Solubilization of single-walled carbon nanotubes by using polycyclic aromatic ammonium amphiphiles in water - Strategy for the design of high-performance solubilizers

Article abstract of DOI:10.1002/chem.200501176

We describe the design of polycyclic aromatic compounds with high performance that dissolve single-walled carbon nanotubes (SWNTs). Synthetic amphiphiles trimethyl-(2-oxo-2-phenylethyl)-ammonium bromide (1) and trimethyl-(2-naphthalen-2-yl-2-oxo-ethyl)-am

Full text of DOI:10.1002/chem.200501176

FULL PAPER
DOI: 10.1002/chem.200501176
Solubilization of Single-Walled Carbon Nanotubes by using Polycyclic
Aromatic Ammonium Amphiphiles in Water—Strategy for the Design of
High-Performance Solubilizers
Yasuhiko Tomonari,[a] Hiroto Murakami,[a] and Naotoshi Nakashima*[b]
Abstract: We describe the design of
polycyclic aromatic compounds with
high performance that dissolve single-
walled carbon nanotubes (SWNTs).
Synthetic amphiphiles trimethyl-(2-
Transmission
electron
microscopy
that of SDS and HTAB. Near-IR pho-
toluminescence measurements revealed
that the chiral indices of the SWNTs
dissolved in an aqueous solution of 4
were quite different from those ob-
tained byusing micelles of SDS and
HTAB; for a SWNTs/4 solution, the in-
tensityof the (7,6), (9,5), and (12,1) in-
dices were strong and the chiralitydis-
tribution was narrower than those of
the micellar solutions. This indicates
that the aqueous solution of 4 has a
tendencyto dissolve semiconducting
SWNTs with diameters in the range of
0.89–1.0 nm, which are larger than
those SWNTs (0.76–0.97 nm) dissolved
in the aqueous micelles of SDS and
HTAB.
(TEM), as well as visible/near-IR, fluo-
rescence, and near-IR photolumines-
cence spectroscopies were employed to
reveal the solubilization properties of 4
in water, and to compare these results
with those obtained byusing sodium
dodecyl sulfate (SDS) and hexadecyl-
trimethylammonium bromide (HTAB)
as solubilizers. Compound 4 solubilized
both the as-produced SWNTs (raw-
SWNTs) and purified SWNTs under
mild experimental conditions, and the
solubilization abilitywas better than
oxo-2-phenylethyl)-ammonium
bro-
mide (1) and trimethyl-(2-naphthalen-
2-yl-2-oxo-ethyl)-ammonium bromide
(2) carrying a phenyl or a naphtyl
moietywere not able to dissolve/dis-
perse SWNTs in water. Bycontrast, tri-
methyl-(2-oxo-2-phenanthren-9-yl-
ethyl)-ammonium bromide (3) solubi-
lized SWNTs, although the solubiliza-
tion abilitywas lower than that of tri-
methyl-(2-oxo-2-pyrene-1-yl-ethyl)-am-
monium bromide (4) (solubilization be-
havior observed byusing 4 was descri-
Keywords: adsorption
cence spectroscopy · nanotubes · pi
interactions · solubilization
·
fluores-
bed
brieflyin
reference [4a]).
Introduction
Since their first report in 1991,[1] carbon nanotubes (CNTs)
have received much attention in nanoscience and nanotech-
nologybecause of their unique physical, chemical, mechani-
cal, and electronic properties.[2] Due to the insolubilityof
CNTs in both aqueous and organic solvents, their applica-
tions are limited. Our interest has focused on noncovalent
sidewall-functionalized soluble carbon nanotubes. Figure 1
Figure 1. Strategy to solubilize SWNTs by physical adsorption of polycy-
clic aromatic molecules carrying a solvophilic moiety onto the surfaces of
SWNTs.
[a] Y. Tomonari, Prof. Dr. H. Murakami
Department of Material Science
Graduate School of Science and Technology, Nagasaki University
Bunkyo, Nagasaki 852–8521 (Japan)
shows the general idea for the solubilization of CNTs in sol-
ution,[3] in which the solubilizers are functional compounds
that are composed of an aromatic moietyand a solvophilic
moiety. In this concept, solubilizers carrying a hydrophilic or
hydrophobic moiety are expected to dissolve CNTs in dipo-
lar solvents, such as water and alcohols, and in nonpolar sol-
[b] Prof. Dr. N. Nakashima
Department of Chemistryand Biochemistry
School of Engineering, Kyushu University
744 Motooka, Fukuoka 819–0395 (Japan)
Fax : (+81)92-802-2840
E-mail: nakashima-tcm@mbox.nc.kyushu-u.ac.jp
Chem. Eur. J. 2006, 12, 4027 – 4034
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4027

[4]
and diameter of raw-SWNTs were approximately1–10 mm and 0.7–
1.2 nm, respectively.
vents, respectively. Reports describing solubilization by
using aromatic compounds carrying pyrene,[4a,b,e,5] anthra-
cene,[6] and porphyrin[4c,7] have been published.
Syntheses
Trimethyl-(2-oxo-2-phenylethyl)-ammonium bromide (1): An excess
amount of trimethylamine gas was introduced into 2-bromoacetophenone
(0.112 g, 0.56 mmol) in dryTHF (15 mL) at RT, and the solution was stir-
red for 1 d. The precipitate was separated and dried in a vacuum to give
1 as a white solid, 84%. 1H NMR (300 MHz, [D4]CD3OD): d=7.5–8.4
(m, 5H; PhH), 3.4 ppm (s, 9H; (CH3)3N+); IR (KBr): n˜ =3054, 3000,
1689 cmÀ1; elemental analysis calcd (%) for C11H16NOBr: C 51.61, H
6.25, N 5.43; found: C 50.91, H 6.25, N 5.39.
