A Numerical Evidence for
Nonframework Cation Redistribution
Upon Water Adsorption in Faujasite
Zeolite
Christle Beauvais, Anne Boutin, and Alain H. Fuchs*[a]
Aluminosilicate faujasite-type zeolites are nanoporous solids,
which are widely used in gas-adsorption and separation processes. In faujasite, the presence of aluminum atoms introduces
charge defects that are compensated with some nonframework cations (sodium, potassium, calcium, barium, etc.), which
are usually located in well-defined crystallographic sites. Depending on the detailed chemical composition of the solid, a
selective adsorption can be observed in favor of one or the
other component of a binary mixture. This is the case, for instance, for sodium Y faujasite (Y meaning that the silicon-toaluminum ratio, Si:Al, is larger than 1.5), which retains, selectively, m-xylene rather than p-xylene when a sample of NaY
solid is exposed to a mixture of these two isomers. It has been
shown that the adsorption selectivity in this case is driven by
equilibrium (thermodynamics) considerations rather than by a
difference in transport properties of the two isomers.[1, 2] This
kind of separation process is interesting in many ways: Isomers
are often difficult to separate by means of conventional methods such as distillation. There is also an energetic interest since
nanoporous solid-based separations can often be performed
close to room temperature. This explains why a large research
effort has been undertaken in the past decade to understand
adsorption selectivity in these systems at the molecular
level[3, 4] and to better predict which kind of zeolitic (or other
open framework inorganic) material would be most appropriate for a given separation.
The issue addressed in the present Communication is the
effect of a small amount of water on the separation processes
of hydrocarbon mixtures. It was observed for some time that
the (often unwanted) presence of preadsorbed water in the
nanoporous sample affected the adsorption selectivity with respect to the hydrocarbon mixture to be separated.[5] Most
often, traces of water tended to “kill” the adsorption selectivity
observed in the dry solid thus severely hindering the intended
process. From time to time though, selectivity enhancement
was observed, as in the case of the above-mentioned separation of xylene isomers. It was found, for instance, that 5 %
(weight) of water enhanced the selectivity of barium-X faujasite in favor of p-xylene, compared to the dry solid situation.[6]
The mechanism responsible for these effects is poorly understood. Being able to understand and predict the effect of
[a] C. Beauvais, Dr. A. Boutin, Prof. A. H. Fuchs
Laboratoire de Chimie Physique, UMR 8000,
CNRS–Universit, Paris-Sud
B-timent 349, 91405 Orsay (France)
Fax: (+ 33) 169-156-188
E-mail: fuchs@lcp.u-psud.fr
ChemPhysChem 2004, 5, 1791 –1793
DOI: 10.1002/cphc.200400195
water on hydrocarbon adsorption is considered a key challenge in the adsorption community today.[5]
Adsorption properties in zeolites are closely related to the
location of nonframework cations and to their accessibility to
adsorbed molecules. The way these cations are distributed
among the available sites does not usually change during the
course of the (nonpolar) hydrocarbon adsorption process. On
the other hand, cation redistribution is suspected to occur
upon adsorption of polar molecules.[7] Recently, for instance,
Mellot-Draznieks et al.[8] carried out a neutron scattering study
of CFCl3 adsorption in NaY and observed cation redistribution
together with a new and previously unknown cation location.
We recently developed a Monte Carlo simulation method that
enables to predict the cation distributions in dry zeolites for a
given Si:Al ratio;[9] here we use this method to predict the
cation distributions in sodium faujasite in presence of preadsorbed water.
The cation distribution in faujasite is usually described as follows (see Figure 1): Na + ions can occupy sites I, which are lo-
Figure 1. Schematic view of a faujasite supercage with the site I, I’, II, and III locations.
cated in the hexagonal prisms connecting the so-called sodalite cages. Sites I’ are inside the sodalite cages facing sites I.
Sites II are in front of the six-rings inside the supercages. Sites
III are also in the supercages, near the four-rings of the sodalite
cages. Site I has a multiplicity of 16 per unit cell, sites I’ and II
have a multiplicity of 32, and site III has a multiplicity of 48 per
unit cell. Site III is believed to be of higher potential energy
than sites I, I’, and II. At low occupancy (Si:Al > 2), cations are
usually believed to occupy sites I, I’, and II only.[11]
We have considered here the case of NaY faujasite (Si:Al = 3;
48 cations per unit cell, a unit cell being composed of eight supercages), for which extensive experimental and molecular
simulation studies on the separation of xylene isomers
exist.[1, 2, 12] The cation distribution in a dry Na48Y sample corresponds to a full occupancy of sites I (16 cations) and sites II
(32 cations), while all sites I’ and III remain empty.[9]
We performed molecular simulations by using the “canonical
replica-exchange Monte Carlo” method,[10] which was recently
implemented for the study of cation redistributions in zeolites.[9] In this method, several independent realizations of the
system are simulated simultaneously, each differing in temperature. As the simulation proceeds, systems at adjacent temperatures are allowed to interchange configurations from time to
C 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1791
time, subject to specific acceptance criteria. These swap-moves
improve considerably the sampling of configuration space. The
higher temperatures are chosen in such a way that the system
can easily overcome energy barriers and thus provide the lowtemperature systems with configurations that cover a broad
range of the configuration space. This method was used here
to compute the stable-cation distribution of Na48Y at room
temperature for different water contents, which ranged from 0
to 370 molecules per unit cell.
