d
o
E
A
e
lu
2
s
Corresponding author: e-mail nemethk@szfki.hu, Phone: þ36 1 392 2222, Fax: þ36 1 392 2219
anions. The products were characterized bywide range infrared

2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
nanotubes: (A) metal reduction by melted potassium, tubes – before any further process – were annealed in
10
6mbar) at 250 8C for 18 h in order to
d other volatile contaminants and to cure
idewalls.
), the nanotubes were annealed with
uum-sealed glass vial at 200 8C for 12 h.
The intercalated nanotubes were dispersed in 60mL of
0dependent changes of electronic structure and selectivity of
the reactions. In order to investigate these effects, we usedanalogous to the procedure previously applied to graphite
[9] and (B) reduction by naphthalenide anion produced in
THF solution, both followed by reacting the tubes with
methanol (Fig. 1). We are interested in the chirality-
dynamic vacuum (
get rid of water an
the defects in the s
In method (A
potassium in a vacby using glow-discharge [5–7], or proton bombardment [8].
We tried two different reductive methods on the
and dried by vacuum cryo-distillation over K-Na alloy.
One hundred milligrams of HiPCo single-walled nano-modified Birch reduction [3, 4], with atomic hydrogenation1 Introduction Since 1991, the discovery of carbon
nanotubes by Iijima [1], these materials have been in the
focus of research. Besides their unique physical properties
based on their extreme one-dimensionality, the possibility of
special chemical reactions was also considered and
attempted from the early days of carbon nanotube science
[2].
Up to now, covalent hydrogenation of single-walled
carbon nanotubes (SWNTs) was carried out by classical and
Our aim was to compare the sidewall hydrogenation of
HiPCo nanotubes – which have smaller average diameter
and therefore higher reactivity than most of the other
available carbon nanotubes [10] – by these two different
methods.
2 Experimental Purified HiPCo SWNTs were from
Carbon NanoTechnologies Inc. Methanol, 99.8%, sodium,
and potassium were purchased from Aldrich and used as
received. Toluene, 99.8% was also purchased from Aldrichopt
spe
con
detTwo different reductive synthetic methods were applied to
hydrogenate the sidewalls of HiPCo single-walled carbon
nanotubes (SWNTs). In the first one, the reductive agent was
melted potassium which doped and exfoliated the nanotube
bundles, so that before hydrogenation all of the tubes had been
converted to metallic ones. In the second method, doping
occurred just before hydrogenation by naphthalenide radicalical spectroscopy on the products. Optical absorption
ctra are influenced by both diameter and metal/semi-
ductor character, therefore this method is very useful in
ecting selectivity by these parameters.(30–52 000 cm
1) spectroscopy with special emphasis on the
selectivity of the twomethods.We found that in the first case the
controlling factor is the bandgap, and in the second case the
diameter. This difference suggests the importance of the p–p
interaction between naphthalenide and the nanotube surface.Investigation of hydrogenate
nanotubes by infrared spectr
Katalin Ne´meth*,1, A´ron Pekker1, Ferenc Borondics1,
and Sa´ndor Pekker1
1Research Institute for Solid State Physics and Optics, Hungarian
2 Institute of Materials and Environmental Chemistry, Chemical R
1525 Budapest, Hungary
3GFMC, Departamento de Fisica Aplicada III, Universidad Comp
Received 12 May 2010, revised 28 July 2010, accepted 5 August
Published online 13 September 2010
Keywords carbon nanotubes, diameter selectivity, infrared spectro
*
Phys. Status Solidi B 247, Nos. 11–12, 2855–2858 (2010) / DOI 10.1HiPCo
scopy
mma Jakab2, Norbert M. Nemes3, Katalin Kamara´s1,
cademy of Sciences, PO Box 49, 1525 Budapest, Hungary
search Center, Hungarian Academy of Sciences, P.O. Box 17,
tense de Madrid, 28040 Madrid, Spain
010
copy, sidewall functionalization
02/pssb.201000329 p s s
basic solid state physics
b
st
a
tu
s
so
li
d
i
www.pss-b.comp
h
y
si
caanhydrous toluene in an argon dry box. One milliliter of
methanol dissolved in 20mL of anhydrous toluene was
added dropwise during mild sonication. Sonication was
continued for another 2 h. The mixture was filtered on a

