs treatment.
XPS measurements were carried out on five different locations of the same sample, the slight
standard deviation of the relative nitrogen atom concentration showed that the whole amount
of CNTs was homogenously treated. The total amount of grafted elements per CNTs is
smaller when the treatment is performed in powder than in pellets probably due to the
differences in the atomic nitrogen residence time at the surface of each carbon nanotube.
68
Element
Binding
Energy (eV)
At. %
284.2
98.0
399.7
0.9
531.8
1.1
C 1s
NN1s
1s
403 401 399 397 395
OO1s
1s
a
539 537 535 533 531 529
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
Binding Energy (eV)

Figure 4 - 7 : XPS spectra of (a) pristine CNTs and (b) CNTs exposed to atomic nitrogen
in the Ar + N2 microwave plasma post-discharge chamber
The above discussed results suggest that nitrogen atom grafting takes place via atomic
nitrogen interaction with CNTs thus presenting ca. 1 at.% of nitrogen.
The next step consisted in the selective grafting of primary amine functions. Before this step,
the origin of the oxygen contamination was determined.
IV.A.2.d. Origin of oxygen surface contamination

As previously discussed, the -wave plasma post-discharge treatment allows the
covalent grafting of nitrogen groups, but also the introduction of oxygen enriched groups on
the carbon nanotube surface. Oxygen contamination sources generally observed in plasma
surface modification processes 15,16 are most often the residual oxygen and water vapour
present in the plasma. The water vapour is produced by desorption of water from the plasma
set-up walls 17. However, unlike the grafted nitrogen content, a variation of the relative
percentage of oxygen atoms, between samples treated with identical plasma parameters was
observed. Oxygen contamination can thus be provided from another source such as postIV. RESULTS AND DISCUSSION
69
plasma reaction on the surface of carbon nanotubes. To shed some light on the origin of this
contamination, freshly treated CNT pellets were put under atmospheric pressure either using
air or oxygen. The microwave plasma post-discharge treatments were carried out for 45
minutes with the plasma parameters presented in Table 4-1. The surface treated CNT pellets
were then characterized by XPS and the results are presented in Table 4-2.
Air
Element
C1s
N1s
O1s
Binding
Energy (eV)
284.2
399.8
533.0
O2
b
At. %
Element
83.8
11.5
4.7
C1s
N1s
O1s
Binding
Energy (eV)
284.3
399.9
533.0
At. %
83.1
9.4
7.5
Table 4 - 2 : Surface element content of CNTs pellets treated under same treatment plasma parameters
and conserved under atmospheric pressure in air (a) or oxygen (b)
When the freshly treated nanotubes were placed under atmospheric pressure via oxygen, the
grafted oxygen quantity proved more important while nitrogen content was lower than when
placed in air. It seems thus that reactive species remain on the surface of CNTs after
functionalization process, which can react with oxygen or water vapour present in ambient
atmosphere. In order to explain this observation, we suggest that these reactive species are
radicals remaining on the CNTs surface after plasma post-discharge treatment. Interestingly,
the presence of a significant amount of radicals on the freshly treated CNTs surface has been
clearly evidenced by chemical reaction with 2,2-diphenyl-1-picrylhydrazyl (DPPH) (Figure 48). DPPH is widely used to test the ability of compounds to act as free radicals scavengers or
radical hydrogen donors. In the presence of such species, DPPH turns from purple to pale
yellow color corresponding to the reduced form of DPPH 18.
DPPH yellow
DPPH purple
Figure 4 - 8 : Reaction with free radicals turns DPPH from purple to yellow color
70
A solution of DPPH in methanol with a concentration of 5.10-7M has been prepared and 20ml
of this solution was added into test tubes containing each 20mg of pristine CNTs and freshly
treated CNTs.
When freshly treated CNTs (t-CNTs) are contacted with the DPPH solution, the supernatant
turns from purple to yellow color, while no evolution of the DPPH purple color is observed
when it is added onto pristine CNTs (Figure 4-9).
DPPH
+ CNTs
+ t-CNTs
Figure 4 - 9 : Qualitative assessment of the presence of radicals on the surface of freshly treated CNTs.
The samples are (a) DPPH solution alone, (b) DPPH solution in presence of pristine CNTs and (c) DPPH solution in
presence of freshly treated CNTs (t-CNTs)
These observations were further confirmed by UV-visible absorption spectroscopy as

presented in Figure 4-10. The molecule of DPPH is characterized as a stable free radical due
to the delocalisation of the spare electron over the aromatic molecule. This delocalisation
gives rise to the deep purple color, characterized by an absorption band in methanol solution
centered at the wavelength of about 520 nm18. This absorption peak is conserved when the
DPPH solution is added to pristine CNTs showing that any chemical reactions occurred. The
observed decrease of peak intensity can be related to the decrease of DPPH concentration due
to partial physisorption of DPPH aromatic rings on the graphitic CNTs surface. When DPPH
solution is added to freshly treated CNTs, the characteristic absorption peak disappears
completely meaning that DPPH is under reduced form after reaction with free radicals present
on the t-CNT surface.
71
DPPH alone
DPPH + CNTs
DPPH + t-CNTs
Absorbance (arb. unit)
a
b
c
300
350
400
450
500
550
600
650
700
750
800
850
Wavelength (nm)
Figure 4 - 10 : Absorption spectrophotometric analysis of (a) DPPH solution alone, (b) DPPH solution in presence of
pristine CNTs and (c) DPPH solution in presence of freshly treated CNTs (t-CNTs)
The DPPH solution was also added to CNTs treated one day earlier, no color change was
observed showing that all remaining radicals did react under ambient atmosphere.
Taking into account these observations, one may assume that oxygen contamination of
the treated CNT surface is partially due to post-plasma reaction between the radicals
remaining active at the CNT surface and ambient air (O2) atmosphere. Accordingly, all wave plasma post-discharge treated CNTs were recovered samples under ambient atmosphere
with argon gas for one hour right after the treatment.
IV.A.3. Quantification of primary amine functions grafted on CNTs

In the -wave post-discharge plasma treatment, atomic nitrogen species interact with
CNTs leading to covalent grafting of nitrogenated groups on their surface without damage to
their structure. However, among these nitrogenated functions, primary amines have attracted
our attention because they can be used as initiation sites for promoting lactone ring-opening
polymerization yielding polyester-grafted MWNT nanohybrids.
72
It is important to remark that the used plasma parameters for all future treatment of CNTs
pellets and powders will be identical in order to allow the comparison between results. These
standard parameters are represented in Table 4-3.
Microwave plasma parameters

Power parameters
Power in the pulse
Average power
Period
Pulsation time
1600 W
400 W
900 s
225 s
Flow rate
Argon
Nitrogen
1750 sccm
50 sccm
Pressures
Discharge tube
Post-discharge chamber
36 Torrs
3 Torrs
Table 4 - 3 : Standard plasma parameters for the post-discharge treatment of the CNTs
IV.A.3.a. Limits of XPS N1s peak deconvolution

Quantification of amine functions can be done by determining their contribution to the
XPS N1s peak. The positions of XPS peaks in term of binding energy provide information
about the chemical state of near surface elements. The nature and the percentage of the
different nitrogenated functions grafted on the CNTs pellet surface could be thus obtained by
N1s curve deconvolution 19. However, this technique presents some limitations for exact
quantification of primary amines. Indeed, one problem is that a good fit is always achieved by
a sufficient number of Gaussian-Lorentzian curves when optimized without constraints.
However, all these curves do not always correspond to existing chemical groups.
Understanding the chemistry of the material before analysis is thus important as it suggests
the number of chemical states and therefore the number of fitting curves. The introduction of
constraint parameters to restrict widths and relative intensities of the fitting peaks is therefore
necessary to force the peak model to correspond to the chemistry of the analyzed material
surface. In our study, we know the grafting support, i.e. the carbon nanotubes, but not the
number and nature of all nitrogenated functional groups. Moreover, the XPS resolution of the
73
spectrometer used in this study is limited to 0.8 eV, it is thus impossible to clearly distinguish
primary, secondary and ternary (nitrogen atoms incrusted in the graphene lattice) amines. An
example of obtained N 1s peak is presented in Figure 4-11, the binding energies of the
different nitrogenated functions are close to each other and the form of the nitrogen peak does
not make the peak fitting easy.
Nitrile
Amine
Amide
Oxime
409
407
405
403
401
399
397
395
393
391
Binding Energy (eV)
Figure 4 - 11 : Example of N 1s peak dconvolution.

This peak can be fitted by many ways with different number of components
IV.A.3.b. Chemical derivatization of primary amines

In brief, even if XPS allows the quantification of superficial elements, the
determination of structural features from the binding state of elements is more delicate.
A solution may consist in the chemical derivatization process, which can transform the
problem of functional groups analysis into the determination of element concentrations. This
technique was developed to selectively label surface functional groups of interest with
molecules containing an element that is not present on the surface prior to the derivatization
reaction. Elements like fluorine or iodine are most useful because they can be determined by
XPS with very high sensitivity 20. In this study, 4-(trifluoromethyl)benzaldehyde (TFBA) has
been used as derivatization agent to detect primary amines 21,22,23. The condensation reaction
consists in a nucleophilic addition on carbonyl group (C=O) that converts NH2 group into
imine covalent bond (C=N) stabilized by conjugation with aromatic rings of CNTs. The
74
aldehyde group of TFBA reacts exclusively with primary amine functions, as shown by
Chevallier et al 24. Accordingly, XPS has been used to analyze the TFBA derivatized surfaces
to determine the fluorine content thus related to the primary amine content. Moreover, the use
of TFBA allows for tripling intensity of the measured signal because three fluorine atoms per
primary amine group are detected (Figure 4-12) and fluorine atoms are efficiently detected by
XPS. Indeed, the atomic sensitivity factors of fluorine and nitrogen are equivalent to 1 and
0.42 respectively 25, one fluorine atom will be thus more than two times easily detected by
XPS than one nitrogen atom. By reaction with TFBA, the XPS signal of primary amine
functions is thus multiplied in average by a factor of six.
F F
NH2
R
F
H
N C
F F
H
N C
NH2
R = nitrile, amide .
F
F F
Figure 4 - 12 : Scheme of the specific derivatization reaction of TFBA with primary amines
grafted on the CNTs surface
Primary amine concentration is derived as follows:
[2] = /( 8/3) 100%

3
(IV.1)
where [NH2] is the percent of carbon atoms carrying primary amines, F and C are the fluorine
and carbon atomic concentrations determined by XPS after the reaction with TFBA. F/3
corresponds to the number of NH2, 8F/3 represents the amount of carbon introduced with
TFBA molecules. The [NH2]/[N] (or F/3N) ratio is considered as a measure of NH2 groups
grafting selectivity, namely the relative concentration of N atoms bonded in primary amines
(in % of N). The latter is therefore particularly useful for comparing samples with different
[N] values and determining the selective primary amination efficiency of treatment. It has
been shown that the TFBA derivatization reaction saturates efficiently all the superficial
primary amines20.
In order to compare quantification values obtained from different XPS analyses, it is
necessary to normalize these values since the measured atomic content are relative values that
75
can vary by introduction of non desired elements like oxygen for example. The normalization
of the element content consists to divide the measured value of this element by the value of
the element present in majority, carbon atoms in the case of CNTs. In this work, normalized
values will be represented with square brackets, [N] for nitrogen for example.
An important parameter of the derivatization technique is the time of exposition of each
sample to TFBA agent. The derivatization reaction between primary amines and TFBA has to
be quantitative. A kinetic study of NH2 derivatization has to be performed by evaluation of
introduced fluorine content as a function of derivative reaction time. The first kinetic study
has been carried out on pressed CNTs pellet because the quantity of grafted nitrogen groups
proved higher than via the powder treatment (see previous discussion).
Firstly, pristine CNTs pellet was derivatized with TFBA for 12 h and analyzed by XPS. No
fluorine traces were detected confirming the absence of primary amines on the surface of
pristine carbon nanotubes.
A CNTs pellet was then treated under standard plasma parameters for 45 minutes and
then divided in several parts, which were exposed each to a TFBA saturated atmosphere in a
derivatization chamber at room temperature for different reaction times. We have studied in
detail the primary amine derivatization kinetics by evaluating the surface chemical
composition (F content) and accordingly the primary amines selectivity as a function of
derivatization time (Figure 4-13). We found that the surface composition stabilizes after
0.5h of derivatization when the F concentration reaches a constant value of 0.9 at.%. The
primary amines selectivity is thus equal to c.a. 2.5 %. To be sure that all primary amines have
reacted with TFBA, a minimal derivatization time of 4h has been considered appropriate for
quantitatively determine the relative content of primary amine functions.
76
3,5
2,0
1,8
1,6
2,5
1,4
1,2
2,0
1,0
1,5
0,8
0,6
1,0
-NH2 selectivity (%)

F content (at.%)
0,5
0,0
0,0
0,5
1,0
1,5
2,0
F content (at.%)
3,0
0,4
0,2
0,0
Derivatization time (h)

Figure 4 - 13 : Evolution of fluorine content and primary amine group selectivity with derivatization time
for CNTs pellet treated in the post-discharge chamber of an Ar + N2 microwave plasma
Even if the quantity of nitrogen groups is important (11,5 at.%), the amount of primary
amines remains limited to 0.3 at.% at the surface of CNTs pellet treated for 45 minutes under
standard plasma parameters. The NH2 selectivity value is thus limited to 2.5% which means
that only one fortieth of nitrogen functions are primary amines. The quantity and the
selectivity of primary amines are thus very low on the surface of CNT pellet treated under
these parameters. After the treatment of CNTs powder, no significant traces of fluorine could
be detected by XPS while 1 at.% of nitrogen was measured confirming the very low primary
amine group selectivity. These observations can be explained knowing that the hydrogen
source to yield amine functions is only constituted of residual water molecules and hydrogen
traces present in the discharge tube and post-discharge chamber. A solution to reach a
selective NH2 grafting seems to introduce hydrogen gas directly in the plasma set-up.
77
IV.A.3.c. Post-plasma reaction with hydrogen gas

The reaction of radicals remaining on the treated CNTs surface with atmospheric oxygen was
observed previously. It has been thus suggested to expose freshly treated CNTs with hydrogen
atmosphere in order to verify if an increase of the efficiency and the selectivity of primary
amine grafting could be observed. The treatment of pressed carbon nanotube pellets was
performed in the conditions reported in the section IV.A.2.4. but here the sample was
contacted with hydrogen gas under atmospheric pressure after treatment and afterwards
derivatized with TFBA for 16h. The XPS analyses are compared in Table 4-4 with results
obtained for the sample simply recovered in air under atmospheric pressure.
Air
Element
C1s
N1s
O1s
F1s
Binding
Energy (eV)
284.2
399.8
533.0
688.2
H2
b
At. %
Element
83.8
11.4
4.8
/
C1s
N1s
O1s
F1s
Binding
Energy (eV)
284.3
399.9
533.0
688.1
At. %
84.5
11.2
4.3
/
Table 4 - 4 : Surface element content of TFBA derivatized CNTs pellets treated with same treatment plasma
parameters (see section IV.A.2.4) and contacted under atmospheric pressure to air (a) and hydrogen gas (b)
In both cases, only traces of fluorine atom are detected showing the low quantity of grafted
primary amines. The relative amount of elements is quite similar, a slightly lower oxygen
content is observed when hydrogen is used. This result indicates that remaining nitrogen
radical species on the surface of freshly treated CNTs are not enough reactive to interact with
molecular hydrogen and to form amine functions. It seems thus necessary to activate the
hydrogen dissociation in H in order to increase the selectivity and the efficiency of primary
amines grafting on the surface of CNTs.
IV.A.3.d. Introduction of H2 in the -wave plasma discharge

Hydrogen gas was introduced at the top of the discharge tube together with argon and
nitrogen gases (Figure 4-14). The hydrogen molecules are easily dissociated in the -wave
plasma. A large majority of the molecules are thus transformed in atomic hydrogen H13,
which can react with nitrogen species to form reduced nitrogen radicals in the plasma.
78
Figure 4 - 14 : Schematic representation of the introduction of H2 directly in the discharge tube
Different quantities of hydrogen were introduced directly in the discharge tube, the
flow varied from 0 to 56.5 sccm, which was the maximal flow allowed by the hydrogen mass
flow controller. The values of relative quantity of the elements present onto the surface of the
treated CNTs pellets, as derivatized by TFBA for 16h, were measured by XPS on every
samples thus treated under increasing hydrogen flow in the discharge (Table 4-5).
H2
(sccm)
CNTs pellet surface elemental composition

(at.%)
[N]
(%)1
[-NH2]
(%)2
-NH2
selectivity
(%)3
88.6
6.8
4.1
0.5
7.7
0.2
2.5
5.65
89.7
6.5
2.5
1.3
7.2
0.5
6.7
11.3
91.1
5.0
2.2
1.7
5.5
0.6
11.3
28.25
91.7
3.8
2.3
2.2
4.1
0.8
19.3
39.55
92.4
3.3
2.0
2.3
3.6
0.9
23.2
56.5
92.5
3.2
1.8
2.5
3.5
26
Note: 1[N]=(N/C)*100; 2 determined using equation IV.1; 3 ([F]/3[N])*100

Table 4 - 5 : XPS analysis of CNTs pellet surface in function of H2 flow introduced in the -wave plasma discharge
chamber: relative content in [N] and [NH2] and primary amine selectivity
79
An evolution of the elemental composition of the CNT pellet surface is observed with
the increase of the hydrogen flow in the discharge chamber. The quantity and selectivity of
primary amines in function of introduced H2 flow are presented in Figure 4-15 as well as the
evolution of nitrogen content. With the increase of the hydrogen flow, the quantity of nitrogen
atoms decreased while the content in primary amine functions is enhanced resulting in an
increase of primary amine selectivity from 2.5% to 26% for the maximal hydrogen flow. It is
interesting to note that the oxygen content decreases with the hydrogen flow as well. Different
treatment mechanisms can explain this behavior. The used parameters allowed to reach a
dissociation rate of N2 of 20% in the Ar + N2 glow. However, the added hydrogen molecules
are in majority dissociated in the plasma and this reaction modifies completely the plasma
chemistry, which can limit the N2 dissociation rate in atomic nitrogen. The passivation of the
CNT surface by atomic hydrogen can also explain the limited nitrogen and oxygen atoms
grafting.
100
N
(-NH2)
(-NH2 selectivity)
Element concentration (at.%)
7
6
80
60
4
40
3
2
20
1
0
10
20
30
40
50
60
Discharge hydrogen flow (sccm)