Faujasite is known to be a very stable zeolite (even in its dehydrated form), and hydration has very little impact on its
overall structure.[13] We thus used a rigid framework system, as
in our previous studies.[2, 9, 12] As no hydrolysis takes place upon
water adsorption,[14] due to the weak bonding of water with
the faujasite framework, we used the simple TIP4P effective
potential model for water.[15]
The computed stable-sodium-cation distributions at 300 K
are shown in Figure 2. For low water content, we observed a
Figure 2. Sodium-cation distributions for different water contents: sites I (*),
sites I’ (&) and sites II (~).
solvation of the site-II cations in the supercages. Above
60 molecules per unit cell (i.e., roughly two water molecules
per site-II cation), a redistribution of site-I and site-I’ cations
takes place. Sodium cations progressively move from sites I to
neighboring sites I’ in the sodalite cage. This is accompanied
by a progressive occupancy of the sodalite cages by water
molecules (Figure 3).
Although these data were obtained using equilibrium
Monte Carlo simulations (i.e., we have no direct information
on the cation and water dynamics), it seems clear that we are
faced with a correlated cation–water motion. To begin with,
water molecules are preferentially adsorbed in sites II (only
site-II cations are directly accessible). By the time each site-II
cation is solvated by roughly two water molecules, it becomes
energetically interesting for water molecules to solvate other
cations. As the hexagonal prism that connects two sodalite
cages is too small to accommodate a water molecule, cations
in sites I will progressively move to sites I’, enabling water molecules to adsorb in sodalite cages. Landau free-energy and correlation-factor calculations (not shown here) have quantitatively confirmed this mechanism. This is, to our best knowledge,
the first numerical evidence of nonframework cation redistribution upon water adsorption in a zeolitic material. In a previous
1792
Figure 3. Number of water molecules in the sodalite cages as a function of the
site I’ occupancy. The slope of the solid line is close to two molecules per site I’
cation.
molecular dynamics (MD) study devoted to hydrated zeolite
4A, Faux[16] observed a relatively slow motion of some of the
Na + ions (one order of magnitude slower than the water diffusion in the same system). Using the same type of method, Jaramillo et al.[17] also observed the diffusion of some of the cations in a model of NaX upon adsorption of two polar hydrofluorocarbons. Standard molecular dynamics is obviously not
well-suited for studying equilibrium cation distributions (and
redistributions), since cation diffusion takes place on a timescale which is larger (or comparable) to the MD timescale.
We carried out the same Monte Carlo simulations of cation
distributions in Na48Y with different water contents in the presence of preadsorbed m-xylene molecules (four molecules per
supercage, which corresponds to the maximum equilibrium
loading). The results are shown in Figure 4. At low water content, both xylene and water molecules are preferentially attracted by site-II cations. Since sodalite cages are too small to
accommodate a xylene molecule, there is a steric and entropic
effect that favors xylene adsorption at sites II and water adsorption in the sodalite cages. For this reason I–I’ cation redis-
Figure 4. Sodium-cation distributions for different water contents in presence
of four m-xylene molecules per supercage: sites I (*), sites I’ (&) and sites II (~).
The dotted lines and open symbols correspond to the data shown in Figure 2.
C 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org
ChemPhysChem 2004, 5, 1791 –1793
tribution takes place faster in this case (i.e., at low water content) than in the previous one (dotted line in Figure 4). This
demonstrates that a very small amount of water in a hydrocarbon–zeolite host–guest system can lead to a redistribution of
the nonframework cations.
Finally, an adsorption selectivity analysis was undertaken for
an equimolar mixture of m- and p-xylene by using the methods developed earlier in the case of dry faujasites.[12] We found
that the cation redistribution observed in this system for a
very low water content [roughly 2 % (weight), see Figure 4]
leads to a decrease by a factor of four in the adsorption selectivity of NaY in favor of m-xylene. This phenomenon has to do
with the competition of water and xylene molecules for adsorbing in front of site-II cations. The presence of a water molecule close to site II tends to displace the xylene molecules towards the center of the supercage. Site II is thus destabilized
for xylene adsorption in almost the same way as explained in
ref. [12], when a sodium cation is replaced by a larger potassium ion. This explains why the adsorption selectivity decreases
in this case. Work is in progress to provide a complete molecular mechanism for this process.