In this paper, we describe the preparation of sidewall-
functionalized water-soluble single-walled nanotubes
(SWNTs) with four different polycyclic aromatic ammonium
amphiphiles (1–4). Compounds 1 to 4 are water-soluble am-
Trimethyl-(2-naphthalene-2-yl-2-oxo-ethyl)-ammonium bromide (2): An
excess amount of trimethylamine gas was introduced into 2-bromo-2’-ace-
tonaphtone (0.134 g, 0.53 mmol) in dryTHF (15 mL) at RT, and the solu-
tion was stirred for 1 d. The precipitate was separated and dried in a
vacuum to give
[D4]CD3OD): d=7.6–8.6 (m, 7H; PhH), 3.4 ppm (s, 9H; (CH3)3N+); IR
(KBr): n˜ =3054, 3008, 1697 cmÀ1
elemental analysis calcd (%) for
2 as a
white solid, 73%. 1H NMR (300 MHz,
;
C15H18NOBr: C 58.44, H 5.89, N 4.54; found: C 58.90, H 6.06, N 4.50.
Trimethyl-(2-oxo-2-phenanthrene-9-yl-ethyl)-ammonium bromide (3): An
excess amount of trimethylamine gas was introduced into 9-(2-bromoace-
tyl)phenanthrene (0.100 g, 0.33 mmol) in dry THF (15 mL) at RT, and
the solution was stirred for 1 d. The precipitate was separated and dried
in a vacuum to give 3 as a white solid, 88%. 1H NMR (300 MHz,
[D4]CD3OD): d=7.7–8.9 (m, 9H; PhH), 3.5 ppm (s, 9H; (CH3)3N+); IR
phiphiles possessing a condensed cyclic number 1, 2, 3, and
4 for 1, 2, 3, and 4, respectively, and are suitable to demon-
strate our general idea for the solubilization of CNTs in sol-
ution. We also used sodium dodecyl sulfate (SDS) and hexa-
decyltrimethylammonium bromide (HTAB) and compared
their solubilization abilities toward SWNTs with those of 1–
4. Carbon nanotubes have been reported to be soluble in
aqueous micellar solutions of such surfactants.[8]
A preliminaryreport obtained byusing compound 4 has
been published elsewhere.[4a] We have also reported that the
pyrene-carrying ammonium vinyl monomer and its copoly-
mers,[4b] as well as some porphyrins,[4c] can solubilize SWNTs
in water or organic solvents, and have described the impor-
tance of p–p interactions between the aromatic moieties
(KBr): n˜ =3000, 1697 cmÀ1
;
elemental analysis calcd (%) for
C19H20NOBr: C 63.68, H 5.63, N 3.91; found: C 63.50, H 5.69, N 3.84.
Trimethyl-(2-oxo-2-pyrene-1-yl-ethyl)-ammonium bromide (4): The syn-
[4a]
thetic method of 4 was brieflydescribed in a previous paper.
An
excess amount of trimethylamine gas was introduced into 1-(bromoace-
tyl)pyrene (0.191 g, 0.59 mmol) in dry THF (15 mL) at RT, and the solu-
tion was stirred for 1 d. The precipitate was separated and dried in
vacuum to give
4 as a
yellow solid, 87%. 1H NMR (300 MHz,
[D4]CD3OD): d=8.1–9.1 (m, 9H; PyH), 4.1 (s, 2H; CH2N+), 3.6 ppm (s,
9H; (CH3)3N+); IR (KBr): n˜ =3045, 3012, 1674 cmÀ1; elemental analysis
calcd (%) for C21H20NOBr·1.2H2O: C 62.45, H 5.59, N 3.47; found: C
62.16, H 5.21, N 3.06.
Purification of raw-SWNTs:[13] Raw-SWNTs (HiPco) were heated at
2258C in humid air, and then the samples were sonicated byusing an ul-
trasonic cleaner (Branson 5510) in concd aqueous HCl. The nanotubes
were then collected byusing a porous filter (Advantec PTFE; pore size,
100 nm), and then washed with sodium hydrogen carbonate. SWNTs
were collected byusing a porous filter (Advantec, PTFE; pore size,
100 nm), and then dried at 508C to obtain the purified SWNTs (p-
SWNTs), which were used for further experiments. Analysis of these
nanotubes byX-rayphotoelectron spectroscopy(XPS) revealed no peak
for Fe2p3/2 in the region of 706–714 eV, indicating that the Fe was almost
removed bythis purification procedure.
and SWNT sidewalls toward the solubilization of SWNTs in
[9]
solvents. Veryrecently, Guldi et al.
fabricated nanohybrids
composed of SWNTs, compound 4, and water-soluble metal-
loporphyrins, and reported that a rapid interhybrid charge
separation causes the reduction of the electron-accepting
SWNTs and the oxidation of the electron-donating porphyr-
in. Theyalso described the photoelectrochemical behaviors
at an indium tin oxide (ITO) electrode modified with a
nanocomposite of SWNTs–compound 4–polythiophene car-
boxylate.[10] We[4c] reported that zinc protoporphyrin IX sol-
ubilizes SWNTs in DMF. Sun et al.[11] described that
5,10,15,20-tetrakis(hexadecyloxyphenyl)-21H,23H-porphine
selectivelyinteracts with semiconducting SWNTs in chloro-
form. Paloniemi et al.[12] examined the solubilization behav-
iors of SWNTs byusing several kinds of aromatic com-
pounds and found that naphthalene groups carrying amino
and sulfonyl groups can dissolve SWNTs in water.