Figure 4 - 15 : Quantity and selectivity of primary amines and nitrogen content in function of H2 flow
directly introduced in the discharge chamber
Figure 4-15 suggests that the grafting of primary amines tends to some saturation with the
addition of hydrogen in the plasma chamber and the decrease of nitrogen content tends to
stabilize as well at ca. 40 sccm of added hydrogen flow.
80
These results demonstrate that the selectivity of primary amines grafting can be enhanced by
the introduction of H2 in the Ar + N2 -wave plasma. The relative quantity of grafted NH2
groups tend to be equal to 1%, which is certainly a very acceptable value compared to data
reported in the literature and for their use as further initiating sites for polymerization process.
As the quantity of grafted primary amines is more important, the kinetic of the
derivatization was studied to verify whether a TFBA derivatization time as long as 16h
remains sufficient for quantitatively determining the NH2 content.
Figure 4-16 shows that all grafted primary amine functions have reacted with TFBA already
after 2h, the studied derivatization time is thus well appropriate.
10
6
2
4
-NH2 selectivity (at.%)

F content (%)
10
12
14
Derivatization duration (h)
16
18
F content (at.%)
Figure 4 - 16 : Evolution of fluorine content and primary amines selectivity with derivatization time
for MWNTs pellet treated in the post-discharge chamber of an Ar + N2 + H2 microwave plasma
In a next step, the treatment of 0.5g of CNT-powder was carried out with addition of
56.5 sccm of hydrogen gas in the discharge tube. The CNTs were treated for 2h under
standard plasma conditions and, after compression in pellet, were derivatized with TFBA for
16h. XPS analysis results are represented in Table 4-6.
81
H2
(sccm)
CNTs elemental composition

(at.%)
[N]
(%)1
[-NH2]
(%)2
-NH2
selectivity
(%)3
98.0
0.9
1.1
traces
0.9
traces
56.5
98.8
<0.2
0.6
0.4
<0.2
0.1
Note: 1[N]=(N/C)*100; 2 determined using equation IV.1; 3 [F]/3[N]

Table 4 - 6 : XPS analysis of CNTs in function of H2 flow introduced in the -wave plasma discharge chamber:
relative content in [N] and [NH2] and primary amine selectivity
Evaluation of the quantity of nitrogen atoms grafted on the CNT-surface was not
possible by XPS analysis since the nitrogen content proved much lower than the minimum
nitrogen content detectable by XPS. However, under the same treatment conditions, but
without any addition of hydrogen gas in the discharge chamber, the grafted nitrogen atoms
was equal to ca. 1 at.%. Owing to their higher XPS sensibility after the TFBA derivatization
process, the content in primary amine functions could be determined as proved to be close to
0.1at.%. In comparison with pristine CNTs, no variation in the oxygen content was observed.
However, an increase from 0.5 to 1.1 at.% has been observed by performing the plasma
treatment without addition of hydrogen in the plasma discharge set-up.
The quantity of grafted primary amines is thus reduced when hydrogen is added in the Ar +
N2 discharge tube. Even if this addition of H2 in the -wave plasma allows an increase of
quantity and selectivity in primary amine functions, the quantity of nitrogen (and primary
amines) is reduced by the passivation of the CNT surface by atomic hydrogen. As hydrogen
cannot be detected by the XPS set-up, it has not been possible to determine the grafted
hydrogen content. The properties of CNTs, such as electrical conductivity, can be modified by
important hydrogen grafting. It has been reported that the surface of CNTs can be chemically
modified by atomic hydrogen that can also create additional defects along the CNTs 26.
In conclusion, the introduction of hydrogen gas in the discharge tube during Ar + N2 -wave
plasma post-discharge treatment of CNTs allows the increase in the selectivity of primary
amines grafted at the CNT surface. However, the total amount of -NH2 grafting is reduced due
to atomic hydrogen passivation of CNTs. The extent of this passivation cannot be quantified
and can impact on the CNT-properties if the creation of defects in the structure of carbon
nanotubes gets too important. As will be discussed below, an alternative process can thus be
proposed, that is the introduction of hydrogen gas directly in the post-discharge chamber
during the plasma treatment of the CNTs.
82
IV.A.3.e. Introduction of H2 in the -wave plasma post-discharge

Hydrogen gas was directly introduced in the post-discharge chamber (Figure 4-17)
instead of in discharge tube to avoid the formation of atomic hydrogen, which can passivate
the CNTs surface. In this strategy, no changes occur in the Ar + N2 -wave plasma allowing
the preservation of the N2 high dissociation rate of 20% in the used standard plasma
parameters. It was expected that the interaction of carbon nanotubes with atomic nitrogen in
hydrogen atmosphere could result in an increase of both quantity and selectivity of primary
amine grafting.
Figure 4 - 17 : Hydrogen gas is directly introduced in the post-discharge chamber
Different quantities of hydrogen gas were introduced in the post-discharge chamber,

the flow varied again from 0 to 56.5 sccm. The values of relative quantity of the elements
present onto the surface of the treated CNTs pellets, as derivatized by TFBA for 16h, were
measured by XPS on every samples thus treated under increasing hydrogen flow in the postdischarge (Table 4-7).
83
H2
(sccm)

(at.%)
[N]
(%)1
[-NH2]
(%)2
-NH2
selectivity
(%)3
88.6
6.8
4.1
0.5
7.7
0.2
2.5
5.65
88.7
6.7
3.9
0.7
7.6
0.3
3.0
11.3
88.2
7.0
4.0
0.8
7.9
0.3
3.8
28.25
88.0
7.1
3.7
1.2
8.1
0.5
5.6
39.55
88.3
6.8
3.4
1.5
7.7
0.6
7.4
56.5
87.7
7.2
3.2
1.9
8.2
0.8
8.3

Table 4 - 7 : XPS analysis of CNTs pellet surface in function of H2 flow introduced in the -wave plasma
post-discharge chamber: relative content in [N] and [NH2] and primary amine selectivity
At the difference of the CNT treatment with addition of hydrogen directly in the
discharge, there is no evolution of the grafted nitrogen content, which remained relatively
constant around 7.9 % whatever the H2 flow added in the post-discharge chamber
(Figure 4-18). However, the total primary amine content increases continuously from 0.2% to
0.8%, resulting in an enhancement of the NH2 selectivity from 2.5% to 8.3%. The relative
percentage of oxygen atom decreases also with the hydrogen flow introduced in the postdischarge chamber but the decrease is less important with respect to the evolution recorded in
Table 4-5.
By contrast to the treatment of CNTs in the post-discharge of the Ar + N2 + H2 -wave
plasma, the quantity and the selectivity in primary amines do not saturate at higher hydrogen
flow values and increase continuously.
84
25
20
N
(-NH2)
(-NH2 selectivity)
15
4
10
3
2
1
0
10
20
30
40
50
60
Discharge hydrogen flow (sccm)

Figure 4 - 18 : Quantity and selectivity of primary amine groups and nitrogen content in function of H2 flow
introduced in the post-discharge chamber
The treatment of 0.5g of CNT-powder was thus carried out with addition of 56.5 sccm
of hydrogen gas in the post-discharge chamber. The CNTs were treated for 2h using standard
plasma parameters and, after their compression in pellet, they were derivatized with TFBA for
16h. XPS results are represented in Table 4-8.
H2
(sccm)
CNTs elemental composition

(at.%)
[N]
(%)1
[-NH2]
(%)2
-NH2
selectivity
(%)3
98.0
0.9
1.1
traces
0.9
traces
56.5
96.1
1.0
0.9
2.0
1.0
0.7
66.7

Table 4 - 8 : XPS analysis of CNTs in function of H2 flow introduced in the -wave plasma post-discharge chamber:
Like in the case of pressed CNT-pellets, the quantity of grafted nitrogen atoms remains
identical to the value obtained after the treatment of CNT-powder without addition of
hydrogen. The quantity of oxygen also presents the same behavior with a slight decrease from
1.1 to 0.9 at.% with addition of hydrogen. However, the quantity and the selectivity of
85
primary amines are very different compared to the CNT pellet treatment. Indeed, the
measured quantity of NH2 is equal to 0.7% resulting in a selectivity as high as 66.7%, which
means that 2/3 of grafted nitrogenated groups are constituted by primary amine functions.
A quantity of grafted NH2 of 0.7% is certainly high enough to initiate the polymerization of
lactone 27. Moreover, the low content of elements grafted on the CNTs during the -wave
plasma post-discharge treatment, 1 at.% of nitrogen and 0.3 at.% of oxygen, should limit the
damages on the their conjugated structure allowing the preservation of their physical
properties 28. The functionalized carbon nanotubes used for the preparation of nanohybrids and
nanocomposites will be thus treated under these specific conditions.
IV.A.3.f.
Stability of primary amines grafted on CNTs
A major concern when primary amines are grafted on a surface is their long term
stability, commonly referred to as ageing. For example, the density of primary amine
functions grafted on a polymer surface via plasma treatment can decrease by more than 50%
during the first few days 29,30 due to post-plasma oxidation that converts amines into amides 31
and due to an important thermodynamically driven reorientation of amine functions into the
polymer bulk18. It was thus important to verify if primary amines grafted on CNTs can
undergo this ageing process. A treatment of the CNT pellet was carried out using standard
plasma parameters with introduction of hydrogen flow of 56.5 sccm in the -wave plasma
post-discharge chamber. This treated CNT-pellet was divided in four parts, which were
derivatized by TFBA for 16h after different storage times under ambient atmosphere
(Figure 4-19).
No evolution in either quantity or selectivity in grafted primary amines with storage time has
been observed attesting for a good stability of these functions.
86
25
20
N
(-NH2)
(-NH2 selectivity)
5
4
15
10
3
2
-NH2 Selectivity (%)
Element Concentration (at.%)
1
0
25
50
75
100
125
150
0
175
Storage Time (h)

Figure 4 - 19 : Time-evolution of the surface-compositions of treated pressed CNTs pellets with storage time
under ambient atmosphere
In order to take into account such stability, it may be suggested that primary amine
functions could be stabilized by a delocalization of the nitrogen non-bonding electron pair
into the aromatic rings of the graphene structure of the carbon nanotubes. This effect can be
compared to the resonance delocalization observed in aniline (Figure 4-20).
Figure 4 - 20 : The different resonance structures of aniline
87
IV.A.4. Influence of treatment time on nitrogen and NH2 functions grafting

The time of the -wave plasma post-discharge treatment is one important parameter in
determining the quantity of grafted elements on the surface of the CNTs. The evolution of the
nitrogen content and primary amine functions quantity (and selectivity) was studied in
function of the treatment time. Both CNT powder and pellet has been studied in order to
control the influence of the surface area that interacts with reactive species.
In both experiments, the treatment was carried out with standard plasma parameters and with
introduction of hydrogen (56.5 sccm) in the post-discharge during the process for a time
ranging from 0 to 120 minutes. XPS values, measured for the CNT pellet treatment, are
represented in Table 4-9 and Figure 4-21.
Treatment
time
(min)

(at.%)
[N]
(%)1
[-NH2]
(%)2
-NH2
selectivity
(%)3
99.4
0.0
0.6
0.0
0.0
96.8
0.9
1.0
1.3
0.9
0.4
48.1
15
93.8
2.9
1.8
1.5
3.1
0.5
17.2
30
90
5.7
2.4
1.9
6.3
0.7
11.1
45
87.6
7.2
3.2
2.0
8.2
0.8
9.3
60
86.2
8.4
3.5
1.9
9.7
0.7
7.5
120
81.9
11.4
4.5
2.2
13.9
0.9
6.4

Table 4 - 9 : XPS analysis of CNTs pellet surface in function of -wave plasma post-discharge treatment time:
Saturation in the content of grafted primary amine is observed after few minutes of treatment
of the CNT pellet while the nitrogen content increases with the reaction time. The resulting
primary amine selectivity presents a sharp decrease from 48.1% to 11.1% between 5 and 30
minutes and then slightly decreases at longer treatment time (Figure 4-21).
88
60
N
(-NH2)
(-NH2 selectivity)
15
55
50
45
40
35
30
10
25
20
15
20
10
5
0
10
20
30
40
50
60
70
80
90
100 110 120
Treatment duration (min)

Figure 4 - 21 : Evolution of nitrogen content and primary amines quantity and selectivity on CNT pellet
for increasing treatment time
The results obtained after treatment of 0.5g of the CNT powder are presented in Table
4-10 and the Figure 4-22.
CNTs surface elemental composition
(at.%)
Treatment
time
(min)
[N]
(%)1
[-NH2]
(%)2
-NH2
selectivity
(%)3
99.4
0.0
0.6
0.0
0.0
0.0
15
98.4
0.3
0.7
0.6
0.3
0.2
66.7
30
97.9
0.5
0.8
0.8
0.5
0.3
53.3
45
97.3
0.6
0.8
1.3
0.6
0.4
72.2
60
96.8
0.8
0.9
1.5
0.8
0.5
62.5
90
96.5
0.8
1.0
1.7
0.8
0.6
70.8
120
96.1
1.0
0.9
2.0
1.0
0.7
66.7

Table 4 - 10 : XPS analysis of CNTs in function of -wave plasma post-discharge treatment time:
In the case of the CNT powder treatment, the content in nitrogen and primary amines
changes in the same proportion with the treatment time. The selectivity in grafted primary
amines remains relatively high and constant at ca. 66%.
89
N
(-NH2)
(-NH2 selectivity)
2,0
1,6
80
1,4
1,2
60
1,0
0,8
40
0,6
0,4
1,8
100
20
0,2
0,0
10
20
30
40
50
60
70
80
90
100 110 120
Treatment duration (min)

Figure 4 - 22 : Evolution of nitrogen content and primary amines quantity and selectivity on CNT powder
with the treatment time
Two different trends in the evolution of the content in nitrogen and primary amines can be
observed for CNT-powder and pellets. In Figures 4-21 and 4-22, we can observe that the
primary amine selectivity has a similar value after few minutes of the CNT pellet treatment
and for the whole time range in the treatment of CNT powder.
It is known that the covalent functionalization of carbon nanotubes is performed preferentially
onto defects present along the nanotube structure 32,33. One hypothesis can be extracted from
these results: at the beginning of the treatment, atomic nitrogen N and reduced nitrogen
radicals such as NH and NH2 react with structural defects of CNTs. This first part of the
treatment consists thus to a functionalization via defect-side chemistry. When all defects are
saturated, the process continues via a sidewall functionalization where reactive species create
new defects on the surface of CNTs. However, during this second phase, it can be suggested
that only atomic nitrogen displays enough reactivity to create new defects. The difference of
reactivity between these radical species would explain the decrease of primary amine
selectivity during pellet treatment with atomic nitrogen continuing to graft nitrogenated
functions on CNTs while the interaction with NH and NH2 species, and thus primary amines
grafting, would be reduced. Indeed, we have demonstrated that molecular hydrogen can not
react with the nitrogen radicals present on the surface of the CNTs and, thus, the grafting of
90
atomic nitrogen does not allow the incorporation of primary amine functions. In the case of
the CNT powder, the overall surface is by far too much important for treatment time to reach
the sidewall functionalization phase. Moreover, carbon nanotubes are permanently agitated,
involving a limited residence time of the active species for one given CNT, which does not
trigger the creation of new defects by atomic nitrogen.
The functionalization of 0.5g of acid treated CNT powder, which presents a high
density of defects, was carried out using the same -wave plasma post-discharge treatment
conditions. A higher quantity of nitrogen is grafted ([N] = 2%) with a selectivity in primary
amine functions around 60% ([-NH2] = 1,2%) indicating the influence of the structural defects
on the efficiency of the CNT amino-functionalization.
IV.A.5. Conclusion
Surface treatment of CNTs in the Ar + N2 -wave plasma post-discharge is an original
way to readily functionalize in one step CNTs with nitrogenated functions. In particular, the
results give credit for an efficient and selective grafting of primary amines on their surface
when H2 is added in the post-discharge chamber along with Ar + N2 -wave plasma postdischarge treatment. The grafted primary amines proved stable in time.
In the case of the powder treatment, we can suggest that the amines grafting takes
place mainly at defects on the structure of CNTs, limiting any damages on the carbon
nanotubes and more likely preserving their physical properties such as electrical conductivity.
Moreover, due to the use of a diaphragm during the functionalization in the post-discharge,
the structure of carbon nanotubes is protected from UV irradiation and high energy particle
interaction during the treatment leading to the preservation of the CNTs morphology.
The introduction of primary amines on the nanotube surface by -wave plasma postdischarge treatment gives access to a large number of functional exploitations by the
consecutive transformation of the amine functions. Among others, this provides the unique
possibility to anchor groups for further initiation of polymerization reactions in order to
produce nanohybrids in which the CNT bundles can be disaggregated.
91
IV.B. Preparation and characterization of CNTs-g-PCL nanohybrids

Owing to its biodegradation and biocompatibility, poly(-caprolactone) (PCL) has
been widely studied in scientific research and used by industrials for producing either
biomaterials, e.g. for drug delivery systems, or environmentally friendly films and coatings,
e.g. in polyester urethane compositions 34. The grafting of PCL onto CNT surfaces via the
grafting from approach to form CNTs-g-PCL nanohybrids is very attractive. Indeed, PCL is
one of the most important biodegradable and biocompatible polymer easily synthesized,
paving the way to potential applications for so-functionalized carbon nanotubes in the field of
bionanotechnology 35. Moreover, PCL has the rarely property of miscibility with a variety of
commercial polymers such as polyvinyl chloride (PVC) 36, poly(styrene-co-acrylonitrile)
(SAN) 37 or chlorinated polyethylene (CPE) 38 for example. Therefore, the PCL chains grafted
at the surface of CNTs could represent a remarkable opportunity to disperse the nanotubes in
such matrices. In other words, PCL chains will act as interfacial compatibilizers 39,40. Several
examples of PCL grafting on the surface of particles such as silica nanoparticles 41, -Fe2O3
nanoparticles 42, titanate nanotubes 43 or carbon nanofibers 44, to form nanohybrids are reported
in the literature.
The covalent grafting of PCL on the carbon nanotube surface has already been carried
out from carbon nanotube surface via the grafting to 45,46 and the grafting from27,35
approaches where multistep and time-consuming purification steps proved necessary to graft
reactive functional groups such as carboxylic acids on the CNT sidewall, which in some case
led to some physical deteriorations of the CNTs.
PCL grafting from primary amine functions grafted on the surface of CNTs as treated
by -wave plasma post-discharge treatment has thus been carried out in order to yield CNTsg-PCL nanohybrids presenting a significant disaggregation of the carbon nanotube bundles
while preserving the CNT structure and physical properties.
IV.B.1. Preparation of CNTs-g-PCL nanohybrids

