Journal of Colloid and Interface Science 286 (2005) 216223 www.elsevier.

com/locate/jcis

An explanation of dispersion states of single-walled carbon nanotubes in solvents and aqueous surfactant solutions using solubility parameters
Hyeong Taek Ham a , Yeong Suk Choi b , In Jae Chung a,
a Department of Chemical and Biomolecular Engineering, KAIST (Korea Advanced Institute of Science and Technology), 373-1 Guseong-dong, Yuseongu,

Daejeon, South Korea

b SAIT (Samsung Advanced Institute of Technology), Mt. 14-1, Nongseo-Ri, Giheung-Eup, Yongin-Si, Gyeonggi-do, South Korea

Received 15 September 2004; accepted 5 January 2005 Available online 9 March 2005

Abstract Dispersions of single-walled carbon nanotubes in various solvents and aqueous surfactant emulsions were investigated to correlate the 2 + 2 + 2 ). It was found that the nanotubes were dispersed or suspended degree of dispersion state with Hansen solubility parameters (t2 = d p h very well in the solvents with certain dispersive component (d ) values. They were precipitated in the solvents with high polar component (p ) values or hydrogen-bonding component (h ) values. The solvents in the dispersed group occupied a certain region in a 3-dimensional space of three components. The surfactants with a lipophilic group equal to and longer than decyl, containing 9 methylene groups and 1 methyl group, contributed to the dispersion of nanotubes in water. The surfactants in the dispersed group had a lower limit in the dispersive component (d ) of the Hansen parameter. 2005 Elsevier Inc. All rights reserved.
Keywords: Single-walled carbon nanotubes; Dispersion; Solubility parameter

1. Introduction Iijima and Ichihashi synthesized single-walled carbon nanotubes using graphite and metallic catalysts under high vacuum in 1993 [1]. The nanotubes are produced using various catalysts, such as molybdenum, cobalt, ironnickel bimetal, and metallic oxides. The general shape of the nanotubes is a cylindrical material, which is composed of sixmembered carbon rings. The nanotubes have average diameters and lengths of about 1.4 nm and 10 m, the diameters depending on the particle sizes of the catalysts [2]. Intensive research on single-walled carbon nanotubes is important because the nanotubes have a high exibility, chemical stability, and electric conductivity [36]. The unique properties result from the high ratio of length to diameter and the connections among adjacent carbons,
* Corresponding author. Fax: +82-42-869-3910.

E-mail address: chung@kaist.ac.kr (I.J. Chung). 0021-9797/$ see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2005.01.002

so the nanotubes have several potential applications, such as nanodevices, or reinforcing materials for polymer composites. Even though carbon nanotubes have promising applications, fundamental problems still exist: how to (1) remove impurities, such as amorphous carbons and metallic catalysts, and (2) obtain uniform dispersions of the carbon nanotubes in dispersing media or polymer solutions. The impurities in unpuried carbon nanotubes severely reduce the mechanical or electrical properties. Heating unpuried nanotubes about 300 C in air effectively eliminates amorphous carbons, and acid treatment excludes metallic catalysts from the nanotubes. The dispersion states of the nanotubes involve complicated phenomena, since the carbon nanotubes are produced in bundles or bundle aggregations. The states are affected by at least two competitive interactions: (1) the interactions of van der Waals forces, among carbon nanotube threads, and (2) the interactions between carbon nanotube threads and dispersion medium. Sonication of single-walled carbon nanotubes leads mostly to thin bundles plus some individ-

