Probing Distinct Fullerene Formation Processes from Carbon


Probing Distinct Fullerene Formation Processes from Carbon...

1 downloads 92 Views 2MB Size

Article pubs.acs.org/ac

Probing Distinct Fullerene Formation Processes from Carbon Precursors of Different Sizes and Structures Jong Yoon Han,† Tae Su Choi,†,‡ Soyoung Kim,‡,§ Jong Wha Lee,†,‡ Yoonhoo Ha,∥ Kwang Seob Jeong,† Hyungjun Kim,∥ Hee Cheul Choi,‡,§ and Hugh I. Kim*,† †

Department of Chemistry, Research Institute for Natural Sciences, Korea University, Seoul, 02841, Republic of Korea Department of Chemistry, Pohang University of Science and Technology (POSTECH), Pohang, 37673, Republic of Korea § Center for Artificial Low Dimensional Electronic Systems, Institute for Basic Science (IBS), Pohang 37673, Republic of Korea ∥ Graduate School of Energy, Environment, Water, and Sustainability (EEWS), Korea Advanced Institute of Science and Technology, Daejeon, 34141, Republic of Korea ‡

S Supporting Information *

ABSTRACT: Fullerenes, cage-structured carbon allotropes, have been the subject of extensive research as new materials for diverse purposes. Yet, their formation process is still not clearly understood at the molecular level. In this study, we performed laser desorption ionization-ion mobility-mass spectrometry (LDI-IM-MS) of carbon substrates possessing different molecular sizes and structures to understand the formation process of fullerene. Our observations show that the formation process is strongly dependent on the size of the precursor used, with small precursors yielding small fullerenes and large graphitic precursors generally yielding larger fullerenes. These results clearly demonstrate that fullerene formation can proceed via both bottom-up and top-down processes, with the latter being favored for large precursors and more efficient at forming fullerenes. Furthermore, we observed that specific structures of carbon precursors could additionally affect the relative abundance of C60 fullerene. Overall, this study provides an advanced understanding of the mechanistic details underlying the formation processes of fullerene.

S

form C60.15 Results from theoretical studies suggested that C2 molecules that dissociate from precursors undergo selfassembly to form large spherical shaped fullerenes by a sizeup process. The size is further adjusted by size-down processes involving C2 ejection to form stable C60 fullerenes (“shrinking hot giant model”).16−19 A recent study, based on transmission electron microscopy (TEM), presented a different view of the formation process of fullerenes.20 In this report, fullerene formation was observed to follow sequential steps: first, loss of carbon atoms at the edge of a graphene flake, and then, closure of the flake into a fullerene. Several other recent studies showed that fullerene formation occurs through shrinkage of larger carbon molecules (Cn or CnHm, n > 60).21−23 These processes can be classified as topdown, since fullerenes originate from shrinking of large fragments rather than from coalescence of small building blocks. However, despite the experimental evidence in support of top-down processes,20−23 detailed examination of the conditions that favor top-down over bottom-up processes has not been performed. In this study, we probed and characterized fullerene formation processes starting from different carbon precursors

ince the discovery of buckminsterfullerene (C60) in 1985 by Kroto, Smalley, Curl, and co-workers,1 fullerenes have been widely investigated to understand the chemistry of their spherical carbon cage structures. The physicochemical characteristics of fullerenes and their derivatives have been highlighted for potential applications as new electronic devices,2 superconductive materials,3 and antiradical agents.4,5 The applications of fullerenes have expanded further toward purposes as diverse as drug delivery,6,7 artificial photosynthesis,8 and optical limiter.9 However, despite the extensive studies of their application and characterization for 30 years,2−9 the formation mechanisms of fullerene itself are yet to be fully established. C60 formation from graphite has long been considered to proceed via the coalescence of small carbon clusters (Cn) of linear (for n ∼ 3−10) and ring (for n ∼ 5−40) structures.10,11 This concept of C60 formation has been developed into paradigms that are classified as bottom-up models. The “pentagon road” model, suggested by Smalley, describes the transformation of an open-cage fullerene of intermediate size by adding curvature through pentagon formation, enclosing the cage to form a C60 fullerene.12 “Ring stacking” model, based on stacking of carbon rings, was further proposed to explain the conversion process of open to close caged fullerenes.13,14 Meanwhile, Heath proposed another mechanism, the “fullerene road” model, in which C2 units are sequentially incorporated into close caged fullerenes of small sizes (Cn, n ∼ 40−60) to © XXXX American Chemical Society

