Innovative Cavitand-Based Sol−Gel Coatings for the Environmental

Jul 19, 2008 - Bellefonte, Pennsylvania 16823, and Sigma Aldrich Italia S.R.L, via Gallarate 154, 20151 Milan, Italy. An innovative and very selective...
0 downloads 0 Views 356KB Size
Download PDF
Anal. Chem. 2008, 80, 6423–6430

Innovative Cavitand-Based Sol-Gel Coatings for the Environmental Monitoring of Benzene and Chlorobenzenes via Solid-Phase Microextraction Federica Bianchi,*,† Monica Mattarozzi,† Paolo Betti,‡ Franco Bisceglie,† Maria Careri,† Alessandro Mangia,† Leonard Sidisky,§ Stefano Ongarato,| and Enrico Dalcanale*,‡ Dipartimento di Chimica Generale ed Inorganica, Chimica Analitica, Chimica Fisica, Universita` degli Studi di Parma, Viale Usberti 17/A, 43100 Parma, Italy, Dipartimento di Chimica Organica e Chimica Industriale and INSTM, Udr Parma, Universita` degli Studi di Parma, Viale Usberti 17/A, 43100 Parma, Italy, Supelco, 595 North Harrison Road, Bellefonte, Pennsylvania 16823, and Sigma Aldrich Italia S.R.L, via Gallarate 154, 20151 Milan, Italy An innovative and very selective solid-phase microextraction coating synthesized by sol-gel technology was developed for the determination of environmental pollutants such as aromatic hydrocarbons at trace levels in air, water, and soil samples. The obtained fibers, composed of quinoxaline-bridged cavitand units, were characterized in terms of film thickness, morphology, thermal stability, and pH resistance. Fibers, characterized by an average thickness of 56 ( 6 µm, exhibited an excellent thermal stability until 400 °C and a very good fiber-to-fiber and batch-to-batch repeatability with RSD lower than 6%. Finally, the capabilities of the developed coating for the selective sampling of aromatic hydrocarbons were proved, obtaining LOD values in the subnanogram per liter range. Extraction efficiency at least 2-fold higher than that obtained using commercial devices was proved for chlorobenzenes sampling in river water, obtaining extraction recoveries ranging from 87.4 ( 2.6% to 94.7 ( 1.9%. The selective desorption of benzene in the presence of high amounts of other airborne pollutants was also demonstrated. The development of new analytical devices characterized by enhanced capabilities for the extraction and preconcentration of organic pollutants at trace levels from liquid, solid, and gaseous samples is a matter of great concern. In this field, the development of new materials is mainly aimed at the achievement of superior selectivity with regard to target analytes or specific classes of compounds. Compared to traditional sample preparation procedures, like liquid-liquid or Soxhlet extraction, solid-phase microextraction (SPME) is a widely used technique for the analysis of volatile and nonvolatile compounds of food, environmental, and * Corresponding authors. Federica Bianchi: e-mail, [email protected]; phone, +39 0521 905446; fax, +39 0521 905556. Enrico Dalcanale: e-mail, [email protected]; phone, +39 0521 905463; fax, +39 0521 905472. † Dipartimento di Chimica Generale ed Inorganica, Chimica Analitica, Chimica Fisica, Universita` degli Studi di Parma. ‡ Dipartimento di Chimica Organica e Chimica Industriale and INSTM, Udr Parma, Universita` degli Studi di Parma. § Supelco. | Sigma Aldrich Italia S.R.L. 10.1021/ac800881g CCC: $40.75  2008 American Chemical Society Published on Web 07/19/2008

