Designing Novel Materials through Functionalization of Carbon

ARTICLE pubs.acs.org/JPCA

Designing Novel Materials through Functionalization of Carbon Nanotubes for Application in Nuclear Waste Management: Speciation of Uranyl Mahesh Sundararajan† and Swapan K Ghosh*,†,‡ † ‡

Theoretical Chemistry Section, Bhabha Atomic Research Centre, Mumbai 400 085, India Homi Bhabha National Institute, Mumbai, India ABSTRACT: Understanding the behavior of radioactive nuclide elements in different environmental conditions is an active area of research. In this work, we have investigated the possible interaction mechanism between carbon nanotubes and uranyl using density functional theory. It is shown that functionalized carbon nanotubes can be used to bind uranyl ions much more efficiently as compared to their unfunctionalized counterpart. The uranyl binding energies are sensitive to the nature of the functional groups rather than the carbon nanotube itself. The binding takes place preferably at the functionalized sites, although pH could determine the strength of uranyl binding. Our predicted results correlate well with the recent experimental uranyl sorption studies on carbon nanotubes. These finding are new and can open up a new era for actinide speciation and separation chemistry using carbon nanotubes.

’ INTRODUCTION With ever-increasing energy demands, nuclear power becomes one of the major sources of energy.1 Besides the radioactive hazards and the high cost of equipping laboratories for work, the chemistry of actinides gained significant attention due to their catalytic properties.2
5 It is widely known that uranyl species can be used for catalyzing many reactions. However, on the other hand, the chemistry of uranium and other transuranics is receiving increased attention due to the environmental danger posed by such species.6 One possible way to control their release into groundwater is to take advantage of the differing solubility of their different oxidation states.7,8 Strategies for reducing the mobility of these species usually center on their reduction to less soluble, lower oxidation state species, which, in the case of uranium, involves reduction of soluble U(VI) to insoluble U(IV). This process is facilitated by multiheme cytochromes,9,10 humic substances,11 and iron-containing mineral surfaces.12 Carbon nanotubes (CNTs) are novel and interesting graphitic carbon materials that, since their discovery, have attracted considerable attention due to their unique structural and physiochemical properties.13 Very recently, researchers became interested in knowing whether CNTs can play a role in the field of nuclear waste management.14 Some experimental strudies were carried out to understand the interactions between nanotubes and metal ions.15
23 Interactions of chromium, lead, copper, and cadmium ions with CNTs have been recently investigated.15
19 Thamavaranukup et al.20 reported that uranyl acetates can be filled inside of single walled (SW) CNT, whereas Wang et al.21 claim that an americium molecule (AmIII) is sorbed to multiwalled (MW) CNTs. However, in both cases, the exact r 2011 American Chemical Society

nature of binding is not known for these species. Wang et al.22 proposed that CNTs can be used for the recovery of radionuclides and separation of actinides and lanthanides. Recently, Schierz and Zanker23 investigated the uranium sorption in surface-oxidized CNTs and found an increase in uranyl sorption at functionalized SW-CNTs in semineutral pH as compared to that at the unfunctionalized form and at low pH. However, poor solubility of the CNT restricts its use in an aqueous environment. Fortunately, upon functionalization of the CNT, one can overcome the poor solubility of the CNT.24
29 In fact, Wong and co-workers have shown that the solubility of CNTs can be modified when organic and inorganic groups are attached to oxidized CNTs.27
29 There has been a large increase in the number of computational investigations of actinide complexes, particularly those of uranium.30
41 The electronic structure and molecular geometry of the dioxo cation of these three elements have been studied at the Hartree
Fock (HF),35 density functional theory (DFT),32,36 and correlated ab initio levels.37 Theoretical studies have also modeled these cations, especially uranyl, using various methods for the inclusion of aqueous solvation effects including both variants of continuum models and by explicitly including the solvent molecules.39,40 The most studied structures are those that arise when the uranyl ion is hydrated, in particular, UO2(H2O)n2þ (n = 4, 5, and 6), with the penta-aqua being generally preferred to be the dominant species.24,30,33,37,38 The effect of bulk solvent is usually included via a continuum model, but there Received: April 21, 2011 Published: May 20, 2011 6732

