Reaction Mechanisms for Graphene and Carbon Nanotube

3340

J. Phys. Chem. C 2010, 114, 3340–3345

Reaction Mechanisms for Graphene and Carbon Nanotube Fluorination Sı´lvia Osuna,†,‡ Miquel Torrent-Sucarrat,† Miquel Sola`,‡ Paul Geerlings,† Christopher P. Ewels,*,§ and Gregory Van Lier*,† Research Group of General Chemistry (ALGC), Vrije UniVersiteit Brussel s Free UniVersity of Brussels (VUB), Pleinlaan 2, B-1050 Brussels, Belgium, Institut de Quı´mica Computacional and Departament de Quı´mica, UniVersitat de Girona, E-17071 Girona, Catalonia, Spain, and Institut des Mate´riaux Jean Rouxel, CNRS, UniVersite´ de Nantes, 2 rue de la Houssinie`re, 44322 Nantes, France ReceiVed: September 14, 2009; ReVised Manuscript ReceiVed: December 19, 2009

Via density functional theory calculations, we explore the process of graphene and carbon nanotube fluorination. Contrary to graphene, HF-catalyzed carbon nanotube fluorination adds fluorine pairs at (1,4) spacing, naturally leading to a self-organized C4F coverage, which can subsequently rearrange into dense axial C2F lines by fluorine diffusion upon solubilization or heat treatment. Controlling the fluorination process and subsequent nanotube processing results in fluorinated material types with unexplored electronic and chemical properties. Introduction There exist few techniques to date that allow variation of the properties of single carbon nanotubes along their length in a controlled manner or selective modification of regions of a graphene surface. Of the different chemical modification techniques available, fluorination takes a special place as it can provide for full outerwall functionalization up to C2F for nanotubes and CF for graphene.1 A range of functionalization methods is available for carbon nanotubes using either atomic fluorine,2 F2 gas,3 inorganic fluorides such as BrF34,5 or XeF2,6 and radio frequency (RF) plasma functionalization.7 In particular, the latter fluorination method is easy, fast, and upscalable to industrial levels with good control over the fluorination degree. Fluorination is used to debundle and disperse carbon nanotubes, enabling their solubilization in, for example, alcohols,8 expected to be due to hydrogen bonding between the surface fluorine and the alcohol molecules.8,9 Potential applications for fluorinated nanotubes include the use in lithium ion batteries,10–13 supercapacitors,14 and as lubricants.15–17 Recently, they have been applied as a new route for polymer reinforcement, where they provide important improvement of the mechanical properties of polymers as compared to pristine nanotube fillers.18,19 Subsequent defluorination is possible through sonication with anhydrous hydrazine. Also, upon heat treatment, defluorination can be achieved,2,8 which provides for a method of producing solid conducting carbon nanotube blocks with good biocompatibility.20,21 Fluorinated parts of nanotubes can also be etched away upon heat treatment, resulting in short “cut” nanotubes.22 There has been much debate on the transport properties of functionalized carbon nanotubes.23 Although a decrease of the conductivity has been measured for fluorinated nanotubes,24 the band gap has been predicted to vanish for metallic nanotubes with an axial fluorine addition pattern.25,26 The study of the possible addition patterns remains therefore an important issue. Self-organized surface fluorination should allow tuning of carbon * To whom correspondence should be addressed. E-mail: [email protected]. Phone: +32-(0)26293516 (G.V.L.); chris.ewels@ cnrs-imn.fr. Phone: +33-240376407 (C.P.E.). † Vrije Universiteit Brussel - Free University of Brussels. ‡ Universitat de Girona. § Universite´ de Nantes.