Solubilization/dispersion of SWNTs: Typical solubilization procedures
are as follows: Raw-SWNTs or p-SWNTs (ꢀ1.0 mg) were placed in an
aqueous solution of 1 (or 2, 3, 4, each 1 mm), HTAB (29 mm), or SDS (10
and 36 mm), followed bysonication in a bath-tpye sonicator (Branson
5510) for 1 h, and then optional sonication in a cup-horn sonicator (SMT,
UH-300) for 10 min at RT. After sonication, the samples were centri-
fuged (SIGMA, 3K30C or HITACHI KOKI, CS100GXL) at a given g-
value. The upper 70–80% of supernatant was then carefullydecanted.
Resolubilization of SWNTs/4 nanocomposite: An aqueous dispersion/sol-
ution of SWNTs/4 was passed through a filter (Advantec, PTFE; pore
size, 100 nm), and a residue (solid) on the filter paper was rinsed well
with D2O. The collected solid was placed in D2O, and then sonicated in a
bath-type sonicator (Branson 5510) for 10 min to produce a black, trans-
parent aqueous solution/dispersion.
Experimental Section
Transmission electron microscopy (TEM) and atomic force microscopy
(AFM): Typical procedures are as follows: A carbon-coated TEM grid
(Ouken-Shoji, 200-A mesh) was immersed in an aqueous solution of
SWNTs/4 for a few seconds and then air-dried. TEM measurements were
conducted byusing a Jeol JEM-100S electron microscope. An aqueous
Materials: SDS, HTAB, and as-produced SWNTs (raw-SWNTs) were
purchased from Nacalai Tesque, Wako Chemical Industries, and Carbon
Nanotechnologies, respectively, and were used as received. The length
4028
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4027 – 4034

Solubilization of Single-Walled Carbon Nanotubes
FULL PAPER
solution of SWNTs/4 was dropped onto a cleaved mica substrate, rinsed
with pure water (Milli-Q Plus Ultrapure water system, Millipore), and
then dried in vacuum before measurement. AFM images were recorded
byusing a SPI 3800N (Seiko Instruments) with a Si 3N4 cantilever (SN-
AF01).
solutions. We measured Raman spectra for SWNTs obtained
from solutions of SWNTs/3 and SWNTs/4 and for insoluble
SWNTs separated from solutions of 1 and 2 containing
SWNTs. As shown in Figure 3, the four spectra obtained (b–
Raman spectroscopy: Raman spectra
for SWNTs were recorded byusing a
Renishaw Ramanscope System 1000
(excitation wavelength, 514.5 nm of
an Ar ion laser).
Vis/near-IR (NIR) absorption and
fluorescence spectroscopy of individ-
ually dissolved SWNTs: Raw-SWNTs
and an aqueous solution of 4 (or SDS
or HTAB) were sonicated in a bath-
type sonicator (Branson 5510) for
1 h, and then in a cup-horn sonicator
(SMT, UH-300) for 10 min at RT.
After
centrifugation
(HITACHI
KOKI, CS100GXL) for 4 h at
118000 g, the supernatant was care-
fullyseparated. NIR fluorescence
Figure 3. Raman spectra for raw-SWNTs (a), insoluble raw-SWNTs (b and c) separated from solutions of
SWNTs/1 and SWNTs/2, respectively, and the solid products (d and e) obtained from solutions of raw-
SWNTs/3 and raw-SWNTs/4, respectively.
spectra were recorded byusing
a
HORIBA SPEX Fluorolog-3-NIR
spectrofluorometer equipped with a
liquid-nitrogen-cooled InGaAs near-IR detector. Excitation and emission
wavelengths were 500–900 and 900–1300 nm, respectively.
Thermogravimetric analysis (TGA): TGA curves (108CminÀ1) were re-
corded byusing a Shimazu TGA-50H instrument.
e) resemble that of the original SWNTs (a) and the intensity
ratios of the G-band/D-band of the five spectra differ only
slightly, indicating that functionalization did not damage the
tubes themselves.
Figure 4 shows typical visible/near-IR spectra of the four
different
solutions/dispersions
containing
SWNTs/1,
Results and Discussion
Solubilization/dispersion of raw-SWNTs: The solubilization/
dispersion of raw-SWNTs in water containing 1, 2, 3, or 4
was carried out byusing the typical procedure described in
the Experimental Section to obtain SWNTs/1, SWNTs/2,
SWNTs/3, and SWNTs/4. A photograph of the four solu-
tions obtained is shown in Figure 2. Although the photo of
the SWNTs dissolved in an aqueous solution of 4 has al-
Figure 4. Vis/near-IR absorption spectra of D2O solutions/dispersions of
raw-SWNTs/1 (a), raw-SWNTs/2 (b), raw-SWNTs/3 (c), and raw-SWNTs/
4 (d). Optical cell length: 1 mm, concentration of 1, 2, and 3: 1.0 mm.
SWNTs/2, SWNTs/3, and SWNTs/4, in which deuterated
water was used in place of normal water, as water absorbs
in the near-IR region. Aqueous solutions of SWNTs/3 and
SWNTs/4 showed characteristic absorption bands in the visi-
Figure 2. Aqueous solutions of raw-SWNTs/1 (a), raw-SWNTs/2 (b), raw-
SWNTs/3 (c), and raw-SWNTs/4 (d).
