Journal of Colloid and Interface Science 328 (2008) 421–428

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Comparative study of carbon nanotube dispersion using surfactants

Richa Rastogi a,∗, Rahul Kaushal a, S.K. Tripathi b, Amit L. Sharma a, Inderpreet Kaur a, Lalit M. Bharadwaj a

a Biomolecular Electronics and Nanotechnology Division (BEND), Central Scientific Instruments Organisation (CSIO), Sector 30C, Chandigarh, Indiab Centre of Advanced Studies in Physics, Punjab University, Sector 14, Chandigarh, India

a r t i c l e i n f o a b s t r a c t

Article history:Received 8 April 2008Accepted 1 September 2008Available online 10 October 2008

Keywords:DispersionAbsorbanceMultiwalled nanotubesSurfactants

Dispersion of carbon nanotubes (CNTs) is a challenging task for their utilization in nanoscale device ap-plications. This account reports a comparative analysis on dispersion of multiwalled carbon nanotubes(MWNTs) with four surfactants—Triton X-100, Tween 20, Tween 80, and sodium dodecyl sulfate (SDS).Among the four surfactants, Triton X-100 and SDS provide maximum and minimum dispersion, respec-tively. Dispersion of MWNTs has been characterized with UV–vis spectroscopy and transmission electronmicroscopy (TEM). TEM results are in agreement with the UV–vis measurements. The experimentally ob-served trend of dispersing power of surfactants is consistent with their chemical structures. An optimumCNT-to-surfactant ratio has been determined for each surfactant. This parameter is shown to affect thenanotube dispersion significantly. Surfactant concentration above or below this ratio is shown to deteri-orate the quality of nanotube dispersion. TEM analysis of a high-surfactant-concentration sample enablesus to construct a plausible mechanism for decrease in CNT dispersion at high surfactant concentration,consistent with the UV–vis observations. Temperature stability of the surfactant is another importantfactor affecting the quality of CNT dispersion.

© 2008 Elsevier Inc. All rights reserved.

1. Introduction

Due to their tremendous optical, electrical, and mechanicalproperties, carbon nanotubes enjoy a preeminent status in thepanoply of nanomaterials, finding wide range of applications inbiosensors [1], composites [2], field emission devices [3], elec-tronic components [4], probe tips [5], etc. Delocalization of π -electrons renders them conducting and alleviates adsorption ofvarious chemical moieties via π–π stacking interaction [6]. A highaspect ratio makes them prone to entanglement and bundling. Par-ticularly, carbon nanotubes are bundled with strong van der Waalsinteraction energy of ca. 500 eV/μm of tube–tube contact [7]. Suchhigh interaction energy renders CNT dispersion a challenging task.

Currently two approaches are widely used in nanotube disper-sion—the mechanical approach and the chemical approach. Themechanical approach includes ultrasonication and high-shear mix-ing. These processes are time-consuming and less efficient. Fur-thermore, as reported earlier by Lu et al. [8], ultrasonication canresult in fragmentation of CNTs, in turn, decreasing their aspect ra-tio. Besides this, the stability of the dispersion is poor. On the otherhand, the chemical approach includes both covalent and noncova-lent methods. Covalent methods involve functionalization with var-ious chemical moieties to improve solubility in solvents. However,

* Corresponding author. Fax: +91 172 2657267.E-mail address: richa.bend@gmail.com (R. Rastogi).

0021-9797/$ – see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2008.09.015

aggressive chemical functionalization at high temperature createsdefects at the nanotube surface, consequently altering the electri-cal properties of carbon nanotubes [9]. In contrast, a noncovalentapproach involves adsorption of the chemical moieties onto thenanotube surface, either via π–π stacking interaction such as inDNA, uncharged surfactants, etc., or through coulomb attraction inthe case of charged chemical moieties. The noncovalent approachis superior in the sense that it does not alter the π -electron cloudof graphene, in turn preserving the electrical properties of carbonnanotubes.