nanomaterial properties such as band gap and conduction channels for transport behavior; however, in practice, studies are lacking on precise control of the functionalization procedure and addition pattern characterization. Fluorination of graphite has long been performed experimentally (for an overview, see ref 27) and allows further functionalization and solubilization,28 with possible application for field effect transistors29 and battery applications.27,30 In fluorinated graphite, electrical conduction can be enhanced by promoting ionic bonding between fluorine and carbon atoms, which increases the number of hole carriers, whereas covalent bonding produces a decrease in the overall carrier concentration.31 For gas-phase fluorination of carbon nanotubes, two primary phases have been described in the literature, which are obtained upon increasing fluorination temperature; these are a low-density phase up to a 20-25% F/C ratio and a higher density phase up to to 50% F/C ratio,32 with the transition temperature depending on the specific treatment characteristics and carbon nanotube type.2,33 While both average surface coverage below 25% and up to 50% has been observed by X-ray photoelectron spectroscopy for plasma and gas-phase fluorinated tubes depending on treatment parameters, experimental observation by scanning tunneling microscopy (STM) for low surface coverage has only observed fluorinated bands corresponding to C2F coverage separated by pristine nanotube parts,34 attributed to the most stable equatorial addition pattern forming C2F bands with sharp equatorial edges.35 Nonetheless, these specific phases have been predicted as consisting of isolated benzenoid rings separated by individual fluorinated carbons for the low-density phase, resulting in a maximal C4F ratio with local aromaticity playing a relevant role in the addition pattern,36 and consisting of alternating lines of fluorinated and nonfluorinated carbons over the whole nanotube surface for the high-density C2F phase.37 Notably, experimental evidence of the structure, formation process, and properties of C4F material is lacking. While graphite fluorination is a fundamentally different process since it involves possible fluorine access to both sides of the carbon layer, we might expect graphene to fluorinate with a similar addition mechanism to carbon nanotubes. However, despite the potential interest of such a material, proper characterization of the addition patterns and mechanisms for fluorinated graphene are also

10.1021/jp908887n  2010 American Chemical Society Published on Web 02/10/2010

Graphene and Carbon Nanotube Fluorination

J. Phys. Chem. C, Vol. 114, No. 8, 2010 3341

lacking to date. We therefore present here density functional theory (DFT) calculations to identify the precise fluorination route and full free-energy reaction barriers in HF-catalyzed gasphase fluorination of graphene and carbon nanotubes. We confirm that this results in the C4F structure reported above at lower temperatures and explore the formation routes and properties of hybrid C4F-C2F nanotubes. The fluorination of graphene is predicted to behave very differently with mixed surface C2F/C4F functionalization. Computational Details All DFT calculations were performed using the Gaussian03 program.38 The fluorine addition mechanism for graphene was studied using coronene as a test system and evaluated at the ab initio B3LYP39–41 level of theory with the standard 6-31G(d)42,43 basis set. In this case, all stationary points were characterized as either minima or transition states by computing the vibrational harmonic frequencies.44 Moreover, the study of full Gibbs free energies for the whole process was performed since all transition-state (TS) structures involve three molecules, and thus, the fluorination reaction is expected to be entropically driven. The nanotube fluorine addition was studied using an armchair (7,7) nanotube (C264H24) as a test system. The reaction mechanism was studied within the two-layered ONIOM approach due to the system size, which was recently shown to provide for comparable results for reaction and activation energies as full B3LYP calculations.45,46 The low level of the system was treated with the DFT XR exchange (R ) 2/3) functional in conjunction withtheelectron-gascorrelationfunctional(inVosko-Wilk-Nusair parametrization)47,48 using a 3-21G basis set,42,43 while a core high-level region of 24 carbon atoms (coronene patch) was treated using the hybrid B3LYP method with 6-31G(d). Gibbs free-energy calculations were performed on a coronene test system curved to mirror the curvature of a (7,7) nanotube due to the high computational cost. Migration barriers were studied using the synchronous transitguided quasi-Newton (STQN) method as implemented by Schlegel and co-workers,49 and the QST2 and QST3 approaches were used at the DFT B3LYP/3-21G level of theory again for an armchair (7,7) nanotube (C264H24). Solvent effects were included using the polarizable continuum model (PCM),50 where extra spheres were added inside of the nanotube to keep the inner core solvent-free. All minimum structures were reoptimized with ethanol as the solvent; however, only single-point energy calculations were performed for all transition structures due to the computational cost for this system size. Results and Discussion 1. Fluorination Process for Graphene and for Carbon Nanotubes. First, we analyze the gas-phase fluorination process for graphene. Hereby, coronene is taken as a test system to allow the study of full Gibbs free energies, which are important since all TS structures involve three molecules, and the fluorination reaction is thus expected to be entropically driven. Figure 1 presents the fluorine addition mechanism, showing all reaction intermediates, TSs, and reaction products found. The F2 molecule alone preferentially sits orthogonal to the coronene surface (int1). When HF is also present, a hydrogen bond (int2, 1.66 Å) is formed between HF and the “upper” fluorine atom of the F2 molecule, leading to the dissociation of F2 and the formation of the TS1 transition state, with an exothermic reaction energy of -5.8 kcal/mol (see Figure 1). However, the Gibbs free energy for this dissociation describes a clearly endothermic process (+13.1 kcal/mol), showing that this

