applied surface science Applied Surface Science 121/122 (1997) 167-170
Quantum chemical calculation on clay-water A. Chatterjee
T. Iwasaki, T. Ebina, H. Hayashi
Inorganic Material Section, Tohoku National Industrial Research Institute, AIST, 4-2-l Nigatake, Miyagino-ku,
Sendai 983, Japan
Received 1 November 1996; accepted 2 1 February 1997
Abstract We use a molecular description of the solvent and clay sheet to model the clay-water interactions. Quantum chemical calculations both semiempirical (MNDO - modified neglect of differential overlap) and first principle (DFT - density functional theory) were performed on localized cluster models of montmorillonite to study the interaction of one water molecule near the vicinity of the clay surface. The minimized orientation of water molecules with respect to the clay surface is determined. It is observed that lower energy is obtained when the two hydrogens of the water molecule point towards the clay surface i.e. the oxygen of water molecules is going away from the clay surface. The effect of tetrahedral substitution in the clay matrix on the water interaction has also been studied to show that tetrahedral substitution in clays has a distinct effect on the clay-water interaction with respect to stabilization energy. 0 1997 Elsevier Science B.V. Keywords:
Montmorillonite; Water; Tetrahedral substitution; MNDO; DFT calculations
1. Introduction Smectite clays are layer aluminosilicates showing a large variety of physicochemical properties. Dioctahedral smectite is a member of the smectite family which shares the common feature that two tetrahedral sheets sandwitch a sheet of octahedrally coordinated metal ions. Substitution of a trivalent metal ion (Al”+ > for the tetravalent Si results in a net negative charge and the interaction with positive ions (the exchangeable cations) to form an interlayer hydrated phase. There are many experimental investigations on the clay-substrate interactions using different techniques as thermal analysis [I], calorimetry , mechanical measurements , neutron diffraction , XRD [S], IR , UV , EPR  and NMR 
_ Corresponding author. Tel.: +81-22-2375211; fax: + 81-22. 2366839; e-mail: [email protected]
spectroscopic methods. A justified modeling of all the observed data requires molecular description of the interacting species. Computer simulation can contribute significantly in achieving this goal. A number of Monte Carlo simulations have been performed on the solvation of clay materials [lO,l l]. This is the first quantum chemical calculation to the clay-water system to study the interaction of clay framework and water molecule. In this communication we perform both semiempirical (MNDO) and first principle (DFT) quantum chemical calculations on localized cluster models of dioctahedral smectite to study the interaction of one water molecule near the vicinity of the model of montmorillonite. The minimized orientation of water molecules with respect to the clay surface is determined. The effect of the tetrahedral substitution in the clay matrix on the water interaction has also been studied.
0169-4332/97/$17.00 0 1997 Elsevier Science B.V. All rights reserved. PZZ SO169-4332(97)00280-S
A. Chatterjee et al./Applied Surface Scienee 121 / 122 (1997) 167 170
2. M e t h o d and m o d e l
The ideal formula of montmorillonite, one of the species of dioctahedral smectites used in this study, is (Na +, nH20) (A14_xMg ~) SisO2o(OH) 4. The montmorillonite structure has been generated from the crystal structure of well defined dioctahedral pyrophyllites having the formula SisA14020(OH) 4 which has been discussed in detail elsewhere . To calculate the clay-water energies, a localized cluster model of the montmorillonite framework is used. Semiempirical (MNDO) calculations were performed on the cluster having the formula AI 28i6024 H ~8 and DFT calculations were performed on the cluster having the formula A1Si4Ol6H]0 . These clusters are schematically shown in Fig. 1 without the additional hydrogen used for capping the dangling bonds for visual clarity. To reproduce the symmetry of the cavity of silicates, the clay fragment contains a hexagonal cavity of silicates plus two
inner octahedral alumina. We have also considered the tetrahedral substitution of one Si TM by one A1[[I, introducing a negative charge in the dioctahedral smectite framework, whose mineral name is beidellite. All dangling bonds were saturated by hydrogen atoms. The calculation includes only one water molecule in the vicinity of a model of the dioctahedral smectite. The MNDO calculations have been performed using MOPAC of Biosym version 3.00. DFT calculations were performed using the DMO1 package of Biosym version 3.00. The geometry optimization calculations were carried out using a minimal numerical basis set . The total energy for the final optimized geometry was then evaluated using a double numerical polarization basis set . A JMW local type functional  was used for the exchange-correlation energy terms in the total energy expression. The static visualization of molecular configurations and the optimized molecular geometries
01 Si, )24 ?
Fig. 1. The cluster models chosen for quantum chemical calculations: ( l ) AI 2 Si 6024 H 18 and (2) AlSi4Oi6 H j j.
