Simulated Adsorption Properties and Synthesis Prospects of

Langmuir 2005, 21, 2267-2272

2267

Simulated Adsorption Properties and Synthesis Prospects of Homochiral Porous Solids Based on Their Heterochiral Analogs Louis A. Clark,‡ Shaji Chempath,† and Randall Q. Snurr* Department of Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208 Received September 12, 2004 Molecular simulations of chiral molecules in porous heterochiral materials were performed to investigate fundamental adsorption properties and possibilities for production of homochiral porous solids. Zeolite BEA polymorph A and zeotype UCSB-7K each provide separated pores of opposite chirality. Single enantiomer and racemic mixture adsorption results are presented and indicate that significant equilibrium enantiomeric excesses of 40-70% in UCSB-7K and 10% in BEA can be achieved. Larger, better-fitting molecules display higher enantiomeric excesses. For dimethylallene, which moves on molecular dynamics time scales in UCSB-7K, self-diffusivities vary by almost an order of magnitude between the two oppositehanded UCSB-7K pores for a given enantiomer. The predicted properties indicate that equilibrium and nonequilibrium strategies using related homochiral materials for separations may be successful. To this end, a discussion of strategies for selectively blocking pores of one chirality on the basis of enantiomer segregation is provided.

1. Introduction The availability of an inexpensive and robust porous chiral solid could open new possibilities for development of enantioselective adsorption separations and heterogeneous catalysis.1-5 Current methods for preparation of heterogeneous chiral catalysts are restricted to grafting expensive and sometimes leaching-prone organometallics to achiral solids.6,7 A solid with an inherently chiral pore network may have the potential to selectively adsorb chiral molecules, provide a chiral reaction environment, and even influence reaction selectivity during diffusion out of the pores. Progress toward using naturally chiral metallic surfaces has been reported.8,9 Here, we focus on porous materials. Recently, a few examples of chiral organic10 and metalorganic11-13 porous solids have appeared in the literature. One of the metal-organic solids achieved ca. 10% ee in a chiral separation when tested as a column packing.12 Considerable success has also been achieved by imprinting polymerizable materials. Molecular imprinting of organic polymers is done during polymerization of monomers in * Author to whom correspondence should be addressed. E-mail: [email protected]. ‡ [email protected]. † [email protected]. (1) Davis, M. E. Microporous and Mesoporous Mater. 1998, 21, 173. (2) Bein, T. Curr. Opin. Solid State Mater. Sci. 1999, 4, 85. (3) Hutchings, G. J. Chem. Commun. 1999, pp 301-306. (4) Thomas, J. M.; Maschmeyer, T.; Johnson, B. F. G.; Shephard, D. S. J. Mol. Catal. A 1999, 141, 139. (5) Brunet, E. Chirality 2002, 14, 135. (6) Bianchini, C.; Barbaro, P. Top. Catal. 2002, 19, 17. (7) Song, C. E.; Lee, S. G. Chem. Rev. 2002, 102, 3495. (8) Sholl, D. S.; Asthagiri, A.; Power, T. D. J. Phys. Chem. B 2001, 105, 4771. (9) Hazen, R. M.; Sholl, D. S. Nat. Mater. 2003, 2, 367. (10) Krische, M. J.; Lehn, J. M.; Cheung, E.; Vaughn, G.; Krische, A. L. C. R. Acad. Sci. Ser. IIc: Chim. 1999, 2, 549. (11) Kepert, C. J.; Prior, T. J.; Rosseinsky, M. J. J. Am. Chem. Soc. 2000, 122, 5158. (12) Evans, O. R.; Ngo, H. L.; Lin, W. B. J. Am. Chem. Soc. 2001, 123, 10395. (13) Cui, Y.; Ngo, H. L.; White, P. S.; Lin, W. B. Inorg. Chem. 2003, 42, 652.