The second step of this work consisted thus in the Ring Opening Polymerization
(ROP) of -caprolactone (-CL) from the primary amine groups grafted on the CNTs surface
by -wave post-discharge treatment. The presence of these functions allows the direct
grafting of poly-(-caprolactone) from the surface of CNTs. Indeed, primary amine functions
92
are known to be effective initiators for ROP of cyclic esters in the presence of
triethylaluminium through a coordination-insertion mechanism 47. The alkylaluminium
activates the carbonyl group of the monomer and accordingly favors the nucleophilic addition
of the primary amine function, which is the actual initiation step (Figure 4-23).
AlEt3
O
NH2
H
N
OAl
+ C 2H 6
(= ROAlEt2)
Figure 4 - 23 : ROP initiation step of -caprolactone ROP from primary amines grafted on CNT surface
The resulting grafted PCL chains are thus linked by an amide bond to the carbon
nanotube surface.
After the reaction of the first -caprolactone molecule, an aluminium alkoxyde is formed
which subsequently promotes the polymer chain propagation through further -CL monomer
insertion into the Al-OR bond (Figure 4-24).
The coordinative ROP of -CL using aluminium alkoxyde as initiator is known to be a
living/controlled polymerization when carried out in homogeneous solution. The living term
is related to polymerization processes without irreversible chain transfer and termination 48.
93
Al
O
R
O
AlO R
OR
O
Al
O
RO
RO
Al
O
OAl
(CH2)5
O
n-2
H3O+
O
H
N
O
H
Figure 4 - 24 : Coordination-insertion mechanism of ROP of -CL with aluminum alkoxyde

leading to the formation of CNTs-g-PCL nanohybrids
94
The amino-functionalized CNTs have thus been produced via Ar + N2 -wave plasma
post-discharge treatment, in presence of hydrogen, allowing the grafting of 1% of nitrogen
groups whose 2/3 proved to be primary amine functions. A negligible quantity of grafted
hydroxyl groups (estimated to less than 0.1% by XPS) can also initiate the ROP of -CL. The
amino-CNTs (0.5g) were suspended in toluene and then sonicated for few seconds to reach
good dispersion. 1 ml of AlEt3 (1M) was added to the suspension to activate the primary
amine functions. Then, 4 ml of -CL (4.12g) was added to the solution and the ROP of -CL
occurred for 24h, the reaction system became viscous, suggesting that the polymerization took
place. The reactive middle was poured in a large excess of heptane to selectively precipitate
the PCL-grafted CNTs and any free PCL that could be formed directly in solution via
homogeneous initiation from hydroxyl impurities present in the medium. After drying, the
obtained gray powder contained a mixture of CNTs-g-PCL nanohybrids and free PCL
chains. The overall monomer conversion, as determined by gravimetry, was estimated to
85%. The free PCL chains could be removed out by washing of product with toluene, via
Soxhlet extraction for one day, allowing to selectively recovered CNTs-g-PCL nanohybrids.
The grafting from approach and thus the ROP of -CL onto the CNT surface are
expected to form PCL domains grafted and precipitated on the nanotube surface
(Figure 4-25). Accordingly, separation of the CNTs from each other should take place making
easier their dispersion within polymer matrices by melt blending.
H
O
(CH2)5
Surface
NH2
1. AlEt3
2. -CL
O=C
NH
3. H3O+
4. Soxhlet
extraction
MWNT
domains
PCL islets
Figure 4 - 25 : Scheme for -caprolactone ROP as initiated by primary amine groups grafted onto the CNT surface
95
IV.B.2. CNTs-g-PCL nanohybrids characterization

IV.B.2.a. Content in PCL grafted on CNTs surface and free PCL chains
Figure 4-26 shows the thermal degradation (under helium atmosphere) of pristine
CNTs, amino-functionalized CNTs (f-CNTs) and CNTs-g-PCL nanohybrids, before and after
Soxhlet extraction of free PCL chains with toluene. Thermogravimetrical analysis (TGA) is
an extensively used method to determine the amount of organic groups on the functionalized
carbon nanotubes 49. TGA have been also used to determine the relative amount of polymer
grafted onto the nanotubes in nanohybrids.
CNTs
f-CNTs
+ CNTs-g-PCL
CNTs-g-PCL + PCL
Figure 4 - 26 : Thermogravimetric analyses under helium flow (at 20K/min) of pristine CNTs (),
functionalized CNTs (), CNTs-g-PCL nanohybrids without (+) and with free PCL chains ()
The pristine CNTs are stable until 550C with a negligible weight loss of only 1 wt%, which
might arise from decomposition and volatilization of impurities. Amino-functionalized CNTs
present a weight loss starting slightly at 200C, with a maximum decomposition rate at
350C. The difference in the weight loss between pristine and functionalized CNTs can be
attributed to the decomposition of the functional groups grafted on their surface16. The degree
96
of functionalization of CNTs can thus be obtained by measuring the difference in weight loss.
In the temperature range from 200 to 450C, f-CNTs showed a weight loss of 2.8 wt%
corresponding to nitrogenated and oxygenated functional groups. It is worth reminding that
the atomic elemental composition of these functions was constituted by 1% of nitrogen and
0.9% of oxygen as determined by XPS analysis. If these atomic percentages are normalized
by their atomic mass on the atomic mass of carbon, the weight loss of nitrogen and oxygen
elements is estimated to 2.4wt% (1.2 wt% for each element). If we consider the hydrogen
element, which is not directly detectable by XPS analysis and is present at the level of 0.7 %
in -NH2, the weight loss attributed to all functional groups can be estimated to 2.5 wt%. This
value is thus very close to the weight loss determined by TGA. It is important to note that
hydroxyl content, which can also initiate the ROP of -CL, can be assumed as negligible.
Indeed, the deconvolution of XPS O1s peak shows that the majority of oxygenated groups are
carbonyl and carboxylic functions and less than 0.1% correspond to -OH groups.
From these results, the weight percentage of primary amine groups is estimated to 1.1 wt%.
This corresponds to ca. 0.7 mmol NH2 group per gram of f-CNTs, or ca. 8.8 primary amine
groups per 1000 carbon atoms. As the CNTs are constituted on average of 9 walls, we can
consider that there are about 8 NH2 groups per 100 carbons on the CNTs surface.
In the case of nanohybrids, the polyester chains, grafted and free PCL, are degraded in the
range from 200 to 350C, which fits the degradation temperature range reported for pure
PCL 50. Thus, for each sample, the residual fraction above 450C corresponds to CNTs.
With such a wide difference in thermal stability, it is possible to measure the amounts of
polymer and CNTs in a given sample. Accordingly, these results indicate that about 20 wt%
of the crude CNTs-g-PCL nanohybrids are composed by CNTs. After Soxhlet extraction of
free PCL chains with toluene, the grafted PCL accounts for ca. 25 wt%.
To determine the number-average molecular weight (Mn) of grafted polymer chains, a
solution consists to break the bond between CNTs and the grafted polymer and to analyse the
so-formed homopolymer by Gel Permeation Chromatography (GPC) 51. In our system, this
measurement method is not possible because the polymer chains are linked via a covalent
amide bond onto the CNT surface and amide bonds are known as more stable than ester
bonds. However, Mn and polydispersity index of free PCL chains, i.e., as recovered by
selective solubilisation in toluene, were determined by GPC to get some indications about the
97
reactive system. The Mn of free PCL chains is equal to 12000 g/mol with a polydispersity
index (PDI) of 2.1.
Knowing the quantity of PCL chains covalently grafted per gram of CNTs and the primary
amine content per gram of CNTs, assuming that all NH2 groups participate successfully in
the initiation of -CL polymerization, the number average molecular weight of grafted PCL
chains can be estimated to 500 g/mol. This value is much lower that the 12000 g/mol of free
PCL chains. Such a difference shows the influence of the heterogeneity of the reactive system
on the ROP of -CL.
It is worth pointing out that the efficiency of the ROP initiation from the primary
amine functions has been highlighted by a tentative polymerization of -CL carried out in the
presence of pristine CNTs, thus without any amine functions. By TGA, it is recorded a weight
loss of about 0.5% (after Soxhlet extraction) confirming that in absence of NH2, no grafting
reaction can take place at the CNTs surface.
IV.B.2.b. Solvent dispersibility of CNTs-g-PCL nanohybrids

The addition of CNTs-g-PCL nanohybrids, with or without free PCL chains, in
chloroform solvent, thus a good solvent of PCL, is a simple method to characterize their
dispersion ability.
Since CNTs are fully insoluble in chloroform, a very fast sedimentation is observed 52.
Therefore, if longer sedimentation times are recorded in case of nanohybrids, this means that
the covalent grafting of the polyester chains allows for improving the dispersibility in
chloroform, individualizing the carbon nanotubes from each other. Purposely, 15mg of
pristine CNTs or nanohybrids, with and without free PCL chains, were added to 15ml of
chloroform followed by a sonication for 15 seconds. A remarkable difference in
sedimentation was observed as illustrated in Figure 4-27. While a quick sedimentation is
observed in chloroform, CNTs-g-PCL nanohybrids remained dispersed in this solvent for a
longer period of time. In presence of the free PCL chains, the crude CNTs-g-PCL
nanohybrids remained dispersed in chloroform for a still longer period of time.
98
Figure 4 - 27 : Suspension test in chloroform of CNTs (left), CNTs-g-PCL nanohybrids (middle)

and CNTs-g-PCL + free PCL chains (right) in toluene after 15 seconds (a), 1 minute (b) and 1 hour (c)
This observation shows the positive influence of grafted PCL chains on the dispersibility of
CNTs owing to CNTs pre-disaggregation and solvation of grafted PCL in chloroform.
Interestingly enough, the PCL-grafted carbon nanotubes presented even better dispersion
ability in presence of not covalently linked PCL chains. A similar effect could thus be expected
in polyester matrix. This behavior is promising for a good processing ability of the crude
nanohybrids (with free PCL chains), which is important for their use as CNT nanofillers in
polymer nanocomposites.
The morphology of the different nanohybrid samples was thus studied by TEM.
IV.B.2.c. Morphology of CNTs-g-PCL nanohybrids

TEM is a very useful tool to characterize the structure of polymer grafted on the
surface of carbon nanotubes. Figure 4-28 represents TEM images recorded over pristine
CNTs thus before the -wave plasma post-discharge treatment.
These pristine CNTs have an average diameter of ca. 9.5 nm and their average length is
approximately equal to 1.5 m as announced by the supplier (Nanocyl S.A.).
99
100 nm
Figure 4 - 28 : TEM images of pristine CNTs (www.nanocyl.com): general overview (a)

and zoom on entangled CNTs (b)
On TEM images, one can observe CNT aggregates and the zoom image presents entangled
CNTs.
Figure 4-29 shows a representative image of CNTs-g-PCL nanohybrids, after Soxhlet
extraction of free PCL chains.
100
200 nm
Figure 4 - 29 : Representative TEM image of CNTs-g-PCL
On the micrograph of Figure 4-29, it can be found that CNTs appear less entangled and an
important part of them have polymer islets on their surface. Some CNTs appears to be stained
with PCL and they were glued together at the intersection points. We can also observe some
nanotubes without polymer at their surface, which is probably due to their confinement in a
bundle during the either functionalization step or polymerization process and, thus, they can
not initiate the -CL polymerization from their surface.
100 nm
Figure 4 - 30 : TEM image of one CNT with PCL islets like pearl-necklace structure
Figure 4-30 shows a zoom on an isolated PCL-grafted CNT presenting a pearl-necklace

structure. The pearl-necklace structure of the PCL grafted CNTs shows also that the primary
amine groups functionalization took place efficiently not only at the tips but also over the
whole surface of CNTs.
The morphology of CNTs-g-PCL nanohybrids in presence of free PCL chains (thus
before Soxhlet extraction) has been studied as well as is presented on Figure 4-31. Indeed, it
is known that physisorbed polymer can also play some role in the pre-destructuration of the
CNTs 53.
101
1 m
200 nm
Figure 4 - 31 : Representative TEM images of CNTs-g-PCL nanohybrids still in presence of the free PCL chains
102
On Figure 4-31, less aggregated CNTs are observed with a high quantity of polymer, which
appears like spider web around nanotubes. The large black zones are also constituted by PCL.
A zoom on a carbon nanotube is presented in the next Figure 4-32.
200 nm
Figure 4 - 32 : TEM image of one CNT embedded by a thin layer of PCL
In Figure 4-32, two phases are observed with a different image contrast. The outer layer with
lower gray is assigned to the polymer phase which is composed by physisorbed and
chemisorbed PCL chains which form a layer of few nanometers around the nanotube. The
presence of this layer is highlighted by the drawing of two parallel white straight lines at three
different locations on the TEM image but it is observable along the whole nanotube. From
the different degree of gray, one can see thus a core-shell structure for the PCL-coated CNTs.
Morphological studies by TEM of the CNTs-g-PCL nanohybrids in the presence or not
of free PCL chains show that the CNTs are pre-disaggregated. The CNTs-g-PCL
nanohybrids, in presence of free PCL chains, are particularly disaggregated. These studies
confirm the previous observations in terms of solvent dispersion ability.
103
IV.B.3. Conclusion
In conclusion, CNTs have been successfully modified by PCL as result of the
grafting from polymerization of -caprolactone initiated from primary amine groups present
on the CNT surface.
Concerning the amination of the CNTs, TGA and XPS analyses allowed to estimate the
number of primary amine groups, i.e. 8 functions every 100 superficial carbon atoms.
Moreover, the TEM morphology study of CNTs-g-PCL nanohybrids showed that the primary
amine groups were successfully grafted onto the whole surface of CNTs and grafted by
polyester chains.
The CNTs-g-PCL nanohybrids are easily dispersed in chloroform, a good solvent of the
grafted CPL chains. Furthermore, in presence of free PCL chains, these nanohybrids proved
to be characterized by an even higher extent of pre-disaggregation. These results are
promising for the use of such pre-disaggregated CNTs-g-PCL nanohybrids as nanofillers in
polymer nanocomposites.
104
IV.C. Carbon nanotubes-based nanocomposites

In this work, the third challenge was to obtain an homogeneous dispersion of a tiny
amount of multi-walled carbon nanotubes (CNTs) in thermoplastic matrices by conventional
melt blending. The goal is to obtain polymer nanocomposites with enhanced electrical
properties at low CNTs content, without altering the other physical properties. For that
purpose and taking advantage of the nanotube pre-disaggregation, the previously synthesized
and characterized CNTs-g-PCL nanohybrids have been further considered as nanofillerenriched masterbatch to be dispersed/melt blended within two selected polymers.
The poly(-caprolactone) (PCL) and high-density polyethylene (HDPE) have been
chosen as matrices (Figure 4-33).
O
(CH2)5
b
n
CH2
CH2
Figure 4 - 33 : Molecular formula of (a) poly(-caprolactone) (PCL) and high density polyethylene (HDPE)
PCL is a linear aliphatic polyester characterized by a crystallinity of about 50%, a relative low
glass transition temperature (Tg ~ -60C) and it is produced by ring opening polymerization
(ROP) of -CL. The PCL can be degraded by hydrolysis or enzymatic cleavage of its ester
linkages along the macromolecular chains in physiological conditions 54. While it has good
water, oil, solvent resistance, PCL is fully biodegradable, biocompatible and non toxic to
living organisms27. This polyester is therefore known for applications in the biomedical field,
such as its use in the development of controlled drug delivery systems as well as in surgical
sutures and other resorbable fixation devices 55. We have also mentioned that PCL displays the
rarely met property to be miscible (or compatible) with a large range of polymers such as
Polyvinyl
chloride
(PVC)36,
Poly(styrene-co-acrylonitrile)
(SAN)37
or
chlorinated
polyethylene (CPE)38, for example. Main disadvantages of PCL are its low melting
temperature (Tm ~ 60C), low modulus and abrasion resistance. Moreover, its poor barrier
properties to water and gases are also a further obstacle on its use in biodegradable packaging
materials. One solution to expand its potential applications may rely upon the incorporation of
finely dispersed carbon nanotubes into PCL matrix. In the literature, three commonly used
methods have been envisaged. The first is the in-situ polymerization where CNTs are added
105
directly within the -caprolactone monomer (with our without a solvent) and then the
polymerization is activated either thermally or in presence of a catalyst 56. The CNTs are
mechanically dispersed into the solution and upon polymerization reaction, in some cases,
they can be chemically linked to the polymer chains 57,58. The second technique is the solution
method where a solvent is used to dissolve PCL chains and CNTs are added to the resulting
solution. The solvent is then evaporated involving the formation of the polyester/CNTs
composite 59. However, the solvent assisted mixture of PCL or -CL with CNTs proved to
result in poorly dispersed system. Surfactants can be used to improve the dispersion ability of
CNTs in the solvent, however with limited positive effect 60. The third method consists of the
melt processing where CNTs are mechanically dispersed into molten PCL 61,62. The central
idea is to use fluid shear forces to break CNTs aggregates or prevent their formation in the
final nanocomposite materials.
Polyethylene (PE) is also a semi-crystalline polymer and it is among the most important
commercial thermoplastics. PE is used in a wide variety of applications because, based on its
structure, it can be produced in many different forms. The first type of PE to was
commercially exploited is called low density polyethylene (LDPE) or branched polyethylene.
This polymer is characterized by a large degree of branching, forcing the molecules to be
packed rather loosely forming a low density material. LDPE is soft and pliable and has
applications ranging from plastic bags, soft containers, textiles, and electrical insulation, to
coatings for packaging materials. Another form of polyethylene differing from LDPE only in
structure is high density polyethylene (HDPE) or linear polyethylene. This form demonstrates
little or no branching, enabling the molecules to be tightly packed. HDPE is much more rigid
than branched polyethylene and is used in applications where rigidity is important. HDPE has
a good combination of properties such as toughness, stiffness, solvent resistance and electrical
insulation. HDPE is therefore used as an insulating material for electric wire or for producing
different containers like mate plastic tubing, bottles, or bottle caps, for example 63. There are
other forms of polyethylenes based on different density and branching depending of the aimed
applications but their description is out of the scope of this work.
In the literature, HDPE/CNTs (nano)compositions are generally prepared via melt
processing 64,65,66. Another technique is the incorporation of CNTs into polyethylene by high
energy ball milling 67.
106
In this work, the combination of two methods is proposed in order to prepare PCL or
HDPE nanocomposites filled with finely dispersed CNTs. The first step is the predisaggregation of amino-functionalized carbon nanotubes via the grafting of PCL chains
directly on their surface to get CNTs-g-PCL nanohybrids. The second step is the use of the
formed nanohybrids like nanofiller-enriched masterbatch, which is then dispersed in the
polymer matrices via conventional melt processing. The used CNTs-g-PCL nanohybrids are
constituted by 20wt% of CNTs and 80wt% of PCL of which ca. 10wt% are covalently grafted
on the CNT surface. Comparative systems based on nanocomposites prepared via melt
processing of PCL with either pristine CNTs or amino-functionalized CNTs (anchoring of 1
at.% of nitrogenated groups with a primary amine groups selectivity of 67%) will be also
prepared.
The PCL and HDPE matrices have been chosen in function of their (non) miscibility
properties with the polyester chains thus grafted on the CNT surface. Since HDPE is not
miscible with PCL, this specific composition has been chosen for sake of comparison in order
to clearly discriminate between a simple dispersion ability of the surface-grafted nanotubes
(in non miscible HDPE matrix) and the interfacial compatibilization that can result from the
chain miscibility, entanglements and possibly crystallinity (in the PCL matrix). The goal is
thus to determine the influence of the CNTs-g-PCL nanohybrid approach on the physical
properties, and more especially on percolation threshold of this new type of nanocomposites.
The used polymer matrices are PCL-CAPA6500 (supplied by Solvay Interox) having a
Mn of ~50000g/mol and HDPE-FINATHENE 6410 (from Total Petrochemicals) having a
MFI of 33.
IV.C.1. Thermal properties of CNTs-based nanocomposites