biological interest.1–3 Despite the recognized advantages (it is a user-friendly, solvent-free method that can be easily automated or used for field analyses), some drawbacks are related to the physical deposition of the coatings, thus reducing the thermal and chemical stability of commercial devices. The lack of an adequate chemical bonding of the coating with the fiber surface requires the development of innovative procedures able to overcome the mentioned problems. Sol-gel technology has been proposed as an innovative alternative to the commonly utilized grafting procedures owing to its capabilities of producing a great variety of inorganic networks from silicon or metal alkoxide monomer precursors under mild conditions.4–8 Taking into account that the inorganic sol-gel polymerization-type process is extremely versatile, all the parameters involved in the sol-gel reaction, i.e., temperature, type, and concentration of reagents, nature and concentration of catalyst, and solvent, drying methodologies can be properly tuned, thus obtaining materials with the desired analytical properties in terms of texture, specific surface, and pore diameters. Supramolecular receptors can be introduced in the sol-gel process to impart selectivity to SPME fibers, thus producing highly stable and selective materials to be used as trapping devices. The supramolecular approach to analytical chemistry9 is particularly appealing due to the possibility of designing selective receptors as a function of the analytes to be detected. The rational design of highly selective receptors requires a molecular level understanding of the receptor-analyte interactions. Another essential feature is the reversibility of the responses, which requires recourse to weak interactions, since the formation (1) Pawliszyn, J. Solid Phase Microextraction; Wiley-VCH: New York, 1997. (2) Bianchi, F.; Careri, M.; Mangia, A.; Musci, M. J. Chromatogr., A 2006, 1102, 268–272. (3) Bianchi, F.; Careri, M.; Maffini, M.; Mangia, A.; Mucchino, C. J. Anal. At. Spectrom. 2006, 21, 970–973. (4) Liu, M.; Liu, Y.; Zeng, Z.; Peng, T. J. Chromatogr., A 2006, 1108, 149– 157. (5) Shende, C.; Kabir, A.; Townsend, E.; Malik, A. Anal. Chem. 2003, 75, 3518– 3530. (6) Bigham, S.; Medlar, J.; Kabir, A.; Shende, C.; Alli, A.; Malik, A. Anal. Chem. 2002, 74, 752–761. (7) Bianchi, F.; Bisceglie, F.; Careri, M.; Di Berardino, S.; Mangia, A.; Musci, M. J. Chromatogr., A, 2008, 1196–1197, 15-22. (8) Kumar, A.; Gaurav; Malik, A. K.; Tewary, D. K.; Singh, B. Anal. Chim. Acta 2008, 610, 1–14. (9) Anslyn, E. V. J. Org. Chem. 2007, 72, 687–699.

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6423

of covalent or ionic bonds would result in an irreversible saturation of the receptors. By contrast to the widespread use of supramolecular chemistry in sensing10,11 and in analytical separative methodologies,12–15 this approach has not been deeply exploited in extraction techniques. As for SPME, up until now only macrocyclic coatings based on the use of crown ethers,16–18 β-cyclodextrins,19,20 calix[4]arenes,21–23 and molecular imprinting polymers (MIPs)24,25 have been developed for the analysis of environmental, biological, and food matrixes. In this study, for the first time, a quinoxaline-bridged cavitand (QxCav) is proposed as a superior SPME sol-gel coating for the selective extraction of volatile and semivolatile compounds like aromatic hydrocarbons (benzene, toluene, ethylbenzene, xylene, BTEX, and chlorobenzenes) in environmental samples. The choice of QxCav as recognition unit is based on the molecular recognition properties of such a receptor toward aromatic hydrocarbons both in the gas phase26 and in the solid state.27 In both environments aromatic CH-π interactions28 are responsible of the observed complexation, both with the quinoxaline cavity walls29 and with the resorcinarene scaffold.30 These multiple weak interactions, made possible by the complete confinement of the guest within the cavity, render QxCav the receptor of choice for selecting aromatic over aliphatic hydrocarbons. The transfer of these complexation properties at the gas-solid interface has been already proven in gas sensing31 using both mass32,33 and surface plasmon resonance transducers.34 Extraction of micropollutants from water using QxCav in pure form has also been demonstra(10) Lavigne, J. J.; Anslyn, E. V. Angew. Chem., Int. Ed. 2001, 40, 3118–3130. (11) Pirondini, L.; Dalcanale, E. Chem. Soc. Rev. 2007, 36, 695–706. (12) Zhang, L.; Chen, L.; Lu, X.; Wu, C.; Chen, Y. J. Chromatogr., A 1999, 840, 225–233. (13) Panda, S. K.; Schrader, W.; Andersson, J. T. J. Chromatogr., A 2006, 1122, 88–96. (14) Wang, J.; Yuan, Q.; Evans, D. G.; Yang, L.; Zheng, G.; Sun, W. J. Chromatogr., B 2007, 850, 560–563. (15) Yang, W.; Yu, X.; Yu, A.; Chen, H. J. Chromatogr., A 2001, 910, 311–318. (16) Zeng, Z.; Qiu, W.; Huang, Z. Anal. Chem. 2001, 73, 2429–2436. (17) Cai, L.; Gong, S.; Chen, M.; Wu, C. Anal. Chim. Acta 2006, 559, 1–10. (18) Yu, J.; Wu, C.; Xing, J. J. Chromatogr., A 2004, 1036, 101–111. (19) Zhou, J.; Yang, F.; Cha, D.; Zeng, Z.; Xu, Y. Talanta 2007, 73, 870–877. (20) Zhou, J.; Zeng, Z. Anal. Chim. Acta 2006, 556, 400–406. (21) Li, X.; Zeng, Z.; Gao, S.; Li, H. J. Chromatogr., A 2004, 1023, 15–25. (22) Li, X.; Zeng, Z.; Zhou, J.; Gong, S.; Wang, W.; Chen, Y. J. Chromatogr., A 2004, 1041, 1–9. (23) Zhou, X.; Li, X.; Zeng, Z. J. Chromatogr., A 2006, 1104, 359–365. (24) Koster, E. H. M.; Crescenzi, C.; den Hoedt, W.; Ensing, K.; de Jong, G. J. Anal. Chem. 2001, 73, 3140–3145. (25) Hu, X.; Pan, J.; Hu, Y.; Huo, Y.; Li, G. J. Chromatogr., A 2008, 1188, 97– 107. (26) Vincenti, M.; Dalcanale, E.; Soncini, P.; Guglielmetti, G. J. Am. Chem. Soc. 1990, 112, 445–447. (27) Soncini, P.; Bonsignore, S.; Dalcanale, E.; Ugozzoli, F. J. Org. Chem. 1992, 57, 4608–4612. (28) Hunter, C. A.; Lawson, K. R.; Perkins, J.; Urch, C. J. J. Chem. Soc., Perkin Trans. 2 2001, 651–669. (29) Vincenti, M.; Dalcanale, E. J. Chem. Soc., Perkin Trans. 2 1995, 1069– 1076. (30) Bianchi, F.; Pinalli, R.; Ugozzoli, F.; Spera, S.; Careri, M.; Dalcanale, E. New J. Chem. 2003, 27, 502–509. (31) Zampolli, S.; Betti, P.; Elmi, I.; Dalcanale, E. Chem. Commun. 2007, 2790– 2792. (32) Dalcanale, E.; Hartmann, J. Sens. Actuators, B 1995, 24, 39–42. (33) Hartmann, J.; Hauptmann, P.; Levi, S.; Dalcanale, E. Sens. Actuators, B 1996, 35, 154–157. (34) Feresenbet, E. B.; Busi, M.; Ugozzoli, F.; Dalcanale, E.; Shenoy, D. K. Sens. Lett. 2004, 2, 186–193.

