Performance of Chiral Zeolites for Enantiomeric Separation Revealed by MolecularSimulation

J. M. Castillo,†,‡ T. J. H. Vlugt,‡ D. Dubbeldam,§ S. Hamad,† and S. Calero*,†

Department of Physical, Chemical, and Natural Systems, UniVersity Pablo de OlaVide, Ctra. Utrera km. 1,41013 SeVille, Spain, Process & Energy Laboratory, Delft UniVersity of Technology, Leeghwaterstraat 44,2628CA Delft, The Netherlands, and Van’t Hoff Institute of Molecular Sciences, UniVersity of Amsterdam,Science Park 904, Amsterdam, The Netherlands

ReceiVed: August 21, 2010; ReVised Manuscript ReceiVed: NoVember 4, 2010

We used molecular simulations to test the adsorption selectivity for enantiomers in the chiral zeolites SOF,STW, and ITQ-37. This work is the first simulation study which demonstrates the chirality of SOF, STW,and ITQ-37. We obtain information at the molecular level that explains the nature of the chiral selectivity.We selected CHBrClF and 4-ethyl-4-methyloctane as the probe molecules for the study. We calculated purecomponent and racemic mixture adsorption isotherms, Henry coefficients, adsorption enthalpies, and radialdistribution functions. While STW is enantioselective for the probe molecules studied, ITQ-37 is likely to beenantioselective for other chiral molecules, and SOF is not a candidate for chiral separation. Appropriatepore geometry, resulting in differences in the adsorption behavior between the two isomers, is required forobserving chiral selectivity. We provide insight that could be used for the synthesis and characterization ofnew chiral structures aimed to separate specific isomers. Using the methodology described here, it could bepossible to easily determine if a new zeolite structure would show adsorption selectivity for a given chiralmolecule.

Introduction

The separation of a racemic mixture is of crucial importancefor the pharmaceutical and agrochemical industries.1,2 Forexample, one of the isomers of a chiral molecule may have atherapeutic effect and the other one may have toxicologicalimpact.3 During the past few years a wide variety of methodshave been developed to achieve chiral separation, such ascrystallization, asymmetric catalysis, or liquid chromatography.4-7

Chiral porous materials usually have potential to selectivelyadsorb chiral molecules or to serve as a chiral reactionenvironment. The number of chiral porous materials showingenantioselectivity is relatively limited,8 although there has beena great advance in the synthesis of new chiral materials duringthe last years. Chirality has been induced in silicas and sol-gelmaterials.9 New materials have been synthesized using chiraltemplates, such as polymerized organogels by the use of chiralorganic gelators.10 Chirality can also be induced in crystals withlarge surface area, such as zeotypes,11 zeolites,11 aluminophos-phates,11 and metal-organic frameworks (MOFs).12,13 Fre-quently, the product obtained after the synthesis of thesematerials is an intergrowth of different polymorphs.11 Therefore,new approaches have been developed to create homochiralporous structures.

Zeolites are preferred to other porous materials in catalysisand separation applications due to their relatively low cost,permanent porosity, and high thermal and chemical stability.14

Zeolites can be modified to act as asymmetric catalysts. Themost common methods to induce a chiral reaction environmentin the zeolite are: (1) adsorption of chiral molecules in an achiral

zeolite (chiral inductor method);15,16 (2) covalent bonding ofchiral molecules to the achiral zeolite structure (chiral auxiliarymethod);15,16 and (3) a combination of the chiral inductor andthe chiral auxiliary methods.15,16 The confined space and thenonframework cations present in the zeolite are believed to beresponsible for the chiral induction.15,16 It has also beendemonstrated that zeolite defects in the form of silanol groupscan improve enantioselectivity in chirally modified zeolites.17

Chiral catalysis is also possible in mesoporous materials suchas MCM-41 and SBA-15.18

Only a few as-synthesized zeolite-like materials were identi-fied as chiral, namely, BEA,19 CZP,20 GOO,21 OSO,22 andBSV.23 Recently, 20 other well-known zeolites have beenidentified as chiral, due to the fact that they have chiralasymmetric unit cells.24 Computational methods have beendeveloped to help in the design of new chiral materials.25,26 Themain conclusion of these computational studies is that in orderto obtain an energetically stable chiral zeolite, some of the Tatoms of the structure should be different from silicon andsubstituted by another type of atoms such as B, Be, Zn Ge, orS.26 Following this approach, three new chiral germanosilicatematerials were synthesized recently: SOF,27 STW,27 and ITQ-37.28 These zeolites crystallize in a chiral space group, so theyhave a gyroidal structure. It has been reported that SOF andSTW are not stable in their pure silica form,29 so the inclusionof Ge in these frameworks is vital for their stability. SOFconsists of straight, elliptical channels, which are crossed byzigzag channels. The twist of the zigzag channels defines thehandedness of the structure (see Figure 1a). STW is formed byrectangular cages connected by two narrow and two widewindows. The cages have three different orientations, and thechirality is determined by the relative orientation of the cagesand the windows between cages (see Figure 1b). ITQ-37 is alow density zeolite with extra large 30-ring windows. It consists

