ARTICLE pubs.acs.org/JPCA
Hybrid Quantum Mechanical/Molecular Mechanics Study of the SN2 Reaction of CH3Cl+OH in Water Hongyun Yin, Dunyou Wang,*,† and Marat Valiev*,‡ † ‡
College of Physics and Electronics, Shandong Normal University, Jinan 250014, People's Republic of China Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, MS-IN: K8-91, P.O. Box 999, Richland, Washington 99352, United States ABSTRACT: The SN2 mechanism for the reaction of CH3Cl + OH in aqueous solution was investigated using combined quantum mechanical and molecular mechanics methodology. We analyzed structures of reactant, transition, and product states along the reaction pathway. The free energy profile was calculated using the multilayered representation with the DFT and CCSD(T) level of theory for the quantum-mechanical description of the reactive region. Our results show that the aqueous environment has a significant impact on the reaction process. We find that solvation energy contribution raises the reaction barrier by ∼18.9 kcal/mol and the reaction free energy by ∼24.5 kcal/mol. The presence of the solvent also induces perturbations in the electronic structure of the solute leading to an increase of 3.5 kcal/mol for the reaction barrier and a decrease of 5.6 kcal/mol for the reaction free energy, respectively. Combining the results of two previous calculation results on CHCl3 + OH and CH2Cl2 + OH reactions in water, we demonstrate that increase in the chlorination of the methyl group (from CH3Cl to CHCl3) is accompanied by the decrease in the free energy reaction barrier, with the CH3Cl + OH having the largest barrier among the three reactions.
’ INTRODUCTION The chlorinated hydrocarbons (CHCs) present one of the major sources of environmental pollution. They are widely used in industry and agriculture in making synthetic solvents, electrical insulators, and pesticides. Many of these compounds are either known or suspected carcinogens posing significant health risks. The long degradation time of these compounds2,3 obstruct the experimental characterization, leading to significant uncertainty in understanding reaction mechanisms. Several theoretical47 and experimental810 investigations about their reaction process with OH have been done. However, only gas phase studies have been performed so far, including two recent electronic structure studies.1,11 Although the recent study12 has taken the effect of solvent into account, it only focuses on the reaction of CH2Cl2 + OH in water. As these investigations show us the reaction mechanisms of CHCs + OH in their actual environment, aqueous solution, need to be explored. Taking the explicit solvent into consideration makes the calculation of reaction process quite formidable due to sampling over a large number of degrees of freedom introduced by the solvent. Direct quantum mechanical (QM) description is impractical because of the huge number of calculations during solvent averaging. This problem can be simplified by using hybrid quantum mechanical/molecular mechanics (QM/MM) methodology.1315 Here the chemical system is partitioned into reactive region (solute) that requires a quantum treatment (QM) and the remainder (solvent) whose chemical identity preserved and can be represented r 2011 American Chemical Society
through classical description (MM). Nonetheless, even with this simplified description, the calculation of the free energy makes the calculation still rather expensive. So in many practical cases, the QM theory is restricted to density functional theory (DFT) or semiempirical methods. This problem is circumvented in this work using layered16,17 QM/MM methodology for free energy calculations. This approach combines Coupled Cluster theory (CC),18,19 density functional theory (DFT),20 and molecular mechanics (MM) methods to construct a sequence of multiple representations of the reactive region. This allows us to get a higher accuracy of the free energy by treating the QM region with the CC method and shift the bulk of statistical sampling to less expensive and more efficient descriptions. Following our previous work CHCl3 + OH and CH2Cl2 + OH reactions in water, we apply this approach to study the second-order bimolecular nucleophilic substitution (SN2) reaction, CH3Cl + OH f CH3OH + Cl in water. So far, the reaction has only been studied in gas phase,1,11 and little is known about its behavior in aqueous environment. We investigate the structures of the reactant, transition, and product states for the reaction in aqueous phase, explore the reaction mechanism using the nudged elastic band (NEB)21 method, and compute the free Received: August 10, 2011 Revised: September 14, 2011 Published: September 26, 2011 12047
dx.doi.org/10.1021/jp2076808 | J. Phys. Chem. A 2011, 115, 12047–12052
The Journal of Physical Chemistry A
ARTICLE
energy profile along reaction pathway up to CCSD(T) level of theory for the QM description.
