- Email: [email protected]
Computational studies of hydration in Na+-mordenite zeolites
1606
Studies in Surface Science and Catalysis, volume 154 E. van Steen, L.H. Callanan and M. Claeys (Editors) © 2004 Elsevier B.V. All rights reserved.
COMPUTATIONAL STUDIES OF HYDRATION IN Na^-MORDENITE ZEOLITES Maurin, G . \ Bell, R . G . \ Devautour, S.^, Henn, F.^ and Giuntini, J.C.^ ^The Davy Faraday Research Laboratory, Royal Institution of Great Britain, 21 Albemarle Street, London WIS 4BS, United Kingdom. ^Laboratoire de Physicochimie de la Matiere Condensee, UMR 5617 CNRS, University of Montpellier II, PI. E. Bataillon, 34095 Montpellier cedex 05, France.
ABSTRACT Computer modeling has been used to explore the effect of water on both the distribution of the extraframework cations and on their interactions with the framework in Na^-Mordenites. Atomistic simulations based on interatomic potentials and minimisation techniques have been used to determine the location of both cations and water molecules as a function of the hydration level. Our calculations showed two different cation behaviours depending on the type of channels that they occupy, the positions of the cations in the main channels being substantially perturbed upon the sorption of water molecules whereas those of the cations located in the small channels being only slightly shifted. Furthermore, Molecular Dynamics simulations allowed us to compare the mobility of the cations in the hydrated and dehydrated states and confirmed the results expected from the previous static calculations. This modelling has been successfully compared with experimental data obtained by dielectric relaxation spectroscopy.
INTRODUCTION A knowledge of the partition of the extra-framework cations among the different sites available in zeolites is a crucial point because it influences the distribution of the electron density in the framework [1] and hence the interactions with the adsorbed molecules, leading to changes in the adsorption and catalytic properties of the zeolites [2]. Here we have focused our attention on a typical zeolite Na^-Mordenite characterised by the ideal composition of the unit cell Na8Al8Si4o096, nH20 with n ranging from 0 (dehydrated state) to 24 (totally hydrated state) [3]. Its structure is orthorombic (Cmcm) with unit cell parameters a= 18.1 A, b= 20.5 A, and c= 7.5 A [3]. The framework has a porous structure which consists of main channels having a slightly elliptical cross section with 12 TO4 tetrahedron units (T=Si, Al) and connected with small channels constituted by 8 TO4 cross section (figure 1 a). The first part of this work was thus to simulate the equilibrium location of the extra-framework cations both in the totally dehydrated and hydrated states by combining pair potential model and energy minimisation techniques. Their positions have been previously studied by X-ray diffraction [4,5] and by our own investigations using dielectric relaxation spectroscopy (DRS) [6-8]. It has been shown that the monovalent cations are mainly located in three different sites I, IV and VI as it is reported in figure lb. From a practical point of view, this study was performed because water plays a key role in many applications involving adsorption and, more particularly, in ion-exchange which is carried out in aqueous solution [9]. In this latter case, it is well known that water improves the efficiency of this process by co-ordinating the cations and hence increasing their mobility [9-10]. There have been some atomistic simulations to study the location of extra-framework cations both in dehydrated [11-13] and hydrated [14-17] zeolites using, for example. Molecular Dynamics and packing and energy minimization procedures. But as far as we know this is the first attempt which proposes to model the whole hydration process and to provide a microscopic description of the effect of water on the exchangeable cations' behaviour in accordance with our experimental data which are reported elsewhere in these proceedings. The second part of this work was to study the mobility of the extra-framework cations as a function of the water content by means of Molecular Dynamics simulations in order to explore the evolution of the nature of their interactions with the zeolite framework. This modelling is motivated by experimental investigations using DRS which showed a weakening of the interactions between the extra-framework cations and the framework with increasing water content [8].
1607 The originality of this investigation was then to compare, at each step of the simulation, the calculated results to our own experimental data obtained by DRS [8].
