About two centuries ago, halogen bonding (XB) story started with Colin's discoveries about the complexes of iodine-amilose and iodine-ammonia [1, 2]. Thereafter, a lot of efforts were made toward understanding of the nature of XB interactions by different experimental techniques and theoretical calculations, and the conceptual influences and applications of XB were found in crystal engineering and molecular assembling . In 2013, the International Union of Pure and Applied Chemistry recommended a definition that “A halogen bond occurs when there is evidence of a net attractive interaction between an electrophilic region associated with a halogen atom in a molecular entity and a nucleophilic region in another, or the same, molecular entity" . A typical XB interaction can be denoted as R-X···Y, where the three dots represent the intermolecular interaction, R-X is the XB donor, X is a halogen atom (F, Cl, Br, I) covalently bound to the R group, and usually Y is N, O, S, Se, ··· atom in the XB acceptor. Our previous study indicated that the XB interaction of X···Y was primarily the electrostatic attraction, on the basis of interaction energy decomposition analyses with symmetry-adapted perturbation theory .
Most of the previous theoretical studies focused on the details of XB interaction of R-X···Y complex in gas phase , very few investigations about its properties in liquid or solution were reported although it was originally found in the liquid system [1, 2]. In an aqueous system, another important intermolecular interaction, hydrogen bonding (HB), is presented. It is interesting about the competition between the HB and XB interactions in the R-X···water system. However, the strength of XB interaction is usually much weaker than the HB interaction. Therefore, in the aqueous solution, the HB interaction usually plays an overwhelming role in the molecular assembling and crystallization, even though the XB interaction possibly co-exists.
In this work, the XB interaction of Cl···O is investigated for the pre-reactive system ClF-water in liquid phase. The predominant products of the reactions between ClF and water were HF, ClO2, Cl2O , implying that the XB interaction between Cl and O atoms should be responsible for the products of ClO2 and Cl2O. In our previous study, the XB interaction energy -4.16 kcal/mol (at CCSD(T) level of theory) of Cl···O in the H2O···ClF complex was predicted , while the HB interaction of Cl···H is quite weak. Here the molecular structures of the aqueous ClF system are studied with classical molecular dynamics simulations, particularly we pay more attentions to the coexistence of the XB and HB interactions.II. Theoretical Method
In the present simulations, all-atom parameters of the force fields for ClF and water were used. The TIP3P model of water molecule was adopted directly . To obtain the parameters of ClF, firstly we optimized the geometrical parameters of ClF and water molecules at the second-order Mller-Plesset perturbation level of theory (MP2) within the frozen core approach. The aug-cc-pVTZ basis set was employed. Then, as shown in Fig. 1, three potential energy curves (PEC1-3) for the ClF···H2O were calculated at the same level and corrected by the basis-set superposition error. The PECs were fitted to the Lennard-Jones (12-6) plus Coulomb potential,
where the parameters σij and εij were calculated following the rules as, σij=(σiσj)1/2 and εij = (εiεj)1/2. The σi and εi values were optimized for Cl and F atoms but those for the atoms of water were cited from the TIP3P model. The atomic charges q of ClF molecule were fitted using the Mulliken charges. The data fitting quality was inspected with the penalty function ,
where N was the total number of data and wk was a weight factor being set constantly to 1 kcal-2·mol2. The quantum mechanical calculations were carried out by the Gaussian 09 revision D.01 suite program .