[4a]
readybeen reported,
it is displayed again in Figure 2d for
ble/near-IR regions, due to the interband transition between
[14]
the purpose of comparison with the solutions obtained by
using 1–3. Compound 3, a tricyclic aromatic amphiphile,
gave a pale-black solution (Figure 2c). In contrast, as can be
seen in Figure 2a and b, the use of 1 and 2 gave colorless
the mirror image spikes in the densityof states (DOSs)
of
the SWNTs. In contrast, almost no absorbance was detected
upon use of compounds 1 and 2, indicating that these com-
pounds did not act as SWNT solubilizers. The number of
Chem. Eur. J. 2006, 12, 4027 – 4034
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4029

N. Nakashima et al.
condensed polycyclic aromatic rings in 1–4 is crucial for the
solubilization of the SWNTs in water. It is evident that both
the phenyl and naphthalene groups are noneffective chro-
mophores in the design of SWNT solubilizers. Recently, Pal-
oniemi et al.[12] reported that naphthalene derivatives with
an amino group could disperse SWNTs in solvent, and they
indicated the contribution of a charge-transfer interaction
between the amino group and the SWNTs.
We have examined the effect of solvent on the solubiliza-
tion of SWNTs. Unfortunately, the solubility of compound 4
is limited, namelyit is soluble in water, water/methanol, and
water/DMF, but insoluble in methanol, DMF, and other
manysolvents. Therefore, we carried out experiments by
using D2O/CD3OD (9:1, v/v) and D2O/DMF (9:1, v/v) in
place of D2O, and found that SWNTs can be solubilized in
these solvents (data not shown).
Transmission electron microscopy(TEM) and atomic
force microscopy(AFM) were performed to reveal the
structures of the raw-SWNTs dissolved in water containing
3 or 4. We used a cationic surfactant, HTAB, and compared
the results to those obtained byusing 3 and 4. Typical exam-
ples of the TEM images obtained are shown in Figure 5, in
which manycatalyst particles of iron are seen upon using
HTAB as a solubilizer. Bycomparison, such particles were
scarce in the solutions of both SWNTs/3 and SWNTs/4, indi-
cating that compounds 3 and 4 were effective for not only
the dissolution, but also in purification of the raw-SWNTs.
Figure 5d and e show images of insoluble raw-SWNTs sepa-
rated from solutions of 1 and 2, respectively, in which
SWNTs are seen to contain a lot of metal nanoparticles.
These images resemble that of the used raw-SWNTs in the
absence of the functionalization procedure. We conducted
thermogravimetric analysis (TGA) experiments for the raw-
SWNTs and for the SWNTs obtained from a solution con-
taining 4 and SWNTs after centrifugation at 60000 g for 2 h.
The Fe contents in these two samples were 28 and 10%, re-
spectively.
As we described elsewhere,[4a] the absorption maxima of
the UV-visible spectra of the SWNTs/4 solution appeared at
234, 288, and 368 nm together with a shoulder at 400 nm,
which are almost identical to those of the 4 aqueous solution
only. This result suggests that the spectrum was governed by
4 in the bulk solution. The UV-visible spectrum of SWNTs/3
was also virtuallyidentical to that of the aqueous solution of
3 containing no SWNTs. The spectral shift of 4 bound to the
surfaces of the SWNTs is discussed below. Compound 4 is
not a micelle-forming amphiphile in water, and neither is
compound 3 because, from surface-tension measurements,
no critical micelle concentration (cmc) was observed in an
aqueous solution of 3. Therefore, the solubilization of the
SWNTs obtained byusing compounds 3 and 4 is not by
means of the so-called “micelle dissolution”, but instead,
physical adsorption through p–p interaction of the polycy-
clic aromatic moietyand the sidewalls of the SWNTs is sug-
gested to playan important role in solubilization. The de-
tails of the mechanism of the solubilization of the SWNTs
are discussed below.
A typical AFM image on mica obtained from a SWNTs/4
solution is shown in Figure 6. From the height profile of the
image, we observe a large percentage of the SWNTs whose
diameters are approximately1–1.5 nm, indicating that the
SWNTs are individuallydissolved in water.
Comparison of raw-SWNTs and purified (p)-SWNTs:
Hauge et al.[13] reported that raw-SWNTs have a tendency
to form strong bundled structures during purification pro-
cesses. Hence, it is suggested that the solubilization behaviors
of the raw-SWNTs and p-SWNTs are different. We exam-
ined the solubilization abilityof 4 toward the raw- and p-
SWNTs and compared the results with those obtained by
using SDS. As shown in Figure 7, following sonication in a
bath-type sonicator, the solubilization behaviors of the raw-
Figure 5. TEM images of aqueous solution/dispersions of raw-SWNTs/3 (a), raw-SWNTs/4 (b), raw-SWNTs/HTAB (c), and insoluble raw-SWNTs (d and
e) separated from solutions of 1 and 2, respectively.
4030
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4027 – 4034

Solubilization of Single-Walled Carbon Nanotubes
FULL PAPER
IR spectrum (spectrum f in Figure 7) that was virtuallyiden-
tical to reported data.[8c] However, for the p-SWNTs, the
spectrum obtained was broad (spectra d and e in Figure 7).
Evidently, compound 4 is a better solubilizer than SDS for
p-SWNTs.
The effect of centrifugation on the dissolution of the raw-
SWNTs byusing 4 was examined. We varied the g-values
from 1000 to 120000 g. Approximate SWNT solubility
values obtained in aqueous solutions of 4 prepared bycen-
trifugation at 1000 and 60000 g were 0.028 and
0.010 mgmLÀ1, respectively. As exhibited in Figure 8, the ab-
sorbance for the four SWNT aqueous solutions decreased as
the g-values increased, and all gave resolved spectra whose
shapes and peak maxima varied little. The fact that centrifu-
gation at much lower g-values (~1000 g) was effective for
the preparation of individually
Figure 6. Typical AFM image of an aqueous solution of raw-SWNTs/4.
dissolved SWNTs in water is
important for obtaining large
amounts of individuallydis-
solved SWNTs.
