Noncovalent Interactions between Cytosine and SWCNT: Curvature

May 18, 2007 - The eclipsed tangential (ET) conformation with respect to the six-membered ring of cytosine and the central ring of SWCNT fragments is ...
0 downloads 0 Views 342KB Size
Download PDF
6520

J. Phys. Chem. B 2007, 111, 6520-6526

Noncovalent Interactions between Cytosine and SWCNT: Curvature Dependence of Complexes via π‚‚‚π Stacking and Cooperative CH‚‚‚π/NH‚‚‚π Yixuan Wang*,† and Yuxiang Bu*,‡ Department of Natural Science, Albany State UniVersity, Albany, Georgia 31705, and School of Chemistry, Shandong UniVersity, Jinan 250100, People’s Republic of China ReceiVed: January 3, 2007; In Final Form: March 11, 2007

Fragments of C24H12, adapted from a variety of armchair [(n,n), (n ) 5, 7, and 8)] and zigzag [(m,0) (m ) 8, 10, and 12)] single-walled carbon nanotube (SWCNT), are used to model corresponding SWCNTs with different diameters and electronic structures. The parallel binding mainly through π‚‚‚π stacking interaction, as well as the perpendicular binding via cooperative NH‚‚‚π and CH‚‚‚π between cytosine and the fragments of SWCNT have been extensively investigated with a GGA type of DFT, PW91LYP/6-311++G(d,p). The eclipsed tangential (ET) conformation with respect to the six-membered ring of cytosine and the central ring of SWCNT fragments is less stable than the slipped tangential (ST) conformation for the given fragment; perpendicular conformations with NH2 and CH ends have higher negative binding energy than those with NH and CH ends. At PW91LYP/6-311++G(d,p) level, two tangential complexes are less bound than perpendicular complexes. However, as electron correlation is treated with MP2/6-311G(d,p) for PW91LYP/ 6-311++G(d,p) optimized complexes, it turns out there is an opposite trend that two tangential complexes become more stable than three perpendicular complexes. This result implies that electron correlation, a primary source to dispersion energy, has more significant contributions to the π‚‚‚π stacking complexes than to the complexes via cooperative NH‚‚‚π and CH‚‚‚π interactions. In addition, it was found for the first time that binding energies for two tangential complexes become more negative with increasing nanotube diameter, while those for three perpendicular complexes have a weaker dependence on the curvature; i.e., binding energies are slightly less and less negative. The performance of a novel hybrid DFT, MPWB1K, was also discussed.

1. Introduction Applications of carbon nanotubes (CNTs) into biotechnology have recently emerged, raising great potential in such a few areas as biosensors, DNA, and protein transporters for therapy purpose.1-7 CNTs themselves have a very low solubility in aqueous solutions as well as in organic solvents, which had been a major barrier for a variety of potential applications. Strategic approaches toward solubilization of CNTs have been developed mainly through their surface functionalization of either covalent or noncovalent attachments to the sidewalls of CNTs.2,8-13 For instance, the high solubility of CNTs in a wide range of solvents could be achieved through 1,3-dipolar cycloaddition reactions of azomethine yield in the presence of R-amino acid and an aldehyde.14,15 This covalent functionalization scheme enables CNTs to carry ammonium groups to further conjugate with therapeutic molecules. Other covalent schemes of CNTs include acidic oxidization,16 amidation, and esterification of the nanotube-bound carboxylic acid.17,18 The covalent modification involving chemical reactions between CNTs and molecules somewhat impairs the structural and electronic properties of CNTs. Therefore, noncovalent functionalization of CNT has attracted increasing attention. Noncovalent functionalization could not only enhance solubility of SWCNT but also maintain their attractive geometric, electronic, and mechanical properties. * Corresponding authors. E-amils: [email protected]; byx@ sdu.edu.cn. † Albany State University. ‡ Shandong University.

Among numerous functional species for solublizing CNTs, biological and bioactive materials are of special importance. The fundamental components in living system, such as carbohydrates, proteins, and nucleic acids, as well as their precursors have been explored to noncovalently functionalize CNTs toward the delivery of therapeutic agents. Several recent review papers deal with the functionalization of SWCNTs with biomolecules and their biomedical applications.2,3,19 Two mechanisms for the CNTs functionalization with DNA were suggested, via wrapping around the sidewall or encapsulating into the cavity of CNTs. Zheng et al. have reported that poly T(thymine) single-strand DNAs (ssDNA) could more efficiently disperse SWCNTs in water than poly A (adenine) and poly C (cytosine) because the latter two poly bases could strongly self-stack in aqueous solution.20 Their further molecular dynamics simulation demonstrated that poly (T) ssDNA wrap around (10,0) SWCNT mainly via π-π stacking interaction between bases of nucleic acid and SWCNT, while the hydrophilic sugar-phosphate backbones point to water media to achieve solubility in water. Using an uncapped armchair (10,10) SWCNT (2.95 and 1.36 nm in respective length and diameter) and a ssDNA oligonucleotide with 8 adenine bases, Gao et al. employed a classical molecular dynamics simulation for SWCNT-DNA conjugate in aqueous environment.21 They found an opposite mechanism that ssDNA spontaneously encapsulates into (10,10) SWCNT, mainly driven by the van der Waals and hydrophobic interactions with the former dominant. They suggested that the diameter (1.08 nm) of (8,8)

