Simulating Adsorption of Organic Pollutants on Finite (8,0) Single

Jul 20, 2012 - Understanding the mechanism and thermodynamics of the adsorption of chemicals on carbon nanotubes (CNTs) is important to risk ...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/est

Simulating Adsorption of Organic Pollutants on Finite (8,0) SingleWalled Carbon Nanotubes in Water Mingying Zou, Jinduo Zhang, Jingwen Chen,* and Xuehua Li Key Laboratory of Industrial Ecology and Environmental Engineering (MOE), School of Environmental Science and Technology, Dalian University of Technology, Dalian 116024, P. R. China S Supporting Information *

ABSTRACT: Understanding the mechanism and thermodynamics of the adsorption of chemicals on carbon nanotubes (CNTs) is important to risk assessment and pollution control of both CNTs and chemicals. We computed the adsorption of cyclohexane, benzene derivatives, and polycyclic aromatic hydrocarbons (PAHs) on (8,0) single-walled carbon nanotubes by the M05−2X of density functional theory. The computed adsorption energies (Ea) in the aqueous phase are lower than those in the gaseous phase, indicating that the adsorption in the aqueous phase is more favorable. The contribution of π−π interactions and the enhancing effect of a −NO2 substituent on the adsorption were quantified. For a hypothetical aromatic with the same hydrophobicity (logKOW) to cyclohexane, π−π interactions contribute ca. 24% of the total interactions as indicated by Ea. −NO2 enhances the π−π interactions due to its electron withdrawing effects, and contributes 24% to the value of Ea. Simple linear regression showed the computed Gibbs free energy changes for the adsorption correlate significantly with the experimental values (r = 0.97, p < 0.01). The correlation together with the computed thermodynamic parameters may be employed to predict the adsorption affinity of other chemicals. The study may pave a new way for evaluating/predicting the adsorption affinity of organic compounds on SWNTs and probing the adsorption mechanisms.



INTRODUCTION

combinations of the interactions, whereas the relative contribution of the different interactions remains unclear. Given the huge number of existing and emerging chemical pollutants in the environment, it becomes impossible to empirically test the adsorption of the exhaustive numbers of chemicals to CNTs, due to the high workload and cost. For ecological risk assessment and pollution prevention of both CNTs and organic chemicals, it becomes necessary to develop a computational method that can simulate and predict the adsorption of chemicals on CNTs. To date, various molecular modeling methods, e.g. molecular mechanics,21,22 molecular dynamics,23 Monte Carlo simulation,24 and quantum mechanics,25−27 have been employed to simulate the interactions between CNTs and chemical species. These studies laid a foundation for the emerging research field of structure−activity relationships of nanomaterials.28,29 An appropriate computational method should compromise between accuracy and computational cost. In 2006, Zhao et al.30 presented a new hybrid meta exchange-correlation functional M05−2X based on the density functional theory (DFT), and

In recent years, the adsorption of organic chemicals on carbon nanotubes (CNTs) has attracted increasing attention, as CNTs have been considered promising materials for molecular and ionization sensors, solid-phase extraction, pollutants removal, and even nuclear waste management.1−6 Furthermore, the production and industrial application of CNTs are booming. About 350 tons of CNTs were produced in 2007/2008, and an even higher future production is estimated.7,8 Once discharged into the environment, due to their adsorption affinity to organic pollutants, CNTs will have impacts on the environmental behavior, mobility, bioavailability, and toxicity of organic pollutants.9−11 Thus, understanding the adsorption mechanism and affinity is of crucial importance for exploring the application of CNTs and assessing the ecological risks of both CNTs and pollutants. To date, the adsorption of a series of organic chemicals on CNTs, including cyclohexane,12 benzene derivatives,12−15 polycyclic aromatic hydrocarbons (PAHs),13,15,16 dyes,17,18 and pharmaceuticals,19,20 has been investigated experimentally in aqueous solutions. According to recent reviews,1,3,4 the main adsorption mechanisms involve hydrophobic, π−π, hydrogen bond, and electrostatic interactions. These interactions may exist simultaneously, with different adsorbates having different © 2012 American Chemical Society

Received: Revised: Accepted: Published: 8887

April 5, 2012 July 17, 2012 July 20, 2012 July 20, 2012 dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

Table 1. Studied Chemicals and Their Physicochemical and Adsorption Parametersa aqueous phase

gaseous phase

compound

logKOW

d

Δq

Ea

ΔH

ΔS

ΔGcal

cyclohexane benzene toluene chlorobenzene 1,2-dichlorobenzene 1,3-dichlorobenzene 1,2,4-trichlorobenzene 1,2,4,5-tetrachlorobenzene nitrobenzene 1,3-dinitrobenzene 4-nitrotoluene 2,4-dinitrotoluene phenol aniline naphthalene anthracene fluorene phenanthrene pyrene

3.44 2.17 2.69 2.78 3.40 3.47 4.06 4.72 1.85 1.49 2.38 2.00 1.44 0.95 3.33 4.68 4.32 4.57 5.13