Fluorescence from pyrene:
Pyrene-carrying
compounds
are often used as fluorescence
probes because the fluores-
cence quantum yield of pyrene
is high.[15] We measured the
fluorescence spectra of aque-
ous solutions of 4 in the ab-
sence or presence of the raw-
SWNTs. At lower concentra-
tions of 4 below 10À6 m, only
monomer emission spectra
were observed at around
421 nm, and at concentrations
above 10À5 m, excimer emis-
Figure 7. Vis/near-IR absorption spectra for D2O solutions of SWNTs/4 (A) and SWNTs/SDS (B). The p-
SWNTs and the raw-SWNTs were used for spectral measurements a, b, d, e, and c, f, respectively. Sample-
preparation conditions for (a)–(f): sonication in a bath-type sonicator for 1 h, followed by centrifugation at
1000 g for 30 min (a and d); sonication in a bath-type sonicator for 1 h and then in a cup-horn sonicator for
10 min, followed bycentrifugation at 1000 g for 30 min (b and e); sonication in a bath-type sonicator for 1 h
and then in a cup-horn sonicator for 10 min, followed bycentrifugation at 60000 g for 4 h (c and f). Optical
cell length, 1.0 mm.
SWNTs (spectrum c) and p-SWNTs (spectrum a) were quite
different. The S1 band of the p-SWNTs appeared in the
region of 1200–1700 nm, which was shifted largelyto the
longer wavelength compared to that of the raw-SWNTs, and
the band was not well-resolved. These results suggested that
the p-SWNTs formed bundled structures under these experi-
mental conditions. The effect of the sonication power on the
combination of p-SWNTs and the solubilizer 4 was exam-
ined. This revealed that additional sonication for 10 min in a
cup-horn type sonicator following sonication in the bath-
type sonicator caused a drastic change in the spectral shape
and, as shown in spectrum b in Figure 7, the spectrum ob-
tained was almost identical to that of the raw-SWNTs. We
wish to emphasis that both the raw- and p-SWNTs can be
readilysolubilized under the “mild” experimental condi-
Figure 8. Vis/near-IR absorption spectra of aqueous solutions of raw-
SWNTs/4 prepared bysonication in a bath-type sonicator for 1 h, fol-
lowed bycentrifugation at 1000 g for 2 h (a), 10000 g for 2 h (b), 60000 g
for 2 h (c), and 118000 g for 2 h (d). Optical cell length, 1.0 mm for (a)–
(c) and 10 mm for (d).
tions. The use of SDS micelles for the dissolution of the
raw-SWNTs and sonication in both a bath-type sonicator
and a cup-horn sonicator, followed bycentrifugation at
60000 g produced dissolution giving the well-resolved near-
Chem. Eur. J. 2006, 12, 4027 – 4034
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4031

N. Nakashima et al.
sions were observed at around 501 nm together with mono-
mer emissions at around 421 nm. Unfortunately, the ob-
served fluorescence behaviors of the raw-SWNTs/4 solution
were virtuallythe same as those of the 4 aqueous solution
alone, suggesting that the fluorescence from the SWNTs/4
solution is derived from the nonbound (free) 4 in the solu-
tion. To remove the free 4 molecules (not bound to the
SWNTs), the solution was filtered byusing filter paper. The
resulting solid was then resolubilized in pure water bysoni-
cation in a bath-type sonicator (for details, see Experimental
Section). The UV/visible/near-IR spectrum for the resolubi-
lized solution obtained is shown in Figure 9, together with
Figure 10. Fluorescence spectra of an aqueous solution of
4 ([4]=
0.24 mm) (a) and a resolubilized aqueous solution of SWNTs/4 (b) (con-
centration of 4 was estimated to be 0.24 mm). Excitation wavelengths for
(a) and (b) are 367 and 357 nm, respectively.
compared the fluorescence intensityof solutions in the ab-
sence and presence of SWNTs possessing the same absorb-
ance of the pyrene chromophore. We have already reported
a similar fluorescence quenching of zinc protoporphyrin IX
adsorbed on the surface of the SWNTs in a DMF solu-
tion.[4c] Sun et al.[17] synthesized SWNTs tethered with pyr-
enes and observed the quenching of pyrene emission in solu-
tion. This quenching was explained byenergytransfers from
the tethered pyrene moieties to the nanotubes. They also
synthesized SWNTs tethered with porphyrins and described
the quenching of porphyrin emission as well as energy trans-
fers from the tethered porphyrins to the nanotubes. On the
[6a]
other hand, Murrayet al.
reported that anthracene adsor-
Figure 9. UV/Vis/near-IR absorption spectra for aqueous solutions of 4
(a), raw-SWNTs/4 (b and d), and resolubilized raw-SWNTs/4 (c and e).
The spectra of (d) and (e) are magnifications of (b) and (c), respectively.
Optical cell length, 1.0 mm.
bed on SWNTs in THF is stronglyluminescent. Hedderman
et al.[6b] also described that the fluorescence from anthracene
adsorbed on the SWNTs was not much different from that
observed in the absence of SWNTs. The fluorescence behav-
iors from fluorophores adsorbed on the surface of nanotubes
remain complex.
the spectra for a SWNTs/4 solution before resolubilization
and for an aqueous solution of 4 only. From the data, the
concentration of 4 in the resolubilized solution was calculat-
ed to be approximately0.24 m m, which means a decrease of
75% from the original concentration. The resolubilized sol-
ution obtained was stable for more than three months in a
refrigerator, that is, no precipitate was produced during this
period. On the other hand, the decrease in absorbance of
the SWNTs in the resolubilized solution in the near-IR
region was only11%. The peak maxima of pyrene moieties
in the resolubilized solution appeared at 232, 287, and
357 nm, which are somewhat different from the original sol-
ution (peak maxima: 232, 288, and 367 nm). For the resolu-
bilized solution, the observed blue-shift of approximately
10 nm in the region of 360 nm together with the decrease in
the shoulder peak near 400 nm relative to the original solu-
tion is due to a p–p interaction between the SWNT side-
walls and the pyrene moiety.