10.1021/jp0700433 CCC: $37.00 © 2007 American Chemical Society Published on Web 05/18/2007

Interactions between Cytosine and SWCNT is an onset for the encapsulation of ssDNA. Such encapsulation mechanism was also evidenced experimentally.22 The interaction strength of SWCNT-DNA conjugate may also depend on the diameters and the electronic properties (i.e., metallic or semiconducting) of SWCNT, and the type of the bases of DNA. Little is known about whether the two different types of mechanism are resulted from SWCNTs with different electronic structures or they are due to different bases or just different diameters. To address these questions, it is necessary to investigate the interactions between SWCNT with different electronic structures and different diameters and a variety of bases at a molecular level. As a part of the project, in the present paper the interactions between cytosine and the fragments of SWCNTs are extensively investigated with density functional theories (DFT) as well as wavefunction-based MP2 method. The purposes of the present paper are 2-fold. One is to understand the noncovalent interaction between cytosine and SWCNTs, including the major source of such interaction, charge transfer, and the effect of electron correlation. On the other hand, we try to shed light on the diameter dependence of such weak interactions. 2. Models and Theoretical Calculations All of the calculations were performed using the G03 program, revision D02.23 Accurate description for noncovalent weak interaction systems is still a challenge for DFT, a promising quantum mechanics method to the large systems, although considerable improvements have been achieved over LDA and such conventional hybrid DFT as B3LYP in the recent years.24,25 In the previous studies, one of us had applied Perdew-Wang 1991 (PW91) exchange functional with LYP correlation functional, PW91LYP, to the noncovalent interaction systems via C-H‚‚‚O such as alkyl carbonate dimer and trimers.26 The synonym of PW91LYP, mPWLYP, was also applied to noncovalent system by Truhlar et al.27 Recently, the novel exchange density functionals modified from PW91 exchange functional together with a meta correlation, MPWB1K, provide relatively reasonable results for π‚‚‚π stacking systems.28 On the basis of these applications, PW91, together with the LYP correlation functional and 6-311++G(d,p) basis set, denoted as PW91LYP/6-311++G(d,p), was used to optimize complexes of cytosine and fragments of nanotube. The novel hybrid meta GGA type of DFT, MPWB1K, was also applied to the complexes of cytosine and (5,5) fragment for comparison. It is found that MPWB1Kmethod has convergence problem for the current systems with the basis set 6-311++G(d,p), while it works well with cc-pVDZ for optimization but basis set superposition error correction. In order to discuss electron correlation effect on binding energies for the present noncovalent systems via π‚‚‚π stacking and NH/CH‚‚‚π interactions, MP2/ 6-311G(d,p) single-point calculations with the geometries optimized with PW91LYP/6-311++G(d,p) and MPWB1K/ccPVDZ, denoted as MP2/6-311G(d,p)//PW91LYP/6-311++G(d,p) and MP2/6-311G(d,p)//MPWB1K/cc-pVDZ, were also carried out for the complexes of cytosine and (5,5). Atomic net charges were calculated using the Mulliken, CHELPG,29 and NPA30 schemes. In order to individually explore curvature effect of SWCNT on the binding of cytosine to SWCNT, the fragments of C24H12 are adapted from a variety of armchair [(n,n), (n ) 5, 7, and 8)] and zigzag [(m,0) (m ) 8, 10, and 12)] SWCNTs. The geometries of the fragments were frozen as they are in the corresponding SWCNTs with C-C distance of 1.41 Å and C-C-C angle of approximately 119.0°, yet the peripheral