3.78 3.29 3.32 3.24 3.22 3.18 3.24 3.24 3.25 3.12 3.26 3.29 3.20 3.22 3.33 3.25 3.27 3.23 3.21

−0.006 0.007 0.009 0.011 0.013 0.013 0.011 0.019 0.010 −0.001 0.010 0.014 0.010 0.003 0.008 0.019 0.021 0.021 0.018

−34.30 −36.48 −36.95 −41.91 −46.93 −45.51 −50.08 −53.69 −45.64 −51.97 −46.63 −53.77 −40.80 −43.87 −45.60 −52.28 −49.69 −52.09 −55.29

−36.78 −38.96 −37.64 −44.39 −49.40 −49.21 −52.56 −56.17 −48.12 −57.47 −49.11 −56.25 −43.28 −46.35 −48.07 −54.76 −52.17 −54.57 −57.77

−123.80 −128.59 −118.03 −142.21 −143.18 −142.41 −151.40 −158.34 −150.90 −164.41 −141.70 −157.04 −131.32 −142.92 −144.52 −150.26 −145.26 −154.20 −144.69

0.13 −0.62 −2.45 −1.99 −6.71 −6.75 −7.42 −8.96 −3.13 −8.45 −6.86 −9.43 −4.13 −3.73 −4.98 −9.96 −8.86 −8.60 −14.63

ΔGexp −0.4012 −3.4212 −5.1812 −13.6312 −9.9613 −13.2912 −14.1312 −6.7912 −14.3312 −15.1712

−17.6513

Ea −18.74 −21.90 −25.05 −27.19 −30.57 −28.42 −36.06 −38.75 −28.45 −44.07 −40.21 −46.58 −23.11 −26.96 −28.97 −30.31 −27.18 −29.10 −33.50

Ea (ref) −25.0925 −25.4725 −29.9126

a logKOW is logarithm of the n-octanol/water partition coefficient, and the values are from ref 37. d is the smallest distance (Å) from an atom of the studied compounds to a SWNT atom. Δq (e) is the charge transfer from the molecule to the SWNT (the positive values indicate that the SWNTs accept charges). ΔH, ΔS, and ΔG are the changes of adsorption enthalpy (kJ/mol), entropy (J/(mol·K)), and Gibbs free energy (kJ/mol), respectively. The ΔGexp values were calculated from the experimental data.12,13 Ea is the adsorption energy (kJ/mol), and the Ea (ref) values are from the references indicated.

demonstrated that the functional had the best performance for noncovalent interactions without increasing computational cost. Burns et al.31 found the M05−2X functional behaved well for dispersion (van der Waals) dominated complexes, by comparing M05−2X with other DFT functionals. The M05− 2X functional was also successfully employed to simulate the stacking interaction between DNA bases and single-walled carbon nanotubes (SWNTs),32,33 and to predict the Gibbs free energy changes of adsorption (ΔG) for substituted benzenes on coronene.34 Nevertheless, most of the previous studies simulated the adsorption of organics on CNTs under vacuum conditions. To predict aqueous adsorption of organic pollutants on CNTs, the effects of water as a solvent should be considered. It was the purpose of this study to probe the interactions between SWNTs and organic compounds and to develop a computational method for estimating the thermodynamic adsorption coefficient (Kd) of aqueous aromatic pollutants on SWNTs by the calculation of ΔGcal. The M05−2X functional was employed to compute the equilibrium geometry, charge transfer, interaction energy, and some thermodynamic parameters.

Kd =

γ C αs = s s αe γe Ce

(1)

where αs and αe are the activities of the adsorbed solute and the solute in solution at equilibrium, respectively; Cs (mmol/L) is the amount of the adsorbate (mmol) adsorbed on the SWNTs per liter of the solution at equilibrium, Ce (mmol/L) is the equilibrium adsorbate concentration; and γs and γe represent the activity coefficients of the adsorbed solute and the solute in solution, respectively. Cs was calculated by Cs = q·m/V, where q (mmol/kg) is the surface concentration of the adsorbates on the SWNTs, m (kg) is the mass of the SWNTs, and V (L) is the volume of the solution. When the solute concentration approaches zero, the activity coefficient approaches unity. Equation 1 can be reduced as follows: Kd =

αs C = s αe Ce

(2)

Values of lnKd were obtained by plotting ln(Cs/Ce) versus Cs, and extrapolating Cs to zero.35,36 Theoretically, when Cs approaches zero, Ce approaches zero too. Thus, Kd is a concentration-independent and dimensionless thermodynamic parameter. The free energy change, ΔGexp, for the interaction was calculated by



COMPUTATIONAL METHODS Adsorbates. Cyclohexane and benzene derivatives (Table 1) were selected for the study as their experimental adsorption isotherms on SWNTs were determined at uniform conditions,12,13 which could be employed to estimate the experimental adsorption free energy changes (ΔGexp) for comparison with the calculated ones. In addition, phenol, aniline, and four PAHs were included as adsorbates so as to compare the effects of different functional groups on the adsorption. Estimation of ΔGexp. Theoretically, Kd is defined as