[8c]
Photoluminescence from SWNTs: Smalleyet al.
and
Weisman et al.[8d] reported that raw-SWNTs dissolved in an
aqueous micelle of SDS showed individual photolumines-
cence in the near-IR region. Since these reports were pub-
lished, considerable attention has been focused on these
unique optical behaviors. We measured the near-IR photolu-
minescence spectra of SWNTs dissolved in an aqueous (deu-
terated) solution of 4, and compared the results to those ob-
tained byusing aqueous micelles of SDS and HTAB. Three
sample solutions were prepared under the same experimen-
tal conditions, namely, by sonication in a bath-type sonicator
for 1 h followed bysonication in a cup-horn sonicator for
10 min and then ultracentrifugation at 118000 g for 4 h. The
three samples obtained exhibited photoluminescence in the
near-IR region and the contour plots of the excitation wave-
length (500–900 nm)/emission wavelength (900–1300 nm)
profile are demonstrated in Figure 11. The contour plot for
SWNTs/SDS proved the existence of SWNTs with chiral in-
Fluorescence spectra provided direct evidence for the in-
teraction between compound 4 and the SWNTs. As shown
in Figure 10, the excimer emission from the pyrene moiety
in the resolubilized solution was drasticallyquenched com-
pared to that of the original solution, suggesting an energy
transfer from the pyrene moiety on 4 to the SWNTs. We
dices of (7,6), (8,6), (9,4), and (9,5), which resembled those
[18]
previouslyreported.
However, there were some differen-
ces: in the literature, the (9,5) index was not detected and
the intensityof (7,5) in our SDS sample was weaker than
4032
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4027 – 4034

Solubilization of Single-Walled Carbon Nanotubes
FULL PAPER
Figure 11. Contour plots of photoluminescence spectra for D2O solutions of raw-SWNTs/SDS (a), raw-SWNTs/HTAB (b), and raw-SWNTs/4 (c) as a
function of excitation and emission wavelengths.
that reported in the literature. This difference could be due
to the experimental procedure and lot numbers of HiPco
SWNTs obtained from Carbon Nanotechnologies. As shown
in Figure 11a and b, SWNTs/SDS and SWNTs/HTAB gave
verysimilar contour plots, whereas SWNTs/ 4 produces a dif-
ferentlypatterned contour plot (Figure 11c), in which the in-
tensityof the (7,6), (9,5), and (12,1) indices were strong and
the intensityof (8,6) was weaker than those from SWNTs/
SDS and SWNTs/HTAB. Notably, the index (12,1) in Figur-
e 11c is not detected in plots a and b. The diameter of
SWNTs with the index (12,1) corresponds to 0.995 nm,
which is the largest diameter among the observed indices.
The p–p interaction between the (12,1) SWNTs and com-
pound 4 might playa role in the solubilization of the nano-
tubes. One more characteristic feature observed for SWNTs/
copyrevealed that the excimer emission from the pryene
moietyon 1 in the resolubilized solution was quenched dras-
ticallyrelative to that of the original solution. Taken togeth-
er, one possible model for the solubilization with 4 is pre-
4 is a red-shift in the spectra relative to those of SWNTs/
[8f]
SDS and SWNTs/HTAB. Smalleyet al.
measured the
near-IR photoluminescence of individuallydissolved
SWNTs in micelles of various surfactants. Theyfound that
SDS and sodium dodecylbenzene sulfate gave the bluest-
shifted spectra, and HTAB and Brij 700 showed a red-shift.
Theydescribed that water at the nanotubes surface causes a
shift in the fluorescence peak.
Figure 12. Possible mechanism for the solubilization of SWNTs with 4.
sented in Figure 12, in which the adsorption of the monomer
and dimer of the solubilizer onto the surfaces of SWNTs is
drawn schematically. Here, the adsorption of the pyrene
moietyonto the SWNTs through a p–p interaction plays an
important role in solubilization. A minor contribution of the
cation–p interaction[12,19] between the ammonium moiety
and the SWNTs might need to be considered. The simple
mechanism presented would be applicable to a varietyof ar-
omatic amphiphiles, such as compound 3.
Mechanism of solubilization: As we described above for the
solubilization/dispersion of the raw-SWNTs, the mechanism
for solubilization with 4 is different from the so-called “mi-
cellar solubilization” with surfactants. In the latter, the ad-
sorption of a hemimicelle onto the nanotube is believed to
be important for the solubilization of the nanotubes, and the
surfactant molecules are in dynamic equilibrium between
the bulk phase and the sidewalls of the SWNTs. We con-
ducted a dialysis experiment for a SWNTs/4 aqueous solu-
tion byusing membrane tubing (molecular weight cut-off
3500, Spectrapor Spectrum Medical Industries). The solubil-
izer 4 leaked graduallyfrom the inside of the tubing to the
outer water phase and caused precipitation of the SWNTs
within the tubing. These results indicate that molecules of 4
are in a state of dynamic equilibrium between the surfaces
of the SWNTs and the bulk solution. Fluorescence spectros-
Conclusion
We have described the importance of the condensed polycy-
clic aromatic moietyin the design of carbon-nanotube solu-
bilizers. Ammonium amphiphiles carrying a phenyl or naph-
thyl group did not act as carbon-nanotube solubilizers; in-
Chem. Eur. J. 2006, 12, 4027 – 4034
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4033

N. Nakashima et al.
10234; g) M. ꢄlvaro, P. Atienzar, J. L. Bourdelande, H. Garcꢅa,
Chem. Phys. Lett. 2004, 384, 119; h) L. S. Fifield, L. R. Dalton, R. S.
