Stable Solutions of Multiwalled Carbon Nanotubes Using an

J. Phys. Chem. C 2010, 114, 14347–14352

14347

Stable Solutions of Multiwalled Carbon Nanotubes Using an Azobenzene Dye Guadalupe Dı´az Costanzo,† Silvia Ledesma,†,‡ In˜aki Mondragon,§ and Silvia Goyanes*,‡,| LPI, Department de Fı´sica, LP&MC, Departamento de Física and Departamento de Ing. Química FCEyN-UBA, Pab. 1, Ciudad UniVersitaria 1428 Bs. As., Argentina, CONICET, Argentina, Department Ing. Quı´mica y M. Ambiente, Escuela Polite´cnica, UniVersity of the Basque Country, Pza. Europa 1. 20018 Donostia-San Sebastia´n, Spain ReceiVed: March 22, 2010; ReVised Manuscript ReceiVed: July 23, 2010

We present a method to achieve stable dispersions of multiwalled carbon nanotubes (MWCNT) in tetrahydrofuran for several days. The physical chemistry of this method is based on the π-π stacking interactions between MWCNT and an azobenzene derivative, Disperse Orange 3 (DO3). Evidence of this interaction was obtained using UV-vis absorption spectroscopy. Our results suggest that there is a weight ratio DO3/MWCNT beyond which interactions are strong enough to achieve stable dispersions of MWCNT for at least 45 days. 1. Introduction As it is well-known, carbon nanotubes (CNT) have unique one-dimensional structure and electronic properties. Among other possibilities, there is a great interest focused on their capability to reinforce composite materials and their potential use in the development of electronic and optical devices. The excellent mechanical and electrical properties of CNT are partially lost when CNT aggregate. The fact that CNT tend to aggregate makes difficult to put them into a matrix. Until now, achieving stable dispersions of CNT is still a challenge. In this sense, there have been several attempts of functionalizing CNT to improve their dispersion into polymer matrices. One of the covalent modifications includes reaction with diazonium salts.1 The covalent functionalization has been reported to damage their electronic properties due to the loss in the one-dimensional electronic features, which is a clear disadvantage for future applications. Indeed, the noncovalent functionalization is preferred to improve CNT solubility due to the weaker modification of the intrinsic properties of carbon nanotubes.2 Noncovalent functionalization includes polymer wrapping,3,4 the use of biocompatible surfactants,5 the encapsulation by supramolecular systems of small molecules, like surfactant micelles,6 and the π-stacking by rigid, conjugated macromolecules. These attempts with noncovalent functionalization have their basis in the sp2 molecular structure of carbon nanotubes, which in fact can be seen as an extended π-system. This outstanding feature leads to the possibility of a noncovalent attachment to other molecular π-systems. In fact, in a theoretical frame molecules with aromatic rings such as benzene or anthracene can interact with the sidewalls of carbon nanotubes7 via π-π stacking interactions. In these interactions, both π-electron systems of both systems are combined leading to minor changes in the electronic structure of the nanotubes.2 In the past few years, several works studied the dispersion of carbon nanotubes. Ji et al.8 focused their attention in the * To whom correspondence should be addressed. Fax: +54 11 45763357. E-mail address: [email protected]. † LPI, Department de Fı´sica, Ciudad Universitaria. ‡ CONICET. § University of the Basque Country. | LP&MC, Department de Fı´sica, Ciudad Universitaria.

