A Study of the Interaction between Single-Walled Carbon Nanotubes

10418

J. Phys. Chem. C 2008, 112, 10418–10422

A Study of the Interaction between Single-Walled Carbon Nanotubes and Polycyclic Aromatic Hydrocarbons: Toward Structure-Property Relationships Sourabhi Debnath,* Qiaohuan Cheng, Theresa G. Hedderman, and Hugh J. Byrne FOCAS Institute/School of Physics, Dublin Institute of Technology, KeVin Street, Dublin 8, Ireland ReceiVed: February 29, 2008; ReVised Manuscript ReceiVed: April 28, 2008

The interaction between polycyclic aromatic hydrocarbons (PAHs) and single walled carbon nanotubes (SWNTs) has been studied by fluorescence spectroscopy. Fluorescence quenching as a function of concentration is modeled, revealing regions where the interaction causes a debundling of the carbon nanotubes. The binding energies between PAHs and SWNTs are obtained from the model. It is found that there is a linear correlation between the binding energy of PAHs with SWNTs and the molecular weight of the PAHs. By plotting these values of binding energy against the molecular weight of these PAHs, a good structure-property relationship governing the interaction has been demonstrated. Introduction Single-walled carbon nanotubes (SWNTs) are proposed to be the most promising of all nanomaterials, with unique electronic, thermal, and mechanical properties1–3 which potentially lend themselves to a variety of applications,4,5 but there are some fundamental problems that hinder the development of the applications of SWNTs. There are several production processes of SWNTs. Despite sustained efforts, all currently known SWNT bulk synthetic techniques generate significant quantities of impurities. In bulk production, SWNTs also grow in bundles or ropes and are largely insoluble in common organic solvents. SWNTs can have a range of structures, and their electronic properties depend on their structure and as well as their diameter. Currently, there is no production process that can isolate only one particular type of SWNTs. For these reasons, carbon nanotubes have been slow to reach real potential applications. In an attempt to speed up the potential applications of SWNTs, it is important to purify, increase the solubility of, disperse, and separate SWNTs with respect to their electronic properties. It has been found that organic solvents in the presence of organic molecules such as polycyclic aromatic hydrocarbons (PAHs) can easily interact with SWNTs, and the solubility of SWNTs is enhanced in the solvent.6,7 Atomic force microscopy and Raman spectroscopy have shown that the PAHs effectively debundle the SWNTs at intermediate and low concentrations. It has further been demonstrated that anthracene predominantly solubilizes metallic nanotubes, whereas p-terphenyl predominantly solubilizes semiconducting nanotubes.8 It is therefore of interest to investigate other molecules of the polyacene and polyphenyl oligomer series to further understand the interaction and optimize the selective solubilization procedure. To understand the interaction between SWNTs and PAHs, a fluorescence model based on the adsorption/desorption equilibrium of SWNTs and PAHs has been used.9 In this model, the quenching of fluorescence represents the interaction between SWNTs and organic polymers. This model can be used to estimate the binding energy between polymers and SWNTs. The ratio of the maximum fluorescence intensity of the composite sample, which contains bound and unbound molecules, to the * Corresponding author. Tel: +353 1 402 7932. Fax: +353 1 402 7901. E-mail: [email protected].

maximum fluorescence of the molecular solutions, which consists solely of unbound molecules, is plotted as a function of concentration. The model is presented for low concentration; when the system is in equilibrium, the adsorption rate equals the desorption rate.9 According to this model, as the fraction of free molecules changes over the concentration range, this fraction of free molecules can be described by a characteristic concentration9 and the concentration of SWNTs.

NF ⁄(NF+NB) ) 1 ⁄ (1 + CNT ⁄ C0) ) Flcomp ⁄Flpolymer

(1)

where

C0 ) π2υFbunAbune-EB ⁄kT/(48Df)

(2)