* To whom correspondence should be addressed, scalero@upo.es.† Department of Physical, Chemical, and Natural Systems, University

Pablo de Olavide.‡ Process & Energy Laboratory, Delft University of Technology.§ Van’t Hoff Institute of Molecular Sciences, University of Amsterdam.

J. Phys. Chem. C 2010, 114, 22207–22213 22207

10.1021/jp1079394 2010 American Chemical SocietyPublished on Web 12/01/2010

of two different types of cavities, each of them connected toother three cavities to form a complex, gyroidal channel system(see Figure 1c). One major advantage of these materials withrespect to other chiral zeolites is that they are homochiral aftersynthesis without further modifications. This means that singlecrystals of each of these zeolites exhibit only one hand, andtherefore they have an intrinsic chiral structure. After synthesiswe would obtain a racemic mixture of these crystals. Experi-mental studies have been performed to favor one crystalpolymorph in intergrown chiral zeolites.30 Recently, a simpleabrasive grinding method has been successfully used to obtainsingle chiral crystals from an initially racemic mixture ofconglomerate crystals.31,32 Similar procedures may be used toobtain only one type of zeolite crystal instead of a racemicmixture. To the best of our knowledge, experimental adsorptionisotherms of chiral molecules in any of these three chiral zeoliteshave not been reported yet, and it is not clear if they showadsorption enantioselectivity for typical chiral guest molecules.It has been demonstrated experimentally that other chiral zeolites

may provide enantioselectivity large enough to achive chiralseparation, such as GOO24 and BEA.30

Computer simulations in chiral porous materials have studieddifferent aspects of asymmetric catalysis and chiral selectivity.Due to the complexity of these systems, computer simulationson chiral adsorption selectivity are scarce. Molecular simulationsin intrinsically homochiral zeolites are not available in theliterature. Most simulations studies are centered around chirallymodified zeolite Y and the different isomers of BEA. Averyet al.33 reviewed computer simulation studies on the structuralaspects of enantioselective asymmetric catalysis, concluding thatsmall energy differences along the reaction pathways to chiralproducts are decisive. Sivasubramanian et al.34 optimized thegeometries of cation complexes in a chiral inductor and a chiralauxiliary modified zeolite Y using quantum chemistry calcula-tions to explain the enantioselectivity of different chiralmolecules and demonstrated the importance of the zeolite cationsin asymmetric reactions. Willock et al.35 used hybrid MonteCarlo (MC) and Molecular Dynamics (MD) simulations to

Figure 1. Pore systems of the zeolites SOF (a), STW (b), and ITQ-37 (c). In the energy representations on the left, the interior of the pores iscolored in red, the exterior in gray. In (a) and (b) the energy representations are rotated with respect to the view on the right to show clearly thechannels and cages of the structures, respectively. STW is formed by identical, rectangular cages oriented in three different directions. Every cageconnects to other cages by two large windows and two small windows. The channel system of ITQ-37 is quite complex, with intersection channelsin three different directions.

22208 J. Phys. Chem. C, Vol. 114, No. 50, 2010 Castillo et al.

explain the chiral modifier stability in protonated zeolite Y.Jirapongphan et al.36 studied the interactions of chiral modifierswith chiral adsorbed molecules in zeolite Y by molecularsimulations, relating the different binding motifs to the observedenantioselectivity. Bao et al.37 obtained large separation factorsfor chiral hydrocarbons in the chiral MOF Cd-BINOL by MCsimulations. In their study these authors also highlighted theimportance of the different adsorption sites and their accesibilityduring chiral separations, concluding that the adsorbate-hostinteraction is the most important factor for enantioselectivityin this system. Van Erp et al.38-40 used molecular simulationsto show that adsorbed scalemic mixtures of chiral hydrocarbonsinduce an enantioselective environment in alumina-substitutedzeolites. Clark et al.41 studied the adsorption of chiral moleculesin zeolite BEA and zeotype UCSB-7. They showed that thereis a large enantioselectivity in the different channels of BEAdue to small differences in adsorption energies and adsorptionorientations of the different isomers.