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

217

ual nanotubes and shortens the nanotubes in 100300 nm lengths with open ends. Carboxylic acid groups are introduced at the tube ends and side walls that can connect nanotubes with organic materials forming covalent bondings via acid chloride condensation [711]. This method has many advantage for obtaining nanotube dispersions because the van der Waals forces among nanotubes are eliminated and the surface characters are changed by acid. However, the oxidation by acids causes defects on the nanotube walls and makes the nanotubes short, which diminish the electrical and mechanical properties [11]. Dispersing the nanotubes without dilapidating connections or lengths can facilitate the use of carbon nanotubes [12,13]. Nanotube dispersions based on physical adsorption of surfactants or polymers on graphene sheets are a representative method [11,1428]. The adsorptions relate to the chemical or physical structures of dispersing media. Several surfactants or dispersing materials have been reported independently. However, the dispersion states were not explained systematically, yet. 2 + 2 + 2 , will be Hansen solubility parameters, t2 = d p h useful in explaining the dispersion state, because the parameters successfully illustrated the miscibility of hydrophobic polymers and polar solvents or the dispersion state of organoclays in organic solvents [2931]. Since the clay has a high aspect ratio and maintains strong interactions among individual clay layers or clay particles via van der Waals forces and hydrogen bondings, the clay dispersion state has some analogy with the carbon nanotubes [32,33]. The adsorption of a molecule on the clay is a local interaction between the hydrophilic part of the molecule and the clay surface [33]. Thus, we analyzed which solubility parameter most affects the dispersion states of the nanotubes in solvents. We expected that the dispersion component (d ) would be the principal factor determining the dispersion states of the nanotubes in solvents, because the nanotubes have mainly nonpolar and hydrophobic surfaces. In this work, we used various solvents to explain the dispersion of single-walled carbon nanotubes, after sonication. Through this examination, we determined that the dispersion state was related to the dispersive component (d ) of the solvent. To compare the dispersive component with the other components, we selected surfactants with diverse alkyl chain lengths and hydrophilic heads. Hydrophilelipophile balance (HLB) values were also considered for comparing the components.

o-methoxyphenol, 1-methyl-2-pyrrolidone, 1-pentyl alcohol, dichloromethane, octylsulfate sodium salt, octadecylsulfate sodium salt, octanoic acid sodium salt, 1-octyl alcohol, hexylamine, decylamine, dodecylamine, octadecylamine, and triethylamine were purchased from Aldrich and used as received. Acrylonitrile, methyl methacrylate, and styrene from Aldrich were passed through basic alumina in order to remove inhibitor as described in another paper [34]. 2.2. Dispersion of single-walled carbon nanotubes Single-walled carbon nanotubes produced by the HiPco process were purchased from Carbon Nanotechnologies Inc. Amorphous carbons were removed by heating the nanotubes at 300 C for 1 h in air atmosphere and water-soluble impurities were removed by the following method. The nanotubes were sonicated in deionized water (100 ml) for 10 h at room temperature and mechanically stirred before ltration. The ltered nanotubes were dispersed in deionized water (30 ml), and then freeze-dried to avoid the aggregation. The experiments for dispersion of carbon nanotubes were performed in the following ways. We selected several solvents whose total solubility parameter values were quite different. The prewashed nanotubes (1 mg) were added into a vial containing a solvent (15 g). The concentration of the nanotubes was 6.7 105 wt%, which was diluted to prevent the aggregation. The vial was sealed with a cap and sonicated for 20 h at 40 kHz and 100 W. After sonication, the vial was left in a quiescent state for 480 h, with pictures taken every 24 h. The dispersion stability was determined by the variation of UV absorbance intensity with respect to time. Anionic and cationic surfactants and long-chain alcohols were selected for further examination of solubility parameter effects on the nanotube dispersions. Aqueous HCl or NaOH solution was added to an aqueous surfactant solution (0.1 M, 10 ml) to adjust the pH for saponication, and then the prewashed nanotubes (1 mg) were added to the surfactant solution. The concentration of the nanotubes was 1 104 wt%. The nal pH of the dispersion was 3 for the cationic surfactants and 10 for the anionic surfactants. The nanotube surfactant solution was sonicated, and then UV absorbance was measured by the method described in the previous section. 2.3. Measurements

2. Experimental 2.1. Materials Methyl alcohol, ethyl alcohol, 2-propyl alcohol, n-hexane, benzene, tetrahydrofuran, acetone, chloroform, sodium dodecyl sulfate salt, and toluene were purchased from Merck and used as received. N ,N -Dimethylformamide, dodecylbenzenesulfonic acid sodium salt, dimethyl sulfoxide, Sonication was carried out with a bath-type Branson 3510 sonicator with a 445 cm2 bottom area. Infrared spectra were recorded on a Bomem 102 FT-IR spectrometer with KBr pellets. A total of 40 scans taken at 4 cm1 of resolution were averaged. Thermogravimetric analyses (TGA) were carried out with a PerkinElmer thermobalance from room temperature to 800 C, with a rate of 10 C/min under air atmosphere.

218

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

Fig. 1. Thermal gravimetric curves for (a) neat single-walled carbon nanotubes, and (b) puried nanotubes via heating and washing under air circumstance.

