Matériaux & Techniques 104, 607 (2016)
©c EDP Sciences, 2017 Matériaux
DOI: 10.1051/mattech/2017009 &Techniques
www.mattech-journal.org
Dispersion and functionalization of single-walled carbon
nanotubes (SWCNTS) for nanocomposite applications
A. Yaya1,2,a, Abraham Tekley3, E. Annan1, J.K. Efavi1, E. K Tiburu4,
B. Onwona-Agyeman1 and Lars R. Jensen5
Received 24 August 2016, Accepted 15 February 2017
Abstract – We present an easy and reliable method of functionalizing and dispersing single-walled carbon
nanotubes (SWCNTs) for nanocomposite production. This was achieved via a mixture of the concentrated
acids HNO3/H2SO4, followed by sonication in appropriate solvents (Ethanol, Acetone and/or Dimethyl-
formamide). SWCNTs were functionalized to attach carboxylic moieties on the walls of the nanotubes,
which aided in the production of SWCNTs suspensions in acetone, ethanol and dimethylformamide in
which nanotubes remained dispersed for a period of two (2) weeks. Energy dispersion analysis (EDS),
scanning electron microscopy (SEM) and Raman spectroscopy were used to characterize the functionalized
SWCNTs.
Key words: SEM / EDS / Raman / Functionalization / SWCNTs
1 Introduction removes the caps at the end and also shortens the SWC-
NTs so that polymer resin molecules will fill the tube to
Surface modification of carbon nanotubes (CNTs) is improve the bonding and surface properties between nan-
an important step in understanding the mechanical prop- otubes and the polymer matrix.
erties of uniformly dispersed nanotubes. The procedure The cutting of CNTs could be carried out by the use of
also ensures effective interfacial bonding between the concentrated acids, fluorination, ultra-sonication or sim-
matrix and nanotubes for load transfer in the polymer ply by mechanical milling or grinding.
nanocomposite [1]. Strong acids such as H2SO4 and HNO3 can be used tocut CNTs [10]. It was reported by Farkas et al. [11] that
Carbon nanotubes always appear in agglomerate form although it is possible to cut long and tangled CNTs using
due to their higher surface area [2]. Hence, better methods these acids, it might not be effective to cut nanotubes with
of dispersion need to be formulated to improve surface smaller diameters in the range (0.7–0.8 nm).
properties of the nanotubes within the matrix material. Cutting CNTs by fluorination was achieved by blend-
Since the side walls of carbon nanotubes are arranged ing SWCNTs with fluorine and then pyrolysing them at
in a seamless hexagonal ring with no dangling bonds, high temperature (∼1000 ◦C) under inert atmosphere.
the end caps of CNTs are more reactive than their side This resulted in the modification and cleavage of nan-
walls [3–7]. The commonest method of improving the sur- otube fragments with different lengths [10].
face properties of CNTs is by functionalization, which in- Another method used to cut and modify CNTs is
volves either opening the end caps of the tubes to induce ultra-sonication. This is a common method of dispersing
defects and enrich its chemistry or attaching functional CNTs in a solvent (resin matrix) as well as shortening
groups to the side walls of the carbon nanotubes [1]. Cut- of SWCNTs [12]. However, there are some limitations
ting CNTs is one of the approaches used for functional- in utilizing ultrasonic waves, as they cause defects such
izing nanotubes [8]. As shown in Figure 1, this method as buckling, bending and sometimes holes in the sidewalls
a Corresponding author: ayaya@ug.edu.gh
1 Department of Materials Science & Engineering, CBAS, University of Ghana, Ghana
2 Institut des Materiaux Jean Rouxel, CNRS-Université de Nantes, France
3 Department of Energy Technology, Aalborg University, Denmark
4 Department of Biomedical Engineering, CBAS, University of Ghana, Ghana
5 Department of Mechanical Engineering, Aalborg University, Denmark
Article published by EDP Sciences
A. Yaya et al.: Matériaux & Techniques 104, 607 (2016)
Fig. 1. Computer Simulation showing the filling of a polymer matrix into SWCNTs after cutting (Adapted from Shiren Wang,
Thesis, 2006) [9].
of CNTs that might compromise the mechanical integrity
of the nanotubes [13]. Long-time sonication has also
been reported to cause permanent damage to the nan-
otubes [14–16].