’ COMPUTATIONAL METHODOLOGY AND SYSTEM SETUP In this work all the calculations of the reaction were performed using the hybrid QM/MM approach in the NWChem computational chemistry package.22 Detailed explanation of the QM/ MM approach has been given in the previous publications.17,23,24 Here the method is presented briefly. For a combined solute and solvent system, the total energy can be expressed as a sum of QM (solute) and MM (solvent) contributions, E ¼ Eqm ðr, R; ψÞ þ Emm ðr, RÞ
ð1Þ
where r and R are the coordinates of the QM and MM regions, respectively, and ψ represents the solute electronic wave function corresponding to the ground state. The QM energy, Eqm, can be further decomposed into the internal and external contributions,17,23,24 ext Eqm ¼ Eint qm ðr; ψÞ þ Eqm ðr, R; FÞ
ð2Þ
coincides with the Here the first part of the QM energy, expression for solute energy in the gas phase. The external part of the QM energy in eq 2 can be approximated by the equivalent effective electrostatic potential (ESP) representation, Eint qm,
Eext qm ¼
∑i, I jR I I iri j EESP ðr, R; Q Þ ZQ
ð3Þ
Here the QM atoms are represented by ESP charges (Q i) such that the electrostatic potential outside the solute region is the same as that produced from the full electron density F. To reduce the computational expense associated with free energy calculations, especially at the CC-level of theory for the QM description, we utilize layered QM/MM representations consisting of CCSD(T)/MM, DFT/MM, and ESP/MM.25 The first two means that we treat the QM region using CCSD(T) and DFT level theory, respectively. The ESP/MM representation is obtained by replacing Eqm in eq 1 by the effective charge energy EESP (eq 3). Using thermodynamic cycles, the burden of statistical sampling can then be shifted to a less expensive representation of the solutesolvent interactions.25,28 The desired free energy difference ΔWAB at the CC/MM level of theory can be then represented as follows:17 CC f DFT CC f DFT ΔWAB ¼ ðΔWAA ΔWBB Þ DFT f ESP DFT f ESP ESP þ ðΔWAA ΔWBB Þ þ ΔWAB
ð4Þ
The first and second terms in brackets represent free energy difference for changing the description from the CC to DFT representations and from DFT to a classical ESP representations at the fixed solute configuration (A or B), respectively. These terms are approximated in our work through the internal energy differences.17 The third term gives the free energy difference for changing solute configuration from A to B within classical ESP/ MM description, and is calculated through explicit sampling. In our investigation the reactive (QM) region consisted of CH3Cl/OH complex. Its initial geometry was constructed based on the structures obtained in previous studies1,11 and solvated in a 35.2 Å cubic box consisting of 1438 water molecules, treated as the MM region. The QM treatment was based on DFT
and CCSD(T) levels of theory, leading correspondingly to DFT/ MM or CC/MM representations. DFT calculations utilized the B3LYP26,27 exchange correlation functional, and the aug-ccpvDZ basis is used for both the DFT and CCSD(T) calculations. Standard Amber force field29 van der Waals parameters are used for the quantum region. The classical MM representation for water solution was based on SPC/E model,30 and the cutoff radius for classical interactions is 15 Å. After full optimization of the system, molecular dynamics simulation was carried out to equilibrate the solvent for 40 ps at 298.15K. During this step the QM region was fixed and represented by the point ESP charges calculated in the prior optimization step. After the equilibration, the full system was again fully reoptimized to produce the initial optimized reactant complex in water. In the next step we determined the transition state using initial reactant complex obtained in the above step. First, an initial guess for reaction pathway was constructed by simulating the bond breaking of CCl and the bond formation of the C atom to the O atom in OH using constrained optimization, while all the other degrees of freedom in the QM region are freely evolving. Second, the initial pathway from the reactant to product was refined using nudged elastic band (NEB)21 method as described in the previous QM/MM approach.14,15 Third, the geometry on the top of the NEB reaction pathway was isolated and used for the saddle point search calculation. The located transition state structure was verified through the numerical frequency calculations that showed one imaginary frequency. Reactant and product complexes were then determined by optimizing initial structures obtained by the corresponding displacements of the transition state along the negative frequency mode. Then using the obtained final reactant and product complexes, the NEB pathway was constructed again with 10 geometries along the reaction profile. Molecular dynamics simulations were performed on each NEB bead along the reaction pathway to equilibrate the solvent for 40 ps. We followed this by carrying out another round of the NEB reaction pathway calculation until the results converge. Finally, the free energy profile along the converged reaction pathway was calculated under the CCSD(T)/MM presentation. Since we adopted a multilayered, ESP, DFT, CC, representation, the free energy under CCSD(T)/MM presentation consists of the quantum internal energy difference from the DFT representation to the CCSD(T) representation, quantum internal energy difference from the DFT representation to the ESP representation and the free energy contribution from the MM region (See eq 4). As mentioned above, 10 points are chosen along the free energy profile. Between each two points on the reaction pathway, the free energy difference between two adjacent NEB points for the ESP/MM representation was sampled for 50 ps to reach convergence.