6424

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

ted.35,36 In the latter case, selectivity in the inclusion has been attributed to the hydrophobicity of the guest, which prefers cavity inclusion to water solvation. The cavitand-based materials prepared were evaluated in terms of film thickness, porosity, thermal, and pH stability. The capabilities of a selective desorption of benzene in the presence of high amounts of other pollutants in air as well as the superior selectivity with respect to the commonly used commercial fibers were finally demonstrated. EXPERIMENTAL SECTION Chemicals and Materials. Triethoxysilane (95% purity), platinum(0)-1,3-divinyl-1,1,3,3-tetramethyldisiloxane 0.1 M in poly(dimethylsiloxane) (Kurstedt’s catalyst), tetraethoxysilane (TEOS, 98% purity), poly(dimethylsiloxane) hydroxy terminated (OHTSO), poly(methylhydroxysiloxane) (PMHS), octane (98% purity), and nonane (99% purity) were purchased from Sigma-Aldrich (Milan, Italy). Resorcinol (98% purity) were purchased from Acros (Geel, Belgium). Hexane (95% purity), cyclohexane (99% purity), ethanol (99.9% purity), acetone, and toluene (99.5% purity) were purchased from J.T.Baker (Deventer, The Netherlands). Benzene (99.9% purity), ethylbenzene, p-xylene-d10 (used as an internal standard), 1,2,4-trichlorobenzene, and trifluoroacetic acid, TFA, (all 99% purity), o-, m-, p-xylene, chlorobenzene, 1,2-dichlorobenzene, and 2-chlorotoluene (all 98% purity) were purchased from Fluka (St. Gallen, Switzerland). Heptane (>99% purity) and dichloromethane (99.8% purity) were from VWR (Lutterworth, Leicestershire, U.K.) whereas pentane (>99% purity) was from Riedel-de-Hae¨n (Seelze, Germany). SPME bare fused silica fibers with and without assembly, PDMS 7 µm, Carboxen-PDMS 75 µm, PDMS-DVB 65 µm, and DVB/Carboxen/PDMS 2 cm-50/30 µm fibers were purchased from Supelco (Supelco, Bellefonte, PA). QxCav Synthesis. For the synthesis, all solvents were dried over 3 and 4 Å molecular sieves. Resorcinarene (R: C10H18) was prepared according to literature procedure.37 1H NMR spectra were recorded on a Bruker Avance 300 (300 MHz) spectrometer (Bruker, Karlsruhe, Germany), and all chemical shifts (δ) were reported in parts per million (ppm) relative to proton resonances resulting from incomplete deuteration of NMR solvents. The electrospray (ESI) mass spectra were acquired on a Waters ACQUILITY SQD detector equipped with a ESCi multimodeAPCI/ESI-ionization (Waters, Milford, MA). Column chromatography was performed using silica gel 60 (Merck 70-230 mesh). Flash chromatography was performed using the Versaflash System (Supelco). Cavitand 1. Cavitand 1 (Figure 1A) was prepared according to a published procedure.30 Cavitands 2 and 3. To a solution of cavitand 1 (680 mg, 0.440 mmol) in dry cyclohexane, under argon, triethoxysilane (0.487 mL, 2.64 mmol) and platinum(0)-1,3-divinyl-1,1,3,3-tetramethyldisiloxane complex (Karstedt’s catalyst, 0.088 mL, 0.1 M in poly(dimethylsiloxane)vinyl terminated) were added. The solution was stirred at 68 °C for 24 h. The solvent was removed under vacuum, and the crude product was purified by column chroma(35) Dalcanale, E.; Costantini, G.; Soncini, P. J. Inclusion Phenom. Mol. Recognit. Chem. 1992, 13, 87–92. (36) Ferrari, M.; Ferrari, V.; Marioli, D.; Taroni, A.; Suman, M.; Dalcanale, E. IEEE Trans. Instrum. Meas. 2006, 55, 828–834.