Fig. 2. FT-IR spectra of (a) neat single-walled carbon nanotubes, and (b) puried nanotubes via heating and washing.

3. Results and discussion The shape of the nanotube was examined by a Philips CM-20 transmission electron microscope (TEM) and a JEOL JSM 5610 scanning electron microscope (SEM). A sample for TEM analysis was prepared by depositing the nanotube dispersion on a carbon-coated copper grid and blotted by a lter paper to form a thin deposition and the water was evaporated slowly. A sample for SEM analysis was not gold coated. UV absorbance was obtained by a Jasco UV/visible spectrometer V-530. UV absorbance was scanned within a survey range from 190 to 500 nm [22], and then a maximum intensity wavelength was chosen for measuring concentration variation. The absorbance of each solvent was subtracted from that of nanotube dispersion in order to remove the solvent effect on the intensity. Fig. 1 shows the thermal degradation of carbon nanotubes. Amorphous carbons in the neat nanotube (a) burnt away within 350 C, and carbon nanotube threads degraded around 350600 C. The amount of amorphous carbon in the neat nanotube was determined using the residual weight difference between the neat nanotube and the prewashed nanotube from thermogravimetric analysis at 300 C. The residual weight percentage of the neat nanotube (a) at 300 C was 94 wt%, and that of the prewashed nanotube (b) was 97 wt%, respectively. The amount of amorphous carbon in the neat nanotube was about 3 wt% which agreed with the manufacturers report. Fig. 2 shows FT-IR spectra of (a) the neat nanotube, and (b) the nanotube after heating and washing. If the nanotube

(a) Fig. 3. SEM (a) and TEM (b) micrographs of the puried nanotubes.

(b)

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

219

Fig. 4. Time-resolved UV absorbance intensities of nanotubes (a) in N ,N -dimethylformamide, (b) in dimethyl sulfoxide, and (c) in water. (a), (b), (c) represent dispersed group, swollen group, and sedimented group solvents, respectively. The left inset picture was taken at 20 h of sonication, and the right inset picture at 480 h after sonication for each group.

has defects or carboxylic acids after heating and washing, we can discern new absorbance bands due to CH stretching (about 2940 cm1 ) or C=O stretching (about 1740 cm1 ). The spectrum (b) of the nanotube shows no absorbance bands assigned to CH stretching or C=O stretching. The result indicates no formation of carboxylic acids or defects after heating and washing. The physical damage and lengths of the nanotubes after heating and washing were examined by using SEM (a)

and TEM (b) micrographs. Nanotube bundles are observed in Figs. 3a and 3b. Metallic catalyst particles were in 30 40 nm in size and the particles were enveloped by the tube ends. So we expect that the catalyst will not affect the dispersion states of the nanotubes. The micrographs indicate that the heating and washing process eliminates amorphous carbons without physical damage. The chromaticity and the pictures of the dispersions were exploited for the criteria of dispersion states. The dispersion states of the nanotubes in various solvents were classied into three groups as dispersed, swollen, and sedimented. Fig. 4 shows three representative states. Dispersed means that the liquid maintains a uniform black color without aggregation and precipitation of the nanotubes during sonication. UV absorbance intensity of the dispersed group decreases from 1.8 to 1.2 for 480 min after sonication. But after 15,000 min, almost no nanotube aggregate is observed. Swollen means that aggregated nanotubes and phase separation are observed during sonication and the aggregation is maintained in a quiescent state. The absorbance intensity of the dispersion decreases from 1.6 to 0.2. Sedimented means that big particles are scattering in solvent during sonication, and all nanotubes are precipitated within 10 min after sonication. The concentration drops quickly. Hansen solubility parameters of solvents and nanotube dispersion states are listed in Table 1. The solvents classied as the dispersed group have moderate or low hydrogenbonding strength. The dispersion state does not depend on the total solubility parameter (t ) of the solvent, because the solvent containing almost the same t value shows three different dispersion states. To explain the nanotube dis-