Figure la. Representation of the Mordenite framework.
M\I
Figure lb. Representation of the unit cell of Mordenite and distribution of the extra-framework cations among the three distinct crystallographic sites named I, IV and VI. COMPUTATIONAL METHODOLOGY Interatomic potential The first, and very crucial, step of this work, is to provide a description of the potential energy surface by selecting an appropriate pair potential model able to reproduce well the interactions within the whole system. For this purpose, we used the potential reported both by Sanders et al. [18] and Jackson and Catlow [19], to describe the interactions within the framework, and between the framework and the exchangeable Na^ cations. In this model the framework is treated as flexible by introducing three body terms for 0-Al-O and O-Si-O and polarisability is taken into account by using a core-shell model for the oxygen atoms. Among the various potentials available to describe the water interactions [20], we have selected those developed by De Leeuw et al [21] which successfully model hydration on zeolite A. This complex potential which includes a core-shell model on the water oxygen atoms, consists of a Morse potential O-H and three body term H-O-H to describe the intramolecular interactions, and Lennard-Jones and Buckingham potentials between 0 - 0 and O-H respectively to define the intermolecular interactions. Finally, the pair potentials between water and
1608 framework-exchangeable cations were taken from the work of Lewis et a.l [15] in which the interactions between the oxygen of the water and both the exchangeable cations and the oxygen of the framework are described by Buckingham terms. The parameters for each of these potentials are fully detailed in our previous paper [22]. Energy minimisation procedure Once the potential energy surface was defined, we selected a typical aluminium configuration generated in our previous works [23,24] related to the description of the (Si,Al) distribution within the framework. The first step was to simulate the positions of the extra-framework cations in the dehydrated mordenite. The calculation was performed by using periodic boundary conditions and the unit cell represented in figure lb in space group PI to avoid any symmetry constraint. The Ewald summation [25] was used to calculate the coulombic interactions and the cut-off was set at 16 A for the short range contributions. We then used the various energy minimisers available in GULP (General Utility Lattice Program) [26]. From this initial configuration, we then built the structure of the fiilly hydrated Na^-mordenite by adding the water molecules step by step. Employing this systematic and computationally-intense approach allowed us to obtain the lowest energy configuration for each hydration level ranging from 1 to 24 H2O/U.C. by using a stepwise minimisation procedure fully described elsewhere [22]. Molecular dynamics simulation The complex potential previously described was implemented in the DLPOLY program [27]. We selected the optimised structures obtained by the minimisation procedure as starting configurations and the minimised cell dimensions were kept fixed during the Molecular Dynamics (MD) runs. The actual cell used for MD was doubled in the c direction, giving a repeat distance in that direction of 15.0 A, and thereby avoiding unrealistic interactions between images of the same particle. The MD simulations were run in a NVT ensemble using the isokinetic Evans thermostat [25,27] for 200 picoseconds including 50 picoseconds equilibration period, during which the Na^ cations were maintained fixed in their initial positions. It is noteworthy that the framework was treated as being flexible in these simulations. Furthermore, the shell model, which is included on both framework and water oxygen, was treated by the adiabatic shell model approach which consists of assigning a small mass to the shell (0.2 u.m) [25]. Due to this latter point, and in order to maintain a stable system, we used a very small time step of 0.2 femtosecond. Positions were then stored every 0.1 ps which allowed us to evaluate the Mean Square Displacements (MSD) of the extra-framework cations in both the dehydrated and totally hydrated mordenites by means of the following classical equation [25]: MSD(t) = ( A r / ( t ) ) = l | : A r ^ 2 ( t ) = l | ; ( r j ( t ) - r j ( 0 ) ) 2 where N corresponds to the number of extra-framework ions considered in the estimation of the MSD and the position of the origin (rj(0)) is allowed to change in order to improve the statistics of the calculation.