Figure 1. Schematic illustration of the steps involved in the coating process: (A) QxCav hydrosilylation; (B) sol-gel coating preparation with plasticizer; and (C) sol-gel coating preparation without plasticizer.

tography on silica gel by using hexane/ethanol (9:1 v/v) as eluant to give a mixture of trifunctionalized 2 and difunctionalized 3 cavitands (Figure 1A) as a yellow solid (480 mg, about 54%). Fiber Preparation. The following fiber preparation procedures were developed in our laboratory. QxCav-Based Fiber with Plasticizer (Figure 1B). The same procedure used by Zeng and co-workers was applied.38 Briefly, the sol solution was prepared by mixing TEOS (100 µL, 0.448 (37) van Velzen, U. E. T.; Engbergsen, J. F. J.; Reinhoudt, D. N. Synthesis 1995, 989–997.

mmol), 85 µL of OH-TSO (used as plasticizer), 10 µL of PMHS (used as end-capper), and 19 µL of TFA (containing 5% of water) with the mixture of cavitands 2 and 3 (21 mg, 0.010 mmol) in 200 µL of CH2Cl2. The mixture was stirred for 3 min, then centrifugated at 12 000 rpm for 5 min. The resulting solution was used for the dip coating. QxCav-Based Fiber without Plasticizer (Figure 1C). The sol solution was prepared by mixing TEOS (75 µL, 0.336 mmol), 5 µL of water, and 5 µL of TFA with the mixture of cavitands 2 and 3 (70 mg, 0.034 mmol) in 250 µL of CH2Cl2. The solution was Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6425

stirred for 3 min, then centrifugated at 12 000 rpm for 5 min. The resulting solution was used for the dip coating. Before the coating process, the SPME fibers were activated by immersion in 40% HF (v/v)7,39 for 1 min and rinsed with distilled water in order to quickly expose a great number of silanol groups for the further immersion in the prepared coating solution. The resulting sol solutions proved to be stable preserving a proper viscosity for about 1 h, thus allowing one to perform the dipping procedure in a reproducible way. The coating was then obtained by vertically dipping the fiber in the sol solution for about 30 s. After the fiber was dried for some minutes at 40 °C, the procedure was repeated from 3 to 10 times. Fiber Characterization. Thermogravimetric analysis (TGA) was performed using a TGA 7 instrument (Perkin-Elmer, Walthan, MA) over the temperature range 30-400 °C (heating rate: 5 °C min-1) under inert (N2) atmosphere. Coating thickness and surface morphology were investigated by using scanning electron microscopy (SEM) with a Leica 430i instrument (Leica, Solms, Germany). Elemental analysis was performed on a CHNS-O EA1108 (Carlo Erba, Milan, Italy) elemental analyzer. 29Si and 13 C solid-state NMR spectra were recorded on a Bruker Avance 400 WB 2-channel solid-state spectrometer. Fiber bleeding was investigated by desorbing the fibers in the GC injection port for 2 min at 100 and 300 °C, respectively. pH resistance was evaluated by sampling 100 ng L-1 of chlorobenzene in water at pH 2, pH 7, and pH 10. Ten replicated measurements for each pH value were performed. Fiber-to-fiber and batch-to-batch repeatability were evaluated both for headspace and for immersion analysis by preparing five fibers for each case. A mixture of benzene and p-xylene was analyzed in the case of headspace analysis, whereas 1,2,4-trichlorobenzene was investigated in the case of immersion analyses. Eight replicated measurements for each fiber were always performed. The same experiments were carried out by using the commercial Carboxen-PDMS 75 µm fibers (Supelco). SPME Analysis. All the SPME experiments were performed by using a manual device. Prior to use, all the fibers were conditioned in the GC injection port at 310 °C for 2 h under a helium flow. Aliphatic hydrocarbons, BTEX, and chlorobenzenes sampling was performed by exposing the QxCav in the headspace of spiked aqueous solutions (10 mL) for 30 min at 50 °C. A constant magnetic stirring was always applied. The same procedure was applied using the PDMS 7 µm, Carboxen-PDMS 75 µm, and DVB/Carboxen/PDMS 2 cm-50/30 µm (Supelco). GC/MS Analysis. A HP 6890 Series Plus gas chromatograph (Agilent Technologies, Palo Alto, CA) equipped with a MSD 5973 mass spectrometer (Agilent Technologies) was used. Helium was used as the carrier gas at a constant flow rate of 1 mL min-1; the gas chromatograph was operated in splitless mode for 1 min with the PTV injector (Agilent Technologies) maintained at the temperature of 250 °C and equipped with a 1.5 mm i.d. multibaffled liner (Agilent Technologies). Chromatographic separation was performed on a 30 m × 0.25 mm, df 0.25 µm SLB-5 ms capillary column (Supelco). The transfer line and source were maintained at the temperatures of 280 and 150 °C, respectively. Preliminarily, full scan EI data were acquired to determine appropriate masses (38) Li, X.; Zeng, Z.; Zhou, J. Anal. Chim. Acta 2004, 509, 27–37. (39) Kang, J. K.; Musgrave, C. B. J. Chem. Phys. 2002, 116, 275–280.