In eq1, C0 is called the characteristic concentration, as defined in ref 9, and CNT is the concentration of the SWNTs. Because SWNTs and PAHs are used in a 1:1 ratio by weight, CNT can also represent the concentration of the molecules. NF is the number of free molecules, NB is the number of bound molecules, Flcomp is the fluorescence of the composite, and Flpolymer is the fluorescence of the polymer, which gives the fraction of free molecules in solution. In equation 2, υ is a pre-exponential frequency factor, Fbun is the SWNT bundle mass density, Abun is the bundle surface area, EB is the binding energy, k is the Boltzmann constant, T is the absolute temperature, f is the spatial integral, D is the diffusion coefficient, and D ) kT/(6πηa), where η is the solvent viscosity and a is the molecular hydrodynamic radius.9 From eq 2, the binding energy, EB, between the molecules and the SWNTs can be obtained.9 The aim of this paper is to probe the interactions between SWNTs and PAHs and to establish structure-property relationships governing this interaction as a guide to optimizing selective solubilization. The interactions of SWNTs with PAHs are investigated through spectroscopic methods such as fluorescence. The fluorescence concentration dependence studies define the concentration range where aggregated PAHs and isolated PAHs exist. In this work, PAHs of two oligomer series, the polyacene series and the polyphenyl series, are used, and chloroform is used as a solvent. Sample Preparation. Two sets of solutions in the range of ∼1.22 × 10-6 mol/L to ∼2.50 × 10-3 mol/L for naphthalene,

10.1021/jp8017925 CCC: $40.75  2008 American Chemical Society Published on Web 06/19/2008

Interaction between SWNTs and PAHs

J. Phys. Chem. C, Vol. 112, No. 28, 2008 10419

∼1.907 × 10-8 mol/L to ∼2.5 × 10-3 mol/L for anthracene, ∼1.8626 × 10-11 mol/L to ∼3.125 × 10-4 mol/L for tetracene, ∼2.38 × 10-9 mol/L to ∼5 × 10-3 mol/L for biphenyl, ∼1.164 × 10-12 mol/L to ∼1.952 × 10-5 mol/L for p-terphenyl, ∼3.637 × 10-14 mol/L to ∼1.25 × 10-3 mol/L for p-quaterphenyl, and ∼2.22 × 10-18 mol/L to ∼3.12 × 10-4 mol/L for p-quinquephenyl were prepared. To one of the two sets, HiPco SWNTs obtained from Carbon Nanotechnologies, Inc. (16200 Park Row, Houston, TX 77087) were added in a 1:1 ratio by weight (w/w), SWNTs/PAHs. At the highest concentration, both PAHs and composite solutions were sonicated using a sonic tip (Ultrasonic processor VCX 750 W) for 3 × 10 s at 38% of the power output, then these samples were serially diluted by a factor of 2 with pure chloroform down to the lowest concentration. Each diluted sample was further tip-sonicated as above and was then allowed to settle for 72 h, after which the supernatant liquid from the composite samples was carefully withdrawn. The suspensions were allowed to settle for an additional 24 h before being characterized by fluorescence spectroscopy (Perkin-Elmer LS55). It was noted that for suspensions above the dispersion limit, described below, a considerable amount of precipitation of SWNTs was observed, so the solubilization is only partial. At concentrations at which the SWNTs are well-dispersed, no precipitate was found. Because composite solutions were prepared by using SWNTs/PAHs in a 1:1 ratio by weight (w/ w), it can be stated that below the dispersion limit, the concentrations of PAHs and SWNTs are the same. Concentrations quoted are those of the as-prepared solutions in terms of PAH concentration. Results and Discussion As pointed out by Coleman et al.,9 in the case of polymer/ SWNT composite solutions, there are two forms of polymer molecules: free polymers and those that are bound to the SWNTs. In the polymer/SWNT solutions, assuming that the interaction with the nanotubes quenches the fluorescence, the observed fluorescence is due to the free polymer only.10 Measurement of the fluorescence of the polymer solution in the absence of SWNTs represents the total amount of polymer present (NTotal). Hence, NTotal ) NBound + NFree, where NFree is the amount of free polymers in composite solution. Thus, the ratio of the fluorescence intensity for a polymer/SWNT solution to that of a polymer solution of equivalent concentration is a measure of the fraction of free polymer molecules in the composite solution at that concentration.10 Adapting Coleman’s model9 and applying eq 1, a plot of the fraction of free oligomer molecules as a function of concentration was achieved (Figures 1, 2, and 3). All curves show the same general trend, in agreement with that seen by Coleman et al.9 and Hedderman et al.6 At the highest concentrations, the molecularly dispersed PAHs interact with SWNT bundles. With decreasing concentration, the PAHs act to disperse the SWNTs, decreasing the bundle size (as confirmed by AFM6,9,11), increasing the surface area per SWNT available to the molecules for adsorption, and resulting in a concentration-dependent decrease in free PAH molecules, as represented by the ratio of eq 1. This decrease reaches a minimum at the point when the SWNTs are maximally dispersed. Similar concentration-dependent debundling has been observed and verified using AFM in organic solvent12 and water/surfactant11 dispersions of SWNTs. The concentration of the maximum dispersion has been termed the dispersion limit.11,12 From this point, decreasing the concentration, the surface area per SWNT is constant, and the fluorescence