Figure 1. The fluorination reaction mechanism for the coronene system is depicted. All optimized structures are also represented, including the most relevant geometrical parameters. All energies are expressed in kcal/mol, and bond distances are in Å. Free Gibbs energies are given in red between brackets.

reaction is indeed driven by the entropy. After the formation of the first fluorine bond with graphene (int3), transition states leading to both (1,2) and (1,4) addition products are found (TS2 and TS3, respectively; see Figure 1). Both possibilities are equally favored as they present approximately the same activation barrier, with again Gibbs free energies being substantially higher. From a kinetic point of view, the fact that both transition states have about the same energy indicates that fluorine addition for a flat system such as graphene is not regioselective. However, the (1,2) addition is thermodynamically more favored by 10 kcal/mol for the final intermediates, and also, the (1,2) addition product is more stable by 10 kcal/mol than the (1,4) product. These calculations thus suggest that HF-catalyzed F2 addition to graphene will result in a mixture of (1,2) and (1,4) addition patterns. We note that the addition of the second fluorine atom has a comparable barrier to that of the first, clearly suggesting a contiguous addition mechanism. This is confirmed when we consider a single fluorine atom bonded to graphene and the isolated FHF- intermediate; a much higher energy of 62.2 kcal/ mol is found with respect to the initial reactants (coronene + F2 + HF). Therefore, although for HF-catalyzed F2 gas-phase fluorination the role of HF is to dissociate the fluorine gas molecule to form F+ and FHF- intermediates, our calculations show that a concerted mechanism is followed rather than fluorine adding in an ionic process.

3342

J. Phys. Chem. C, Vol. 114, No. 8, 2010

Osuna et al.

TABLE 1: Different Stabilities of the Product Intermediates Found for the (7,7) Nanotube Corresponding to the Addition of Two Fluorine Atoms Forming a Hydrogen Bond with the Catalyst HFa

F configuration (1,2) (1,4) (1,2) (1,4)

equatorial axial axial equatorial

∆ER

d(H · · · F1)

d(H · · · F2)

-74.6 -71.3 -67.0 -70.4

1.979 1.776 2.002 2.045

1.958 3.683 1.967 2.046

a d(H · · · F1) and d(H · · · F2) are expressed in Å. All energies are given in kcal/mol.

The carbon nanotube fluorination is studied using an armchair (7,7) nanotube (C264H24). Although we started from the conformations of F2 and HF in the intermediate and TS structures found for graphene, since we expect that the reaction mechanism for nanotube fluorination will be analogous, we were not able to find any transition state, and only the (1,2) and (1,4) final intermediates were located. This is likely due to the potential energy surface being very flat. In Table 1, the relative energies for the different product intermediates are given. As can be seen, (1,2) equatorial and (1,4) axial additions are the most favorable, with reaction energies of -74.6 and -71.3 kcal/mol, respectively. The HF catalyst forms hydrogen bonds with both fluorine atoms in every case considered except for the (1,4) axial adduct. In this case, the fluorine atoms are situated farther apart due to curvature, and HF can form a hydrogen bond with only one fluorine (see Table 1). Since all attempts to find the intermediates and TSs starting from the int2, int3, TS1, TS2, and TS3 conformations found for graphene always led to the final (1,4) intermediate product on the nanotube, this clearly suggests that curvature plays an important role. Since the importance of entropic effects for graphene fluorination (see above), we performed calculations on a coronene test system, curved to mirror the curvature of a (7,7) nanotube, where a constrained optimization was performed imposing the same C-F, F-F, and F-H distances as those found previously for the TS2 and TS3 conformations, coinciding with a (1,2) and (1,4) addition pattern, respectively. The electronic activation barriers (∆E°) found in this case are 5.9 kcal/mol for (1,2) and 2.5 kcal/mol for the (1,4) configuration, which shows that the introduction of curvature leads to a preference for (1,4) addition. When the entropy is also included, these barriers (∆G°) at 298 K shift further to 9.0 kcal/mol for the (1,2) and 5.5 kcal/mol for the (1,4) conformation. The energy difference between the latter TSs leading to either (1,2) or (1,4) addition is reduced from 3.5 kcal/mol at 298 K to 3.0 kcal/mol at 600 K.

