Monte Carlo Simulation for the Adsorption and Separation of Linear

May 16, 2006 - have been investigated using Monte Carlo simulation. For pure linear .... (18) Garberoglio, G.; Skoulidas, A. I.; Johnson, J. K. J. Phy...
0 downloads 0 Views 331KB Size
Download PDF
5702

Langmuir 2006, 22, 5702-5707

Monte Carlo Simulation for the Adsorption and Separation of Linear and Branched Alkanes in IRMOF-1 Jianwen Jiang*,† and Stanley I. Sandler‡ Department of Chemical & Biomolecular Engineering, National UniVersity of Singapore, 4 Engineering DriVe 4, Singapore 117576, and Department of Chemical Engineering, UniVersity of Delaware, Newark, Delaware, 19716 ReceiVed February 22, 2006. In Final Form: April 14, 2006 The adsorption and separation of linear and branched alkanes in the isoreticular metal-organic framework IRMOF-1 have been investigated using Monte Carlo simulation. For pure linear alkanes (C1-nC5), the limiting adsorption properties exhibit linear behavior with the alkane carbon number; the long alkane is preferentially adsorbed over the short alkane at low fugacities, whereas the reverse is found at high fugacities. For pure branched alkanes (C5 isomers), the linear isomer adsorbs more than its branched analogue. The adsorbed amounts of pure alkanes in IRMOF-1 are substantially greater than in a carbon nanotube bundle and in silicalite. For a five-component mixture of C1 to nC5 linear alkanes, the long alkane adsorption first increases and then decreases with increasing fugacity, whereas short alkane adsorption continually increases and progressively replaces the long alkane at high fugacity due to the size entropy effect. For a three-component mixture of C5 isomers, the adsorption of each isomer increases with increasing fugacity until saturation, though there is less adsorption of the branched isomer due to the configurational entropy effect. The adsorption selectivity among the alkanes in IRMOF-1 is smaller than in a carbon nanotube bundle and in silicalite.

I. Introduction

* Corresponding author. E-mail: [email protected]. Tel: (65) 6516 5083. Fax: (65) 6779 1936. † National University of Singapore. ‡ University of Delaware.

CO2 over N2, CH4, and Ar and suggested that this MOF may act as a highly selective molecular sieve to remove CO2 from natural gas, and to recover H2 from gas mixtures. Uchida and Mizuno10 demonstrated that a zeotype ionic crystal of Cs5[Cr3O(OOCH)6(H2O)3][R-CoW12O40]‚7.5H2O exhibits shape-selective adsorption of H2O, which could be used for the removal of water from the H2O/C2H5OH azeotropic mixtures. Takamizawa et al.11 used a transformable 1D host [Rh2(O2CPh)4(pyz)]n for the inclusion of CO2 inside the crystal, which might lead to new techniques for CO2 storage. Pan et al.12,13 found the high sorption capacities of large hydrocarbons (nC6 and cyclo-C6) in a thermally stable MOF and synthesized a microporous MOF that is capable of separating nC4 from higher linear alkanes and olefins. Alternatively, molecular simulations have been performed to investigate fluid behavior in MOFs. Vishnyakov et al.14 simulated the adsorption isotherms of Ar in Cu-BTC and identified the adsorption sites and the sequence of pore filling. Sarkisov et al.15 examined the adsorption and diffusion of C1, n-alkanes, cycloC6, and benzene in IRMOF-1 and bipyridine molecular squares and provided insight into the microscopic interactions in these MOFs. Du¨ren et al.16 explored the adsorption characteristics of C1 in IRMOFs and molecular squares and proposed new IRMOF structures with even higher adsorption capacities. Du¨ren and Snurr17 assessed the adsorption of pure and mixed C1 and nC4 in five different IRMOFs and suggested that IRMOFs are promising materials for the separation of hydrocarbons. Gar-