dx.doi.org/10.1021/jp203723t | J. Phys. Chem. A 2011, 115, 6732–6737

The Journal of Physical Chemistry A

ARTICLE

Figure 2. HOMO of an (a) armchair SW-CNT and (b) functionalized armchair SW-CNT in solution (COSMO) derived from the B3LYP functional.

Figure 1. Optimized structure of a functionalized SW-CNT.

is an increasing interest in defining the actual nature of the interaction with the nearest solvent molecules.40 Schlosser et al.41 have performed a study of monocarboxylate complexes of uranyl, comparing predicted structures with EXAFS data. On the CNT side, major computational studies were devoted to understand the enhanced reactivities of certain molecules upon confinement.42
45 Due to the importance of dispersion effects, binding of pollutant aromatics on CNTs have been investigated by Ramraj et al.46 Simeon et al.47 investigated the interactions of indium ions with the defect sites of SW-CNTs. In this paper, we present the first theoretical investigation of uranyl binding motifs to SW-CNTs using DFT and propose various uranyl binding sites to the SW-CNT. In particular, we have functionalized the CNT with carboxylate functional groups at the tail and also at the wall, such that the binding affinity of uranyl species to CNTs increases drastically. Additionally, we have also investigated the sensitivity of uranyl binding energies to (a) different CNT lengths, (b) diameters, (b) the influence of defects at the CNT site, (d) other functional groups, and (e) two different CNT tubes (armchair and zigzag).

’ COMPUTATIONAL DETAILS The armchair (8, 8) conformations (11.4 Å diameter) with four unit cell lengths (10.5 Å length) and penta-aquo uranyl ([UO2(H2O)5]2þ) are chosen as models for the SW-CNT and uranyl, respectively. This uranyl complex [UO2(H2O)5]2þ was inserted into the armchair SW-CNT with three varying diameters chosen to be (7,7), (8,8), and (9,9). The dangling bonds were saturated by adding hydrogen atoms. If the diameter of the SWCNT is less than ∼11 Å (as in the (7,7) SW-CNT case), dissociation of the uranyl water complex is noticed due to the strain imposed by the SW-CNT. However, if the diameter of the SW-CNT is more than ∼11 Å (as in (8,8) and (9,9) SW-CNT cases), no severe geometric distortions take place. In the encapsulated adduct, the interaction between the waters bonded to uranyl and those of carbons of the SW-CNT are noncovalent. If the SW-CNT diameter is very large, the confinement stabilization will be decreased, and the uranyl species will behave similar to that in the gas phase. Further, we find that the binding energies are similar for both (8,8) and (9,9) SW-CNTs (70 and 77 kcal mol
1 for (8,8) and (9,9) SW-CNT in the gas phase). Hence, for computational efficiency, all of the uranyl binding studies are reported here for (8,8) only. The SW-CNT is