Surfactants and polymers are extensively used for carbon nan-otube dispersion via noncovalent approach. Both of them get ad-sorbed onto nanotube surface, rendering them soluble in aqueousand organic solvents. Dispersion of nanotubes in polymer matri-ces may not be a proper choice for electronic device applications,as polymer itself can participate in electrical events [10]. Disper-sion using surfactants diminishes such anomalies, as they can beremoved easily by washing. To date, a wide variety of surfactantshave been investigated for dispersion of carbon nanotubes, suchas sodium dodecyl benzenesulfonate (SDBS) [11], dodecyltrimethyl-ammonium bromide (DTAB) [12], hexadecyltrimethylammoniumbromide (CTAB) [13], octyl phenol ethoxylate (Triton X-100) [14],and sodium dodecyl sulfate (SDS) [15]. In view of the large numberof surfactants available for dispersion, it is imperative to conducta systematic study of different parameters such as concentration,nature, stability, etc. in order to choose the right surfactant for aparticular application. Until recently, few attempts have been made

422 R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428

to compare dispersing power of different surfactants. Moore et al.[16] have done comparative analysis of various surfactants on thebasis of their spectral properties. Hertel et al. [17] and Yurekli et al.[18] have reported changes in phase behavior of carbon nanotubeson the basis of the nature, concentration, and type of interactionof surfactants. Still, there is a lack of systematic study to optimizedifferent parameters influencing dispersion of nanotubes.

The present study confronts a comparative analysis of foursurfactants—Triton X-100, SDS, Tween 20, and Tween 80 [19–22]—for nanotube dispersion. The dispersing power of the surfactantshas been analyzed experimentally as well as theoretically on thebasis of their structural organization. This study provides insightsinto some parameters for optimization of dispersion of MWNTsusing surfactants. UV–vis spectroscopy and transmission electronmicroscopy (TEM) have been employed to analyze the dispersionability of these surfactants. The significance of using a particularratio of surfactants and MWNTs has been established for obtain-ing optimum dispersion, which may be cited as a relatively newfinding in this area of research. From our study, the optimum CNT-to-surfactant ratio turns out to be the most important parameterin nanotube dispersion. We have optimized this ratio for these sur-factants. Apart from the above, temperature stability of surfactantis also shown to influence nanotube dispersion.

2. Materials and method

2.1. Materials

Purified MWNTs (diameter 40–70 nm, length 2–5 μm, purity>95%) were procured from Amorphous & Nanostructure Ltd., USA.The surfactants—Triton X-100, Tween 20 (USB Corp., USA), Tween80, and sodium dodecyl sulfate (Bio Basic, USA)—were used as re-ceived. All solutions were prepared in 18 M� cm deionized water.

2.2. Preparation of dispersion of MWNTs in various surfactants

In order to compare the dispersing power of the four sur-factants, dispersions of MWNTs were prepared at concentrationsspanning from 15 to 50 mg/L in steps of 5 mg/L, keeping theconcentration of surfactant (1%) constant. These 32 samples wereultrasonicated for 2 h in order to get surfactant-coated MWNTs.Dispersions were analyzed with UV–vis spectroscopy and the max-imum extractable concentration of MWNTs (at 1% surfactant con-centration) was determined for each of the surfactants.

In order to find the optimum CNT-to-surfactant ratio for eachsurfactant, a second set of experiments were carried out. In theseexperiments, concentration of surfactants was varied from 1.1 to1.9% in steps of 0.1%, keeping the amount of MWNTs constant.Constant MWNT concentrations chosen in these experiments werethe maximum extractable concentrations of MWNTs (determinedin the first set of experiments for 1% surfactant concentration).Again, these samples were analyzed using UV–vis spectroscopy.

2.3. Characterization of dispersion

2.3.1. UV–vis spectroscopyThe dispersions of MWNTs in surfactants were characterized us-

ing UV–vis spectrophotometer (Hitachi U-2800) operating betweenthe ranges of 300–1100 nm. In the first set of experiments, baselinecorrection was carried out using pure 1% solutions of the four sur-factants so that their absorbance values got subtracted from thatof MWNTs dispersions. In the second set of experiments, the base-line was equilibrated every time a new sample was analyzed withcorresponding surfactant concentration.