Figure 1. A plot of the fraction of free naphthalene (a) anthracene (b) and tetracene (c) as a function of concentration. The solid line is the best fit to eq 1 and gives C0 (characteristic concentration) values of ∼6 × 10-5, ∼1.75 × 10-5, and ∼4 × 10-7 mol/L for naphthalene, anthracene, and tetracene, respectively.

quenching behavior can be fitted by equation 1 using C0 (characteristic concentration) as a fitting parameter. In Figure 1a, b, and c, for naphthalene between ∼2.5 × 10-3 and ∼1.17 × 10-4 mol/L, for anthracene between ∼3.12 × 10-4 and ∼3.90 × 10-5 mol/L, and tetracene between ∼4.88 × 10-6 and ∼2.28 × 10-7 mol/L, the increased quenching of the fluorescence with decreasing concentration represents the debundling process. The bundles are dispersed progressively until the nanotubes are fully dispersed at concentration ranges from ∼7.81 × 10-5 to ∼6.10 × 10-7 mol/L (naphthalene), ∼1.95 × 10-5 to ∼3 × 10-7 mol/L (anthracene), and ∼1.14 × 10-7 to ∼5.96 × 10-10 mol/L (tetracene). These concentration ranges were fitted using eq 1 and with C0 as a fitting parameter. The solid line is the best fit to eq 1 and gives C0 values of ∼6 ×

10420 J. Phys. Chem. C, Vol. 112, No. 28, 2008

Figure 2. A plot of the fraction of free biphenyl (a) and p-terphenyl (b) as a function of concentration. The solid line is the best fit to eq 1 and gives C0 (characteristic concentration) values of ∼1.5 × 10-5 and ∼2.5 × 10-9 mol/L for biphenyl and p-terphenyl, respectively.

10-5, ∼1.75 × 10-5, and ∼4 × 10-7 mol/L for naphthalene, anthracene, and tetracene, respectively. In Figure 2a, b, for biphenyl, between ∼5 × 10-3 and ∼1.6 × 10-4 mol/L, and for p-terphenyl, between ∼1.95 × 10-5 and ∼3.8 × 10-8 mol/L, SWNT debundling occurs. From concentrations ∼7.8 × 10-5 to ∼2.4 × 10-9 mol/L (biphenyl) and ∼1.90 × 10-8 to ∼4.65 × 10-12 mol/L (p-terphenyl), isolated PAH molecules and fully dispersed SWNTs exist. These concentration ranges were fitted using eq 1 and with C0 as a fitting parameter. The solid line is the best fit to eq 1 and gives C0 values of ∼1.5 × 10-5 and ∼2.5 × 10-9 mol/L for biphenyl and p-terphenyl, respectively. Similarly, in Figure 3a, b, for p-quaterphenyl, between ∼3.8 × 10-8 and ∼6.0 × 10-10 mol/L, and for p-quinquephenyl, between ∼3.0 × 10-10 and ∼1.1 × 10-15 mol/L, SWNT debundling occurs. From concentrations of ∼3.0 × 10-10 to ∼7.3 × 10-14 mol/L (p-quaterphenyl) and ∼5.7 × 10-16 to ∼2.2 × 10-18 mol/L (p-quinquephenyl), isolated PAH molecules exist. These concentration ranges were fitted using eq 1 and with C0 as a fitting parameter. The solid line is the best fit to eq 1 and gives C0 values of ∼7 × 10-11 and ∼4 × 10-16 mol/L for p-quaterphenyl and p-quinquephenyl, respectively. Figure 4 shows the relationship between the estimated dispersion limit of HiPco SWNTs and the molecular weight of the PAHs. This plot is an indication of the ease with which HiPco SWNTs are dispersed by the different PAHs, and it is found that with the increasing length of the PAHs, the concentration at which HiPco SWNTs are dispersed is reduced. This is an indication that the smaller molecules are more effective at dispersing the SWNTs. The polyacene oligomers are more effective than the polyphenyl counterparts, and this

Debnath et al.