Received: May 27, 2016 Accepted: July 19, 2016

A

DOI: 10.1021/acs.analchem.6b02076 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

with poly-DL-alanine (procured from Sigma-Aldrich, St. Louis, MO, U.S.A.) following a method reported by Thalassinos et al.25 Figure S2 shows the calibration curve of the correlation between the corrected CCS values and corrected arrival times of poly-DL-alanine ions. Arrival times of carbon cluster ions formed from various precursors were converted to the experimental CCS values based on the calibration curve. We constructed a calibration curve based on a power relationship that had been earlier reported to yield good estimates of CCS values of large molecules.26 For the theoretical models we chose sheet-structured clusters, (14,0) single-wall carbon nanotubes (SWNT; diameter ∼ 1 nm), double-wall carbon nanotubes (DWNT) composed of (8,0) and (17,0) nanotubes, and fullerenes. Theoretical CCS values of these models carrying +1 charge-state were calculated using the trajectory model in MOBCAL.27 Molecular Dynamics (MD) Simulation. MD simulations were performed to simulate fullerenes using Gromacs 4.5.5.28 The potential energy of interactions were described with OPLS-aa force field.29 The C−C bond and C−C−C angle of the carbon atoms were taken from the values for describing phenyl ring of phenylalanine. The initial structures for the simulations were prepared by placing a C180 fullerene on the surface of a model MLG/MWNT. The positions of all the atoms of MLG/MWNT were weakly restrained using a harmonic potential with a force constant, k = 10 kJ mol−1 nm−2 to represent carbon precursors that can flutter but not evaporate. The temperature of the simulations was set to 4000 K and cutoff radius for nonbonded interactions was 20 nm.

using laser desorption ionization-ion mobility-mass spectrometry (LDI-IM-MS). We found that the size distributions of fullerenes formed are significantly affected by the size of the carbon precursor used. As representatives of small carbon precursors, we utilized three polycyclic aromatic hydrocarbons (PAHs), namely, triphenylene (Tri, C18H12), anthracene (Ath, C14H10), and phenanthrene (Phn, C14H10). To represent large carbon precursors, we chose three different graphitic carbons, namely, multilayered graphene (MLG), single-walled carbon nanotube (SWNT), and multiwalled carbon nanotube (MWNT). Further, we performed LDI of C60 fullerene for comparative analyses of the distributions of the range of fullerene species formed during processing of each substrate. Our results show that graphitic carbon precursors generate large fullerenes of sizes that roughly follow a Gaussian distribution, whereas PAHs coalesce into product ions of sizes that display a decaying distribution. The differences in the distribution patterns indicate that fullerene formation mechanisms vary depending on the size of precursors. Our experiments suggest that large graphitic carbon precursors undergo processing mainly via the top-down mechanism, whereas small precursors follow bottom-up growth in which the coalescence of small molecules generates large fullerenes. Due to the need for multiple coalescence and ejection steps in the bottom-up process, the top-down process is generally more efficient. In addition, large graphitic carbon precursors showed differing abundances of C60 fullerene depending on the precursor structures, which is discussed based on the extent of reaction of carbon clusters during laser desorption processing.