IV.C.1.a. Influence of CNTs on the thermal degradation of nanocomposites
The thermal properties of polymer/CNTs composites have attracted much attention
from polymer scientists because of the generally reported increase of thermal stability with
respect to the pristine polymer matrix 68. Even if the general trends concern improved thermal
stability, one can find some reports attesting for some reduction of thermal stability. For
instance, Yang et al. 69 reported the effect of single-walled CNTs on the thermal stability of
107
acrylonitrile-butadiene-styrene (ABS)/CNTs composites and showed the degradation of these

composites at lower temperatures due to the destabilization of ABS by the addition of CNTs.
Probst et al. 70 reported that multi-walled CNTs promote also the degradation of polyvinyl
alcohol. In contrast and more generally, it has been reported that the addition of CNTs to
polypropylene (PP) 71, poly(methyl methacrylate) (PMMA) 72, poly(ethylene-co-vinyl acetate)
(EVA) 73 increases the thermal stability of the recovered nanocomposites. Clearly this
behavior is observed in the majority of polymer nanocomposites filled with CNTs and several
mechanisms of action have been reported so far 74. Indeed, dispersed CNTs could hinder the
flux of degradation product and therefore delay the onset of degradation. Moreover, the
polymer chains near the CNTs might degrade more slowly involving a higher temperature of
maximum weight loss. Another proposed mechanism attributes the improved thermal stability
to the effect of higher thermal conductivity in the nanotube/polymer composites that
facilitates heat dissipation within the entire volume of the composite materials 75.
In the following paragraphs, the influence of the different types of investigated carbon
nanotubes (CNTs, f-CNTs and CNTs-g-PCL nanohybrids) on the thermal stability of PCL
and HDPE matrices will be determined.
IV.C.1.b. Thermal degradation of PCL nanocomposites

The thermogravimetric analyses (TGA) under helium flow of the PCL nanocomposites
based on 3wt% of CNTs are presented in Figure 4-34 and the measured temperatures of
decomposition onset (Tonset) and at maximum weight loss rate (Tpeak) are gathered in
Table 4-11.
The dark blue line () in the Figure 4-34 represents the temperature dependence of the PCL
matrix weight loss. The TGA curve displays one main degradation beginning at 343C with
an inflection point at 414C. However, the derivative weight loss curve is slightly asymmetric
at lower temperature. Indeed, it is known that PCL presents a thermal degradation involving
two consecutive mechanisms50. The first process implies a statistical rupture of the polyester
chains via ester pyrolysis reaction (Figure 4-35).
108
PCL matrix
+ 3wt% CNTs
+ +3wt% f-CNTs
+ 3wt% CNTs-g-PCL
+ 3wt% CNTs-g-PCL + free PCL
Figure 4 - 34 : Thermogravimetric analysis curves of () PCL matrix, () PCL/CNTs, (+) PCL/f-CNTs,

() PCL/CNTs-g-PCL and () PCL/CNTs-g-PCL + free PCL with 3 wt% of inorganics; in helium at 20C/min
Figure 4 - 35 : First degradation step of PCL via pyrolysis reactions and subsequent chain cleavage
randomly distributed all along the chain50
109
The second principal step leads to the formation of -CL cyclic monomer as result of an
unzipping depolymerization process (Figure 4-36).
Figure 4 - 36 : Second degradation step of PCL via unzipping depolymerization process50
With the incorporation of pristine CNTs, the degradation is shifted towards higher
temperature at 424C, showing also the formation of a thermally stable char. The addition of
functionalized CNTs involves the same effect showing that the presence of primary amine
groups on the surface of CNTs does not influence the thermal stability of the nanocomposite.
The addition of CNTs-g-PCL nanohybrids, mixed with free PCL chains, causes the
nanocomposite to degrade at lower temperature contrary to the incorporation of pristine and
functionalized CNTs. Indeed, the formed nanocomposite begins to degrade at 260C against
343C for PCL matrix while the temperature of maximum weight rate decreases from 414 to
401C (Table 4-11). The presence of PCL chains (free and/or surface-grafted) in the
nanohybrid used for the production of the CNTs-filled composite thus significantly influences
the polyester matrix degradation.
Sample
Tonset
Tpeak
PCL matrix
343C
414C
+ 3wt% CNTs
359C
424C
+ 3wt% f-CNTs
357C
425C
+ 3wt% CNTs-g-PCL
310C
416C
260C
401C
Table 4 - 11 : The decomposition onset temperature, Tonset, and the temperature at maximum weight loss rate,
Tpeak, of the different CNTs filled PCL nanocomposites
110
To determine the impact of free PCL chains from the one of grafted PCL chains, the
thermogravimetric analysis of nanocomposites filled with the CNTs-g-PCL nanohybrid,
previously deprived of the free PCL chains, has been also carried out and is represented as the
red line
( ) in Figure 4 -34. The onset decomposition temperature appears in between the
values recorded for PCL filled with the crude nanohydrid (at lower temperature) and the
unfilled PCL (at higher temperature, at. 343C). The free PCL chains present a much lower
number-average molecular weight (Mn) with respect to the PCL matrix (~12000g/mol to be
compared to ~50000g/mol). As a consequence, more PCL chain ends are present and can
promote the degradation of PCL via the unzipping depolymerization mechanism. In the
CNTs-g-PCL filled nanocomposites, the grafted PCL chain ends can also degrade at lower
temperature. Accordingly, Tonset is decreased by about 30 and 80C when CNTs-g-PCL,
without and with free PCL, respectively, are dispersed in PCL while Tpeak is decreased by ca.
10C in compositions filled with the crude CNTs-g-PCL (thus with free PCL chains) and
remains constant in absence of any free (short) PCL chains. This broad difference in the
evolution of degradation temperatures can be clearly attributed to the quicker unzipping
depolymerization of the introduced short PCL chains (free or surface-grafted).
Another factor that can affect the thermal stability of the so-produced nanocomposites is the
presence of residual polymerization catalysts, which can promote thermal degradation of
PCL 76. To check whether aluminum residues have such an effect on the thermal (in)stability
of PCL, commercial PCL (CAPA 6500) was mixed with triethylaluminium (AlEt3), the
relative content of which being comparable to the quantity used for promoting the -CL ROP
in presence of functionalized CNTs. Interestingly, no significant difference in TGA curves
could be observed meaning that the presence of aluminum catalyst residues (at least in the
conditions used in this study) does not modify the thermal stability behavior of the PCL
matrix.
IV.C.1.c. Thermal degradation of HDPE nanocomposites

The thermogravimetric analyses (TGA) under helium flow of the HDPE nanocomposites
based on 3wt% of CNTs are presented in Figure 4-37 and the measured temperatures of
decomposition onset (Tonset) and at maximum weight loss rate (Tpeak) are represented in
Table 4-12.
111
HDPE matrix
+ 3wt% CNTs
+ +3wt% f-CNTs
+ 3wt% CNTs-g-PCL + freePCL
Figure 4 - 37 : Thermogravimetric analysis curves of () HDPE matrix, () HDPE/CNTs, (+) HDPE/f-CNTs,

and () HDPE/CNTs-g-PCL + free PCL with 3 wt% of inorganics in helium at 20C/min
Sample
Tonset
Tpeak
HDPE matrix
372C
460C
+ 3wt% CNTs
439C
491C
+ 3wt% f-CNTs
442C
260C and 441C
492C
354C and 490C
Table 4 - 12 : The decomposition onset temperature, Tonset, and the temperature at maximum weight loss rate,
Tpeak, of the different CNTs filled HDPE nanocomposites
The degradation of the pure HDPE matrix starts at 372C with a maximum weight loss rate at
460C. The mechanism of thermal degradation of polyethylene has been shown to occur via
random scission type reactions 77. After the incorporation of pristine CNTs, the degradation is
shifted towards higher temperature (491C), showing also the formation of a thermally stable
char. The Tonset and Tpeak of f-CNTs filled nanocomposite are identical to the values recorded
112
for CNTs filled nanocomposite proving that the presence of primary amine groups on the
surface of CNTs does not influence the thermal stability of the nanocomposite.
The nanocomposites filled with CNTs-g-PCL, with free PCL chains, present two degradation
steps. The first step can be attributed to the degradation of short PCL chains with a Tpeak very
close to the degradation temperature (at 350C) of the crude nanohybrids containing free PCL
chains. The second degradation step corresponds to the degradation of polyethylene. The
Tonset and Tpeak of this second degradation are equivalent to the values measured for CNTs and
f-CNTs filled nanocomposites showing that the degradation of PCL chains has no effect on
the thermal stabilization role of CNTs in the HDPE matrix.
In conclusion, we have observed that the incorporation of each type of studied CNTs
increases the thermal stability of PCL and HDPE matrices except for the introduction of
CNTs-g-PCL nanohybrid in the PCL matrix where the presence of shorter PCL chains favors
polyester degradation via depolymerization mechanism.
IV.C.1.d. DSC analyses of PCL nanocomposites

The DSC analyses of PCL nanocomposites have been carried out to determine the
influence of CNTs and derived/functionalized nanofillers on the melting temperature (Tm) and
crystallinity rate (c) of the PCL matrix.
Hereafter are presented the thermograms obtained for PCL nanocomposites. The Tm of CNTs
filled PCL nanocomposites and neat PCL are obtained from DSC thermograms as shown in
Figure 4-38.
113
PCL matrix
+ 3wt% CNTs
+ +3wt% f-CNTs
Figure 4 - 38 : DSC curves of neat PCL and PCL nanocomposites filled with 3wt% of CNTs, f-CNTs
and CNTs-g-PCL nanohybrids and a zoom on the melting zone
The measured Tm and crystallinity degree (c) determined from Figure 4-38 are summarized
in Table 4-13.
114
The crystallinity degree (c) of the PCL component within the nanocomposites has been
determined from the melting peak area as follows:
c =
(IV.1)
0
where is the enthalpy of fusion of the sample,
is the heat of fusion for 100%
crystalline PCL (taken as 136 J/g 78) and f is the weight fraction in PCL.
Sample
Tc
Tm
Hma
PCL matrix
35C
59C
70.9 J/g
52.2 %
+ 3wt% CNTs
45C
58C
71.6 J/g
54.3 %
+ 3wt% f-CNTs
44C
59C
71.9 J/g
54.5 %
44C
58C
72.1 J/g
54.7 %
Table 4 - 13: DSC data for PCL nanocomposites
From Table 4-13, the melting temperature of the nanocomposites is very close to the value
recorded for neat PCL at ca. 59C. Moreover, all CNT fillers promote the nucleation of PCL
crystallization with an increase of about 10C for the crystallization temperature as observed
in the literature60. However, the crystallinity degree seems to be only slightly increased with
the addition of the different CNTs.
An advanced thermal analysis method based on quasi-isothermal crystallization
experiments has been also used on neat PCL and CNTs filled nanocomposites 79. It is shown
that the incorporation of CNTs into the PCL matrix strongly increases the amount of
crystalline-amorphous interface. However, the crystal stability is reduced in the presence of
CNTs due to constraints imposed by the presence of the filler particles promoting the growth
of less stable subsidiary PCL crystal lamellae.
Interestingly, polarized optical micrographs show a modification of the crystalline
structure of PCL with addition of only 0.5 wt% of CNTs (Figure 4-39). Indeed, the presence
of CNTs promotes an overall smaller crystalline morphology of the PCL. This morphological
change can be partially induced by the nucleating effect of the CNTs. However, it can also be
related to the presence of inorganic barriers preventing the unconstrained growth of PCL
115
crystalline lamellae. This last effect can be especially pronounced in view of the high aspect
ratio of CNTs79. With the comparison of the thermal characteristics of PCL nanocomposites
filled with different nanofillers (such as POSS and sepiolite), it has been also shown that the
extent to which the thermal behavior of the PCL matrix is altered evolves strongly with the
nanoparticles aspect ratio.
Figure 4 - 39 : Polarized optical micrographs for unfilled PCL (a) and PCL filled with 0,5 wt% CNTs (b)
Scale bar: 50 m31
IV.C.1.e. DSC analyses of HDPE nanocomposites

The melting temperature (Tm) and the crystallinity degree (c) of CNTs filled HDPE
nanocomposites and neat HDPE have been also obtained from DSC thermograms as shown in
Figure 4-40. The measured values are presented in Table 4-14.
116
HDPE matrix
+ 3wt% CNTs
+ +3wt% f-CNTs
Figure 4 - 40 : DSC curves of neat HDPE and HDPE nanocomposites filled with 3wt% of CNTs, f-CNTs
and CNTs-g-PCL nanohybrids and a zoom on the melting zone
The crystallinity degree (c) of the HDPE component in the nanocomposites has been
0
determined via relation IV.1 with
, the heat of fusion for 100% crystalline polyethylene,
taken as 290 J/g 80 (Table 4-14).
117
Sample
Tc
Tm
Hma
HDPE matrix
121C
135C
215.1 J/g
74.2 %
+ 3wt% CNTs
121C
135C
217.1 J/g
77.2 %
+ 3wt% f-CNTs
122C
135C
220.0 J/g
78.2 %
43C
57C
11.4 J/g
69.9 %
122C
134C
165.8 J/g
67.3 %
Table 4 - 14 : DSC data for HDPE nanocomposites
The incorporation of CNTs does not influence Tc and Tm of HDPE matrix. However, the
crystallinity degree of HDPE seems to be increased with the addition of CNTs and f-CNTs.
On the contrary, the CNTs-g-PCL nanohybrids filled HDPE nanocomposite presents a lower
HDPE crystallinity degree. The endothermic melting and exothermic crystallization peaks of
the PCL phase can be also observed in the CNTs-g-PCL(+ free PCL)-based nanocomposite.
The measured Tc and Tm are close to the values obtained with PCL-based nanocomposites.
However, the crystallinity degree is around 15% superior to that one of PCL nanocomposites.
The increase of the PCL crystallinity degree with the decrease of the c HDPE can be
attributed to the presence of CNTs embedded within PCL and globally dispersed in the HDPE
matrix. The PCL is thus essentially in contact with CNTs surface promoting nucleation of
crystals. With the presence of PCL, the nanofillers present a more important diameter and
thus a diminution of the aspect ratio. Moreover, the dispersion of CNTs can be also improved.
The decrease of aspect ratio and the improvement of CNTs dispersion could prevent the
formation of large HDPE crystalline lamellae and could involve a decrease of the crystallinity
degree.
IV.C.1.f. Conclusion
The incorporation of the different CNT nanofillers in PCL and HDPE matrices improves their
thermal stabilities. One exception concerns the addition of CNTs-g-PCL nanohybrids, in
presence of free PCL chains, in the PCL matrix where the thermal degradation is observed at
lower temperature. This might be explained by the lower molecular weight of the PCL chains
(free or surface-grafted) known to undergo faster unzipping-type depolymerization reaction.
118
The DSC analyses demonstrate that the presence of CNTs induces morphological changes in
polymer matrices. The influence of the PCL coating on the crystallographic properties of
HDPE matrix has been also highlighted.
IV.C.2. Electrical properties of CNTs-filled nanocomposites

The multi-walled carbon nanotubes (MWNTs) are known to be electrically conductive
and to have an electrical conductivity that can reach a value of 103 S/cm 81. Compared to
composites filled with carbon black or carbon fibers as conductive reinforcements, polymeric
composites filled with CNTs can form conductive patch at a relatively very low content,
owing to their high aspect ratio of 100-1000 and high specific surface areas 82. Therefore, their
use in various applications, such as antistatic films or electromagnetic shielding materials, has
been anticipated.
In the next study, the effect of the primary amine grafting and the PCL coating on the
CNTs surface has been determined on the electrical properties of the corresponding polymer
CNTs nanocomposites.
IV.C.2.a.
Electrical properties of nanocomposites and percolation theory
The electrical conductivity of a material is defined as:
(IV.2)
where is the resistivity, commonly expressed in .cm and the unit of electrical conductivity
is S/cm. For a body of length L, cross sectional area A and resistance R, the electrical
conductivity is given by:
(IV.3)
In polymers with chemically saturated structures, the electrons are tightly bound in the
-bonds between atoms. Since all the available electrons are fixed in the -bonds, there are no
electrons free to carry an electrical current. Hence, saturated polymers are characteristically
119
electrical insulators. In order to achieve conduction in a saturated polymer material,