We have reported, in this work, the first clear-cut numerical
evidence of nonframework cation redistribution upon water
adsorption in sodium Y faujasite zeolite. We suggest a mechanism responsible for this effect, which is based on a correlated
motion of water molecules (from supercages to sodalite cages)
and sodium cations (from site I in the hexagonal prisms to site
I’ in the sodalite cages). When the faujasite supercages are
filled with xylene molecules, the cation redistribution takes
place for very low water content, and the adsorption selectivity
is modified by a factor of four. This is in accordance with the
experimental observation of a dramatic effect of water traces
on some separation processes. One should be aware that the
described redistribution mechanism is not a general one. It
will, obviously, depend on the Si:Al ratio and on the nature of
the zeolite framework.
The molecular simulation tools used in this work can, in
principle, be extended to any type of guest–host systems.
These methods enable to disclose the molecular mechanisms
of cation redistribution and adsorption selectivity in a given
multicomponent system. This should help, in the near future,
in the rational design of nanoporous materials for separation
purposes.
subjected to the external field imposed by the rigid faujasite
framework host. For each water content, eight independent realizations of the system were simulated simultaneously. The chosen
temperatures ranged between 300 and 2325 K. Simulations runs
lasted for 30 to 40 million steps. Swap between replica were attempted with a rate of 0.1 %, the acceptance rate being 3 %.
Acknowledgement
We acknowledge Florent Calvo for fruitful discussions.
Keywords: adsorption · cations · molecular modeling ·
statistical thermodynamics · zeolites
[1] V. Cottier, J.-P. Bellat, M.-H. Simonot-Grange, A. MNthivier, J. Phys. Chem.
B 1997, 101, 4798.
[2] V. Lachet, B. Tavitian, A. Boutin, A. H. Fuchs, Langmuir 1999, 15, 8678.
[3] A. H. Fuchs, A. K. Cheetham, J. Phys. Chem. B 2001, 105, 7375.
[4] B. Smit, R. Krishna, Chem. Eng. Sci. 2003, 58, 557.
[5] F. Meunier, in Proceedings of the 7th Conference on Fundamentals of Adsorption, (Eds.: K. Kaneko, H. Kanoh, Y. Hanzawa), IK International, 2002,
1.
[6] C. Pichon, PhD Thesis, UniversitN de Bourgogne (France), 1999.
[7] C. P. Grey, F. I. Poshni, A. F. Gualtieri, P. Norby, J. C. Hanson, D. R. Corbin,
J. Am. Chem. Soc. 1997, 119, 1981.
[8] C. Mellot-Draznieks, J. Rodriguez-Carvajal, D. E. Cox, A. K. Cheetham,
Phys. Chem. Chem. Phys. 2003, 5, 1882.
[9] C. Beauvais, X. Guerrault, F.-X. Coudert, A. Boutin, A. H. Fuchs, J. Phys.
Chem. B 2004, 108, 399.
[10] D. A. Kofke, J. Chem. Phys. 2002, 117, 6911.
[11] G. Vitale, C. F. Mellot, L. M. Bull, A. K. Cheetham, J. Phys. Chem. B 1997,
101, 4559
[12] V. Lachet, S. Buttefey, A. Boutin, A. H. Fuchs, Phys. Chem. Chem. Phys.
2001, 3, 80.
[13] D. W. Lewis, A. R. Ruiz-Salvador, N. Almora-Barrios, A. Gomez, M. Mistry,
Molec. Sim. 2002, 28, 649.
[14] J. C. MoOse, J. P. Bellat, A. MNthivier, Microporous Mesoporous Mater.
2001, 43, 91.
[15] W. L. Jorgensen, J. Chadrasekhar, J. D. Madura, R. W. Impey, M. L. Klein,
J. Chem. Phys. 1983, 79, 926.
[16] D. A. Faux, J. Phys. Chem. B 1999, 103, 7803.
[17] E. Jaramillo, C. P. Grey, S. M. Auerbach, J. Chem. Phys. 2001, 105, 12 319.
[18] E. Jaramillo, S. M. Auerbach, J. Chem. Phys. 1999, 103, 9589.
Received: April 29, 2004
Revised: July 22, 2004
Computational Section
The cation force field was adapted from the work of Jaramillo and
Auerbach[18] in the way described in ref. [9]. The cation-framework
potential consists of an exp-6 repulsion–dispersion term that acts
between the cation and the oxygen atoms of the faujasite and a
Coulombic term that acts between the cation and both the
oxygen and the T atoms of the framework. Sodium cations interact
with each other through a single Coulombic term. The TIP4P potential was used for water molecules.[15] Lorentz–Berthelot combining rules were used to obtain the cross framework–water interaction terms. Ewald sums were used to calculate the long-range Coulombic terms. We performed replica exchange canonical Monte
Carlo simulations of the sodium cations and TIP4P water molecules
ChemPhysChem 2004, 5, 1791 –1793 www.chemphyschem.org
C 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1793