RESULTS AND DISCUSSIONS From the selected (Si,Al) configuration, we have shown that the extra-framework cations are located half in the main channels (sites IV and VI) and half in the small channels (site I) as is reported in figure lb. This result is in good agreement with X-ray diffraction [4,5] and DRS data [6,7]. We then obtained the lowest energy structures corresponding to hydration levels ranging from 1 to 24 water molecules per unit cell. The structure of the fiilly hydrated Na^-mordenite is given in figure 2. From this procedure, we observed that the two first water molecules do not interact strongly with the extra-framework cations. Next, the cations of the main channels are progressively extracted from their initial sites at 3, 6, 7 and 11 H2O/U.C. By contrast, the cations in the small channels are only slightly perturbed by water occupying neighbouring side pockets and remain trapped in the same initial positions whatever the hydration level. These two different types of cation behaviour are well summarised by figure 3 where, for the sake of clarity, only the cation positions in the fully hydrated and dehydrated Na^-mordenites are shown in an extended section of mordenite structure. These results are in good qualitative agreement with Dielectric relaxation spectroscopy data [8] which predicts that the population of site I remains constant whatever the H2O content, and that the occupations of sites IV and VI dramatically decrease in the range 3,2 - 8,7 H2O per unit cell.
1609
Figure 2. Representation of the structure of the fully hydrated Na^-Mordenite.
Figure 3. Representation of the extra-framework cation positions in the dehydrated and in the fiilly hydrated Na^-mordenite. For the sake of clarity, the unit cell is extended in 2 dimensions and the water molecules are not shown. Analysis of the MSD plots calculated from the molecular dynamics run at 300 K, show that both cations in small and main channels are essentially immobile at this temperature in the dehydrated zeolite, as we can see in figures 4 and 5 respectively. Then, we investigated the MSD for the two types of cations in the hydrated state. We can observe that the cations located in the small channels are not displaced, presenting only small vibrational amplitudes around their mean positions (figure 4). By contrast, the cations in the main channels are characterised by motions on a large scale (figure 5) corresponding to displacements throughout the whole cavity.
1610
0
20
40 60 Time (picoseconds)
80
100
Figure 4. MSD plots for the sodium cations located in the small channels at 300 K in dehydrated and hydrated Na^-mordenite.
Ilvclralcd sV:\'
Dehydrated state ^0
40 ^ 60 ' Time (picoseconds)
80"
100
Figure 5. MSD plots for the sodium cations located in the main channels at 300 K in dehydrated and hydrated Na^-mordenite. We have further tried to correlate these results with the detrapping energy of the cations from their initial sites which have been measured experimentally by DRS [8]. In the case of the cations located in the main channels, we observed a well-pronounced decrease of this energy as the number of water molecules increases. This result can be interpreted microscopically by an increase of mobility of these extra-framework cations, the water interacting directly, pulling them away from the framework and thus leading to weaker cation-framework interactions. On the other hand, we have also observed a decrease of the detrapping energy for the cations in the small channels but less pronounced than for the main channel cations. This decrease cannot be explained by an enhancement of the mobility of the cations induced by direct interactions with water molecules but rather by cation-water interactions over longer distances, induced by water located in neighbouring side pockets. These interactions are much weaker than those found for the cations in the main channels, leading to a smaller decrease in the detrapping energy.
1611
CONCLUSION From the results of our simulations, we provide a microscopic description of the effect of hydration in Na^-mordenites, which plays a key role in ion exchange processes in zeolites. We identified two different types of cation behaviour, in terms of mobility as a ftinction of hydration level, depending on their initial positions in the main or small channels. The originality of this study was to compare these simulated results with experimental data. This successful comparison allowed us to obtain a good qualitative agreement between experiment and theory.