2.3.2. Transmission electron microscopyThe dispersions of MWNTs were also characterized using a

transmission electron microscope (Hitachi H-7500) operating at80 kV. MWNTs dispersed in surfactants were filtered, dried, andredispersed in distilled water at a concentration of 0.01 mg/ml.A drop of 15 μl of the above suspension was dropped onto carbon-coated TEM grids (300 mesh, 3 mm, purchased from TAAB Labora-tories, England) and viewed under a transmission electron micro-scope.

3. Results and discussion

3.1. Comparison of dispersing power of surfactants using UV–visspectroscopy

Bundled carbon nanotubes are not active in the UV–vis region[23,24]. Only individual carbon nanotubes absorb in this region.Therefore, dispersion of carbon nanotubes can be characterized us-ing UV–vis absorption spectroscopy. To characterize the dispersionof MWNTs in surfactants using UV–vis spectroscopy, absorbancevalues were recorded at 500 nm as reported in previous studies[25–29]. This wavelength is virtually unaffected by ambient con-ditions of nanotubes. The Lambert–Beer law is well obeyed byMWNTs at this wavelength. Concentration of MWNTs dissolved ordispersed into the solution can then be determined using the spe-cific extinction coefficient of carbon nanotubes at 500 nm, ε500 =28.6 cm2 mg−1 [29], in the Lambert–Beer law. With this knowl-edge, percentage recovery into the solution can be calculated as

% extractability = c1

c× 100, (1)

where c1 = concentration of MWNTs recovered in solution andc = concentration of MWNTs originally taken in surfactant. Thisparameter is the measure of dispersion of carbon nanotubes in so-lution.

Fig. 1 depicts the UV–vis spectra of MWNTs with varying con-centrations of nanotubes in surfactant solutions. The inset showsthe Lambert–Beer dependence of absorption at 500 nm on theconcentration of surfactant. Background absorption in spectra isdue to the presence of bundles and mats of CNTs [13]. In orderto compare the dispersing power of surfactants in context, per-centage extractability was calculated at different concentrations ofnanotubes (Table S1 of the supplementary material).

In all surfactants, percentage extractability vs concentration fol-lows a Gaussian trend (Fig. 2), which depicts a linear increasewith increase in concentration of MWNTs until the maximum ex-tractability limit is achieved. This is presumably because at lowCNT concentrations, the amount of surfactant is sufficient to coatthe carbon nanotube surface evenly. Eventually, a concentrationvalue is attained for which the surfactant amount is just sufficientto disperse the carbon nanotubes; i.e., the maximum extractabil-ity limit is achieved at this point. For subsequent increases inthe concentration of nanotubes, the surfactant amount turns in-sufficient to fully disperse the agglomerates of CNTs, therefore de-creasing the percentage extractability at high concentrations. Themaximum amount of MWNTs is extracted in the case of Triton X-100, where percentage extractability has gone as high as 90.03%.It is estimated to be 89.98, 84.61, and 84.89% for Tween 80, SDS,and Tween 20, respectively. It is noteworthy that maximum ex-tractability limit is attained earlier for SDS in comparison to othersurfactants. Maximum extractable MWNT concentration is foundto be 25, 30, 30, and 40 mg/L for SDS, Tween 20, Tween 80, andTriton X-100, respectively, for the constant surfactant concentra-tion (1%). Thus, the same amount of Triton X-100 can disperselarge amounts of MWNTs as compared to other surfactants. Al-though Tween 20 and Tween 80 show equivalent behavior in terms

R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428 423

Fig. 1. UV–vis spectra of carbon nanotubes in (a) SDS, (b) Tween 80, (c) Tween 20, and (d) Triton X-100 (Beer–Lambert curves inset).

of maximum extractable MWNTs concentration, Tween 80 provesto be better than Tween 20 in terms of percentage extractability.Hence, according to our experimental results, the dispersing powerof the four surfactants follows the trend

SDS < Tween 20 < Tween 80 < Triton X-100−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−→Dispersing power