Additionally, mechanical ball milling or grinding is a
solid state process which can be used to cut or shorten
CNTs. During this process, the formation of kinks in the
CNTs results in bond fracture when the local strain ex-
ceeds the critical threshold.
Generally, the methods of cutting CNTs outlined
above present some challenges in controlling the length
of the shortened CNTs and preserving the original side
wall structure of the nanotubes.
The covalent modification of CNTs results in the for-
mation of covalent bonds on the surface of nanotubes.
This method, however, has some drawbacks; the me-
chanical and electrical properties of covalently modified Fig. 2. The possible defects induced in SWCNTs. (a) Pen-
CNTs were shown to decrease significantly compared to tagon or heptagon carbon rings in the CNT that lead to a3
the original behaviour of the nanotubes. This was at- bend in the nanotube. (b) SP hybridized defects (R = H and
tributed to the disruption of the CNT conjugation sys- OH). (c) –COOH groups created by oxidation. (d) Open end
of the SWCNTs terminated with –COOH groups [21].
tem during bond formation and breakage as a result of
functionalization.
The reactivity on the end caps of nanotubes is driven
by strain relief which stems from the pyramidaliza- walls of fluorinated SWCNTs, which provides a synthetic
tion and misalignment of pi-orbitals that causes local route to bind amino acids to the side walls.
strain [17]. As the side walls of the nanotubes are inert to Another strategy for surface modification via co-
most chemicals, they need to be activated prior to surface valent method is by inducing defects on the walls of
modification. CNTs through oxidation. This helps in generating defects
Fluorination of CNTs is the earliest method observed and other functional groups through which a chemical
to form a surface bond between the carbon of the nan- species could be linked [1]. Treating CNTs with strong
otubes and fluorine at higher temperatures [18]. The C- acids like HNO3, KMnO4/H2SO4, oxygen gas, O3, and
F bonds established at the surface are reported to be K2Cr2O7/H2SO4 causes defects (Stone-Wales defects, va-
weaker, thus the C-F bonds can serve as a starting point cancies) and terminal Carboxylic groups, as shown in
for surface modification through subsequent addition of Figure 2.
various chemical moieties. Once the nanotubes is fluori- The carboxyl groups generated by oxidation of CNTs
nated, this then form the basis for the attachment of sev- are important sites for linking nanotubes to polymers in
eral groups along the sidewalls of the tube; alkyl groups the processing of nanocomposites.
(alkylation) [19] and amino groups [20]. The side wall The reactivity of terminal carboxyl groups could
alkylation of fluorinated SWCNTs were shown to precede be further enhanced by reacting with thionyl chloride
the attachment of other functional moities either by us- (SOCl2) to yield an acyl chloride on the surface of the
ing alkyl lithium species or alkyl magnesium bromide in nanotube which can add several groups at the carbonyl
tetrahydrofuran (Grignard synthesis) as alkylation pre- position. This is due to the labile nature of the C-Cl
cursors. Amino groups can also be attached to the side bond [17].
607-page 2
A. Yaya et al.: Matériaux & Techniques 104, 607 (2016)
Furthermore, the non-covalent method of surface 200S) for 30 min. The short sonication time was cho-
modification is advantageous, thereby maintaining the sen to avoid damage of the SWCNTs due to prolonged
original properties of the nanotube. In this method, there sonication, as these nanotubes will further be sonicated
is no bond formation or breakage, but only dispersion and for dispersion purposes. The resulting solution was rinsed
wrapping of linear long chain molecules, like surfactants well with deionised water till PH ∼ 7, after leaving the
and polymers, on the surface of CNTs. mixture in reaction time for about 5 h. The function-
Finally, surface properties of SWCNTs were reported alized nanotubes (SWCNTs-COO−) were then collected
to be modified by using surfactants such as benzalkonium after filtration through a 0.45 µm millipore membrane fil-
chloride and sodium dodecylsulfate (SDS) in an aqueous ter. Finally, the collected solutes were dried overnight in
phase, causing well dispersed nanotubes [23]. The surface an oven at 100 ◦C to get rid of water.