RESULTS AND DISCUSSION Effect of Laser Desorption on C60 Fullerene. We examined the distribution of carbon clusters generated by LDI of C60 fullerene. The mass spectrum of the resulting carbon clusters (Figure 1A) shows singly charged C60 ions (C60+) as the dominant product, formed by direct ionization of the C60 substrate. Other species observed in the spectrum are also predominantly singly charged and found to be separated by intervals of 24 Da (Figure 1B), which corresponds to the mass of a C2 unit. This shows that the formation of even-numbered carbon clusters is favorable and that C2 unit migration is an important step in this process (Figure 1C). To verify the structures of the observed carbon clusters, we used IM-MS. The experimental arrival times of the carbon cluster ions were converted into CCS values by the method of calibration (Figure 2).25 A monotonic increase of CCS with the number of carbon atoms was observed, indicating that all the observed carbon clusters bear significant structural relationship with each other (Figure 2A). Among the various theoretical models (sheets, nanotubes, and fullerenes; Figure 2B), the experimental CCS values of the carbon clusters most closely agreed with the theoretical CCS values of fullerenes, showing that the observed carbon clusters are fullerenes. Interestingly, despite the use of C60 as a substrate, two other major distributions are observed in the mass spectrum around C118+ and C170+ (Figure 1A). The sizes of the fullerenes in these distributions are approximately double or triple the size of the original C60 substrate. Therefore, it suggests direct coalescence of C60 induces forming of dimeric and trimeric states of fullerenes of larger sizes.30,31 The decrease in the abundance of C118+ and C170+ compared to C60 can be explained by recognizing the necessity for multiple reactions for forming the higher order structures, which would occur with a low

EXPERIMENTAL SECTION Sample Preparation. Buckminsterfullerene (C60) and multiwalled carbon nanotubes (O.D. × I.D. × length: 30−50 nm × 5−15 nm × 0.5−200 μm) from Sigma-Aldrich (St. Louis, MO, U.S.A.), triphenylene, anthracene, and phenanthrene from Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan), singlewalled carbon nanotubes (mean diameter of 1.6 nm and length of 1.5 μm24) from Iljin Nanotech Co. Ltd. (Seoul, Korea) were purchased. Multilayered graphene of thickness ∼5 nm (Supporting Information (SI) and Figure S1) was provided by the Research Institute of Industrial Science and Technology (Pohang, Korea). All the carbon materials were dispersed or dissolved using methylene chloride purchased from J. T. Baker (Phillipsburg, NJ, U.S.A.). Then, aliquots of the sample were spotted on a MALDI plate and dried in the ambient air. Laser Desorption Ionization−Ion Mobility−Mass Spectrometry (LDI-IM-MS). LDI-MS was performed with Waters Synapt G2 HDMS quadrupole traveling wave ion mobility (TWIM) orthogonal time-of-flight mass spectrometer (Waters, Manchester, U.K.). For laser desorption, diode pumped Nd:YAG UV laser (355 nm wavelength) operating at 1 kHz was used at the maximum power (50 μJ pulses of 3 are considered. Inset shows a magnified view. (B) Structural models used for calculation of theoretical CCS values. Experimental CCS values of the carbon clusters were compared with the theoretical CCS values predicted for various model structures carrying +1 charge state. The result implies that the carbon clusters have the structure of fullerenes. Similar agreement was observed for carbon clusters formed from other precursors (see Figure S3).

probability. Taken together, migration of C2 units and direct coalescence of substrates are important pathways for the formation of large fullerenes from the C60 precursor. Formation of Fullerene from Small-Sized Carbon Precursors. We examined the distribution of fullerenes generated from small-sized carbon precursors. For small precursors, their coalescence is the only conceivable pathway to fullerenes, which are composed of larger numbers of the carbon atoms than those of the original precursor molecules (bottom-up). It was previously reported that fullerene can be produced from small PAHs by thermal32−34 and laser pyrolysis.35 Thus, we utilized LDI-IM-MS to investigate fullerene formation from Tri, Ath, and Phn as examples of small PAH precursors. LDI mass spectra of the three substrates (Figure 3) exhibit peaks corresponding to the substrate and to assemblies of two and three substrate molecules. This shows that direct coalescence of the precursor molecules is an important pathway for these molecules. In addition, the peaks for the hydrocarbon ions (CnHm+) were observed densely in a broad range of m/z (50−5000), represented by the dark lines

in Figure 3, suggesting that the migration of hydrocarbon fragments and ejection of H2 units also occurred actively. In the mass spectrum of Tri, C60+ was also observed in high abundance (Figure 3A). We verified that the ion is in the form of a spherical fullerene by employing IM-MS (Figure S3). Among the four most protruding peaks in the mass spectrum of Tri, three correspond to the monomeric, dimeric, and trimeric states of Tri and the other corresponds to C60. An analysis of the Tri substrate using UV/vis spectroscopy verified that C60 is not present in the original material (Figure S4). This indicates that C60 fullerene was generated by coalescence of Tri molecules. In addition, peaks of large carbon clusters were observed in the m/z range over 1000 with 24 Da intervals. C

