Characterization and Testing of Periodic Mesoporous Organosilicas

Sep 29, 2009 - David C. Calabro, Corporate Strategic Research, ExxonMobil Research & Engineering Company, Annandale, NJ 08801; Tel: (908) 730-3713; Fa...
0 downloads 10 Views 2MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Characterization and Testing of Periodic Mesoporous Organosilicas as Potential Selective Benzene Adsorbents )

W. G. Borghard,† D. C. Calabro,*,† F. P. DiSanzo,† M. M. Disko,† J. W. Diehl,† J. C. Fried,† M. A. Markowitz,*,‡,§ M. Zeinali,‡, B. J. Melde,*,‡ and A. E. Riley† ExxonMobil Research and Engineering Company, Annandale, New Jersey 08801, and ‡Code 6900, Naval Research Laboratory, Washington, DC 20375. §Current address: Division of Materials Science & Engineering, Office of Basic Energy Sciences, U. S. Department of Energy, Washington, D.C. 20585. Current address: Environmental, LLC, P.O. Box 1126, Columbia, MD 21044. )



Received April 15, 2009. Revised Manuscript Received July 17, 2009 The effects of surface imprinting on the adsorption and desorption properties of benzene- and diethylbenzene-bridged periodic mesoporous organosilicas (PMOs) acting as GC stationary-phase preconcentration sorbents for benzene and xylene were examined. Surface-imprinted and nonimprinted PMOs with diethylbenzene (DEB), benzene (BENZ), and ethane (BTSE) bridges and nonimprinted mesoporous silica (MCM-41) were prepared via well-established surfactant templating synthetic methods. The imprinted materials were synthesized using a surfactant demonstrated to produce trinitrotoluene (TNT) selective sorbents with increased adsorption capacity for cresol and 4-nitrophenol as well as TNT. Powder XRD and nitrogen sorption measurements revealed that all of the materials were mesoporous with the DEB materials having a random pore structure and lower surface area than the other materials which had ordered pore structures. Results for maximum uptake of benzene and p-xylene indicate a small but consistent positive effect on the adsorption of benzene and p-xylene due to surface imprinting. Comparing the surface area normalized uptakes (mg/m2) for materials having the same organic bridge with and without imprinting (DEB vs TDMI-DEB and BENZ vs TDMI-BENZ) shows that in seven of eight comparisons the imprinted analogue had a higher aromatic uptake. The imprinted samples showed higher weight normalized uptakes (mg/g) in five of eight cases. When used as a GC stationary phase, the organosilica materials yield more symmetrical chromatographic peaks and better separation than MCM-41, indicating superior trapping of BTX analytes, particularly at low concentrations. Additionally, these materials rapidly desorb the preconcentrated compounds.

Introduction A key element for improving the trace detection capability of sensors is the development of preconcentrator materials with high capacity and selectivity for sorption of the target analyte molecule(s). Preconcentration enables the collection of enough sample mass to obtain detectable signals from a sensor array. In addition to selectivity and high capacity, preconcentrating sorbents must have rapid adsorption/desorption kinetics, stability to environmental conditions, and low fouling. While some sorbents have been developed as sensor preconcentrators,1-11 effective *Corresponding authors. Michael A. Markowitz, Office of Basic Energy Sciences, U. S. Department of Energy; Tel: (301) 903-6779; Fax: (301) 903-9513; Email: [email protected]. David C. Calabro, Corporate Strategic Research, ExxonMobil Research & Engineering Company, Annandale, NJ 08801; Tel: (908) 730-3713; Fax: (908) 730-3198; Email: dccalabro@exxonmobil. com. Brian J. Melde, Code 6900, Naval Research Laboratory, Washington, D.C. 20375: Tel: (202) 767-0591; Fax: (202) 767-9594; Email: [email protected]. (1) Jenkins, T. F.; Miyares, P. H.; Myers, K. F.; McCormick, E. F.; Strong, A. B. Anal. Chimi. Acta 1994, 289, 69–78. (2) Harvey, S. D.; Clauss, T. R. W. J. Chromatogr., A 1996, 753, 81–89. (3) Smith, M.; Collins, G. E.; Wang, J. J. Chromatogr., A 2003, 991, 159–167. (4) Walcarius, A.; Mandler, D.; Cox, J. A.; Collinson, M.; Lev, O. J. Mater. Chem. 2005, 15, 3663–3689. (5) Carrington, N. A.; Yong, L.; Xue, Z. Anal. Chim. Acta 2006, 572, 17–24. (6) Carrington, N. A.; Xue, Z.-L. Acc. Chem. Res. 2007, 40, 343–350. (7) Zheng, F.; Baldwin, D. L.; Fifield, L. S.; Anheier, N. C.; Aardahl, C. L.; Grate, J. W. Anal. Chem. 2006, 78(7), 2442–2446. (8) Davis, C. E.; Ho, C. K.; Hughes, R. C.; Thomas, M. L. Sens. Actuators, B 2005, 104, 207–216. (9) Lu, C.-J.; Zellers, E. T. Anal. Chem. 2001, 73, 3449–3457. (10) Panavaite, D.; Padarauskas, A.; Vickackaite, V. Chemija 2005, 16(2), 24. (11) Bianchi, F.; Pinalli, R.; Ugozzoli, F.; Spera, S.; Careri, M.; Dalcanale, E. New J. Chem. 2003, 27(3), 502.

