Molecular Imprinting at Walls of Silica Nanotubes for TNT Recognition

Dec 19, 2007 - Sung Kuk Kim , Jong Min Lim , Tuhin Pradhan , Hyo Sung Jung , Vincent ... Guijian Guan , Suhua Wang , Renyong Liu , and Zhongping Zhang...
0 downloads 0 Views 416KB Size
Download PDF
Anal. Chem. 2008, 80, 437-443

Molecular Imprinting at Walls of Silica Nanotubes for TNT Recognition Chenggen Xie, Bianhua Liu, Zhenyang Wang, Daming Gao, Guijian Guan, and Zhongping Zhang*

Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei, Anhui 230031, China

This paper reports the molecular imprinting at the walls of highly uniform silica nanotubes for the recognition of 2,4,6-trinitrotoluene (TNT). It has been demonstrated that TNT templates were efficiently imprinted into the matrix of silica through the strong acid-base pairing interaction between TNT and 3-aminopropyltriethoxysilane (APTS). TNT-imprinted silica nanotubes were synthesized by the gelation reaction between APTS and tetraethylorthosilicate (TEOS), selectively occurring at the porous walls of APTSmodified alumina membranes. The removal of the original TNT templates leaves the imprinted cavities with covalently anchored amine groups at the cavity walls. A high density of recognition sites with molecular selectivity to the TNT analyte was created at the wall of silica nanotubes. Furthermore, most of these recognition sites are situated at the inside and outside surfaces of tubular walls and in the proximity of the two surfaces due to the ultrathin wall thickness of only 15 nm, providing a better site accessibility and lower mass-transfer resistance. Therefore, greater capacity and faster kinetics of uptaking target species were achieved. The silica nanotube reported herein is an ideal form of material for imprinting various organic or biological molecules toward applications in chemical/biological sensors and bioassay. Synthetic host materials capable of selectively binding target molecules have attracted considerable research interests due to their importance in separation,1 catalysis,2 chemical/biological sensors,3-6 and biomedical materials.5,6 The specific recognition * To whom correspondence should be addressed. E-mail: [email protected]. (1) (a) Wulff, G. Chem. Rev. 2002, 102, 1. (b) Wulff, G. Angew. Chem., Int. Ed. 1995, 34, 1812. (c) Wang, J. P.; Cormack, A. G.; Sherrington, D. C.; Khoshdel, E. Angew. Chem., Int. Ed. 2003, 42, 5336. (2) (a) Wulff, G.; Gross, T.; Scho¨nfeld, R. Angew. Chem., Int. Ed. 1997, 36, 1962. (b) Liu, J. Q.; Wulff, G. J. Am. Chem. Soc. 2004, 126, 7452. (3) (a) Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495. (b) Batra, D.; Shea, K. J. Curr. Opin. Chem. Biol. 2003, 7, 434. (c) Stephenson, C. J.; Shimizu, K. D. Polym. Int. 2007, 56, 482. (d) Sellergren, B. Molecularly Imprinted Polymers. Man-Made Mimics of Antibodies and Their Application in Analytical Chemistry; Elsevier: New York, 2001. (4) (a) Henry, O.; Cullen, D. C.; Piletsky, S. Anal. Bioanal. Chem. 2005, 382, 947. (b) Takeuchi, T.; Dobashi, A.; Kimura, K. Anal. Chem. 2000, 72, 2418. (c) Panasyuk, T. L.; Mirsky, V. M.; Pilesky, S. A.; Wolfbeis, O. S. Anal. Chem. 1999, 71, 4609. (5) Shi, H.; Tsai, W.; Garrison, M. D.; Ferrari, S.; Ratner, B. D. Nature 1999, 398, 593. (6) (a) Hayden, O.; Mann, K. J.; Krassnig, S.; Dickert, F. L. Angew. Chem., Int. Ed. 2006, 45, 2626. (b) Hayden, O.; Dickert, F. L. Adv. Mater. 2001, 13, 1480. (c) Hayden, O.; Lieberzeit, P. A.; Blaas, D.; Dickert, F. L. Adv. Funct. Mater. 2006, 16, 1269. 10.1021/ac701767h CCC: $40.75 Published on Web 12/19/2007