the presence of chiral templates and can produce chiral porous solids.14 These materials are beginning to be used as catalysts15 and in practical enantiomeric separations applications.16-18 Imprinting of more thermally stable materials such as silica has also been achieved but is still in development.19 As an alternative to direct synthesis of a homochiral solid, we examine here possibilities for the use of heterochiral solids, that is, materials having separated pores of opposite chirality within a given sample. Adsorption into such solids should show that individual chiral pores do indeed selectively adsorb chiral molecules. In contrast to the homochiral solids and imprinted polymers mentioned above, more thermostable zeolite,20,21 zeotype,22-25 and ordered, but noncrystalline26 heterochiral materials are available. The pores of these materials are often racemic (equimolar) mixtures of two nonintersecting opposite-handed networks and can often be described as helical. Often, the crystals are members of the Ia3 h d space group. We will examine adsorption into two materials, zeotype UCSB-7K and zeolite beta (BEA). The two materials differ in the length scale on which the pores form a racemic mixture and in the pore size. Polymorph A of BEA has a helical structure (Figure 1A) with elliptical (7.3 Å × 6.1 (14) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812. (15) Wulff, G. Chem. Rev. 2002, 102, 1. (16) Sellergren, B. TrAC, Trends Anal. Chem. 1997, 16, 310. (17) Andersson, L. I. J. Chromatogr. B 2000, 739, 163. (18) Maier, N. M.; Franco, P.; Lindner, W. J. Chromatogr. A 2001, 906, 3. (19) Katz, A.; Davis, M. E. Nature 2000, 403, 286. (20) Treacy, M. M. J.; Newsam, J. M. Nature 1988, 332, 249. (21) Newsam, J. M.; Treacy, M. M. J.; Koetsier, W. T.; de Gruyter, C. B. Proc. R. Soc. London A 1988, 420, 375. (22) Anderson, M. W.; Terasaki, O.; Ohsuna, T.; Philippou, A.; Mackay, S. P.; Ferreira, A.; Rocha, J.; Lidin, S. Nature 1994, 367, 347. (23) Gier, T. E.; Bu, W.; Feng, P.; Stucky, G. D. Nature 1998, 395, 154. (24) Bu, X.; Feng, P.; Gier, T. E.; Zhao, C.; Stucky, G. D. J. Am. Chem. Soc. 1998, 120, 13389. (25) Lin, Z. E.; Yao, Y. W.; Zhang, J.; Yang, G. Y. J. Chem. Soc.,Dalton Trans. 2002, 4527, 4527. (26) Anderson, M. W. Zeolites 1997, 19, 220.

10.1021/la047722e CCC: $30.25 © 2005 American Chemical Society Published on Web 12/16/2004

2268

Langmuir, Vol. 21, No. 6, 2005

Figure 1. Zeotype materials used in the simulations shown with framework as sticks and pore structure as space-filling spheres. (A) One unit cell of BEA polymorph A (right-handed pore), (B) Two unit cells of UCSB-7K in the z direction without adsorbed cations and water. The two non-interconnecting pore types in UCSB-7K are distinguished by the yellow and blue coloring.

Å) pores running in three dimensions.20,21,27 The singlechirality domain size is unknown but must be greater than a single unit cell because only one pore chirality type is present per unit cell. Attempts to synthesize enantiomerically enriched BEA polymorph A with chiral structuredirecting agents have been largely unsuccessful.28 In contrast, the UCSB-7K zeotype has a racemic mixture of pores in each unit cell with elliptical (4.0 Å × 5.6 Å) pores and cavities in the 9-10 Å range.23,24,27 Figure 1B shows the intertwined, yet noninterconnecting, nature of the UCSB-7K pore network. The UCSB-7K material is employed in the simulations without the stabilizing cations and water molecules. The independent nature of the two opposite-handed pore structures suggests a possible strategy for production of a homochiral solid. It may be possible to selectively block one pore type such that the remaining unblocked portion is chiral. From this viewpoint, the precursor material can be termed ‘prochiral’ since an operation on it could result in a chiral structure. Suggestions for blocking one pore type will be given in the discussion. 2. Methods All adsorption simulations were done with the Music code29 using grand canonical Monte Carlo (GCMC). For rigid molecules, energy biasing and combined cavity and orientational biasing were used to speed the convergence.30 Simulation of flexible molecules required development of two types of hybrid Monte Carlo methods that have been detailed elsewhere.31 The molecular dynamics (MD) simulations were carried out in the NVT ensemble