properties were also modified by DNA molecules, enabling Raman spectra were obtained from the raw (r-
efficient dispersion of CNTs due to the chain flexibility SWCNTs), purified (p-SWCNTs) and functionalized (f-
and backbone charge of the DNA molecule [24]. SWCNTs) nanotubes using an argon (Ar+) laser excita-
In this paper, we present a soft and simple chem- tion of 785 nm with a resolution of 2 cm−1. The procedure
ical technique for the dispersion and functionalization for obtaining the purified SWCNT samples has been de-
of single-walled carbon nanotubes with the formation of scribed in Reference [26].
fewer defects on the nanotube side walls which can be
tailored for nanocomposites and other applications.
2.4 Results and discussion
2 Experimental procedure Due to the substantial van der Waals attractions,
CNTs often aggregate and form bundles of nanotubes.
2.1 Chemicals When employed as reinforcing materials, efficient and
effective dispersion plays a significant role in order to
HCl (37% fuming) supplied by Sigma-Aldrich Inc., achieve high performance in nanocomposites. Therefore,
H2SO4 (97%) and HNO3 (65%) supplied by Merck, ab- it is desirable to functionalize the sidewalls of CNTs,
solute Ethanol (99.9%) supplied by Kemetyl, dimethyl- thereby generating CNTs-derivatives that are compatible
formamide (DMF) and acetone were the chemicals used with solvent as well as polymeric matrix materials.
for the purification, dispersion and functionalization pur- The SWCNTs were functionalised by the chemical
poses. method of oxidation using H2SO4/HNO3 mixtures. Since
chemical functionalization of nanotubes is believed to
start at defect sites (heptagon-pentagon pair; a junction
2.2 Single-walled carbon nanotubes at the walls of nanotube where the hexagon pairs are
changed to a heptagon and pentagon pairs), the extent of
The as-prepared SWCNTs were supplied by Carbolex oxidation depends on the amount of defects in the start-
Inc., and they were manufactured by the Arc method. ing materials. Already, the SWCNTs had some defects
These tubes are supposed to have a purity of 50–70 vol%, as a result of the production methods employed. More-
as provided in the supplier information [25]. According to over, additional defects (bends and opening of the end
the manufacturer’s information, these carbon nanotubes caps) introduced during purification make functionaliza-
have an average diameter of 1.4 nm and are found in tion easier. Hence, it is expected that carboxylic func-
“ropes” which are typically ∼20 nm in diameter. Impu- tional groups and other oxygen bearing groups will exist
rities within the products include approximately 35 w% at the end caps and defect sites, thus decorating the nan-
residual catalyst (Ni, Y) and some amorphous carbon on otubes with larger numbers of oxygenated functionalities.
the outer surfaces of the ropes. The single-walled carbon Although it is possible to introduce carboxyl func-
nanotubes were purified by a method described elsewhere tional groups on the surface of single wall carbon nan-
in the text [26]. otubes using HNO3 alone, the oxidizing acid mixture of
H2SO4/HNO3 (3:1 by volume) has been reported to at-
tach almost three times more carboxylic acid sites relative
2.3 Functionalization of purified SWCNTs to HNO3 [22].
The H2SO4/HNO3 mixture can generate NO+2 ions
For the preparation of functionalized SWCNTs, the which are strong electrophiles with the ability to attack
oxidizing acid treatment methods were used. Accordingly, not only the existing defect sites but also the graphene-
500 mg of purified SWCNTs were immersed in 250 ml of like structure of nanotubes to generate new defect sites.