’ RESULTS AND DISCUSSIONS Reactant State. Figure 1 shows the final optimized structure of the reactant state complex for the CH3Cl + OH reaction in water. It has a structure similar to that obtained in gas phase.1,11 The presence of the solvent makes little impact on the internal structure of CH3Cl. The CCl bond length is 1.822 Å, compared to 1.821.83 Å in gas phase; the three CH bond lengths are 1.092 Å, 1.092 Å and 1.093 Å, which are very close to the value of 1.12 Å of the CH bond in gas phase. The biggest difference 12048
dx.doi.org/10.1021/jp2076808 |J. Phys. Chem. A 2011, 115, 12047–12052
The Journal of Physical Chemistry A
ARTICLE
Figure 1. The structure of the reactant, transition, and product states for the reaction, CH3Cl + OH f CH3OH + Cl in aqueous phase. The indicated distances are in Angstroms.
Figure 2. The structure of reaction CH3Cl + OH f CH3OH + Cl along the reaction pathway in water, No. 1 is for the structure of the reactant state, No. 5 for the transition state, and No. 10 represents the product state for the reaction. The indicated distances are in Angstroms.
between the gas phase and in aqueous phase is the H...OH distance, it is much longer at 2.149 Å in aqueous phase compared to 1.77 Å gas phase. The OH forms four stronger hydrogen bonds with average hydrogen bond distance of around 2.60 Å and also forms a fifth weaker hydrogen bond with a distance of 2.87 Å. The similar situation is also observed in other experimental31 and theoretical32 studies of OH in aqueous phase. In both studies,31,32 the authors found that the OH directly bonds with four water molecules with an additional fifth water molecule weakly bound to the hydroxyl group. This indicates that our methodology applied here correctly describes the solvation of the hydroxyl group. Here the solvent shielding weakens the H...OH bond leading to a longer bond length. This shows that water solution reduces the interaction between CH3Cl and OH. Transition State. The transition state, also shown in Figure 1, is the No. 5 bead on the free energy NEB reaction pathway. Numerical frequency calculations confirm the presence of single imaginary frequency of 425i cm1. The structure of the transition state is characterized by the formation of a nearly planar CH3 group. The most noticeable changes are the loss of hydrogen bond from the OH, which rotates to face the C atom with the CO distance at 2.29 Å, and the elongation of CCl distance from 1.82 Å to 2.57 Å . These two distances are both longer than
in the gas phase1,11 with the attacking group RCO = 2.24 Å and the leaving group RCCl =2.06 Å . This is, again, caused by the solvation shielding effects. We also found, that the charge is still mostly concentrated on the OH, 90% compared to the reactant. The solvation pattern for OH consists of three donor and one acceptor hydrogen bonds. Product State. At the product state, the nucleophilic substitution process is finished with the Cl totally detached from the C atom with the distance at 3.916 Å. The leaving Cl group now has the full negative charge and is surrounded by seven water molecules with average ClO distance of 3.2 Å. This feature is consistent with theoretical33 and experimental34 estimates of Cl solvation. On the other hand, the attacking OH group, which formed CH3OH, only has two weak acceptor bonds (2.752 Å and 2.893 Å). We should note that in aqueous solution solvent caging effects play a key role in the stabilization of the Cl...CH3OH complex, which is otherwise unstable in gas phase. Reaction Pathway. In order to examine the reaction mechanism in detail, the entire reaction process is shown in Figure 2. This figure shows the CCl bond breaking and CO bond forming as ten snapshots along the NEB reaction pathway. We observe that the formation of the transition state from the reactant complex is accomplished through the rotation of OH group, the 12049
dx.doi.org/10.1021/jp2076808 |J. Phys. Chem. A 2011, 115, 12047–12052
The Journal of Physical Chemistry A
ARTICLE
Figure 3. The evolution of CO and ClC bond distances (Å) along the NEB reaction pathway.