J. Phys. Chem. B, Vol. 111, No. 23, 2007 6521 carbons are saturated with H-atoms. PW91LYP/6-311++G(d,p) was employed to fully optimize adsorptions of cytosine on SWCNTs, and the binding energies were then corrected with MP2/6-311G(d,p)//PW91LYP/6-311++G(d,p). 3. Results and Discussions 3.1. Geometries and Binding Energies of Cytosine with (5,5) Fragment. The tangential binding through π‚‚‚π stacking interaction, as well as the perpendicular binding via NH‚‚‚π and CH‚‚‚π between cytosine and the fragments of SWCNT, are extensively investigated. The relevant calculations show that covalent interaction does not exist, but there does exist noncovalent interaction between cytosine and the fragment of (5,5). In this study seven complexes are located, including two tangential through π‚‚‚π stacking interaction, four perpendicular ones through the cooperation of NH‚‚‚π or NH2‚‚‚π and CH‚‚‚π, and another perpendicular one via NH2‚‚‚π, which are shown in Figure 1 for (5,5) armchair SWCNT fragment. The first structure (1a) is an eclipsed tangential (denoted as ET) complex with respect to the cytosine ring and the central sixmembered (C6) ring of the (5,5) fragment, and the second one (1b) corresponds to a slipped tangential (ST) in which the cytosine ring considerably deviates from the central C6 ring. In the perpendicular structures 1c and 1d, the NH2 end of cytosine points toward nanotube fragments, the cytosine ring being parallel and perpendicular to the nanotube axis, respectively; while the NH end sits above the fragments in structures 1e and 1f. For structure 1g, the NH2 group symmetrically resides above the fragment. All complexes are strongly bound at PW91LYP/ 6-311++G(d,p), and the BSSE-corrected binding energies (BE) are in a range from -1.86 to -3.63 kcal/mol. The least stable complex is 1a (BE ≈ -1.86 kcal/mol), followed by another tangential one ST 1b (-2.28 kcal/mol). The trend is the same as benzene dimers (BEs predicted for parallel and slipped parallel ones by MCG3-MPWB ≈ -1.60 vs -2.88 kcal/mol31). The perpendicular structures, 1c-1f, have more negative binding energies than the two tangential complexes. Of 1c-1f, 1e is the most stable structures with a BE of -3.63 kcal/mol. Besides the cooperative interactions of CH‚‚‚π and NH‚‚‚π, the relative short distance (∼3.7 Å) in structure 1e from carbonyl oxygen and hydrogen of the fragment may be responsible for its highest stability, resulting in significant electrostatic attraction between cytosine and the fragment, which is absent in other structures. The structure 1f has less negative binding energy than 1e by 0.64 kcal/mol (-2.99 vs -3.63 kcal/mol). The structures 1c and 1d are mainly stabilized by cooperative interactions of CH‚‚‚π and NH‚‚‚π but from NH2 ends. Since a purine or pyrimidine base is linked through a β-N-glycosidic linkage to C-1 of a pentose sugar in ribonucleotides and DNA, NH2 is still a relatively naked group and complexes 1c and 1d are therefore more important than 1e and 1f. A more negative binding energy of 1c than 1d by approximately 0.3 kcal/mol may be due to the fact that N-H and C-H groups are closer to the carbon atoms in the fragment as cytosine ring is oriented in parallel to nanotube axis in 1c. 3.2. The Performance of MPWB1K. Because of convergence problem of MPWB1K/6-311++G(d,p) in the present systems, cc-pVDZ basis set was used with MPWB1K method. Basis set effect was first addressed for PW91LYP. The binding energies before BSSE correction for 1c, 1d and 1f predicted by PW91LYP/cc-pVDZ are -4.43, -4.13, and -4.04 kcal/mol, which are very similar to those from PW91LYP/6-311++G(d,p) (-4.38, -4.20, and -3.98 kcal/mol). As expected, the DFT method exhibits a rather small basis set effect as basis set

6522 J. Phys. Chem. B, Vol. 111, No. 23, 2007

Wang and Bu

Figure 1. The complexes are resulted from the noncovalent binding of cytosine to (5,5) fragment. ET and ST are eclipsed and slipped tangential structures with respect to cytosine ring and the central C6 ring, respectively. P1-P4 are perpendicular structures through cooperative interactions of either NH‚‚‚π or NH(2)‚‚‚π and CH‚‚‚π, and P5 is a perpendicular structure via NH2‚‚‚π. A red ball stands for oxygen atom, blue for nitrogen, gray for carbon, and white for hydrogen atoms.

TABLE 1: Binding Energies (BE, kcal/mol) for the Complexes of Cytosine and the Fragment of (5,5) Predicted from Various Theoretical Methodsa MP2/C// HF/C// complexes PW91L YP/Ab PW91L YP/B MPWB 1K/B PW91L YP/A MPWB 1K/B -1.86(-3.53) -2.27(-3.88) -3.38(-4.38) -3.44(-4.43) -3.09(-4.20) -3.02(-4.13) -3.63(-4.73) -2.99(-3.98) -3.00(-4.04) -2.43(-3.22)

1a 1b 1c 1d 1e 1f 1g

HF/A// MP2/A// Ecorrc MPWB 1K/B MPWB 1K/B

MP2/C// MPWB 1K/B

(-4.22) (-4.70) (-5.13) (-3.99)

-6.53(-9.33) 1.44(-0. 32) -6.68(-9.23) 0.57(-1. 16) -6.04(-8.39) 0.22(-1. 15) -5.92(-7.97) -0.48(-1.45)

-7.27(-10.61) -7.83(-11.13) -6.96(-10.53) -6.26(-8.88)

(-3.35)

-5.56(-7.42) -1.18(-2.00) -5.67(-7.71)

Ecorr

-8.71 -8.40 -7.18 -5.78

1.28 0.47 0.37 -0.33

-8.76 -9.25 -7.61 -6.81

-10.04 -9.70 -7.98 -6.48

-4.49

-1.03

-6.12

-5.09

a A, B, and C represent 6-311++G(d,p), cc-pVDZ, and 6-311G(d,p) basis sets, respectively. b The data outside and inside of the parenthesis refer to binding energies with and without BSSE corrections. c MP2/C correlation interaction energy defined as difference between MP2/C and HF/C.