2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

shown here), the functionalization rate of each sample is of
the order of 1mass.% of hydrogen, corresponding to one
proton per 8–10 carbon atoms. The degree of functionaliza-
ig
p
h
ys
ic
a ssp status solid
i b
Figure 2 (online color at: www.pss-b.com) Baseline-corrected
optical conductivity of functionalized HiPCo nanotubes. Upper:
samples made by reaction (A); lower: made by reaction (B). Curves
1,2, and 3 correspond to increasing order in the degree of function-
alization.Millipore1 0.5mm pore sized PTFE membrane, washed
with methanol, distilled water, some more methanol, and
dried in dynamic vacuum (200 8C, 12 h) in order to remove
the residual solvents. This procedure was repeated two more
times with the formerly functionalized nanotubes.
Method (B) was the one using potassium naphthalenide
described in Ref. [3]. The reaction was repeated three times
in this case as well.
Infrared spectra were measured with two FTIR instru-
ments (Bruker IFS 66v/S and Bruker Tensor37) on self-
supportingfilms,whichweremade by vacuumfiltration [11].
Spectra were recorded between 30 and 52 000 cm
1. From
these wide-range spectra, we determined the optical
conductivity by Kramers–Kronig transformation [12]. In
the area of the transitions, we applied a baseline-correction
procedure described earlier [13].
Optical conductivity is appropriate to fit with a Drude–
Lorentz model.We used the fitted oscillators for background
corrections. First we subtracted the contribution of the
oscillators which belong to the amorphous carbon back-
ground and to the p–p transition. Then we assigned the
2856 K. Ne´meth et al.: Invest
Figure 1 Reaction schemes of functionalization.remaining Lorentzians to different transitions (free carriers
so-called M00, S11, S22, etc.). For the analysis of a particular
transition we considered the Lorentzians which do not
belong to that specific transition as background and
subtracted them from the spectrum. We performed this
calculation for the S11, S22, and M11-S33 peaks, in the latter
case the different contributions were not separable. Figures 2
and 3 show the spectra related to the transitions. Each of them
contains information related to only one particular set of
peaks.
For Figs. 2 and 3 we consider as reference material the
hydrogenated nanotube sample produced in the 1st step. We
do not compare the spectra to the unreacted material (e.g.,
after annealing at 250 8C), because many manipulations
happened in the 1st step compared to this initial state, apart
from the reaction, that could not be investigated by infrared
spectroscopy.
3 Results and discussion According to thermogra-
vimetry-mass spectrometry (TG-MS) measurements (not

2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimation of hydrogenated HiPCo nanotubes by IR spectroscopytion increases with the number of steps in the order 1st step<
3rd step< 2nd step for reaction (A), and in the order 1st step
< 2nd step < 3rd step for reaction (B). The reason for this
behavior is probably an equilibrium process which will be
discussed elsewhere.
The optical conductivity of the functionalized nanotubes
in selected regions of the spectrum is shown in Figs. 2 and 3.
The changes are slight, according to the low degree of
functionalization, but there are distinct differences between
the products obtained by the two reaction types.
The decrease in optical conductivity upon functionaliza-
tion at the transitions between Van Hove singularities is
caused by the transformation of sp2 into sp3 carbon atoms.
We observe that this decrease is not uniform and depends on
the reaction mechanism: for reaction type (A), the low-
energy part is decreasing faster, while for reaction type (B),
the opposite happens. The effect is visible for both S11 and
S22 transitions (Fig. 2). In Fig. 3 we enlarged the S11 region
(6000–8500 cm
1) and illustrate the changes on these
transitions.
www.pss-b.com