Langmuir 2009, 25(21), 12661–12669

sorbents are still lacking for many analytes. Consequently, new materials are key to the development of preconcentration methods for trace detection of chemical species.12 One method of imparting selectivity to organic and inorganic materials is molecular imprinting.12-15 The generic process of molecularly imprinting a polymer to have selectivity for a specific analyte is to mix that molecule with the polymer precursors and allow the precursors to assemble around the imprint molecule during polymerization. In some cases, the precursors may have pendant functional groups that noncovalently bind to complementary functional groups of the imprinting analyte. After the reaction is complete, the imprint molecule is extracted, leaving a polymer with molecular recognition sites selective for the shape and functional group orientation of the analyte. Mesoporous organosilicas with open-ended pore structures have generated interest as preconcentration sorbents because of their high adsorption capacities and the rapid diffusion of small molecules through their pores.4,16 Template-directed molecular imprinting was developed as a way of imparting selectivity for small molecules primarily into the high surface area of (12) Guan, W. N.; Xu, F.; Liu, W. M.; Zhao, J. H.; Guan, Y. F. J. Chromatogr., A 2007, 1147, 59–65. (13) Haupt, K.; Mosbach, K. Trends Biotechnol. 1998, 16, 468. (14) Wulff, G. Chem. Rev. 2002, 102, 1. (15) Alexander, C; Andersson, H. S.; Andersson, L. I.; Ansell, R. J; Kirsch, N.; Nicholls, I. A.; O’Mahony, J.; Whitcombe, M. J. J. Molec. Recognition 2006, 19, 106. (16) Hoffmann, F.; Cornelius, M.; Morell, J.; Froba, M. Angew. Chem., Int. Ed. 2006, 45, 3216–3251.