TABLE 2: The Distances and Binding Energies for the Complexes of Cytosine and the (5,5) Fragment with Different Theoretical Methods distancesa (Å)

binding energies (kcal/mol)

complexes

PW91LYP/A

MPWB1K/B

MP2/B

MP2/C// PW91LYP/A

1a 1b 1c 1d 1f

3.83 4.05 3.82 4.15 3.93

3.76 3.90 3.72 3.95 3.87

3.14 3.35 3.36 3.77 3.59

-6.53 -6.68 -6.04 -5.92 -5.56

MP2/C// MPWB1K

MP2/C// MP2/B

-7.27 -7.83 -6.96 -6.26 -5.67

-8.00 -9.46 -7.09 -6.38 -6.06

a The distances for the complexes 1a and 1b refer to the separations between the center of the six-membered ring of cytosine and the center of the uppermost C6 ring of the (5,5) fragment, while those for the complexes 1c, 1d, and 1f are the distances from the N atoms of NH2 or NH to the center of neighbor six-membered ring in the fragment of (5,5).

is beyond double-ζ. Major parameters for 1a-1d and 1f predicted with PW91LYP/6-311++G(d,p), MPWB1K/cc-pVDZ, and MP2/cc-PVDZ are listed in Table 2. For the two π‚‚‚π stacking complexes 1a and 1b, the distances, between the center of the six-membered ring of cytosine and the center of the uppermost C6 ring of the (5,5) fragment, from MPWB1K/cc-pVDZ are only slightly shorter by 0.07-0.15 Å than those from PW91LYP/ 6-311++G(d,p) (3.76 and 3.90 Å vs 3.83 and 4.05 Å for 1a and 1b). Since it is well-known that MP2 method considerably overestimates the binding for noncovalent interaction systems,

the corresponding distances for the two π‚‚‚π stacking complexes predicted by MP2/cc-pVDZ are rather shorter by ∼0.7 Å than the ones from PW91LYP/6-311++G(d,p), and by 0.40.6 Å than those predicted by MPWB1K/cc-pVDZ. However, these differences are dramatically less than those for parallel benzene dimer, reported by Johnson et al.,32 where approximately 2.0 Å difference of the benzene ring separation, predicted by PW91/6-31G* and MP2/aug-cc-pVDZ, may be partially due to too small basis set of PW91/6-31G* level. Another factor that may be responsible for this significant

Interactions between Cytosine and SWCNT difference is distinct origins in the two kinds of parallel complexes. It is expected that electrostatic interactions contribute relatively more to the present π‚‚‚π stacking complexes than to the benzene parallel dimers. Electrostatic interactions could be more readily account for by DFT than dispersion interaction, an important component for van der Waals interaction. For the three investigated perpendicular complexes (1c, 1d, and 1f), all of the distances from the N atoms of NH2 or NH to the center of neighbor six-membered ring in the fragment of (5,5), predicted by PW91LYP/cc-pVDZ are only longer by 0.06 Å (1f), 0.10 Å(1c), and 0.20 Å (1c) than those from MPWB1K/ cc-pVDZ. The first two differences are somewhat less than the distance difference of the center of mass for NH3-benzene dimer, predicted by Zhao et al.33 with PW91 and MPB1K (3.82 vs.3.58 Å), and the last one is comparable to that in NH3-benzene dimer. Because of the same reason as for the complexes 1a and 1b, the MP2/cc-pVDZ method again provides much shorter distances than PW91LYP/6-311++G(d,p) by 0.34-46 Å, and than MPWB1K/cc-PVDZ by 0.18-0.36 Å. It seems that the binding energies are more sensitive to the applied methods than geometries, especially for the two tangent π‚‚‚π complexes, where the binding energies before BSSE correction of 1a and 1b with MPWB1K/cc-pVDZ are 0.7-0.8 kcal/mol more negative than those with PW91LYP/6-311++G(d,p). The binding energies of 1c, 1d, and 1f from MPWB1K/ cc-pVDZ exhibit the same trend as those from PW91LYP/ccpVDZ: 1c the most stable with binding energy before BSSE correction of -4.78 kcal/mol, followed by 1d (BE: -3.99 kcal/ mol) and 1f (BE: -3.35 kcal/mol). To more accurately account for dispersion for the current systems, MP2/6-311G(d,p) method was applied to the complexes predicted by two DFT methods. The binding energies predicted by MP2/6-311G(d,p)//MPWB1K/ cc-pVDZ are more negative by 0.1-1.2 kcal/mol than those from MP2/6-311G(d,p)//PW91LYP/6-311++G(d,p), which approximately account for 2-15% of the total binding energies. This indicates that MPWB1K could more accurately describe noncovalent interaction systems. However, it is interesting to note that MP2/6-311G(d,p)//PW91LYP/6-311++G(d,p) yields the same sequence of binding energy for 1a-1d and 1f as that from MP2/6-311G(d,p)//MPWB1K/cc-pVDZ, that is, the stability of the involved complexes decrease in the order of 1b > 1a > 1c > 1d > 1f. According to Table 2, MP2/6-311G(d,p)// MP2/cc-pVDZ binding energies are more negative than those from the two DFT-optimized geometries, especially for the two π‚‚‚π stacking complexes, which is consistent with the variation of the relevant distances. The major objective of the present work is to investigate curvature dependence of the complexes on SWCNTs. For this purpose the most important is the relative binding of cytosine to different SWCNTs. Although MP2/6311(d,p)//PW91LYP/6-311++G(d,p) provides less accurate binding energies for the specific complex than MP2/6-311(d,p)// MPWB1K/cc-pVDZ, because of the same sequence the former method is also reliable for the current objective. 3.3. Contribution of Electron Correlation to the Binding Energy. At HF/6-311G(d,p)//MPWB1K/cc-pVDZ level, two π‚‚‚π stacking complexes (1a and 1b) and complex 1c via cooperative CH‚‚‚π and NH(2)‚‚‚π interactions are unbound; while the complexes 1d and 1f via cooperative CH‚‚‚π and NH(2)‚‚‚π interactions are stable yet their binding energies are quite low negative (-0.48 and -1.18 kcal/mol). When electron correlation is well taken into account with MP2/6-311G(d,p)// MPWB1K/cc-pVDZ, an opposite stability sequence to that from PW91LYP/6-311++G(d,p) is turned out, i.e., two π‚‚‚π stacking complexes (BE of 1a and 1b: -7.27 and -7.83 kcal/mol)