concentrated H2SO4/HNO3 in a ratio of 3:1 v/v % respec- Contrarily, this mixture can also cut the nanotubes and
tively. As reported in the functionalizing of CNTs [27,28], may cause the nanotubes to be reduced to amorphous
an hour’s sonication and short time acid treatments were carbons [29, 30]. This phenomenon was briefly described
deemed to be optimal for oxidizing and preserving the as- by Zhang et al. [8]. Furthermore, the oxidation process is
pect ratio of CNTs. Hence, the acid/SWCNT mixtures accompanied by defect generating and defect consuming
were sonicated using a horn ultra-sonicator (type UP steps. In the former step, electrophilic reactions attacks
607-page 3
A. Yaya et al.: Matériaux & Techniques 104, 607 (2016)
Table 1. EDS Elemental analysis of SWCNTs before and 52000 G
after oxidative functionalization. (This is only qualitative data
46800
averaged for 10 samples).
41600
Atom (%) C O S Ni Y 36400
Before functionalization 95.03 3.15 – 1.25 0.51 RBM
31200
After functionalization 80.20 18.73 0.25 0.71 0.11
26000
20800 D D' G'
p-SWCNT
the graphene structure to create active sites such as – 15600
COOH and –C = O. In the latter step (defect consuming), 10400 f-SWCNT
the graphene structure of the tube is destroyed due to 5200 r-SWCNT
further oxidation of the active sites generated by the first 0
step. Thus, as the defect generating steps were found to be 400 800 1200 1600 2000 2400 2800 3200Raman Shift (cm-1)
faster than the consuming steps, higher carboxylic groups
were found on the surface of nanotubes when oxidized in Fig. 3. Raman spectra for raw (r-SWCNTs), functionalised
H SO /HNO mixture. (f-SWCNTs) and purified (p-SWCNTs) showing radial breath-2 4 3
nd
In Table 1, representative energy dispersive spec- ing mode (RBM), disorder mode (D), 2 order D-mode, tan-
troscopy (EDS) values of the SWCNTs before and af- gential mode (G) and second order tangential mode (G’).
ter functionalization are given. This qualitative data in-
0,20
dicates that there is an increase in atomic percentage of
oxygen from 3.15% to 18.73%, which could signify the
presence of carboxylic functional groups attached to the
SWCNTs. The EDS gives the relative concentration of 0,15
each element in the sample; nevertheless, it does not con-
firm the presence of carbonyl group (-C=O) on the side
walls of nanotubes. In order to confirm the presence of a
carboxyl functional groups, Fourier Transform Infra-red 0,10
spectroscopy would have been the best method, as it can
detect the inter-band transition to determine the effect of
bonding on the band structure. Moreover, band frequen- 0,05
cies around ∼1750 cm−1 and ∼1640 cm−1 give a finger-
print for the presence of C=O stretching vibration modes
of carboxylic group and the asymmetric stretching vibra-
tion of the carboxylate (-COO−) respectively [31–33]. Un- 0,00 r-SWNT p-SWNT f-SWNT
fortunately, we were not able to do any FTIR analysis in
this work. SWCNT Type
As functionalized SWCNTs were expected to carry Fig. 4. The intensity of ID/IG ratio for the raw(r-SWCNTs),
several carboxylic acid groups, this offers the opportu- purified (p-SWCNTs) and functionalized (f-SWCNTs) single-
nity to form well dispersed colloids in ethanol. Further- walled carbon nanotubes.