Figure 2. Optimized TS geometries for curved coronene (a) at B3LYP/ 6-31G(d) and for the (7,7) nanotube (b) at the ONIOM(B3LYP/631G(d):SVWN/3-21G) level of theory, where the high-level coronene patch considered in the ONIOM approach is indicated in black. Both TSs involve the migration of the fluorine atoms from the (1,4) position to the (1,2) thanks to the formation of a hydrogen bond with the catalyst HF. Most relevant distances have been represented in Å.

The above calculations therefore indicate that for HFcatalyzed gas-phase fluorination of carbon nanotubes, the introduction of curvature leads to a clear preference for a (1,4) addition pattern, especially in a lower-temperature regime. We note that this also suggests that for nonplanar graphene (due to folding, local buckling, etc), (1,4) F2 addition will also be favored over a (1,2) addition pattern. If only (1,4) addition occurs on a carbon nanotube surface, only up to 25% of the carbon atoms can be functionalized, corresponding to a C4F ratio.37 This corresponds to the 20-25% observed experimentally for HF-catalyzed gas-phase fluorination at temperatures up to 150 °C.51 However, when performed at higher temperatures, fluorination is shown experimentally to result in a maximal C2F ratio.1 We have indeed found a transition state for the nanotube and the curved coronene that involves migration of the fluorine atom from the (1,4) position to the (1,2) thanks to the formation of a hydrogen bond with the catalyst HF (see Figure 2). This transition state has an associated activation barrier of 23.0 kcal/mol for the curved coronene and 22.2 kcal/mol for the (7,7) nanotube (see Figure 2), showing that at higher temperatures, fluorination can result in (1,2) addition due to the catalytic effect of HF, in accordance with experimental results. We note that here, both fluorine atoms added originate from the F2 molecule. 2. Fluorine Migration Process. The question now remains whether migration of (1,4) to (1,2) can be obtained upon subsequent treatment of low-temperature domain fluorinated nanotubes. Therefore, migration barriers were studied for an armchair (7,7) nanotube (C264H24) for migration from the (1,4) to the (1,2) position, either through the less-stable (1,3) arrangement or directly from the (1,4) to the (1,2) position, both from the axial to the equatorial and from the equatorial to the axial positions (see Figure 3).52 Consistent with previous calculations,37 the calculated barriers are around 45 kcal/mol in all cases (see Figure 4), showing that these migrations can only occur at higher temperatures. We note that for fluorine

Graphene and Carbon Nanotube Fluorination

J. Phys. Chem. C, Vol. 114, No. 8, 2010 3343

Figure 3. Representation of the different fluorine migrations studied through the nanotube axis (marked with a blue arrow). We have located the direct migration from the (1,4) position to the (1,2), that is, TS(1,4)f(1,2), and the indirect mechanism going through the unstable (1,3) arrangement (TS(1,4)f(1,3), TS(1,3)f(1,2)). The same migrations but following the equatorial axis of the nanotube have also been studied (red arrow).

Figure 5. Representation of the different fluorine addition patterns considered for a (7,7) nanotube. Colored dots represent fluorinated carbon atoms, and the arrow indicates the nanotube axis. The reaction energy per added fluorine pair is given in kcal/mol.