The classical molecular dynamics simulations were performed with the GROMACS 4.5.7 program package [10, 11]. Initially, ClF was solvated by 1000 water molecules in a cubic box imposed with the periodic boundary condition. Then, 1.0 ns simulation in the NPT ensemble was performed at 1 atm and 300 K with a 2 fs step. The Berendsen coupling algorithm  was used to impose the constant temperature and pressure conditions. Finally, 10.0 ns simulation in the NPT ensemble was carried out at 1 atm and 300 K with a 2 fs step. The Nosé-Hoover thermostat [13, 14] with a time constant of 0.5 ps was employed to regulate the temperature. The Parrinello-Rahman coupling algorithm  was used to impose the constant pressure. The cutoff distance for van der Waals interactions was 12.0 Å, and long-range electrostatic interactions were treated by the particle-mesh Ewald method  implemented in GROMACS. The LINCS algorithm  was used in all the simulations, to constrain all bonds and angles, keeping the rigid structures of the solute and solvent molecules.III. Results and discussion
As shown in Fig. 1, the most attractive interaction occurs when ClF approaches the O atom along the angular bisector of water molecule (PEC1); while the interaction when H2O molecule attacks ClF in a direction perpendicular to the ClF bond is nearly repulsive (PEC3). The minima are -4.60, -0.70, -0.27 kcal/mol on PEC1, PEC2, PEC3, respectively. The attraction energy of -4.60 kcal/mol is a little larger than that predicted at the higher level of theory CCSD(T) before . When fitting the PECs to obtain the atomic ε and σ parameters of ClF, the parameters for water from the TIP3P model  should be kept simultaneously. First the Lennard-Jones parameters for ClF were obtained by fitting all points on the PECs lower than 2.5 kcal/mol with a penalty f value of 0.11. All individual error is lower than 0.2 but more pronounced for PEC2 (f=0.18). The fitting errors for PECs 1 and 2 are mainly influenced by the σ values of Cl and F atoms jointly. The fitted potentials are compared with the MP2 calculated data in Fig. 1, exhibiting a general satisfactory. The parameters for ClF together with TIP3P of water in Table I represent the average values to minimize the overall f functions.
The solvation structure around ClF molecule are depicted as the radial distribution function g(r) in Fig. 2(a). The g(r) curve varies with the distance (r) between the center-of-mass of ClF and H2O, exhibiting a maximum at r=3.0 Å and a shoulder at r=3.5 Å. This solvation shell structure is closely related to the details of XB and HB interactions discussed in the following parts.
Figure 2(b) shows more details about the hydration structure around Cl and F atoms. The g(r) peak positions for the HB (solid curve) and XB (broken curve) interactions are distinctly different, and the values are listed in Table II. A sharp peak at 2.6 Å represents the XB interactions of Cl···O, while a broad peak at 3.3 Å corresponds to the HB interactions of Cl···H. It is noted that the Cl···O bonding distance of 2.6 Å is in good agreement with 2.5174 Å predicted at the higher level of theory CCSD(T)  and the 2.608 Å determined with the experiment for the ClF···H2O complex in the gas phase . The average number 〈N〉 and average lifetime 〈τ〉 of HB and XB are listed Table II. They are calculated in statistics with the whole molecular dynamics trajectories performed in the NVT ensemble. Their criteria for an intermolecular bonding interaction are: the distance between Cl (or F) and O (or H) atoms is less than 4.0 Å and the angle O (or H) Cl-F is less than 30°. The g(r) curves show the statistical structure of a large quantity of water molecules, while the relatively small quantity of the water molecules surrounding the solute can be clearly depicted. One can find that the values of 〈N〉 and 〈τ〉 for the HB interaction of F···H are relatively larger, while those for the HB interaction of Cl···H are comparable with the XB interaction of Cl···O. The values for the F···O interactions are too small, thus the XB between F and O can be ignored. This is in line with a rule established well that the XB strength scales with the polarizability of the XB donor atom, that is, F<Cl<Br<I .
Although the g(r) curves provide us useful information on a radial structure of solution, to fully understand the spatial structures, the angular dependent information is also necessary, in particular, for the solutions including the polar molecules. The most elegant way is to explore spatial distribution function which enables us to have much better insights into the three-dimensional local solvation structure. The three-dimensional spatial distributions of O and H atoms of water for the first hydration shell are depicted in Fig. 3, with the red solid surfaces representing the density distribution of O atoms at 2.3 times of the average level and the green wire-frames representing the density distribution of H atoms. Since the distances between F and H (or O) are much longer (see Fig. 2(b)), here only the HB interactions of Cl···H and XB interactions of Cl···O can be found in Fig. 3. It is more clearly in Fig. 3(b) that the distribution of some O atoms around Cl atom in the Cl-F bond direction are attributed to the XB interactions of Cl···O, while the other O atoms are due to the HB interactions of Cl···H. The combination effect of the spatial distributions of these HB and XB interactions is to form the peak at 3.0 Å of g(r) in Fig. 2(a). Such an anisotropic hydration structure is a result both of HB and XB interactions between ClF and H2O molecules. To the best of our knowledge, there are no reports prior to this work on this unique hydration structure of ClF.IV. Conclusion
On the basis of the force field parameters for ClF optimized from $ab$ $initio$ calculations, we perform all-atom molecular dynamics simulations of the aqueous solution of ClF. The hydration structures are derived from the simulation data. The coexistence of hydrogen bonding and halogen bonding interactions is proved, and these two types of intermolecular interactions have the comparable contributions to the first hydration shell around ClF. This work should be helpful toward understanding of the halogen-bond related molecular assembling and crystallization.V. Acknowledgments
This work is supported by the National Natural Science Foundation of China (No.20673105).