ΔGexp = −RT ln Kd

(3)

where R is the universal gas constant (8.314 J/mol·K), and T is the room temperature in the experiment. We took T = 298.15 K for the calculation. The experimental adsorption isotherms data,12,13 kindly provided by prof. Wei Chen at the Nankai University, were adopted to calculate Kd. As the adsorption isotherms deviate from linearity at relatively high concentrations, only the 6 8888

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

Figure 1. Equilibrium configurations for adsorption of the simulated compounds on the (8,0) SWNTs (dark gray: C; light gray: H; blue: N; red: O; green: Cl).

of imaginary frequencies. The thermodynamic parameters, including enthalpy, entropy, and Gibbs free energy, were calculated at 298.15 K and 1 atm. Atomic charges were calculated by the natural bond orbital (NBO)45 analysis to estimate the charge transfer (Δq)46 in the adsorption systems. Implicit solvent effects were taken into account by the integral equation formalism of the polarizable continuum model (IEFPCM)47 with the dielectric constant of water (78.39) to simulate the water environment. The IEFPCM can mimic the bulk solvent polarization effect of water.48 We also noted that explicit water molecules could have some influences on the configurations and energies. To evaluate the explicit solvent effects, we calculated the Ea and ΔGcal values for nitrobenzene− SWNT complex by adding 1−2 water molecules. The results (detailed in the SI) indicate that the explicit solvent molecules have marginal effects on the adsorption energy. Zhao et al.30 performed calculations on noncovalent complexes with and without the counterpoise corrections49 for the basis set superposition error (BSSE), and concluded that the M05−2X functional without the counterpoise corrections gave good agreement with the experimental results. Michalkove et al.34 also suggested not applying the BSSE correction in the prediction of thermochemical properties for noncovalent interacting systems when using the hybrid metageneralized gradient approximation density functional like M05−2X. Thus, the BSSE correction was not applied in this study. Thermodynamic Parameters. The adsorption energy (Ea), changes in the Gibbs free energy (ΔGcal), enthalpy (ΔH), and entropy (ΔS) were calculated by the following equations:

experimental data points at relatively low concentrations were employed. Thus, ΔGexp is an integrative and macroscopic thermodynamic parameter. Adsorbent. The pristine (8,0) zigzag SWNT (length 0.93 nm, outer diameter 0.63 nm) composed of 96 carbon and 16 saturated hydrogen atoms was selected as the adsorbent model. Such a zigzag SWNT with a similar size was also employed in the previous computational studies on adsorption of small molecules.38 As some studies indicated that truncation of a nanotube to a finite length could affect the reactivity of SWNTs,39,40 we also calculated the adsorption of benzene and nitrobenzene on (8,0) SWNTs with different length (0.93, 1.15, and 1.36 nm). The results (detailed in the Supporting Information (SI)) indicate that the current model is appropriate for the current study. DFT Computation. The Gaussian 09 program package41 with the M05−2X functional and cc-pVDZ basis set was employed for the computation. We tested the accuracy of M05−2X by comparing it with other methods such as B97-D,42 B3LYP, and B3LYP-D,43 and comparing the calculated ΔGcal values with ΔGexp for benzene, chlorobenzene, and nitrobenzene. The results (presented in the SI) indicated that the M05−2X method is adequate for the current study. Previous gaseous simulations on the interactions between CNTs and benzene,23,25 toluene,25 1,2-dichlorobenzene,26 and nitrobenzene25,44 showed that the phenyl plane tended to orientate parallel to the nanotube surface, forming π-stacked structures. Thus, the original configuration with phenyl (or cyclohexane) planes of the studied compounds paralleling to the nanotube surfaces were considered in the computation so as to decrease the computational load. The Cartesian coordinates of all the initial configurations are provided in the SI. All the structures were subjected to full geometry optimizations without any constraints. The minimum energy geometries were determined to be the true minima by absence 8889

Ea = ESWNT − M − ESWNT − EM

(4)

ΔGcal = GSWNT − M − GSWNT − GM

(5)

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

ΔH = HSWNT − M − HSWNT − HM

(6)

ΔS = (ΔH − ΔG)/T

(7)

where E stands for the sum of electronic and thermal energies, H represents the sum of electronic and thermal enthalpies, and G is the sum of electronic and thermal free energies for the optimized configurations; the subscript SWNT-M stands for the SWNT−small molecule interacting system, while M stands for organic molecules.