Figure 3. A plot of the fraction of free p-quaterphenyl (a) and p-quinquephenyl (b) as a function of the concentration. The solid line is the best fit to eq 1 and gives C0 (characteristic concentration) values of ∼7 × 10-11 and ∼4 × 10-16 mol/L for p-quaterphenyl and p-quinquephenyl, respectively.

Figure 4. Estimated dispersion limit of HiPco SWNTs as a function of the molecular weight of the PAHs.

may be an indication that the more planar and rigid polyacenes are more effective at debundling than the polyphenyl oligomers in which the phenyl units are free to rotate relative to each other. Employing eq 2, the binding energy between the PAHs and SWNTs is calculated. For the calculation of binding energy, the value of υ is taken as 1018 Hz,8 the value of Fbun for HiPco SWNTs is taken as 1.6 × 10-3 Kg/m-3,8 the value of Abun is taken as 1.5 × 10-15 m2,8 the value of η for chloroform is taken as 0.536 × 10-3 s Pa,9 the value of T is taken as 298 K, the value of k is taken as 1.38 × 10-23 J/K, and the value of other parameters chosen are given in Table 1, where the values of a for different PAHs are obtained by molecular modeling (ChemOffice 2004). To date, no molecular modeling of the interaction of PAHs with SWNTs, particularly including diameter and energy band

Interaction between SWNTs and PAHs

J. Phys. Chem. C, Vol. 112, No. 28, 2008 10421

TABLE 1: Parameters for the Calculation of Binding Energy PAHs

C0 (Kg/m3)

a (m)

EB (eV)

naphthalene anthracene tetracene biphenyl p-terphenyl p-quaterphenyl p-quinquephenyl

7.68 × 10 3.11 × 10-3 9.13 × 10-5 2.31 × 10-3 5.75 × 10-7 2.14 × 10-8 4 × 10-16

2.4 × 3.6 × 10-10 4.8 × 10-10 4.47 × 10-10 7.06 × 10-10 9.65 × 10-10 12.23 × 10-10

0.46 0.49 0.58 0.50 0.72 0.81 1.12

-3

10-10

It is further noted that the values calculated for anthracene and p-terphenyl are significantly lower in this work, using chloroform as solvent, than those calculated for toluene solutions of SWNTs produced by the laser vaporization method.6 The difference may point toward the potential importance of the solvent in mediating the interaction, although it is also likely that differing initial bundle sizes and densities of SWNTs produced through different mechanisms can significantly affect the process.8 Conclusion

gap dependence, has been reported. However, the interactions of PAHs with themselves and graphite have been modeled using density functional theory and experimentally determined using thermal desorption spectroscopy.13,14 The quasicrystalline arrangements of closely packed SWNTs are like graphite, and they are held together by a long-range van der Waals interaction. Because both graphite and SWNTs possess similar properties, it is reasonable to expect similar degrees of interaction for SWNTs with PAHs. Theoretical models predict and thermal desorption spectroscopy confirms that an increase in the carbon number of PAHs increases the binding energy between PAHs, with larger PAHs having similar interlayer distances and binding energies to graphite.13,14 The binding energy between PAHs and graphite are predicted to be larger than the binding energy of a corresponding PAH dimer. The cohesive energy is determined by thermal desorption measurements of PAHs on the graphite surface and reveals a binding energy of 52 ( 5 meV per carbon atom with a contribution of 27 meV per hydrogen.13,14 Figure 5 shows the binding energy between PAHs and HiPco SWNTs determined above as a function of molecular weight of the PAHs. From this graph, it is found that the binding energy between PAHs and SWNTs increases approximately linearly with increasing molecular weight of the PAHs. The binding energy of naphthalene, anthracene, tetracene, biphenyl, p-terphenyl, p-quaterphenyl, and p-quinquephenyl are found to be ∼0.46, ∼0.49, ∼0.58, ∼0.50, ∼0.72, ∼0.81, and ∼1.12 eV, respectively, with HiPco SWNTs. From this result, it can be said that oligomers of the polyphenyl series bind more tightly to the HiPco SWNTs surface than those of the polyacene series. This is because per ring they have more carbon and hydrogen atoms. Figure 5 shows a linear fit to the data from which a value of 20 ( 5 meV per carbon atom and 28 ( 5 meV per hydrogen atom are calculated. The values are in approximate agreement with those calculated for the interaction of PAHs with graphite, although it is noted that the curvature of the SWNTs should have the effect of reducing the binding interaction.