© 2008 American Chemical Society

systems can be spatially organized by imprinting template molecules in a polymeric organic/inorganic matrix through template-monomer complexes such as covalent and noncovalent interactions. The removal of templates from the cross-linked matrix generates recognition cavities complementary to the shape, size, and functionality of the templates. Imprinted polymeric materials with the molecular selectivity approaching that of an enzyme or antibody, however, are intrinsically more robust and stable than biological counterparts, facilitating their applications in harsh environments, such as in the presence of acids or base and organic solvents, and at high temperatures. In particular, the introduction of synthetic design into the molecular imprinting strategy can even make a host element suitable for the analyte for which a natural receptor does not exist. These have been driving the novel development of imprinted materials to the promising use as recognition elements in separation, catalysis, bioassay, sensors, and drug delivery.3,4 The effectiveness of molecular imprinting in the cross-linked polymeric matrix is greatly dependent on the bond nature of the template-monomer complex,7-9 the form of imprinted materials,10-12 and the rigidity of the polymeric matrix.8,9 Although various imprinted materials have been synthesized by different strategies, the imprinted materials ideally suitable for molecular recognition elements have yet to be explored, because of their small binding capacity, slow binding kinetics, and irregular materials shape. In principle, one main challenge to the traditional imprinted materials is that the extraction of original templates located at the interior area of bulk materials is quite difficult due to the high cross-linking nature of imprinted materials,9,13-15 reducing the capacity of rebinding target analyte. Furthermore, if the generated cavities (7) (a) Zimmerman, S. C.; Wendland, M. S.; Rakow, N. A.; Zharov, I.; Suslick, K. S. Nature 2002, 418, 399. (b) Mertz, E.; Zimmerman, S. C. J. Am. Chem. Soc. 2003, 125, 3424. (8) (a) Katz, A.; Davis, M. E. Nature 2000, 403, 286. (b) Bass, J. D.; Katz, A. Chem. Mater. 2003, 15, 2757. (9) Ki, C. D.; Oh, C.; Oh, S.-G.; Chang, J. Y. J. Am. Chem. Soc. 2002, 124, 14838. (10) (a) Yilmaz, E.; Haupt, K.; Mosbach, K. Angew. Chem., Int. Ed. 2000, 39, 2115. (b) Bossi, A.; Piletsky, S. A.; Piletska, E. V.; Righetti, P. G.; Turner, A. P. F. Anal. Chem. 2001, 73, 5281. (11) Schmidt, R. H.; Mosbach, K.; Haupt, K. Adv. Mater. 2004, 16, 719. (12) (a) Sellergren, B.; Ruˇckert, B.; Hall, A. J. Adv. Mater. 2002, 14, 1204. (b) Sulitzky, C.; Rˇ uckert, B.; Hall, A. J.; Lanza, F.; Unger, K.; Sellergren, B. Macromolecules 2002, 35, 79. (c) Titirici, M. M.; Sellergren, B. Chem. Mater. 2006, 18, 1773. (13) (a) Markowitz, M. A.; Kust, P. R.; Deng, G.; Schoen, P. E.; Dordick, J. S.; Clerk, D. S.; Gaber, B. P. Langmuir 2000, 16, 1759. (b) Rao, M. S.; Dave, B. C. J. Am. Chem. Soc. 1998, 120, 13270. (14) Gao, D. M.; Zhang, Z. P.; Wu, M. H.; Xie, C. G.; Guan, G. J.; Wang, D. P. J. Am. Chem. Soc. 2007, 129, 7859.