6426

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

for selected-ion monitoring mode (Table S-1 in the Supporting Information) under the following conditions: ionization energy, 70 eV; mass range, 35-350 amu; scan time, 3 scan s-1; electron multiplier voltage, 1600 V. Signal acquisition and data handling were performed using the HP Chemstation (Agilent Technologies). Method Validation. Method validation was performed according to EURACHEM guidelines40 following the same procedure reported in a previous study.7 Surface water taken from a mountain river was used as blank for validation purposes. RESULTS AND DISCUSSION QxCav Hydrosilylation. To succeed in the aim of developing new sol-gel coatings for SPME, the quinoxaline cavitand required the introduction of a suitable substituent at the lower rim. Among the available sol-gel precursors, TEOS was chosen because of both the mild polymerization conditions and of the regioselective introduction on the terminal double bond of the alkyl chains (R) of the QxCav via hydrosilylation (Figure 1A). In order to obtain the highest number of silylated groups at the lower rim, a new synthethic procedure was developed and optimized, thus requiring a proper selection both of the catalyst and of the reaction solvent. In contrast to a previous study,30 in which both H2PtCl6 and Wilkinson’s catalyst were used, the Karstedt’s catalyst was selected for the hydrosilylation reaction owing to its capability of reducing the occurrence of internal isomerization of the terminal olefins, thus obtaining higher amounts of polysilylated derivatives. As for the reaction solvent, toluene, which is the solvent commonly used for hydrosilylation via Karstedt’s catalyst, had to be discharged as a consequence of its possible inclusion in the cavity. Cyclohexane was used as a valid alternative owing both to the possibility of an easy removal under reduced pressure and to the low complexation within the QxCav cavity. By operation under the developed conditions, a noticeable improvement of the hydrosilylation reaction with respect to the use of the Wilkinson’s catalyst was achieved: the presence of both the di- and trifunctionalized products was proved by ESI-MS spectra, with the trisilylated being the predominant product, as determined by 1H NMR spectra (Table S-2). Sol-Gel Coating Preparation. Different reaction pathways were carried out: the first approach followed a protocol used by Zeng and co-worker38 based on the use of two silicon polymers as plasticizers in order to obtain a flexible coating and avoid cracking phenomena. The optimization of this sol-gel process mainly in terms of choice of the solvent and ratio among plasticizer/matrix precursor/cavitand was a critical aspect. As for the solvent, taking into account the insolubility of the synthesized cavitand in polar solvents, CH2Cl2 was proposed as an alternative solvent due to its ability of dissolving QxCav and to its low boiling point, necessary to guarantee a rapid drying of the sol solution. Regarding the ratio among the sol-gel reagents, several experiments were carried out to maximize the concentration of the cavitand in the coating while maintaining a reasonable gelation time to allow an adequate and reproducible dip-coating of the fiber. The QxCav, being mainly trisilylated, is able to act like a cross(40) The Fitness for Purpose of Analytical Methods: A Laboratory Guide to Method Validation and Related Topics, EURACHEM Guide, 1st English ed.; LGC Ltd.: Teddington, U.K., 1998; http://www.eurachem.ul.pt/.