J. Phys. Chem. B, Vol. 111, No. 23, 2007 6523

Figure 2. Plots of the electrostatic potential for the fragment C24H12 of (5,5) SWCNT (left) and cytosine (right) (PW91LYP/6-311++G(d,p) on the isodensity surface of 0.001. The color is coded as red for strong negative, and blue for strong positive.

TABLE 3: Total Atomic Charges of Cytosine in Cyosine-(5,5) Complexes with PW91LYP/6-311++G(d,p) complexes

NPA

CHELPG

MULLIKE

cytosine 1a 1b 1c 1d 1f

0.00 0.012 0.008 0.004 0.001 0.004

0.00 -0.094 -0.097 -0.210 -0.020 -0.190

0.00 -0.180 -0.160 -0.068 -0.130 -0.089

have higher negative binding energies than the three perpendicular structures (BE of 1c, 1d, and 1f: -6.96, -6.26, and -5.67 kcal/mol). This difference shows that electron correlation effects contribute more to the π‚‚‚π stacking complexes (1a and 1b) than to the complexes via cooperative NH‚‚‚π and CH‚‚‚π interactions. The contribution of electron correlation to the binding energy, Ecorr, is defined as the difference between the binding energies predicted by MP2/6-311G(d,p)//MPWB1K/cc-pVDZ and HF/ 6-311G(d,p)//MPWB1K/cc-pVDZ and is listed in the eighth column in Table 1. Dispersion energy in the noncovalent systems is primarily determined by Ecorr. Thus, the high negative values of Ecorr (-8.71 and -8.40 kcal/mol) for the tangential complexes indicate that the dispersion interaction may be the dominant source of attraction in these complexes; while the low negative values of Ecorr (-7.18, -5.78, and -4.49 kcal/mol) for the perpendicular complexes imply that the perpendicular complexes may be mainly determined by another long-range attraction, electrostatic interaction. In order to examine the effect of diffuse functions, MP2/6311++G(d,p)//MPWB1K/cc-pVDZ has been employed to the above complexes consisting of cytosine and the fragemnet of (5,5). As shown in Table 1, the diffuse functions further stabilize the complexes with more negative binding energies than those yielded by MP2/6-311G(d,p)//MPWB1K/cc-pVDZ. The binding energies for two π‚‚‚π stacking complexes become more negative by approximately 1.5 kcal/mol (-8.76 vs -7.27 kcal/ mol for 1a; -9.25 vs -7.83 kcal/mol for 1b), while the perpendicular complexes are less sensitive to the diffuse functions than the π‚‚‚π stacking complexes, evidenced by only ∼0.5 kcal/mol variation for binding energies (-7.61 vs -6.96 kcal/mol for 1c; -6.81 vs -6.26 kcal/mol for 1d; and -6.12 kcal/mol vs -5.67 kcal/mol for 1f). Moreover, the most interesting issue for MP2/6-311++G(d,p) is that the diffuse function involved basis set provides the same stability sequence as MP2/6-311G(d,p). For electron correlation effects, the same conclusion can also be reached that the effect plays a more important role on the π‚‚‚π stacking complexes than on the perpendicular ones.

6524 J. Phys. Chem. B, Vol. 111, No. 23, 2007

Wang and Bu

Figure 4. Binding energy variations for ET, ST, P1, P2, and P4 complexes with diameters of SWCNT, predicted by PW91LYP/6311++G(d,p) (top) and MP2/6-311G(d,p)//PW91LYP/6-311++G(d,p) (bottom).

Figure 3. Orbital interaction in complex P1 at PW91lYP/6-311++G(d,p) level. The isosurface value is 0.02 au.

To further discuss the binding between cytosine and the nanotube fragment, their electrostatic potentials (ESP) are presented in Figure 2. Regions of strong positive potential (blue and light blue color) of cytosine are found close to NH2/CH as well as NH/CH groups, carbonyl and neighbor nitrogen show strong negative (red), and the cytosine ring shows relatively strong negative (yellow to green). Internal regions of (5,5) fragment show strong negative potential (red to yellow). This at least provides an explanation, why the electrostatic attraction plays more important roles on stabilizing the perpendicular complexes than on the tangential complexes. The BE of 1b predicted by MP2/6-311G(d,p)//MPWB1K/ cc-PVDZ is much more negative by over 5 kcal/mol than that of slipped parallel benzene dimer (-7.83 vs -2.1934 kcal/mol). This difference is in part attributed to van der Waals and electrostatic interactions between NH/NH2 groups and nanotube fragment. Carbonyl carbon and the carbon connecting NH2 carry high positive charge (+0.995 and +0.901e), which may also induce significant electrostatic attraction. In addition, although oxygen (charge: -0.657e) of cytosine is quite far from neighbor H atoms of C24H12 (>4.0 Å), long-range electrostatic attraction between O and H atoms may be another factor for the higher stability of complex 1b. NH‚‚‚π interaction in T-shaped NH3-benzene dimer and CH‚‚‚π interaction existing in the T-shaped benzene dimer were estimated to be -1.45 and -2.12 kcal/mol at MP2/6-311G(d,p) level.34,35 Sum of the two interactions (-3.57 kcal/mol) is still considerably far from the binding energy of the complex 1c (-6.96 kcal/mol). The net electrostatic charge analysis shows