ACKNOWLEDGMENTS This research was supported through a European Communities Marie Curie Individual Fellowship: G.M. with EC Contract Number HPMF-CT-2001-01379. We thank J. Gale for providing us GULP, and N. De Leeuw for fruitful and helpful discussions. R.B. acknowledges the Leverhulme Trust for funding. REFERENCES 1. Bosch, E., Huber, S., Weitkamp, J., Knozinger, H., Phys. Chem. Chem. Phys., 1 (1999), 579. 2. Vitale, G., Bull, L.M., Morris, R.E., Cheetham, A.K., Toby, B.H., Coe, C.G., Mac Dougall, J.E., J. Phys. Chem. B, 99 (1995), 16087. 3. Meier, W.M., Z. Kristallogr., 115 (1961),439. 4. Mortier, W.J., Compilation of Extraframework Sites in Zeolites, Butterworth, Guildford, (1982). 5. Coughlan, B., Carrol, W.M., Mc Cann, A., J. Chem. Soc. Faraday Trans., 73 (1977), 1612. 6. Devautour, S., Vanderschueren, J., Giuntini, J.C, Henn, F., Zanchetta, J.V., Ginoux, J.L., J. Phys. Chem. B, 102(1998), 3749. 7. Pamba, M., Maurin, G., Devautour, S., Vanderschueren, J., Giuntini, J.C, Di Renzo, F., Hamidi, F., Phys. Chem. Chem. Phys., 113, 11 (2000), 4498. 8. Devautour, S., Abdoulaye, A., Giuntini, J.C, Henn, F., J. Phys. Chem. B, 105 (2001), 9297. 9. Pissis, P., Daoukaki-Diamanti, D., J. Phys. Chem. Solids, 54 (1993), 701. 10. Artioli, G., Smith, J.V., Kvick, A., Pluth, J.J., Stahl, K., Zeolites, Ed. Dram., B., Hocevar., S., Paovnik, S., Elsevier, Amsterdam, (1985) 249. 11. Vitale, G., Mellot, CF., Bull, L.M., Cheetham, A.K., J. Phys. Chem. B, 101 (1997), 4559. 12. Lignieres, J., Newsam, J.M., Microporous and Mesoporous Materials, 28 (1999), 305. 13. Grillo, M.E., Carrazza, J., J. Phys. Chem. B, 100 (1996), 12261. 14. Faux, D., J. Phys. Chem. B, 103 (1999), 7803. 15. Dewis, D.W., Ruiz Salvador, A.R., Almora Barios, N., Gomez, A., Mistry, M., Molecular Simulation, 28, 6-7 (2002), 649. 16. Almora-Barios, N., Gomez, A., Ruiz Salvador, A.R., Mistry, M., Lewis, D.W., Chem. Commun., (2001), 531. 17. Faux, D., Smith, W., Forester, T.R., J. Phys. Chem. B, 101 (1997), 1762. 18. Sanders, M.J., Leslie, M, Catlow, C.R.A., J. Chem. Soc. Chem. Commun. (1984), 1271. 19. Jackson, R.A., Catlow, C.R.A., Molecular Simulation, 1, 207, 1988. 20. Lau, K.F., Alper, H.E., Thatcher, S., Stouch, T.R., J. Phys. Chem. B., 98 (1994), 8785. 21. Manon Higgins, F., de Leeuw, N.H., Parker, S.C, J. Mater. Chem., 12 (2002), 124. 22. Maurin, G., Bell, R., Devautour, S., Henn, F., Giuntini, J.C, J. Phys. Chem. B, in press. 23. Maurin, G., Senet, P., Devautour, S., Henn, F., Giuntini, J.C, Van Doren, V.E., Computational Materials Science 22 (2001), 106. 24. Maurin, G., Senet, P., Devautour, S., Gaveau, P., Henn, F., Van Doren, V., Giuntini, J.C, J. Phys. Chem. B, 105 (2001), 9157. 25. Allen, M.F., Tildesley, D., Computer Simulation of Liquids, Oxford, Science Publication, Oxford (1987). 26. Gale, J.D., J. Chem. Soc. Faraday. Trans., 93 (1997), 629. 27. Smith, W., Forester, T., J. Mol. Graphics, 14 (1996), 136.