dispersion of MWCNT in a generic nonpolar solvent with the aim of having the nanotubes dispersed in a silicone matrix. They found that a pyrene surfactant leads to stable dispersions of MWCNT in petroleum ether with a 10 mg/mL solubility. Iamsamai et al.9 used the noncovalent functionalization of MWCNT with chitosan improving their dispersion in aqueous solutions. Particularly, it has been demonstrated that a diazo dye molecule, Congo Red, efficiently improves the dispersion of MWCNT in water10 without damaging the molecular characteristics of carbon nanotubes and thus preserving their properties. To go further in the understanding of solubility of carbon nanotubes, Zhang et al.11 recently studied the effect of methyl substitutes of organic dyes in the dispersion of MWCNT, and Ham et al.12 related Hansen solubility parameters of different solvents with molar volumes of the surfactants groups with the dispersion of single-walled carbon nanotubes. In other respects, one of the applications of carbon nanotubes is to introduce them in a polymer.13-16 Carbon nanotubes are introduced in polymers mainly to improve electric, mechanical, or thermal behavior. In particular, in epoxy matrices MWCNT are used to obtain certain degree of electrical conductivity without modifying their mechanical properties. The most frequently used solvent in epoxy resins, which are not watersoluble, is tethrahydrofuran (THF). For this reason, it is relevant to find proper mechanisms to disperse carbon nanotubes in THF. Among many others, one dye molecule that has benzene rings is the Disperse Orange 3 (DO3). DO3 is an azobenzene derivative with an amine and a nitro group that is highly soluble in THF and that could interact with carbon nanotubes via π-π stacking interactions, thus helping to achieve stable dispersions in THF. In addition, the fact that DO3 has a terminal amine group could be helpful for a future covalent bonding with an epoxy resin. Moreover, one of the main characteristics of azobenzene, as well as its derivatives, is that they can undergo a photoisomerization process. In this sense, the addition of carbon nanotubes with DO3 to polymers could lead to potential applications in the development of optical data storage devices. Taking into account this kind of application, in this work we pay our attention only to the MWCNT dispersion and interaction with DO3 in THF to further effectively disperse the nanotubes in epoxy resins without focusing on the possible variations in

10.1021/jp102580b  2010 American Chemical Society Published on Web 08/06/2010

14348

J. Phys. Chem. C, Vol. 114, No. 34, 2010

Figure 1. Molecular structure of Disperse Orange 3.

thermal or electrical conductivity. However, this could be an interesting task to study in order to develop devices with different applications. In this work, we show that the addition of DO3 to dispersions of MWCNT in THF in specific quantities efficiently helps to debundle MWCNT making dispersions stable for several days. We report optical absorption experiments that suggest an interaction between MWCNT and DO3 molecules that would explain the stability observed in the suspensions observed. 2. Experimental Methods In our study, we worked with commercial MWCNT from NANOCYL (NC3100). Their length is about 1.5 µm and their diameter is around 20-40 nm as observed by field emission scanning electron microscopy. MWCNT were dried in a vacuum oven at 120° C during 3 h (to remove adsorbed water). MWCNT were left in vacuum at room temperature until used. As a dye, we use DO3, which molecular structure is shown in Figure 1. It has the appearance of orange powder and was used as received from Sigma-Aldrich. The employed solvent was tetrahydrofuran from Sintorgan (HPLC analytical grade) in all cases. In this study, three different kinds of samples were prepared: (1) samples in which only DO3 was dissolved in THF at different concentrations (named, reference samples); (2) samples in which a fixed amount, 6 µg/ mL, of MWCNT was added to each reference sample; and (3) samples in which a different fixed amount, 10 µg/mL, of MWCNT was added to each reference sample. The concentrations of DO3 for the reference samples were 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 µg/mL. All the obtained results for the samples containing MWCNT and DO3 were compared to the corresponding reference sample. All samples containing MWCNT were sonicated during four periods of 15 min each one. Three different sets of samples (reference samples and samples with MWCNT added) were prepared to ensure the results were repetitive. Stability of the dispersion in samples with MWCNT was qualitatively evaluated by capturing images at different times. Images were taken with a commercial Sony DSC-W120 camera.