Table 1 Dispersion of nanotubes in various solvents with different values of Hansen parameters of the solvents Organic solvents N ,N -Dimethylformamide Chloroform 1-Methyl-2-pyrrolidone 2-Propyl alcohol 1-Pentyl alcohol Tetrahydrofuran Toluene o-Methoxyphenol Dichloromethane Benzene Dimethyl sulfoxide Styrene Methyl methacrylate Methanol Hexane Acetone Water Ethanol Acrylonitrile d a (MPa1/2 ) 17.4 17.8 18.0 15.8 16.0 16.8 18.0 18.0 18.2 18.4 18.4 18.6 13.7 15.1 15.3 15.5 15.6 15.8 16.4 p a (MPa1/2 ) 13.7 3 .1 12.3 6 .1 4 .5 5.7 1 .4 8 .2 6 .3 0 .0 16.4 1.0 9 .8 12.3 0 .0 10.4 16.0 8 .8 17.4 h a (MPa1/2 ) 11.3 5.7 7.2 16.4 13.9 8.0 2.0 13.3 6.1 2.0 10.2 4.1 6.1 22.3 0.0 7.0 42.3 19.4 6.8 t b (MPa1/2 ) 24.8 19.0 22.9 23.5 21.7 19.4 18.2 23.8 20.3 18.6 26.7 19.0 17.9 29.6 15.3 20.0 47.8 26.5 24.8 Molecular weight (g/mol) 73.10 119.38 99.13 60.10 88.15 72.11 92.14 124.14 84.93 78.11 78.13 104.15 100.12 32.04 86.18 58.08 18.02 46.07 53.06 Structure HCON(CH3 )2 CHCl3 HN((CH2 )3 CO), cyclo (CH3 )2 CHOH CH3 (CH2 )4 OH (CH2 )4 O, cyclo C6 H5 CH3 2-(CH3 O)C6 H4 OH CH2 Cl2 C6 H6 (CH3 )2 SO C6 H5 CH=CH2 H2 C=C(CH3 )CO2 CH3 CH3 OH CH3 (CH2 )4 CH3 CH3 COCH3 H2 O C2 H5 OH H2 C=CHCN Dispersion state Dispersed Dispersed Dispersed Swollen Swollen Swollen Swollen Swollen Swollen Swollen Swollen Swollen Sedimented Sedimented Sedimented Sedimented Sedimented Sedimented Sedimented

of Hansen solubility parameters.

a A.F.M. Barton, CRC Handbook of Solubility Parameters and Other Cohesion Parameters, CRC Press, Boca Raton, FL, 1991 [30]. b 2 = 2 + 2 + 2 , where is total Hildebrand solubility parameter, is dispersion component, is polar component, is hydrogen bonding component t p d h p t d h

220

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

Fig. 6. Time-resolved UV absorbance intensities of nanotubes (a) in decylamine, (b) in dodecybenzene sulfonic acid, sodium salt, and (c) in hyxylamine solutions. (a), (b), (c) represent dispersed group, swollen group, and sedimented group surfactants, respectively. The left inset picture was taken at 20 h of sonication, and the right inset picture at 480 h after sonication for each group.

(c)

Fig. 5. Plot of dispersion states versus Hansen solubility parameters of various solvents: (a) dispersion state, the dispersive components relationship; (b) the polar components, the hydrogen bonding components relationship; and (c) 3D plot of the three components of Hansen parameters.

persion, we plot the dispersion states against the Hansen parameters. The dispersion states are plotted against the dispersion component (d ), as shown in Fig. 5a, the swollen group solvents are distributed in d value ranges of 15.8 17 or 1819 MPa1/2 , and the dispersed group in the range of 1718 MPa1/2 . The plot illustrates that the dispersion state partially relates to the d value of the solvent, while hydrogen components (h ) and polar components (p ) affect the states in other ways. For example, we expected that o-methoxyphenol would belong to the dispersed group, since its d value is 18 MPa1/2 , but it is in the swollen group. We think that o-methoxyphenols high h value (13.3 MPa1/2 ) may affect the dispersion state. In other cases, acrylonitrile (d = 16.4) and ethanol (d = 15.8) make nanotubes sedimented, and they have a high p value and a high h value. In Fig. 5b, the dispersed group solvents have p values less than 14 MPa1/2 , and h values less than 12 MPa1/2 . They seem to be distributed randomly, but we can nd upper limits of p and h values to which the dispersed group solvents belong. In Fig. 5c, a 3-dimensional plot of Hansen parameters, there seems to be a certain space where the dispersed group solvents locate in a threecomponent axis. This phenomenon indicates that d is important for dispersing nanotubes in solvents. It also indicates that Hansen solubility parameters rather than the total solubility parameter can explain the interactions between the nanotubes and the solvents. To investigate the effect of d in detail, surfactants with various structures or different alkyl chain lengths with similar structures were used. The dispersion states of nanotubes in aqueous surfactant solutions were also divided into three groups as dispersed, swollen, and sedimented. Fig. 6 shows the three representative groups. The dispersed group does

