pubs.acs.org/Langmuir © 2009 American Chemical Society
Microporous Coordination Polymers As Selective Sorbents for Liquid Chromatography Rashid Ahmad,†,‡ Antek G. Wong-Foy,† and Adam J. Matzger*,† †
Department of Chemistry and the Macromolecular Science and Engineering Program, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109-1055, and ‡Chemistry Division, Directorate of Science, PINSTECH, Nilore, Islamabad, Pakistan Received June 24, 2009. Revised Manuscript Received August 5, 2009
We evaluate the potential of microporous coordination polymers (MCPs) to act as the stationary phase in liquid chromatographic separations. MCPs derived from carboxylates coordinated to copper (HKUST-1) and zinc (MOF-5) were studied. The shape and size selective separation of organic compounds including benzene, ethylbenzene, styrene, naphthalene, anthracene, phenanthrene, pyrene, 1,3,5-triphenylbenzene, and 1,3,5-tris(4-bromophenyl)benzene was performed, and in most cases excellent separation was achieved based on a combination of molecular sieving and adsorption effects.
Introduction Most liquid chromatographies rely on the partitioning of an analyte between solution and an adsorbed state on a solid. Commonly, silica gel or alumina, with appropriate derivatization for reversed phase conditions, serves as the solid adsorbent. These are porous materials with modest surface areas (200-800 m2/g) and pore sizes in the range of 50-250 A˚.1 The porosity in these materials offers access to the surface area such that reasonable capacity can be achieved; however, for separations involving small molecules, selective inclusion of molecules in the pores based on size is a very minor contributor to retention differences, and instead the adsorption energies are most important. This stands in stark contrast to gel permeation chromatography (GPC), where polymers and oligomers are separated by size. GPC supports such as cross-linked styrenes, although having nanoscale features, are amorphous and present a broad distribution of pore sizes. Therefore an approach to highly selective separations of small molecules would be to employ sorbent materials with defined nanostructure on the molecular level. Zeolites can provide such properties in the small pore regime, but are unable to accommodate many molecules of interest in pharmaceutical and fine chemical applications; furthermore, zeolites suffer from poor transport kinetics due to restricted pore apertures.2 Microporous coordination polymers (MCPs) offer an attractive alternative because of their tunable pore sizes and high levels of porosity (>1 cc/g); we have previously demonstrated that polycyclic dyes3 and organosulfur compounds4 can be adsorbed by MCPs. These, and experiments by other groups, suggest the considerable potential of MCPs to act as novel *Corresponding author. E-mail:
[email protected]. (1) Cserhati, T.; Valko, K. Chromatographic Determination of Molecular Interactions: Applications in Biochemistry, Chemistry and Biophysics; CRC Press: Ann Arbor, 1994. (2) Martens, J. A.; Jacobs, P. A. Stud. Surf. Sci. Catal. 2001, 137, 633–671. (3) Chae, H. K.; Siberio-Perez, D. Y.; Kim, J.; Go, Y. B.; Eddaoudi, M.; Matzger, A. J.; O’Keffe, M.; Yaghi, O. M. Nature 2004, 427, 523–527. (4) Cychosz, K. A.; Wong-Foy, A. G.; Matzger, A. J. J. Am. Chem. Soc. 2008, 130, 6938–6939. (5) Alaerts, L.; Maes, M.; Jacobs, P. A.; Denayer, J. F. M.; De Vos, D. E. Phys. Chem. Chem. Phys. 2008, 10, 2979–2985. (6) Owen, R. E.; Helen, L. N.; Wenbin, L. J. Am. Chem. Soc. 2001, 123, 10395– 10396.
Langmuir 2009, 25(20), 11977–11979
separation media.5-7 With these successes in liquid phase adsorption and the recent commercial availability of several MCPs, this is an appropriate time for an evaluation of the potential of these novel sorbents for applications where traditional media are inadequate.