Dı´az Costanzo et al. Interactions between MWCNT and DO3 were studied by comparing the UV visible spectra of reference samples to those containing MWCNT and DO3. UV-vis spectra were measured in all samples by using a HP 8453 spectrometer (wavelength resolution: 1 nm). All the absorption spectra were obtained at room temperature. For the three kinds of samples, three different set of samples were prepared to ensure the results were repetitive. The effects of DO3 on the MWCNT morphology were studied with a field emission scanning electron microscopy (FE-SEM Zeiss LEO 982 GEMINI). 3. Results and Discussion We qualitatively followed the stability of the dispersions containing 6 and 10 µg/mL MWCNT and DO3 at different concentrations for several days. From direct observation, two different behaviors were observed when progressively increasing DO3 concentration to MWCNT in THF: one for the lower DO3 concentrations and other for the higher DO3 concentrations. All these samples were followed for at least 45 days. For low DO3 concentrations, MWCNT still aggregated but in a different way as they did in THF. For high DO3 concentrations, there was no MWCNT aggregation. A remarkable fact arised for DO3 concentrations higher than 4 µg/mL since almost no aggregation was observed for more than 45 days. Effective dispersion of MWCNT would suggest that MWCNT would be interacting with a large number of DO3 molecules, this fact leading to a more efficient dispersion. This effect could be attributed to π-π stacking interactions between MWCNT and DO3 molecules. To exemplify these behaviors, we show in Figure 2 a picture taken after 45 days of the preparation of the samples. A zoom of the area of interest is also shown in order to clarify the extent of aggregation. It is worth to note that even for a low DO3 concentration, 3 µg/mL, there was a significant change in the way MWCNT started to aggregate. When MWCNT were in THF, they created very large bundles that decanted within the first hours. On the contrary, when DO3 was added in low concentrations, MWCNT aggregation still happened but in an absolutely different way as the nanotubes did aggregate and finally decant but in smaller aggregates as if they were little rocks. This fact suggests the existence of interactions between DO3 molecules and MWCNT. Similar results were evidenced when samples with MWCNT were settled for days. MWCNT started to aggregate within the first hours when dispersed in THF and also a certain degree of aggregation was noticed for

Figure 2. Samples with 6 µg/mL of MWCNT. From left to right: MWCNT in THF only, MWCNT in THF and DO3, 3 µg/mL, and MWCNT in THF and DO3, 10 µg/mL.

Stable Solutions of MWCNT Using an Azobenzene Dye

Figure 3. Absorbance spectra of reference samples. Numbers against each curve refer to the DO3 concentration of the sample. From bottom to top, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 µg/mL.

the lower DO3 concentrations (from 1 to 5 µg/mL). The nonaggregation was only achieved when DO3 concentration was higher than 6 µg/mL. The UV-vis absorbance spectra of the reference samples for all the different concentrations are shown in Figure 3. There are two absorption bands, one at 276 nm and the other at 443 nm. In the UV-vis energy range, DO3 should have three electronic transitions, the one at low wavelengths, a π-π* electronic transition that corresponds to the benzene ring, and two transitions, π-π* and n-π*, associated to the azo group. But for the kind of azo dyes used in this work, because of the terminal substituents like the amino or the nitro group, the strong π-π* electronic transition overlaps the n-π* transition. Thus, two absorption bands did appear though there were three electronic transitions. The band at 443 nm corresponds to the stronger π-π* electronic transition of the azo group and it is associated to the trans isomer. With the aim of analyzing how the interactions between MWCNT and azo dye molecules did influence the UV-vis spectrum of DO3, in principle we took into account both possibilities: either MWCNT interact with DO3 molecules via the azo group, or MWCNT stack via the benzene ring by π-π