more, carboxylic functional groups on the functionalized
nanotube walls can undergo esterification reactions with
other polymers such as epoxy resins to form covalent As reported by M.S. Dresselhaus et al. [23], the inten-
bonds which could improve the mechanical properties of sity of the D-band of Raman spectra increases with de-
the composites. creasing length of nanotubes and the ratio ID/IG showsan increasing trend with decreasing nanotube length
Generally, functionalization of SWCNTs could en- (L ) of nanotubes, showing the dependence as;
hance the efficiency of reinforcement by improving solu- tube
bility and dispersion and also by forming chemical bonds ID 1
with the matrix [34]. ∼ (1)IG Ltube
Raman spectroscopy studies were carried out on the
functionalized SWCNTs to ascertain the degree of func- By similar extension, this concept can be applied to our
tionalization in comparison to raw and purified powders study to investigate whether purification and functional-
of the SWCNTs. This is shown in Figure 3. As can be ization affect the length of SWCNTs. Based on the av-
seen in Figure 3, the increase in intensity and width of the erage Raman spectra intensities of the raw, purified and
Raman D-band of functionalized SWCNTs (f-SWCNTs) functionalized SWCNTS, the data values of ID/IG were
as compared to the purified ones (p-SWCNTs) could be plotted against the type (raw, purified and functionalized)
due to the change of the nanotubes to graphitic struc- of SWCNTs, showing the extent to which the length of
ture (and amorphous carbon) owing to the exposure of nanotubes was shortened during functionalization, as the
the nanotubes to strong oxidizing agents (HNO3/H2SO4 values relate inversely to nanotubes length (1/Ltube) as
mixture). shown in Figure 4. The ID/IG ratio increases from raw
607-page 4
ID/IG Intensity (a.u)
A. Yaya et al.: Matériaux & Techniques 104, 607 (2016)
A B 
200 nm 200 nm 
 
Fig. 5. SEM images of SWCNTs; (A) Pristine SWCNTs, (B) Purified SWCNTs.
to purified nanotubes simply because a fraction of amor- polar functional groups which could create better inter-
phous carbon is removed during purification. The ratio action (miscibility) with the relatively polar solvents (po-
increases from purified to functionalized nanotubes, not larity index of Ethanol, DMF and Acetone is 5.2, 6.4 and
only because of the shortening, but simply because of the 5.1 respectively) and for comparison, the polarity index
functionalization of the nanotube sidewalls. of water (polar), and hexane (non-polar) is also 9 and
However, the increase in ID/IG intensity for function- 0 respectively [37]. After the dispersion by sonication,
alised SWCNTs results from the shortening of the nan- the functionalized SWCNTs were observed to have bet-
otubes which increased the intensity of the Raman D- ter stability than the raw SWCNTs as shown in Figure 6.
band, but this needs to be verified with other electron The stability of the functionalized SWCNTs in the sol-
microscopy techniques like AFM and TEM which was not vent could be because the polar groups attached to the
done in this studies. nanotube surface might create better interaction with the
Dispersion and interfacial bonding phenomena are polar solvent.
the two most important factors that control the effi- It is not surprising to see that it is difficult to dis-
ciency of load transfer between the nanotubes and poly- perse functionalized or purified nanotubes. It is known
mer matrix in polymer nanocomposites. We tackled these that nanotubes often aggregate irreversibly after drying.
problems by dispersing the nanotubes in an appropriate Samples are sometimes freeze-dried so that they can be
solvent such as ethanol, which was followed by the son- more easily processed.