DOI: 10.1021/acs.analchem.6b02076 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

indeed occurs for these hydrocarbon precursors. For example, generation of C36H18+ ion from Tri (C18H12; Figure 3A) requires the ejection of 3H2. However, the requirement for multiple ejection steps is suboptimal in terms of the efficiency of forming fully dehydrogenated carbon clusters. Homann and co-workers suggested that a pair of PAHs can be joined together at the periphery as a zipper, accompanied by immediate loss of all the hydrogen atoms.36,37 The asymmetric intermediary molecules so formed can then rearrange into spherical fullerenes.22 The high abundance of C60 and large fullerenes generated from Tri indicates that its structure is more readily predisposed toward the zipper mechanism than Ath and Phn. The presence of numerous armchair edges in Tri-based PAHs is most likely to favor assembly via the zipper mechanism (Figure 3D). In addition, the C−H bond energy of armchair edges (296 kJ/mol) is lower than that of zigzag edges (370 kJ/ mol),38 further explaining the efficient loss of hydrogen atoms in Tri. In contrast, due to the relatively small number of armchair edges in Ath and Phn, dehydrogenated fullerenes are formed in low abundance. In addition to the fullerene ions, the abundances of hydrocarbon clusters (dark lines in Figure 3) formed from the small-sized precursors are low and further decay following increases in cluster size. This suggests that bottom-up coalescence of small substrates is generally not efficient because the probability of generating large clusters is limited by the requirement of multiple reactions, in a similar manner to the case of the laser desorption process of C60 fullerene substrate (Figure 1). Formation of Fullerene from Graphitic Carbon Precursors. The formation of C60 fullerene and carbon clusters from large-sized precursors of graphitic carbon (MLG, SWNT, and MWNT) was investigated. In contrast to the small precursor molecules, all three of the graphitic carbon precursors yield the C60+ ion as the dominant species (Figure 4). In addition, broad distributions of carbon clusters were observed at high m/z, with different abundances for each precursor. Analogous to the case of C60+ and Tri, IM-MS experiments were employed to verify that these carbon clusters were spherical-shaped fullerenes (Figure S3). LDI mass spectrum of MLG (Figure 4A) exhibits a C60+ peak in addition to a broad distribution of large fullerene ions. This feature is similar to previously reported LDI mass spectrum of the (0001) plane of highly oriented pyrolytic graphite (HOPG).30 All the peaks of carbon cluster ions observed in Figure 4A showed mass intervals of 24 Da, indicating that C2 migration had occurred during LDI processing of graphitic precursors. LDI mass spectra of SWNT and MWNT (Figure 4B,C) also exhibit peaks of highly abundant C60+ with broad distribution of evennumbered carbon clusters (C2n), as observed in MLG. A significant difference between the LDI mass spectra of precursors, small (Figure 3) versus large (Figure 4), is the abundance of multiply charged cluster ions. Ions of +2 and +3 charge states are observed in the spectra of large precursors (Figure S6), and the sizes of these ions can be as high as 10000 Da, showing that fullerenes generated from graphitic precursors are considerably larger than those formed from small PAH molecules. Furthermore, the large-sized fullerenes from graphitic precursors exhibit roughly Gaussian distributions, as opposed to the decaying patterns observed for the ionic aggregates of PAHs. Marshall, Kroto, and co-workers have shown that large fullerenes can be formed by bottom-up processes involving addition of C2 units to closed fullerenes.39

Figure 3. LDI mass spectra of (A) triphenylene (Tri, C18H12), (B) anthracene (Ath, C14H10), and (C) phenanthrene (Phn, C14H10). Inset of (A) corresponds to hydrocarbon distributions (dark) and fullerene distribution with wider m/z intervals (gray); no artificial thickness differences are applied to the figures. Insets of (B) and (C) show hydrocarbon distributions constructed with narrow m/z intervals (dark). (D) Schematic representation of C60 fullerene formation from Tri. The assembly of planar hydrocarbon ejects peripheral hydrogen atoms via the “zipper” mechanism36,37 followed by rearrangement of their structures into spherical-shaped fullerenes.