that H atoms connected to N (H2) and C in cytosine carry higher positive charges than those in NH3 and benzene (+0.38 vs +0.32e; +0.17 vs +0.07 at PW91LYP/6-311++G(d,p)), which will result in higher electrostatic attraction and may be in part responsible for the higher binding energy in complex 1c. On the other hand, the significant higher negative binding energy of complex 1c indicate that, similar to the normal hydrogen bonding, there may be a cooperativity for such weak interactions as NH‚‚‚π and CH‚‚‚π, which is worthy to be further investigated in future. 3.4. Charge Transfer and Orbital Interaction. To analyze charge transfer between cytosine and (5,5), atomic charges of cytosine from the three different schemes are collected in Table 3. Both Mulliken and CHELPG yield negative charges for cytosine, while NPA charges are very small positive. The following discussion will be based on NPA charges. Of all of the investigated complexes (1a-1g), cytosine in the perpendicular complexes (1c, 1d, 1f, and 1g) are almost neutral, suggesting that little electron transfer between the two moieties. Only very small amount of charge of 0.01e was transferred from cytosine to nanotube fragment in π‚‚‚π stacking complexes 1a and 1b. Either neutral or rather small positive charge of cytosine implies that short-range charge-transfer interaction may not play important role on the stability of complexes. In an effort to study frontier orbital interactions to further understand the nature of the complexes, an orbital analysis of P1 was done at PW91LYP/6-311++G(d,p) level. As shown in Figure 3, there is an extremely weak orbital interaction and no significant mixing of the frontier orbitals, which is again different than the conventional H-bond system. Frontier orbitals are only slightly perturbed due to the cooperative CH‚‚‚π and NH‚‚‚π interactions as cytosine and (5,5) fragment form the complex. The LUMO and HOMO basically come from the LUMO and HOMO of nanotube fragment. The HOMO of cytosine mainly contributes to the HOMO-2. This orbital

Interactions between Cytosine and SWCNT

Figure 5. Plots of electrostatic charge (black diamond) carried by cytosine molecule and of the distance (purple triangle) between center of cytosine ring and the center of central ring of SWCNT fragment in ET type of complexes.

Figure 6. Plots of electrostatic charges carried by cytosine molecule (black diamond) and by NH2 groups (purple square), and plot of the distance (blue star) between H atom and the closest C atom of NH‚‚‚π interaction for P2 type of complexes.

interaction pattern is rather different from conventional Hbonding system, where the HOMO from H acceptor considerably mixes with the LUMO of the H-donor. This weak frontier orbital interaction is consistent with a little charge transfer. 3.5. Curvature Dependence of Binding Energies. To explore the curvature or diameter dependence of the noncovalent complexes, the interactions between cytosine and a variety of SWCNTs with different diameters and electronic structures were also extensively investigated with PW91LYP/6-311++G(d,p) for the geometries of complexes 1a-1f. The binding energies were then corrected with MP2/6-311G(d,p)//PW91LYP/6311++G(d,p). The binding energies from the two methods are plotted in Figures 4a,b. The orbital energies of HOMO and LUMO only slightly change and HOMO-LUMO gaps are approximately 0.1 au even for metallic armchair SWCNTs, suggesting that the present fragment C24H12 has disability to model the electronic structures of SWCNTs. However, the interactions between cytosine with the fragment from distinct SWCNTs can reflect the curvature effect on π‚‚‚‚π stacking interaction and NH‚‚‚‚π and CH‚‚‚‚π interactions. The relative strength of the complexes (ET, DT, P1, P2, and P4) for the other investigated fragments qualitatively remains the same as that for (5,5). At PW91LYP/6-311++G(d,p) level, two π‚‚‚π