Figure 4. Representation of the fluorine migration barriers found; blue color is used to represent migrations along the axis, and black is for the equatorial. Those values inside of brackets are the migration barriers in ethanol as the solvent. All energies are expressed in kcal/mol.

migration from the (1,6) to the (1,4) position, we found a comparable activation barrier of 47 kcal/mol. These migration barriers suggest that upon higher-temperature treatment, fluorine added via a (1,4) addition pattern at lower temperature will surface migrate into a denser packing associated with fluorine in (1,2) addition sites. This densification will eventually result in the close-packed maximal addition of a C2F ratio. At lower temperatures, the surface migration cannot occur, and fluorination will be restricted to a maximal uniform coverage of C4F, corresponding to a self-organized addition pattern where nonfluorinated benzenoid rings are surrounded by fluorine atoms in only the (1,4) position. The above results have important implications for the production of new hybrid carbon nanotube fluorides, which have not been studied previously. By selecting the fluorination temperature and subsequent heating, with optionally making repeated fluorination steps, it should be possible to produce carbon nanotubes with alternating bands of pristine C4F or C2F material. For example, fluorinating at low temperatures to give complete C4F coverage, followed by heat treatment (resulting in alternating pristine and C2F bands), followed by repeat

fluorination at room temperature, should result in nanotubes with alternating C4F-C2F bands. At first sight, the above results may seem contradictory with the observation of fluorine banding by STM at fluorination temperatures up to 150 °C, where C2F-fluorinated circumferential bands are alternated with pristine nanotube parts.34 This addition pattern could only correspond to the above-described (1,4) addition mechanism if migration of the fluorine atoms has occurred after the fluorination process. Previous preliminary calculations suggested that this could result from solubilization of the C4F-fluorinated nanotubes, where the fluorine migration barriers are lowered through hydrogen bonding with the solvent.37 We therefore repeated the migration barrier calculations using ethanol as the solvent. These calculations show a decrease in the electronic migration barriers from 11 to 4 kcal/ mol in the presence of solvent, depending on the case (see Figure 4). However, if an explicit ethanol molecule is taken into account, the migration barrier lowers by 10 kcal/mol, showing that solubilization of the nanotubes in alcohol will indeed help the fluorine migration over the surface to a more stable (1,2) addition pattern. We note that entropic effects were not taken into account here and that only one ethanol molecule was considered, whereas multiple ethanol molecules can surround the migrating fluorine atom. A more important lowering of the migration barriers upon solubilization can thus be expected. Migration will also be driven by the reduction of the strain occurring between fluorinated and unfluorinated bands.35 In addition, if an explicit ethanol molecule is taken into account for the TS of Figure 2a, the activation barrier of 23.0 kcal/mol is reduced by 11.6 kcal/mol, leaving the migration barrier at only 11.40 kcal/mol. The above results also indicate that banded nanotubes can be obtained either upon subsequent heat treatment of nanotubes fluorinated at lower temperature or by solubilization. Solubilization of these nanotubes will therefore lead to a material with different addition patterns than the original fluorinated nanotubes.