conductive pathways of particles are required throughout the polymer matrix. The formation
of this way can be described by the percolation theory. This theory quantitatively relates the
electrical conductivity of the composite to the volume fraction of the filler by this scaling law:
= 0 ( c) for V > Vc
(IV.4)
where 0 is the conductivity of the filler, V is the volume fraction of fillers. The critical
volume fraction Vc, also called percolation threshold, is the lowest concentration of filler that
forms continuous conductive pathways throughout the polymer matrix. The critical exponent t
is theoretical uniform for systems of the same dimensionality, it is a universal value. For twodimensional percolating systems, t is equal to 1.33 and for three-dimensional systems, t varies
between 1.6 and 2 depending on theoretical simulations 83. Other theories can be used to
model the electrical conductivity of polymer composites, like effective medium theory, but
they do not allow a so good correlation with experimental results than percolation theory 84. If
polymer and carbon nanotubes density are almost equal, the volume fraction can be replaced
by weight fraction in the equation IV.4.
Since the first observation of percolation-dominated electrical conductivity for a
composite filled with CNTs 85, a lot of polymer matrices have been investigated. CNTs present
unique electrical properties. In general, MWNTs are quite often found to be one-dimensional
conductors with a high electrical conductivity, in average 103S/cm, while conductivity of
SWNTs was experimentally found to vary between 104 and 3.106 S/m81. The metallic
properties of the MWNTs are due to their multiple-shell structure consisting of tubes with
various electrical properties, where additional electronic coupling between shells takes place.
Figure 4-39 shows the non-linear decrease of electrical resistivity of a composite with the
CNTs concentration. At low filler concentration, CNTs are separated from each other and the
electrical properties of the composite are dominated by the polymer matrix. The resistivity
decreases slowly with increasing CNTs concentration until the critical volume fraction is
reached.
At percolation threshold Vc, a very sharp fall of resistivity, and thus a very sharp jump of
conductivity, is obtained in a very small concentration range. From this region, the
conductivity and CNTs concentration have a power-law relationship as given by the
120
equation IV.4. Above percolation threshold, the electrical resistivity of the nanocomposites
levels off and does not decrease significantly with further addition of carbon nanotubes.
Figure 4 - 41 : Schematic representation of the electrical resistivity in carbon nanotube filled polymer composite
in function of filler loading fraction
Even if it allows a good approximation of the electrical conductivity evolution with the filler
concentration, the percolation theory presents some limitations. Firstly, it is rather
complicated to get absolute geometrical information by the estimate of t value 86. Indeed, the
results of statistical percolation theory are derived for ideal systems containing a
homogeneous dispersion of identical particles. The spread of CNTs properties, like length,
diameter, chirality, entanglement or waviness is not taken into account. So t value, obtained
with experimental results fitting, must be considered with precaution.
Secondly, in the theory, 0 value approaches the conductivity of the MWNT itself. However,
the contact resistance between CNTs or their bundles in the system is not considered in the
theory but it decreases the effective conductivity of the CNTs in the polymer matrix 87.
Moreover, electron tunnelling effect between CNTs can also participate to the limitation of 0
value and significantly interacts on the electrical conductivity of nanocomposites. In fact,
carbon nanotubes can be separated by insulating polymer chains that act as a potential barrier.
The electrical conductivity is therefore limited by hopping and/or tunneling of the charge
carriers between conductive fillers. Comparing different matrices, the nanocomposite
electrical conductivity variations are important (much higher than for different nanotube
121
types)37. This can only be explained with the distance dependence of tunneling through
polymer barriers between CNTs. The maximum tunneling distance of electrons is reported to
be about 1.8nm 88.
In polymer CNTs composites, the conduction of electrons by quantum tunnelling can
be described by thermal fluctuation induced tunneling (FIT) model 89, for constant
temperature, by the equation IV.5:
ln
(IV.5)
where w is the gap width in the tunnel junction. Connor and al 90 indicate w is proportional to
-1/3 if the CNTs are homogenously dispersed in the matrix. So, tunneling effect between
carbon nanotubes, separated by a thin isolating layer, should lead to dependence between
electrical conductivity and filler load:
ln 1/3
(IV.6)
This relation can be used to verify if the conduction of electron in nanocomposite systems is
formed by quantum tunnelling 91.
Moreover, nanocomposite electrical conductivity values below = 500.2.7 S/m (obtained by
semi-empirical way from largely published data86) are also indicative of the important
influence of polymer tunnelling barrier between CNTs.
When tunnelling effect is considered in theoretical calculations of percolation relation, the
exponent t loses his universal value and acquires strong filler concentration dependence 92.
This observation can explain the large spread of t value presented in the literature and shows
the importance that can have tunnelling effect in the electrical conductivity of
nanocomposites.
Bauhofer and Kovacs have condensed a large collection of published data in order to
get general dependencies of Vc (c), t and on many parameters like CNT type, CNT
synthesis method, CNT treatment and dimensionality, polymer type and dispersion method86.
Referring to published minimum percolation thresholds and maximum conductivities, type of
polymer and dispersion method seem to have more important effect on these values than type
and production method of CNTs. This review shows also the minimum percolation threshold
122
c tends to 0.1%wt for each type of polymer matrix, in function of the number of publications.
This observation confirms a relation obtained by excluded volume concept for a statistical
distribution of filler particles 93,94. For a homogeneous filler dispersion, these calculations
give:
c 1/
(IV.7)
where is the aspect ratio (length on diameter) of fillers. The typical aspect ration 1000
for CNTs reproduces the above mentioned value of 0.1wt%. This relation, which was also
determined by Monte Carlo simulation 95, shows the strong influence of the CNTs aspect ratio
on electrical properties of nanocomposites, the percolation threshold decreases with
increasing aspect ratio 96. However, lower critical concentrations are presented in the
literature 97,98 but these results can be attributed to another type of percolation, the kinetic
percolation. For this type of percolation, particles are free to move and thereby can form a
conducting network at much lower particle concentrations. Particle movement can be caused
by diffusion, convection, shearing or external fields allowing the flocculation of
homogeneously dispersed particles. So, kinetic percolation can not be linked to tridimensional statistical distribution of fillers in the nanocomposites. Recently, the two
percolation thresholds have been observed in the same CNTs/epoxy system 99.
IV.C.2.b. Electrical resistivity measurements range

The range of volume resistivity of prepared nanocomposites is very large varying between
insulating and moderately conductive materials. This broad range involves that resistivity
measurements might be carried out via two different systems. Two measurement systems are
thus used in this study : two-probe test method, where a multimeter is used to test moderately
conductive composite whose resistivity is in the range 1-107 .cm, and high resistivity meter
(Model 8009 test fixture resistance measurement) that allows to measure volume resistivity of
insulating materials up to 1018 .cm (Figure 4-42).
123
Figure 4 - 42 : Electrical resistivity measurement for polymer nanocomposites
The electrical resistivity was measured for nanocomposites filled with CNTs, f-CNTs and
CNTs-g-PCL nanohybrids (containing free PCL chains, see previous sections), in function
of the filler content.
IV.C.2.c.
Electrical properties of PCL nanocomposites
PCL/CNTs nanocomposites
Table 4-15 presents the measured resistivity values, and the so-calculated conductivity
values, for PCL nanocomposites filled with different quantities of pristine carbon nanotubes.
%CNTs
r (.cm)
(S/cm)
0
0.25
0.4
0.5
0.75
1
2
3
1.9E+12
8.8E+11
3.0E+04
1.0E+03
1.9E+02
1.1E+02
1.8E+01
1.2E+01
5.3E-13
1.1E-12
3.4E-05
9.7E-04
5.3E-03
9.2E-03
5.5E-02
8.2E-02
Table 4 - 15 : Electrical resistivity and conductivity values of PCL/CNTs nanocomposites
124
The resistivity of neat PCL, i.e., CAPA 6500 matrix, and PCL filled with 0.25 wt% of CNTs
are measured with the Model 8009 test fixture. The two-probe system was used for others
samples. The resulting measured electrical conductivities are presented in Figure 4-43.
1,E-01
1,E-03
(S/cm)
1,E-05
1,E-07
1,E-09
1,E-11
1,E-13
0
0,5
1,5
2,5
3,5
%wt CNTs
Figure 4 - 43 : Electrical conductivity of PCL nanocomposites in function of CNTs content
On Figure 4-41, it can be observed that the relation between nanocomposite electrical
conductivity and the filler content is typical of percolation relationship. The percolation
threshold c is estimated at 0.35 0.05 wt%.
To determine the t and 0 values, log10 of electrical conductivity values (superior to
percolation threshold) are placed in function of log10 (-c) to get a linear relation y = ax+b
where slope a is equal to t value and y-intercept b corresponds to log10 0 (Figure 4-44).
125
Log10
0
-0,5
-1
-1,5
-2
-2,5
-3
-3,5
-4
-4,5
-5
y = 1.89094 x 1.71956
-1,5
-1
-0,5
0,5
Log10 ( - c)
Figure 4 - 44 : Nanocomposite electrical conductivity in function of the difference between filler content and
percolation threshold in logarithmic scale
By the linear equation, 0 value can be evaluated with the relation log10 0 = - 1.72 and
directly t value corresponding to the slope value (Table 4-16):
1.9 0.1
0 (S/cm)
1.9 10-2 0.1 10-2
Table 4 - 16 : Critical exponent and CNTs electrical conductivity values for PCL/CNTs nanocomposites
PCL/f-CNTs nanocomposites
Table 4-17 presents the measured resistivity values, and the so-calculated conductivity values,
for PCL nanocomposites filled with different quantities of functionalized carbon nanotubes.
%f-CNTs
r (.cm)
(S/cm)
0
0.25
0.5
0.75
1
2
1.9E+12
1.5E+12
1.9E+03
2.3E+02
9.5E+01
1.3E+01
5.3E-13
6.6E-13
5.2E-04
4.3E-03
1.1E-02
7.5E-02
1.0E+01
9.6E-02
Table 4 - 17 : Electrical resistivity and conductivity values of PCL/CNTs nanocomposites
126
Figure 4-45 represents the electrical conductivity values (in logarithmic scale) in function of
CNTs content. The percolation threshold c is estimated at 0.35 0.05 wt%.
1,E-01
(S/cm)
1,E-03
1,E-05
1,E-07
1,E-09
1,E-11
1,E-13
0
0,5
1,5
2,5
3,5
%wt CNTs
Figure 4 - 45 : Electrical conductivity of PCL nanocomposites in function of f-CNTs content
For the determination of the t and 0 values, the same approach as for PCL/CNTs
nanocomposites has been considered (Table 4-18).
1.9 0.1
0 (S/cm)
2.2 10-2 0.1 10-2
Table 4 - 18 : Critical exponent and CNTs electrical conductivity values for PCL/f-CNTs nanocomposites
PCL/CNTs-g-PCL nanocomposites
Table 4-19 presents the measured resistivity values, and the so-calculated conductivity
values, for PCL nanocomposites filled with different quantities of CNTs-g-PCL nanohybrids
mixed with free PCL chains.
127
%CNTs
r (.cm)
(S/cm)
1.9E+12
5.3E-13
0.25
1.6E+12
6.1E-13
0.5
3.4E+03
2.9E-04
0.75
9.2E+02
1.1E-03
2.4E+02
4.1E-03
1.25
1.8E+02
5.6E-03
7.2E+01
1.4E-02
2.5E+01
4.0E-02
Table 4 - 19: Electrical resistivity and conductivity values of PCL/nanohybrids nanocomposites
Figure 4-46 represents the electrical conductivity values (in logarithmic scale) in function of
CNTs content in PCL nanocomposites filled with CNTs-g-PCL nanohybrids (and free PCL
chains). The percolation threshold c is also estimated at 0.35 0.05 wt%.
1,E-01
1,E-03
(S/cm)
1,E-05
1,E-07
1,E-09
1,E-11
1,E-13
0
0,5
1,5
2,5
3,5
%wt CNTs
Figure 4 - 46: Electrical conductivity of PCL nanocomposites in function of CNTs content in PCL composite
The resulting critical exponent t and filler electrical conductivity are represented in
Table 4-20.
t
1.7 0.2
0 (S/cm)
6.8 10-3 0.1 10-3
Table 4 - 20 : Critical exponent and CNTs electrical conductivity values for PCL/nanohybrids nanocomposites
128
Comparison of results and discussion

One can observe that measured percolation thresholds are identical for each type of
nanocomposites with a value of 0.35 0.05wt%. The comparison of the different curves
representing nanocomposite resistivity (in log scale) in function of inorganic filler content, for
each type of nanofillers, confirms this observation (Figure 4-47).
14
resistivity (.cm)
12
10
8
MWNTs
f-MWNTs
MWNTs-g-PCL
4
2
0
0
0,5
1,5
2,5
3,5
%wt CNTs
Figure 4 - 47 : Nanocomposite electrical resistivity in function of CNTs content in CNTs, f-CNTs
and CNTs-g-PCL nanohybrid filled PCL nanocomposite
The average aspect ratio value of CNTs is close to 160 with an average length of 1.5 m for a
diameter of 9.5 nm 100. The theoretical minimal percolation threshold c is thus estimated to
about 0.63 wt% from the relation IV-7. The measured percolation threshold value of 0.35
wt% for each PCL nanocomposite is lower than the theoretical value, which corresponds to
homogenous filler dispersion. This result means that CNTs, f-CNTs and nanohybrid
dispersions are optimal and the carbon nanotubes tend to be individualized in the PCL matrix.
Since carbon nanotubes, surface-treated or not, are homogenously dispersed in PCL
matrix (see next section), the relation IV-6 can be used to determine if a tunneling effect
influences the conductivity of the studied PCL nanocomposites. The graph of ln in function
of -1/3 is presented in Figure 4-48.
129
-1
-2
ln (S/cm)
-3
-4
MWNTs
-5
f-MWNTs
MWNTs-g-PCL
-6
Linaire (MWNTs)
-7
Linaire (f-MWNTs)
-8
Linaire (MWNTs-g-PCL)
-9
0,6
0,8
1,2
1,4
-1/3 (%wt)
Figure 4 - 48: ln in function of -1/3 for each type of PCL nanocomposite
For every studied nanocomposite, the linear relation indicates that their electrical conductions
are influenced by quantum tunnelling. Figure 4-48 shows that CNTs and f-CNTs filled PCL
nanocomposites have a very similar linear relationship. For CNTs-g-PCL filled composite, the
straight line presents the same slope but with lower electrical conductivity at the same CNTs
content (meaning at the same adjacent distance). The presence of grafted PCL chains on the
CNTs surface seems limiting the conductivity of CNTs within the PCL matrix, which is
confirmed by the recorded 0 and t values. Indeed, while t values are very close from each
other (1.7-1.9), the nanocomposites present different 0 values (Table 4-21).
Filler
CNTs
f-CNTs
CNTs-g-PCL
0 (S/cm)
1.9 10-2
2.2 10-2
6.8 10-3
Table 4 - 21 : Filler electrical conductivity values for each type of PCL nanocomposites
The 0 value of CNTs-g-PCL is two-fold lower than 0 values recorded of CNTs and f-CNTs.
The electrical conductivity of carbon nanotubes does not seem altered by the covalent aminofunctionalization in the post-discharge of microwave plasma and it is even slightly better.
130
Morphology of PCL nanocomposites

The TEM observations of the PCL nanocomposites filled with 3 wt% CNTs or CNTsg-PCL, with free PCL chains, are presented in Figure 4-49. As expected from the
aforementioned electrical properties, a good dispersion of CNTs is observed in both types of
PCL nanocomposites.
200nm
200nm
100nm
100nm
Figure 4 - 49: TEM micrographs at high magnification of PCL nanocomposites filled with 3wt% CNTs (a, c) and
3wt% CNTs-g-PCL+ free PCL (b, d)
It clearly comes out that CNTs can be finely dispersed in a polyester matrix as PCL.
However, the number of remaining CNT aggregates seems to be slightly superior in CNTs
filled PCL nanocomposites (Figure 4-49a) in comparison with the materials filled with CNTsg-PCL nanohybrids (Figure 4-49b). At higher magnification, finely dispersed CNTs are
131
observed in materials filled with CNTs-g-PCL nanohybrids (Figure 4-49d) while a CNT
aggregate appears in CNTs filled PCL nanocomposites (Figure 4-49c).
IV.C.2.d. Electrical properties of HDPE nanocomposites

HDPE/CNTs nanocomposites
The measured resistivity and conductivity values for HDPE filled with CNTs are
represented in Table 4-22.
%CNTs
r (.cm)
0
1
1.5
2
3
4
5
1.7E+17
5.6E+16
1.1E+16
4.3E+14
3.2E+04
3.1E+03
6.6E+02
(S/cm)
5.8E-18
1.8E-17
9.4E-17
2.3E-15
3.2E-05
3.2E-04
1.5E-03
Table 4 - 22 : Electrical resistivity and conductivity values of HDPE/CNTs nanocomposites
The electrical conductivity values are represented in function of CNTs content incorporated in
the HDPE matrix in logarithmic scale (Figure 4-50). The percolation threshold c is estimated
at 2.5 0.1 wt%.
1,E-03
(S/cm)
1,E-06
1,E-09
1,E-12
1,E-15
1,E-18
0
%wt CNTs
Figure 4 - 50: Electrical conductivity of HDPE nanocomposites in function of CNTs content
132
The critical exponent value t is determined by the measurement of the slope and the yintercept of log10 in function of log10(-c) (for values superior to c) allows the
calculation of the filler electrical conductivity 0 (Table 4-23).
t
2.4 0.2
0 (S/cm)
1.5 10-4 0.3 10-4
Table 4 - 23 : Critical exponent and CNTs electrical conductivity values for HDPE/CNTs nanocomposites
HDPE/f-CNTs nanocomposites
Table 4-24 gathers the measured electrical resistivity values and the so-calculated
conductivity values of HDPE filled with functionalized carbon nanotubes.
%CNTs
r (.cm)
(S/cm)
0
1
2
3
4
5
1.7E+17
9.4E+15
8.8E+13
8.4E+03
9.4E+02
2.3E+02
5.8E-18
1.1E-16
1.1E-14
1.2E-04
1.1E-03
4.3E-03
Table 4 - 24 : Electrical resistivity and conductivity values of HDPE/f-CNTs nanocomposites
Figure 4-51 represents the evolution of electrical conductivity of HDPE/f-CNTs

nanocomposites in function of f-CNTs content. The electrical conductivity is presented in
logarithmic scale.
133
1,E-03
(S/cm)
1,E-06
1,E-09
1,E-12
1,E-15
1,E-18
0
%wt f-CNTs
Figure 4 - 51 : Electrical conductivity of HDPE nanocomposites in function of f-CNTs content
The percolation threshold is estimated to 2, 5 0,1 wt%. By the study of the slope and the yintercept of log10 in function of log10(-c) (for values superior to c), the critical exponent
t and the filler electrical conductivity 0 has been also determined (Table 4-25).
2.5 0.1
0 (S/cm)
5.2 10-4 0.3 10-4
HDPE/CNTs-g-PCL nanocomposites
Electrical resistivity and conductivity values of HDPE/CNTs-g-PCL, with free PCL
chains, nanocomposites are presented in Table 4-26.
%CNTs
r (.cm)
0
0.5
1
1.25
1.5
2
3
1.7E+17
7.6E+16
7.4E+09
1.7E+05
1.6E+03
1.6E+02
7.8E+01
(S/cm)
5.8E-18
1.3E-17
1.4E-10
6.0E-06
6.2E-04
6.3E-03
1.3E-02
Table 4 - 26 : Electrical resistivity and conductivity values of CNTs-g-PCL filled HDPE
134
The evolution of electrical conductivity of HDPE/CNTs-g-PCL nanocomposites in function

of inorganic filler content is represented in Figure 4-52.
1,E-02
1,E-04
(S/cm)
1,E-06
1,E-08
1,E-10
1,E-12
1,E-14
1,E-16
1,E-18
0
0,5
1,5
2,5
3,5
%wt CNTs
Figure 4 - 52 : Electrical conductivity of HDPE/CNTs-g-PCL nanocomposites in function of inorganic content
The percolation threshold c is estimated at 1,1 0,1 wt%. By the same procedure as above
reported, the critical exponent t and the filler electrical conductivity 0 values have been
determined as well (Table 4-27).
2.2 0.2
0 (S/cm)
6.2 10-3 0.2 10-3
Comparison of results and discussion