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

221

Table 2 Dispersion of nanotubes in aqueous surfactant solutions and the molecular structures, the types, and the HLB values of the surfactants Surfactants Dodecyl sulfate, sodium salt Decylamine Dodecylamine Hexadecylamine Triethylamine Octylsulfate, sodium salt Dodecylbenzene sulfonic acid, sodium salt Octanoic acid, sodium salt 1-Pentanol 1-Octanol Hexylamine Octadecylamineb Octadecylsulfate, sodium saltb Structures
+ CH3 (CH2 )11 SO 4 Na CH3 (CH2 )9 NH2 CH3 (CH2 )11 NH2 CH3 (CH2 )15 NH2 (CH3 CH2 )3 N + CH3 (CH2 )7 SO 4 Na + CH3 (CH2 )11 C6 H4 SO 3 Na

HLB numbersa 40.0 11.7 10.7 8.8 13.6 41.9 9.5 22.8 6.5 5.1 13.6 7.9 37.2

Surfactant type Anionic Cationic Cationic Cationic Cationic Anionic Anionic Anionic Nonionic Nonionic Cationic Cationic Anionic

Dispersion state Dispersed Dispersed Dispersed Dispersed Swollen Swollen Swollen Sedimented Sedimented Sedimented Sedimented

CH3 (CH2 )6 COO Na+ CH3 (CH2 )4 OH CH3 (CH2 )7 OH CH3 (CH2 )5 NH2 CH3 (CH2 )17 NH2 + CH3 (CH2 )17 SO 4 Na

(hydrophilic group numbers) + (hydrophobic group numbers) + 7. HLB group numbers were listed on other papers [3537]. Amine functionality (NH2 ) was looked on as tertiary amine [38]. b In case of ODA and SODS, the viscosity of their aqueous solution was too high and formed gel during sonication. Table 3 Hansen parameters and molar volumes of the lipophilic groups of surfactants Lipophilie Methylb Ethyla Propylb Butylb Pentylb Hexylb Octylb Decylb Dodecyl (lauryl)a Pentadecyl (palmitate)a Hexadecyla Heptadecyl (stearate)a Octadecyl (stearyl)b Nonadecylb Formula CH3 CH3 CH2 CH3 (CH2 )2 CH3 (CH2 )3 CH3 (CH2 )4 CH3 (CH2 )5 CH3 (CH2 )7 CH3 (CH2 )9 CH3 (CH2 )11 CH3 (CH2 )14 CH3 (CH2 )15 CH3 (CH2 )16 CH3 (CH2 )17 CH3 (CH2 )18 V (cm3 /mol) 34.1 50.3 66.5 82.6 98.8 114.9 147.2 179.5 211.3 260.2 275.7 293.2 308.8 325.0 (MPa1/2 ) d 12.6 14.3 15.1 15.6 15.9 16.1 16.4 16.5 16.7 16.8 16.8 16.8 16.9 16.9 p 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 h 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

a HLB number was calculated by the equation of Davies in accordance with their structures [35]. HLB =

a A.F.M. Barton, CRC Handbook of Solubility Parameters and Other Cohesion Parameters, second ed., CRC Press, Boca Raton, FL, 1991 [31]. b Calculated by the equation of Hansen and Hildebrand in accordance with the molar volume and cohesive energy of methyl and ethyl groups [39,40], 2 + 2 + 2 , = (E/V )1/2 . t2 = d p h

not show nanotube particles even after 480 h in a quiescent state following the sonication. The initial concentration is 1.0 104 wt% , and the initial absorbance is 2.6. The concentration of nanotubes for the dispersed group only drops by 15% after 480 h have passed. Surfactants are classied in three types of nonionic, cationic, and anionic according to the charge types of hydrophilic heads. They may have nonpolar and lipophilic tails with hydrocarbons. The hydrophilelipophile balance between head and tail can be used as a measure to determine the effect of surfactant on the stability of the dispersion of nanotubes in water. HLB values, molecular structures, surfactant types, and dispersion state of nanotubes are listed in Table 2. As shown in Table 2, the dispersion of nanotubes