In this work we explore the enantioselective capacity of thethree newly synthesized chiral germanosilicates, SOF, STW,and ITQ-37, using molecular simulations. We will focus on thepore geometry of the zeolites and not on their chemicalcomposition. We use bromochlorofluoromethane, CHBrClF, thesimplest chiral organic molecule, as well as other halocarbons,and the hydrocarbon 4-ethyl-4-methyloctane as probes to testthe adsorption selectivity of these materials. We explore thepossibility of describing chiral selectivity in zeolites by molec-ular simulations, and its dependence on the characteristics ofthe adsorbed molecules and the pore geometry. We describethe reasons for the different behavior of chiral molecules in thesezeolites, and the pore parameters that determine enantioselec-tivity in these chiral structures. We show that a chiral zeolitecan show chiral selectivity depending on the nature of theadsorbed molecules. The remainder of this paper is organizedas follows: in the next section we present the simulation forcefields and methods. The following section contains the resultsand discussion, and we present our conclusions in the finalsection.

Simulation Details

Details about the simulation parameters and models used canbe found in the Supporting Information (Tables T1-T4). In thiswork, we will consider the pure silica version of the zeolites.This approximation is justified by the fact that the valences ofGe and Si are identical and their electronegativities are verysimilar. The spectroscopic electronegativities of Si and Ge are11.33 and 11.80 eV, respectively.42 Therefore, we can safelyassume that the partial charges of Si and Ge atoms in the zeoliteare identical. In addition, the dispersive interaction of the Tatoms in the zeolite is shielded by the oxygen atoms.43 As aresult, the oxygen atoms of the zeolite are responsible of themost important contribution to the dispersive interaction betweenthe zeolite and the adsorbed molecules, so that differences inthe dispersive interaction of the T atoms can be neglected in afirst approximation.43,44 Our SOF and STW force fields use thisapproximation, which is widely used for pure siliceous zeolites.For simplicity, we will name the molecules of 1-bromo-1-chloro-1-fluoroalkyl halide as CnBrClF, n being the number of carbonatoms in the molecule.

Results and Discussion

The pure component isotherms of the two CHBrClF isomers,as well as the isotherm of their racemic mixture in SOF, STW,and ITQ-37, were calculated as a function of the total fugacity

at 298 K (see Figure 2). For SOF and ITQ-37 there are no largedifferences in adsorption between the two isomers, both for thepure components and in the racemic mixture. These differencesare smaller than the error bars and are only appreciable nearsaturation. The differences in adsorption in SOF and ITQ-37 atsaturation did not increase when the temperature was reduced,so we conclude that these structures do not provide enantioselec-tivity for the two isomers of CHBrClF. On the other hand, wefound adsorption selectivity in STW for the racemic mixture. TheR-CHBrClF isomer is preferentially adsorbed with an enantio-meric excess of 18%. We define the enantiomeric excess (ee) as

Figure 2. Racemic mixture (open symbols) and pure componentisotherms (closed symbols) for the two CHBrClF isomers in SOF (a),STW (b), and ITQ-37 (c) at 298 K calculated by GCMC simulationsas a function of the total fugacity: squares, L-isomer; triangles,R-isomer. The total loading in the racemic mixture is given by thedashed line.

Zeolite Enantiomeric Separation J. Phys. Chem. C, Vol. 114, No. 50, 2010 22209

where θL and θR are the loadings of isomers L and R in the zeolite,respectively. The results provide evidence that the pore geometryof STW is an excellent environment for the adsorption separationof this molecule.

The separation of CHBrClF in STW takes place at highloadings, which suggests that the selectivity is caused by entropiceffects. At low loadings the adsorption is equal, so that there isno selectivity in the Henry regime, and the pure componentisotherms of the two isomers are identical. We have computedthe adsorption enthalpy at zero loading for the two isomers inSTW at different temperatures (Figure 3). At temperatures below150 K the R-isomer is preferentially adsorbed, while attemperatures above 150 K this is the L-isomer. This preferenceis reversed at the saturation loading at 298 K. This shows thatthe two isomers interact differently with the zeolite, in agreementwith previous experimental and simulation studies. Dryzunet al. showed that the isomers of histidine have a differentadsorption enthalpy in several chiral zeolites.24 Avery et al.showed that small differences in the conformation energy ofdifferent isomers in chirally active materials can lead to largeeffects in selectivity.33 In STW, the differences in the adsorptionenthalpy at zero loading are not reflected in differences in thepure component adsorption isotherms, but the slightly differentinteraction with the zeolite can be a reason for the differentadsorption in the mixture. We also computed the Henrycoefficients and adsorption enthalpies at zero loading ofCHBrClF in SOF, STW, and ITQ-37 at different temperatures(Supporting Information, Tables T5, T6, and T7 respectively).In SOF there is not a significant difference between the twoisomers of CHBrClF, so we do not observe any selectivity. Wefind larger differences for ITQ-37 than for STW. We concludethat ITQ-37 does not show adsorption selectivity for CHBrClFbecause its pores are much larger than the pores of STW, sopacking efficiency is less important for this structure.