Confirming data reported elsewhere67, 69,101,102 with percolation thresholds c varying
between 2.5 to 7wt% of CNTs in polyethylene matrix, the herewith electrical measurements
show that a good dispersion of carbon nanotubes in HDPE is difficult to be reached (c =
2.5wt%). The amination of CNTs surface does not influence the percolation threshold value
(c = 2.5wt%). But PCL grafting on CNTs surface has a strong impact on the dispersion of
CNTs in spite of the immiscibility of PCL with HDPE. The percolation threshold of HDPE
135
composite filled with CNTs-g-PCL is equal to 1.1 wt%, thus well below percolation threshold
value for CNTs-filled HDPE composites (Figure 4-53).
20
18
16
r (.cm)
14
12
10
MWNTs
f-MWNTs
MWNTs-g-PCL
4
2
0
0
%wt CNTs
Figure 4 - 53 : Electrical conductivity of each HDPE in function of CNTs content
For HDPE nanocomposites, the relation IV-6 can not be used anymore to determine if
the tunnelling effect can influence the electrical conductivity as it was observed in PCL
nanocomposites. Indeed, the dispersion of CNTs is far from being optimized in HDPE and the
measured percolation threshold values are largely superior to the theoretical value of 0.63
wt% expected for homogenously dispersion of CNTs as predicted by relation IV-7. However,
the measured 0 values for each type of HDPE nanocomposites appear low with respect to the
electrical conductivity of carbon nanotubes. These values seem to be influenced by electron
tunnelling transport.
Filler
CNTs
f-CNTs
CNTs-g-PCL
0 (S/cm)
1.5 10-4
5.2 10-4
6.2 10-3
Table 4 - 28 : Filler electrical conductivity 0 value for HDPE nanocomposites
Unlike PCL nanocomposites, the measured 0 value of the HDPE nanocomposites filled with
CNTs-g-PCL nanohybrid, with free PCL chains, appears higher. The measured 0 value
recorded on nanocompositions filled with f-CNTs is also slightly higher compared to pristine
136
CNTs-based materials, similarly to previously achieved observations in case of PCL

nanocomposites. The 0 values for CNTs and f-CNTs-filled nanocomposites are lower in
HDPE in regards to PCL matrix. This observation can be explained by the poor dispersion of
nanotubes in HDPE and also by electron tunnelling effect. In fact, the HDPE is one of the
most electrically insulating of polymers, 5.10-18 S/cm against 5.10-13 S/cm for PCL matrix.
Interestingly, the 0 value of HDPE-based composites filled with CNTs-g-PCL is identical to
the value measured in PCL matrix where the nanotubes proved to be finely dispersed. It
seems thus that CNTs-g-PCL fillers form an electrical conductive network within the HDPE
matrix. Only the presence of PCL at the nanotube surface (grafted of simply adsorbed) plays
the role of barrier in tunnelling junction between carbon nanotubes like schematically
represented in Figure 4-54.
PCL
HDPE matrix
Figure 4 - 54 : Schematic representation of HDPE nanocomposites filled with CNTs-g-PCL nanohybrids, where
nanotubes are partially embedded with PCL chains
The formation of this pathway can be easily explained by the better interaction between PCL
chains and carbon nanotubes due to partial PCL chains covalent grafting and the ability of
PCL to interact with CNTs as confirmed by their good dispersion in PCL matrix.
137
Morphology of the HDPE nanocomposites

The TEM observations of the HDPE nanocomposites filled with 3wt% of either CNTs
or CNTs-g-PCL, with free PCL chains are shown in Figure 4-55.
500 nm
500 nm
200 nm
200 nm
Figure 4 - 55 : TEM micrographs at low magnification of HDPE nanocomposites filled with 3wt% CNTs (a, c) and
3wt% CNTs-g-PCL+ free PCL (b, d)
Even though PCL and HDPE are known to be non miscible polymers, TEM analysis clearly
indicates that the CNTs-g-PCL nanohybrids are more homogenously dispersed in the HDPE
matrix than the pristine CNTs. Indeed, the HDPE/CNTs composites present large CNTs
aggregates and no detectable trace of individually dispersed nanotubes (Figures 4-53a and c).
In the images of HDPE nanocomposites filled with CNTs-g-PCL nanohybrids, some CNTs
aggregates can be observed but they appear with smaller mean size. Interestingly, many
138
individual CNTs can be detected on the whole micrograph (Figures 4-53b and d) attesting for
a better dispersion within the HDPE matrix.
These observations confirm the electrical conductivity measurements and show that the use of
the pre-disaggregated CNTs-g-PCL nanohybrids significantly facilitates the dispersion of
CNTs in the HDPE matrix.
IV.C.2.e. Conclusions
Electrical conductivity of CNTs-filled PCL and HDPE nanocomposites was characterized in
function of the CNTs content. Typical percolation behavior was observed for each type of
nanocomposites. CNTs display the ability to be homogenously dispersed in PCL matrix but
not in HDPE matrix. In fact, percolation threshold of PCL nanocomposites is even lower than
theoretical value (inverse of the CNTs aspect ratio) for optimal dispersion of CNTs. TEM
morphological study confirms the good dispersion of CNTs in the PCL matrix.
It was also shown that electron tunnelling effect has a strong influence on PCL
nanocomposite electrical conductivity and seems to play also a role in HDPE nanocomposites
conductivity.
The surface-amination of CNTs in the post-discharge of an Ar + N2 -wave plasma, in
presence of H2, does not influence the electrical percolation threshold in both polymer (HDPE
and PCL) matrices. This functionalization does not decrease the electrical conductivity of
carbon nanotubes, which seems even to be slightly higher.
The PCL grafting on the CNTs surface allows improving the CNTs dispersion in HDPE
matrix. This treatment decreases the electrical conductivity of nanotubes in PCL
nanocomposites. In HDPE matrix, the electrical conductivity shown by the dispersion of the
nanohybrids proved higher than the other two studied nanofillers (CNTs and f-CNTs). The
approach involving CNTs-g-PCL nanohybrids seems to create a conductive pathway in the
HDPE matrix and since PCL is known as less insulating than HDPE, electron flux crosses
more easily the PCL tunnelling barrier.
139
References
www.nanocyl.com, 2009-06
Godfroid T., Dauchot J.-P., Hecq M., Surf. Coat. Technol. 2003, 174-175, 1276-1281
3
Ruelle B., Licence memory thesis (UMH) 2005, Fonctionnalisation contrle de nanotubes de carbone et
caractrisation
4
Beamson G., Briggs D., High Resolution XPS of Organic Polymers: The Scienta 300 Database 1992, Wiley,
Chichester
5
Chen P., Wu X., Sun X., Lin J., Ji W., Tan K. L., Phys. Rev. Lett. 1999, 82, 2548-2551
6
Ruelle B., Felten A., Ghijsen J., Drube W., Johnson R. L., Liang D., Erni R., Van Tendeloo G., Peeterbroeck
S., Godfroid T., Dubois P., Hecq M., Bittencourt C., Micron 2009, 40, 85-88
7
Felten A., Bittencourt C., Pireaux J.-J., Van Lier G., Charlier J.C., J. Appl. Phys. 2005, 98, 074308
8
Lim S. C., Jo C. S., Jeong H. J., Shin Y. M., Lee Y. H., Samayoa I. A., Choi J., Jpn. J. Appl. Phys. 2002, 41,
5635-5639
9
Bjrkqvist M., Gthelid M., Grehk T. m., Karlsson U. O., Phys. Rev. B 1998, 57, 2327-2333
10
Hammer P., Lacerda R. G., Droppa R., Alvarez F., Diamond Relat. Mater. 2000, 9, 577-581
11
Xu T., Yang J., Liu J., Fu Q., Appl. Surf. Sc. 2007, 253, 8945-8951
12
Kalita G., Adhikari S., Aryal H. R., Ghimre D. C., Afre R., Soga T., Sharon M., Umeno M., Physica E 2008,
41, 299-303
13
Khare B. N., Meyyappan M., Cassell A. M., Nguyen C. V., Han J., Nano Lett. 2002, 2, 73-77
14
NANOCYLTM NC7000 series - Product Datasheet - 10 March 2009 - V05 (www.nanocyl.com)
15
Lejeune M., Durand-Drouhin O., Charvet S., Grosman A., Ortega C., Benlahsen M., Thin Solid Films 2003,
444, 1-8
16
Kaltofen R., Sebald T., Schulte J., Weise G., Thin Solid Films 1999, 347, 31-38
17
Straboni A., Pichon L., Girardeau T., Surf. Coat. Technol. 2000, 125, 100-105
18
Molyneux P., J. Sci. Technol. 2004, 26 (2), 211-219
19
Kennou S., Logothetidis S., Sygellou L., Laskarakis A., Sotiropoulou D., Panayiotatos Y., Diamond Relat.
Mater. 2002, 11, 1183-1187
20
Snyders R., Zabeida O., Roberges C., Shingel K. I., Faure M.-P., Martinu L., Klemberg-Sapieha J. E., Surf.
Sci. 2007, 601, 112-122
21
Choukourov A., Kousal J., Slavinska D., Biederman H., Fuoco E. R., Tepavcevic S., Saucedo J., Hanley L.,
Vacuum 2004, 75, 195-205
22
Favia P., Stendardo M. V., dAgostino R., Plasmas Polym. 1996, 1, 91-112
23
Meyer-Plath A. A., Schrder K., Finke B., Ohl A., Vacuum 2003, 71, 391-406
24
Chevallier P., Castonguay M., Turgeon S., Dubrulle N., Mantovani D., McBreen P.H., Wittmann J.-C.,
Laroche G., J. Phys. Chem. B 2001, 105, 12490-12497
25
Wagner C. D., Davis L. E., Zeller M. V., Taylor J. A., Raymond R. M., Gale L. H., Surf. Interface Anal. 1981,
3, 211
26
Zhi C. Y., Bai X. D., Wang E. G., Appl. Phys. Lett. 2002, 81, 1690-1692
27
Buffa F., Hu H., Resasco D. E., Macromolecules 2005, 38, 8258-8263
28
29
Truica-Marasescu F., Wertheimer M. R., Plasma Process. Polym. 2008, 5, 44-57
30
Ghasemi M., Minier M., Tatoulian M., Arefi-Khonsari F., Langmuir 2007, 23, 11554-11561
31
Gengenbach T. R., Chatelier R. C., Griesser H. J., Surf. Interface Anal. 1996, 24, 271-281
32
Dyke C. A., Tour J. M., Phys. Chem. A 2004, 108, 11151-11159
33
Bettinger H. F., J. Phys. Chem. B 2005, 109, 6922-6924
34
Goldberg D., J. Environ. Polym. Degrad. 1995, 3, 61-68
35
Zeng H., Gao C., Yan D., Adv. Funct. Mater. 2006, 16, 812-818
36
Lepoittevin B., Pantoustier N., Devalckenaere M., Alexandre M., Calberg C., Jrme R., Henrist C., Rulmont
A., Dubois Ph., Polymer 2003, 44, 2033-2040
2
140
37
Benali S., Olivier A., Brocorens P., Bonnaud L., Alexandre A., Bourbigot S., Espuche E., Gouanve F.,
Lazzaroni R., Dubois Ph., Adv. Mater. Sci. Eng. 2008, doi:10.1155/2008/394235
38
Benali S., Peeterbroeck S., Brocorens P., Monteverde F., Bonnaud L., Alexandre M., Lazzaroni R., Dubois
Ph., Eur. Polym. J. 2008, 44, 1673-1685
39
Sanchez C., Julian B., Belleville Ph., Popall M., J. Mater. Chem. 2005, 15, 3559-3592
40
Duquesne E., Moins S., Alexnadre M., Dubois Ph., Macromol. Chem. Phys. 2007, 208, 2542-2550
41
Carrot G., Rutot-Houze D., Pottier A., Degee Ph., Hilborn J., Dubois Ph., Macromol. 2002, 35, 8400-8404
42
Flesch C., Delaite C., Dumas P., Bourgeat-Lami E., Duguet E., J. Polym. Sci. Part A : Polym. Chem. 2004, 42,
6011-6020
43
Shi Z., Xueping G., Deying S., Zhou Y., Yan D., Polymer 2007, 7516-7522
44
Wang K., Li W., Gao C., J. Appl. Polym. Sci. 2007, 105, 629-640
45
Lou X., Detrembleur C., Sciannamea V., Pagnoulle C., Jrme R., Polymer 2004, 45, 6097-6102
46
Jana R. N., Cho J. W., J. Appl. Polym. Sci. 2008, 110, 1550-1558
47
Dubois Ph., Dege Ph., Jrme R., Teyssi Ph., Macromol. 1992, 25, 2614-2618
48
Penczek S., Biela T., Duda A ., Macromol. Rapid Commun. 2000, 21, 941-950
49
Gao C., He H., Zhou L., Zheng X., Zhang Y., Chem. Mater. 2009, 21, 360-370
50
Persenaire O., Alexandre M., Dege Ph., Dubois Ph., Biomacromol. 2001, 2, 288-294
51
Kong H., Gao C., Yan D., Macromol. 2004, 37, 4022-4030
52
Murakami H., Nakashima N., J. Nanosci. Nanotechnol. 2006, 6, 16-27
53
Bahr J. L., Tour J. M., J. Mater. Chem. 2002, 12, 1952-1958
54
Goldberg D., J. Environm. Polym. Degr. 1995, 3, 61-67
55
Uhrich K. E., Cannizzaro S. M., Langer R. S., Shakesheff K. M., Chem. Rev. 1999, 99, 3181-3198
56
Chrissafis K., Antoniadis G., Paraskevopoulos K. M., Vassiliou A., Bikiaris D. N., Comp. Sci. Technol. 2007,
67, 2165-2174
57
Saeed K., Park S.-Y., J. Appl. Polym. Sci. 2007, 104, 1957-1963
58
Mitchell C. A., Krishnamoorti R., Polym. 2005, 46, 8796-8804
59
Chen E.-C., Wu T.-M., Polym. Degr. Stabil. 2007, 92, 1009-1015
60
Mitchell C. A., Krishnamoorti R., Macromol. 2007, 40, 1538-1545
61
Wu D., Wu L., Sun Y., Zhang M., J. Polym. Sci. : Part B: Polym. Phys. 2007, 45, 3137-3147
62
Thomassin J.-M., Lou X., Pagnoulle C., Saib A., Bednarz L., Huynen I., Jrme R., Detrembleur C., J. Phys.
Chem. C 2007, 111, 11186-11192
63
Yang H.-S., Wolcott M. P., Kim H.-S., Kim S., Kim H.-J., Compos. Struct. 2007, 79, 369-375
64
Tang W., Santare M. H., Advani S. G., Carbon 2003, 41, 2779-2785
65
McNally T., Ptschke P., Halley P., Murphy M., Martin D., Bell S.E.J., Brennan G.P., Bein D., Lemoine P.,
Quinn J.P., Polymer 2005, 46, 8222-8232
66
Zou Y., Feng Y., Wang L., Liu X., Carbon 2004, 42, 271-277
67
Gorrasi G., Sarno M., Di Bartolomeo, Sannino D., Ciambelli P., Vittoria V., J Polym Sc B 2007, 45, 597-606
68
Beyer G., Fire Mater 2002, 26, 291-293
69
Yang S. Y., Castilleja J. R., Barrera E.V., Lozano K., Polym. Deg. Stab. 2004, 83, 383-388
70
Probst O., Moore E. M., Reasaco D. E., Grady B. P., Polymer 2004, 5, 4437-4443
71
Kashiwagi T., Grulke E., Hilding J., Harris R., Awad W., Douglas J., Macromol. Rapid Commun. 2002, 23,
761-765
72
Du F., Fischer J. E., Winey K. I., J. Polym. Sci.: Part B: Polym. Phys. 2003, 41, 3333-3338
73
Peeterbroeck S., Alexandre M., Nagy J. B., Pirlot C., Fonseca A., Moreau N., Philippin G., Delhalle J.,
Mekhalif Z., Sporken R., Beyer G., Dubois Ph., Compos. Sci. Tech. 2004, 64, 2317-2323
74
Moniruzzaman M., Winey K. I., Macromol. 2006, 39, 5194-5205
75
Huxtable S. T., Cahill D. G., Shenogin S., Xue L., Ozisik R., Barone P., Usrey M., Strano M. S., Siddons G.,
Shim M., Keblinski P., Nat. Mater. 2003, 2, 731-734
76
Abe H., Takahashi N., Kim K. J., Mochizuki M., Doi Y., Biomacromol. 2004, 5, 1480-1488
77
Bockhorn H., Hornung A., Hornung U., Schawaller D., J. Anal. Appl. Pyrol. 1999, 48, 93-109
78
Lepoittevin B., Devalckenaere M., Pantoustier N., Alexandre M., Kubies D., Calberg C., Jrme R., Dubois
Ph., Polymer 2002, 43, 4017-4023
141
79
Miltner H. E., Watzeels N., Goffin A.-L., Duquesne E., Benali S., Ruelle B., Peeterbroeck S., Dubois Ph.,
Van Mele B., Macromol. submitted
80
Runt J. P., Encyclopedia Polym. Sci. Eng. 1985, 4, 482-519
81
Ando Y., Zhao X., Shimoyama H., Sakai G., Kaneto K., Int. J. Inorg. Mater. 1999, 1, 77-82
82
Kim Y. J., Shin T. S., Choi H. D., Kwon J. H., Chung Y.-C., Yoon H. G., Carbon 2005, 43, 23-30
83
Lux F.; J. Mater. Sci. 1993, 28(2), 285301
84
Weber M., Kamal M.R., Polym Comp 1997, 18(6), 711-725
85
Coleman J.N., Curran S., Dalton A.B., Davey A.P., McCarthy B., Blau W., Barklie R.C., Phys Rev B 1998,
58(12), 7492-7495
86
Bauhofer W., Kovacs J.Z, Compos Sci Technol 2008, doi:10.1016/j.compscitech.2008.06.018
87
Ounaies Z., Park C., Wise K.E., Siochi E.J., Harison J.S., Comput Sci Tech 2003, 63, 1637-1646
88
Li C.Y., Thostenson E.T., Chou T.W., Appl Phys Lett 2007, 91(22), 223114
89
Sheng P., Phys Rev B 1980, 21(6), 80-95
90
Connor M.T., Roy S., Ezquerra T.A., Balta Calleja F.J., Phys Rev B 1998, 57(4), 2286-2294
91
Kim Y.J., Shin T.S., Choi H.D., Kwon J.H., Chung Y.C., Yoon H.G., Carbon 2005,43 (1), 23-30
92
Johner N., Grimaldi C., Balberg I., Ryser P., Phys Rev B 2008, 77, 174204
93
Balberg I., Anderson C.H., Alexander S., Wagner N., Phys Rev B 1984, 30(7), 3933-3943
94
Celzard A., McRae E., Deleuze C., Dufort M., Furdin G., March J. F., Phys. Rev. B 1996, 53, 6209-6214
95
Foygel M., Morris R. D., Anez D., French S., Sobolev V. L., Phys. Rev. B 2005, 71, 104201
96
Bai J.B., Allaoui A., Compos A 2003, 34(8), 689-694
97
Sandler J.K.W., Kirk J.E., Kinloch I.A., Shaffer M.S.P., Windle A.H., Polymer 2003, 44(19), 5893-5899
98
Bryning M.B., Islam M.F., Kikkawa J.M., Yodh A.G., Adv Mater 2005, 17(9), 1186-1191
99
Kovacs J.Z., Velagala B.S., Schulte K., Bauhofer W., Compos Sci Technol 2007, 67(5), 922-928
100
NANOCYLTM NC7000 series - Product Datasheet Thin Multi-Wall Carbon Nanotubes
101
Zhang Q., Rastogi S., Chen D., Lippits D., Lemstra P.J., Carbon 2006, 44, 778-785
102
Liang G.D., Tjong S.C., Mater Chem. and Phys 2006, 100, 132-137
142
V. GENERAL CONCLUSIONS &