in an aqueous surfactant solution does not depend on HLB values, molecular structures, and surfactant types. To correlate the dispersion states in aqueous surfactant solutions with solubility parameters, solubility parameters and molar volumes of lipophilic portions of surfactants were calculated and are listed in Table 3 in which the lipophilic formula corresponded to the hydrocarbons group of the surfactant in Table 2. Molar volumes of surfactants are linearly increased with the chain lengths, as shown in Fig. 7a. The dispersive component of the Hansen solubility parameter increases and then levels off to about 17 MPa1/2 in Fig. 7b. Nanotubes are well dispersed in surfactant solutions, when the surfactants have lipophilic alkyl chain lengths equal to a decyl group or longer. In other words, stable nanotube

222

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

(a)

Fig. 7. Effect of the lipophilic alkyl chain length of surfactants on (a) dispersion statemolar volume relationship, and (b) dispersion statedispersive component relationship.

(b) Fig. 8. SEM micrographs of nanotubes dispersed (a) in NMP, and (b) with sodium dodecyl sulfate salt in water.

dispersions can be obtained, if the dispersing media have dispersive component values more than 16.5 MPa1/2 and molar volumes larger than 179.5 cm3 /mol. When the surfactants, such as octadecylamine or octadecylsulfate sodium salt, have alkyl chain lengths longer than octadecyl, gelation occurs during sonication. We could not determine that there would be an upper limit of the molar volume and the Hansen solubility parameters according to the hydrophobic chain length. Nanotube dispersion states were veried using SEM images. Fig. 8 shows the SEM images of the nanotubes in solvent 1-methyl-2-pyrrolidone (NMP) (a), and with surfactant sodium dodecyl sulfate salt (SDS) (b). The dispersions were dropped on the silicon wafer and freeze-dried to prevent self-aggregation during the solvent vaporization. By UV/vis absorbance measurements, their states were classied as dispersed group. In both Figs. 8a and 8b, the tubes are scattered in bundled structures with average diameters of 30 nm. Small white dots may be residual catalyst particles. The SEM images give additional information on the

dispersion states. In Fig. 8b, surfactant micelles and their aggregations, sphere with diameter of 100 nm, are located on and outside the tube walls.

4. Conclusion The dispersion states of single-walled carbon nanotubes in various solvents and aqueous surfactant solutions were illustrated with the components in Hansen solubility parameters and molar volumes. The dispersion state of the nanotubes did not depend on the total solubility parameter of solvents or the HLB value of the surfactants, but depended on the dispersive component of the solubility parameter. The nanotubes in the solvents with high polar or hydrogen components were precipitated because nanotubes had hydrophobicity. It was found that there existed a dispersive component region of solvents, where the nanotubes dispersed well. Surfactants dispersed the nanotubes if they had hydrophobic

H.T. Ham et al. / Journal of Colloid and Interface Science 286 (2005) 216223

223

alkyl chains equal to and longer than a decyl group, regardless of their functionality or their hydrophilic heads. The surfactants in the dispersed group had a lower limit in hydrophobic alkyl chain length.

Acknowledgments The authors express their sincere thanks to KOSEF (Korea Science and Engineering Foundation), to CAFPoly (Center for Advanced Functional Polymers), and to the BK 21 program for their nancial support.

References
[1] S. Iijima, T. Ichihashi, Nature 363 (1993) 603. [2] C.N.R. Rao, B.C. Satishkumar, A. Govindaraj, M. Nath, ChemPhysChem 2 (2001) 78. [3] D.S. Bethune, C.H. Kiang, M.S. de Vries, G. Gorman, R. Savoy, J. Vazquez, R. Bevers, Nature 363 (1993) 605. [4] T.W. Ebbesen, Carbon Nanotubes, CRC Press, Boca Raton, FL, 1997. [5] M.S. Dresselhaus, G. Dresselhaus, P.C. Eklund, Science of Fullerenes and Carbon Nanotubes, Academic Press, San Diego, 1996. [6] P.M. Ajayan, T.W. Ebbesen, Rep. Prog. Phys. 60 (1997) 1025. [7] J. Chen, M.A. Hamon, H. Hu, Y. Chen, A.M. Rao, P.C. Eklund, R.C. Haddon, Science 282 (1998) 95. [8] F. Pompeo, D.E. Resasco, Nano Lett. 2 (2002) 369. [9] J.E. Riggs, Z. Guo, D.L. Carrol, Y.P. Sun, J. Am. Chem. Soc. 122 (2000) 5879. [10] Y.-P. Sun, K. Fu, Y. Lin, W. Huang, Acc. Chem. Res. 35 (2002) 1096. [11] D. Tasis, N. Tagmatarchis, V. Georgakilas, M. Prato, Chem. Eur. J. 9 (2003) 4000. [12] G. Viswanathan, N. Chakrapani, H. Yang, B. Wei, H. Chung, K. Cho, C.Y. Ryu, P.M. Ajayan, J. Am. Chem. Soc. 125 (2003) 9258. [13] J.I. Paredes, M. Burghard, Langmuir 20 (2004) 5149. [14] A. Star, D.W. Steuerman, J.R. Heath, J.F. Stoddart, Angew. Chem. Int. Ed. 41 (2002) 2508. [15] R.J. Chen, Y. Zhang, D. Wang, H.J. Dai, J. Am. Chem. Soc. 123 (2001) 3838.