Figure 2. Scanning electron micrographies of QxCav-based coating: (A) with plasticizer and (B) without plasticizer.

linking agent, thus producing, at high concentration levels, extremely rapid gelation times. All the hydroxyl groups present in hydrolysis products allowed the silica network to grow by chemical binding with the three-dimensional network, thus being chemically anchored to the polymeric network. After exposure of the silica fiber to the sol solution, the activated silanol groups on the fiber surface were chemically bonded in the polymeric network also, thus obtaining the desired coating. A schematic illustration of the steps involved in the process is represented in

Figure 1B, where the cavitands are multilayer distributed in a variety of orientations. The morphology of the obtained fibers was investigated by scanning electron microscopy under different magnifications (Figure 2A), revealing a homogeneous and nonporous coating on the entire surface of the fiber. The average thickness of the developed coating was found to be 10 ± 2 µm (n ) 5): further experiments, carried out by increasing the number of dipping did not allow one to obtain higher thicknesses owing to the too high viscosity of the sol-solution. As a consequence, Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6427

coatings thinner than those of commercially available fibers (with the exception of PDMS 7 µm) were obtained. However, the main drawback of the developed SPME coating was related to the reduced amount (10-15%) of QxCav in a silica network containing silicon oil as an additional reagent. Under these conditions, the highly specific molecular recognition properties of the quinoxaline receptor could be diluted by the unspecific ones with the bulk of the matrix. On the basis of these considerations, the challenge was the development of materials characterized by high concentrations of the quinoxaline molecular receptors in a “neutral” matrix, thus minimizing the presence of unspecific moieties like the Si-CH3 groups. Therefore only TEOS and QxCav were the precursors used for the purpose (Figure 1C). The main problem of this approach was to find the optimal reaction conditions to avoid cracking of the network. The absence of a plasticizer in a rigid network such as silica could make the network not suitable to host big molecules such as quinoxalinebridged cavitands. Solvent selection as well as the ratio among the utilized reagents needed to be optimized. Preliminary experiments carried out by using acetone (a good solvent for all the reagents) produced a highly porous and spongelike material, but the main drawback was related to the inhomogeneous distribution on the silica fiber (Figure S-1). Again, the capabilities of CH2Cl2 were evaluated, but the immiscibility with water was the crucial aspect, thus leading to the impossibility of controlling the stoichiometric ratio among the reagents.With the use of an excess of water to ensure the complete hydrolysis of siloxane, a phase separation was obtained, thus requiring the removal of the aqueous layer before the dip-coating in the organic phase. SEM images showed a high porous, spongelike material uniformly distributed on the whole surface of the fiber, but unfortunately, a poor adhesion onto the silica surface was demonstrated (Figure S-2). By a subsequent optimization of the reaction conditions mainly in terms of the amount of catalyst, a very uniform and homogeneous coating was obtained with an average thickness of 56 ± 6 µm (n ) 5) (Figure 2B), 5 times higher than those achieved using the plasticizer-based protocol. The presence of anchored cavitands was confirmed by 29Si and 13C solid-state NMR (Figures S-3 and S-4). Characterization of QxCav Coated Fibers. Taking into account that thermal resistance is a very important parameter for SPME applications in order to obtain a complete desorption of the analytes from the fiber without carry-over effects, the thermal stability of the obtained gels was studied by means of TGA. A very good stability from room temperature to 400 °C with a weight loss lower than 8% was observed. The thermal capabilities of the coatings were also evaluated by conditioning the fibers in the GC injector port under different temperatures: by desorbing the fibers at 100 and 300 °C no significant bleeding was observed, thus confirming the high thermal stability of the QxCav coatings. pH influence was also assessed since it is known that conformational switching between the vase and kite forms of quinoxalinebridged cavitands can be obtained at different pH values.41,42 pH resistance was proved by using the developed fibers for the sampling of chlorobenzene (immersion analysis) in aqueous solutions at different pH for 20 min at 30 °C. More precisely, each pH value was obtained using proper buffer solutions and maintaining the ionic strength constant. ANOVA did not show significant 6428

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

Table 1. Fiber-to-Fiber and Batch-to-Batch Repeatability Using the QxCav and Carboxen-PDMS 75 µm Fiber repeatability (n ) 8) fiber-to-fiber (RSD %) QxCav fiber