This experimentally observed trend for dispersing power of surfac-tants can also be explained on the basis of their chemical struc-tures (Fig. S1 of the supplementary material). In order to dispersenanotubes in water, surfactant molecules orient themselves in sucha fashion that hydrophobic tail groups face toward the nanotubesurface while hydrophilic head groups face toward the aqueousphase, producing a lowering of the nanotube/water interfacial ten-sion. Thus, the dispersing power of the surfactant depends onhow firmly it adsorbs onto the nanotube surface and producesby this adsorption energy barriers of sufficient height to aggre-gation. Molecules having the benzene ring structure adsorb morestrongly to the graphitic surface due to π–π stacking type interac-tion [11,30]. Generally, hydrophobic tail groups tend to lie flat onthe graphitic surface because graphitic unit cells match well withthe methylene units of hydrocarbon chains [31]. Thus, efficiency of

adsorption and consequently dispersing power of surfactants aregreatly affected by the tail length of the surfactant. Longer tailsmeans high spatial volume and more steric hindrance, thus provid-ing greater repulsive forces between individual carbon nanotubes[32]. Besides this, surfactants with unsaturated bonds in their tailgroups contribute more toward nanotube dispersion [32].

As one can see, Tween 80 has the greatest hydrocarbon taillength, while Triton X-100 has the smallest one (Fig. S1 of thesupplementary material). The phenyl ring has an effective lengthof about three and one-half carbon atoms, while carbon atomson branches contribute about one-half the effect of carbon atomson straight alkyl chains. Thus, Triton X-100 has an effective chainlength of nine atoms only, which is the shortest out of the four.Thus, theoretically Triton X-100 should exhibit minimum dispers-ing power, contrary to experimental observations. Such a paradigmdeparture from experimental observations is presumably becauseof the presence of benzene ring in the tail group of Triton X-100.This finding prompts us to conclude that whenever “tail lengthfactor” and “benzene ring factor” compete, the latter contributesmore to dispersion of CNTs. In total, Triton X-100 proves to bethe best among the four surfactants. Among the remaining threesurfactants, Tween 80 has the longest tail. In addition, it has oneC–C unsaturated bond in its tail. Thus, theoretically, it is a better

424 R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428

Fig. 1. (continued)

dispersant than Tween 20 and SDS. Indeed, this is what has beenobserved experimentally. Among the remaining two, the tail of SDShas an extra carbon atom as compared to Tween 20, but the headgroup of Tween 20 is bulkier than that of SDS. Bulkier head groupscause larger steric stabilization, in turn, producing steric energybarriers to aggregation in nonionic surfactants such as Tween 20[33]. It is to be observed that these two factors are competing al-most equally when compared with the experimental results. Thereis just a slight difference between maximum extractability of SDSand Tween 20. This is presumably due to compensation for theeffects of extra carbon atom in SDS by bulkier head group inTween 20. Overall, therefore, structural considerations also supportour experimentally established trend of dispersing power amongthese surfactants.

3.2. Determination of optimum CNT to surfactant ratio using UV–visspectroscopy

In order to determine the optimum CNT-to-surfactant ratiofor each surfactant, a second set of experiments was carried outin which the concentration of surfactants was varied while theMWNT concentration was kept constant. Constant MWNT concen-

trations were selected to be 25, 30, 30, and 40 mg/L for SDS,Tween 20, Tween 80, and Triton X-100, respectively (chosen fromthe first set of experiments). Again, absorbance values were de-termined at 500 nm and percentage extractability of MWNTs wascalculated for each sample (Table S2 of the supplementary mate-rial). There appears to be a polynomial correlation between per-centage extractability and the concentration of surfactant (Fig. 3).Percentage extractability attains maxima at 1.3 (Triton X-100), 1.2(Tween 80), 1.4 (Tween 20), and 1.3% (SDS) surfactant concentra-tions. With this knowledge, optimum CNT-to-surfactant ratio wascalculated to be 1:466.66, 1:520, 1:400, and 1:350 for Tween 20,SDS, Tween 80, and Triton X-100, respectively. Such a large amountof surfactant required with respect to a small amount of MWNTscan be attributed to the large surface-to-weight ratio of carbonnanotubes.