ication and functionalization of nanotube surfaces using
H2SO4/HNO3 mixtures. The advantage of using ethanol
is that it is easier to evaporate during the processing 3 Conclusion
of composites at lower temperature than other solvents
without affecting the property of the polymer. To benefit from interfacial properties of SWCNTs with
Solvents such as acetone, tetrahydrofuran and chloro- polymers, we had functionalized the purified SWCNTs to
form were reported to better disperse SWCNTs [35, 36]; attach carboxylic functional groups on the walls of the
however, they can interact with some polymer networks nanotubes, with the view that these functional groups
which can either affect the mechanical property negatively would enhance the interfacial properties between nan-
when composites are formed without ensuring that they otubes when used as a nanocomposite. Thus, we found
evaporate from the mixture. that, the strong oxidizing agent, H2SO4/HNO3 mixture,
It is true that the functionalization of nanotubes will enabled us to attach a significant amount of carboxylates
contribute more towards addressing the dispersion and to the nanotubes. However, it was necessary to control re-
adhesion needs. The dispersion process of the purified action parameters (temperature, concentration and time
and functionalized nanotubes by sonication in ethanol was of reaction) to reduce the amount of damage on the sur-
however difficult and needed longer sonication time com- face of nanotubes and a shortening of the SWCNTs. Also,
pared to the fluffy raw SWCNTs in our case. This was due although sonication of SWCNTs in solvents improves the
to the highly aggregated and solid nature of the SWC- dispersion process, the little residual amount of solvents
NTs obtained after purification and functionalization, see that remains after evaporation might influence the chem-
Figure 5. ical reaction process which may affect their use in appli-
In order to verify the extent of our functionalization, cations such as nanocomposites if they are not completely
the functionalised tubes were placed in acetone, dimethyl removed.
formamide and ethanol, see Figure 6. The significance Finally, the SWCNTs which were functionalized re-
of this dispersion test was that, it enabled us to judge mained dispersed in different solvents for a period lasting
whether the surface of nanotubes had been attached with two weeks as compared to the pristine and purified tubes.
607-page 5
A. Yaya et al.: Matériaux & Techniques 104, 607 (2016)
A B
  
 
 
 
 
 
i ii iii i ii iii  
C 
e d a
Fig. 6. Dispersion of SWCNTs in ethanol after 1 hour sonication (A) and after 1 week showing the relative stability of the
functionalized SWCNTs (B) where; raw is (i), purified is (ii) & functionalised is (iii). (C) Investigating the state of dispersion
of functionalized SWCNTs in different solvents [e-ethanol, d-dimethyl formamide and a-acetone] for 2 weeks.
Acknowledgements. AY acknowledges support from the [10] Y.C. Chiang, W.H Lin, Y.C. Chang, App. Sur. Sci. 257
Carnegie-NGAA for sponsorship on the write-shop which was (2011) 2401-2410
organised by Prof. Y. Ntiamoah-Badu. We also thank Prof. [11] E. Farkas, M.E. Anderson, Z. Chen, A.G. Rinzler, Chem.
Helen Yitah of the English Dept. University of Ghana for proof Phys. Lett. 363 (2002) 111-116
reading the manuscript. [12] H.L. Pan, L.Q. Liu, Z.X. Guo, L.M. Dai, F.S. Zhang, D.B.
Zhu, Nano lett. 3 (2002) 29–32
[13] G.T. Caneba, C. Dutta, V. Agrawal, M. Rao, JMMCE 9
References (2010) 165-181
[14] Z.M. Huang, Y.Z. Zhang, M. Kotaki, S. Ramakrishna,
[1] L. Chen, H. Xie, W. Yu, Functionalization Methods Comp. Sci. Tech. 63 (2003) 15
of Carbon Nanotubes and Its Applications, Carbon [15] M. Meyyappan, Carbon nanotubes science and applica-
Nanotubes Applications on Electron Devices, Prof. Jose tions, (CRC Press, Taylor & Francis, USA), 2004
Mauricio Marulanda (Ed.), InTech, DOI: 10.5772/18547 [16] L.N. An, W.X. Xu, S. Rajagopalan, C.M. Wang, H.
(2011) Wang, Y. Fan, L.G. Zhang, D.P. Jiang, J. Kapat, L.
[2] M.S. Dresselhaus, G. Dresselhaus, R. Saito, Carbon 33 Chow, B.H. Guo, J. Liang, R. Vaidyanathan, Adv. Mater.
(1995) 7 16 (2004) 22
[3] S.E. Baker, W. Cai, T.L. Lasseter, K.P. Weidkamp, R.J. [17] F. Tournus, S. Latil, M.I. Heggie, J.C. Charlier, Phys.