functionalized both at the tail and wall using ethyl acetate. Such ethyl acetate functionalization to the SW-CNT is experimentally feasible.48 Further, it has been known that 1
3 functional groups can be present per 100 atoms (Figure 1).24 All of the geometries are optimized using the BP86 functional in conjunction with def2-SV(P) basis sets. For uranium, the valence electrons are treated using the def-SV(P) basis set, whereas the core electrons are modeled with the def-ECP pseudopotential. For energetics, we have used the B3LYP functional in conjunction with the TZVP basis for the uranyl complex including ethyl acetate groups, whereas the SW-CNT is treated with def2-SV(P) only (2384 Cartesian basis functions). Long-range solvation effects are incorporated using COSMO continuum solvation model (using ε = 80) as implemented in TURBOMOLE.49 Such a costeffective strategy saves computational time, and the predicted structures, vibrational spectra, and redox properties are comparable to the experimental data.40 The binding energies are evaluated by considering the processes as given by ½Functionalized-SW-CNT
þ ½UO2 ðH2 OÞ5 2þ f ½Functionalized-SW-CNT-½UO2 ðH2 OÞ5 þ ½Functionalized-SW-CNT
þ ½UO2 ðH2 OÞ5 2þ f ½Functionalized-SWNT-½UO2 ðH2 OÞ3 þ þ ðH2 OÞ2 The uranyl species can bind to this functionalized SW-CNT at the tail or at the wall via covalent bonding with carboxylate groups by displacing two water molecules or by encapsulation inside of the functionalized SW-CNT without any chemical bond-making or bond-breaking processes.

’ RESULTS AND DISCUSSION Due to the introduction of carboxylate functional groups, the p
π orbitals of highest occupied molecular orbital (HOMO) of the functionalized SW-CNT is distorted as compared to the SWCNT (Figure 2). Such an electronic structure is expected upon introduction of functional groups to the SW-CNT. Additionally, the gap between the HOMO and lowest unoccupied molecular orbital (LUMO) of the functionalized SW-CNT (0.90 eV) is reduced as compared to the unfunctionalized SW-CNT (1.07 eV). These values indicate that the reactivity of the SWCNT can be altered by proper functionalization. In comparison to the “free” form and irrespective of the binding sites, an elongation in the UdO bond (by 0.03 Å) is noted for all three bound structures (Table 1). Further, the 6733

dx.doi.org/10.1021/jp203723t |J. Phys. Chem. A 2011, 115, 6732–6737

The Journal of Physical Chemistry A

ARTICLE

, bond angle, °), Mulliken Charge, q (e
), and Binding Energies, BE (kcal mol
1), of Table 1. Optimized Structures (bond length Å a Uranyl Species [UO2(H2O)5]2þ bound at

Unbound [UO2(H2O)5]2þ UdO

encapsulated

1.77

1.80 (1.80)

U
OCOO

a

2.43
2.49 (2.44
2.48) 177.2 (176.7)

tail

wall

1.81

1.80

2.40
2.43

2.40
2.41

2.52
2.54 173.6

2.51
2.56 174.0

U
OH2O OdUdO

2.47 180.0

qU

þ1.386

1.295 (1.275)

þ1.131

þ1.123

qO


0.193


0.224 (
0.258)


0.245


0.246

BEGas

0.0


154.5 (
68.5)


193.1


195.0

BECOSMO

0.0

þ54.6 (þ43.8)


20.8


20.8

Values in parentheses are for unfunctionalized (8,8) SW-CNT.

Figure 3. Optimized structure of uranyl bound to a functionalized armchair SW-CNT.

OdUdO bond angle is slightly bent in bound structures but was 180° in the free form, and the extent of bending is large for tailand wall-bound uranyl structures (6°) as compared to that for the encapsulated structure (3°) (Figure 3). Similarly for both walland tail-bound uranyl structures, the U
C bond lengths are very similar (∼2.76 Å). The optimized uranyl complexes at the wall and tail are similar to those of the isolated structures of the uranyl-bound ethyl acetate molecule. Additionally, we also notice the U
OH2O bond length not to be as symmetric as that in the isolated structure for all three binding sites. The water molecules bound to the uranyl complex upon encapsulation are not planar but are slightly tilted due to encapsulation. However, the calculated strain energy for the tilted structure is only 1.2 kcal mol
1 higher than that of the isolated molecule. A similar structure is also noticed for the encapsulated adduct of the unfunctionalized SW-CNT. Further, there is a significant amount of charge transfer from both functionalized and unfunctionalized SW-CNTs to the uranyl moiety, resulting in the elongation of the UdO bond (Table 1). The extent of charge transfer to uranyl is somewhat larger for wall-and tail-bound adducts as compared to that for the encapsulated adduct. It may be noted that the Mulliken charges are not very reliable for charge distribution in many cases. However, here, we are particularly interested in the charge shift at the uranyl centers of bound structures as compared to that of unbound structures. Indeed, we have evaluated the charge shift using other population methods such as natural population analysis (NPA) charges as implemented in TURBOMOLE. Although the absolute charges are different, the values of the change in charge at the uranyl centers between the bound and unbound structures, as calculated by different schemes, are very similar.