(1) Yaghi, O. M.; O’Keeffe, M.; Ockwig, N. W.; Chae, H. K.; Eddaoudi, M.; Kim, J.Nature 2003, 423, 705. (2) Snurr, R. Q.; Hupp, J. T.; Nguyen, S. T. AIChE J. 2004, 50, 1090. (3) Rowsell, J. L. C.; Yaghi, O. M. Microporous Mesoporous Mater. 2004, 73, 3. (4) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’Keefe, M.; Yaghi, O. M. Science 2002, 295, 469. (5) Rosi, N. L.; Eckert, J.; Eddaoudi, M.; Vodak, D. T.; Kim, J.; O’Keeffe, M.; Yaghi, O. M. Science 2003, 300, 1127. (6) Li, H.; Eddaoudi, M.; O’Keeffe, M.; Yaghi, O. M. Nature 1999, 402, 276. (7) Li, Y. W.; Yang, R. T. J. Am. Chem. Soc. 2006, 128, 726. (8) Kitaura, R.; Seki, K.; Akiyama, G.; Kitagawa, S. Angew. Chem., Int. Ed. 2003, 42, 428. (9) Dybtsev, D. N.; Chun, H.; Yoon, S. H.; Kim, D.; Kim, K. J. Am. Chem. Soc. 2004, 126, 32.

(10) Uchida, S.; Mizuno, N. J. Am. Chem. Soc. 2004, 126, 1602. (11) Takamizawa, S.; Nakata, E.; Yokoyama, H.; Mochizuki, K.; Mori, W. Angew. Chem., Int. Ed. 2003, 42, 4331. (12) Pan, L.; Liu, H. M.; Lei, X. G.; Huang, X. Y.; Olson, D. H.; Turro, N. J.; Li, J. Angew. Chem., Int. Ed. 2003, 42, 542. (13) Pan, L.; Liu, H. M.; Lei, X. G.; Huang, X. Y.; Olson, D. H.; Turro, N. J.; Li, J. Angew. Chem., Int. Ed. 2006, 45, 616. (14) Vishnyakov, A.; Ravikovitch, P. I.; Neimark, A. V.; Bulow, M.; Wang, Q. M. Nano Lett. 2003, 3, 713. (15) Sarkisov, L.; Duren, T.; Snurr, R. Q. Mol. Phys. 2004, 102, 211. (16) Duren, T.; Sarkisov, L.; Yaghi, O. M.; Snurr, R. Q. Langmuir 2004, 20, 2683. (17) Duren, T.; Snurr, R. Q. J. Phys. Chem. B 2004, 108, 15703.

Metal-organic frameworks (MOFs) are a new family of porous materials consisting of metal-oxide clusters and organic linkers.1 These structures have extremely high porosities (up to 90%) and well-defined pore sizes and are promising materials for the storage of gases, the separation of mixtures, and ion exchanges.2 In particular, the controllable length of the organic linker allows for tailoring the MOF functionality, pore volume, and size over a wide atomic-scale range.3 A large number of experimental studies of adsorption have been reported in a wide variety of MOFs. Yaghi and co-workers4,5 found high storage capacities of the isoreticular MOF-1 (IRMOF1)6 for CH4 and H2 and established a connection between MOFs and the needs of the fuel gas industry. Li and Yang7 showed significant enhancement of H2 storage in IRMOF-1 and IRMOF-8 via spillover and suggested that MOFs with additional rings in the organic linker could result in even higher H2 storage capacity. Kitaura et al.8 observed that CO2, CH4, O2, and N2 adsorb in Cu(dhbc)2(4,4′-bpy) at specifically tuned gate-open pressures that depend on the intermolecular interactions. Dybtsev et al.9 found that Mn(HCO2)2]‚1/3(C4H8O2) selectively adsorbs H2 and

10.1021/la060506g CCC: $33.50 © 2006 American Chemical Society Published on Web 05/16/2006