PERSPECTIVES
Offering at the same time a high aspect ratio (length-to-diameter) and a low density,
CNTs show strong application potential as advanced filler materials to substitute or
complement the conventional nanofillers, such as nanoclays, in the fabrication of
multifunctional polymer nanocomposites. In this field, the extent of dispersion of the
individual CNTs is an essential parameter to produce nanocomposite materials with largely
improved properties. In order to homogenously disperse the CNTs throughout the polymer
matrix, the CNTs entanglements, formed during their synthesis, and the CNTs bundles,
stabilized by numerous intermolecular van der Waals forces, must be disrupted. One of the
most cited technique to overcome these agglomeration problems is the surface modification of
CNTs.
In this study, CNTs were functionalized via an original covalent surface treatment and the sografted primary amine groups were used as initiating sites for lactone polymerization forming
CNTs-g-polyester nanohybrids allowing the pre-destructuration of the CNTs agglomerates.
The CNTs-based nanocomposites have been prepared by the most simple preparation
technique, the melt blending, of polymer matrices with nanohybrids, used as highly filled
chemical masterbatches, and their thermal, electrical and morphological properties have been
characterized.
The first challenge of our research was to develop a controlled functionalization process
allowing an efficient and selective primary amine groups grafting without damages on the
CNTs structure and in particular on their electrical properties. To achieve this goal, an
original microwave plasma process was developed to graft nitrogenated groups on the CNT
surface in one single step. The use of plasma, which is an ionized gas consisting in ions,
electrons and neutral species, allows to treat the CNTs with a relative low time and
environmentally friendly process without any use of solvent. In our functionalization process,
the CNTs are submitted to an atomic nitrogen reactive flow formed in an Ar + N2 microwave
plasma. It was shown that this process allows grafting efficiently nitrogenated groups on the
surface of CNTs treated under pellet form. We have also verified the covalent character of the
bonds between nitrogenated groups and the CNT surface via high-resolution ultraviolet
V. GENERAL CONCLUSIONS & PERSPECTIVES
143
photoelectron spectroscopy (HR-UPS) carried out at the synchrotron of Hamburg. The

morphological analysis of treated CNTs showed that their graphitic sidewall structure was not
damaged by the treatment as observed by high-resolution transmission electron microscopy
(HR-TEM). Therefore, the increase of the primary amines selectivity grafting on CNTs
powder has been performed via the reduction of oxygen contamination and the addition of
hydrogen in the experimental set-up. The preliminary analyses of the nitrogenated functions
and primary amines grafting on the CNTs surface seemed to indicate that atomic nitrogen
(N) and reduced nitrogen species (NH and NH2) react firstly with defect sites of CNTs and,
then, only atomic nitrogen continues to react on the CNT surface creating defects.
In the second part of this work, the so-grafted amine groups have been used as initiation sites
for promoting the ring opening polymerization (ROP) of -caprolactone (-CL) yielding
polyester-grafted CNT nanohybrids. The quantity of grafted and free poly(-caprolactone)
(PCL) chains has been determined by thermogravimetric analysis (TGA). When compared to
CNTs-g-PCL nanohybrids and pristine CNTs, the CNTs-g-PCL nanohybrids, in presence of
free (non grafted) PCL chains, displayed the best dispersion ability in chloroform, a good
PCL solvent. The morphology of the recovered nanohybrids has been characterized by TEM
showing that PCL islets are homogenously dispersed along the CNT sidewalls. Interestingly
enough, the absolute quantification of primary amine groups have been performed via both
TGA and XPS analyses. The CNTs-g-PCL nanohybrids, mixed with free PCL chains,
presented the best CNTs pre-disaggregation and were chosen as nanofiller candidates.
Another challenge of our research was to disperse homogenously a tiny amount of CNTs in
PCL and HDPE matrices. Due to their pre-disaggregation state, it was hoped that CNTs-gPCL masterbatch increased the dispersion ability of CNTs in both polymer matrices.
Moreover, the last part of this work has also focused on the effect of PCL grafting on the CNT
surface on their dispersion ability. The pristine and functionalized CNTs filled
nanocomposites were also prepared as comparative materials.
It was observed that the incorporation of each type of CNTs increases the thermal stability of
the PCL and HDPE matrices except for the introduction of CNTs-g-PCL nanohybrid in the
PCL matrix where the presence of shorter PCL chains favors the PCL degradation via
unzipping-type depolymerization mechanism. DSC analyses demonstrated that the presence
of CNTs induces morphological changes in the polymer matrices and the influence of PCL
coating on the crystallographic properties of HDPE matrix. Finally, the electrical conductivity
of CNTs-filled PCL and HDPE nanocomposites have been characterized in function of the
144
CNTs content. Typical percolation behavior was observed for each type of nanocomposites.
CNTs displayed strong inherent ability to be homogenously dispersed in PCL matrix but not
in HDPE matrix. In fact, percolation threshold of PCL nanocomposites proved lower than the
theoretical value for optimal dispersion of CNTs. The TEM morphological study confirmed
the good dispersion of CNTs in the PCL matrix. It was also shown that electron tunneling
effect has a strong influence on PCL nanocomposite electrical conductivity and seems to play
also a role in conductivity of HDPE nanocomposites. The PCL grafting on the CNTs surface
involved the improvement of the CNTs dispersion in HDPE matrix. This treatment decreased
the electrical conductivity of CNTs in PCL nanocomposites, unlike within HDPE matrix
where electrical conductivity of nanohybrids is better than the other two nanofillers. The
CNTs-g-PCL nanohybrids created a conductive pathway in the HDPE matrix and, since PCL
is less insulating than HDPE, electron flux crosses more easily the PCL tunneling barrier.
The electrical properties of nanocomposites have been clearly improved by the incorporation
of CNTs-g-PCL nanohybrids in the PCL matrix, it would be interesting to determine the
mechanical performances of these nanocomposites and check the influence of covalent bonds
at the interface CNTs/polymer. The nanohybrid-filled HDPE matrix can be used as
comparative material to determine the influence of miscibility of grafted polymer chains on
these mechanical properties.
The results collected in the frame of this work can be used in several orientations.
Firstly, the amino-functionalization of the CNTs surface gives the opportunities for
manipulating the surface chemistry of CNTs. We have observed that the microwave plasma
post-discharge treatment allows an efficient and selective primary amine grafting without
damages to CNTs morphology and their intrinsic properties, in sharp contrast to a majority of
covalent functionalization processes.
The amino-CNTs present a lot of potential applications, in particular in nanobiotechnology.
Indeed, biological systems, including peptides, proteins, nucleic acids and other therapeutic
agents, can be attached to the amine functions allowing their use in several biological
applications such as drug delivery or biosensor for example. Moreover, amino-CNTs present
a good biocompatibility.
The amination of CNTs surface shows also an interest for reinforced polymer composites. In
this field, our proposed functionalization method provides a flexible and efficient route to
covalently link CNTs to the polymer matrix resulting in improved mechanical performance
without a decrease of the electrical properties. Another interest of our process is that the
145
primary amine functions are directly grafted on the CNT surface with any link between the
function and the graphitic sidewall. This direct link between primary amine functions and
CNT would be interesting for photovoltaic applications. Indeed, the incorporation of CNTs
into the conjugated polymer leads to great improvements in photovoltaic device performance
compared to polymer-only devices. The best improvement of these performances can be
obtained via the creation of covalent bonds between conjugated polymer and CNTs. However,
in the functionalization process reported in the literature, the CNTs present functions on their
surface with an intermediary saturated linker limiting the performances improvement effect.
A study of the grafting of poly(3-hexylthiophene) on CNTs treated via our functionalization
process show very encouraging preliminary results.
The grafting of amino groups onto CNTs is also a suitable method to obtain active and stable
basic catalysts for application in liquid phase reactions. For example, amino-grafted CNTs
have good heterogeneous catalytic performance for the transesterification of triglycerides.
The microwave plasma post-discharge treatment can be used to graft other types of functions.
Indeed, the number of organic functions that can be grafted on the surface of CNTs or other
nanoparticles is relatively large and can be carried out just by the appropriate choice of gas
introduced in the plasma. For example, hydroxyl moieties could be grafted on the CNT
surface via an Ar + O2 microwave plasma post-discharge treatment with introduction of
hydrogen in the post-discharge.
Interestingly enough, the presence of radicals on the CNTs surface after the treatment could
be used to initiate radical polymerization of a large variety of monomers to form nanohybrids
via an original solvent-free process.
The CNTs-g-PCL nanohybrids present also potential applications as chemical
masterbatches to disperse homogenously the CNTs in high resistive polymer matrices to yield
nanocomposites with decreased electrical percolation threshold and enhanced electrical
conductivity due to the PCL tunnel interjunctions between CNTs in the host polymer.
To conclude, this work shows that functionalization of CNTs via microwave plasma
post-discharge process can pave the way to a large range of applications such as the
formation of high electrical performance nanocomposites.
Florian Boon, final masters thesis 2009

Villa A., Tessonnier J.-P., Majoulet O., Su D. S., Schlgl R., Chem. Commun. 2009, 4405-4407
146
VI. EXPERIMENTAL PART &

INSTRUMENTATION
VI.A. Materials
The multiwalled carbon nanotubes (MWNTs) studied in this work were industrially produced
by Nanocyl S.A. (Nanocyl-7000), via a catalytic carbon vapour deposition (CCVD)
process1. Carbon nanotubes have a purity of min. 90% carbon. These nanotubes are
characterized by an average outer diameter of 10nm, corresponding to ca. 9-10 concentric
layers. The average length is ca. 1.5m.
The used polymer matrices are commercial poly(-caprolactone) (PCL), CAPA6500 (Solvay
Interox) with a Mn of ~ 50000g/mol, and high density polyethylene (HDPE), Finathene 6410
(Total Petrochemicals) with a MFI of 33.
VI.B. Preparation of samples

a) Amination of CNTs - Experimental microwave plasma set-up
The CNTs are treated in the post-discharge of an Ar + N2 microwave plasma sustained by a
surface wave launched in a quartz tube via a surfaguide supplied by a 2.45 GHz microwave
generator, which can be pulsed. Hydrogen gas can also be introduced either in the discharge
or in the post-discharge chamber. The experimental set-up is schematically represented in
Figure 6-1.
VI. EXPERIMENTAL PART & INSTRUMENTATION
147
1: microwave generator with impedance

tuning
2: surfaguide
3: discharge tube
4: Post-discharge chamber
5: gate valve with sampling hole
5: chopper;
6: quadrupole mass spectrometer
7: turbomolecular pump
8: to rotary pump and downstream
baratron;
9: gas inlet;
10, 11, 10, 11: inlets and outlets of
cooling water
12, 12: inlet and outlet of cooling oil
13: upstream baratron.
Figure 6 - 1 : Microwave plasma experimental set-up
The plasma is sustained by a surface wave launched in a quartz tube (inner diameter, 14 or 6
mm) via a surfaguide supplied by a 2.45 GHz microwave generator whose power is 2 kW
maximum and which can be pulsed (Muegge Electronic GmbH). The discharge tube is cooled
by silicone oil maintained at low temperature (278288 K), flowing between the discharge
quartz tube and an external concentric pyrex tube. The whole system is surrounded by a
grounded aluminium grid to preserve the laboratory from micro-wave leaks. The length of the
discharge is limited at the two extremities of the discharge tube by silver rings in contact with
the water cooled aluminium pieces which hold all the system tight. These silver rings reflect
the surface wave and therefore allow to deliver the power in a definite volume of the
discharge so that the power density can be varied. The gas flows are regulated by mass flow
controllers. The total flow can be varied from 500 to 2200 sccm. The exit of the discharge
tube to the post-discharge tube is a circular aluminium aperture whose diameter is equal to 9
mm. Two baratron gauges measure the upstream pressure at the entry of the gases and the
downstream pressure in the post-discharge tube. This latter is pumped by a rotary pump which
maintains the pressure at approximately 267 Pa. The mass spectrometer samples the post-
148
discharge composition through a 200 mm hole drilled in a 100 mm gate valve and is pumped
by a turbomolecular pump. Before each experiment, this gate valve is opened, the valve
between the post-discharge and the rotary pump is closed so that all the vessel can be pumped
down to a pressure lower than 1.5 10-4 Pa.
For the treatment of powder, the carbon nanotubes were placed in a crucible directly mounted
on a vibrating speaker membrane in the post-discharge chamber. The CNTs were stirred
permanently during the treatment.
b) Preparation of CNTs-g-PCL nanohybrids

The amino-functionalized CNTs are produced via Ar + N2 -wave plasma post-discharge
treatment, in presence of hydrogen. The -caprolactone (-CL, 99%, Fluka) was dried over
calcium hydride (93+%, Acros) for 48 hours and, then, distilled under reduced pressure and
kept under nitrogen at 4C. The triethylaluminium solution (AlEt3, 25wt% in toluene,
Aldrich) was diluted in dried toluene to obtain a solution 1M which is stored under nitrogen
atmosphere.
The amino-CNTs (0.5g) were suspended in toluene and then sonicated for few seconds to
reach good dispersion. 1 ml of AlEt3 (1M) was added to the suspension to activate the primary
amine functions. Then, 4 ml of -CL (4.12g) was added to the solution and the ROP of -CL
occurred for 24h, the reaction system became viscous, suggesting that the polymerization took
place. The reactive middle was poured in a large excess of heptane to selectively precipitate
the PCL-grafted CNTs and any free PCL that could be formed directly in solution via
homogeneous initiation from hydroxyl impurities present in the medium. After drying, the
obtained gray powder contained a mixture of CNTs-g-PCL nanohybrids and free PCL
chains. The free PCL chains could be removed out by washing of product with toluene, via
Soxhlet extraction for one day, allowing to selectively recovered CNTs-g-PCL nanohybrids.
149
c) Preparation of nanocomposites - Melt blending

Polymer / MWNTs composites were prepared with a Minilab twin-screw mini-extruder
(ThermoHaake).
MWNT were dried mixed with the PCL matrix (total mass of mixture: 6g) in order to reach
different weight fraction of MWNTs (from 0.25 to 3 wt.-% of MWNTs) in the final material.
The mixture was introduced in the addition funnel of a Minilab twin-screw mini-extruder
equipped with co-rotating screws and a closed loop for re-circulation. The temperature of the
mini-extruder was set at 120C. The mixture was introduced in the mini-extruder for 5 min at
30 rpm screw speed. Then the speed of the screws was increased to 60 rpm for 10 min,
allowing the materials to circulate in the closed loop. The nanocomposites were recovered by
opening the mini-extruder, in order to avoid any preferential orientation of the nanotubes
during the extrusion process.
For the HDPE / MWNTs composites preparation, the process is almost identical to that used
for MWNTs/PCL samples. The differences are the MWNTs weight fraction ranges (from 1 to
5 wt% of MWNTs and from 0.5 to 3 wt% of nanohybrids MWNTs-g-PCL + PCL) and the
temperature of melt blending was set at 180C.
VI.C. Instrumentation
a) X-ray Photoelectron Spectroscopy (XPS)
X-ray photoelectron spectroscopy (XPS) is a quantitative and non-destructive spectroscopic
technique that measures the elemental composition, empirical formula, chemical state and
electronic state of the elements that exist within a material. XPS is also known as ESCA, an
abbreviation for Electron Spectroscopy for Chemical Analysis.
Principle: X-rays of energy h are directed onto a material surface where photoelectrons are
created. The emission process is so fast that the photoelectrons leave behind an ionized atom,
which will later relax into the ground state by recombination. Electrons have a very limited
150
free path in matter, which is responsible for the surface selectivity of this method, only
electrons from 10nm deep region can leave the surface without losing energy. During the
measurement, the sample is located in an ultra-high vacuum (UHV) chamber so that
photoelectrons from the surface can freely travel to the analyzer (Figure 6-2).
Figure 6 - 2 : Basic elements of XPS experiment
XPS spectra are obtained by irradiating a material with a beam of X-rays while
simultaneously measuring the kinetic energy and number of core-level electrons that escape
from the top 1 to 10 nm of the material being analyzed.
Since the energy of a particular X-ray wavelength is known, the electron binding energy of
each of the emitted electrons can be determined by using the following equation:
= ( + )
where Ebinding is the binding energy of the electron, Ephoton is the energy of the X-ray photons
being used, Ekinetic is the kinetic energy of the electron as measured by the instrument and is
the work function of the spectrometer (not the material).
151
Figure 6 - 3 : Schematic diagram of a core-level photoelectron emission process
As each element has characteristic core-electron binding energies, it will therefore emit
photoelectrons with a characteristic kinetic energy (Figure 6-3). Element identification can
thus be accomplished by recording the photoelectron energy distribution (spectrum), which
will show intensity peaks corresponding to different elements. XPS detects all elements with
an atomic number (Z) of 3 (lithium) and above. It cannot detect hydrogen (Z = 1) or helium (Z
= 2).
The chemical composition of a sample can be also identified. Indeed, the binding energy of a
core-level electron depends also on surroundings of atoms. Non-equivalence of binding
energies for an element in different chemical compounds can arise from various reasons such
as difference in a formal oxidation state or different molecular environment. The binding
energy shift due to environment effects is usually named the chemical shift and can vary from
a fraction to several eV.
For the analyses, the samples were transferred under UHV into the ESCA analysis chamber.
XPS was performed on a VG-ESCALAB 220iXL spectrometer under a pressure typically
around 10-11 Torr. The atomic composition was derived from peak areas by using calculated
photoionisation cross section2, corrected for the dependence of the escape depth on the kinetic
energy (KE) of the electrons (assumed to have the form =KE0.6) and corrected for the
analyzer transmission function of our spectrometer. The diameter of the Al K X-ray beam is
250m. The samples were analyzed at minimum three different places.
152
Primary amines derivatization reaction
The primary amines grafted onto the surface of CNTs content is estimated by performing
derivatization reactions in gas phase with 4-(trifluoromethyl)benzaldehyde (TFBA, 98%,
Fluka). The flask with the liquid label molecule is connected through a valve to the
derivatization chamber which was evacuated by a rotary pump to a pressure of approximately
1 Pa prior to the reaction. Then, the chamber is closed from pumping and the labeling
molecule is introduced at vapor pressure. After the derivatization time, the chamber is closed
from the flask and gaseous TFBA is evacuated by the pump. The derivatized sample is then
kept under ambient atmosphere before to be placed under vacuum in the XPS set-up to be
analyzed.
b) High Resolution XPS and UPS
High resolution photoemission measurements were performed in the synchrotron of Hamburg