RESULTS AND DISCUSSION Adsorption Configurations and Mechanisms. All the optimized equilibrium configurations for the studied complexes are depicted in Figure 1, and the corresponding thermodynamic parameters are listed in Table 1. The adsorption energy Ea indicates the strength of the adsorption/interactions. We first calculated the gaseous Ea values for the adsorption of a few compounds on the (8,0) SWNTs so as to make a comparison with the previous results. Woods et al.25 and Fagan et al.26 calculated the gaseous Ea of selected compounds on (8,0) SWNTs using the local density approximation (LDA) functional. Our calculated gaseous Ea values for benzene, toluene, 1,2-dichlorobenzene, and nitrobenzene are consistent with their values.25,26 The computed Ea values in the aqueous phase are much lower than those in the gaseous phase (Table 1). Thus, the adsorption of the compounds on the regular surface of the SWNTs in the aqueous phase is more favorable. Kubicki et al.,50 who simulated the adsorption of naphthalene on soot, also found that the presence of water enhanced the interaction between the adsorbate and adsorbent. As the outer surface of individual CNTs provides evenly distributed hydrophobic sites for organic molecules, it was proposed that hydrophobic interactions played an important role in the interactions between CNT and organics including benzene derivatives12 and polycyclic aromatic hydrocarbons (PAHs).4,16 Actually, for benzene, toluene, chlorinated benzenes, and PAHs under study, their Ea values correlate with logKOW significantly (Figure 2):

Figure 2. Relationship between the adsorption energy (Ea) of the aromatics and their logKOW values.

interactions were also proposed by experimental studies52 and other DFT calculations,53 and identified by Raman band analysis.54 According to Figure 3, CH/π interactions also exist in the cyclohexane−SWNT adsorption system. However, the CH/π interaction is much weaker than the π−π interaction.27 At the MP2/cc-pVDZ level of theory, Tsuzuki et al.55 calculated the CH/π interaction energy in the benzene− methane complex, which is only −2.9 kJ/mol. For a hypothetical aromatic with the same logKOW value as cyclohexane, its pseudo Ea value estimated by eq 8 is −45.04 kJ/mol. The difference in the Ea values between the hypothetical aromatic and cyclohexane is −10.74 kJ/mol, implying that π−π interactions contribute ca. 24% of the interactions between the hypothetical aromatic and the SWNT. The four nitroaromatics deviate from the regression line too and exhibit much stronger adsorption than the other aromatics with similar or even higher logKOW values. For example, the adsorption of nitrobenzene is much stronger than that of benzene, toluene, and chlorobenzene, although it is less hydrophobic than these three aromatics. The differences in Ea values between the two nitroaromatics and two hypothetical aromatics with the same logKOW values are indicated in Figure 2, which shows the increase of Ea with the number of −NO2 substituents. Thus, −NO2 substituents significantly enhance the interactions. Rochefort and Wuest56 who employed the DFTLDA method to calculate the interactions of aromatic compounds with graphene, also found that nitrobenzene had stronger interactions with graphene than benzene. Following the same approach as for cyclohexane to calculate the contribution of π−π interactions, we estimated the contribution of −NO2 to Ea by comparing the Ea of nitrobenzene with that of a hypothetical aromatic with the same logKOW as nitrobenzene but without −NO2. The results indicate that −NO2 contributes ca. 24% to the Ea for nitrobenzene. Aniline and phenol have lower logKOW values than benzene. Nevertheless, they show much stronger adsorption than benzene. This implies that the functional groups (−NO2, −NH2, and −OH) can enhance electrostatic interactions between the aromatics and SWNTs. Tsuzuki et al. calculated the interaction between hexafluorobenzene and benzene, and also found that −F could enhance the electrostatic interactions.57 In addition, our calculation results show that

Ea = −22.89 − 6.44log K OW (n = 12, r = 0.97, p < 0.01)

(8)

Thus, the stronger the hydrophobicity of the chemicals, the stronger their interactions with the SWNTs. However, eq 8 does not necessarily imply that the interactions between the aromatics and the SWNTs involve hydrophobic interactions only. Due to the presence of π-electrons, the aromatics definitely have π−π interactions with the SWNTs, as was also pointed out in the previous studies.26,51 Cyclohexane has a similar logKOW value to 1,2-dichlorobenzene or 1,3-dichlorobenzene. According to eq 8, cyclohexane should have a Ea value similar to 1,2-dichlorobenzene/ 1,3-dichlorobenzene. However, the adsorption of cyclohexane is the weakest, as indicated by the lowest absolute Ea value and the longest equilibrium distance (Table 1). The reason is that cyclohexane lacks π−π interactions with the SWNTs. The contour plots of the total charge densities for cyclohexane, benzene, naphthalene, and nitrobenzene adsorbing on the SWNTs are shown in Figure 3. There are specific nonlocal electron correlations between the aromatic π electrons and the SWNTs, indicating the existence of π−π interactions. The π−π 8890

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

Figure 3. Contour plot of the total charge densities of (a) cyclohexane, (b) benzene, (c) naphthalene, and (d) nitrobenzene.