Adsorption and Separation of Alkanes in IRMOF-1

beroglio et al.18 calculated the adsorption of various light gases on a number of MOFs and observed good agreement with experiments in some cases and poor agreement in others. Skoulidas and Sholl19 studied the adsorption and diffusion of light gases in various MOFs and found that the diffusion characteristics in MOFs are analogous to that in zeolites in both magnitude and mechanism. Yang and Zhong20,21 simulated the adsorption and diffusion of H2 in IRMOF-1, IRMOF-8, IRMOF18, and MOF-505 and concluded that MOFs cannot store an appreciable amount of H2 at room temperature and moderate pressures. Many industrial processes involve alkanes and their mixtures, and how to separate and store these fluids is of practical importance in applications. Adsorption is one of the technically feasible and cost-effective technologies toward these ends. With a proper choice of porous materials and operating conditions, alkanes may be separated and stored efficiently. Currently, most experimental and simulation studies of alkane adsorption were carried out in zeolites and carbon materials. There have been scarce simulation studies of the adsorption of alkanes in MOFs. In this work, Monte Carlo (MC) simulation has been used to investigate the adsorption behavior of linear and branched alkanes in IRMOF-1, which is a prototype MOF and also known as MOF-5. First we simulate the adsorption behavior for five pure linear alkanes from C1 to nC5, and three pure C5 isomers, then examine the competitive adsorption for a mixture of C1 to nC5 linear alkanes as a function of chain length, and finally examine the competitive adsorption for a mixture of C5 isomers with a variation in the degree of chain branching. Additionally, the adsorption properties in IRMOF-1 are compared with previous studies in a single-walled carbon nanotube (SWNT) bundle by us22 and in silicalite by Smit and co-workers.23,24 The goal is to determine if IRMOF-1 is a good candidate for the storage and separation of alkanes.

II. Model and Methodology IRMOF-1 has a Fm-3m crystal space group with a lattice constant of 25.832 Å, a crystal density of 0.593 g/cm3, and a free volume of 79.2%.4 For simulation, the atomic coordinates within the crystal were constructed using experimental X-ray crystallographic data.4 Figure 1 shows a two-dimensional view of a unit cell of IRMOF-1 on the (100) plane. The central sphere represents the largest van der Waals sphere that would fit in IRMOF-1 without touching the framework. IRMOF-1 has a formula of Zn4O(BDC)3, where BDC is 1,4-benzenedicarboxylate.6 Each oxide-centered Zn4O tetrahedron is edge-bridged by six carboxylate linkers resulting in an octahedral Zn4O(O2C-)6 building unit, which reticulates into a three-dimensional cubic structure with benzene struts. The simulation box used in the current work consisted of one unit cell of IRMOF-1, and the framework was assumed to be rigid. As our study focused on low-energy equilibrium configurations, flexibility of IRMOF-1 allowing local small movements of the framework atoms would not be expected to give significantly different results. The universal force field (UFF)25 was adopted to describe the Zn, C, O, and H atoms in (18) Garberoglio, G.; Skoulidas, A. I.; Johnson, J. K. J. Phys. Chem. B 2005, 109, 13094. (19) Skoulidas, A. I.; Sholl, D. S. J. Phys. Chem. B 2005, 109, 15760. (20) Yang, Q. Y.; Zhong, C. L. J. Phys. Chem. B 2005, 109, 11862. (21) Yang, Q. Y.; Zhong, C. L. J. Phys. Chem. B 2006, 110, 655. (22) Jiang, J. W.; Sandler, S. I.; Schenk, M.; Smit, B. Phys. ReV. B 2005, 72, 045447. (23) Du, Z. M.; Manos, G.; Vlugt, T. J. H.; Smit, B. AIChE J. 1998, 44, 1756. (24) Vlugt, T. J. H.; Krishna, R.; Smit, B. J. Phys. Chem. B 1999, 103, 1102. (25) Rappe, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A., III.; Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024.

Langmuir, Vol. 22, No. 13, 2006 5703

Figure 1. Two-dimensional view of the (100) plane of a unit cell of IRMOF-1. The central sphere represents the largest van der Waals sphere that would fit in IRMOF-1 without touching the framework.