3344

J. Phys. Chem. C, Vol. 114, No. 8, 2010

Upon fluorine addition at higher temperatures or migration over the surface, different addition patterns can still be formed. For the (7,7) nanotube under consideration, the most stable C2F pattern on carbon nanotubes is to add in alternating fluorinated and nonfluorinated zigzag lines, as close to parallel to the nanotube axis as possible (see Figure 5). This results in a maximal deviation of 30° from the nanotube axis for all nanotube diameters and helicities. Moreover, for the same nanotube length, it is more favorable to have a larger fluorinated band (half of the nanotube fluorinated) rather than a two-banded fluorinated nanotube, with binding energies of -75.2 versus -69.4 kcal/mol per fluorine pair, respectively (see Figure 5). This is mainly attributed to the higher deformation energy found for the two-banded fluorinated structures, consistent with earlier analysis predicting that circumferential banding can be caused by axial fluorine addition rather than circumferential patterns.35 Conclusions and General Perspectives The current study suggests that HF-catalyzed F2 addition to graphene and carbon nanotubes follows a different addition pattern. Whereas for graphene addition both (1,2) and (1,4) additions can occur, resulting in a random surface coverage of up to C2F, the introduction of curvature leads to a clear preference for (1,4) addition. This results in carbon nanotube fluorination leading to only a C4F addition ratio at lower temperatures. This phase features isolated benzenoid rings separated by individual fluorinated carbons and is expected to show significantly different behavior from conventional graphene; for example, it should have a significant band gap with poor conductivity and exhibit localized aromatic reaction chemistry. At higher reaction temperatures, direct contiguous addition or surface fluorine migration to the (1,2) position becomes possible, leading to the close-packed C2F addition upon further fluorination. This transition can also be obtained through solvation, which lowers the migration barriers significantly. The initial state of the low-temperature fluorinated nanotubes thus consists of a nanotube fully covered by fluorine in a spaced C4F addition pattern; this is a new phase that has not yet been observed experimentally. Since solvents will rapidly lead to its degradation into C2F-fluorinated bands, we suggest that nonsolvent-based characterization may allow its identification, for example, fluorinated nanotubes which have already been distributed onto a transmission electron microscopy grid or graphite surface for atomic force microscopy or direct Raman and Fourier transform infrared spectroscopy of untreated fluorinated nanotube powder. Banded nanotubes should have many interesting properties. We note that the fluorinated bands will have a different electronic structure than the pristine segments, raising the intriguing possibility of introducing quantum wells along the nanotube length via fluorine banding. Proper control over the banding mechanism, that is, control over the extent of the fluorine migration over the surface and thus the length of the bands, will result in different properties. This has also implications for cutting fluorinated nanotubes,22 where control over the length of the bands can influence the resultant nanotube length. Banded nanotubes should also exhibit specific solubility properties, with alternating hydrophilic and hydrophobic parts, allowing for, for example, micelle formation. This opens up important new possibilities for colloid formation in material synthesis. Furthermore, a whole range of new fluorinated carbon nanotube materials can now be imagined by controlling surface coverage and banding. For example, mixing banded fluorinated nanotubes into a compatible polymer matrix should make for