Figure 5. Comparison between gas-phase and internal QM/MM energies along the NEB pathway under the CCSD(T)/MM representation using the gas-phase energy of the reactant state as a reference point.
Figure 4. Comparison of free energy profiles calculated at DFT/MM and CCSD(T)/MM levels of theory and solvation contribution using the reactant state (bead 1) as a reference point.
breaking of the hydrogen bond and OH attack on the central C atom, and displacement of the Cl leaving group to form the product state. From Figure 3, we can observe that the reaction is initiated by the attack of OH. The departure of Cl atom commences only when the CO distance reaches 2.5 Å. After that point, both distances change nearly at the same rate until the CO bond is fully formed (point 7). The increase of CCl distance is then continued until it reaches its final value of 3.9 Å at the product state. The overall process is clearly associative as would be expected for the SN2 reaction mechanism. The Free Energy. The free energy profile along the NEB reaction pathway is shown in Figure 4 and includes DFT/MM and CCSD(T)/MM results, as well as solvation free energy (last term in eq 4). Here the number 1 bead is used as the reference energy point, so only relative free energies along the pathway are calculated. In Figure 5 and Figure 6 we present the gas phase energies along the NEB reaction pathway against the internal QM/MM energies (first two terms in eq 4) using the energy of the reactant state in gas phase as a zero reference point. Our results show that free energy activation barriers are 46.4 kcal/mol with DFT/MM and 49.9 kcal/mol with CCSD(T)/ MM representations, which are significantly higher than the corresponding gas phase value of 27.4 kcal/mol (Figure 4). This indicates a much slower reaction rates in solution than in gas
Figure 6. Comparison between gas-phase and internal QM/MM energies along the NEB pathway under the DFT/MM representation using the gas-phase energy of the reactant state as a reference point.
phase for this reaction. The reaction free energy 22.7 kcal/mol at DFT/MM and 20.0 kcal/mol at CCSD(T)/MM are only half of that in gas phase of 44.6 kcal/mol. The impact of aqueous environment on reaction energetics arises from two contributions, one from the solvation energy, the other from solute polarization effects as seen in Figures 4, 5, and 6, respectively. Figure 4 shows that the solvation contribution, using reactant state as the reference energy point, contributes ∼18.9 kcal/mol to the transition state barrier height and ∼24.5 kcal/mol to product state. It does, as one would expect, play a major role in shaping the reaction process in aqueous phase. On the basis of the early published results of the solvation energies of CH3OH (3.19 kcal/mol35,36), CH3Cl (1.3 kcal/mol36), OH(105 kcal/mol)37 and Cl (74.7 kcal/mol),37 we can obtain a solvation reaction energy of 25.8 kcal/mol which agree well with our calculated value of 24.5 kcal/mol . In order to examine the solute polarization effects, which are caused by the perturbation of solute wave function by the solvent, we compare the gas phase and internal QM/MM energy (see eq 2) 12050
dx.doi.org/10.1021/jp2076808 |J. Phys. Chem. A 2011, 115, 12047–12052
The Journal of Physical Chemistry A
ARTICLE
Table 1. Comparison of the Rate Constants in Gas Phase, The Combined Solvation Free Energy and Polarization Effect Contribution to the Reaction Barrier Heights in Water and the Reaction Barrier under CCSD(T)/MM Representation for the Reactions OH + CHCl3 OH + CH2Cl2, and OH + CH3Cl
which according to gas phase experimental data38 (see Table 1), decreases with the reduced chlorination.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected];
[email protected].