IRMOF-1. Alkanes were represented by a united-atom model with TraPPE-UA force field,26,27 as in our previous work for the adsorption of alkanes in a SWNT bundle.22 The interactions between IRMOF-1 atoms and alkane sites were modeled using Lennard-Jones (LJ) potential, in which the cross LJ parameters were estimated using the Lorentz-Berthelot combining rules.28 A spherical cutoff length of 12.9 Å was used for the calculation of the LJ interaction energies, and the usual long-range correction for a homogeneous system was added. The use of the usual long-range correction is an appropriate approximation as it has been shown that the error introduced by assuming homogeneity is small compared with the magnitude of the long-range correction.29 MC simulation in NVT ensemble was used to calculate the limiting adsorption properties (the isosteric heat, the Henry constant, the Helmholtz free energy, enthalpy, energy, and entropy) of each pure alkane at zero coverage in IRMOF-1. MC simulation in the grand canonical ensemble was used to calculate the adsorption of pure alkanes and their mixtures as a function of bulk fugacity. The configurational-bias technique30-32 was implemented in the simulations to improve the efficiency in sampling the conformations of alkane molecules. A typical simulation was performed for a total 20 000 cycles, in which the first 10 000 cycles were used for equilibration and the second 10 000 cycles for ensemble averages. Each cycle consisted of a number of the attempted trial moves, including translation, rotation, regrowth, exchange with the reservoir, and exchange of molecular identity (only for a mixture). Within the statistical uncertainty, the simulation results were found to be independent of the sequence of the trial moves. The details of the simulation method have been described in our previous study.22

III. Results and Discussion Table 1 lists the numerical values of the limiting adsorption properties, including isosteric heats q0st, Henry constants KH, and adsorption entropies ∆S°, for the adsorption of pure linear and (26) Martin, M. G.; Siepmann, J. I. J. Phys. Chem. B 1998, 102, 2569. (27) Martin, M. G.; Siepmann, J. I. J. Phys. Chem. B 1999, 103, 4508. (28) Maitland, G. C.; Rigby, M.; Smith, E. B.; Wakeham, W. A. Intermolecular Forces; Claredon Press: Oxford, 1981. (29) Siperstein, F.; Myers, A. L.; Talu, O. Mol. Phys. 2002, 100, 2025. (30) Siepmann, J. I.; Frenkel, D. Mol. Phys. 1992, 75, 59. (31) Frenkel, D.; Mooij, G. C. A. M.; Smit, B. J. Phys.: Condensed Matter 1992, 4, 3053. (32) de Pablo, J. J.; Laso, M.; Suter, U. W. J. Chem. Phys. 1992, 96, 2395.

5704 Langmuir, Vol. 22, No. 13, 2006

Jiang and Sandler

Figure 2. Limiting adsorption properties of pure alkanes (C1, C2, C3, nC4, nC5, iC5, and neoC5) in IRMOF-1 at 300 K. (a) Isosteric heats, (b) Henry constants, and (c) adsorption entropies. Dotted lines are drawn to guide the eye. Table 1. Limiting Isosteric Heats, Henry Constants, and Adsorption Entropies of Pure Alkanes in IRMOF-1 at 300 K alkane

qst0 (kJ/mol)

KH (mol/dm3 /kPa)

∆S° (J/mol//K)

C1 C2 C3 nC4 nC5 isoC5 neoC5

10.42 15.54 20.17 24.66 29.32 28.39 25.71

0.0028 0.0124 0.0458 0.172 0.654 0.537 0.311

-10.47 -15.37 -20.05 -24.12 -28.65 -27.16 -22.69

branched alkanes in IRMOF-1 at 300 K. Figure 2 shows (a) isosteric heats, (b) Henry constants, and (c) adsorption entropies, respectively, as a function of Nc, the alkane carbon number. For linear alkanes with variation in length, the longer the chain length, the greater the values of all of these limiting properties. This is expected as the number of the interaction sites in a linear alkane increases with increasing length, and thus, the strength of the interaction with the IRMOF-1 increases. Linear relations between the limiting properties and Nc are obtained

q0st ) 4.693Nc + 5.948

(1)

ln KH ) 1.352Nc - 7.160

(2)

∆S° ) -4.510Nc - 6.203

(3)