Osuna et al. low nanotube density percolative networks, which nonetheless interact strongly with their host matrix since the pristine nanotube sections will have a tendency to bind together while the fluorinated segments will repel but be attracted by the host matrix. This should restrict bundling and segregation. New routes to nanotube functionalization can also be imagined due to the possibility of defluorination. For example, nanotubes could be “masked” by fluorine bands, followed by functionalization of the pristine sections. Upon careful defluorination, functionalized bands will have been created. This methodology could even be followed by functionalization by a different addition group on the defluorinated bands, creating ABAB · · · nanotubes. When electron-donating and -accepting groups are alternated, very specific electronic properties can be patterned on the nanotube surface. Finally, we note that although gas-phase fluorination has primarily been described here, fluorine banded nanotubes could also be obtained by other functionalization methods yielding low-coverage functionalization, such as RF plasma functionalization, for which subsequent heat treatment or solubilization will also result in banding. Acknowledgment. C.P.E. acknowledges the NANOSIM_ GRAPHENE project n° ANR-09-NANO-016-01 funded by the French National Agency (ANR) in the frame of its 2009 programme in Nanosciences, Nanotechnologies & Nanosystems (P3N2009). G.V.L. acknowledges the Research Foundation Flanders (FWO) for financial support as a Postdoctoral Research Fellow. S.O. acknowledges the Spanish MICINN for the doctoral fellowship No. AP2005-2992. M.T. acknowledges the European Community for financial help through the postdoctoral Grant MEIF-CT-2006-025362. C.P.E. and G.V.L. also acknowledge funding for scientific exchange between France and the Belgian Flemish community through the TOURNESOL 2009 Project T2009.19. M.S. is grateful to the Spanish MICINN and to the Catalan DIUE for financial support through projects No. CTQ2008-03077/BQU and 2009SGR637, respectively. References and Notes (1) Lee, Y. S. J. Fluorine Chem. 2007, 128, 392–403. (2) Mickelson, E. T.; Huffman, C. B.; Rinzler, A. G.; Smalley, R. E.; Hauge, R. H.; Margrave, J. L. Chem. Phys. Lett. 1998, 296, 188–194. (3) Marcoux, P. R.; Schreiber, J.; Batail, P.; Lefrant, S.; Renouard, J.; Jacob, G.; Albertini, D.; Mevellec, J. Y. Phys. Chem. Chem. Phys. 2002, 4, 2278–2285. (4) Yudanov, N. F.; Okotrub, A. V.; Shubin, Y. V.; Yudanova, L. I.; Bulusheva, L. G.; Chuvilin, A. L.; Bonard, J. M. Chem. Mater. 2002, 14, 1472–1476. (5) Giraudet, J.; Dubois, M.; Claves, D.; Pinheiro, J. P.; Schouler, M. C.; Gadelle, P.; Hamwi, A. Chem. Phys. Lett. 2003, 381, 306–314. (6) Unger, E.; Liebau, M.; Duesberg, G. S.; Graham, A. P.; Kreupl, F.; Seidel, R.; Hoenlein, W. Chem. Phys. Lett. 2004, 399, 280–283. (7) Felten, A.; Bittencourt, C.; Pireaux, J.-J.; Van Lier, G.; Charlier, J.-C. J. Appl. Phys. 2005, 98, 074308. (8) Mickelson, E. T.; Chiang, I. W.; Zimmerman, J. L.; Boul, P. J.; Lozano, J.; Liu, J.; Smalley, R. E.; Hauge, R. H.; Margrave, J. L. J. Phys. Chem. B 1999, 103, 4318–4322. (9) Bettinger, H. F. ChemPhysChem 2005, 6, 1169–1174. (10) Hamwi, A.; Gendraud, P.; Gaucher, H.; Bonnamy, S.; Beguin, F. Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A 1998, 310, 185–190. (11) Mukhopadhyay, I.; Yokoyama, Y.; Okino, F.; Kawasaki, S.; Touhara, H.; Hsu, W. K. Proc. Electrochem. Soc. 2001, 14, 37–40. (12) Peng, H. Q.; Gu, Z. N.; Yang, J. P.; Zimmerman, J. L.; Willis, P. A.; Bronikowski, M. J.; Smalley, R. E.; Hauge, R. H.; Margrave, J. L. Nano Lett. 2001, 1, 625–629. (13) Root, M. J. Nano Lett. 2002, 2, 541–543. (14) Lee, J. Y.; An, K. H.; Heo, J. K.; Lee, Y. H. J. Phys. Chem. B 2003, 107, 8812–8815. (15) Gupta, V. Nano Lett. 2004, 4, 999–999. (16) Hayashi, T. Nano Lett. 2004, 4, 1001–1002.