Published on Web 09/29/2009

DOI: 10.1021/la901334z

12661

Article

organosilica materials.17-23 This approach involves mixing a surface-active molecular imprint of the target analyte with a micelle-forming surfactant. Under proper conditions, this mixture forms a self-assembled structure that is used as a template around which bis-trialkoxyorganosilanes or a mixture of tetraalkoxysilanes and trialkoxyorganosilanes are allowed to condense and polymerize. Subsequent extraction of surfactant and imprint molecule has yielded mesoporous organosilicas with selectivity for specific organophosphonates19-21 and TNT.22,23 In addition, the organosilicas formed using this surface imprinting technique have exhibited significantly higher adsorption capacities for both the imprinted analytes and similarly structured molecules.19-23 Since developing sorbents with selectivity for individual compounds is a costly and time-consuming process, taking advantage of this enhancement of adsorption capacity to boost the adsorption capacities of organosilicas for families of molecules might be an efficient way of enhancing preconcentration effectiveness. In combination with other means of molecular discrimination, such as preferential adsorption by specific organic bridging groups, and controlling the temperature at which a specific molecule within a family desorbs from a sorbent, increasing the capability of a preconcentrator to capture analytes from very dilute concentrations in any environment may sufficiently improve sensor sensitivity to enable trace detection of toxic materials. Of particular interest is the gas-phase detection of benzene at trace levels in mixtures containing similarly structured aromatic molecules such as toluene and xylene as well as other hydrocarbons. The gas-phase analysis of benzene/toluene/xylene (BTX) mixtures at trace levels (e100 ppb) typically employs GC or GC-MS methods either via direct headspace analysis24,25 or in combination with a cryogenic26 or solid-phase microextraction10,27,28 preconcentrator. While this technology excels for laboratory analysis, major technical hurdles must be overcome (e.g., μ-GC) to adapt these methods to real-time personal exposure monitoring. Real-time exposure sensors must be compact, lightweight, and self-contained. One route toward this goal is to develop miniaturized high-capacity preconcentration sorbents. To test the viability of using template-directed molecular imprinting to improve preconcentration performance, we investigated the relative adsorption properties of surface-imprinted and nonimprinted mesoporous silicas and organosilicas having open-ended pore structures, for benzene, toluene, and xylene. Specifically, surface-imprinted and nonimprinted mesoporous organosilanes were prepared with diethylbenzene and benzene bridged bistrialkoxysilane precursors via well-established (17) Markowitz, M. A.; Kust, P. R.; Deng, G.; Schoen, P. E.; Dordick, J. S.; Clark, D. S.; Gaber, B. P. Langmuir 2000, 16, 1759–1765. (18) Markowitz, M. A.; Kust, P. R.; Klaehn, J.; Deng, G.; Gaber, B. P. Anal. Chim. Acta 2001, 435, 177–185. (19) Jayasundera, S.; Zeinali, M.; Miller, J. B.; Velea, L. M.; Gaber, B. P.; Markowitz, M. A. J. Phys. Chem. B 2006, 110, 18121–18125. (20) Markowitz, M. A.; Deng, G.; Burleigh, M. C.; Wong, E. M.; Gaber, B. P. Langmuir 2001, 17, 7085–7092. (21) Markowitz, M. A.; Deng, G.; Gaber, B. P. Langmuir 2000, 16, 6148–6155. (22) Johnson-White, B.; Zeinali, M.; Shaffer, K. M.; Patterson, C. H.; Charles, P. T.; Markowitz, M. A. Biosens. Bioelectron. 2007, 22, 1154–1162. (23) Trammell, S. A.; Zeinali, M.; Melde, B.; Charles, P.; Velez, F.; Dinderman, M. A.; Kusterbeck, A.; Markowitz, M. A. Anal. Chem. 2008, 80, 4627. (24) Rosell, M.; Lacorte, S.; Forner, C.; Rohns, H.-P.; Irmscher, R.; Barcelo, D. Environ. Toxicol. Chem. 2005, 24(11), 2785. (25) Pena, F.; Cardenas, S.; Gallego, M.; Valcarcel, M. J. Chromatogr., A 2004, 1052(1-2), 137. (26) Austin, C. C.; Wang, D.; Ecobichon, D. J.; Dussault, G. J. Toxicol. Environ. Health, Part A 2001, 63(6), 437. (27) Xiong, G.; Koziel, J. A.; Pawliszyn, J. J. Chromatogr., A 2004, 1025(1), 57. (28) Rubiano, D.; del Pilar, M.; Marciales, C. C.; Duarte, M. A. Revista Colombiana de Quimica 2002, 31(1), 33. (29) Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.; Terasaki, O. J. Am. Chem. Soc. 1999, 121, 9611–9614.

12662 DOI: 10.1021/la901334z

Borghard et al.

surfactant templating synthetic methods.16,29-31 The imprinted materials were synthesized using a trinitrobenzene surfactant demonstrated to produce TNT (2,4,6-trinitrotoluene) selective sorbents with increased adsorption capacity for cresol, 4-nitrophenol, and RDX (1,3,5-hexahydro-1,3,5-trinitrohydrazine) as well as TNT.22,23 In addition, nonimprinted mesoporous MCM-4132,33 and an ethane-bridged periodic mesoporous organosilica was prepared. Since desorption is as important as adsorption for effective preconcentration, both relative adsorption and desorption properties of these materials toward benzene, toluene, and xylene as GC stationary-phase sorbents were examined.