The calculated binding energies are very sensitive to environment effects as incorporated via the continuum solvation model.41,50 The formation of functionalized SW-CNT uranyl adducts is strongly favorable by more than 150 kcal mol
1 in the gas phase. However, in the solution phase, only the wall- and tailbound uranyl adducts are favorable (∼21 kcal mol
1). For the corresponding encapsulated adduct formation, it is unfavorable by more than 54 kcal mol
1 in the solution phase. It is interesting to note that in the gas phase, the binding of uranyl to functionalized SW-CNTs is favorable by more than 80 kcal mol
1 as compared to that of unfunctionalized SW-CNTs. This may be partly due to the difference in the overall charge of the CNT prior to binding. Due to functionalization, the overall negative charge of the SW-CNT becomes larger, which can bind to the positively charged uranyl complex more strongly as compared to the neutral form. However, in the solution phase, the formation of an encapsulated adduct of the unfunctionalized SW-CNT is less unfavorable as compared to that of its functionalized SW-CNT counterpart by ∼10 kcal mol
1. Nevertheless, between the encapsulated bound uranyl and covalently bound complex (both at the wall and at the tail), the binding at the carboxylate sites is favorable by 40 and 70 kcal mol
1 in the gas phase and in the solution phase. In our calculations, the COSMO binding energies for the encapsulated adduct are largely unfavorable. A comment has to be made on the calculated COSMO energetics. The solvation free energy of the ion is proportional to q2/d, where q is the charge and d is the radius of a sphere encompassing the ion in the medium. The small molecule [UO2(H2O)5]2þ dication is found to have a solvation free energy of
192 kcal mol
1. If one were to double the radius (through SW-CNT encapsulation), then the solvation free energy would drop nealy by half (
92 kcal mol
1). 6734

dx.doi.org/10.1021/jp203723t |J. Phys. Chem. A 2011, 115, 6732–6737

The Journal of Physical Chemistry A

ARTICLE

Figure 4. Optimized structure of a (a) bare Stone
Wales defect SW-CNT and (b) uranyl bound at the oxidative damaged SW-CNT.

, bond angle, °) and Binding Energies, BE (kcal mol
1) of Uranyl Species in Various Table 2. Optimized Structures (bond length Å SW-CNTs CNT type

unit cell

UdO

U
OCOO

U
OH2O

OdUdO

BECOSMO

SW-CNT (6, 6)

armchair

4

1.80

2.41

2.50
2.53

174.6


23.5

SW-CNT (6, 6)

armchair

5

1.80

2.41

2.51
2.54

173.2


21.7

SW-CNT (6, 6)

armchair

6

1.80

2.41

2.51
2.53

174.4


22.2

SW-CNT (7, 7)

armchair

4

1.80

2.41

2.51
2.53

173.7


22.8

SW-CNT (8, 8)

armchair

4

1.80

2.40
2.43

2.51
2.53

174.2


22.9

SWD-SWCNT(8, 8)