Based on their CCS values from ion mobility measurements, they were verified as being spherical and are also thought to be formed through coalescence of Tri molecules during laser desorption process. In contrast, C60 and large fullerenes could not be detected in the LDI mass spectra of the two other PAH precursors, Ath and Phn (Figure 3B,C). The use of IM-MS allows for the detection of trace quantities of fullerene ions (Cn+ and Cn2+) formed from Ath and Phn (Figure S5). This implies that while fullerene formation from various small precursors is possible, the formation efficiencies can vary. Fullerene formation through coalescence of hydrocarbons requires the ejection of H2 or a hydrocarbon fragment,36,37 which can aid in the formation of the structural curvature necessary to produce the spherical structure of fullerenes. Our LDI mass spectra show that ejection D

DOI: 10.1021/acs.analchem.6b02076 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

“shrinking hot giant road”,16,17 proposes that C2 molecules generated by disintegration of substrates are assembled into large, closed fullerenes (size-up), which are then shrunk to form C60 (size-down).16,17 It is expected from this model that all graphitic substrates should display similar spectra including the abundance of large fullerenes because substrates will be completely disintegrated into common building blocks. However, LDI-MS study of Xie et al. showed that the structure of precursors could play a role in the formation mechanism of fullerene.30 The authors observed that perpendicular HOPG generates no fullerenes but only small carbon clusters (Cn, n < 20) by applying laser irradiation. In contrast, polycrystalline graphite generates a little abundance of C60, and HOPG generates high abundance of C60 and large fullerenes. The diversity of products indicates that graphitic substrates do not follow the same process of C2 disintegration and assembly (size-up). Therefore, it is inferred that large graphitic precursors follow an alternative pathway of fullerene formation. Our experiments suggest that top-down mechanisms are important routes to fullerene formation. For large graphitic precursors, generation of large two-dimensional graphene flakes from direct rupturing of the precursors can be a starting point for fullerene formation (Figure 4D). Migration of C2 units observed in this study (Figures 1, 3, and 4) suggests that ejection of C2 molecules from carbonaceous substrates is possible under experimental conditions. Nucleation and propagation of cracks, activated by the high temperature of laser irradiation, can be expected to promote rupturing of graphitic sheets, as has previously been observed under application of mechanical force40 or electron irradiation.41 The graphene flakes can be wrapped into large fullerenes,20 and this distinct pathway allows such large fullerenes to be prominent in the spectra of large graphitic substrates. This pathway does not require complete dissociation of the precursors into C2 units and is therefore relatively efficient compared to bottom-up pathways. Effect of Structural Differences of the Graphitic Carbon Precursors on the Formation of C60 Fullerene. The C60+ fullerene ion is observed as a magic number cluster in the mass spectra of the graphitic carbon precursors investigated in the present study. However, the relative abundances of the C60+ and other large fullerene ions are found to be different in the LDI mass spectra of the three graphitic precursors. Especially, the relative abundance of the C60 fullerene ion in the mass spectrum of MWNT is dramatically higher than that observed in the spectra of MLG and SWNT. A study based on TEM has shown that joule heating (∼2,000 K) diminishes the size of large fullerenes by continuous ejection of C2.42 Since C60 is a uniquely stable molecule in the fullerene family,18 the shrinkage is slowed down after the formation of C60.17 This indicates that the amount of C60 increases with the reaction time of fullerenes for C2 ejection. Therefore, high abundance of the C60+ ion synthesized from MWNT implies that sufficient time was allowed for the reactions generating the carbon clusters from MWNT, whereas shorter time was available for synthesis from MLG and SWNT. It is noteworthy that the relative abundance of C60+ increases with the structural complexity of graphitic precursors, from flat sheets (MLG) to monolayered cylinders (SWNT) to multilayered cylinders (MWNT; Figure 4). The high abundance of C60 fullerene in the case of MWNT (Figure 4C) suggests a correlation between C60 formation and the structure of MWNT. It was reported that fullerenes could be formed in

Figure 4. LDI mass spectra of (A) MLG, (B) SWNT, and (C) MWNT. Distributions of multiply charged fullerene ions plotted with narrow m/z intervals (dark), and distributions of singly charged fullerene ions plotted with wide m/z intervals (gray); no artificial thickness differences were applied to the figures. (D) Schematic representation of fullerene formation from graphitic carbons. Cracks initially nucleate at random points on the graphene structure due to laser irradiation. The cracks propagate to the edges inducing the formation of carbon flakes. These flakes finally wrap into fullerenes.