Such linear relations were previously found for the adsorption of linear alkanes in a SWNT bundle,22 in Carbopack C and C-HT,33 and in silicalite.24,34-40 From these relations, we can estimate the limiting adsorption properties of longer linear alkanes. For C5 isomers, the greater the degree of branching, the smaller the values of the limiting adsorption properties. This is a consequence of the decreased packing efficiency within IRMOF-1 as the degree of branching increases. Compared to nC5 and iC5, neoC5 has a much lower absolute value of ∆S°, even lower than ∆S° of nC4. Again, this is a result of the alkane configurations in that neoC5 is a pseudospherical molecule, unlike the linear alkanes. Upon adsorption and confinement in IRMOF-1, the configuration of neoC5 is not appreciably changed from that in bulk fluid, and therefore the entropy loss upon confinement is less than that of the other C5 isomers. Figure 3 shows the adsorption and desorption isotherms of pure linear alkanes (C1, C2, C3, nC4, and nC5) in IRMOF-1 at 300 K as a function of bulk fugacity. These isotherms are completely reversible for adsorption and desorption. The dotted lines indicate the saturation conditions of bulk alkanes. Consequently, the isotherms correspond to vapor adsorption below saturation and to liquid adsorption above saturation. At low fugacities, there is considerable adsorption of the longest alkane and almost no adsorption of the shortest alkane. The number of admolecules increases with increasing alkane length because the long alkane interacts with IRMOF-1 more strongly as a result of the increased number of interaction sites. However, at high fugacities, there is greater adsorption of the short alkanes, and the amount adsorbed and the saturation coverage increase with decreasing alkane size. This is due to the dominance of the size entropy effect at high coverages, in which small molecules can fit into partially filled pores more easily, so that a given volume can hold more small molecules. Also, the long alkane is found to approach adsorption saturation at a lower fugacity than the short alkane. For example, the adsorption of nC5, nC4, and C3 are nearly saturated at 103 kP, though not for C1 and C2. This behavior is similar to the adsorption of pure linear alkanes in a SWNT bundle22 and in silicalite.24,36,38,41 Table 2 gives the gravimetrically and volumetrically adsorbed amounts of pure linear alkanes in IRMOF-1, in a SWNT bundle,22 and in silicalite23,24 at 300 K and 104 kPa. In terms of the gravimetric amount, the adsorption in IRMOF-1 is the largest, about 5 times that in a SWNT bundle and about 8 times that in silicalite. In terms of the volumetric amount, the adsorption in IRMOF-1 is again the largest, about 2 times that in a SWNT bundle and about 3 times that in silicalite. This suggests that (33) Bruce, C. D.; Rybolt, T. R.; Thomas, H. E.; Agnew, T. E.; Davis, B. S. J. Colloid Interface Sci. 1997, 194, 448. (34) Smit, B.; Siepmann, J. I. Science 1994, 264, 1118. (35) Dubbeldam, D.; Calero, S.; Vlugt, T. J. H.; Krishna, R.; Maesen, T. L. M.; Smit, B. J. Phys. Chem. B 2004, 108, 12301. (36) Calero, S.; Smit, B.; Krishna, R. Phys. Chem. Chem. Phys. 2001, 3, 4390. (37) Calero, S.; Dubbeldam, D.; Krishna, R.; Smit, B.; Vlugt, T. J. H.; Denayer, J. F. M.; Martens, J. A.; Maesen, T. L. M. J. Am. Chem. Soc. 2004, 126, 11377. (38) Krishna, R.; Calero, S.; Smit, B. Chem. Eng. J. 2002, 88, 81. (39) Denayer, J. F.; Souverijns, W.; Jacobs, P. A.; Martens, J. A.; Baron, G. V. J. Phys. Chem. B 1998, 102, 4588. (40) Pascual, P.; Ungerer, P.; Tavitian, B.; Pernot, P.; Boutin, A. Phys. Chem. Chem. Phys. 2003, 5, 3684. (41) Schenk, M.; Vidal, S. L.; Vlugt, T. J. H.; Smit, B.; Krishna, R. Langmuir 2001, 17, 1558.

Adsorption and Separation of Alkanes in IRMOF-1

Langmuir, Vol. 22, No. 13, 2006 5705

Figure 3. Isotherms of pure linear alkanes (C1, C2, C3, nC4, and nC5) in IRMOF-1 at 300 K. Circles are for adsorption, and the crosses are for desorption. The dotted lines indicate saturation conditions of bulk alkanes.

Figure 4. Isotherms of pure C5 isomers (nC5, iC5, and neoC5) in IRMOF-1 at 300 K. Circles are for adsorption, and the crosses are for desorption. The dotted lines indicate saturation conditions of bulk alkanes.