To study the influence of molecular packing in the separationof CHBrClF isomers in STW, we calculated the radial distribu-tion functions (RDF) of the molecules adsorbed at 298 K anda total fugacity of 100 kPa. At these conditions there is a clearseparation in the equimolar mixture. The R-isomer preferentiallyadsorbed (see Figure 2). A given type of molecule packs moreefficiently than others when more molecules of the first typecan fit in the same volume inside the porous material. The RDFis related to the average distance between molecules of the sametype. One molecule will pack more efficiently than another ifthe peaks of the RDF of the former molecule are higher and atlower distances than in the RDF of the latter. The C-C RDFsof the pure component and racemic mixture of CHBrClF isomersare shown in Figure 4. The Br-Br, Cl-Cl, F-F, and H-HRDFs can be found in the Supporting Information (FiguresS2-S5). A mixture of molecules of different handednessprovides the most efficient packing. On the other hand, thepacking of the R-isomer is much more efficient when adsorbedas pure isomer than in the mixture, while there is only a smallimprovement in the packing efficiency for the L-isomer in themixture compared to the pure isomer. As a consequence, theloss of packing efficiency of the R-isomer in the presence ofL-isomer molecules causes a small excess of R molecules inthe adsorbed phase of the racemic mixture. The RDFs also showthat, when the packing efficiency is optimal, the distancebetween Cl atoms of different molecules is maximal (Supporting

Information, Figure S3). This suggests that the adsorptionseparation at saturation conditions is due to steric effects.

To understand the adsorption mechanisms of CHBrClFisomers in STW, in Figure 5 we show snapshots obtained duringthe MC simulations at saturation conditions of the pure isomersL and R. Every cage of STW can contain up to two CHBrClFmolecules. The large windows connecting the cages of STWare marked with dashed lines. The molecules are adsorbed closeto both types of windows of STW. Molecules adsorbed closeto the large windows have one of its atoms in the middle of thewindow, while the rest of the molecule is at the cage. In mostcases, the atom at the large window is Cl in the case of theL-isomer and Br for the R-isomer. Molecules can also beadsorbed with one of its atoms at the middle of a small window,although the orientation of molecules at the small window doesnot follow a clear trend. Nevertheless, this configuration is notfavored due to the reduced size of the small window. The maindifference in the adsorption behavior between isomers takesplace at the center of the large window. This is in agreement

Figure 3. Adsorption enthalpy at zero loading of the different isomersof CHBrClF in STW as a function of temperature: squares, L-isomer;triangles, R-isomer.

Figure 4. Radial distribution functions of the carbon-carbon distancesof CHBrClF adsorbed in STW, obtained by MD simulations in theNVT ensemble. A constant total loading of 12 molecules per unit cellwas imposed, corresponding to a loading of five and seven moleculesper unit cell for the L- and R-isomers in the mixture, respectively. Theimposed loading was made equal to the loading at saturation conditionsin the calculated isotherm at 298 K and a total fugacity of 100 kPa. Inthe figure we show the results for the pure L (solid line) and R (dashedline) isomers, the L (dotted line) and the R (dashed dotted line) isomersin the mixture, and the L-R RDF in the mixture (short dashed line).

ee100

)|θL - θR|

|θL + θR|

22210 J. Phys. Chem. C, Vol. 114, No. 50, 2010 Castillo et al.

with experimental results in chirally modified zeolites, whichdemonstrates that a tight fitting between the reactant and thechiral inductor is needed to achieve enantioselectivity.45 In thecase of CHBrClF, Br and Cl atoms have the larger van derWaals radius, so these atoms interact more tightly with thezeolite and therefore play the most active role in thezeolite-molecule interactions.