(27) Pore sizes are calculated by finding oxygen-oxygen distances that define the pore and subtracting 2.7 Å to account for the oxygen diameter. For the UCSB-7K structure, it was necessary to complement this information with pore diameter ratio from contour maps because oxygen pairs defining the pore diameter were difficult to find. (28) Tsuji, K.; Beck, L. W.; Davis, M. E. Microporous Mesoporous Mater. 1999, 28, 519. (29) Gupta, A.; Chempath, S.; Sanborn, M. J.; Clark, L. A.; Snurr, R. Q. Mol. Sim. 2003, 29, 29. Code available at: http://zeolites.cqe. nwu.edu/Music/music.html. (30) Snurr, R. Q.; Bell, A. T.; Theodorou, D. N. J. Phys. Chem. 1993, 97, 13742. (31) Chempath, S.; Clark, L. A.; Snurr, R. Q. J. Chem. Phys. 2003, 118, 7635.

Clark et al.

Figure 2. Enantiomeric molecular structures used in the simulations. (A) 1,3-Dimethylallene, (B) 1,2-dimethylcyclopropane, and (C) 1,2-dimethylcyclobutane. Color coding: dark gray, CH2 or H; white, CH or C; light gray, CH3. using a Nose´-Hoover thermostat and a rigid molecule constraint algorithm due to Ciccotti et al.32 As is common when modeling adsorbate-zeolite interactions, we assumed that the zeolite could be treated as rigid and that only the lattice oxygen atoms interact with the sorbates. This is a reasonable assumption because they protrude further into the pores than the metal atoms and therefore provide a screening effect. Additionally, since the sorbates lack strongly charged groups, we assumed that all interactions with the adsorbate could be modeled solely with Lennard-Jones interactions. To further simplify the calculations for the molecules in UCSB-7K, we chose to use identical and united atom representations of the -CH, -CH2, and -CH3 groups. The molecules are shown in Figure 2. The crystal structures for UCSB-7K24 and BEA polymorph A21 were taken from the literature. In some cases, as indicated in the presentation of the results, the assumption of molecular rigidity was relaxed and intramolecular interactions were included. Intramolecular potentials for all but the diphenyldimethylallene are taken from the work of Sholl.33 For the diphenyldimethylallene molecule, the intramolecular potentials are identical to those published in our previous work.31,34 All force field parameters are given in the Supporting Information. The adsorption energies, and thus the pressure-dependent onset of adsorption, are not meant to be quantitative predictions because the force fields used in this work have not been calibrated with experimental results specific to the systems studied. However, the force field is similar to that used for other zeoliteadsorbate systems, and the relative results between enantiomers can be considered sufficiently predictive.

3. Results and Discussion To understand chiral molecule adsorption and to make assessments for process and materials design, it is desirable to study adsorption under two different conditions. Adsorption resulting from exposure of the heterochiral porous solid to a single enantiomer of the adsorbate is the least complicated case and may be important in materials preparation schemes. From a practical point of view, adsorption from a racemic mixture is interesting (32) Ciccotti, G.; Ferrario, M.; Ryckaert, J. P. Mol. Phys. 1982, 47, 1253. (33) Sholl, D. S. Langmuir 1998, 14, 862. (34) Chempath, S. Molecular Simulation of Multicomponent Adsorption and Diffusion in Zeolites. Ph.D. Thesis, Northwestern University, Evanston, IL, 2004.

Adsorption Properties of Homochiral Porous Solids

Langmuir, Vol. 21, No. 6, 2005 2269

Figure 4. Enantiomeric excesses calculated from adsorption of racemic mixtures of dimethylcyclobutane in UCSB-7K at 200, 300, and 400 K. Error bars are the standard deviation calculated from the four equivalent definitions in eq 6 in the Supporting Information. Lines are linear fits to the data.