Table 2. Gravimetrically and Volumetrically Adsorbed Amounts of Pure Linear Alkanes in IRMOF-1, in a SWNT Bundle, and in Silicalite at 300 K and 104 kPa

increasing and progressively replaces the long alkane at high fugacities. At fugacities above those considered here, however, adsorption saturation will be achieved. The fugacity at which the maximum adsorption occurs for the long alkane (e.g., nC5) is lower than for a shorter alkane (e.g., nC4). At low fugacities, the preferential adsorption of the long alkane is a result of the energetic contribution. The replacement of the long alkane by the short one at elevated fugacity is due to the previously mentioned size entropy effect. Such an effect has been observed previously, for example, in the simulation studies of the adsorption of linear alkane mixtures in a SWNT bundle22 and in zeolites,23,36,38,42 of a CF4-Xe mixture in a SWNT bundle,43 of N2-O2 mixtures in a SWNT bundle44 and in C168 schwarzite,45 in predictions of the adsorption of hard rods on a linear substrate,46 and of squarewell mixtures in one-dimension,47 and in measurements of the adsorption of N2-O2 mixtures in anatase.48 Figure 6b shows the adsorption selectivities of nC5, nC4, C3, and C2 relative to C1. The adsorption for long alkane, particularly for nC5 the longest one in the mixture, first increases faster with increasing fugacity than for a short alkane and then decreases. As a consequence, the selectivity increases to a maximum at an intermediate fugacity, and then decreases. Not surprisingly, the greater the difference between the two components, the higher the selectivity. Compared to a SWNT bundle,22 the selectivities for the various alkanes in IRMOF-1 are smaller, implying that the separation efficacy in IRMOF-1 is less, though the total amount adsorbed is greater. Figure 7 shows the adsorption of a three-component mixture of C5 isomers in IRMOF-1 at the bulk partial fugacity ratio of nC5: iC5: neoC5 of 1: 1: 1. The isotherms in Figure 7a show competitive adsorption between linear and branched isomers. With increasing fugacity, the adsorption of each isomer increases, but the adsorption of nC5 increases more rapidly and has a higher saturation value. This is due to the configurational entropy effect as the packing efficiency within the IRMOF-1 is greater with a lower degree of branching, in accord with the concept of shape selectivity; that is, a slender alkane is preferentially adsorbed

SWNT22

IRMOF-1

silicalite23,24

mmol/g mmol/cm3 mmol/g mmol/cm3 mmol/g mmol/cm3 alkane 20.9 21.6 17.4 14.6 12.5

12.4 12.8 10.3 8.7 7.4

5.1 4.3 3.4 2.9 2.5

6.8 5.7 4.5 3.9 3.3

2.8 2.6 2.1 1.7 1.5

5.0 4.7 3.8 3.0 2.7

C1 C2 C3 nC4 nC5

IRMOF-1 is the best candidate for the adsorptive storage of alkanes among these three nanoporous materials. Figure 4 shows the adsorption and desorption isotherms of pure C5 isomers (nC5, iC5, and neoC5) in IRMOF-1 at 300 K as a function of bulk fugacity. As for the pure linear alkanes, adsorption and desorption are completely reversible. The dotted lines indicate the saturation conditions of bulk alkanes. Consequently, the isotherms correspond to vapor adsorption below saturation, and to liquid adsorption above saturation. With increased degree of branching, the adsorbed amount decreases. This is generally attributed to the configuration entropy effect, as the slender linear isomer has a less steric hindrance within IRMOF-1 and can pack more efficiently than the bulky branched isomer. This has also been observed in the adsorption of alkanes in a SWNT bundle22 and in silicalite.36,38,42 Figure 5 shows the center-of-mass distributions of nC5 molecules adsorbed in IRMOF-1 at 300 K and three different fugacities 1, 100, and 10000 kPa. The upper part shows the density profiles from the center of a unit cell. The lower part shows the equilibrium snapshots generated by accumulating 50 configurations. From low to high fugacity, the adsorbed density increases slightly. The adsorbate molecules are observed to intercalate only the central cavity within IRMOF-1, which is why the isotherms of the linear and branched alkanes in Figures 3 and 4 have only one step up to saturation. Figure 6 shows the adsorption of a five-component mixture of C1 to nC5 linear alkanes in IRMOF-1. The bulk fugacity ratio of C1:C2:C3:nC4:nC5 is 5:4:3:2:1. The isotherms in Figure 6a show competitive adsorption between long and short alkanes. With increasing fugacity, the long alkane is first adsorbed, the extent of adsorption passes through a maximum, and then decreases, whereas the short alkane adsorption continues (42) Krishna, R.; Smit, B.; Calero, S. Chem. Soc. ReV. 2002, 31, 185.