Analytical Chemistry, Vol. 80, No. 2, January 15, 2008 437

are not at the surface or in the proximity of the materials’ surface, the high resistance to mass transfer will still hinder target species from accessing the deep imprinted cavities, thus slowing the kinetics of binding target analyte.14 Addressing these problems is usually attempted by controlling templates to locate at the surface of materials or in the proximity of the materials’ surface, such as surface imprinting,5,6,10 film/surface-grafted imprinting,11,12 and monomolecular dendritic imprinting.7 Recently, several research groups have begun to explore alternative approaches for developing molecularly imprinted nanomaterials.5,14-17 The molecular imprinting at the nanostructures with high surface-tovolume ratio can provide more complete removal of templates, better site accessibility, lower mass-transfer resistance, and welldefined materials’ shape. Therefore, they will display some clear advantages over traditional forms of bulk materials: larger binding capacity and faster kinetics, as well as easy installation at the surface of the interrogative transducer. Silica has been of considerable interest because it forms the basis of technologically important materials. The ability to functionalize silica materials has prompted a renewed interest in enriching our understanding of its fundamental properties and enhancing its performance in currently existing applications. Here we report an acid-base pairing strategy for the molecular imprinting at the walls of highly uniform silica nanotubes for the recognition of 2,4,6-trinitrotoluene (TNT). It has been demonstrated that TNT templates were efficiently imprinted into the walls of silica nanotubes by the strong acid-base pairing interaction with 3-aminopropyltriethoxysilane (APTS) and the gelation reaction between APTS and tetraethylorthosilicate (TEOS), selectively occurring at the walls of APTS-modified alumina pores. The removal of TNT templates from the tubular walls using acidic solvents leaves the imprinted cavities with covalently anchored aminopropyl groups at the cavity walls. A high density of recognition sites to TNT molecules was created at the walls of highly uniform silica nanotubes. These resultant imprinted sites are highly specific to TNT molecules by the measurements of molecular recognition properties. Because the inside and outside surfaces of silica nanotubes are open to analytes and the thickness of the nanotube wall is only about 15 nm, most of the imprinted sites are at the inside and outside surfaces and in the proximity of surfaces, providing better site accessibility and lower masstransfer resistance. Therefore, the larger capacity and faster kinetics of uptaking target species were achieved. TNT is a highly explosive and environmentally deleterious substance that has been of societal security concern, and therefore the exploration in detection and sensors for ultratrace TNT has attracted considerable research efforts in recent years.18-20 (15) Xie, C. G.; Zhang, Z. P.; Wang, D. P.; Guan, G. J.; Gao, D. M.; Liu, J. H. Anal. Chem. 2006, 78, 8339. (16) (a) Wang, H. J.; Zhou, W. H.; Yin, X. F.; Zhuang, Z. X.; Yang, H. H.; Wang, X. R. J. Am. Chem. Soc. 2006, 128, 15954. (b) Yang, H. H.; Zhang, S. Q.; Tan, F.; Zhuang, Z. X.; Wang, X. R. J. Am. Chem. Soc. 2005, 127, 1378. (c) Li, Y.; Yang, H. H.; You, Q. H.; Zhuang, Z. X.; Wang, X. R. Anal. Chem. 2006, 78, 317. (17) Chronakis, I. S.; Milosevic, B.; Frenot, A.; Ye, L. Macromolecules 2006, 39, 357. (18) (a) Rose, A.; Zhu, Z.; Madigan, C. F.; Swager, T. M.; Bulovic, V. Nature 2005, 434, 876. (b) Andrew, T. L.; Swanger, T. M. J. Am. Chem. Soc. 2007, 129, 7254. (19) Toal, S. J.; Trogler, W. C. J. Mater. Chem. 2006, 16, 2871.