A. Chanerjee et al. /Applied Surface Science 121 / 122 (1997) 167 170
were made with the InsightII code of Biosym on a Silicon graphics INDIGO 2 workstation.
3. Results and discussion
The orientation of the hydrogen in the center of the cavity is obtained by a minimization of the energy of the clay fragment. The OH bond is directed in such a way that the hydrogen pointing away from the plane of the octahedral A1 (as shown in Fig. 1). Table 1 shows the MNDO results of the Mulliken population analysis for the cluster AlzSi6Oe4H~s, the DFT results of the Mulliken population analysis have also been incorporated in Table 1 for the cluster AISi4OI6H io. The Mulliken population analysis results show that the positive charge density is low on octahedral A1 in comparison to tetrahedral Si. It is also observed that the hydroxyl
Table 1 Mulliken population analysis results for the clusters A12 Si6024 His and AISi4OI6HllI MNDO A12Si6024HI8
All Ol 02 HI H2 Sil 03 H3 04 Si2 05 H4 06 07 08 H5 09 H6 Si3 Si4 O10 H7 O11 OI2 O13
1.10 -0.93 0.92 0.53 0.52 1.65 -0.93 0.55 -0.93 1.66 -0.93 0.54 -0.92 -0.92 -0.92 0.53 0.93 0.55 1.67 1.66 -0.92 0.55 -0.91 -0.93 -0.93
H8 OI4 H9 H10 O15 Hll AI2 Si5 Si6 O16 HI2 O17 O18 O19 020 O21 022 HI3 HI4 023 Hl5 024 HI6 HI7 H18
0.54 0.91 0.56 0.54 0.92 0.55 1.10 1.66 1.67 -0.92 0.53 -0.74 -0.72 -0.92 -0.92 -0.92 -0.92 0.43 0.42 -0.92 0.54 -0.92 0.56 0.54 0.53
H1 H2 All Sil Si2 Si3 Si4 O1 02 03 H3 04 H4 05 H5 06 H6 07 H7 08 H8 09 H9 O10 HI0
0.54 0.53 0.95 1.77 1.77 1.78 1.79 -0.92 0.91 0.91 0.64 -0.91 0.66 0.93 0.67 -0.92 0.68 -0.91 0.64 -0.93 0.68 0.91 0.62 -0.95 0.60
Oll O12 O13 O14 O15 O16 H11
-0.64 -0.64 -0.93 -0.93 -0.93 -0.93 0.62
Distance (Angstrom) Fig. 2. Dioctahedral smectite-water interaction, with MNDO (A1 and B1) and DFT (A2 and B2) without tetrahedral substitution. The energy is attractive (B1 and B2) when protons are directed towards clay and repulsive (AI and A2) when oxygen is directed towards clay. The distance of the water molecule (protons or oxygen) from the Si plane is measured in A.