Increase in percentage extractability with increase in surfactantconcentration was on the expected lines. However, the reason be-hind the decrease in percentage extractability after attaining max-ima was not clear. This problem was addressed by TEM observationof a high-concentration surfactant sample (Fig. 4). Interestingly,flocculation of CNTs was observed at some places. Individual CNTswere coated with thick nonuniform surfactant coatings. Chunks of

R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428 425

Fig. 2. Percentage extractability vs concentration trend of carbon nanotubes for (a) Tween 20, (b) SDS, (c) Tween 80, and (d) Triton X-100.

CNTs were aggregated via surfactant molecules. This observationcan be explained in the light of the theory of micelle formationin surfactants. Fig. 5 depicts a schematic illustration of a plausi-ble mechanism of flocculation of CNTs via surfactant molecules. Athigh concentrations, the surfactant molecules form micelles in so-lution. The size of these micelles keeps on increasing with increas-ing surfactant concentration due to interaction between groupsof the same polarity. Likewise, surfactant molecules form multi-layers on nanotube surface when the concentration of surfactantis increased for a constant nanotube concentration. As a conse-quence, surface coverage by surfactant molecules becomes so highthat portions of surfactant molecules extending into the liquidphase start interacting with others on neighboring CNTs [34]. Ifthe orientation of the outermost layer is such that the hydropho-bic groups of surfactant molecules are forced to extend into theaqueous phase, then this interaction favors a reduction in theirsurface energies [35]. This bridging of CNTs via extra surfactantmolecules causes flocculation. Flocculation of CNTs might be thereason behind the decrease in percentage extractability, which inturn decreases the dispersion of nanotubes at high surfactant con-centration.

3.3. Morphology of CNT dispersions

In order to characterize dispersion ability of different sur-factants morphologically, bundle diameters were measured usingTEM. Larger bundle diameter implies less dispersion. Fig. 6 de-picts TEM micrographs of MWNTs at optimum CNT-to-surfactant

ratio in different surfactants. In SDS dispersion, ribbon-like ropes ofMWNTs were observed with bundle diameters ranging from 62.5to 218.75 nm. Organization of these ribbon-shaped ropes could notbe observed due to poor contrast. Small particles of diameter rang-ing between 70 and 100 nm were also observed. These are sug-gested to be SDS crystals and were present due to crystallizationof SDS at low ambient temperature (17 ◦C). Poor temperature sta-bility of SDS solution might also be the reason for poor dispersionof MWNTs in SDS. No entanglement of MWNTs was observed inthe cases of Tween 20, Triton X-100, and Tween 80. Well-isolatedMWNTs were observed in these surfactants. Bundle diameters ofMWNTs ranged from ∼19.52 to 28.20 nm for Tween 20, ∼5.4to 11.43 nm for Tween 80, and ∼4 to 5.35 nm for Triton X-100,respectively. This finding reveals the highest dispersing power inTriton X-100 and the minimum in SDS, consistent with UV–visobservations. Thickness of surfactant coating was observed to be∼6.4–8 nm for Tween 20, ∼3–4 nm for Tween 80, and ∼0.9–1.78nm for Triton X-100, respectively. Stability of Tween 20, Tween 80,and Triton X-100 at low ambient temperature might also be thereason for better dispersion in these surfactants in comparison toSDS.

4. Summary and conclusions

For a comparative study of dispersion parameters of MWNTs,four surfactants, namely Triton X-100, Tween 20, Tween 80, andSDS, were analyzed. The key motif of this study was to analyzethe parameters responsible for the dispersion ability of a surfac-

426 R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428

Fig. 3. Variation of percentage extractability with variation of concentration of surfactant for (a) SDS, (b) Triton X-100, (c) Tween 80, and (d) Tween 20.

Fig. 4. Transmission electron micrograph of a high-surfactant-concentration sample.