armchair

4

1.80

2.41

2.51
2.53

174.2


21.5

SW-CNT (8, 0) SW-CNT (8, 8)-oxo

zigzag armchair

4 4

1.81 1.82

2.40
2.43 2.13

2.51
2.53 2.54
2.57

173.9 171.4


24.2
44.0

That change is quite consistent with the change in binding free energy from the gas phase (
68 kcal mol
1) to the solution phase (þ44 kcal mol
1). A similar observation was reported by Shamov et al. upon actinyl binding to an 18-crown-6 host molecule.50 We have assessed the uranyl binding strength with respect to other solvents. We have calculated the binding energy using three different dielectric constants (80 for water, 37 for acetonitrile, and 5 for chloroform). As expected, the strength of uranyl binding increases with decreasing dielectric constant of the solvents due to the relative destabilization of the functionalized SW-CNT prior to uranyl binding. The results presented herein reveal the nature of binding motifs of uranyl when bound to CNT. It is typical that FT-IR and Raman spectroscopic measurements are used as fingerprints to identify the symmetric and asymmetric uranyl (UdO) stretching frequencies. We believe that if the uranyl is encapsulated or covalently linked to functionalized CNTs, the UdO stretching frequencies are significantly red-shifted as compared to the unbound uranyl species. For instance, the symmetric UdO stretching frequency for the uranyl acetate species was reported51 to be 861 cm
1, whereas for the penta-aqua uranyl complex,52 this value is 870 cm
1. Although the shift is only 9 cm
1, this frequency shift can be probed using advanced spectroscopic techniques and should be useful as vibrational markers to understand the binding behavior in CNTs. However, it is known that if the diameter of the CNT is significantly large or very small as compared to the van der Waals radii of the uranyl species, the binding of uranyl will be via carboxylates rather than encapsulation. Further, if the functionalized SW-CNT is immersed in solution of very low pH, a competitive binding of uranyl and protons with carboxylates is expected.11 The calculated proton affinity at the wall or at the tail is ∼290 kcal mol
1 (in the solution phase), which is ∼25 kcal mol
1 higher than that of the

hydration free energy of the proton determined experimentally (264 kcal mol
1).53 Schierz and Zanker23 carried out uranyl sorption experiments on various CNTs by varying the pH of the solution (2
11). They observed that uranyl sorption is at a maximum for surfaceoxidized CNTs at the pH 4
8 range. When the pH of the solution is increased from 2 to 11, a decrease in the zeta potential from
10 to
35 mV is noted. The negative charges on the surface (due to deprotonation of carboxylates) can bind uranyl strongly as compared to unfunctionalized CNTs. Our calculations also suggest that binding energies at the wall or at the tail are much larger (
21 kcal mol
1) when carboxylates are deprotonated as compared to those of encapsulated (þ54.6 kcal mol
1) or protonated carboxylate forms (þ19.6 kcal mol
1). Thus, these values suggest that, depending on the pH of the solution, the binding of uranyl can be tuned. Hence, in such a low pH scenario, uranyl binding to the CNT will be dominantly physisorption (or very weak chemisorption), whereas at near-neutral pH, the uranyl binds at carboxylates of the CNT. Although our calculations favor uranyl binding at both the functionalized tail and wall, uranyl binding will occur only at the tail-functionalized site due to the ease of functionalization at the edges. Functionalization at the wall will occur only at the defect site such as the Stone
Wales type.29 In this regard, we have modeled the Stone
Wales defect (SWD), whose uranyl optimized structure is shown in Figure 4a. Oxidative damage at the SWD site creates the wall-functionalized species,29 where uranyl binding can take place (Figure 4b). The calculated geometries and binding energy (
21.5 kcal mol
1) are similar to those of the defect-free site (Table 2). Further, Matsuo et al.54 reported that the chemical reactivities of the finite-length tube change periodically as the tube length is elongated. Particularly, finite modeling of the CNT corresponds 6735

dx.doi.org/10.1021/jp203723t |J. Phys. Chem. A 2011, 115, 6732–6737

The Journal of Physical Chemistry A

ARTICLE

functionalized site. We have considered a five-coordinated species, where four waters and an oxo group are coordinated in the equatorial positions of the uranyl ion (Figure 6). The ) is shorter as compared to uranyl
oxo bond length (2.14 Å ). The optimized the uranyl
carboxylate bond length (2.40 Å UdO bond length at the oxo-functionalized site is slightly longer ) as compared to the carboxylated armchair SW-CNT (∼0.01 Å (Figure 6; Table 2). The calculated binding energy is twice (44 kcal mol
1) as large as that of the carboxylated adduct. Hence, the functional group attached to the CNT plays a major role in both solubilizing the CNT and modifying the uranyl binding strengths.