hydrogen has less positive charge ((0.42, 0.43) in MNDO and (0.53, 0.54) in DFT) in comparison to the hydrogens which have been used as capping of the dangling bonds (0.56 in MNDO and 0.61 in DFT) this is due to the difference in O - H distances. The same trend in charge is observed for the oxygen of the bridging hydroxyls. We have calculated the clay-water energy for different configurations of the water molecule, approaching different sites of the clay surface. Initially, to locate the active site on the surface we kept both the geometry of water molecule and clay fragment fixed and then we optimized the orientation of the water molecule with fixed clay configuration. The water O - H bond length is assumed to be 0.96 A with a H - O - H angle of 104.52 °. It is observed that among all the orientations of water the lowest energy is obtained when the two hydrogens of the water molecule point towards the clay surface, while the maximum energy is obtained when the oxygen of the water molecule is pointed towards the clay surface due to repulsion. Fig. 2 shows the energy pattern calculated using MNDO and DFT for the above mentioned cases of water interaction. A1 (MNDO) and A2 (DFT) show the repulsive trend with oxygen of water pointing to the framework, while B1 (MNDO) and B2 (DFT) shows
A. Chatterjee et al. /Applied Surface Science 121 / 122 (1997) 167-170
B -20 "
- 4 0 -~
i -50 i
stituted and unsubstituted case the larger stabilization is reached for a clay water distance of 4 A. Close to the surface, the energy exhibits some large variations. At large distance, the minimum occurs at the site of substitution. As the water molecule approaches towards the surface, the minimum shifts gradually to the center of cavity above the site of substitution. So it can be predicted that the site of substitution behaves like a saddle point for the translation from one cavity to another. At short distance, the oxygen atoms become strongly repulsive in nature.
-60 i Distance (Angstrom)
Fig. 3. Dioetahedral smectite-water interaction, with DFT without (A) and with (B) tetrahedral substitution. The distance of the water molecule (protons or oxygen) from the Si plane is measured in A.
the attractive trend with hydrogen of water pointing to the framework. The trends for both the methods match with a little variation due to the approximation involved in MNDO compared to DFF. Now the clay-water interaction has also been studied with and without tetrahedral substitution. The configuration of water with two hydrogens pointing to the clay surface, i.e. the stable configuration was chosen to study the effect of tetrahedral substitution on the water-clay interaction. The water molecule remains parallel to the plane of the clay surface. The distance between the water molecule and the clay framework for both the substituted and the unsubstituted case has been repeated as before. The energies calculated using DFT, in terms of distances (given in A), between the atom of oxygen of the water molecule and the Si plane of the dioctahedral smectite are shown in Fig. 3. Fig. 3 shows that the stabilization energy for the substituted configuration is more than the unsubstituted one. This is due to the fact that when one Si has been replaced by one A1 the system acquires a net negative charge which favors the interaction with hydrogen of water to a greater extent. Tetrahedral substitution greatly modifies the position of the energy levels, reducing the gap between HOMO and LUMO. Far from the clay surface the energy is nearly constant and does not vary with tetrahedral substitution. For both the sub-
4. Conclusion This is the first attempt at a molecular description of the solvation of the clay surface. It is observed that among all the orientations of water the lowest energy is obtained when the two hydrogens of the water molecule points towards the clay surface. Tetrahedral substitution modifies the position of the energy levels in case of dioctahedral smectite. This model will be useful in providing a better understanding of the physicochemical processes involved in surface diffusion, chemisorption, etc.
References  A. Mackenjie, in: Advanced Techniques for Clay Mineral Analysis, Elsevier, Amsterdam, 1982, ch. 2.  B. Mikhail, N.M. Guindy, S. Hanafi, J. Colloid Interf. Sci. 70 (1979) 282.  D. Collaghan, R.H. Ottewill, Farad. Disc. Chem. Soc. 57 (1974) 110.  F. Cebula, R.K. Thomas, J.W. White, J. Chem. Soc. Farad. Trans. I 76 (1980) 314.  G. Pons, F. Rousseaux, D. Tchoubar, Clay Miner. 16 (1981) 23.  H. Suquet, R. Prost, H. Pezerat, Clay Miner. 12 (1977) 113.  I. Villemure, C. Detellier, A.G. Szabo, J. Am. Chem. Soc. 108 (1986) 4658.  K. Boyd, M.M. Mortland, Nature 316 (1985) 532.  L. Grandjean, P. Laszlo, Clays Clay Miner. 37 (1989) 403.  M. Delville, S. Sokolowski, J. Phys. Chem. 97 (1993) 6261.  N. Delville, J. Phys. Chem. 99 (1995) 2033.  A. Chatterjee, T. Iwasaki, T. Ebina, H. Hayashi, K. Torii, J. Mol. Catal., submitted.  P. Delly, J. Chem. Phys. 94 (1991) 7245.  Q. Hedin, B.L Lundqvist, J. Phys. C 4 (197l) 2064.