Studies in Surface Science and Catalysis, volume 154 E. van Steen, L.H. Callanan and M. Claeys (Editors) © 2004 Elsevier B.V. All rights reserved.
COMPUTATIONAL STUDIES OF HYDRATION IN Na^-MORDENITE ZEOLITES Maurin, G . \ Bell, R . G . \ Devautour, S.^, Henn, F.^ and Giuntini, J.C.^ ^The Davy Faraday Research Laboratory, Royal Institution of Great Britain, 21 Albemarle Street, London WIS 4BS, United Kingdom. ^Laboratoire de Physicochimie de la Matiere Condensee, UMR 5617 CNRS, University of Montpellier II, PI. E. Bataillon, 34095 Montpellier cedex 05, France.
ABSTRACT Computer modeling has been used to explore the effect of water on both the distribution of the extraframework cations and on their interactions with the framework in Na^-Mordenites. Atomistic simulations based on interatomic potentials and minimisation techniques have been used to determine the location of both cations and water molecules as a function of the hydration level. Our calculations showed two different cation behaviours depending on the type of channels that they occupy, the positions of the cations in the main channels being substantially perturbed upon the sorption of water molecules whereas those of the cations located in the small channels being only slightly shifted. Furthermore, Molecular Dynamics simulations allowed us to compare the mobility of the cations in the hydrated and dehydrated states and confirmed the results expected from the previous static calculations. This modelling has been successfully compared with experimental data obtained by dielectric relaxation spectroscopy.
INTRODUCTION A knowledge of the partition of the extra-framework cations among the different sites available in zeolites is a crucial point because it influences the distribution of the electron density in the framework [1] and hence the interactions with the adsorbed molecules, leading to changes in the adsorption and catalytic properties of the zeolites [2]. Here we have focused our attention on a typical zeolite Na^-Mordenite characterised by the ideal composition of the unit cell Na8Al8Si4o096, nH20 with n ranging from 0 (dehydrated state) to 24 (totally hydrated state) [3]. Its structure is orthorombic (Cmcm) with unit cell parameters a= 18.1 A, b= 20.5 A, and c= 7.5 A [3]. The framework has a porous structure which consists of main channels having a slightly elliptical cross section with 12 TO4 tetrahedron units (T=Si, Al) and connected with small channels constituted by 8 TO4 cross section (figure 1 a). The first part of this work was thus to simulate the equilibrium location of the extra-framework cations both in the totally dehydrated and hydrated states by combining pair potential model and energy minimisation techniques. Their positions have been previously studied by X-ray diffraction [4,5] and by our own investigations using dielectric relaxation spectroscopy (DRS) [6-8]. It has been shown that the monovalent cations are mainly located in three different sites I, IV and VI as it is reported in figure lb. From a practical point of view, this study was performed because water plays a key role in many applications involving adsorption and, more particularly, in ion-exchange which is carried out in aqueous solution [9]. In this latter case, it is well known that water improves the efficiency of this process by co-ordinating the cations and hence increasing their mobility [9-10]. There have been some atomistic simulations to study the location of extra-framework cations both in dehydrated [11-13] and hydrated [14-17] zeolites using, for example. Molecular Dynamics and packing and energy minimization procedures. But as far as we know this is the first attempt which proposes to model the whole hydration process and to provide a microscopic description of the effect of water on the exchangeable cations' behaviour in accordance with our experimental data which are reported elsewhere in these proceedings. The second part of this work was to study the mobility of the extra-framework cations as a function of the water content by means of Molecular Dynamics simulations in order to explore the evolution of the nature of their interactions with the zeolite framework. This modelling is motivated by experimental investigations using DRS which showed a weakening of the interactions between the extra-framework cations and the framework with increasing water content [8].
1607 The originality of this investigation was then to compare, at each step of the simulation, the calculated results to our own experimental data obtained by DRS [8].