Experimental Section Basolite C 300 (Cu3(C9H3O6)2) was obtained from Aldrich, and MOF-5 was synthesized as described elsewhere.8 High-performance liquid chromatography (HPLC) columns were obtained from Restek and dry packed in a glovebox according to the previously described procedure.9 HPLC-grade hexanes and CH2Cl2 were used as the mobile phase at a flow rate of 1 mL/min. A 0.01 wt % solution of adsorbate in hexanes (10 μL) was injected for all trials. Separations were carried out using a Waters HPLC consisting of a 600 multi solvent delivery system, 717 auto sampler, and 996 PDA detector. In all cases, the UV-visible spectrum and retention time were used to confirm identity of the peaks.
Results and Discussion To test the notion that MCPs may serve as selective sorbents for liquid chromatography, Basolite C 300, a recently commercialized sorbent that has the structure of HKUST-110 was employed as a stationary phase in HPLC. HKUST-1 (Figure 1i) is one of the earliest examples of a highly porous MCP and these facecentered-cubic crystals contain an intersecting three-dimensional system of large square shaped pores (9 9 A˚) with 40.7% accessible porosity and surface area exceeding 1000 m2/g. Basolite C 300 was chosen as a representative example of this emerging class of sorbents and was applied to the separation of several series of compounds where either size or functionality was varied. To test the potential for size selective separation of small molecules, the acene series benzene, naphthalene, and anthracene (7) Suslick, K. S.; Bhyrappa, P.; Chou, J. H.; Kosal, M. E.; Nakagaki, S.; Smithenry, D. W.; Wilson, S. R. Acc. Chem. Res. 2005, 38, 283–291. (8) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’Keefe, M.; Yaghi, O. M. Science 2002, 295, 469–472. (9) Hamilton, R. J.; Sewell, P. A. Introduction to High Performance Liquid Chromatography, 2nd ed.; Chapman and Hall: New York, 1982. (10) Chui, S. S.-Y.; Lo, S. M.-F.; Charmant, J. P. H.; Orpen, A. G.; Williams, I. D. Science 1999, 283, 1148–1150.
Published on Web 09/15/2009
DOI: 10.1021/la902276a
11977
Letter
Figure 1. Space-filling representations of the structures of (a) benzene, (b) naphthalene, (c) anthracene, (d) phenanthrene, (e) pyrene, (f) 1,3,5-triphenylbenzene, (g) 1,3,5-tris(4-bromophenyl)benzene, (h) MOF-5, and (i) HKUST-1 shown on a common scale to convey the relative sizes of sorbent and sorbate. Numbers represent kinetic diameters (A˚) and are calculated from the minimum cross-sectional diameter of the molecule, assuming a Lennard-Jones 6-12 potential to describe the interaction energy (see ref 17). (Atom (sphere color): C (gray), H (white), O (red), Br (orange), Cu (turquoise), Zn (blue)).
was chosen and passed through a column of 250 4.6 mm dimensions. Benzene, when carried in hexanes at a flow rate of 1 mL/min, has a retention time of 38 min, the retention time of naphthalene is 180 min and anthracene remains adsorbed after more than 4 h.11 The kinetic diameter of benzene is less than the pore aperture of HKUST-1, therefore its diffusion, equilibration, and desorption are fast. Naphthalene more strongly interacts with the framework, and, in the case of anthracene, the size of the sorbate molecule is comparable to the MCP channels leading to strong partitioning.12 The kinetic diameter of these acenes is very similar (Figure 1a-c) because of their essentially identical crosssectional areas and therefore in the second set of experiments shape-selective separation of aromatic compounds was studied. When a mixture of benzene, naphthalene, and 1,3,5-triphenylbenzene was passed through the column, the 1,3,5-triphenylbenzene was essentially unretained, whereas the acenes eluted after more than 30 min (Figure 2, top trace) suggesting a size-exclusion effect. Consistent with this notion is the fact that 1,3,5-tris(4bromophenyl)benzene is also unretained. Phenanthrene, like anthracene, remained adsorbed after 4 h; however, pyrene elutes rapidly suggesting that it is just large enough to be excluded from the pores. Taken together, these data suggest that larger molecules are more retained if they can access the pores efficiently (higher acenes, phenanthrene) but that, above a certain size threshold (pyrene, 1,3,5-triphenylbenzene, and 1,3,5-tris(4-bromophenyl)benzene), the inability to penetrate/diffuse in the pores leads to negligible retention. (11) When the column diameter was changed to 100 2.1 mm, the retention time was 4 and 19.3 min for benzene and naphthalene, respectively, but anthracene remained adsorbed for more than 10 h. (12) The partition coefficient value for naphthalene was 33 cm3 g-1 as compared to 2363 cm3 g-1 for anthracene as determined by the shake flask method.