Fig. 5. Mechanism of flocculation of CNTs via surfactant molecules.

tant. With this objective in mind, UV–vis and TEM studies of thedispersion of MWNTs in the above surfactants were executed. Dis-persion of MWNTs in the four investigated surfactants shows thefollowing trend:

SDS < Tween 20 < Tween 80 < Triton X-100−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−→Dispersing power

The experimental trend was further correlated with the chem-ical structures of the surfactants. Our experimental investigationshows unambiguously that, contrary to the theoretical aspect, Tri-ton X-100 enjoys the highest dispersing power among four sur-factants by virtue of its benzene ring. This analysis enables us

R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428 427

Fig. 6. Transmission electron micrographs of carbon nanotube dispersion in (a) SDS, (b) Tween 80, (c) Tween 20, and (d) Triton X-100.

to conclude that the “benzene ring factor” enjoys reigning statusin comparison to the “tail length factor” in dispersion of MWNTsusing surfactants. Therefore, the dispersing power of different sur-factants is inherent, due to their chemical organization. The keyfinding of the present study is the significance of optimum CNT-to-surfactant ratio. The quality of nanotube dispersion deterioratesbelow or above this ratio. Thus, surfactants should be in concentra-tion just sufficient to coat the nanotube surface, avoiding any ex-cess, as an unnecessarily large amount of surfactant also decreasesthe nanotube dispersion. To ascertain the reason for this decreasein the dispersion, a TEM study was employed. The TEM studypermits the elucidation of a plausible mechanism indicating de-creasing dispersion at high surfactant loading to be caused by theflocculation of CNTs with undesired excess surfactant molecules.Furthermore, a low degree of dispersion of CNTs at low concentra-tions of surfactant is attributed to an insufficient reagent amount.We have standardized this ratio for the surfactants in context. Itwas found to be 1:466.66, 1:520, 1:400, and 1:350 for Tween 20,SDS, Tween 80, and Triton X-100, respectively. Nanotube dispersionwas found to be highest at these ratios. Moreover, thermal stabil-ity of surfactant solution also plays an important role in dispersionof MWNTs. Surfactants such as SDS disperse the nanotubes poorlydue to instability of the solution at low temperature. In total, onecan conclude that proper choice of a suitable surfactant needs con-sideration of its structure, its optimum ratio to nanotubes, andthermal stability of its solution.

Acknowledgments

The authors thank the Department of Information and Technol-ogy (DIT), Ministry of Science and Technology, India, for financial

assistance. The authors are also thankful to the Director, CentralScientific Instruments Organization (CSIO), Chandigarh, for provid-ing necessary infrastructure and support and to Mr. M.L. Sharmaof the Regional Sophisticated Instrumentation Center (RSIC), PanjabUniversity, Chandigarh for transmission electron microscope facil-ity.

Supplementary material

The online version of this article contains additional supple-mentary material.

Please visit DOI: 10.1016/j.jcis.2008.09.015.

References

[1] X. Tang, S. Bansaruntip, N. Nakayama, Y. Erhan, Y.L. Chang, Q. Wang, NanoLett. 6 (2006) 1632.

[2] L. Clayton, T. Gerasimov, M. Meyyappan, J.P. Harmon, Adv. Funct. Mater. 15(2005) 101.

[3] J. Koohsorkhi, Y. Abdi, S. Mohajerzadeh, H. Hosseinzadegan, Y. Komijani, E.A.Soleimani, Carbon 44 (2006) 2797.

[4] E. Frackowiak, F. Beguin, Carbon 40 (2002) 1775.[5] H.J. Dai, J.H. Hafner, A.G. Rinzler, D.T. Colbert, R.E. Smalley, Nature 384 (1996)

147.[6] M. Terrones, Annu. Rev. Mater. Res. 33 (2003) 419.[7] L.A. Girifalco, M. Hodak, R.S. Lee, Phys. Rev. B 62 (2000) 13104.[8] K.L. Lu, R.M. Lago, Y.K. Chen, M.L.H. Green, P.J.F. Harris, S.C. Tsang, Carbon 34

(1996) 814.[9] J. Hilding, E.A. Grulke, Z.G. Zhang, F. Lockwood, J. Dispers. Sci. Technol. 24

(2003) 1.[10] T. Blythe, D. Bloor, Electrical Properties of Polymers, Cambridge University

Press, London, 2005.[11] M.F. Islam, E. Rojas, D.M. Bergey, A.T. Johnson, A.G. Yodh, Nano Lett. 3 (2003)

269.