We have analyzed the influence of pore size distribution inthe host framework on enantioselectivity for the CHBrClFmolecule. We have generated modified SOF, STW, and ITQ-37 structures by stretching and shrinking the distances betweenframework atoms in steps of 10% and calculated their pore sizedistributions (Supporting Information, Figure S6). We havecomputed the racemic mixture isotherms of CHBrClF in thesemodified SOF, STW, and ITQ-37 structures (Supporting Infor-mation, Figure S8). CHBrClF is not adsorbed in SOF shrunkby more than 10%, and STW by more than 20%, due to thereduced pore size of the host frameworks. For SOF we onlyobtain a low enantiomeric excess of 4% when the atomicdistances are increased 10%. For STW, we find adsorptionseparation in all the structures tested, although for the stretchedstructure adsorption only occurs at very large fugacities. Thestretched structure of STW provides a larger enantiomeric excessthan the original and shrunk structures. Finally, the ITQ-37structure provides a large enantiomeric excess of 40% only whenits size is reduced 30% (Supporting Information, Figure S8).According to our results, a tight fitting of the molecules adsorbedis a necessary condition to achieve separation in one porousmaterial. The material has to contain pores with a diameter ofthe same order as the van der Waals diameter of the largerbranches of the chiral molecule. The van der Waals diametersof the Cl and Br atoms in our CHBrClF model are 6.94 and 7.3

Å, respectively. The investigated structures which provideenantioselectivity have pores with mean diameters between 6.3and 8.7 Å. Furthermore, structures which provide large enan-tioselectivity have two different types of pores, both with adiameter similar to the diameter of the larger atoms in themolecule. The difference in diameter of these two types of poresis on the order of 0.2-0.4 Å, which is in the range of thedifference between the van der Waals diameters of the Br andCl atoms. It is noteworthy that, when the difference in the Henrycoefficients and adsorption enthalpies at zero loading of the twoisomers in one zeolite type is larger (see Tables S6, S7, andS8), the enantiomeric excess obtained in its stretched or shrunkversions is also larger. These pore characteristics are necessaryto provide enantioselectivity for the CHBrClF molecule, butthey are not sufficient. MFI has two pores with diameters 8.1and 8.4 Å (Supporting Information, Figure S7) and thereforesatisfy the previous conditions for enantioselectivity. Nonethe-less, this zeolite is not enantioselective for CHBrClF. We believethat, apart from having a tight fitting between the host and theguest molecule, to obtain enantioselectivity the chiral moleculeneeds to encounter a chiral environment. The conclusion is thatonly the zeolites that are chiral can show enantioselectivityprovided that the former conditions are met.

To check the adsorption selectivity of large chiral moleculesin SOF, STW, and ITQ-37, we computed pure component and50/50 mixture isotherms of 4-ethyl-4-methyloctane in thesestructures at 200 and 298 K. 4-Ethyl-4-methyloctane is notadsorbed in SOF. This is due to the small pore diameter andlow pore volume of SOF. The pore volumes of SOF, STW,and ITQ-37 calculated by MC simulations are 0.145, 0.204, and0.719 cm3/g, respectively. The pore volume of SOF is 12%smaller than that for MFI (0.164 cm3/g), another zeolite with asimilar channel system, where 4-ethyl-4-methyloctane moleculescan be adsorbed using the same simulation techniques.38

Although the straight channels of SOF are formed by large 12-rings, the channel sections and the windows between straightand zigzag channels are elliptical, which may hinder theadsorption of large branched molecules.

The racemic mixture isotherms of 4-ethyl-4-methyloctane inSTW at 200 K and ITQ-37 at 298 K are shown in Figure 6.We obtain large error bars for the adsorption of 4-ethyl-4-methyloctane. The acceptance probability for insertion moves,although significant, was low. STW shows a small adsorptionselectivity for 4-ethyl-4-methyloctane, which is only evident atlow temperatures. At 200 K, the adsorption isotherm yields anenantiomeric excess of 20%. In the case of ITQ-37, we do notfind any evidence of differences in adsorption.

In Figure 7 we show a snapshot of a racemic mixture of4-ethyl-4-methyloctane adsorbed at 200 K and at saturationconditions in STW. The C2H5 branches of the molecules arepreferentially pointing in the direction of the small windows.The C3H7 branch of the L-isomer crosses one of the largewindows, entering the neighboring box, while the tip of the C4H9

branch stays close to the center of the other large window ofthe same box. In the case of the R-isomer, the chiral center ismuch closer to one of the large windows, and the C4H9 branchof the molecule enters nearly completely into the neighboringbox. The tip of the C3H7 branch of the R-isomer stays close tothe center of the other large window of the same box. Therefore,the L-isomer can adsorb in STW occupying less space than theR-isomer, so it is the component preferentially adsorbed at highloadings.