Experimental Section Materials. The organosilanes 1,4-bis(triethoxysilyl)benzene, bis(triethoxysilylethyl)benzene (mixture of 1,3 and 1,4 isomers), and 1,2-bis(triethoxysilyl)ethane were purchased from Gelest, Inc. (Tullytown, PA). Brij76, NaOH, HCl, and ethanol were purchased from Sigma-Aldrich (St. Louis, MO). Deionized water was obtained from a Milli-Q water purification system (Millipore). 4-Decylaminetrinitrobenzene22 and MCM-4133 were prepared as previously described. Hydrocarbons were obtained from Sigma Aldrich, and used as received. Preparation of Mesoporous Organosilicas. Imprinted and nonimprinted benzene-bridged and diethylbenzene-bridged polysilsesquioxane mesoporous organosilicas were prepared using Brij76 surfactant in acidic media as previously described.22,34 Briefly, the Brij76 surfactant (8.0 g) was added to 400 mL 1 M aq HCl while stirring. The covered mixture was maintained at 50 °C for 12 h prior to addition of the organosilane precursor. The surface-imprinted mesoporous organosilicas were prepared by adding the imprint surfactant, 4-decylaminetrinitrobenzene (0.5 g/400 mL), after Brij76 surfactant equilibration followed by an additional 6 h of equilibration at 50 °C with stirring. This solution was filtered through a 0.2 μm filter to remove excess imprint molecule. The filtered solution was returned to 50 oC and stirred for an additional 3 h prior to the dropwise addition of 1,4-bis(triethoxysilyl)benzene or bis(triethoxysilylethyl)benzene (mixed meta and para isomers) precursor. Aging, surfactant template removal, and drying were the same for both nonimprinted and imprinted PMOs and have been previously described34 with the exception that these powders were not refluxed in deionized water prior to surfactant removal. Characterization. XRD data were collected on a Scintag DCS2000 except for sample NRL4 for which the data were collected on a Bruker D4 Endeavor using a high-speed detector. Data collection range was from 1° to 30° 2θ. Powder samples were placed in shallow planchets. In both cases, Cu radiation was used. Gas sorption experiments were performed using a Micromeritics Tristar. Nitrogen gas was used as the adsorbate at 77 K, using traditional methods. Thermogravimetric analyses were performed with a TA Instruments TGA 2950 thermogravimetric analyzer. All measurements were made in high-resolution dynamic mode. HC Isotherm Data Collection/Analysis. Data were collected on a TA Instruments TGA 2950 with a sparging system to introduce the hydrocarbons at variable partial pressures. In a typical procedure, 20 mg of sample was dried at 150 °C in flowing (30) Melde, B. J.; Holland, B. T.; Blanford, C. F.; Stein, A. Chem. Mater. 1999, 11, 3302–3308. (31) Asefa, T.; MacLachan, M. J.; Coombs, N.; Ozin, G. A. Nature 1999, 402, 867–871. (32) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (33) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834. (34) Burleigh, M. C.; Jayasundera, S.; Thomas, C. W.; Spector, M. S.; Markowitz, M. A.; Gaber, B. P. Colloid Polym. Sci. 2004, 282, 728–733.

Langmuir 2009, 25(21), 12661–12669

Borghard et al.

Article

Figure 1. Powder X-ray diffraction of (A) DEB, (B) TDMI-DEB, (C) BENZ, (D) TDMI-BENZ, (E) BTSE, (F) MCM-41. He for 2 h and cooled to isotherm temperature, and then the weight gain was monitored as hydrocarbon partial pressure increased in steps. Aromatic uptake weights are normalized (Table 2) to both sample dry weight (mg/g) and surface area (mg/m2). Data were collected at two temperatures (35 and 100 °C) for both benzene and p-xylene. The benzene data were only collected in one range (x-y P/Po), whereas the p-xylene data were collected in two ranges (0.02-0.1 and 0.1-0.55) and then combined. Due to equilibrium effects, the data sets did not match at the crossover point, so the first one or two points of the higher range set was likely not at equilibrium and, therefore, not included in the analysis. Gas Chromatography. An Agilent 5880 equipped with a direct-packed injector and a flame ionization detector (FID) was used. All of the materials were in powder form and used without sieving. The materials were packed manually by tapping into 4 in  1/8 in stainless steel tubes. The amount of material loaded ranged 50-125 mg, depending on the pressure drop. The materials were held in the tube by placing glass wool at both ends. The packed materials were conditioned at 180 °C overnight under helium flow in the gas chromatograph prior to obtaining data. Probe mixtures were (1) paraffins comprising equal mass % of n-pentane, n-heptane, n-decane, n-dodecane, n-tetradecane, and 2,2,4-trimethylpentane; (2) aromatics comprising equal mass % of benzene (Bz), toluene (Tol), o-xylene (o-Xyl), 1,2,3-trimethylbenzene (123-TMT); (3) naphthalene dissolved in toluene; and (4) olefins comprising equal mass % of 1-pentene, 1-hexene, and 1-nonene. Probe mixtures were injected directly without splitting using an autosampler with a 0.1 μL injection size. The mixtures were injected separately at isothermal GC oven temperature (°C) of 170, 150, 140, 125, 70, and 50 with a constant helium carrier gas flow of 15 mL/min. Langmuir 2009, 25(21), 12661–12669