Hamers, Nano lett. 2 (2002) 1413 Rev. B 72 (2005) 075431
[4] S.S. Bale, P. Asuri, S.S. Karajanagi, J.S. Dordick, R.S. [18] L. Bulusheva, Yu. Fedoseeva, A. Okotrub, I. Asanov, V.
Kane, Adv. Mater. 19 (2007) 20 Koroteev, A. Yaya, C.P. Ewels, A. Chuvilin, A. Felten,
[5] A. Dillon, K. Jones, T. Bekkedahl, C. Kiang, D.S. G. Van Lier, D. Vyalikh, Chem. Mater. 22 (2010) 14
Bethune, M.J. Heben, Nature 386 (1997) 377-379 [19] J. Zhao, J.P. Lu, J. Han, C.K. Yang, Appl. Phys. Lett.
[6] P. Eklund, J. Holden, R. Jishi, Carbon 33 (1995) 959 82 (2003) 3746
[7] R. Saito, T. Takeya, T. Kimura, G. Dresselhaus, M. [20] Y. Wang, Z. Iqbal, S. Mitra, Amer. Chem. Soc. 128
Dresselhaus, Phys. Rev. B 57 (1998) 4145 (2005) 95-99
[8] Y.J. Zhang, Y. Wen, Y. Liu, D. Li, J.H. Li, Electrochem. [21] A. Hirsch, Angew. Chem. Int. Ed. 41 (2002) 1853-1859
Commun. 6 (2004) 1180-1184 [22] J.E. Fischer, Carbon nanotubes: structure and proper-
[9] S. Wang, Functionalization of Carbon Nanotubes ties, 2006, pp. 41-75
(Ph.D. dissertation, Department of Industrial and [23] M.S. Dresselhaus, G. Dresselhaus, P.C. Eklund, Science
Manufacturing Engineering) Florida state university, of fullerenes and carbon nanotubes, (1st ed. Academic
USA, 2006 Press, San Diego, USA), 1996
607-page 6
A. Yaya et al.: Matériaux & Techniques 104, 607 (2016)
[24] K.A. Williams, P.T.M. Veenhuizen, B.G.R. De la Torre, [31] T. Belin, F. Epron, Mater. Sci. Eng. B 119 (2005) 105-
Nature 420 (2002) 761-761 118
[25] http://www.carbolex.com/products.html [32] S. Wang, Z. Liang, T. Liu, B.Wang, C. Zhang,
[26] A. Yaya, C.P. Ewels, Ph. Wagner, I. Suarez-Martinez, A. Nanotechnology 17 (2006) 1551-1557
Gebramariam Tekley, L. Rosgaard Jensen, Eur. Phys. J. [33] H. Kataura, Y. kumazawa, Y. Maniwa, I. Umezu, S.
Appl. Phys. 54 (2011) 10401 Suzuki, Y. Ohtsuka, Y. Achiba, Synth. Mets. 103 (1999)
[27] P.M. Ajayan, S. Lijima, Nature 358 (1992) 23 2555-2558
[28] M.S. Dresselhaus, G. Dresselhaus, Ph. Avouris, Carbon [34] P. Liu, Eur. Polym. J 41 (2005) 2693-2703
Nanotubes, Springer-Verlag Berlin Heidelberg, 2001 [35] G. Gao, T. Cagin, W.A. Goddard, Nanotechnology 9
[29] Z.J. Shi, Y.F. Lian, F.H. Liao, X.H. Zhou, Z.N. Gu, Y. (1998) 184
Zhang, S. Iijima, Solid State Commun. 35 (1999) 112 [36] N. Grobert, Mater. Today 10 (2007) 28-35
[30] B. Liu, T. Wagberg, E. Olssen, R. Yang, H. Li, S. Zhang, [37] Y.T. Liu, W. Zhao, Z.Y. Huang, Y.F Gao, X.M. Xie, X.H.
H. Yang, G. Zou, B. Sundqvist, Chem. Phys. Lett. 320 Wang, X.Y. Ye, Carbon 44 (2006) 1613-1616
(2000) 365
607-page 7