Apart from size, the main difference between the CHBrClFand the 4-ethyl-4-methyloctane molecules is that the former

Figure 5. Snapshots of pure L (top), and pure R (bottom) CHBrClFisomers adsorbed in STW at 298 K and a total fugacity of 100 kPa.The central carbon atom of the molecule is colored red for the L-isomerand light blue for the R-isomer. The different atoms of the moleculeare also colored: Br ) gray; Cl ) yellow; F ) green; H ) white. Thelarge windows perpendicular to the page are marked with a dashedline. The large windows parallel to the page are marked with a dashedcircle.

Zeolite Enantiomeric Separation J. Phys. Chem. C, Vol. 114, No. 50, 2010 22211

contains pseudoatoms carrying a partial charge. We checkedthe influence of the charge in the separation efficiency ofCHBrClF isomers in STW. For this purpose, we performed MCsimulations in the canonical ensemble. We use as a startingconfiguration the final configuration of the grand-canonical MCsimulation of CHBrClF in STW at 298 K and 100 kPa (5.75molecules/uc of L-isomer, 6.25 molecules/uc of R-isomer). In

these simulations, we scaled all the partial charges of theCHBrClF pseudoatoms by factors ranging from zero to 1.4. Themolecule remained charge neutral in this process. The enan-tiomeric excess obtained as a function of the charge scalingfactor is shown in Figure 8. The enantiomeric excess follows anearly linear trend, with one sudden increase at a scaling factorof 0.6. When the partial charges are zero, we still obtainadsorption separation, although at a low enantiomeric excessof 1.2%. This confirms that the STW structure studied here isenantioselective for molecules that have a tight fitting in its poresand that the selectivity increases when the interaction of theisomers adsorbed is larger.

The adsorption isotherms of the racemic mixtures of C2BrClFand C3BrClF in STW and 298 K are shown in Figure 9. Theadsorption isotherms of the racemic mixtures of CnBrClFmolecules up to n ) 6 in SOF, STW, and ITQ-37 at 298 K canbe found in the Supporting Information (Figures S9-S12). Wedo not observe any adsorption selectivity for these molecules,while there is clear adsorption selectivity for the two differentisomers of CHBrClF. We believe that one chiral molecule, apartfrom having a tight fitting with a chiral environment, needs tohave certain asymmetry in order to present adsorption selectivity.The only difference between CHBrClF and C2BrClF is the

Figure 6. Adsorption isotherms for the racemic mixture of 4-ethyl-4-methyloctane isomers in STW at 200 K (a) and ITQ-37 at 298 K (b)as a function of the total fugacity: squares, L-isomer; triangles,R-isomer.

Figure 7. Snapshot of the mixture of L and R 4-ethyl-4-methyloctaneisomers adsorbed in STW at 200 K and a total fugacity of 0.1 Pa. TheL-isomer is colored in red and the R in blue. The large windowsperpendicular to the page are marked with a dashed line. The largewindows parallel to the page, and crossed by one adsorbed molecule,are marked with a dashed circle.

Figure 8. Variation of the enantiomeric excess of racemic mixturesof CHBrClF isomers adsorbed in STW at 298 K, and a constant loadingof 12 molecules/uc, as a function of the CHBrClF point charge values.The horizontal axis indicates the factor by which all the point chargesof the CHBrClF pseudoatoms are scaled. The error bars lie within thesymbol size.

Figure 9. Adsorption isotherms for the racemic mixture of C2BrClF(squares) and C3BrClF (triangles) isomers in STW at 298 K: closedsymbols, L-isomer; open symbols, R-isomer.

22212 J. Phys. Chem. C, Vol. 114, No. 50, 2010 Castillo et al.

substitution of the H of the first molecule by a CH3 group. Thesigma parameter for the H pseudoatom is 2.5 Å while the CH3

Lennard-Jones size parameter is 3.76 Å. This last value is similarto the sigma parameters of Cl (3.47 Å) and Br (3.65 Å).Therefore, in the C2BrClF molecule there are three brancheswith a very similar van der Waals diameter, destroying theoriginal asymmetry of the CHBrClF molecule.

Conclusions

Our calculations show that the STW structure is enantiose-lective for both CHBrClF and 4-ethyl-4-methyloctane isomers,while SOF and ITQ-37 are not. We obtain adsorption selectivitywithout making any distinction between the Ge and the Si atomsof the structure. Therefore, the adsorption selectivity is causedonly by the pore geometry of the zeolite and not by the Gesubstitution in the zeolite. Radial distribution functions showthat the reason for the adsorption selectivity is the differentpacking efficiency of the different isomers when adsorbed aspure components or in the mixture. Furthermore, the differentisomers adsorb differently at the points where the interactionwith the zeolite is tighter. As a result, the separation factor isproportional to the point charges of the molecules, so that theelectrostatic interaction of the molecule with the zeolite isstronger.