Figure la. Representation of the Mordenite framework.
M\I
Figure lb. Representation of the unit cell of Mordenite and distribution of the extra-framework cations among the three distinct crystallographic sites named I, IV and VI. COMPUTATIONAL METHODOLOGY Interatomic potential The first, and very crucial, step of this work, is to provide a description of the potential energy surface by selecting an appropriate pair potential model able to reproduce well the interactions within the whole system. For this purpose, we used the potential reported both by Sanders et al. [18] and Jackson and Catlow [19], to describe the interactions within the framework, and between the framework and the exchangeable Na^ cations. In this model the framework is treated as flexible by introducing three body terms for 0-Al-O and O-Si-O and polarisability is taken into account by using a core-shell model for the oxygen atoms. Among the various potentials available to describe the water interactions [20], we have selected those developed by De Leeuw et al [21] which successfully model hydration on zeolite A. This complex potential which includes a core-shell model on the water oxygen atoms, consists of a Morse potential O-H and three body term H-O-H to describe the intramolecular interactions, and Lennard-Jones and Buckingham potentials between 0 - 0 and O-H respectively to define the intermolecular interactions. Finally, the pair potentials between water and
1608 framework-exchangeable cations were taken from the work of Lewis et a.l [15] in which the interactions between the oxygen of the water and both the exchangeable cations and the oxygen of the framework are described by Buckingham terms. The parameters for each of these potentials are fully detailed in our previous paper [22]. Energy minimisation procedure Once the potential energy surface was defined, we selected a typical aluminium configuration generated in our previous works [23,24] related to the description of the (Si,Al) distribution within the framework. The first step was to simulate the positions of the extra-framework cations in the dehydrated mordenite. The calculation was performed by using periodic boundary conditions and the unit cell represented in figure lb in space group PI to avoid any symmetry constraint. The Ewald summation [25] was used to calculate the coulombic interactions and the cut-off was set at 16 A for the short range contributions. We then used the various energy minimisers available in GULP (General Utility Lattice Program) [26]. From this initial configuration, we then built the structure of the fiilly hydrated Na^-mordenite by adding the water molecules step by step. Employing this systematic and computationally-intense approach allowed us to obtain the lowest energy configuration for each hydration level ranging from 1 to 24 H2O/U.C. by using a stepwise minimisation procedure fully described elsewhere [22]. Molecular dynamics simulation The complex potential previously described was implemented in the DLPOLY program [27]. We selected the optimised structures obtained by the minimisation procedure as starting configurations and the minimised cell dimensions were kept fixed during the Molecular Dynamics (MD) runs. The actual cell used for MD was doubled in the c direction, giving a repeat distance in that direction of 15.0 A, and thereby avoiding unrealistic interactions between images of the same particle. The MD simulations were run in a NVT ensemble using the isokinetic Evans thermostat [25,27] for 200 picoseconds including 50 picoseconds equilibration period, during which the Na^ cations were maintained fixed in their initial positions. It is noteworthy that the framework was treated as being flexible in these simulations. Furthermore, the shell model, which is included on both framework and water oxygen, was treated by the adiabatic shell model approach which consists of assigning a small mass to the shell (0.2 u.m) [25]. Due to this latter point, and in order to maintain a stable system, we used a very small time step of 0.2 femtosecond. Positions were then stored every 0.1 ps which allowed us to evaluate the Mean Square Displacements (MSD) of the extra-framework cations in both the dehydrated and totally hydrated mordenites by means of the following classical equation [25]: MSD(t) = ( A r / ( t ) ) = l | : A r ^ 2 ( t ) = l | ; ( r j ( t ) - r j ( 0 ) ) 2 where N corresponds to the number of extra-framework ions considered in the estimation of the MSD and the position of the origin (rj(0)) is allowed to change in order to improve the statistics of the calculation.