11978 DOI: 10.1021/la902276a
Ahmad et al.
Figure 2. Separations achieved using Basolite C 300 as the stationary phase. Chromatogram of a mixture of 1,3,5-triphenylbenzene, benzene, and naphthalene (absorbance is the average at 210 and 220 nm, top trace). Chromatogram of styrene and ethyl benzene (absorbance is the average of that at 210 and 215 nm, bottom trace).
Figure 3. Chromatogram of 1,3,5-tris(4-bromophenyl)benzene, naphthalene, and pyrene using MOF-5 as the stationary phase (absorbance is the average at 220, 238, and 240 nm).
To test the hypothesis that size-selective separation is being achieved in HKUST-1, MOF-58 (Figure 1h), the best studied MCP, and one with comparatively larger pore dimensions, was applied for the separation of hydrocarbons to support the proposed mechanism. MOF-5 possesses an open cubic structure that consists of face-sharing cubic cages that extend in all three dimensions. Benzene, naphthalene, anthracene, phenanthrene, and pyrene were passed through a column of 150 4.6 mm dimensions. Their retention times were 4.1, 4.4, 8.7, 8.3, and 11.2 min, respectively, indicating the overall less retentive nature of this structure compared to HKUST-1. When a mixture of 1,3,5-tris(4-bromophenyl)benzene, naphthalene, and pyrene was passed through the column, the 1,3,5-tris(4-bromophenyl)benzene passed unretained whereas naphthalene and, in contrast to HKUST-1, pyrene were both retained (Figure 3). The generally Langmuir 2009, 25(20), 11977–11979
Ahmad et al.
longer retention times of benzene, naphthalene, anthracene, and phenanthrene in HKUST-1 can be rationalized by the comparatively larger pores of MOF-5 (compare h to i in Figure 1) offering a weaker attractive potential, although the coordinatively unsaturated metal sites in HKUST-1 or the more electron deficient nature of the tricarboxylate linker may also contribute to these differences. For the larger analytes 1,3,5-triphenylbenzene, and 1,3,5-tris(4-bromophenyl)benzene the retention times are 4.7 and 1.6 min, respectively, in MOF-5. The poor retention of these molecules clearly stems from their larger size precluding diffusion into the pores albeit with some penetration of 1,3,5-triphenylbenzene as judged by the longer residence time; in HKUST-1, both compounds are completely unretained. These results are the first demonstration of a molecular sieving effect for the HPLC separation of large hydrocarbons by MCPs and these findings are clearly very promising in demonstrating strong size selectivity in a heretofore inaccessible size range. The above results suggest that adsorption behavior in MCPs is a hybrid of GPC-like size selectivity and HPLC-like adsorption. The strong partitioning of anthracene and phenanthrene into HKUST-1 is therefore understood as arising from these three ring molecules having a small enough size to access the internal surface area of the adsorbent and a strong affinity relative to the mobile phase. If true, the retention behavior should be tunable by controlling solvent polarity and therefore 10% v/v of CH2Cl2 in hexanes was investigated as a mobile phase with a Basolite C 300 column of dimensions 250 4.6 mm. The retention times for benzene, naphthalene, anthracene, phenanthrene, pyrene, 1,3,5triphenylbenzene, and 1,3,5-tris(4-bromophenyl)benzene were 24.7, 19, 51, 13.5, 3.8, 3.8, and 4.2 min, respectively; anthracene and phenanthrene, which were difficult to elute from the column with pure hexanes, now display reasonable retention times, and this result demonstrates that MCPs, like other stationary