J. Phys. Chem. C, Vol. 114, No. 34, 2010 14349 stacking interaction. However, it is clear that the latter would be a stronger interaction. This means that the absorption band of DO3 around 276 nm associated to the benzene ring should have more noticeable changes. In Figure 4a,b the absorbance spectra of samples containing 6 and 10 µg/mL MWCNT, respectively, are shown. From bottom to top, the curves corresponding to 1, 2, 3, 4, 5, 6, 7, 8, 10, and 12 µg/mL of DO3 for samples with 6 µg/mL MWCNT (Figure 3a) and 1, 2, 4, 5, 6, 7, 8, 9, 10, 11, and 12 µg/mL of DO3 for samples with 10 µg/mL MWCNT (Figure 3b) are presented. The insets show a typical absorption spectrum of MWCNT in THF at the same MWCNT concentrations (6 and 10 µg/mL, respectively). As can be seen, no absorption band appeared in the UV-vis range for MWCNT in THF. Recently, Lidorikis and co-workers17 reported the energy values of the π-plasmon absorption of MWCNT of varying diameters. They found π-plasmon absorption of MWCNT to be centered around 4.5 eV, UV range. The fact that no absorption band was observed in the MWCNT spectra is a consequence of their aggregation of the MWCNT. Taking into account that larger surface area of MWCNT leads to higher absorbance, it would be expected that a more efficient dispersion will produce an increment in absorption. In Figure 5, the maxima of the absorbance bands at 443 nm for the different kinds of samples are plotted against DO3 concentration. Absorbance of samples with MWCNT was corrected subtraction of MWCNT absorbance in all cases. Absorbance maxima at 443 nm of the reference samples grew linearly with concentration, thus indicating it followed Beer’s law. Also for samples with MWCNT, no significant changes did appear in the maxima at 443 nm though there were some differences for the higher DO3 concentrations tested. This could be attributed to the appearance of the characteristic absorption bands of carbon nanotubes in this wavelength range.18,19 These bands, that are present only when the nanotubes are dispersed, could eventually make absorbance stronger for the higher DO3 concentrations. As can be seen below, the concentrations for which there is a subtle difference in absorbance around 443 nm are the same concentrations for which dispersion by DO3 is effective. In Figure 6 we show the absorbance at 276 nm for the reference samples and those containing MWCNT at the two fixed concentrations [(a) 6 µg/mL and (b) 10 µg/mL]. In this

Figure 4. UV-vis absorption spectra of the two different set of samples with fixed MWCNT concentration: (a) 6 µg/mL and (b) 10 µg/mL and different DO3 concentrations. Numbers against each curve refer to the DO3 concentration in µg/mL of the sample. Insets in (a) and (b): MWCNT absorption background in THF at 6 µg/mL and 10 µg/mL respectively.

14350

J. Phys. Chem. C, Vol. 114, No. 34, 2010

Dı´az Costanzo et al.

Figure 5. Absorbance at 443 nm for the reference samples and samples with MWCNT at 6 and 10 µg/mL (MWCNT samples are corrected by MWCNT absorbance).

Figure 7. Samples of MWCNT 48 h after of being sonicated: (a) MWCNT, 0.1 mg/mL, in THF only; (b) MWCNT, 0.1 mg/mL, with DO3 in a weight ratio DO3/MWCNT ) 2; (c) MWCNT, 1 mg/mL, with DO3 in a weight ratio DO3/MWCNT ) 2.

case, absorbance of samples with MWCNT was once again corrected by subtraction of the MWCNT absorbance. For the absorbance band at 276 nm the influence of the MWCNT was clearly evident as this band was higher than that for the reference samples in both cases. It is important to note, that also for the two fixed MWCNT concentrations studied, the results in absorbance at 276 nm are similar and have two characteristic features. First, there exists a fixed DO3 concentration beyond which absorbance at 276 nm increases keeping a high linear correlation, and second, the displacement in absorbance is corresponds to a certain quantity greater than the statistical error in both cases. In samples with MWCNT, 6 µg/mL, for DO3 concentrations under 5 µg/mL there almost did not exist differences among the absorbance of the reference samples and the samples with MWCNT whereas for DO3 concentrations equal or beyond 5 µg/mL the differences in absorbance were around 10%. This value is beyond the 3% considered by the statistical dispersion of the different set of measurements of the reference samples. This effect was even more evident for samples in which MWCNT concentration was fixed at 10 µg/mL. In this case,