CarboxenPDMS

batch-to-batch (RSD %) QxCav fiber

benzene p-xylene

Headspace Analysis 3.5 8.4 4.2 8.8

3.8 5.4

1,2,4-trichlorobenzene

Immersion Analysis 2.7 8.5

4.9

differences (p > 0.05) among the obtained mean responses (n ) 10 for each pH value), thus assessing the capabilities of the developed coatings for the sampling of solutions under different pH conditions. The performance of the developed fibers was also evaluated in terms of fiber-to-fiber (by dipping different fibers in the same sol) and batch-to-batch (by dipping different fibers in different sol) repeatability. As shown in Table 1, the QxCav fiber allowed to obtain RSD lower than 6% both for headspace and for immersion analyses, whereas RSD values higher than 8% were obtained using the Carboxen-PDMS fiber. The obtained results assessed the feasibility of the proposed coating procedure in the development of highly chemically and thermally stable coatings. Extraction Capabilities of the QxCav-Based Coatings. The host-guest interactions of the quinoxaline-based fibers, both with and without plasticizer, were exploited for the selective sampling of aromatic hydrocarbons at trace levels in environmental samples. Concentration levels of aliphatic compounds 3 times higher than those of when aromatic guests were used. More precisely, the aim was to make the most of the interactions among QxCav and aromatic guests in order to achieve a superior selectivity during the sampling, then acting on the desorption temperatures to obtain a further gain especially for the selective detection of benzene. The performance of the fibers obtained using the plasticizer as reagent was not satisfying. As expected, the QxCav coating showed a reduced adsorption of the investigated compounds with respect to the commonly used DVB/Carboxen/PDMS and the Carboxen-PDMS fibers, whereas a behavior similar to that of the PDMS 7 µm was observed. Taking into account that the goal of the work was the development of new materials able to selectively desorb benzene in the presence of other aromatic and aliphatic pollutants, the selectivity of the coating resulted to be compromised by the plasticizer. As shown by data depicted in Figure 3A and obtained by performing consecutive desorption steps in the 50-250 °C range, the lack of selectivity of the PDMS coating was demonstrated, being all the investigated analytes simultaneously desorbed already at low temperatures. As for the QxCav-based fiber with plasticizer (Figure 3B), the selective detection of benzene was interfered with by the presence of other aliphatic and aromatic compounds owing to unspecific interactions with the siloxane matrix, especially with the Si-CH3 moiety of the (41) Skinner, P. J.; Cheetham, A. G.; Beeby, A.; Gramlich, V.; Diederich, F. Helv. Chim. Acta 2001, 84, 2146–2153. (42) Lagugne´-Labarthet, F.; An, Y. Q.; Yu, T.; Shen, Y. R.; Dalcanale, E.; Shenoy, D. K. Langmuir 2005, 21, 7066–7070.

Figure 4. Performances of the QxCav fiber without plasticizer vs the PDMS-DVB fiber (aqueous solution of benzene, toluene, and chlorobenzenes: 5 ng L-1 each). Table 2. LOD Values Obtained with the QxCav and the PDMS-DVB 65 µm Coatings LOD (ngL-1) benzene toluene chlorobenzene 2-chlorotoluene 1,2-dichlorobenzene 1,2,4-trichlorobenzene

Figure 3. Performances of (A) PDMS 7 µm; (B) QxCav fiber with plasticizer; (C) QxCav fiber without plasticizer (BTEX, 50 ng L-1 each; aliphatic compounds, 150 ng L-1 each). Subsequent desorption steps in the 50-250 °C range.

substrate during the sampling. However, it can be observed that by performing subsequent desorption experiments, the other aromatic compounds were mostly released at higher temperatures owing to the interactions with the quinoxaline cavity, absent in the PDMS 7 µm phase. As for the extraction mechanisms, owing to the predominant presence of the PDMS phase with respect to the QxCav, the prevalent mechanism for the “plasticizer” coating is absorption, being the analytes mainly extracted by partitioning into a “liquidlike” phase. However, the interactions with the Qx cavities are not to be disregarded, being the QxCav is able to selectively interact with the aromatic guests. By contrast, excellent results in terms of a selective desorption of benzene were obtained using the fibers without plasticizer (Figure 3C). Aliphatic compounds were not retained by QxCav due to the lack of unspecific interactions with the matrix support. In addition, the use of low desorption temperatures allowed one to selectively desorb benzene, while retaining toluene and the other aromatic compounds into the cavity as a consequence of the strongest conditions required for their release. With the new coating, GC responses at least 2-fold higher than those achieved using the fiber with the plasticizer were obtained. This behavior could be ascribed both to the larger surface of the coating and to the greater amount of immobilized cavitand, suggesting the use of the new fibers for the detection of benzene