Results and Discussion Characterization of Materials. The data in this study were obtained with five mesoporous organosilicas prepared using diethylbenzene (DEB), benzene (BENZ), and ethane (BTSE) bridged bistriethoxyorganosilane precursors, and MCM-41. Surface-imprinted diethylbenzene (TDMI-DEB) and benzene (TDMI-BENZ) bridged sorbent materials were prepared via template-directed molecular imprinting using a mixture of Brij76 and 4-decylaminotrinitrobenzene to direct structure and adsorption site formation. The structures of the polysilsesquioxane materials are shown in Figure S1 (Supporting Information). It is important to note that TDMI-DEB and TDMI-BENZ were imprinted to be selective for trinitrotoluene and not the hydrocarbons used in this study in order to test the premise that the enhanced sorbent adsorption capacity obtained using this procedure would improve the preconcentration capabilities of these sorbents for benzene, toluene, and xylene. Also, the solubility of the 4-decylaminetrinitrobenzene is quite low ( TDMI-DEB. BENZ and TDMI-BENZ show a step in the 35 °C p-xylene sorption data at about 0.1 P/Po attributed to capillary condensation in the mesopore volume, while BTSE shows this feature more intensely at 0.3 P/Po. The difference in position of these capillary condensation steps indicates that the pores may be somewhat occluded in the benzene-bridged PMOs relative to BTSE. For the gas adsorption to show different step positions, it is necessary to have different pore volumes. While all three of these samples have similar pore sizes and fundamental diffraction peaks, BTSE has a significantly larger pore volume and better pore ordering compared to the benzene-bridged PMOs suggesting that the difference in the relative positions of the adsorption step is due to the respective differences in pore ordering. As with the 100 °C data, normalizing the p-xylene uptakes at 35 °C to the BET surface area (Figure 4b) nearly inverts the uptake order. On a mg/m2 basis at maximum P/Po (Table 2), TDMI-DEB > DEB > BTSE> TDMI-BENZ > BENZ. Again, 12666 DOI: 10.1021/la901334z

the measured uptake per unit surface area which emphasizes the relative sorbate/sorbent affinity shows the surface-imprinted DEB linker to have the highest p-xylene uptake. Figures 5 and 6 show benzene gas adsorption isotherms for the organosilicas and for MCM-41. At 100 °C, all of the organosilica materials exhibit higher benzene uptake than MCM-41 on both a weight and surface area basis (Figure 6a). At 35 °C (Figure 6b), the most structurally disordered samples (DEB and TDMI-DEB) have much lower gravimetric uptakes than the others. As seen in the p-xylene isotherms, capillary condensation uptake steps are observed for all samples except DEB and TDMI-DEB. This feature occurs at P/Po ∼ 0.7 for BTSE and 0.4-0.45 for BENZ, TDMI-BENZ, and MCM-41. Note that the benzene and p-xylene uptake curves are similar in shape at low P/Po, indicating little difference in the nature of how these aromatics interact with the materials in the small pore range. This is consistent with a mesoporous material having negligible microporosity. Also, while the aromatics adsorb similarly at low P/Po for the organosilicas, they consistently exceed that obtained with MCM-41 and demonstrate enhanced uptake due to the organic functionalization. The condensation steps for BENZ, TDMI-BENZ, and BTSE occur at much higher P/Po for benzene than for p-xylene, in accordance with their relative boiling points. The presence of this step at comparable pressure in MCM-41 indicates that it can be attributed to mesopore filling in BENZ and TDMI-BENZ. The high P/Po (∼0.7) condensation step observed for BTSE in the 35 °C benzene data is usually indicative of condensation between grain boundaries in a powder. It is unclear why this sample does not show a step in the adsorption data related to mesoporosity at P/Po ∼ 0.4. The benzene isotherms at 35 °C (Figure 5b) also show that MCM-41 has uptake capacity spanning the entire range of the organosilica materials. This indicates that the silsesquioxane functionality does not impart any significant enhancement of benzene adsorption based on a gravimetric uptake. We can conclude that, at 35 °C where pore filling dominates the isotherm, structural ordering is a more important factor than wall content for adsorption in these materials. Nonetheless, when the benzene isotherms at both temperatures are normalized to surface area (Figure 6) the effect of the organic functionality can be seen. Similar to what was observed with p-xylene, the relative benzene uptakes of DEB and TDMI-DEB are greatly increased. When compared on this basis, all of the organic-functionalized materials have greater benzene uptake at 100 °C, and the highly structurally disordered DEB and DEB-TDMI have comparable or greater benzene uptake per unit surface area than the highly ordered MCM-41 at 35 °C. Retention and Desorption. GC is a rapid technique for determining the adsorption capacity and strength of target analytes on these materials under dynamic adsorption/desorption conditions. Initial studies indicated that the mesoporous materials possessed very strong adsorption for C5-C10 hydrocarbons as Langmuir 2009, 25(21), 12661–12669