|||Colin and H. Gaultier de Claubry M., Ann. Chim. 90 , 87 (1814).|
|||M. Colin, Ann. Chim. 91 , 252 (1814).|
|||Metrangolo and G. Resnati P., Halogen bonding:fundamentals and applications. Berlin: Springer (2008).|
|||G. R. Desiraju, P. S. Ho, L. Kloo, A. C. Legon, R. Marquardt, P. Metrangolo, P. Politzer, G. Resnati, and K. Rissanen, Pure Appl. Chem. 85 , 1711 (2013).|
|||H. B. Li, Y. B. Bai, S. X. Tian, and J. L. Yang, J. Theor. Comp. Chem. 8 , 615 (2009). DOI:10.1142/S0219633609004964|
|||K. O. Christe, Inorg. Chem. 11 , 1220 (1972). DOI:10.1021/ic50112a012|
|||W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey, and M. L. Klein, J. Chem. Phys. 79 , 926 (1983). DOI:10.1063/1.445869|
|||F. Fu and S. X. Tian C., J. Chem. Phys. 132 , 174507 (2010). DOI:10.1063/1.3426029|
|||M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, A. Montgomery Jr. J., J. E. Peralta, F. Ogliaro, Bearpark M.J., J. Heyd, Brothers E.N., Kudin K.N., Staroverov V.N., R. Kobayashi, J. Normand, K. Raghavachari, Rendell A.P., J. C. Burant, Iyengar S.S., J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Farkas Ö., J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian 09 (2009).|
|||Van Der Spoel D., E. Lindahl, B. Hess, G. Groenhof, A. E. Mark, and J. C. Berendsen H., J. Comput. Chem. 26 , 1701 (2005). DOI:10.1002/(ISSN)1096-987X|
|||B. Hess, C. Kutzner, Van der Spoel D., and E. Lindahl, J. Chem. Theor. Comput. 4 , 435 (2008). DOI:10.1021/ct700301q|
|||J. C. Berendsen H., P. M. Postma J., DiNola A., and J. R. Haak, J. Chem. Phys. 81 , 3684 (1984). DOI:10.1063/1.448118|
|||Nosé S., Mol. Phys. 52 , 255 (1984). DOI:10.1080/00268978400101201|
|||W. G. Hoover, Phys. Rev A31 , 1695 (1985).|
|||Parrinello and A. Rahman M., J. Appl. Phys. 52 , 7182 (1981). DOI:10.1063/1.328693|
|||U. Essmann, L. Perera, M. L. Berkowitz, T. Darden, H. Lee, and L. G. Pedersen, J. Chem. Phys. 103 , 8577 (1995). DOI:10.1063/1.470117|
|||B. Hess, H. Bekker, J. C. Berendsen H., and G. E. M. Fraaije J., J. Comput. Chem. 18 , 1463 (1997). DOI:10.1002/(ISSN)1096-987X|
|||S. A. Cooke, Cotti C. M. Evans G., J. H. Holloway, Z. Kisiel, A. C. Legon, and M. A. Thumwood J., Chem. Eur. J. 7 , 2295 (2001). DOI:10.1002/(ISSN)1521-3765|
|||E. Parisini, P. Metrangolo, T. Pilati, G. Resnati, and G. Terraneo, Chem. Soc. Rev. 40 , 2267 (2011). DOI:10.1039/c0cs00177e|