the stretching vibrational frequencies of O−H and N−H bonds change after the adsorption, from 3846.93 to 3842.85 cm−1 for phenol and from 3599.84 to 3543.48 cm−1 for aniline. This may indicate that O/N−H···π interactions exist in the phenol− and aniline−SWNT systems. The charge transfer results for the adsorption systems show that Δq in the complex formation is negligible (Table 1). It is worth mentioning that Kah et al.58 reported a linear relationship between CNT−water distribution coefficients (logKCNT) of PAHs in the low concentration range and logKOW. However, Pan et al.1 found a poor correlation between the adsorption coefficients of more than 20 organic chemicals with different functional groups on CNTs and their KOW. To observe the correlation between logKOW and logKd (or logKCNT) among different compounds, two basic conditions should be met: (a) logKd should be deduced from very low concentrations of adsorbates (Cs→0) as indicated by eq 2; and (b) for different adsorbates, their adsorption isotherms should be determined at consistent experimental conditions to eliminate errors from different experimental batches. Thermodynamic Parameters. The computed ΔGcal, ΔH, and ΔS values in the aqueous phase are listed in Table 1. The range of ΔH values indicates that the adsorption is physical and exothermic.59 All the ΔS values are 0.63 nm, it is to be expected that the experimentally determined macroscopic ΔGexp values were lower than the ΔGcal values. In addition, other experimental factors such as the ionic strength of the solution, and introduction of −CO, −COOH, and −OH functional groups by pretreatment of the SWNTs12 can also change the adsorption affinity of the SWNTs, which was not considered in the current study.

As ΔGexp and experimental Kd values are only available for a dozen organic compounds, the correlation can be employed to estimate Kd values of other chemicals, especially for the aromatics including the same functional groups as in this study, by calculating ΔGcal values. Nevertheless, it can be seen from Figure 4 that the absolute ΔGcal values are generally lower than the absolute ΔGexp values, indicating that the calculated Kd values are lower than the experimental Kd values. The calculation method and the implicit solvent model used here have uncertainties. As pointed out by Zhao et al.,30 the mean errors for the π−π interaction and the overall noncovalent interactions calculated by the M05−2X functional are ca. 2.51 and 1.38 kJ/mol, respectively. 8891

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

(2) Mauter, M. S.; Elimelech, M. Environmental applications of carbon-based nanomaterials. Environ. Sci. Technol. 2008, 42 (16), 5843−5859. (3) Ren, X. M.; Chen, C. L.; Nagatsu, M.; Wang, X. K. Carbon nanotubes as adsorbents in environmental pollution management: a review. Chem. Eng. J. 2011, 170 (2−3), 395−410. (4) Yang, K.; Xing, B. S. Adsorption of organic compounds by carbon nanomaterials in aqueous phase: Polanyi theory and its application. Chem. Rev. 2010, 110 (10), 5989−6008. (5) Niu, H. Y.; Cai, Y. Q.; Shi, Y. L.; Wei, F. S.; Liu, J. M.; Jiang, G. B. A new solid-phase extraction disk based on a sheet of single-walled carbon nanotubes. Anal. Bioanal. Chem. 2008, 392 (5), 927−935. (6) Belloni, F.; Kuetahyali, C.; Rondinella, V. V.; Carbol, P.; Wiss, T.; Mangione, A. Can carbon nanotubes play a role in the field of nuclear waste management? Environ. Sci. Technol. 2009, 43 (5), 1250−1255. (7) Rakov, E. G. The current status of carbon nanotube and nanofiber production. Nanotechnol. Russ. 2008, 3 (9−10), 575−580. (8) Mueller, N. C.; Nowack, B. Exposure modeling of engineered nanoparticles in the environment. Environ. Sci. Technol. 2008, 42 (12), 4447−4453. (9) Petersen, E. J.; Huang, Q. G.; Weber, W. J. Bioaccumulation of radio-labeled carbon nanotubes by Eisenia foetida. Environ. Sci. Technol. 2008, 42 (8), 3090−3095. (10) Petersen, E. J.; Pinto, R. A.; Landrum, P. F.; Weber, W. J. Influence of carbon nanotubes on pyrene bioaccumulation from contaminated soils by earthworms. Environ. Sci. Technol. 2009, 43 (11), 4181−4187. (11) Wiesner, M. R.; Lowry, G. V.; Jones, K. L.; Hochella, M. F.; Di Giulio, R. T.; Casman, E.; Bernhardt, E. S. Decreasing uncertainties in assessing environmental exposure, risk, and ecological implications of nanomaterials. Environ. Sci. Technol. 2009, 43 (17), 6458−6462. (12) Chen, W.; Duan, L.; Zhu, D. Q. Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environ. Sci. Technol. 2007, 41 (24), 8295−8300. (13) Chen, W.; Duan, L.; Wang, L. L.; Zhu, D. Q. Adsorption of hydroxyl- and amino-substituted aromatics to carbon nanotubes. Environ. Sci. Technol. 2008, 42 (18), 6862−6868. (14) Yang, K.; Wu, W. H.; Jing, Q. F.; Zhu, L. Z. Aqueous adsorption of aniline, phenol, and their substitutes by multi-walled carbon manotubes. Environ. Sci. Technol. 2008, 42 (21), 7931−7936. (15) Chen, J. Y.; Chen, W.; Zhu, D. Q. Adsorption of nonionic aromatic compounds to single-walled carbon nanotubes: effects of aqueous solution chemistry. Environ. Sci. Technol. 2008, 42 (19), 7225−7230. (16) Yang, K.; Zhu, L. Z.; Xing, B. S. Adsorption of polycyclic aromatic hydrocarbons by carbon nanomaterials. Environ. Sci. Technol. 2006, 40 (6), 1855−1861. (17) Ghaedi, M.; Hassanzadeh, A.; Kokhdan, S. N. Multiwalled carbon nanotubes as adsorbents for the kinetic and equilibrium study of the removal of alizarin red S and morin. J. Chem. Eng. Data 2011, 56 (5), 2511−2520. (18) Mishra, A. K.; Ramaprabhu, S.; Arockiadoss, T. Study of removal of azo dye by functionalized multi walled carbon nanotubes. Chem. Eng. J. 2010, 162 (3), 1026−1034. (19) Ji, L. L.; Chen, W.; Duan, L.; Zhu, D. Q. Mechanisms for strong adsorption of tetracycline to carbon nanotubes: a comparative study using activated carbon and graphite as adsorbents. Environ. Sci. Technol. 2009, 43 (7), 2322−2327. (20) Oleszczuk, P.; Pan, B.; Xing, B. S. Adsorption and desorption of oxytetracycline and carbamazepine by multiwalled carbon nanotubes. Environ. Sci. Technol. 2009, 43 (24), 9167−9173. (21) Efremenko, I.; Sheintuch, M. Predicting solute adsorption on activated carbon: phenol. Langmuir 2006, 22 (8), 3614−3621. (22) Abir, H.; Sheintuch, M. Atomistic calculation of adsorption in activated carbon with pore-size distribution. J. Colloid Interface Sci. 2010, 342 (2), 445−454. (23) Shim, Y.; Jung, Y.; Kim, H. J. Carbon nanotubes in benzene: internal and external solvation. Phys. Chem. Chem. Phys. 2011, 13 (9), 3969−3978.