rate constants aqueous environment total free energy in gas phase contribution to barrier activation barrier (cm3/molec-s)
reaction
CHCl3 + OH
8.8 10
14 38
CH2Cl2 + OH 8.2 1014 38 CH3Cl + OH 3.1 1014 38
(kcal/mol) 17
height (kcal/mol) 17
10.7
29.3
17.312 22.4
31.812 49.9
under CCSD(T)/MM presentation in Figure 5. Here the reactant state energy of the gas-phase is used as the reference energy point. The results indicate that the CCSD(T)/MM level of theory solute polarization effects raise the energy of the reactant state by ∼9.2 kcal/mol, transition state by 12.7 kcal/mol, and product state by 3.6 kcal/mol compared to gas-phase. The net result is the increase of the transition state barrier height about 3.5 kcal/mol and reduction in the reaction energy about 5.6 kcal/mol comparing to the gas phase. With the DFT/MM level of theory (see Figure 6), the polarization effects are somewhat different ∼10.7 kcal/mol for the reactant complex, 10.8 kcal/mol for the transition state, and 8.2 kcal/mol for the product. The net result is only 0.1 kcal/mol increase on the barrier height and 2.5 kcal/mol deduction for the reaction energy. We should note, that the impact of the polarization effects of the water solvent, including the first hydration shell, is captured by statistical averaging over different solvent configurations where water molecules are allowed to dynamically rearrange themselves in the proper field of the solute.
’ CONCLUSIONS Using a QM/MM approach, we investigated the reaction mechanism of CH3Cl + OH. Reactant, transition state, and product state structures were characterized, and the free energy profile was calculated using CCSD(T)/MM and DFT/MM representations. Our results show that the free energy activation barrier in solution is significantly higher (49.9 kcal/mol CCSD(T)/ MM) than that in the gas phase (27.4 kcal/mol CCSD(T)/MM), suggesting long degradation times. As expected the main effect of the aqueous environment comes from the solvation energy, which disfavors charge transfer from OH to Cl. The DFT description of the reactive region seems to provide adequate description of the overall energetics with the average error of around 3 kcal/mol. Compared to previous studies of chlorinated hydrocarbons, we clearly see a trend of the increase in free energy activation barriers (see Table 1) with the decreased chlorination of the methyl group (CH4‑nCln, n = 1,2,3). One of the effects is related the solvation energy contribution. For all three compounds (CHCl3, CH2Cl2, CH3Cl) the net effect of solvation energy is to raise free energy activation barrier, however the amount of this increase lessens with the increased chorination (see Table 1). This can be rationalized based on the electronegativity of H and Cl substituents with more electronegative Cl increasing the charge polarization of the transition state complex, which is favorable in terms of the solvation effects. Further contribution comes from the intrinsic reactivity of chlorinated methyl group,
’ ACKNOWLEDGMENT D.W. thanks the National Natural Science Foundation of China (Grant No. 11074150) and Tanshai Scholarship funding for supporting this work. Part of computational work was carried out at the Shanghai Supercomputer Center. Work at PNNL was supported by the U.S. Department of Energy’s (DOE) Office of Basic Energy Sciences, Chemical Sciences program, and was performed in part using the Molecular Science Computing Facility (MSCF) in the William R. Wiley Environmental Molecular Sciences Laboratory, a DOE national scientific user facility located at the Pacific Northwest National Laboratory (PNNL). PNNL is operated by Battelle for DOE. ’ REFERENCES (1) Suyong, R.; Morokuma, K. Theor. Chem. Acc. 2004, 112, 59. (2) Vogek, T. M.; Criddle, C. S.; McCarty, P. L. Environ. Sci. Technol. 