RESULTS AND DISCUSSIONS From the selected (Si,Al) configuration, we have shown that the extra-framework cations are located half in the main channels (sites IV and VI) and half in the small channels (site I) as is reported in figure lb. This result is in good agreement with X-ray diffraction [4,5] and DRS data [6,7]. We then obtained the lowest energy structures corresponding to hydration levels ranging from 1 to 24 water molecules per unit cell. The structure of the fiilly hydrated Na^-mordenite is given in figure 2. From this procedure, we observed that the two first water molecules do not interact strongly with the extra-framework cations. Next, the cations of the main channels are progressively extracted from their initial sites at 3, 6, 7 and 11 H2O/U.C. By contrast, the cations in the small channels are only slightly perturbed by water occupying neighbouring side pockets and remain trapped in the same initial positions whatever the hydration level. These two different types of cation behaviour are well summarised by figure 3 where, for the sake of clarity, only the cation positions in the fully hydrated and dehydrated Na^-mordenites are shown in an extended section of mordenite structure. These results are in good qualitative agreement with Dielectric relaxation spectroscopy data [8] which predicts that the population of site I remains constant whatever the H2O content, and that the occupations of sites IV and VI dramatically decrease in the range 3,2 - 8,7 H2O per unit cell.
1609
Figure 2. Representation of the structure of the fully hydrated Na^-Mordenite.
Figure 3. Representation of the extra-framework cation positions in the dehydrated and in the fiilly hydrated Na^-mordenite. For the sake of clarity, the unit cell is extended in 2 dimensions and the water molecules are not shown. Analysis of the MSD plots calculated from the molecular dynamics run at 300 K, show that both cations in small and main channels are essentially immobile at this temperature in the dehydrated zeolite, as we can see in figures 4 and 5 respectively. Then, we investigated the MSD for the two types of cations in the hydrated state. We can observe that the cations located in the small channels are not displaced, presenting only small vibrational amplitudes around their mean positions (figure 4). By contrast, the cations in the main channels are characterised by motions on a large scale (figure 5) corresponding to displacements throughout the whole cavity.
1610
0
20
40 60 Time (picoseconds)
80
100
Figure 4. MSD plots for the sodium cations located in the small channels at 300 K in dehydrated and hydrated Na^-mordenite.
Ilvclralcd sV:\'
Dehydrated state ^0
40 ^ 60 ' Time (picoseconds)
80"
100
Figure 5. MSD plots for the sodium cations located in the main channels at 300 K in dehydrated and hydrated Na^-mordenite. We have further tried to correlate these results with the detrapping energy of the cations from their initial sites which have been measured experimentally by DRS [8]. In the case of the cations located in the main channels, we observed a well-pronounced decrease of this energy as the number of water molecules increases. This result can be interpreted microscopically by an increase of mobility of these extra-framework cations, the water interacting directly, pulling them away from the framework and thus leading to weaker cation-framework interactions. On the other hand, we have also observed a decrease of the detrapping energy for the cations in the small channels but less pronounced than for the main channel cations. This decrease cannot be explained by an enhancement of the mobility of the cations induced by direct interactions with water molecules but rather by cation-water interactions over longer distances, induced by water located in neighbouring side pockets. These interactions are much weaker than those found for the cations in the main channels, leading to a smaller decrease in the detrapping energy.
1611
CONCLUSION From the results of our simulations, we provide a microscopic description of the effect of hydration in Na^-mordenites, which plays a key role in ion exchange processes in zeolites. We identified two different types of cation behaviour, in terms of mobility as a ftinction of hydration level, depending on their initial positions in the main or small channels. The originality of this study was to compare these simulated results with experimental data. This successful comparison allowed us to obtain a good qualitative agreement between experiment and theory.