this is especially clear for DO3 concentrations higher than 6 µg/mL. It is also worth noting that the DO3 concentration in which there existed a larger difference in absorbance with respect to the reference samples depended on the MWCNT concentration. For samples in which DO3 concentration was higher than 5 µg/mL, differences in absorbance were beyond 30%. Only to make it clearer, in Figure 6 there is a line joining each set of experimental data. To further analyze the behaviors observed in the solutions of MWCNT with DO3 in THF, we studied the stability of solutions with higher MWCNT concentration to see whether the effect of DO3 could be extended to higher concentrations. In agreement with our previous observations, we found that stable solutions of MWCNT and DO3 are obtained above certain weight ratios of DO3/MWCNT. These observations are exemplified in Figure 7 where we show three samples after 48 h of being sonicated. In Figure 7, from left to right: (a) MWCNT 0.1 mg/mL in THF; (b) MWCNT, 0.1 mg/mL, DO3/MWCNT ) 2; (c) MWCNT, 1 mg/mL, DO3/MWCNT ) 2. When MWCNT are in THF only, the stability disappears within the first hours. On the contrary, Figure 7b,c shows the stability of

Figure 6. Absorbance at 276 nm for the reference samples and (a) samples with MWCNT at 6 µg/mL and (b) 10 µg/mL (Absorbance of MWCNT samples is corrected by MWCNT absorbance).

Stable Solutions of MWCNT Using an Azobenzene Dye

J. Phys. Chem. C, Vol. 114, No. 34, 2010 14351

Figure 8. SEM analysis of two representative samples: (a,c) MWCNT in THF at a magnification of 100 000× and 300 000×, respectively, (b,d) MWCNT with DO3 in a DO3/MWCNT weight ratio equal 2 at a magnification of 100 000× and 300 000×, respectively.

the solutions when DO3 is added in the weight ratio DO3/ MWCNT ) 2. This in fact shows the efficiency of DO3 as a dispersing agent of MWCNT until MWCNT concentrations of 1 mg/mL. It is worth mentioning that having stable solutions of MWCNT, 1 mg/mL, supports the possibility of using DO3 as dispersing agent for MWCNT in the development of new nanomaterials. When comparing UV-vis results with the direct observation of the samples, there existed a good correlation. At low concentrations, DO3 interacted with MWCNT by modifying the way they aggregated suggesting the existence of interactions. However, these interactions are so weak that they cannot be distinguished by analyzing a spectrum. But as DO3 was added, the interactions became more important as MWCNT were better dispersed and thus the exposed area increased. As already mentioned, MWCNT have an absorption band around 4.5 eV (around 274 nm). This band is associated to MWCNT π-plasmon electronic transition and it becomes stronger when MWCNT have better dispersion as the superficial area increases. Our results are conclusive and reveal that π-π stackings interaction between the benzene rings of DO3 molecules and MWCNT sidewalls do exist. These interactions lead to stable dispersions of the nanotubes for days. We show FE-SEM micrographs in Figure 8 to visualize the efficiency in the noncovalent functionalization of the MWCNT with DO3 through changes in the morphology of MWCNT. In Figure 8, we show the FE-SEM micrographs of samples of MWCNT with and without DO3. Figure 8a,c corresponds to the sample without DO3 with two different magnifications, 100 000× and 300 000×, respectively. Figure 8b,d shows the micrographs of the sample containing DO3 in a weight ratio DO3/MWCNT ) 2/1 with two different magnifications, 100 000× and 300 000×, respectively. To compare the micrographs, the same volume of samples with MWCNT was