Addleman, R. A. Galhotra, M. H. Engelhard, G. E. Fryxell, C. L.
Aardahl, J. Phys. Chem. B 2004, 108, 8737; i) A. B. Artyukhin, O.
Bakajin, P. Stroeve, A. Noy, Langmuir 2004, 20, 1442.
stead, the phenanthlyl and pyrenyl groups play an important
role in the solubilization. Pyrene ammonium (4), an amphi-
phile that does not form micelles, acted as an excellent solu-
bilizer for both the raw-SWNTs and p-SWNTs. Aqueous
solutions of SWNTs/4 exhibited interesting fluorescence
properties. The efficient fluorescence quenching of the
pyrene chromophore suggested energy transfer from the
pyrene group chemisorbed on the SWNTs to the SWNTs.
The most interesting feature is the selective dissolution of
SWNTs by 4, indicated bythe near-IR photoluminescence
measurements that revealed that the aqueous SWNTs/4 pos-
sesses (7,6), (9,5), and (12,1) indices. This means that aque-
ous solution of 4 has a tendencyto dissolve semiconducting
SWNTs with diameters in the range of 0.89–1.0 nm. These
were larger than those of aqueous micelles of SDS and
HTAB (diameter range of 0.76–0.97 nm). This fine discrimi-
nation in the diameters of the SWNTs is of interest in the
design of nanotube solubilizers that recognize a single nano-
tube chiral index, as the synthesis of SWNTs with a single
chiral index is presentlydifficult, and also for their potential
applications in manyfields of science and technology.
[6] a) J. Zhang, J.-K. Lee, Y. Wu, R. W. Murray, Nano Lett. 2003, 3, 403;
b) T. G. Hedderman, S. M. Keogh, G. Chambers, H. J. Byrne, J.
Phys. Chem. B 2004, 108, 18860.
[7] a) D. M. Guldi, G. N. A. Rahman, J. Ramey, M. Marcaccio, D. Pao-
lucci, F. Paolucci, S. Qin, W. T. Ford, D. Balbinot, N. Jux, N. Tagma-
tarchis, M. Prato, Chem. Commun. 2004, 2004; b) D. M. Guldi, H.
Taieb, G. M. A. Rahman, N. Tagmatarchis, M. Prato, Adv. Mater.
2005, 17, 871; c) D. Baskaran, J. W. Mays, X. P. Zhang, M. S. Bratch-
er, J. Am. Chem. Soc. 2005, 127, 6916; d) A. Satake, Y. Miyajima, Y.
Kobuke, Chem. Mater. 2005, 17, 716.
[8] a) V. Krstic, G. S. Duesberg, J. Muster, M. Burghard, S. Roth, Chem.
Mater. 1998, 10, 2338; b) N. Nakashima, H. Kobae, T. Sagara, H.
Murakami, ChemPhysChem 2002, 3, 456; c) M. J. OꢆConnell, S. M.
Bachio, C. B. Huffman, V. C. Moore, M. S. Strano, E. H. Haroz,
K. L. Rialon, P. J. Boul, W. H. Noon, C. Kittrell, J. Ma, R. H. Hauge,
R. B. Weisman, R. E. Smalley, Science 2002, 297, 593; d) S. M. Bach-
ilo, M. S. Strano, C. Kittrell, R. H. Hauge, R. E. Smalley, R. B. Weis-
man, Science 2002, 298, 2361; e) M. F. Islam, E. Rojas, D. M. Bergey,
A. T. Johnson, A. G. Yodh, Nano Lett. 2003, 3, 269; f) V. C. Moore,
M. S. Strano, E. H. Haroz, R. H. Hauge, R. E. Smalley, Nano Lett.
2003, 3, 1379; g) R. B. Weisman, S. M. Bachilo, Nano Lett. 2003, 3,
1235; h) C. Richard, F. Balavoine, P. Schultz, T. W. Ebbesen, C.
Mioskowski, Science 2003, 300, 775; i) K. Yurekli, C. A. Mitchell, R.
Krishnamoorti, J. Am. Chem. Soc. 2004, 126, 9902; j) M. S. Strano,
C. B. Huffman, V. C. Moore, M. J. OꢆConnell, E. H. Haroz, J. Hub-
bard, M. Miller, K. Rialon, C. Kittrell, S. Ramesh, R. H. Hauge,
R. E. Smalley, J. Phys. Chem. B 2003, 107, 6979.
Acknowledgements
We thank Prof. H. Furuta of Kyushu University for the use of
a
HORIBA SPEX Fluorolog-3-NIR spectrofluorometer and Prof. S. Mar-
uyama and Mr. Y. Miyauchi of The University of Tokyo for their techni-
cal assistance in the preparation of the contour plots of the photolumi-
nescence data. This work was supported byGrants-in-Aid from the Japa-
nese Ministryof Education, Science, Sports, and Culture.
[9] a) D. M. Guldi, G. M. A. Rahman, N. Jux, N. Tagmatarchis, M. Prao,
Angew. Chem. 2004, 116, 5642; Angew. Chem. Int. Ed. 2004, 43,
5526; b) D. M. Guldi, G. M.A. Rahman, N. Jux, D. Balbinot, N. Tag-
matarchis, M. Prato, Chem. Commun. 2005, 2038.