In their experiments, laser irradiation of fullerene-casted graphite rod ejected C2 molecules that were progressively incorporated into the precursor fullerene ions. The large-sized fullerenes generated following such a mechanism should exhibit a distribution that decays from the m/z value of the precursor fullerene used.39 This is because the involvement of multiple steps in the transformation process renders the generation of large fullerenes less probable. In contrast, the presence of large fullerenes of roughly Gaussian distributions in Figure 4 suggests a different pathway for the formation of fullerene from graphitic precursors used in this study. Another bottom-up model, the E

DOI: 10.1021/acs.analchem.6b02076 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry the inner cavity of MWNT by application of a high current.42,43 The cylindrical structure of the MWNT used in this study has wide inner and outer diameters of 5−15 and 30−50 nm, respectively. The laser (355 nm) used in our experiments can penetrate through a 20 nm thick layer of graphitic sheets without much loss of power (see the section on Discussion, SI), implying that fullerenes can be formed in the inner cavity of MWNT by laser irradiation. Under high temperatures generated by laser irradiation, the fullerenes are entrapped inside MWNT experiencing longer residence time than those molecules produced from MLG and SWNT. This may allow a larger number of fullerenes to shrink and crossover to the C60+ state. Although SWNT also has a cylindrical structure, it is composed of only a single layer and its diameter is too small (∼1.6 nm24) to accommodate fullerenes. Therefore, it is expected that the carbon flakes or fullerenes are readily formed in the outer surface of SWNT, in a manner similar to that observed with MLG. We performed MD simulations to investigate the residence time of fullerenes in MLG and MWNT (Figure 5). A model

fullerene, C180, was chosen for the MD simulations because it is one of the most abundant of species among the family of fullerenes (Figure 5). MLG was modeled as square-shaped double layers with sides of length ∼ 10 nm (Figure 5A), and the MWNT (Figure 5B) was represented as double-walled cylinders with cross-sectional diameter of ∼10 nm and length of ∼5 nm. It should be noted that the physical length of the MWNT used in the experiments was 0.5−200 μm, but for computational efficacy, the length chosen for the simulations was approximately a thousandth smaller. The MD trajectories show that C180 placed on the MLG surface readily escapes into the vacuum within 10 ps (Figure 5C,D). On the other hand, C180 positioned inside the MWNT cannot escape outside to the vacuum due to its relative immobilization from physical entrapment inside the MWNT. Its residence time increases to 80 ps. This simulation indicates that the tube structure allows large-sized fullerenes more time for transformation into C60 fullerene. Because the lateral length of MWNT used in the experiments is much greater than the simulated MWNT, the actual difference in the residence times of fullerenes inside the nanostructures are expected to be considerably greater than that observed in the simulations. Consequently, the differences in residence times resulting from structural characteristics of the carbon precursors and the extent of confinement of the intermediate products can modulate the final distribution of fullerenes generated by laser irradiation of the precursors.



CONCLUSIONS We have characterized the processes underlying fullerene formation from carbon precursors of different molecular sizes and structures using LDI-IM-MS and theoretical approaches. Our results demonstrate that small molecules form fullerene by bottom-up coalescence and C2 migration, whereas top-down processing for the synthesis of fullerene is possible from largesized graphitic carbon precursors, which are found to produce distinct features in their mass spectra. The requirement for multiple coalescence and adequate structures limits the efficiency of bottom-up processes when compared to topdown processes. In addition, we observed that transformation of large fullerenes into C60 fullerene via progressive shrinkage becomes feasible as reaction time increases. These results suggest a possibility for controlling the yield of fullerenes by modulating the size and structure of the carbon precursors. Detailed mechanistic, structural, and kinetic insights will further aid the development of efficient processing routes for fullerene production.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b02076. Additional discussion, experimental data, and supporting figures (PDF).