428 R. Rastogi et al. / Journal of Colloid and Interface Science 328 (2008) 421–428

[12] E.A. Whitsitt, A.R. Barron, Nano Lett. 3 (2003) 775.[13] A.G. Ryabenko, T.V. Dorofeeva, G.I. Zvereva, Carbon 42 (2004) 1523.[14] H. Wang, W. Zhou, D.L. Ho, K.I. Winey, J.E. Fischer, C.J. Glinka, E.K. Hobbie, Nano

Lett. 4 (2004) 1789.[15] J. Yu, N. Grossiord, C.E. Koning, J. Loos, Carbon 45 (2007) 618.[16] V.C. Moore, M.S. Strano, E.H. Haroz, R.H. Hauge, R.E. Smalley, Nano Lett. 3

(2003) 1379.[17] T. Hertel, A. Hagen, V. Talalaev, K. Arnold, F. Hennrich, M. Kappes, Nano Lett. 5

(2005) 511.[18] K. Yurekli, C.A. Mitchell, R. Krishnamootri, J. Am. Chem. Soc. 126 (2004) 9902.[19] J. Liu, A.G. Rinzler, H.J. Dai, J.H. Hafner, R.K. Bradley, P.J. Boul, A. Lu, T. Iver-

son, K. Shelimov, C.B. Huffman, F.J. Rodriguez-Marcias, Y.S. Shon, T.R. Lee, D.T.Colbert, R.E. Smalley, Science 280 (1998) 1253.

[20] M.J. O’Connell, S.M. Bachilo, C.B. Huffman, V.C. Moore, M.S. Strano, E.H. Haroz,Science 297 (2002) 593.

[21] L. Jiang, L. Gao, J. Sun, J. Colloid Interface Sci. 260 (2003) 89.[22] W. Wenseleers, I.I. Vlasov, E. Goovaerts, E.D. Obraztsova, A.S. Lobach, A.

Bouwen, Adv. Funct. Mater. 14 (2004) 1105.[23] Y. Junrong, N. Grossiord, C.E. Koning, J. Loos, Carbon 45 (2007) 618.[24] J.S. Laurent, C. Voisin, G. Cassabois, C. Delalande, P. Roussignol, O. Jost, L. Capes,

Phys. Rev. Lett. 90 (2003) 057404.

[25] W. Huang, S. Taylor, K. Fu, Y. Lin, D. Zhang, T.W. Hanks, A.M. Rao, Y.P. Sun, NanoLett. 2 (2002) 311.

[26] K.D. Ausman, R. Piner, O. Lourie, R.S. Ruoff, M. Korobov, J. Phys. Chem. B 104(2000) 8911.

[27] O.K. Kim, J. Je, J.W. Baldwin, S. Kooi, P.E. Pehrsson, L.J. Buckley, J. Am. Chem.Soc. 125 (2003) 4426.

[28] V.A. Sinani, M.K. Gheith, A.A. Yaroslavov, A.A. Rakhnyanskaya, K. Sun, A.A.Mamedov, J.P. Wicksted, N.A. Kotov, J. Am. Chem. Soc. 127 (2005) 3463.

[29] A. Ikeda, T. Hamano, K. Hayashi, J. Kikuchi, Org. Lett. 8 (2006) 1153.[30] J.F. Liu, W.A. Ducker, Langmuir 16 (2000) 3467.[31] D.M. Cyr, B. Venkataraman, G.W. Flynn, Chem. Mater. 8 (1996) 1600.[32] D.H. Napper, Polymeric Stabilization of Colloidal Dispersion, Academic Press,

London, 1983.[33] V.C. Moore, M.S. Strano, E.H. Haroz, R.H. Hauge, R.E. Smalley, Nano Lett. 3

(2003) 1379.[34] M.J. Rosen, Surfactants and Interfacial Phenomena, Wiley–Interscience, New

York, 1978.[35] P. Somasundaran, T.W. Healy, D.W. Fuerstenau, J. Colloid Interface Sci. 22 (1966)

599.