[10] a) D. M. Guldi, G. M.A. Rahman, M. Prato, N. Jux, S. Qin, W. Ford,
Angew. Chem. 2005, 117, 2051; Angew. Chem. Int. Ed. 2005, 44,
2015; b) G. M. A. Rahman, D. M. Guldi, R. Cagnoli, A. Mucci, L.
Schenetti, L. Vaccari, M. Prato, J. Am. Chem. Soc. 2005, 127, 10051.
[11] H. Li, B. Zhou, Y. Lin, L. Gu, W. Wang, K. A. S. Fernando, S.
Kumar, L. F. Allard, Y.-P. Sun, J. Am. Chem. Soc. 2004, 126, 1014.
[12] H. Paloniemi, T. ꢇꢈriralo, T. Laiho, H. Like, N. Kocharova, K. Haa-
pakka, F. Terzi, R. Seeber, J. Lukkari, J. Phys. Chem. B 2005, 109,
8634.
[1] S. Iijima, Nature 1991, 354, 56.
[2] a) Carbon Nanotubes and Related Structures (Ed.: P. J. F. Harris),
Cambridge UniversityPress, Cambridge, 1999; b) The Science and
Technology of Carbon Nanotubes (Eds.: K. Tanaka, T. Yamabe, K.
Fukui), Elsevier, Oxford, 1999; c) S. Niyogi, M. A. Hamon, H. Hu,
B. Zhao, P. Bhowmik, R. Sen, M. E. Itkis, R. C. Haddon, Acc.
Chem. Res. 2002, 35, 1105; d) A. Hirsch, Angew. Chem. 2002, 114,
1933; Angew. Chem. Int. Ed. 2002, 41, 1853; e) S. Banerjee, M. G. C.
Kahn, S. S. Wong, Chem. Eur. J. 2003, 9, 1893; f) D. Tasis, N. Tagma-
tarchis, V. Georgakilas, M. Prato, Chem. Eur. J. 2003, 9, 4000;
g) Carbon Nanotubes (Eds.: S. Reich, C. Thomsen, J. Maultzsch),
Wiley-VCH, Berlin, 2004.
[13] I. W. Chiang, B. E. Brinson, A. Y. Huang, P. A. Willis, M. J. Broni-
kowski, J. L. Margrave, R. E. Smalley, R. H. Hauge, J. Phys. Chem.
B 2001, 105, 8297.
[14] T. Ando, J. Phys. Soc. Jpn. 1997, 66, 1066.
[15] a) Photophysics of Aromatic Molecules (Ed.: J. B. Birks), John
Wiley, New York, 1970; b) Principles of Fluorescence Spectroscopy
(Ed.: J. R. Lakowicz), Kluwer Academic/Plenum, 1999.
[16] a) L. Qu, R. B. Martin, W. Huang, K. Fu, D. Zweifel, Y. Lin, Y.-P.
Sun, C. E. Bunker, B. A. Harruff, J. R. Gord, L. F. Allard, J. Chem.
Phys. 2002, 117, 8089; b) R. B. Martin, L. Qu, Y. Lin, B. A. Harruff,
C. E. Bunker, J. R. Gord, L. F. Allard, Y.-P. Sun, J. Phys. Chem. B
2004, 108, 11447.
[17] H. Li, R. B. Martin, B. A. Harruff, R. A. Carino, L. F. Allrd, Y.-P.
Sun, Adv. Mater. 2004, 16, 896.
[18] Y. Miyauchi, S. Chiashi, Y. Murakami, Y. Hayashida, S. Maruyama,
Chem. Phys. Lett. 2004, 387, 198.
[3] N. Nakashima, Int. J. Nanosci. 2005, 4, 117.
[4] a) N. Nakashima, Y. Tomonari, H. Murakami, Chem. Lett. 2002,
638; b) N. Nakashima, S. Okuzono, Y. Tomonari, H. Murakami,
Trans. Mater. Res. Soc. Jpn. 2004, 29, 525; c) H. Murakami, T.
Nomura, N. Nakashima, Chem. Phys. Lett. 2003, 378, 481; d) N. Na-
kashima, S. Okuzono, H. Murakami, T. Nakai, K. Yoshikawa, Chem.
Lett. 2003, 32, 456; e) N. Nakashima, Y. Tanaka, Y. Tomonari, H.
Murakami, H. Kataura, T. Sakaue, K. Yoshikawa, J. Phys. Chem. B
2005, 109, 13076.
[5] a) R. J. Chen, Y. Zhang, D. Wang, H. Dai, J. Am. Chem. Soc. 2001,
123, 3838; b) H. Xin, A. T. Woolley, J. Am. Chem. Soc. 2003, 125,
8710; c) F. J. Gꢁmez, R. J. Chen, D. Wang, R. M. Waymouth, H.
Dai, Chem. Commun. 2003, 190; d) P. Petrov, F. Stassin, C. Pag-
noulle, R. Jꢂrꢃme, Chem. Commun. 2003, 2904; e) L. Liu, T. Wang,
J. Li, Z.-X. Guo, L. Dai, D. Zhang, D. Zhu, Chem. Phys. Lett. 2003,
367, 747; f) K. A. S. Fernando, Y. Lin, W. Wang, S. Kumar, B. Zhou,
S.-Y. Xie, L. T. Cureton, Y.-P. Sun, J. Am. Chem. Soc. 2004, 126,
[19] a) J. C. Ma, D. A. Dougherty, Chem. Rev. 1997, 97, 1303; b) T. Fu-
kushima, A. Kosaka, Y. Ishimura, T. Yamamoto, T. Takigawa, N.
Ishii, T. Aida, Science 2003, 300, 2072.
Received: September 24, 2005
Published online: March 21, 2006
4034
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4027 – 4034