Figure 3. Adsorption data for single enantiomer RH dimethylcyclobutane in UCSB-7K at 200 and 300 K. A rigid model for the adsorbate was used unless otherwise specified. (A) Isotherms. At 200 K, the isotherm in the RH channel is compared to results using a flexible sorbate model and shown to be similar. At 300 K, the loading of the LH dimethylcyclobutane in the RH channel is shown to be identical to the RH dimethyl cyclobutane loading in the LH channel. (B) Siting enantiomeric excess (ee′) calculated from single-enantiomer adsorption into a pair of enantiomeric pores. See definition of ee′ in the Supporting Information.

because real separation processes are most likely to encounter unenriched feeds. Results are given here for both single enantiomer and racemic mixtures. 3.1. Adsorption in UCSB-7K. The choice of test molecules for adsorption in UCSB-7K was a subset of those chosen by Sholl33 to study chiral adsorption on metal surfaces. No attempt was made to optimize the molecular identities to increase the selectivity. The united-atom representations of the test molecules employed in the simulations are shown in Figure 2. Enantiomers are distinguished by their left-handed (LH or L) or righthanded nature (RH or R). Comparison of 1,2-dimethylcyclopropane and 1,2-dimethylcyclobutane adsorption allows investigation of bulkiness effects. Simulation of 1,3-dimethylallene enantiomers permits comparison of shape effects and enables molecular dynamics simulations, since these molecules diffuse on an accessible time scale in UCSB-7K. Adsorption isotherms at two temperatures for the RH dimethylcyclobutane molecule in UCSB-7K are shown in Figure 3A. The total isotherms are broken into contributions from adsorption in each pore type and clearly display enantioselectivity. From these site-specific isotherms, it can be seen that the right-handed molecule prefers to be in the right-handed pore type. It can also be seen that the LH enantiomer in the right channel data at 300 K is nearly identical to that of the RH enantiomer in the left channel, verifying that the required adsorption symmetry is present (see the Supporting Information). Siting specificity can be substantial (50-70 ee′) at low pressures and loadings, as shown in Figure 3B, but decays to near zero as the

loading increases toward saturation. This type of decay is typical in adsorption systems and is due to initial preferential filling of the most favorable sites.35 At higher temperatures where entropic effects are more important, a decrease in the maximum siting enantiomeric excess is seen. This selectivity damping effect indicates that the selectivity mechanism is dominated by potential energy contributions such as geometric complementarity interactions. Simulations using a flexible model of dimethylcyclobutane revealed little difference in loading or siting enantiomeric excess between rigid and flexible forms of the molecule (Figure 3). Adsorption of a racemic mixture was tested at 200, 300, and 400 K. Figure 4 shows the results in terms of enantiomeric excess. Note that ee and ee′ are identical for this case, as described in the Supporting Information. The consistency check discussed in the Supporting Information was used to produce error estimates for the points. The enantiomeric excesses remain fairly constant throughout the pressure range. Adsorption of each enantiomer into its preferred sites partially blocks adsorption of the other enantiomer in its nonpreferred sites and keeps the enantiomeric excess high even at higher loadings. This is an example of a so-called reconciliation-type adsorption process;35 one which typically produces comparatively large segregation. One might hypothesize that this behavior could be due to creation of new chiral adsorption sites between adsorbate and pore wall. However, the molecules adsorb singly in the intersections and are therefore too isolated from each other to produce new adsorption sites. Adsorption of the flexible molecule was tested at 300 K and shows that the enantiomeric excess deviates by less than 10% from the rigid molecule case. Table 1 compares the difference in enantiomer potential energies in the two pore types during a racemic mixture simulation at low loading. Results are presented for three different molecules. There is a clear and logical correlation between higher energy differences and enantiomeric excesses. The results indicate that an enantiomeric excess of ca. 50% can be achieved even if there is only a 4-5 kJ/mol difference in adsorption energies. In the work of Sholl and colleagues, an energy difference of ca. 3 kJ/mol for 1,2-dimethylcyclobutane on Pt(643) was found to give (35) Clark, L. A.; Gupta, A.; Snurr, R. Q. J. Phys. Chem. B 1998, 102, 6720.

2270

Langmuir, Vol. 21, No. 6, 2005

Clark et al.