Figure 5. Optimized structure of uranyl bound to a functionalized zigzag SW-CNT.

’ CONCLUSIONS We have addressed here the possible speciation of uranyl at various SW-CNTs. Our calculations have addressed many issues and proposed various binding sites of uranyl at the SW-CNT. We have shown that by proper functionalization, the binding affinities of uranyl to the SW-CNT can be enhanced significantly. Our calculations suggest that binding of uranyl at the peripherals is preferred over encapsulation, although pH plays a vital role in uranyl sorption, as observed by Schierz and Zanker.23 The calculated binding energies of uranyl to the CNT at different lengths and diameters are essentially the same. However, the uranyl binding affinities can be altered by different functionalization, such as an oxo group, which was used here. Hence, the strength of uranyl binding depends on the actual functional groups rather than the nature of the SW-CNT. We believe that the present study is extremely useful for understanding the speciation of the uranyl ion at the CNT at the molecular level. We are working in this direction for other radionuclides in our computational laboratory. ’ AUTHOR INFORMATION

Figure 6. Optimized structure of uranyl bound to oxo-functionalized SW-CNT.

to three different classes referred to as Kekule, incomplete Clar, and Clar networks. In this regard, the sensitivity of uranyl binding of tail-functionalized SW-CNTs with respect to different unit cell lengths and diameters was carried out. To ease the computation, we have taken the (6,6) SW-CNT using four, five, and six unit cell lengths. Although the chemical reactivities are sensitive to the actual finite length of the CNT, the uranyl binding energies (∼21 kcal mol
1) were largely unaffected (Table 2). We have also investigated the effect of curvature of the SWCNT by functionalizing the tail of different diameters of the armchair CNT at a four unit cell lengths. Here again, we find that both geometries at the uranyl site and the binding energies are largely unaltered. As most nanotube samples are mixtures of armchair and zigzag CNTs, we have also considered the uranyl binding at the zigzag CNT. Here, we have functionalized the (8,0) zigzag SW-CNT using four unit cell lengths. Although the optimized structure at the uranyl site is similar to that of armchair type (Figure 5), the actual binding energy is slightly larger by 2 kcal mol
1 as compared to that of the armchair type (Table 2). Finally, we have functionalized the armchair (8,8) SW-CNT with an oxo group at the tail whose structures and binding energies are somewhat different from those of the carboxylated

Corresponding Author

*E-mail: [email protected]. Tel.: þ91 22 25595092. Fax: þ91 22 25505151.

’ ACKNOWLEDGMENT We thank the KSKRA fellowship for funding and Dr. T. Mukherjee for his kind support and the BARC computer center for providing the high-performance paraller computing facility (Ameya and Ajeya Systems). Funding from the INDO-EU project MONAMI is also gratefully acknowledged. ’ REFERENCES (1) Katz, J. J.; Morss, L. R.; Seaborg, G. T. The Chemistry of the Actinide Elements, 2nd ed.; Chapman and Hall: London, 1986. (2) van Axel Castelli, V.; Dalla Cort, A.; Mandolini, L. J. Am. Chem. Soc. 1998, 120, 12688. (3) Barnea, E.; Eisen, M. S. Coord. Chem. Rev. 2006, 250, 855. (4) Fox, A. R.; Bart, S. C.; Meyer, K.; Cummins, C. C. Nature. 2008, 455, 341. (5) Lin, Z.; Marks, T. J. J. Am. Chem. Soc. 1990, 112, 5515. (6) Taylor, D. M. Rev. Environ. Health 1997, 12, 147. (7) Anderson, R. T.; Vrionis, H. A.; Ortiz-Bernad, I.; Resch, C. T.; Long, P. E.; Dayvault, R.; Karp, K.; Marutzky, S.; Metzler, D. R.; Peacock, A.; White, D. C.; Lowe, M.; Lovley, D. R. Appl. Environ. Microbiol. 2003, 69, 5884. 6736