QxCav

PDMS-DVB

0.28 0.12 0.02 0.02 0.01 0.004

0.55 0.58 0.055 0.046 0.033 0.011

at trace levels in environmental samples. In this case an adsorptive mechanism is involved due to the high interconnectivity conferred by the TEOS and promoted by densification of the gel at 300 °C. Under these conditions, most probably a relatively dense, highly cross-linked, and less permeable network results with the larger surface and the presence of a great amount of immobilized cavitand contributing to the interaction with the analytes. The complexation capabilities of QxCav were also exploited toward the sampling of chlorobenzenes in water at trace levels. Aqueous solutions spiked with 5 ng L-1 of each compound were analyzed by performing three replicated measurements. Because of the presence of electron withdrawing substituents on the aromatic guests able to strengthen CH-π interactions with the quinoxaline cavity,29 a superior extraction efficiency with respect to commercial fibers (Figure 4) was observed obtaining chromatographic responses 2/3-fold higher than those achieved using commercial devices. It should also to be noticed that an increase of the selectivity corresponds with the presence of a high number of chlorine atoms, thus confirming that guest hydrophobicity is the driving force for complexation at the solid-water interface.35 In order to assess the capabilities of the QxCav-coating, a method for the determination of chlorobenzenes was finally validated and applied for the quantification of these compounds in water. Excellent results were obtained with LOD values in the subnanogram per liter range (Table 2), thus proving the potential of the method for the determination of these compounds at ultratrace levels and satisfying the restrictive quality criteria for surface water established for the year 2015.43 The achieved LOD values were lower than those reported in previous studies,44–46 (43) D.M. no. 367 06/11/2003 Official Italian Gazette no. 5-08/01/2004. (44) Almeida, C. M. M.; Vilas Boas, L. J. Environ. Monit. 2004, 6, 80–88. (45) Vidal, L.; Canals, A.; Kalogerakis, N.; Psillakis, E. J. Chromatogr., A 2005, 1089, 25–30.

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6429

thus demonstrating the enhanced performance of the developed coating with respect to other devices. The removal of the plasticizer resulted to highly improve selectivity by minimizing aspecific interactions with the coated materials and increasing the number of available binding sites per volume. Good linearity was proved in the 0.1-50 ng L-1 range for all the analytes by applying Mandel’s fitting test. Method precision was evaluated testing two concentration levels, i.e., at 1 and 20 ng L-1. Good results were obtained both in terms of intraday repeatability and intermediate precision: RSD values lower than 4% at the highest concentration and lower than 5% at the lowest one were calculated for intraday repeatability, whereas between-day precision was evaluated verifying homoscedasticity and performing ANOVA on the data acquired over 3 days. ANOVA showed that mean values were not significantly different among the 3 days obtaining p values > 0.05. RSD lower than 6% at both concentration levels were calculated. Extraction recoveries ranging from 87.4 ± 2.6% to 94.7 ± 1.9% (n ) 3) were calculated at 5 ng L-1, thus showing the good efficiency of the developed method in terms of extraction recovery as well as of precision. The developed method was finally applied for the analysis of surface water samples (river water). Four samples out of six showed chlorobenzene levels ranging from 0.55 (±0.02) to 20 (±0.4) ng L-1, thus demonstrating the capabilities of the method for the analysis of ultratrace levels. Finally, it can be stated that proposed sol-gel procedure can be really advantageous also in terms of fiber lifetime, allowing the use of the developed coating for more than 150 times, with a negligible loss of efficiency. The performance of the proposed QxCav coating has been validated in the analysis of a complex matrix of soil. By contrast to previous studies in which LOD values in the nanogram per gram range were obtained,47,48 the excellent performances of the developed coating allowed one to achieve detection limits in the nanogram per kilogram range for all the investigated aromatic compounds (Table S-3), thus also confirming the remarkable selectivity toward chlorinated aromatic hydrocarbons coupled to (46) He, Y.; Wang, Y.; Lee, H. K. J. Chromatogr., A 2000, 874, 149–154. (47) Li, X.; Zeng, Z.; Xu, Y. Anal. Bioanal. Chem. 2006, 384, 1428–1437. (48) Llompart, M.; Li, K.; Fingas, M. Talanta 1999, 48, 451–459.

6430

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

the absence of aliphatic hydrocarbons uptake in the analysis of a complex matrix like soil samples. CONCLUSIONS A novel SPME gel-coating based on QxCav as a supramolecular receptor was developed and proposed as a valid alternative to the commercial fibers for the selective determination of benzene and chlorobenzenes at ultratrace levels in environmental air and water samples. Compared to commercially available fibers, the main features were the excellent thermal (400 °C) and chemical stability, a very good fiber-to-fiber and batch-to-batch repeatability, and the possibility of a selective desorption of benzene in the presence of high amounts both of aliphatic and other aromatic hydrocarbons. The efficacy of the developed coating was proven also at the solid-liquid interface by sampling chlorobenzenes in water, obtaining GC responses at least 2-fold higher than those obtained using commercial devices. An excellent selectivity and sensitivity toward chlorine-substituted compounds was also proved in complex matrixes like soil. This remarkable analytical result was achieved by embedding the right receptor in the appropriate material. None of the two items taken alone would have achieved the desired analytical performance. This integrated approach to analytical sampling materials can be seen as a general strategy to boost analytical performances by imparting selectivity to analytical techniques from one side and by tailoring selectivity of the adsorbed material toward the desired class of analytes from the other side. SUPPORTING INFORMATION AVAILABLE Analytical and spectral characterizations of the developed coating like 1H NMR, ESI, 29Si and 13C solid-state NMR spectra, and scanning electron micrographies to better support the characterization of the developed materials and a detailed description of the GC/MS experimental conditions used in this work. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review April 30, 2008. Accepted June 18, 2008. AC800881G