Table 1. Results of Low-Loading Racemic Mixture Adsorption Simulations enantiomeric Pair

∆Ua (kJ/mol)

eeb (%)

1,2-dimethylcyclopropane (300 K) 1,2-dimethylcyclobutane (300 K) 1,3-dimethylallene (200 K) 1,3-dimethylallene (300 K) 1,3-dimethylallene (400 K)

0.58 2.5-3.0 0.93 0.84 0.84

11 50 22 13 9

a Difference between mean potential energies in the two helices for a single enantiomer. Here, results are for rigid molecules. b Enantiomeric excess, ee or ee′, as defined in the Supporting Information.

Figure 5. The two enantiomers of dimethyldiphenylallene.

significant enantiomeric excesses33,36 The dimethylallene molecule also shows chiral adsorption selectivity despite its smaller size. The enantiomeric excess for the dimethylallene molecule increases from 9% to 22% as the temperature is lowered from 400 to 200 K. Visualization of the siting for the various adsorbates indicates that the ring molecules sit in the intersections and the dimethylallenes sit in the middle of the pores. Since adsorption is enantiomerically selective, both sites can be said to have some chirality. For dimethylcyclobutane, simulations performed at higher loadings showed that the potential energy difference between the two pore types was essentially constant with changing loading. 3.2. Adsorption in BEA Polymorph A. The sizes of the pore chirality and the molecular shape must be commensurate in order to see chiral adsorption selectivity. The molecules that show such selectivity in UCSB-7K are too small or of the wrong shape to show it in BEA polymorph A. Some carbon nanotubes are also chiral but have been predicted not to adsorb enantiomers with chiral selectivity.37 Presumably, this is due to the lack of chiral definition on the relatively flat pore walls. To match the length scale and shape of the helical BEA polymorph A pores, a bulkier substituted allene molecule was chosen. The two enantiomers of dimethyldiphenylallene shown in Figure 5 were simulated in the zeolite structure singly and as a racemic mixture. Isotherms are presented in Figure 6. One line shows the results for LH molecules in the LH pores and RH molecules in RH pores. The loading is 64-72% higher than for RH molecules in LH pores and LH molecules in RH pores. Since convergence is more difficult to achieve with the larger molecules, the simulations were run at a higher temperature and in the linear region of the isotherms. The expected symmetry (see eq 5 in the Supporting Information) was used to provide averages and standard deviation error bars. It is clear that the chirality of the molecule/pore pairs produces a difference in loading. Siting enantiomeric excesses cannot be calculated from this single-component data because, in contrast to UCSB-7K, each BEA polymorph A structure only posseses a single pore type. (36) Power, T. D.; Sholl, D. S. J. Vac. Sci. Technol. A 1999, 17, 1700. (37) Power, T. D.; Skoulidas, A. I.; Sholl, D. S. J. Am. Chem. Soc. 2002, 124, 1858.

Figure 6. Isotherm adsorption data for single-enantiomer rigid dimethyldiphenylallene in BEA polymorph A at 425 K. Error bars are calculated using adsorption symmetry (see Supporting Information). Higher adsorbing ‘like’ molecule-pore pairs are LH molecule, LH pore and RH molecule, RH pore; ‘unlike’ pairs lead to lower adsorption.

Figure 7. Atomic-siting information for rigid dimethyldiphenylallene enantiomers in LH BEA polymorph A at 425 K and a fugacity of 1.0 × 104 kPa. Data from two different single enantiomer runs were combined to produce the figure. The centers of mass are represented by yellow and blue densities (spheres) corresponding to the LH and RH enantiomers, respectively, and are located only at the intersections. The white and light blue densities show the position of the LH and RH enantiomer phenyl groups.

For comparison with Table 1, energy differences for the enantiomers in their favored and unfavored BEA structures were calculated. The differences range from 2.1 to 3.1 kJ/mol, similar to that seen for dimethylcyclobutane in the two pore types of UCSB-7K. Visualization of the average adsorbate positions during the adsorption simulations indicates that the molecules orient to take advantage of the pore twist around the channel intersections. Figure 7 shows the preferred locations of the center of mass and the phenyl groups for the two enantiomers. In both cases, the center of mass (allene portion) is localized only in the intersections and the bulky phenyl ends hang into the neighboring pores. However, it can be seen that the two enantiomers have distinctly different adsorption orientations. The probability densities representing the likely center of mass positions are similar but rotated for the two different enantiomers (dark blue and yellow, in Figure 7). More evidence for a different tilt for each enantiomer is seen in the positions of the phenyl groups, which are also different (light blue and white, in Figure 7).