dx.doi.org/10.1021/jp203723t |J. Phys. Chem. A 2011, 115, 6732–6737

The Journal of Physical Chemistry A (8) Lovley, D. R.; Phillips, E. J. P.; Gorby, Y. A.; Landa, E. Nature. 1991, 350, 413. (9) Renshaw, J. C.; Butchins, L. J. C.; Livens, F. R.; May, I.; Charnock, J. M.; Lloyd, J. R. Environ. Sci. Technol. 2005, 39, 5657. (10) Sundararajan, M.; Campbell, A. C.; Hillier, I. H. J. Phys. Chem. A. 2008, 112, 4451. (11) Shanbhag, P. M.; Chopping, G. R. J. Inorg. Nucl. Chem. 1981, 43, 3369. (12) Lefevre, G.; Noinville, S.; Fedoroff, M. J. Colloid Interface Sci. 2006, 296, 608. (13) Niyogi, S.; Hamon, M. A.; Hu, H.; Zhao, B.; Bhowmik, P.; Sen, R.; Itkis, M. E.; Haddon, R. C. Acc. Chem. Res. 2002, 35, 1105. (14) Belloni, F.; Kutahyali, C.; Rondinella, V.; Carbol, P.; Wiss, T. Environ. Sci. Technol. 2009, 43, 1250. (15) Hu, J.; Chen, C.; Zhu, X.; Wang, X. J. Hazard. Mater. 2009, 162, 1542. (16) Moradi, O.; Zare, K.; Yari, M. Int. J. Nano Dimens. 2011, 1, 203. (17) Ruparelia, J. P.; Duttagupta, S. P.; Chatterjee, A. K.; Mukherji, S. Dasalination 2008, 232, 145. (18) Atieh, M. A.; Bakather, O. Y.; Tawabini, B. S.; Bukhari, A. A.; Khaled, M.; Alharthi, M.; Fettouhi, M.; Abuilaiwi, F. A. J. Nanomater. 2010, 1. (19) Moradi, O.; Zare, K.; Monajjemi, M.; Aghaie, H. Fullerenes, Nanotubes, Carbon Nanostruct. 2010, 18, 285. (20) Thamavaranukup, N.; Hoppe, H. A.; Ruiz-Gonzalez, L.; Costa, P. M. F. J.; Sloan, J.; Kirkland, Green, M. L. H. Chem. Commun. 2004, 1686. (21) Wang, X.; Chen, C.; Hu, W.; Ding, A.; Xu, D.; Zhou, X. Environ. Sci. Technol. 2005, 39, 2856. (22) Wang, P.; Patent Article. IPC8 Class: AG21F906FI, USPC, 588 18. (23) Schierz, A; Zanker, H. Environ. Pollut. 2009, 157, 1088. (24) Balasubramanian, K.; Burghard, M. Small. 2005, 1, 180. (25) Sun, Y.-P.; Fu, K.; Lin, Y.; Huang, W. Acc. Chem. Res. 2002, 35, 1096. (26) Tasis, D.; Tagmatarchis, N.; Georgakilas, V.; Prato, M. Chem.— Eur. J. 2003, 9, 4000. (27) Kahn, M. G. C.; Banerjee, S.; Wong, S. S. Nano Lett. 2002, 2, 1215. (28) Banerjee, S.; Wong, S. S. J. Am. Chem. Soc. 2002, 124, 8940. (29) Hemraj-Benny, T.; Banerjee, S.; Wong, S. S. Chem. Mater. 2004, 16, 1855. (30) Vallet, V.; Wahlgren, U.; Schimmelpfennig, B.; Szabo, Z.; Grenthe, I. J. Am. Chem. Soc. 2001, 123, 11999. (31) Schreckenbach, G.; Hay, P. J.; Martin, R. L. J. Comput. Chem. 1999, 20, 70. (32) Kaltsoyannis, N. Coord. Soc. Rev. 2003, 32, 9. (33) Buhl, M.; Diss, R.; Wipff, G. J. Am. Chem. Soc. 2005, 127, 13506. (34) Tsushima, S.; Wahlgren, U.; Grenthe, I. J. Phys. Chem. A 2006, 110, 9175. (35) Craw, J. S.; Vincent, M. A.; Hillier, I. H.; Wallwork, A. L. J. Phys. Chem. A 1995, 99, 10181. (36) Shamov, G. A.; Schreckenbach, G. Acc. Chem. Res. 2010, 43, 19. (37) Hagberg, D.; Kalstrom, G.; Roos, B. J.; Gagliardi, L. J. Am. Chem. Soc. 2005, 127, 14250. (38) Hay, P. J.; Martin, R. L.; Schreckenbach, G. J. Phys. Chem. A 2000, 104, 6259. (39) Shamov, G. A.; Schreckenbach, G. J. Phys. Chem. A 2005, 109, 10961. (40) Austin, J. A.; Sundararajan, M.; Vincent, M. A.; Hillier, I. H. Dalton Trans. 2009, 5902. (41) Schlosser, F.; Kruger, S.; Rosch, N. Inorg. Chem. 2006, 24, 1480. (42) Wang, L.; Zou, H.; Yi, C.; Xu, J.; Xu, W. J. Nanosci. Nanotechnol. 2011, 11, 3298. (43) Li, Z.; Chen, Z.-X.; Kang, G.-J.; He, X. Catal. Today 2011, 165, 25. (44) Garg, I.; Sharma, H.; Dharamvir, Jindal, V. K. J. Phys. Chem. C 2010, 114, 18762. (45) Malijevsky , A.; Lísal, M. J. Chem. Phys. 2009, 130, 164713.