J. Phys. Chem. B, Vol. 111, No. 23, 2007 6525 stacking complexes are less stable than the complexes via cooperative NH(2)/CH‚‚‚π interaction; while MP2/6-311G(d,p)// PW91LYP/6-311++G(d,p) reverses this order. However, as shown in Figure 4 (top and bottom), binding energies for two tangential complexes (ET and ST) become more negative with an increase of nanotube diameter; while those for three perpendicular complexes (P1, P2, and P4) have a weaker dependence on the curvature; i.e., binding energies are slightly less and less negative with increasing nanotube diameter. According to Figure 5, for the eclipsed tangential complexes (ET), the distances from the center of cytosine ring to the center of C6 ring of nanotubes are in a short range of 3.82-3.92 Å. The electrostatic charges carried by cytosine become slightly less negative with the increase of nanotube diameter. Such two factors imply that electrostatic attraction between cytosine and the fragment of nanotube may not significantly change with the increase of nanotube diameter. Thus, the fragment of nanotube with bigger diameter should favor another factor, electron correlation, which is a primary contribution to dispersion attraction energy, and variation of electron correlation with diameter may be responsible for the binding energy trend of tangential complexes. The electrostatic charges of cytosine, charges of NH2 group of cytosine and the distance between H atom and the closest C atom of NH‚‚‚‚π interaction are presented in Figure 6 for P2 type of complexes. The electrostatic charges of cytosine become less negative from -0.200e for (8,0) to -0.144e for (8,8), which is consistent with the variation of HOMO of nanotubes, suggesting that electron transfer from naotube to cytosine become less with increasing diameters of nanotubes. The similar distances (∼2.7 Å) from donor H atoms of NH‚‚‚‚π and CH‚‚‚π interactions to the closest neighbor C of nanotube fragments suggest that electrostatic attractions should be close for different nanotubes. This may qualitatively explain the trend that the binding energies for the perpendicular complexes have a weak dependence on diameters of nanotubes. To quantitatively illustrate the origin of variation pattern for tangential and perpendicular complexes with nanotube diameter, energy decomposition analysis (EDA)36 at MP2/6-311G(d,p) level will be carried out, which will provide detailed insights into various contributions to the total binding energy, such as electrostatic interaction energy, Pauli exchange repulsion, charge transfer interaction energy and dispersion energy. 4. Conclusions PW91LYP, MPWB1K, and MP2 methods are employed to intensively investigate the interactions of cytosine and the fragment of (5,5) SWCNT, and then the interactions are extensively explored between cytosine and fragments of a variety of armchair [(n,n), (n ) 5, 7, and 8)] and zigzag [(m,0) (m ) 8, 10, and 12)] SWCNT. The results show that covalent interaction does not exist, but there does exist noncovalent interaction between cytosine and the fragments of SWCNT. Electron correlation, a major component of dispersion attraction, is a dominant source for tangential π‚‚‚π stacking conformations, while electrostatic attractions play more important roles on the stability of the perpendicular complexes via cooperative NH‚‚‚π and CH‚‚‚π interactions than on that of the tangential complexes. In addition, it is found for the first time, that binding energies for tangential complexes become more and more negative with increasing nanotube diameter, while those for three perpendicular complexes have a weak dependence on the curvature; i.e., binding energies are slightly less and less negative with increasing nanotube diameter.

6526 J. Phys. Chem. B, Vol. 111, No. 23, 2007 Acknowledgment. The financial support from the Faculty Development Award (Y. Wang) at the Albany State University through grants P20 MD001085 and G11 HD049635 from the National Institutes of Health (NIH), and computational time from the National Center for Supercomputing Applications (NCSA) are appreciated. References and Notes (1) Martin, C. R.; Kohli, P. The emerging field of nanotube biotechnology. Nat. ReV. Drug DiscoV. 2003, 2, 29-37. (2) Lin, Y.; Taylor, S.; Li, H. P.; Fernando, K. A. S.; Qu, L. W.; Wang, W.; Gu, L. R.; Zhou, B.; Sun, Y. P. Advances toward bioapplications of carbon nanotubes. J. Mater. Chem. 2004, 14, 527-541. (3) Bianco, A.; Kostarelos, K.; Prato, M. Applications of carbon nanotubes in drug delivery. Curr. Opin. Chem. Biol. 2005, 9 (6), 674679. (4) Kam, N. W. S.; Dai, H. J. Carbon Nanotubes as intracellular protein transporters: Generality and biological functionality. J. Am. Chem. Soc. 2005, 127, 6021-6026. (5) Kam, N. W. S.; O’Connell, M.; Wisdom, J. A.; Dai, H. J. Carbon nanotubes as multifunctional biological transporters and near-infrared agents for selective cancer cell destruction. Proc. Natl. Acad. Sci. U.S.A. 2005, 102 (33), 11600-11605. (6) Wu, W. W. S.; Pastorin, G.; Benincasa, M.; Klumpp, C.; Briand, J. P.; Gennaro, R.; Prato, M.; Bianco, A. Targeted delivery of amphotericin B to cells by using functionalized carbon nanotubes. Angew. Chem., Int. Ed. 2005, 44 (39), 6358-6362. (7) Kam, N. W. S.; Liu, Z. A.; Dai, H. J. Carbon nanotubes as intracellular transporters for proteins and DNA: An investigation of the uptake mechanism and pathway. Angew. Chem., Int. Ed. 2006, 45, 577581. (8) Star, A.; Liu, Y.; Grant, K.; Ridvan, L.; Stoddart, J. F.; Steuerman, D. W.; Diehl, M. R.; Boukai, A.; Heath, J. R. Noncovalent side-wall functionalization of single-walled carbon nanotubes. Macromolecules 2003, 36 (3), 553-560. (9) Chen, R. J.; Zhang, Y.; Wang, D.; Dai, H. J. Noncovalent sidewall functionalization of single-walled carbon nanotubes for protein immobilization. J. Am. Chem. Soc. 2001, 123 (16), 3838-3839. (10) Dyke, C. A.; Tour, J. M. Covalent Functionalization of SingleWalled Carbon Nanotubes for Materials Applications. J. Phys. Chem. A 2004, 108 (51), 11151-11159. (11) Bianco, A.; Prato, M. Can carbon nanotubes be considered useful tools for biological applications? AdV. Mater. 2003, 15 (20), 1765-1768. (12) Huang, W.; Taylor, S.; Fu, K.; Lin, Y.; Zhang, D.; Hanks, T. W.; Rao, A. M.; Sun, Y.-P. Attaching proteins to carbon nanotubes via diimideactivated amidation. Nano Lett. 2002, 2 (4), 311-314. (13) Lin, Y.; Allard, L. F.; Sun, Y. P. Protein-affinity of single-walled carbon nanotubes in water. J. Phys. Chem. B 2004, 108 (12), 3760-3764. (14) Georgakilas, V.; Kordatos, K.; Prato, M.; Guldi, D. M.; Holzinger, M.; Hirsch, A. Organic functionalization of carbon nanotubes. J. Am. Chem. Soc. 2002, 124 (5), 760-761. (15) Georgakilas, V.; Tagmatarchis, N.; Pantarotto, D.; Bianco, A.; Briand, J. P.; Prato, M. Amino acid functionalisation of water soluble carbon nanotubes. Chem. Commun. 2002, 24, 3051-3052. (16) Liu, J.; Rinzler, A. G.; Dai, H.; Hafner, J. H.; Bradley, R. K.; Boul, P. J.; Lu, A.; Iverson, T.; Shelimov, K.; Huffman, C. B.; Rodriguez-Macias, F.; Shon, Y.-S.; Lee, T. R.; Colbert, D. T.; Smalley, R. E. Fullerene pipes. Science 1998, 280, 1253-1256. (17) Sun, Y. P.; Fu, K. F.; Lin, Y.; Huang, W. J. Functionalized carbon nanotubes: Properties and applications. Acc. Chem. Res. 2002, 35 (12), 1096-104. (18) Lin, Y.; Rao, A. M.; Sadanadan, B.; Kenik, E. A.; Sun, Y. P. Functionalizing multiple-walled carbon nanotubes with aminopolymers. J. Phys. Chem. B 2002, 106 (6), 1294-1298.