Adsorption Properties of Homochiral Porous Solids

To allow calculation of enantiomeric excesses, racemic mixture simulations at 425 K were performed. Convergence was problematic for these runs, and reliable results were only obtainable below a total loading of 1.0 molecules/ unit cell. Shape complementarity is also seen in these runs; the LH molecule adsorbs most strongly in the LH pore structure. In contrast to the UCSB-7K system, the enantiomeric excess is lower, at 10-14%, but still relatively constant across the range of studied loadings. It is interesting to note that the enantiomeric excess is significantly smaller in this case, despite similar differences in the average potential energies for the two enantiomers. Part of this difference is due to the higher temperature used in the BEA simulation, but this is not the only factor because at 400 K enantiomeric excesses of 40% are seen in UCSB-7K. The difference in enantiomeric excesses must thus be attributed to diminished moleculepore shape complementarity. 3.3. Potential for Homochiral Solid Production and Usage. A homochiral porous material could be produced from a heterochiral precursor if one of the racemic pore types were blocked selectively. In the UCSB7K material, one would fill only a pore of a single handedness, and for BEA, one would fill only one enantiomer of polymorph A. The potential of this idea depends on the reliability of the process used to fix blockers and the siting enantiomeric excess of the blocker under those conditions. The simplest concept for selective blocking is adsorption of a single enantiomer followed by fixation. Fixation of the adsorbate could be accomplished by flash thermal degradation to form coke, photoinduced polymerization, or any other method that does not unfavorably change the siting enantiomeric excess. Under this challenging scheme, molecules with both high chiral adsorption selectivities and well-placed polymerizable end-groups are necessary. Olefin functional groups might be useful because they can oligomerize easily under acidic conditions. If one pore type can be blocked, then the resulting material could be used for adsorptive separations. An estimate for the attainable enrichment can be had by artifically blocking one pore type of the UCSB-7K structure or considering only one enantiomer of BEA polymorph A during adsorption simulations. The enrichment can be quantified by considering the ratio of adsorbed-phase mole fraction to the fluid-phase mole fraction. Representative results are shown in Figure 8. For 1,2-dimethylcyclobutane in UCSB-7K with the right-handed channel blocked at 200 K, enrichments can reach 350% or more. At a higher temperature, a similar test in BEA gives enrichments for dimethyldiphenylallene as high as 160% depending on the fluid-phase mole fraction. Experimental verification of this concept could be obtained by adsorbing a single enantiomer, heat shocking the system to quickly coke adsorbates, and then exposing to a racemic mixture. Measurement of the chiral enrichment in the fluid phase would indicate which enantiomer adsorbs best and to what degree. It may also be possible to take advantage of diffusive transport differences between enantiomers or between a single enantiomer in two different pore environments. Nonequilbrium preparation strategies can be envisioned, and once a homochiral porous solid is produced, diffusivity differences could be exploited in kinetic separation processes such as pressure-swing adsorption. To provide some support for the assertion that diffusivity differences can work to increase the enantiomeric excess, we performed molecular dynamics simulation for enantiomers

Langmuir, Vol. 21, No. 6, 2005 2271

Figure 8. Equilibrium adsorbed-phase mole fraction versus bulk mole fraction (XY) plots for 1,2-dimethylcyclobutane in UCSB-7K with one pore blocked at 200 K and dimethyldiphenylallene in BEA polymorph A at 425 K. Larger deviation from the diagonal indicates greater potential for use in separation processes.