Figure 5. Snapshots from MD simulations of a fullerene positioned on (A) MLG and (B) MWNT. (C) Z-axis distance between the center of mass of fullerene and MLG/MWNT. The escape time of fullerene is shorter in MLG than MWNT. (D) Interaction energy between fullerene and MLG/MWNT. The interaction energy between the fullerene and precursor diminishes with time as the fullerene moves away from the MLG/MWNT.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. F

DOI: 10.1021/acs.analchem.6b02076 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry



(27) Mesleh, M. F.; Hunter, J. M.; Shvartsburg, A. A.; Schatz, G. C.; Jarrold, M. F. J. Phys. Chem. 1996, 100, 16082. (28) Hess, B.; Kutzner, C.; Van Der Spoel, D.; Lindahl, E. J. Chem. Theory Comput. 2008, 4, 435. (29) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. J. Am. Chem. Soc. 1996, 118, 11225. (30) Xie, Z. X.; Liu, Z. Y.; Wang, C. R.; Huang, R. B.; Lin, F. C.; Zheng, L. S. J. Chem. Soc., Faraday Trans. 1995, 91, 987. (31) Budyka, M. F.; Zyubina, T. S.; Ryabenko, A. G.; Muradyan, V. E.; Esipov, S. E.; Cherepanova, N. I. Chem. Phys. Lett. 2002, 354, 93. (32) Taylor, R.; Langley, G. J.; Kroto, H. W.; Walton, D. R. Nature 1993, 366, 728. (33) Osterodt, J.; Zett, A.; Vögtle, F. Tetrahedron 1996, 52, 4949. (34) Crowley, C.; Kroto, H. W.; Taylor, R.; Walton, D. R.; Bratcher, M. S.; Cheng, P.-C.; Scott, L. T. Tetrahedron Lett. 1995, 36, 9215. (35) Armand, X.; Herlin, N.; Voicu, I.; Cauchetier, M. J. Phys. Chem. Solids 1997, 58, 1853. (36) Baum, T.; Löffler, S.; Löffler, P.; Weilmünster, P.; Homann, K. H. Ber. Bunsen-Ges. Phys. Chem. 1992, 96, 841. (37) Ahrens, J.; Bachmann, M.; Baum, T.; Griesheimer, J.; Kovacs, R.; Weilmünster, P.; Homann, K.-H. Int. J. Mass Spectrom. Ion Processes 1994, 138, 133. (38) Serp, P.; Figueiredo, J. L. Carbon Materials for Catalysis; John Wiley & Sons, 2009. (39) Dunk, P. W.; Kaiser, N. K.; Hendrickson, C. L.; Quinn, J. P.; Ewels, C. P.; Nakanishi, Y.; Sasaki, Y.; Shinohara, H.; Marshall, A. G.; Kroto, H. W. Nat. Commun. 2012, 3, 855. (40) Budarapu, P.; Javvaji, B.; Sutrakar, V.; Roy Mahapatra, D.; Zi, G.; Rabczuk, T. J. Appl. Phys. 2015, 118, 064307. (41) Kim, K.; Artyukhov, V. I.; Regan, W.; Liu, Y.; Crommie, M.; Yakobson, B. I.; Zettl, A. Nano Lett. 2012, 12, 293. (42) Huang, J. Y.; Ding, F.; Jiao, K.; Yakobson, B. I. Phys. Rev. Lett. 2007, 99, 175503. (43) Huang, J. Y.; Chen, S.; Jo, S. H.; Wang, Z.; Han, D. X.; Chen, G.; Dresselhaus, M. S.; Ren, Z. F. Phys. Rev. Lett. 2005, 94, 236802.