(ii) The adsorption surfaces could in practice be quite heterogeneous due to the presence of various adsorption sites (e.g., groove and interstitial areas, inner pores) and impurities (e.g., MWNTs, CNTs with surface functional groups and defect sites). The groove or interstitial areas could enhance the adsorption affinity, as both sides of aromatic molecules could be involved in the π−π interactions with CNT bundles. For example, Michalkova et al.34 found that when benzene, naphthalene, and anthracene were moved into the middle between two coronene layers, their gaseous phase ΔGcal values decreased by 30.12, 26.36, and 39.33 kJ/mol, respectively. As SWNTs almost always have higher sorption capacities than MWNTs,1,12 it should be the other factors such as the groove and interstitial areas, but not the MWNTs, that cause the macroscopic ΔGexp values to be lower than the ΔGcal values. Computational toxicology and structure−activity relationships on nanomaterials have emerged as a challenging research field.28,29,64,65 In this study, we first calculated the aqueous adsorption configurations and thermodynamic parameters for a series of organic compounds on the SWNTs. The computed adsorption energies (Ea) indicate that the adsorption in the aqueous phase is more favorable than in gaseous phases. The contribution of the −NO2 enhancing effects and π−π interactions to the adsorption energy was quantified. A statistically significant correlation between the calculated microscopic ΔGcal and macroscopic ΔGexp values was reported, which together with the other thermodynamic parameters, could be employed to evaluate the adsorption of other compounds with no experimental adsorption data. Thus, this study may pave a new way for evaluating/predicting the adsorption of organic compounds on SWNTs and probing the adsorption mechanisms.



ASSOCIATED CONTENT

S Supporting Information *

Description of selection of the calculation method, effects of the SWNT length and the explicit presence of water molecules on the adsorption energy, Cartesian coordinates of the adsorption systems. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone/fax: +86-411-84706269; e-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Prof. Baoshan Xing (University of Massachusetts Amherst) for constructive comments and suggestions, prof. Wei Chen (Nankai University) for providing the experimental data, and prof. Willie Peijnenburg (RIVM) for improving the English expression. The study was supported by the National Natural Science Foundation of China (21007008, 20890113, 21137001) and the High-tech Research and Development Program of China (2012AA06A301).



REFERENCES

(1) Pan, B.; Xing, B. S. Adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 2008, 42 (24), 9005−9013. 8892