(Hasylab). X-ray photoemission measurements (XPS at a photon energy of 3300 eV) were
performed on the BW2 beamline, described in details by Drube et al. 3. Ultra-violet
photoelectron spectroscopy (UPS) experiments were performed at the FLIPPER II beamline4,
using a photon energy of 110 eV, the overall resolution of the system (source + analyser) was
0.2 eV. Binding energies were calibrated with respect to the 4f level (in the case of XPS) or to
the Fermi edge (in the case of UPS) of a continuous thick evaporated gold film.
c) Transmission electron microscopy (TEM)
Transmission electron microscopy (TEM) is an imaging technique whereby a beam of

electrons is focused onto a specimen causing an enlarged version to appear on a fluorescent
screen or layer of photographic film, or to be detected by a CCD camera. Electrons are
generated by a process known as thermionic discharge in the same manner as the cathode in a
cathode ray tube, or by field emission; they are then accelerated by an electric field and
focused by electrical and magnetic fields onto the sample. The electrons can be focused onto
the sample providing a resolution far better than is possible with light microscopes, and with
153
improved depth of vision. Details of a sample can be enhanced in light microscopy by the use
of stains; similarly with electron microscopy, compounds of heavy metals such as lead or
uranium can be used to selectively deposit heavy atoms in the sample and enhance structural
detail, the dense electron clouds of the heavy atoms interacting strongly with the electron
beam. The electrons can be detected using a photographic film, or fluorescent screen among
other technologies. The TEM is used heavily in both material science/metallurgy and the
biological sciences. In both cases the specimens must be very thin and able to withstand the
high vacuum present inside the instrument. In material science/metallurgy the specimens tend
to be naturally resistant to vacuum, but must be prepared as a thin foil, or etched so some
portion of the specimen is thin enough for the beam to penetrate. The thin sample is deposited
on copper grids. TEM observations were performed on a Philips CM200 apparatus using also
an acceleration voltage of 100 kV. Before analyses, the samples were cut in sections between
75 and 90 nm of thickness using an Ultra-cryomicrotome Leica UCT at -130C. This
equipment is available at Materia Nova research centre in Mons.
High-resolution transmission electron microscopy (HRTEM) was carried out using a Philips
CM30 FEG instrument operated at 300 kV in EMAT at the University of Antwerp.
d) Thermogravimetric analysis (TGA)
The thermogravimetric analysis allows to measure the weight loss of a sample submitted to a
controlled heating program. The observed weight changes can therefore result from several
processes such as the thermal decomposition (pyrolysis or oxidation), the desorptionadsorption or sublimation of volatile by-products, and solid state reactions releasing a volatile
product.
The weight loss measurement is performed by using a microbalance that records the sample
weight variations during the time. A purge gas is generally employed during experiments in
order to evacuate the produced gases at the immediate surface of the sample. This gas is the
most often helium, air or oxygen, depending on which sample degradation mechanism is
studied, thermolysis or thermo-oxidation. The oven can heat the material up to 1000C using
a maximal heating ramp of 50C/min. The oven walls are isolated and cooled in order to
avoid heat transfer to the balance.
154
Thermogravimetric analyses were performed in this study using a Hi-Res TGA 2950
thermogravimetric analyser from TA Instruments with a heating ramp of 20C/min under
helium flow (74 cm3/min) from room temperature to 800C.
e) Differential scanning calorimetry (DSC)
The differential scanning calorimetry is a technique where the energy difference to provide
between a sample and a reference is measured in function of temperature, when the sample
and the reference are submitted to the same controlled temperature program. More precisely,
the instrument measures the difference of heat flow, which is directly proportional to specific
heat (Cp) of the sample. It is therefore possible to obtain a graphic (thermogram) presenting
the specific heat evolution of the sample in function of temperature.
Thermal behaviour has been measured with a DSC Q 50 from TA Instruments, with a heating
and cooling ramp of 10K/min from 80 to 90C for PCL/MWNTs nanocomposites and 0 to
200C for HDPE/MWNTs nanocomposites, under nitrogen flow. The values were generally
recorded during the second heating scan.
f) Electrical conductivity measurements
The electrical conductivity measurements were carried out in the University of Salerno. Due
to broad range of volume resistivity of prepared nanocomposites, the methods and test
fixtures for resistivity measurements vary significantly. Two measurement systems are used
in this study : two-probe test method, where multimeter is used to test moderately conductive
composite whose resistivity is in the range 1-107 cm, and high resistivity meter that allows
to measure volume resistivity of insulating materials up to 1018 cm.
Two-probe measurement
For two-probe method, there are error resistance measurements including contact resistance,
between sample and probes, and probe resistance itself. However, these resistance errors can
be considered as negligible for moderately conductive materials resistance measurements. To
155
measure resistance samples, the constant voltage method is used, a known voltage is sourced
and an electrometer measures the resulting current. A high output voltage power supply is
prepared as voltage source to drive the samples and a sensitive electrometer is used to
measure the resulting current. A short-circuit guard is added to protect the measurement
equipment.
A high resistance meter is used like voltage source and, in the same time, electrometer (Model
6517A, Keithley Instruments, Inc). Data acquisition is performed by the LabView program.
The sample is cleaned with methanol and, after, it is relied to measurement system with
crocodile clips (corresponding to probes). Dimensions of sample are resumed in Table 6-1.
Dimension
Value (cm)
Length
10
Width
0.5
Thickness
0.02-0.05
Table 6- 1 : Dimensions of moderately conductive samples
For each nanocomposite, three samples are prepared. In average, five I/V curves are measured
by sample for different voltage value range (generally 1V-21V by 1V step). An example of
resistance measurement is showed in Figure 6-4:
2,5E-03
I (A)
2,0E-03
1,5E-03
1,0E-03
5,0E-04
0,0E+00
0,00
10,00
20,00
30,00
V (V)
Figure 6 - 4 : Example of I/V measurement
Resistance is determined by the calculation of the I/V curve slope in ohmic region.
156
Then, for each sample, the average resistance can be transformed in resistivity by means of
sample dimensions:
where is the resistivity (.cm), V the voltage (V), I the current intensity (A), A is the crosssectional area (equal to width multiplied by thickness, cm), d the distance between probes on
the sample (cm) and R the measured resistance (). We consider d is equal to the length less
crocodile clip width (1cm).
Finally, the volume resistivity value is obtained by samples resistivity average. The maximum
observed standard deviation is estimated to 10% of electrical resistivity value.
High electrical resistivity measurements

When the two-probe measured resistivity value is superior to 107 cm or when its not
possible to get ohmic relation in I/V curve, the resistivity test fixture is used. This system
allows to measure high volume resistivity from 103 to 1018 cm.
The principle of resistance measurement is identical to the two-probe. A thin sample (max.
0,4mm) is placed in a box, between circle electrodes.
Firstly, we apply a bias-measure sequence in which a test voltage is applied for a
programmed time to permit resistivity to reach equilibrium. After, the volume resistivity
measurement can be made at desired voltage (we apply 100, 300 and 500V successively).
The equation to determine the sample volume resistivity is:
where R is the measured volume resistance (), t is the average thickness of the sample (cm)
and A is the effective area of the guarded electrode, its equal to 22.9 cm for this system. The
dimensions of samples are 8X8cm for the surface and a thickness between 0.01 and 0.03cm.
The samples are cleaned with methanol before measurements. The maximum observed
standard deviation is estimated to 20% of electrical resistivity value.
157
References
1
www.nanocyl.com
Tarniowy A., Mania R., Rekas M., Thin Solid Films 1997, 311, 93-100
3
Drube W., Grehk T.M., Treusch R., Materlik G., J. Electron Spectrosc. Relat. Phenom 1998, 88, 683-687
4
Johnson R.L., Reichardt J., Nucl. Instrum. Methods Phys. Res. 1983, 208, 791796
2
158
LIST OF PUBLICATIONS
1. PLASMA SURFACE TREATMENT OF CARBON NANOTUBES FOR POLYMER
COMPOSITES.
B. Ruelle, C. Bittencourt, Ph. Dubois,
Chapter Contribution in Polymer carbon nanotube composites: synthesis, properties
and applications, in preparation
Eds: T. McNally and P. Potschke
Publisher: Woodhead Publishing
2. EFFECT OF PCL-GRAFTED CARBON NANOTUBES ON ELECTICAL AND
THERMAL PROPERTIES OF PCL AND HDPE NANOCOMPOSITES.
B. Ruelle, S. Peeterbroeck, T. Godfroid, J.-P. Dauchot, C. Bittencourt, R. Snyders,
M. Hecq, Ph. Dubois, In preparation
3. PREPARATION AND CHARACTERIZATION OF POLY(-CAPROLACTONE)
GRAFTED CARBON NANOTUBES AND DERIVED NANOCOMPOSITES.
4. SELECTIVE PRIMARY AMINES GRAFTING OF CARBON NANOTUBE VIA
PLASMA POST-DISCHARGE SURFACE TREATMENT.
5. FUNCTIONALIZATION OF MWNTS WITH ATOMIC NITROGEN.
B. Ruelle, A. Felten, J. Ghijsen, S. Peeterbroeck, T. Godfroid, W. Drube, R.L.
Johnson, D. Liang, R. Erni, G. Van Tendeloo, Ph. Dubois, M. Hecq, C. Bittencourt,
Micron 2009, 40, 85-88
6. THE EFFECT OF NANO-SIZED FILLER PARTICLES ON THE INTERPHASE
FORMATION AND THERMAL PROPERTIES IN MELT-BLENDED POLYESTER
NANOCOMPOSITES.
H. Miltner, N. Watzeels, A.-L. Goffin, E. Duquesne, S. Benali, B. Ruelle, S.
Peeterbroeck, Ph. Dubois, B. Van Mele, Macromolecules, submitted (2008)
7. FUNCTIONALIZATION OF MWNTS WITH ATOMIC NITROGEN:
ELECTRONIC STRUCTURE.
B. Ruelle, A. Felten, J. Ghijsen, W. Drube, R.L. Johnson, D. Liang, R. Erni, G. Van
Tendeloo, Ph. Dubois, M. Hecq, C. Bittencourt, J. Phys. D: Appl. Phys. 2008, 41,
045202 (4pp)
8. FUNCTIONALIZATION OF CARBON NANOTUBES BY ATOMIC NITROGEN
FORMED IN A MICROWAVE PLASMA AR + N2 AND SUBSEQUENT POLY(CAPROLACTONE) GRAFTING.
B. Ruelle, S. Peeterbroeck, R. Gouttebaron, T. Godfroid, F. Monteverde, J.-P.
Dauchot, M. Alexandre, M. Hecq, Ph. Dubois, J. Mater. Chem. 2007, 17, 157-159
COMMUNICATIONS IN SCIENTIFIC CONFERENCES

1. Selective primary amine groups grafting on multi-walled carbon nanotubes with an
original non-destructive plasma process. (Poster contribution, Best poster awards)
B. Ruelle, S. Peeterbroeck, C. Bittencourt, T. Godfroid, G. Van Tendeloo, J.-P.
Dauchot, R. Snyders, M. Hecq, Ph. Dubois, jSRC, Wpion, Belgium, March 2009
2. Selective primary amine groups grafting on multi-walled carbon nanotubes with an
original non-destructive plasma process. (Poster contribution)
Dauchot, R. Snyders, M. Hecq, Ph. Dubois, Matine des jeunes chercheurs, Mons,
Belgium, March 2009
3. Functionalization of MWNTs by atomic nitrogen produced by Ar-N2 -waves plasma.
(Oral presentation)
B. Ruelle, C. Bittencourt, T. Godfroid, S. Peeterbroeck, F. Monteverde, J.-P. Dauchot,
R. Snyders, Ph. Dubois, M. Hecq, The 11th International Conference on Plasma
Surface Engineering PSE 2008, Garmisch Partenkirchen, Germany, September 2008
4. Selective amination of carbon nanotubes in the post-discharge of an Ar + N2 -waves
plasma. (Poster contribution)
Dauchot, R. Snyders, M. Hecq, Ph. Dubois, Biannual conference on modification,
degradation and stabilisation of polymers (MODEST2008),, Lige, Belgium,
September 2008
5. Selective amination of multi-walled carbon nanotubes. Use of -NH2 as anchoring site
for polyester grafting. (Poster contribution)
Dauchot, R. Snyders, M. Hecq, Ph. Dubois, Annual Meeting of the Belgian Polymer
Group, De Haan, Belgium, May 2008
6. Amination of multi-walled carbon nanotubes and subsequent poly(-caprolactone)
grafting. (Poster contribution)
Dauchot, R. Snyders, M. Hecq, Ph. Dubois, 4th international Symposium on
Nanostructured and Functional Polymer-based Materials and Nanocomposite, Rome,
Italy, April 2008
7. Photoemission studies of atomic nitrogen functionalized MWNTs. (poster
contribution)
B. Ruelle, A. Felten, J. Ghijsen, J.-J. Pireaux, R.L. Johnson, M. Felicissimo, P.
Rudolf, Ph. Dubois, M. Hecq, C. Bittencourt, MRS Spring Meeting, San Francisco,
USA, March 2008
8. Functionalization of MWNTs and subsequent polyester grafting. (Poster contribution)
B. Ruelle., S. Peeterbroeck, F. Monteverde, R. Gouttebaron, T. Godfroid, J-P.
Dauchot, M. Alexandre, M. Hecq, Ph. Dubois, Annual Meeting of the Belgian
Polymer Group BPG 2007, Houffalize, Belgium, May 2007
9. Functionalization of multi-walled carbon nanotubes and poly(-caprolactone) grafting,

(Poster contribution, Best poster awards)
B. Ruelle., S. Peeterbroeck, F. Monteverde, R. Gouttebaron, T. Godfroid, J-P.
Dauchot, M. Alexandre, M. Hecq, Ph. Dubois, 5ime Journe de Rencontre des Jeunes
Chimistes de la SRC, Herbeumont, March 2007
10. Functionalization of MWNTs using atomic nitrogen: effects on electronic structure
(Poster contribution),
B. Ruelle, A. Felten, W. Drube, R.L. Johnson, D. Liang, G. Van Tendeloo, Ph.
Dubois, M. Hecq, C. Bittencourt, E-MRS Fall Meeting, Warsaw, Poland, September
2007
11. Amination of carbon nanotubes by atomic nitrogen produced in an Ar-N2 -waves
plasma, (Oral presentation)
B. Ruelle, C. Bittencourt, T. Godfroid, S. Peeterbroeck, F. Monteverde, J.-P. Dauchot,
Ph. Dubois, M. Hecq, IX School on Synchrotron Radiation: Fundamentals, Methods
and Applications, Trieste, Italy, September 2007
12. Controlled functionalization of multi-walled carbon nanotubes and characterization.
(Poster contribution)
B.Ruelle, S. Peeterbroeck, F. Monteverde, R. Gouttebaron, T. Godfroid, M.
Alexandre, M. Hecq, Ph. Dubois, The 10th International Conference on Plasma
Surface Engineering PSE 2006, Garmisch Partenkirchen, Germany, September 2006
13. Functionalization of multi-walled carbon nanotubes and characterization of grafted
primary amine groups. (Poster contribution)
Alexandre, M. Hecq, Ph. Dubois, Annual Meeting of the Belgian Polymer Group BPG
2006, Ostende, Belgium, May 2006
14. Comparison between different types of amine groups grafting on carbon nanotubes
surface. (Poster contribution)
Alexandre, M. Hecq, Ph. Dubois, Annual Meeting of the Belgian Polymer Group BPG
2005, La Roche-en-Ardenne, Belgium, May 2005

Ruelle

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Ruelle

Transféré par

Droits d'auteur :

Formats disponibles

Functionalization of carbon nanotubes

via plasma post-discharge surface treatment:

Laboratoire de Chimie Inorganique et

Merci au F.R.I.A. (Fonds pour la Formation la Recherche dans lIndustrie et dans

Functionalization of carbon nanotubes ......................................................................18

II.B.1. Interests and types of functionalization...................................................................18

II.C. Plasma treatment of carbon nanotubes ......................................................................28

III. OBJECTIVES & PROPOSED RESEARCH STARTEGIES ............................54

IV. RESULTS & DISCUSSION ...........................................................................................58

IV.A.3. Quantification of primary amine functions grafted on CNTs ..............................72

IV.B.3. Conclusion ................................................................................................................104

IV.C.1.e. DSC analyses of HDPE nanocomposites ............................................................116

IV.C.2. Electrical properties of CNTs-filled nanocomposites .........................................119

V. GENERAL CONCLUSIONS & PERSPECTIVES .............................................143

Figure 1- 1 : TEM micrograph of CNTs bundle/aggregate

Iijima S., Nature 1991, 354, 56-58

II. STATE OF THE ART

II. STATE OF THE ART

II.A.1. Structure of carbon nanotubes

Figure 2- 2 : Formation of a nanotube by a graphene rolling up 9

II. STATE OF THE ART

II. STATE OF THE ART

II. STATE OF THE ART

II.A.2. Carbon nanotubes: Methods of production

II. STATE OF THE ART

High pressure conversion of carbon monoxide (HiPCO) technique is considered as

II. STATE OF THE ART

limit the quantity in produced CNTs.

II.A.3. Properties of carbon nanotubes

II. STATE OF THE ART

Youngs Modulus (TPa)

Tensile Strength (GPa)

Elongation at Break (%)

Table 2- 1 : Comparison of mechanical properties of different materials

II. STATE OF THE ART

Figure 2- 8 : HR-TEM micrograph and computer simulation of nanotube bend47

II. STATE OF THE ART

II.A.4. A large range of applications

II. STATE OF THE ART

. The small size of these electron guns enables miniaturisation of elements in

Figure 2- 10 : CNTs-based nano-emissive display (NED) prototype

II. STATE OF THE ART

II. STATE OF THE ART

For other applications requiring non-aligned carbon nanotubes, the functionalization of

II.B. Functionalization of carbon nanotubes

II. STATE OF THE ART

Figure 2- 13 : Functionalization possibilities for CNTs: defect functionalization (A),

II.B.2. Non-covalent functionalization with surfactant or polymer

II. STATE OF THE ART

Figure 2- 14 : Interaction of nanotubes with pyrene derivatives 93

Figure 2- 15 : Schematic representation of the polymerization-filling technique (PFT) applied to CNTs96

II.B.3. Covalent functionalization

II. STATE OF THE ART

Figure 2- 16 : Typical defects in a SWNT

II. STATE OF THE ART

II. STATE OF THE ART

II. STATE OF THE ART

A large majority of covalent sidewall functionalizations is carried out in organic

II. STATE OF THE ART

The first studied sidewall functionalization is the fluorination of CNTs. Carbon

II. STATE OF THE ART

II. STATE OF THE ART

II.C. Plasma treatment of carbon nanotubes

II. STATE OF THE ART