ACKNOWLEDGMENTS We gratefully acknowledge supply of MLG from Research Institute of Industrial Science and Technology (Pohang, Korea). This work was supported by the Basic Research Program (Grant No. NRF-2016R1A2B4013089) through the National Research Foundation (NRF) of Korea funded by the Ministry of Science, ICT, and Future Planning (MSIP), the Basic Science Research Program through the National Research Foundation (NRF) of Korea funded by the Ministry of Education (Grant No. 20100020209), a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare of Korea (Grant No. HT13C0011-040013), and the Institute for Basic Science (IBS; IBSR014) in Korea. T.S.C. acknowledges support from TJ Park Fellowship.



REFERENCES

(1) Kroto, H. W.; Heath, J. R.; O’Brien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985, 318, 162. (2) Prato, M. J. Mater. Chem. 1997, 7, 1097. (3) Haddon, R. C. Acc. Chem. Res. 1992, 25, 127. (4) McEwen, C. N.; McKay, R. G.; Larsen, B. S. J. Am. Chem. Soc. 1992, 114, 4412. (5) Xiao, L.; Takada, H.; Gan, X. H.; Miwa, N. Bioorg. Med. Chem. Lett. 2006, 16, 1590. (6) Bakry, R.; Vallant, R. M.; Najam-ul-Haq, M.; Rainer, M.; Szabo, Z.; Huck, C. W.; Bonn, G. K. Int. J. Nanomed. 2007, 2, 639. (7) Montellano, A.; Da Ros, T.; Bianco, A.; Prato, M. Nanoscale 2011, 3, 4035. (8) Garg, V.; Kodis, G.; Chachisvilis, M.; Hambourger, M.; Moore, A. L.; Moore, T. A.; Gust, D. J. Am. Chem. Soc. 2011, 133, 2944. (9) Kojima, Y.; Matsuoka, T.; Takahashi, H.; Kurauchi, T. Macromolecules 1995, 28, 8868. (10) von Helden, G.; Hsu, M.-T.; Gotts, N.; Bowers, M. T. J. Phys. Chem. 1993, 97, 8182. (11) von Helden, G.; Gotts, N.; Bowers, M. T. J. Am. Chem. Soc. 1993, 115, 4363. (12) Smalley, R. E. Acc. Chem. Res. 1992, 25, 98. (13) Wakabayashi, T.; Achiba, Y. Chem. Phys. Lett. 1992, 190, 465. (14) Hunter, J. M.; Fye, J. L.; Roskamp, E. J.; Jarrold, M. F. J. Phys. Chem. 1994, 98, 1810. (15) Heath, J. R. ACS Symp. Ser. 1992, 481, 1. (16) Irle, S.; Zheng, G.; Elstner, M.; Morokuma, K. Nano Lett. 2003, 3, 1657. (17) Irle, S.; Zheng, G.; Wang, Z.; Morokuma, K. J. Phys. Chem. B 2006, 110, 14531. (18) Curl, R. F.; Lee, M. K.; Scuseria, G. E. J. Phys. Chem. A 2008, 112, 11951. (19) Saha, B.; Irle, S.; Morokuma, K. J. Phys. Chem. C 2011, 115, 22707. (20) Chuvilin, A.; Kaiser, U.; Bichoutskaia, E.; Besley, N. A.; Khlobystov, A. N. Nat. Chem. 2010, 2, 450. (21) Berne, O.; Tielens, A. G. G. M. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 401. (22) Zhang, J.; Bowles, F. L.; Bearden, D. W.; Ray, W. K.; Fuhrer, T.; Ye, Y.; Dixon, C.; Harich, K.; Helm, R. F.; Olmstead, M. M. Nat. Chem. 2013, 5, 880. (23) Zhen, J.; Castellanos, P.; Paardekooper, D. M.; Linnartz, H.; Tielens, A. G. Astrophys. J., Lett. 2014, 797, L30. (24) Nirmalraj, P. N.; Lyons, P. E.; De, S.; Coleman, J. N.; Boland, J. J. Nano Lett. 2009, 9, 3890. (25) Thalassinos, K.; Grabenauer, M.; Slade, S. E.; Hilton, G. R.; Bowers, M. T.; Scrivens, J. H. Anal. Chem. 2009, 81, 248. (26) Scarff, C. A.; Thalassinos, K.; Hilton, G. R.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2008, 22, 3297. G

DOI: 10.1021/acs.analchem.6b02076 Anal. Chem. XXXX, XXX, XXX−XXX