Borghard et al.

Article

Figure 7. Log retention time (Tr = min) vs BP of aromatic (( benzene, toluene, o-xylene, 1,2,3-trimethylbenzene), paraffin (0 pentane, heptane, decane, 2,2,4-trimethylpentane), and olefin (Δ pentene, hexane, nonene) probes at 170 °C and 15 mL/min He flow for MCM-41 (A), DEB (B), TDMI-DEB (C), and TDMI-BENZ (D).

Figure 8. Estimated capacity factors (k) of TDMI-DEB, DEB, and MCM-41 for benzene, toluene, and xylene.

evidenced by long retention times, particularly at lower temperatures. Therefore, only short-packed columns were used to reduce retention times to 10-120 min. The unsieved small particle sizes of the materials created a high pressure drop. For comparison, constant flow control was used to maintain a relatively constant flow of 15 mL/min as temperature was increased from 70 °C to approximately 170 °C. The packed columns were conditioned at 180 °C overnight in a flow of helium before obtaining the data. The materials appeared very stable and yielded little or insignificant “bleed” from decomposition during conditioning. Mixed feeds of aromatics, paraffins, and olefins were eluted through packed columns of the test materials at 170 °C isothermal. The capacity factor used normalized all component retention Langmuir 2009, 25(21), 12661–12669

times relative to that of the least retained compound (n-pentane) at 170 °C to compensate for variations (tube lengths, packing particle sizes, etc.) in the packed columns. For convenience, npentane, which elutes relatively quickly, was used instead of methane to approximate the “dead volumes” of the packed tubes in order to calculate k0 . The capacity factor was defined as38 k 0 ¼ TRbtx=½TRbtx -TRpentane where TRbtx ¼ retention time of BTX compounds in minutes TRpentane ¼ retention time of n-pentane at 170 °C DOI: 10.1021/la901334z

12667

Article

Borghard et al.

Figure 9. Desorption of benzene, toluene, o-xylene, and 1,2,3-trimethylbenzene from TDMI-DEB as a function of temperature. Arrows indicate that the retention exceeds 120 min past the last point.

Figure 10. Separation of benzene (1), toluene (2), o-xylene (3), and 1,2,3-trimethylbenzene (4) on GC liquid stationary phases coated with TDMI-DEB (A) and MCM-41 (B).

Figure 7 shows the retention characteristics, plotted as the log of the capacity factor log TR vs boiling point of each compound in the feed mixtures. The results indicate that within a hydrocarbon family (paraffin, olefin, aromatic) there is a relatively smooth functional behavior as encountered for most GC stationary phases. For a given boiling point, there is overlap among the various hydrocarbon family types, indicating that the five materials studied do not offer significant separation selectivity based on thermal desorption of one class of hydrocarbon compounds over another. All of the hydrocarbon probes exhibit very strong retention, particularly as the temperature is lowered when compared to other GC liquid stationary phases. As a result, DEB, TDMIDEB, and MCM-41 were studied in more detail as a function of temperature for benzene, toluene, and o-xylene. Figure 8 shows capacity factor retention characteristics of BTX, and the data indicate that strong retention or adsorption of BTX occurs as the temperature of the materials is lowered; DEB and TDMI-DEB do not show a significant difference in the relative capacity factors for the temperature range investigated; MCM-41 shows a slightly 12668 DOI: 10.1021/la901334z