(43) Byl, O.; Kondratyuk, P.; Forth, S. T.; FitzGerald, S. A.; Chen, L.; Johnson, J. K.; Yates, J. T. J. Am. Chem. Soc. 2003, 125, 5889. (44) Jiang, J. W.; Sandler, S. I. Langmuir 2004, 20, 10910. (45) Jiang, J. W.; Sandler, S. I. Langmuir 2003, 19, 5936. (46) Talbot, J. AIChE J. 1997, 43, 2471. (47) Heuchel, M. Langmuir 1997, 13, 1150. (48) Arnold, J. R. J. Am. Chem. Soc. 1949, 71, 104.

5706 Langmuir, Vol. 22, No. 13, 2006

Jiang and Sandler

Figure 5. Distributions of the centers-of-mass of nC5 molecules adsorbed in IRMOF-1 at 300 K and 1, 100, and 10 000 kPa. Upper: density profiles from the center of a unit cell. Lower: equilibrium snapshots.

Figure 6. Adsorption of a mixture of linear alkanes (C1:C2:C3: nC4:nC5 ) 5:4:3:2:1) in IRMOF-1 at 300 K. (a) Isotherms and (b) selectivities with respect to C1.

over a bulky one. Such behavior has been observed in the adsorption of C5 isomers in a SWNT bundle,22 and of linear and branched alkane mixtures in MFI zeolite.24,36,38,41,42 Figure 7b shows the selectivities of nC5 and iC5 relative to neoC5. With increasing fugacity, the selectivity first increases and then decreases slightly to a constant value near saturation. Though

Figure 7. Adsorption of a mixture of C5 isomers (nC5:iC5:neoC5 ) 1:1:1) in IRMOF-1 at 300 K. (a) Isotherms and (b) selectivities with respect to neoC5.

the values of selectivity here are not as large as in Figure 6b, adsorptive separation between the C5 isomers is still possible. Again, as for the mixture of linear alkanes, the selectivities in IRMOF-1 are smaller than those in a SWNT bundle.

Adsorption and Separation of Alkanes in IRMOF-1

IV. Conclusions We have investigated the adsorption of pure linear and branched alkanes and their mixtures in IRMOF-1 using Monte Carlo simulations. The limiting adsorption properties for pure linear alkanes are found to be linearly related to their carbon number, which permits the estimation of the limiting properties for even longer linear alkanes. Competitive adsorption occurs in alkane mixtures as a consequence of size and/or configurational differences between the components. For a mixture of linear alkanes, the energetic contribution prevails at low fugacities, and the long alkane is preferentially adsorbed; however, the size entropy effect becomes more important at high fugacities, and the short alkane progressively replaces the long alkane. For a mixture of linear and branched isomers, the configurational entropy effect dominates, and there is greater adsorption of the linear isomer. The storage capacities of IRMOF-1 are found to

Langmuir, Vol. 22, No. 13, 2006 5707

be greater than a SWNT bundle and silicalite, but the separation efficacies are less. To date, a large number of MOFs have been synthesized. Although the adsorption of alkanes only in IRMOF-1 was considered here, the study of other MOFs is also of interest. The wide variety of the metal oxides and organic linkers available provides considerable opportunities to develop nanostructured MOFs with unique functionalities, specified pore sizes, and geometries. In this way, it may be possible to produce tailored MOFs optimized for the storage and/or separation of specific alkanes. Acknowledgment. Support from the National University of Singapore to J.J. and from the U.S. Department of Energy and U.S. National Science Foundation to S.I.S. is graciously acknowledged. LA060506G