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

(24) Minami, D.; Ohkubo, T.; Kuroda, Y.; Sakai, K.; Sakai, H.; Abe, M. Structural optimization of arranged carbon nanotubes for hydrogen storage by grand canonical Monte Carlo simulation. Int. J. Hydrogen Energy 2010, 35 (22), 12398−12404. (25) Woods, L. M.; Badescu, S. C.; Reinecke, T. L. Adsorption of simple benzene derivatives on carbon nanotubes. Phys. Rev. B 2007, 75 (15), 155415. (26) Fagan, S. B.; Souza, A. G.; Lima, J. O. G.; Mendes, J.; Ferreira, O. P.; Mazali, I. O.; Alves, O. L.; Dresselhaus, M. S. 1,2dichlorobenzene interacting with carbon nanotubes. Nano Lett. 2004, 4 (7), 1285−1288. (27) Kar, T.; Bettinger, H. F.; Scheiner, S.; Roy, A. K. Noncovalent π-π stacking and CH···π interactions of aromatics on the surface of single-wall carbon nanotubes: an MP2 study. J. Phys. Chem. C 2008, 112 (50), 20070−20075. (28) Burello, E.; Worth, A. Computational nanotoxicology predicting toxicity of nanoparticles. Nat. Nanotechnol. 2011, 6 (3), 138−139. (29) Godwin, H. A.; Chopra, K.; Bradley, K. A.; Cohen, Y.; Harthorn, B. H.; Hoek, E. M. V.; Holden, P.; Keller, A. A.; Lenihan, H. S.; Nisbet, R. M.; Nel, A. E. The University of California Center for the Environmental Implications of Nanotechnology. Environ. Sci. Technol. 2009, 43 (17), 6453−6457. (30) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. Design of density functionals by combining the method of constraint satisfaction with parametrization for thermochemistry, thermochemical kinetics, and noncovalent interactions. J. Chem. Theory Comput. 2006, 2 (2), 364− 382. (31) Burns, L. A.; Vazquez-Mayagoitia, A.; Sumpter, B. G.; Sherrill, C. D. Density-functional approaches to noncovalent interactions: a comparison of dispersion corrections (DFT-D), exchange-hole dipole moment (XDM) theory, and specialized functionals. J. Chem. Phys. 2011, 134 (8), 084107. (32) Shukla, M. K.; Dubey, M.; Zakar, E.; Namburu, R.; Leszczynski, J. Density functional theory investigation of interaction of zigzag (7,0) single-walled carbon nanotube with Watson-Crick DNA base pairs. Chem. Phys. Lett. 2010, 496 (1−3), 128−132. (33) Shukla, M. K.; Dubey, M.; Zakar, E.; Namburu, R.; Czyznikowska, Z.; Leszczynski, J. Interaction of nucleic acid bases with single-walled carbon nanotube. Chem. Phys. Lett. 2009, 480 (4− 6), 269−272. (34) Michalkova, A.; Gorb, L.; Hill, F.; Leszczynski, J. Can the Gibbs free energy of adsorption be predicted efficiently and accurately: an M05−2X DFT study. J. Phys. Chem. A 2011, 115 (11), 2423−2430. (35) Biggar, J. W.; Cheung, M. W. Adsorption of picloram (4-amino3,5,6-trichloropicolinic acid) on Panoche, Ephrate, and Palouse soils: a thermodynamic approach to the adsorption mechanism. Soil Sci. Soc. Amer. Proc. 1973, 37, 863−868. (36) Yan, X. M.; Shi, B. Y.; Lu, J. J.; Feng, C. H.; Wang, D. S.; Tang, H. X. Adsorption and desorption of atrazine on carbon nanotubes. J. Colloid Interface Sci. 2008, 321 (1), 30−38. (37) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Envrionmental Orgainc Chemistry, 2nd ed.; John Wiley & Sons, Inc.: New York, 2003. (38) Peles-Lemli, B.; Matisz, G.; Kelterer, A. M.; Fabian, W. M. F.; Kunsagi-Mate, S. Noncovalent interaction between aniline and carbon nanotubes: effect of nanotube diameter and the hydrogen-bonded solvent methanol on the adsorption energy and the photophysics. J. Phys. Chem. C 2010, 114 (13), 5898−5905. (39) Bettinger, H. F. Effects of finite carbon nanotube length on sidewall addition of fluorine atom and methylene. Org. Lett. 2004, 6 (5), 731−734. (40) Hu, X. B.; Liu, C. Y.; Wu, Y. T.; Zhang, Z. B. Density functional theory study on nitrogen-doped carbon nanotubes with and without oxygen adsorption: the influence of length and diameter. New J. Chem. 2011, 35, 2601−2606. (41) Frisch, M. J.; TrucksG. W.Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;

Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A.; Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; 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.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision a.02; Gaussian, Inc.: Wallingford, CT, 2009. (42) Antony, J.; Grimme, S. Structures and interaction energies of stacked graphene-nucleobase complexes. Phys. Chem. Chem. Phys. 2008, 10, 2722−2729. (43) Grimme, S. Semiempirical GGA-type density functional constructed with a long-range dispersion correction. J. Comput. Chem. 2006, 27 (15), 1787−1799. (44) Badescu, S. C.; Alldredge, E. S.; Bajwa, N.; Perkins, F. K.; Snow, E. S.; Reinecke, T. L. Adsorption of nitro-substituted aromatics on single-walled carbon nanotubes. Phys. Rev. B 2010, 82 (12). (45) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Intermolecular interactions from a natural bond orbital, donor-acceptor viewpoint. Chem. Rev. 1988, 88 (6), 899−926. (46) Shtogun, Y. V.; Woods, L. M.; Dovbeshko, G. I. Adsorption of adenine and thymine and their radicals on single-wall carbon nanotubes. J. Phys. Chem. C 2007, 111 (49), 18174−18181. (47) Cances, E.; Mennucci, B.; Tomasi, J. A new integral equation formalism for the polarizable continuum model: theoretical background and applications to isotropic and anisotropic dielectrics. J. Chem. Phys. 1997, 107 (8), 3032−3041. (48) Miertus, S.; Scrocco, E.; Tomasi, J. Electrostatic interaction of a solute with a continuum. A direct utilization of AB initio molecular potentials for the prevision of solvent effects. Chem. Phys. 1981, 55 (1), 117−129. (49) Boys, S. F.; Bernardi, F. Calculation of small molecular interactions by differences of separate total energies - some procedures with reduced errors. Mol. Phys. 1970, 19 (4), 553−566. (50) Kubicki, J. D. Molecular simulations of benzene and PAH interactions with soot. Environ. Sci. Technol. 2006, 40 (7), 2298−2303. (51) Long, R. Q.; Yang, R. T. Carbon nanotubes as superior sorbent for dioxin removal. J. Am. Chem. Soc. 2001, 123 (9), 2058−2059. (52) Gotovac, S.; Hattori, Y.; Noguchi, D.; Miyamoto, J.; Kanamaru, M.; Utsumi, S.; Kanoh, H.; Kaneko, K. Phenanthrene adsorption from solution on single wall carbon nanotubes. J. Phys. Chem. B 2006, 110 (33), 16219−16224. (53) Zhao, J. J.; Lu, J. P.; Han, J.; Yang, C. K. Noncovalent functionalization of carbon nanotubes by aromatic organic molecules. Appl. Phys. Lett. 2003, 82 (21), 3746−3748. (54) Gotovac, S.; Honda, H.; Hattori, Y.; Takahashi, K.; Kanoh, H.; Kaneko, K. Effect of nanoscale curvature of single-walled carbon nanotubes on adsorption of polycyclic aromatic hydrocarbons. Nano Lett. 2007, 7 (3), 583−587. (55) Tsuzuki, T.; Honda, K.; Uchimaru, T.; Mikami, M.; Tanabe, K. The magnitude of the CH/π interaction between benzene and some model hydrocarbons. J. Am. Chem. Soc. 2000, 122 (15), 3746−3753. (56) Rochefort, A.; Wuest, J. D. Interaction of substituted aromatic compounds with graphene. Langmuir 2009, 25 (1), 210−215. (57) Tsuzuki, S.; Uchimaru, T.; Mikami, M. Intermolecular interaction between hexafluorobenzene and benzene: Ab initio calculations including CCSD(T) level electron correlation correction. J. Phys. Chem. A 2006, 110 (5), 2027−2033. (58) Kah, M.; Zhang, X. R.; Jonker, M. T. O.; Hofmann, T. Measuring and modeling adsorption of PAHs to carbon nanotubes over a six order of magnitude wide concentration range. Environ. Sci. Technol. 2011, 45 (14), 6011−6017. (59) Atkins, P. W.; Paula, J. D. Atkins’ Physical Chemistry, 7th ed.; Higher Education Press: P. R. China, 2002. 8893

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894

Environmental Science & Technology

Article

(60) Shen, X. E.; Shan, X. Q.; Dong, D. M.; Hua, X. Y.; Owens, G. Kinetics and thermodynamics of sorption of nitroaromatic compounds to as-grown and oxidized multiwalled carbon nanotubes. J. Colloid Interface Sci. 2009, 330 (1), 1−8. (61) Chin, C. J. M.; Shih, M. W.; Tsai, H. J. Adsorption of nonpolar benzene derivatives on single-walled carbon nanotubes. Appl. Surf. Sci. 2010, 256 (20), 6035−6039. (62) Stepanian, S. G.; Karachevtsev, M. V.; Glamazda, A. Y.; Karachevtsev, V. A.; Adamowicz, L. Stacking interaction of cytosine with carbon nanotubes: MP2, DFT and Raman spectroscopy study. Chem. Phys. Lett. 2008, 459 (1−6), 153−158. (63) Tournus, F.; Charlier, J. C. Ab initio study of benzene adsorption on carbon nanotubes. Phys. Rev. B 2005, 71, 16. (64) Kavlock, R. J.; Ankley, G.; Blancato, J.; Breen, M.; Conolly, R.; Dix, D.; Houck, K; Hubal, E.; Judson, R.; Rabinowitz, J.; Richard, A.; Setzer, R. W.; Shah, I.; Villeneuve, D.; Weber, E. Computational toxicology - a state of the science mini review. Toxicol. Sci. 2008, 103 (1), 14−27. (65) Rusyn, I.; Daston, G. P. Computational Toxicology: Realizing the promise of the toxicity testing in the 21st century. Environ. Health Perspect. 2010, 118 (8), 1047−1050.

8894

dx.doi.org/10.1021/es301370f | Environ. Sci. Technol. 2012, 46, 8887−8894