The SOF and ITQ-37 structures were scaled in size in orderto obtain a tight fitting environment for the probe molecules.While ITQ-37 showed adsorption selectivity when it was shrunkby a factor of 30%, we did not observe selectivity in SOF forany of the scaled structures tested. There are two possibleexplanations for this difference: the zeolite has to containadequate pore geometries to provide enantioselectivity, or thetwo isomers have to interact differently with the zeolite. Thislast point can be revealed by comparing the Henry coefficientsand adsorption enthalpies at zero loading of the two isomers.Finally, we found that if the branches of the chiral moleculeare made symmetrical in terms of interactions with the zeolite,the enantioselectivity disappears. Therefore, a chiral zeolite mayshow chiral selectivity depending on the nature of the moleculesadsorbed.

Acknowledgment. This work is supported by the Spanish“Ministerio de Educacion y Ciencia (MEC)” (CTQ2007-63229),the Junta de Andalucıa (P07-FQM-02595), and by the NationalScience Foundation (CTS-0507013). T. J. H. Vlugt acknowl-edges financial support from The Netherlands Organization forScientific Research (NWO-CW) through a VIDI grant. J. M.Castillo thanks EuroSim for his PhD fellowship.

Supporting Information Available: Simulation details,figures showing radial distribution function, pore size distribu-tions, and adsorption isotherms, and tables of Lennard-Jonesparameters, bond stretching ineractions, bond bending potentials,torsion potentials, and Henry coefficients. This material isavailable free of charge via the Internet at http://pubs.acs.org.

References and Notes

(1) Rouhi, A. M. Chem. Eng. News 2004, 82, 47.(2) Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Kesseler, M.;

Sturmer, R.; Zelinski, T. Angew. Chem., Int. Ed. 2004, 43, 788.(3) Hoffmann, R. The same and not the same; Columbia University

Press: New York, 1997.(4) Rekoske, J. E. AIChE J. 2001, 47, 2.

(5) Eckhardt, C. J.; Peachey, N. M.; Swanson, D. R.; Takacs, J. M.;Khan, M. A.; Gong, X.; Kim, J. H.; Wang, J.; Uphaus, R. A. Nature 1993,362, 614.

(6) Mikami, K.; Matsukawa, S. Nature 1997, 385, 613.(7) Hare, P. E.; Gilav, E. Science 1979, 204, 1226.(8) Blaser, H. U.; Muller, M. Enantioselective catalysis by chiral solids:

approaches and results. In Heterogeneus catalysis and fine chemicals II;Elsevier: Amsterdam, 1991; Vol. 59.

(9) Marx, S.; Avnir, D. Acc. Chem. Res. 2007, 40, 768.(10) Fu, X. J.; Yang, Y.; Wang, N. X.; Wang, H.; Yang, Y. J. J. Mol.

Recognit. 2007, 20, 238.(11) Yu, J.; Xu, R. J. Mater. Chem. 2008, 18, 4005.(12) Ke, Y. X.; Collins, D. J.; Sun, D. F.; Zhou, H. C. Inorg. Chem.

2006, 45, 1897.(13) Cao, X. Y.; Li, Z. J.; Zhang, J.; Qin, Y. Y.; Cheng, J. K.; Yao,

Y. G. Crystengcomm 2008, 10, 1345.(14) Handbook of zeolites: structure, properties and applications; Nova

Science Publishers, Inc.: New York, 2009.(15) Shailaja, J.; Sivaguru, J.; Uppili, S.; Joy, A.; Ramamurthy, V.

Microporous Mesoporous Mater. 2001, 48, 319.(16) Joy, A.; Kaanumalle, L. S.; Ramamurthy, V. Org. Biomol. Chem.

2005, 3, 3045.(17) Szabo, L. P.; Selmeczi, K.; Pilter, Z.; Szecsi, A.; Bodnar, J.; Lippai,

E. H.; Kallo, D.; Szotyory, L. Chromatographia 1998, 48, 140.(18) Li, C.; Zhang, H. D.; Jiang, D. M.; Yang, Q. H. Chem. Commun.