1987, 21, 722. (3) Schwarzenbach, R. P.; Gashwend, P. M. in Aquatic Chemical Kinetics; Stumm,W., Ed.; Wiley-Intersicence; New York, 1990; pp 199233. Schwarzenback, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry; Wiley: New York, 1993. (4) Evanseck, J. D.; Blake, J. F.; Jorgensen, W. L. J. Am. Chem. Soc. 1987, 109, 2349. (5) Ohta, K.; Morokuma, K. J. Am. Chem. Soc. 1987, 89, 5845. (6) Pliego, J. R.; Almeida, W. R. D. J. Phys. Chem. 1996, 100, 12410. Pliego, J. R.; Almeida, W.R. D. Chem. Phys. Lett. 1996, 249, 136. (7) Marrone, P. A.; Arias, T. A.; Peters, W. A.; Tester, J. W. J. Phys. Chem. A 1998, 102, 7013. (8) Hine, J. J. Am. Chem. Soc. 1954, 72, 2438. Hine, J.; Dowell, A. M. J. Am. Chem. Soc. 1954, 76, 2688. (9) Henchman, M.; Hierl, P. M.; Paulson, J. F. J. Am. Chem. Soc. 1985, 107, 2812. (10) Hierl, P. M.; Paulson, J. F.; Henchman, M. J. Phys. Chem. 1995, 99, 15655. (11) Borisov, Y. A.; Arcia, D. E.; Mielke, S. L.; Garrett, B. C.; Dunning, T. H., Jr. J. Phys. Chem. A 2001, 105, 7724. (12) Wang, D.; Valiev, M.; Garrett, B. C. J. Phys. Chem. A 2011, 115, 1380–1384. (13) Bowie, J. H. Acc. Chem. Res. 1980, 13, 76. (14) Gao, J. L.; Truhlar, D. G. Annu. Rev. Phys. Chem. 2002, 53, 467–505. (15) Warshel, A. Ann. Rev. Biophys. Biomol. Struct. 2003, 32, 425–443. (16) Valiev, M.; Kowalski, K. J. Chem. Phys. 2006, 125, 211101. (17) Valiev, M.; Garrett, B. C.; Tsai, M.-K.; Kowalski, K.; Kathmann, S. M.; Schenter, G. K.; Dupuis, M. J. Chem. Phys. 2007, 127, 051102. (18) Bartlett, R. J.; Stanton, J. F. in Reviews of Computational Chemistry; Lipkowitz, K. B., Boyd, D. G., Eds.; VCH Publisher: New York, 1995. (19) Bartlett, R. J.; Musial, M. Rev. Mod. Phys. 2007, 79, 271. (20) Hohenberg, P.; Kohn, W. Phys. Rev. 1964, 136, B864. Kohn, W.; Sham, L. J. Phys. Rev. 1965, 140, A1133. (21) Henkelman, G.; Uberuaga, B. P.; Jonsson, H. J. Chem. Phys. 2000, 113, 9901. (22) Valiev, M; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T. P.; van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L.; de Jong, W. A. Comput. Phys. Commun. 2010, 181, 1477. 12051
dx.doi.org/10.1021/jp2076808 |J. Phys. Chem. A 2011, 115, 12047–12052
The Journal of Physical Chemistry A
ARTICLE
(23) Valiev, M.; Bylaska, E. J.; Dupuis, M.; Tratnyek, P. G. J. Phys. Chem. A 2008, 112, 2713. (24) Valiev, M.; Yang, J.; Adams, J. A.; Taylor, S. S.; Weare J. Phys. Chem. B 2007, 111, 1345. (25) Zhang, Y. K.; Liu, H. Y.; Yang, W. T. J. Chem. Phys. 2000, 112, 3483. (26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (27) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (28) Henkelman, G.; Uberuaga, B. P.; Jonsson, H. J. Chem. Phys. 2000, 113, 9901. (29) Fox, T.; Kollman, P. A. J. Phys. Chem. B 1998, 102, 8070. (30) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. J. Phys. Chem. 1987, 91, 6269. (31) Botti, A.; Brunti, F.; Imberti, S.; Ricci, M. A.; Soper, A. K. J. Mol. Liq. 2005, 117, 81. (32) Tuckerman, M. E.; Marx, D.; Parrinello, M. Nature 2002, 417, 925. (33) Koneshan, S.; Rasaiah, J. C.; Lynden-Bell, R. M.; Lee, S. H. J. Phys. Chem. B 1998, 102, 4193. Tongraar, A.; Rode, B. M. Phys. Chem. Chem. Phys. 2003, 5, 357. (34) Ohtaki, H.; Radnai, T. Chem. Rev. 1993, 93, 1157. (35) Bylaska, E. J.; Dixon, D. A.; Felmy, A. R. J. Phys. Chem. A 2000, 104, 610. (36) Munz, C.; Roberts, P. V. Environ. Sci. Technol. 1986, 20, 830. (37) Pliego, J. R., Jr; Riveros, J. M. Chem. Phys. Lett. 2000, 332, 597. (38) Hsu, K.-J.; DeMore, W. B. Geophys. Res. Lett. 1994, 21, 805–808.
12052
dx.doi.org/10.1021/jp2076808 |J. Phys. Chem. A 2011, 115, 12047–12052