The interaction between SWNTs and PAHs has been studied by using a fluorescence model based on the adsorption/ desorption equilibrium of SWNTs and PAHs. The results indicate a significant structural dependence of the interaction between SWNTs and PAHs. It is found that the dispersion limit of HiPco SWNTs decreases with increasing molecular weight of the PAHs. This points to the structure-property relationship governing how effectively different PAHs disperse HiPco SWNTs. One can state that smaller PAHs debundle SWNTs at higher concentration than the larger PAHs. It is also shown that the polyacene series are more effective at debundling the SWNTs than their polyphenyl counterparts. By using the model mentioned above, the binding energy between SWNTs and different PAHs as a function of the molecular weight of the corresponding PAH molecules is obtained. It is found that the binding energy between SWNTs and PAHs increases as the molecular weight of the PAHs increases, thereby establishing a linear relationship between the binding energy of SWNTs with PAHs and the molecular weight of corresponding PAHs. It can be stated that because the number of carbon and hydrogen of polyphenyl PAHs is greater than that of their counterpart, the binding energy of polyphenyl PAHs with SWNTs is greater than that between polyacene PAHs and SWNTs. This further supports the observation regarding the debundling process: the rigid acene oligomers bind less readily to the surfaces of bundles and intercalate more easily into the bundles, whereas the polyphenyl oligomers bind more strongly to the surface of the bundles and, due to their rotational freedom, do not intercalate as easily into the bundles to disperse them. However, once dispersed, the polyphenyl oligomers bind more strongly to the dispersed SWNTs than their polyacene counterparts. These results give a strong indication that structure-property relationships for the interaction of PAHs with SWNTs do exist and are relatively easily accessible experimentally. Further studies will explore the structural dependence of selective solubilization. Acknowledgment. This project is funded under the Science Foundation Ireland Research Frontiers Program PHY037 2006. References and Notes

Figure 5. Binding energies between PAHs and HiPco SWNTs as a function of the molecular weight of the PAHs.

(1) Hone, J.; Whitney, M.; Piskoti, C.; Zettl, A. Phys. ReV. B 1999, 59, R2514. (2) Thess, A.; Lee, R.; Nikolaev, P.; Dai, H. J.; Petit, P.; Robert, J.; Xu, C. H.; Lee, Y. H.; Kim, S. G.; Rinzler, A. G.; Colbert, D. T.; Scuseria, G. E.; Tomanek, D.; Fischer, J. E.; Smalley, R. E. Science 1996, 273, 483. (3) Yu, M. F.; Files, B. S.; Arepalli, S.; Ruoff, R. S. Phys. ReV. Lett. 2000, 84, 5552. (4) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Science 2002, 297, 787. (5) Javey, A.; Guo, J.; Farmer, D. B.; Wang, Q.; Yenilmez, E.; Gordon, R. G.; Lundstrom, M.; Dai, H. J. Nano Lett. 2004, 4, 1319. (6) Hedderman, T. G.; Keogh, S. M.; Chambers, G.; Byrne, H. J. J. Phys. Chem. B 2006, 110, 3895. (7) Hedderman, T. G.; Keogh, S. M.; Chambers, G.; Byrne, H. J. J. Phys. Chem. B 2004, 108 (49), 18860.

10422 J. Phys. Chem. C, Vol. 112, No. 28, 2008 (8) Hedderman, T. G. Ph.D. Thesis; Dublin Institute of Technology, Dublin, Ireland, 2006. (9) Coleman, J. N.; Fleming, A.; Maier, S.; O’Flaherty, S.; Minett, A. I.; Ferreira, M. S.; Hutzler, S.; Blau, W. J. J. Phys. Chem. B 2004, 108, 3446. (10) Giordani, S.; Bergin, S. D.; Drury, A.; Mhuircheartaigh, E. N.; Ruther, M.; Coleman, J. N.; Blau. W. J. Proc. SPIE 2005, SPIE-5824, 42.

Debnath et al. (11) (12) M. M.; (13) (14)

Priya, B. R.; Byrne, H. J. J. Phys. Chem. C 2008, 112, 332. Giordani, S.; Bergin, S. D.; Nicolosi, V.; Lebedkin, S.; Kappes, Blau, W. J.; Coleman, J. N. J. Phys. Chem. B 2006, 110, 15708. Chakarova, S. D.; Schroder, E. J. Chem. Phys. 2005, 122, 054120. Zacharia, R.; Ulbricht, H.; Hertel, T. Phys. ReV. B 2004, 69, 155406.

JP8017925