phases for liquid chromatography, follow the general rules of retention in that the more retained larger solutes are more susceptible to mobile phase influence than smaller solutes.13 The elution time of pyrene, 1,3,5-triphenylbenzene, and 1,3,5-tris(4-bromophenyl)benzene remain similar to those in pure hexanes because they are mainly excluded from the pores and display the expected insensitivity of retention time to solvent identity that is typically observed in GPC. The above experiments demonstrate the ability of MCPs to act as size-selective sorbents in the liquid phase giving rise to behavior similar to GPC but with sharp cutoffs around molecular dimensions. The limitation of this approach is that size selectivity will (13) Schoenmakers, P. J.; Billet, H. A. H.; Galan, L. d. Chromatographia 1982, 15, 205–214. (14) Takahashi, A.; Yang, R. T. AIChE J. 2002, 48, 1457–1468.
Langmuir 2009, 25(20), 11977–11979
Letter
fail for many important separations; for example, the separation of ethyl benzene and styrene is an industrially significant process where the molecules are of very similar dimensions. Styrene is produced by the dehydrogenation of ethyl benzene obtained from petroleum refining. Their separation by distillation is a very energy-intensive process motivating the development of alternative separation techniques. When a mixture of ethyl benzene and styrene was passed through the Basolite C 300 column (250 4.6 mm) in hexanes at a flow rate of 1 mL/min, the retention times for ethyl benzene and styrene were 35 and 125 min, respectively (Figure 2, bottom trace). This constitutes a very efficient separation for two molecules of approximately identical size. Unlike GPC, where the separation is governed by molecular size, the separation in this case is proposed to be due to the coordinative interaction of styrene with copper in the stationary phase, i.e., the π-complexation mechanism.14 To test this hypothesis, an ethyl benzene and styrene mixture was passed through the MOF-5 column in hexanes at a flow rate of 1 mL/min. The components coeluted at approximately 3 min in the absence of a sorbent with coordinatively unsaturated metal sites.
Conclusion In conclusion, MCPs can be effectively used as stationary phases in liquid chromatography, and this will be an important application of these materials. The combination of accessible pores and surface areas now exceeding 5000 m2/g16 is ideal for the chromatographic separation process. By selecting specific MCPs, target compounds can be separated by design. Additionally, MCPs are superior to cross-linked polystyrene used in GPC because a smaller size regime is accessible, and the size cutoffs are more defined.15 Finally, utilizing the unique chemistry of the MCPs can separate even molecules with very similar sizes if coordination differences can be exploited, and this approach has the potential for yielding chemically tailored separation media. Acknowledgment. We thank Dalvin Mendez for conducting initial investigations. R.A. was supported in part by a fellowship from the Higher Education Commission of Pakistan. This work was supported by Grant #454 of the 21st Century Jobs Trust Fund received through the SEIC Board from the State of Michigan. (15) Ahuja, S. Chromatography and Separation Science; Academic Press: New York, 2003; Vol. 4, p 203. (16) Koh, K.; Wong-Foy, A. G.; Matzger, A. J. J. Am. Chem. Soc. 2009, 131, 4184–4185. (17) Breck, D. W. Zeolite Molecular Sieves; Krieger Publishing Co.: Malabar, FL, 1984; Chapter 8.
DOI: 10.1021/la902276a
11979