greater relative retention at all temperatures when compared to the retentions on DEB and TDMI-DEB. As noted previously, the surface-imprinted materials were designed to be specific for TNT, not benzene, toluene, or o-xylene. Recently, we reported that desorption of TNT from these TNT-surface imprinted materials is heavily influenced by pore structure with TNT desorbing more rapidly from TDMI-BENZ than from TDMI-DEB.23 Given the significant contribution of pore structure to desorption and the specific analytes used in this study, it is not possible in this instance to determine what influence, if any, that surface imprinting has on the desorption process. Figure 9 shows a plot of absolute retention times in minutes for BTX and 2,3,4-trimethylbenzene as a function of isothermal temperatures that confirms the strong retention observed in Figure 8. As seen in Figure 9, benzene, the least retained, has a significant retention even at the lowest temperature of 70 °C. The retention of benzene at this temperature is approximately 60 min in the 4 in short tube at 15 mL/min helium flow. This retention is significantly greater than what is normally observed with the retention of benzene on coated GC liquid stationary phases and demonstrates the unique adsorption partitioning of BTX compounds on the organosilica and MCM41 materials. This strong adsorption is a plus for developing a short trap using these materials for concentration purposes. Figure 10 shows a typical example comparing hydrocarbon desorption from the organosilica and MCM-41 materials. One interesting observation when comparing the organosilica materials with MCM-41 is that the organosilica materials yield more symmetrical peaks and better separation. In particular, the symmetrical peak shapes suggest that the organosilica materials have fewer residual adsorption sites to bind with the BTX compounds. The longer retention times and increasing peak width as hydrocarbon boiling point increases are indicative of strong partitioning with the DEB material. In preconcentration, desorption is as important as adsorption since incomplete release of adsorbed analyte will result in decreased capacity and possibly false positive signals due to leaching of residual analyte. Therefore, the improved separation performance for the organosilica materials indicate that they may act as better trapping materials, as it may be less difficult to release adsorbed compounds such as BTX particularly at low concentrations. Similar results have been obtained in HPLC separations using periodic mesoporous organosilica as the stationary phase. Spherical phenylene-bridged Langmuir 2009, 25(21), 12661–12669

Borghard et al.

Article

PMO particles with a narrow population of particle size have been used as a HPLC stationary phase to achieve baseline separations of different mixtures of compounds with similar structure, similar octanol-water partition coefficients, and varying polarity.39

Conclusion To assess the potential of benzene- and diethylbenzenebridged periodic mesoporous materials imprinted for an aromatic molecule to act as preconcentrators for BTX, we compared their benzene and p-xylene adsorptions relative to each other and to the all-inorganic MCM-41 and an ethylenebridged periodic mesoporous organosilica. In the absence of molecular sieving effects, the total surface area normalized uptake should largely reflect the relative strength of the aromatic/sorbent interaction. Of particular interest, the imprinted aromatic functionalized organosilicas showed enhanced uptake relative to ethylene-bridged PMO (BTSE) and non(39) Rebbin, V.; Schmidt, R.; Froba, M. Angew. Chem., Int. Ed. 2006, 45, 5210– 5214.

Langmuir 2009, 25(21), 12661–12669

functionalized MCM-41 on a surface area basis. These results demonstrate the feasibility of using template-directed molecular imprinting to improve preconcentration performance based on achieving enhanced adsorption capacity and sensitivity via surface imprinting. In order to take this approach further, more soluble imprint molecules will need to be investigated to greatly increase the density of surface imprinted sites and thus maximize the absorption capacity. Acknowledgment. The authors would like to acknowledge Dr. Kirk Schmitt for suggesting the testing of these materials as stationary GC phases, and the helpful assistance of Doug Colmyer and Richard Ernst in collecting the adsorption isotherm, XRD, TGA, and SEM data for this report. M. A. Markowitz, M. Zeinali, and B. J. Melde were supported by the Office of Naval Research. Supporting Information Available: Additional information as described in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

DOI: 10.1021/la901334z

12669