2007, 547.(19) Higgins, J. B.; Lapierre, R. B.; Schlenker, J. L.; Rohrman, A. C.;

Wood, J. D.; Kerr, G. T.; Rohrbaugh, W. J. Zeolites 1988, 8, 446.(20) Rajic, N.; Logar, N. Z.; Kaucic, V. Zeolites 1995, 15, 672.(21) Rouse, R. C.; Peacor, D. R. Am. Mineral. 1986, 71, 1494.(22) Cheetham, A. K.; Fjellvag, H.; Gier, T. E.; Kongshaug, K. O.;

Lillerud, K. P.; Stucky, G. D. Very open microporous materials: fromconcept to reality. In. Stud. Surf. Sci. Catal. 2001, 135, 158.

(23) Gier, T. E.; Bu, X. H.; Feng, P. Y.; Stucky, G. D. Nature 1998,395, 154.

(24) Dryzun, C.; Mastai, Y.; Shvalb, A.; Avnir, D. J. Mater. Chem. 2009,19, 2062.

(25) Baburin, I. A.; Leoni, S.; Seifert, G. J. Phys. Chem. B 2008, 112,9437.

(26) Li, Y.; Yu, J. H.; Wang, Z. P.; Zhang, J. N.; Guo, M.; Xu, R. R.Chem. Mater. 2005, 17, 4399.

(27) Tang, L. Q.; Shi, L.; Bonneau, C.; Sun, J. L.; Yue, H. J.; Ojuva,A.; Lee, B. L.; Kritikos, M.; Bell, R. G.; Bacsik, Z.; Mink, J.; Zou, X. D.Nat. Mater. 2008, 7, 381.

(28) Sun, J. L.; Bonneau, C.; Cantin, A.; Corma, A.; Diaz-Cabanas, M. J.;Moliner, M.; Zhang, D. L.; Li, M. R.; Zou, X. D. Nature 2009, 458, 1154.

(29) Sastre, G.; Corma, A. J. Phys. Chem. C , 114, 1667.(30) Xia, Q. H.; Shen, S. C.; Song, J.; Kawi, S.; Hidajat, K. J. Catal.

2003, 219, 74.(31) Noorduin, W. L.; van der Asdonk, P.; Meekes, H.; van Enckevort,

W. J. P.; Kaptein, B.; Leeman, M.; Kellogg, R. M.; Vlieg, E. Angew. Chem.,Int. Ed. 2009, 48, 3278.

(32) Noorduin, W. L.; Vlieg, E.; Kellogg, R. M.; Kaptein, B. Angew.Chem., Int. Ed. 2009, 48, 9600.

(33) Avery, K. A.; Mann, R.; Norton, M.; Willock, D. J. Top. Catal.2003, 25, 89.

(34) Sivasubramanian, K.; Kaanumalle, L. S.; Uppili, S.; Ramamurthy,V. Org. Biomolecular Chem. 2007, 5, 1569.

(35) Willock, D. J.; Bethell, D.; Feast, S.; Hutchings, G. J.; King, F.;Page, P. C. B. Top. Catal. 1996, 3, 77.

(36) Jirapongphan, S. S.; Warzywoda, J.; Budil, D. E.; Sacco, A.Chirality 2007, 19, 508.

(37) Bao, X. Y.; Broadbelt, L. J.; Snurr, R. Q. Mol. Simul. 2009, 35,50.

(38) van Erp, T. S.; Caremans, T. P.; Dubbeldam, D.; Martin-Calvo,A.; Calero, S.; Martens, J. A. Angew. Chem., Int. Ed. 2010, 49, 3010.

(39) van Erp, T. S.; Dubbeldam, D.; Caremans, T. P.; Calero, S.;Martens, J. A. J. Phys. Chem. Lett. 2010, 1, 2154.

(40) Caremans, T. P.; van Erp, T. S.; Dubbeldam, D.; Martens, J. A.;Castillo, J. M.; Calero, S. Chem. Mater. 2010, 22, 4591.

(41) Clark, L. A.; Chempath, S.; Snurr, R. Q. Langmuir 2005, 21, 2267.(42) Mann, J. B.; Meek, T. L.; Allen, L. C. J. Am. Chem. Soc. 2000,

122, 2780.(43) Kiselev, A. V.; Lopatkin, A. A.; Shulga, A. A. Zeolites 1985, 5,

261.(44) Bezus, A. G.; Kiselev, A. V.; Lopatkin, A. A.; Du, P. Q. J. Chem.

Soc., Faraday Trans. 2 1978, 74, 367.(45) Sundarababu, G.; Leibovitch, M.; Corbin, D. R.; Scheffer, J. R.;

Ramamurthy, V. Chem. Commun. 1996, 2159.

JP1079394

Zeolite Enantiomeric Separation J. Phys. Chem. C, Vol. 114, No. 50, 2010 22213