immediately pippeted after sonication. It must be clarified that large quantities of MWCNT were deposited onto the silicon substrate in order to clearly observe the noncovalent functionalization. Several drops of each solution were deposited on the silicon substrate. Drops were put taking care that the previous drop was already dried. A change in the morphology of the carbon nanotubes can be clearly seen when DO3 was added. When comparing Figure 8 panels a and b it can be appreciated that DO3 covers the MWCNT all along the sample. Micrographs show that MWCNT were wrapped by DO3, thus suggesting an efficient noncovalent modification of their surface, as already reported.20 Micrographs c and d are a magnification of the same area of each of the samples in THF and with DO3, respectively. In these micrographs, the changes in the morphology of carbon nanotubes are even more noticeable. Figure 8b,d allows one to see the DO3 coverage onto MWCNT. Taking into account that DO3 is soluble in THF, it could be expected that a good dispersion of MWCNT in THF will occur. It must be noted that FE-SEM micrographs depend on how the sample is prepared and they are not conclusive to evaluate the efficiency of DO3 as a dispersing agent for MWCNT. The experimental results showing the efficiency of DO3 as dispersing agent are mainly based on Figures 2 and 7 and further supported by the UV-vis absorbance measurements. 4. Conclusions In this work, the interactions between an azobenzene derivative and MWCNT have been investigated. Stable dispersions of MWCNT and DO3 in THF were achieved during at least 45 days for weight ratios of DO3/MWCNT higher than two-thirds. UV-vis absorption spectroscopy results indicated that interactions between MWCNT and DO3 molecules do exist. These interactions were noticeable in the absorption band around 276

14352

J. Phys. Chem. C, Vol. 114, No. 34, 2010

nm associated to the benzene ring of the DO3 molecules. This fact suggested a successful noncovalent functionalization of the MWCNT by π-π interactions with DO3. In contrast to the changes in the absorption band of the benzene ring, the electronic transitions of the azo group were almost unaffected. Taking into account these results altogether, we conclude that π-π stacking interactions existing between DO3 molecules and MWCNT are high enough to obtain stable dispersions of the nanotubes for several days. Acknowledgment. The research leading to these results has received funding from the European Community’s Seventh Framework Programme (FP7/2007-2013) under Grant 213939. The authors are thankful for funds of Buenos Aires University (UBACyT X094 and X118), CONICET PIP 2010-2012, Project 11220090100699 and CONICET PIP 2009-2011, Project 11220080103047, ANPCYT 2008-2010, Project 2284, Basque Country Government (Grupos Consolidados -IT-365-07) and ETORTEK/inanoGUNE projects. References and Notes (1) Usrey, M. L.; Chaffee, A.; Jeng, E. S.; Strano, M. S. Application of polymer solubility theory to solution phase dispersion of single-walled carbon nanotubes. J. Phys. Chem. C 2009, 113, 9532–9540. (2) Tournus, F.; Latil, S.; Heggie, M. I.; Charlier, J. C. π stacking interaction between carbon nanotubes and organic molecules. Phys. ReV. B 2005, 72, 75431–5. (3) Yi, W.; Malkovskiy, A.; Chu, Q.; Sokolov, A. P.; Colon, M. L.; Meador, M.; Pang, Y. Wrapping of Single-Walled Carbon Nanotubes by a π-Conjugated Polymer: The Role of Polymer Conformation-Controlled Size Selectivity. J. Phys. Chem. B 2008, 112, 12263–9. (4) De Falco, A.; Fascio, M. L.; Lamanna, M. E.; Corcuera, M. A.; Mondragon, I.; Rubiolo, G. H.; DSˇAccorso, N. B.; Goyanes, S. Thermal treatment of the carbon nanotubes and their functionalization with styrene. Physica B 2009, 404, 2780–3. (5) Piret, J. P.; Detriche, S.; Vigneron, R.; Vankoningslooans, S.; Rolin, S.; MejiaMendoza, J. H.; Masereelans, B.; Lucas, S.; Delhalle, J.; Luizi, F.; Saout, C.; Toussaint, O. Dispersion of multi-walled carbon nanotubes in biocompatible dispersants. J. Nanopart. Res. 2010, 12, 75–82. (6) Angelikopoulos, P.; Gromov, A.; Leen, A.; Nerushev, O.; Bock, H.; Campbell, E. E. B. Dispersing Individual Single-Wall Carbon Nanotubes in Aqueous Surfactant Solutions below the cmc. J. Phys. Chem. C 2010, 114, 2–9. (7) Lu, J.; Nagase, S.; Zhang, X.; Wang, D.; Ni, M.; Maeda, Y.; Wakahara, O. T.; Nakahodo, T.; Tsuchiya, T.; Akasaka, T.; Gao, Z.; Yu,