Graphene and Carbon Nanotube Fluorination (17) Hayashi, T.; Terrones, M.; Scheu, C.; Kim, Y. A.; Ruhle, M.; Nakajima, T.; Endo, M. Nano Lett. 2002, 2, 491–496. (18) Khabashesku, V. N.; Margrave, J. L.; Barrera, E. V. Diamond Relat. Mater. 2005, 14, 859–866. (19) Owens, F. J. J. Mater. Chem. 2006, 16, 505–508. (20) Sato, Y.; Ootsubo, M.; Yamamoto, G.; Van Lier, G.; Terrones, M.; Hashiguchi, S.; Kimura, H.; Okubo, A.; Motomiya, K.; Jeyadevan, B.; Hashida, T.; Tohji, K. ACS Nano 2008, 2, 348–356. (21) Sato, Y.; Yokoyama, A.; Kasai, T.; Hashiguchi, S.; Ootsubo, M.; Ogino, S. I.; Sashida, N.; Namura, M.; Motomiy, K.; Jeyadevan, B.; Tohji, K. Carbon 2008, 46, 1927–1934. (22) Gu, Z.; Peng, H.; Hauge, R. H.; Smalley, R. E.; Margrave, J. L. Nano Lett. 2002, 2, 1009–1013. (23) Charlier, J. C.; Blase, X.; Roche, S. ReV. Mod. Phys. 2007, 79, 677–732. (24) Dettlaff-Weglikowska, U.; Skakalova, V.; Meyer, J.; Cech, J.; Mueller, B. G.; Roth, S. Curr. Appl. Phys. 2007, 7, 42–46. (25) Bettinger, H. F. ChemPhysChem 2003, 4, 1283–1289. (26) Seifert, G.; Kohler, T.; Frauenheim, T. Appl. Phys. Lett. 2000, 77, 1313–1315. (27) Nakajima, T. J. Fluorine Chem. 1999, 100, 57–61. (28) Worsley, K. A.; Ramesh, P.; Mandal, S. K.; Niyogi, S.; Itkis, M. E.; Haddon, R. C. Chem. Phys. Lett. 2007, 445, 51–56. (29) Mori, T.; Kikuzawa, Y.; Takeuchi, H. Org. Electron. 2008, 9, 328– 332. (30) Giraudet, J.; Claves, D.; Guerin, K.; Dubois, M.; Houdayer, A.; Masin, F.; Hamwi, A. J. Power Sources 2007, 173, 592–598. (31) Bloemink, H. I.; Cooke, S. A.; Holloway, J. H.; Legon, A. C. Angew. Chem., Int. Ed. Engl. 1997, 36, 1340–1342. (32) Lee, Y. S.; Cho, T. H.; Lee, B. K.; Rho, J. S.; An, K. H.; Lee, Y. H. J. Fluorine Chem. 2003, 120, 99–104. (33) Khabashesku, V. N.; Billups, W. E.; Margrave, J. L. Acc. Chem. Res. 2002, 35, 1087–1095. (34) Kelly, K. F.; Chiang, I. W.; Mickelson, E. T.; Hauge, R. H.; Margrave, J. L.; Wang, X.; Scuseria, G. E.; Radloff, C.; Halas, N. J. Chem. Phys. Lett. 1999, 313, 445–450. (35) Van Lier, G.; Ewels, C. P.; Zuliani, F.; De Vita, A.; Charlier, J. C. J. Phys. Chem. B 2005, 109, 6153–6158. (36) Osuna, S.; Torrent-Sucarrat, M.; Ewels, C. P.; Sola`, M.; Geerlings, P.; Van Lier, G. J. Nanosci. Nanotechnol. 2009, 9, 6078–6083. (37) Ewels, C. P.; Van Lier, G.; Charlier, J. C.; Heggie, M. I.; Briddon, P. R. Phys. ReV. Lett. 2006, 96, 216103 (1-4). (38) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,

J. Phys. Chem. C, Vol. 114, No. 8, 2010 3345 X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision C.01; Gaussian, Inc.: Pittsburgh, PA, 2003. (39) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (40) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789. (41) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, J. M. J. Phys. Chem. 1994, 98, 11623–11627. (42) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213– 222. (43) Hehre, W. J.; Ditchfie, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257–2261. (44) All transition states here were characterized by computing the analytical vibrational frequencies to have one (and only one) imaginary frequency corresponding to the approach of both reacting molecules. An intrinsic reaction coordinate (IRC) study was also performed to ensure that all TSs located were connecting the expected reactants and products. (45) Osuna, S.; Houk, K. N. Chem. Eur. J. 2009, 15, 13219–13231. (46) Osuna, S.; Morera, J.; Cases, M.; Morokuma, K.; Sola`, M. J. Phys. Chem. A 2009, 113, 9721–9726. (47) Slater, J. C. Quantum Theory of Molecules and Solids; McGrawHill: New York, 1974; Vol. 4. (48) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200– 1211. (49) Peng, C.; Ayala, P. Y.; Schlegel, H. B.; Frisch, J. M. J. Comput. Chem. 1996, 17, 49–56. (50) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999. (51) An, K. H.; Heo, J. G.; Jeon, K. G.; Bae, D.; Jo, C. S.; Yang, C. W.; Park, C. Y.; Lee, Y. H.; Lee, Y. S.; Chung, Y. S. Appl. Phys. Lett. 2002, 80, 4235–4237. (52) Our B3LYP/3-21G calculations predict the (1,4) final addition product as the most favorable product by approximately 2 kcal/mol, whereas our previous ONIOM calculations indicate that the (1,2) adduct was thermodynamically the most favorable product by approximately 3 kcal/ mol. The slight difference in stability between different (1,2) and (1,4) adducts is likely to change with the method/basis set considered. Although the basis set used for the study of the migration is quite small, a qualitative description of the process is still possible.

JP908887N