ACKNOWLEDGMENTS This research was supported through a European Communities Marie Curie Individual Fellowship: G.M. with EC Contract Number HPMF-CT-2001-01379. We thank J. Gale for providing us GULP, and N. De Leeuw for fruitful and helpful discussions. R.B. acknowledges the Leverhulme Trust for funding. REFERENCES 1. Bosch, E., Huber, S., Weitkamp, J., Knozinger, H., Phys. Chem. Chem. Phys., 1 (1999), 579. 2. Vitale, G., Bull, L.M., Morris, R.E., Cheetham, A.K., Toby, B.H., Coe, C.G., Mac Dougall, J.E., J. Phys. Chem. B, 99 (1995), 16087. 3. Meier, W.M., Z. Kristallogr., 115 (1961),439. 4. Mortier, W.J., Compilation of Extraframework Sites in Zeolites, Butterworth, Guildford, (1982). 5. Coughlan, B., Carrol, W.M., Mc Cann, A., J. Chem. Soc. Faraday Trans., 73 (1977), 1612. 6. Devautour, S., Vanderschueren, J., Giuntini, J.C, Henn, F., Zanchetta, J.V., Ginoux, J.L., J. Phys. Chem. B, 102(1998), 3749. 7. Pamba, M., Maurin, G., Devautour, S., Vanderschueren, J., Giuntini, J.C, Di Renzo, F., Hamidi, F., Phys. Chem. Chem. Phys., 113, 11 (2000), 4498. 8. Devautour, S., Abdoulaye, A., Giuntini, J.C, Henn, F., J. Phys. Chem. B, 105 (2001), 9297. 9. Pissis, P., Daoukaki-Diamanti, D., J. Phys. Chem. Solids, 54 (1993), 701. 10. Artioli, G., Smith, J.V., Kvick, A., Pluth, J.J., Stahl, K., Zeolites, Ed. Dram., B., Hocevar., S., Paovnik, S., Elsevier, Amsterdam, (1985) 249. 11. Vitale, G., Mellot, CF., Bull, L.M., Cheetham, A.K., J. Phys. Chem. B, 101 (1997), 4559. 12. Lignieres, J., Newsam, J.M., Microporous and Mesoporous Materials, 28 (1999), 305. 13. Grillo, M.E., Carrazza, J., J. Phys. Chem. B, 100 (1996), 12261. 14. Faux, D., J. Phys. Chem. B, 103 (1999), 7803. 15. Dewis, D.W., Ruiz Salvador, A.R., Almora Barios, N., Gomez, A., Mistry, M., Molecular Simulation, 28, 6-7 (2002), 649. 16. Almora-Barios, N., Gomez, A., Ruiz Salvador, A.R., Mistry, M., Lewis, D.W., Chem. Commun., (2001), 531. 17. Faux, D., Smith, W., Forester, T.R., J. Phys. Chem. B, 101 (1997), 1762. 18. Sanders, M.J., Leslie, M, Catlow, C.R.A., J. Chem. Soc. Chem. Commun. (1984), 1271. 19. Jackson, R.A., Catlow, C.R.A., Molecular Simulation, 1, 207, 1988. 20. Lau, K.F., Alper, H.E., Thatcher, S., Stouch, T.R., J. Phys. Chem. B., 98 (1994), 8785. 21. Manon Higgins, F., de Leeuw, N.H., Parker, S.C, J. Mater. Chem., 12 (2002), 124. 22. Maurin, G., Bell, R., Devautour, S., Henn, F., Giuntini, J.C, J. Phys. Chem. B, in press. 23. Maurin, G., Senet, P., Devautour, S., Henn, F., Giuntini, J.C, Van Doren, V.E., Computational Materials Science 22 (2001), 106. 24. Maurin, G., Senet, P., Devautour, S., Gaveau, P., Henn, F., Van Doren, V., Giuntini, J.C, J. Phys. Chem. B, 105 (2001), 9157. 25. Allen, M.F., Tildesley, D., Computer Simulation of Liquids, Oxford, Science Publication, Oxford (1987). 26. Gale, J.D., J. Chem. Soc. Faraday. Trans., 93 (1997), 629. 27. Smith, W., Forester, T., J. Mol. Graphics, 14 (1996), 136.