Dı´az Costanzo et al. D.; Ye, H.; Mei, W. N.; Zhou, Y. Selective Interaction of Large or ChargeTransfer Aromatic Molecules with Metallic Single-Wall Carbon Nanotubes: Critical Role of the Molecular Size and Orientation. J. Am. Chem. Soc. 2006, 128, 5114–5118. (8) Ji, Y.; Huang, Y. Y.; Tajbakhsh, A. R.; Terentjev, E. M. Polysiloxane surfactants for the dispersion of carbon nanotubes in nonpolar organic solvents. Langmuir 2009, 25, 12325–31. (9) Iamsamai, C.; Hannongbua, S.; Ruktanonchai, U.; Soottitantawat, A.; Dubas, S. T. The effect of the degree of deacetylation of chitosan on its dispersion of carbon nanotubes. Carbon 2010, 48, 25–30. (10) Hu, C.; Chen, Z.; Shen, A.; Shen, X.; Lin, J.; Hu, S. Water-soluble single-walled carbon nanotubes via non covalent functionalization by rigid, planar and conjugated diazo dye. Carbon 2006, 44, 428–434. (11) Zhang, W.; Ravi, S; Silva, P. The effects of phenolic hydrogens and methyl substitute groups in organic dyes on their dispersion of multiplewalled carbon nanotubes. Carbon 2010, 48, 2063–71. (12) Ham, H. T.; Choi, Y. S.; Chung, I. J. An explanation of dispersion states of single-walled carbon nanotubes in solvents and aqueous surfactant solutions using solubility parameters. J. Colloid Interface Sci. 2005, 286, 216–23. (13) Broza, G. Thermoplastic elastomers with multi-walled carbon nanotubes: Influence of dispersion methods on morphology. Compos. Sci. Technol. 2010, 70, 1006–10. (14) Han, M. S.; Lee, Y. K.; Kim, W. N.; Lee, H. S.; Joo, J. S.; Park, M.; Lee, H. J. Effect of Multi-walled Carbon Nanotube Dispersion on the Electrical, Morphological and Rheological Properties of Polycarbonate/ Multi-walled Carbon Nanotube Composites. Macromol. Res. 2009, 17, 863– 69. (15) Ma, P. C.; Mo, S. Y.; Tang, B. Z.; Kim, J. K. Dispersion, interfacial interaction and re-agglomeration of functionalized carbon nanotubes in epoxy composites. Carbon 2010, 48, 1824–34. (16) Chua, T. P.; Mariatti, M.; Azizan, A.; Rashid, A. A. Effects of surface-functionalized multi-walled carbon nanotubes on the properties of poly(dimethyl siloxane) nanocomposites. Compos. Sci. Technol. 2010, 70, 671–7. (17) Lidorikis, E.; Ferrari, A. C. Photonics with Multiwall Carbon Nanotube Array. ACS Nano 2009, 3, 1238–1248. (18) Goyanes, S.; Rubiolo, G. R.; Salazar, A.; Jimeno, A.; Corcuera, M. A.; Mondragon, I. Carboxylation treatment of multiwalled carbon nanotubes monitored by infrared and ultraviolet spectroscopies and scanning probe microscopy. Diamond Relat. Mater. 2007, 16, 412–41. (19) O’Connell, M. J.; Bachilo, S. M.; Huffman, C. B.; Moore, V. C.; Strano, M. S.; Haroz, E. H.; Rialon, K. L.; Boul, P. J.; Noon, W. H.; Kittrell, C.; Ma, J.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E. Band Gap Fluorescence fron Individual Single-Walled Carbon Nanotubes. Science 2002, 297, 593–596. (20) Jimeno, A.; Goyanes, S.; Eceiza, A.; Kortaberria, G.; Mondragon, I.; Corcuera, M. A. Effects of Amine Molecular Structure on Carbon Nanotubes Functionalization. J. Nanosci. Nanotechnol. 2009, 9, 6222–6227.

JP102580B