ARTICLE

(46) McNamara, J. P.; Sharma, R.; Vincent, M. A.; Hillier, I. H.; Morgado, C. A. Phys. Chem. Chem. Phys. 2008, 10, 128. (47) Simeon, T.; Balasubramanian, K.; Welch, C. R. J. Phys. Chem. Lett. 2010, 1, 457. (48) Peng, H.; Alemany, L. B.; Margrave, J. L.; Khabashesku, V. N. J. Am. Chem. Soc. 2003, 125, 15174. (49) TURBOMOLE, V6.0 2009, a development of University of Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989
2007, TURBOMOLE GmbH, since 2007; available from http://www.turbomole.com. (50) Shamov, G. A.; Schreckenbach, G.; Martin, R. L.; Hay, P. J. Inorg. Chem. 2008, 47, 1465. (51) Nguyen-Trung, C.; Begun, G. M.; Palmer, D. A. Inorg. Chem. 2002, 31, 5280. (52) Toth, L. M.; Begun, G. M. J. Phys. Chem. 1981, 85, 547. (53) Tissandier, M. D.; Cowen, K. A.; Feng, W. Y.; Gundlach, E.; Cohen, M. H.; Earhart, A. D.; Coe, J. V.; Tuttle, T. R., Jr. J. Phys. Chem. A 1998, 102, 7787. (54) Matsuo, Y.; Tahara, K.; Nakamura, E. Org. Lett. 2003, 5, 3181.

6737

dx.doi.org/10.1021/jp203723t |J. Phys. Chem. A 2011, 115, 6732–6737