[16] J. Chen, H. Liu, W.A. Weimer, M.D. Halls, D.H. Waldeck, G.C. Walker, J. Am. Chem. Soc. 124 (2002) 9034. [17] Y. Kang, A. Taton, J. Am. Chem. Soc. 125 (2003) 5650. [18] S. Bandow, A.M. Rao, K.A. Williams, A. Thess, R.E. Smalley, P.C. Eklund, J. Phys. Chem. B 101 (1997) 8839. [19] G.S. Duesberg, M. Burghard, J. Muster, G. Philipp, S. Roth, Chem. Commun. 3 (1998) 435. [20] M. Shim, N.W.S. Kam, R.J. Chen, Y. Li, H. Dai, Nano Lett. 2 (2002) 285. [21] X. Gong, S. Baskaran, R.D. Voise, J.S. Young, Chem. Mater. 12 (2000) 1049. [22] L. Jiang, L. Gao, J. Sun, J. Colloid Interface Sci. 260 (2003) 89. [23] M.J. OConnell, S.M. Bachilo, 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 297 (2002) 593. [24] P.J. Boul, J. Liu, E.T. Mikelson, C.B. Huffman, L.M. Ericson, I.W. Chiang, K.A. Smith, D.T. Colbert, R.H. Hauge, J.L. Margrave, R.E. Smalley, Chem. Phys. Lett. 310 (1999) 367. [25] H. Wang, E.K. Hobbie, Langmuir 19 (2003) 3091. [26] S. Niyogi, M.A. Hamon, D.E. Perea, C.B. Kang, B. Zhao, S.K. Pal, A.E. Wyant, M.E. Itkis, R.C. Haddon, J. Phys. Chem. B 107 (2003) 8799. [27] T. Kanyo, Z. Konya, F. Berger, I. Dekany, I. Kiricsi, Chem. Commun. 21 (2003) 2746. [28] M.R. Diehl, S.N. Yaliraki, R.A. Beckman, M. Barahona, J.R. Heath, Angew. Chem. Int. Ed. 41 (2002) 353. [29] C.M. Hansen, J. Paint Technol. 39 (1967) 104. [30] F.M. Barton, CRC Handbook of Solubility Parameters and Other Cohesion Parameters, second ed., CRC Press, Boca Raton, FL, 1991. [31] J. Brandrup, E.H. Immergut, E.A. Grulke, Polymer Handbook, fourth ed., Wiley, New York, 1999. [32] D.L. Ho, C.J. Glinka, Chem. Mater. 15 (2003) 1309. [33] Y.S. Choi, H.T. Ham, I.J. Chung, Chem. Mater. 16 (2004) 2522. [34] J.J. Cernohous, C.W. Macosko, T.R. Hoye, Macromolecules 31 (1998) 3759. [35] J.T. Davies, Proc. Intern. Congr. Surface Activity 1 (1957) 426. [36] J. Piirma, Polymeric Surfactants, Dekker, New York, 1992. [37] T. Davies, E.K. Rideal, Interfacial Phenomena, second ed., Academic Press, New York, 1963. [38] T. Ban, T. Sato, T.F. Yen, J. Petrol. Sci. Eng. 21 (1998) 223. [39] C.M. Hansen, Ind. Eng. Chem. Prod. Res. Dev. 8 (1969) 2. [40] H. Hildebrand, R.L. Scott, The Solubility of Nonelectrolytes, Reinhold, New York, 1950.