of dimethylallene in UCSB-7K at infinite dilution. Each enantiomer diffuses almost an order of magnitude faster in its nonpreferred pore than in its thermodynamically preferred pore: 1.0 × 10-8 m2/s versus 1.8 × 10-9 m2/s at 300 K. This is a significant difference. Certain properties of the UCSB-7K and BEA materials limit their usefulness as prochiral solids. UCSB-7K is unstable after desorption of the pore-filling water and cations.23,24 Similarly, in a typical BEA sample, not all of the material is the heterochiral polymorph A. The use of both materials with larger chiral molecules may also be limited by their relatively small pore sizes. Fortunately, other materials with heterochiral pore structures are available on more-mesoscopic length scales. The dual racemic noninterconnecting pore morphology is present in many microporous or phase-separated systems. It is called the “double gyroid” morphology.38 Microporous bicontinuous materials often have such morphology.39,40 At least one of the porous amorphous silicas formed from micellular precursors, MCM-48, has a similar structure.26 Other systems include diblock copolymers,41 triblock copolymers,42 and three-component polymer systems.43 These systems can be manipulated. A triblock copolymer with a silica-containing block formed a double gyroid structures and was used to make a porous silica network.44 It is even possible to generate helical tubules from a micellular solution by manipulating flow conditions.45 4. Conclusions We have provided evidence from molecular simulations that adsorption of chiral molecules into microporous chiral materials can occur with significant selectivity. From the examples in zeotype UCSB-7K presented here, it can be expected that 40-70% enantiomeric excess could be (38) Gandy, P. J. F.; Klinowski, J. Chem. Phys. Lett. 2000, 321, 363. (39) Scriven, L. E. Nature 1976, 263, 123. (40) Burban, J. H.; He, M.; Cussler, E. L. AIChE J. 1995, 41, 159. (41) Forster, S.; Khandpur, A. K.; Zhao, J.; Bates, F. S.; Hamley, I. W.; Ryan, A. J.; Bras, W. Macromolecules 1994, 27, 6922. (42) Avgeropoulos, A.; Dair, B. J.; Hadjichristidis, N.; Thomas, E. L. Macromolecules 1997, 30, 5634. (43) Suzuki, J. J. Chem. Phys. 2000, 112, 4862. (44) Chan, V. Z. H.; Hoffman, J.; Lee, V. Y.; Iatrou, H.; Avgeropoulos, A.; Hadjichristidis, N.; Miller, R. D.; Thomas, E. L. Science 1999, 286, 1716. (45) Yang, S. M.; Kim, W. J. Adv. Mater. 2001, 13, 1191.

2272

Langmuir, Vol. 21, No. 6, 2005

achieved for well-fitting molecules. It is also apparent that selective chiral adsorption is not a given; chiral selectivity in microporous solids is sensitive to shape and size matching between the sorbate and sorbent. Materials that are very selective for one set of enantiomers might not show any selectivity toward another pair of molecules of slightly different size and shape. Because of this, the enantiomeric excess for dimethyldiphenylallene molecules in BEA polymorph A is comparatively smaller than the results seen in UCSB-7K. The results presented here indicate that it may be possible to use selective chiral adsorption to block micropores of a specific chirality and produce a chiral porous material. Success will depend on optimizing the fit of the chiral blocker and the fixation method. If these types of partially blocked porous solids can be produced, they will be able to provide chiral reaction environments and have significant separative capabilities. The large differences in diffusivities indicate that nonequilibrium uses of the materials may also be possible. This work suggests other promising avenues of theoretical study for chiral adsorption. The helical pore shape common to both materials studied here is a common chiral

Clark et al.

motif in nature. Helix structures can even arise macroscopically from synthesis.46 However, other motifs may be possible and perhaps more effective since the helical shape of the adsorbing molecule would then be less critical. Selective placement of functional groups or metal centers could also be pursued to enhance enantiomeric adsorption selectivity. Acknowledgment. We gratefully acknowledge the help of X. Bu (UCSB) for his help with the UCSB-7K unit cell coordinates. Funding from the National Science Foundation GOALI program is acknowledged (CTS0302428). L.A.C. appreciates partial support of this work through a fellowship from the Alexander von Humboldt Foundation. We thank David Sholl (Carnegie Mellon) for encouragement and helpful discussions. Supporting Information Available: Calculation of enantiomeric excesses and force field information. This material is available free of charge via the Internet at http://pubs.acs.org. LA047722E (46) Giraldo, O.; Brock, S. L.; Marquez, M.; Suib, S. L.; Hillhouse, H.; Tsapatsis, M. Nature 2000, 405, 38.