Wang and Bu (19) Bianco, A.; Kostarelos, K.; Partidos, C. D.; Prato, M. Biomedical applications of functionalized carbon nanotubes. Chem. Commun. 2005, 571-577. (20) Zheng, M.; Jagota, A.; Semke, E. D.; Diner, B. A.; Mclean, R. S.; Lustig, S. R.; Richardson, R. E.; Tassi, N. G. DNA-assisted dispersion and separation of carbon nanotubes. Nat. Mater. 2003, 2 (5), 338-342. (21) Gao, H. J.; Kong, Y.; Cui, D. X.; Ozkan, C. S. Spontaneous Insertion of DNA Oligonucleotides into Carbon Nanotubes. Nano Lett. 2003, 3 (4), 471-473. (22) Ito, T.; Sun, L.; Crooks, R. M. Observation of DNA transport through a single carbon nanotube channel using fluorescence microscopy. Chem. Commun. 2003, 1482. (23) 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, 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 D.2; Gaussian, Inc.: Wallingford, CT 2004. (24) Zhao, Y.; Truhlar, D. G. Assessment of Model Chemistries for Noncovalent Interactions. J. Chem. Theory Comput. 2006, 2 1009-1018. (25) Xu, X.; Goddard, W. A., III. The X3LYP extended density functional for accurate descriptions of nonbond interactions, spin states, and thermochemical properties. PNAS 2004, 101 (9), 2673-2677. (26) Wang, Y. X.; Balbuena, P. B. Associations of Alkyl Carbonates: Intermolecular C-H‚‚‚O Interactions. J. Phys. Chem. A 2001, 105, 99729982. (27) Dahlke, E. E. T.; Donald, G. Improved Density Functionals for Water. J. Phys. Chem. B 2005, 109, 15677-15683. (28) Zhao, Y.; Truhlar, D. G. How well can new-generation density function methods describe stacking interactions in biological systems? Phys. Chem. Chem. Phys. 2005, 7, 2701-2705. (29) Breneman, C. M.; Wiberg, K. B. Determining atom-centered monopoles from molecular electrostatic potentials. The need for high sampling density in formamide conformational analysis. J. Comput. Chem. 1990, 11, 361. (30) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 899. (31) Zhao, Y.; Truhlar, D. G. Multicoefficient Extrapolated Density Functional Theory Studies of π‚‚‚π Interactions: The Benzene Dimer. J. Phys. Chem. A 2005, 109, 4209-4212. (32) Johnson, E. R.; Wolkow, R. A.; DiLabio, G. A. Application of 25 density functionals to dispersion-bound homomolecular dimers. Chem. Phys. Lett. 2004, 394, 334-338. (33) Zhao, Y.; Tishchenko, O.; Truhlar, D. G. How Well Can Density Functional Methods Describe Hydrogen Bonds to Acceptors? J. Phys. Chem. B 2005, 109 (41), 19046-19051. (34) Tsuzuki, S.; Honda, K. Origin of attraction and directionality of the pi/pi interaction: model chemistry calculations of benzene dimer interaction. J. Am. Chem. Soc. 2002, 124 (1), 104-112. (35) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M.; Tanabe, K. Origin of the attraction and directionality of the NH/pi Interaction: Comparsion, with OH/pi and CH/pi interactions. J. Am. Chem. Soc. 2000, 122, 11450-11458. (36) Kitaura, K.; Morokuma, K. Int. J. Quantum Chem. 1976, 10, 325.