438

Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

EXPERIMENTAL SECTION Materials and Chemicals. The alumina membranes with pore diameter of ∼70 nm were prepared by electrochemical anodization,21 and the aminopropylsilane modification of alumina pore walls was further carried out according to the method reported in our previous work.15 Tetraethylorthosilicate (TEOS, Aldrich) and 3-aminopropyltriethoxysilane (APTS, Aldrich) were used as received. 2,4,6-Trinitrotoluene was supplied by the National Security Department of China and was recrystallized with ethanol before use. 2,4-Dinitrotoluene (DNT) and cetyltrimethylammonium bromide (CTAB) were purchased from Shanghai Chemicals Ltd. Synthesis of TNT-Imprinted Silica Nanotubes. TNTimprinted silica nanotubes were synthesized using the gelation reaction of APTS with TEOS in the pores of alumina membranes. Typically, 0.1 mmol of TNT and 50 µL of APTS were dissolved in 0.2 mL of acetonitrile and stirred at room temperature for 10 min, resulting in the formation of the TNT anions with red color. Under stirring, 4 mL of ethanol, 0.25 mL of TEOS, 50 mg of CTAB, and 0.25 mL of sodium acetate buffer (pH, 5.1) were orderly added into the above solution containing TNT anions and APTS cations. The sol mixture was stirred in ice bath for 30 min to form the sol-gel precursor of silica. Before the preparation of silica nanotubes, the alumina membranes were chemically modified with APTS, leading to the formation of an APTS monolayer at the alumina pore walls. Several pieces of APTS-modified alumina membranes were immersed into the sol-gel precursor for 3 h for a complete filling of membrane pores. After the membranes were taken out, the gelation process was undertaken in an oven of 60 °C for 24 h under nitrogen atmosphere, and then the material was cured in an oven of 150 °C for 6 h. The alumina membranes embedded with silica nanotubes were mechanically polished to remove the silica at the surface of the alumina membranes. The individual silica nanotubes were obtained by dissolving alumina membranes with aqueous phosphoric acid and then washing with distilled water for several times. Nonimprinted silica nanotubes were similarly synthesized within APTS-modified membrane pores in the absence of TNT templates. In order to obtain a more precise comparison with imprinted nanotubes, an equal amount of APTS monomers was also used in the synthesis system of the nonimprinted nanotubes. Meanwhile, the imprinted bulk silica was synthesized using the same sol-gel precursor and reaction procedure as that for the imprinted nanotubes. The imprinted silica monoliths were mechanically ground into small particles with an average diameter of ∼2-3 µm. Measurements of Molecular Recognition Properties. Original TNT templates at the walls of silica nanotubes were extracted with a mixture solvent of ethanol/acetonitrile/HAc (v/v/v, 8:2:1). After removal of the templates, the silica nanotubes were carefully washed with distilled water to remove the residual adsorption of acidic solvent. Steady-state binding capacity of nanotubes to TNT molecules was measured by suspending 1 mg (20) (a) Goldman, E. R.; Medintz, I. L.; Whitley, J. L.; Hayhurst, A.; Clapp, A. R.; Uyeda, H. T.; Deschamps, J. R.; Lassman, M. E.; Mattoussi, H. J. Am. Chem. Soc. 2005, 127, 6744. (b) Medintz, I. L.; Goldman, E. R.; Lassman, M. E.; Hayhurst, A.; Kusterbeck, A. W.; Deschamps, J. R. Anal. Chem. 2005, 77, 365. (21) Masuda, H.; Yamada, H.; Satoh, M.; Asoh, H. Appl. Phys. Lett. 1997, 71, 2770.

Figure 1. Acid-base pairing interaction between TNT and APTS molecules. (A) The evolution of UV-vis absorbance spectra with addition of different amounts of APTS into 20 mL of 1 mM TNT solution (solvent, ethanol/acetonitrile, 8:2). Inset color image shows the corresponding colors of the mixture solutions. (B) Schematic representation for the interaction between TNT and APTS and the formation of TNT anion and APTS cation.

of imprinted nanotubes in 5 mL of the solution with different TNT concentrations (solvent, acetonitrile/ethanol, 2:8). After the samples were incubated on a rocking table at room temperature for 12 h, the silica nanotubes were discarded from the solution phase by centrifugation. The bound amount of TNT in the imprinted nanotubes was determined by measuring the difference between the total TNT amount and the residual amount in the solution phase by a UV spectrometer. Meanwhile, the binding kinetics was tested by monitoring the temporal evolution of TNT concentration in the solution containing the imprinted nanotubes. The binding rates of imprinted nanotube to target analyte were obtained by fitting the time-dependent amount of uptaken TNT molecules. The molecular recognition selectivity of imprinted nanotubes was investigated using the structure-analogous DNT as a compound competitive with TNT. For a comparison, the same measurements were carried out with nonimprinted silica nanotubes and imprinted bulk particles. Characterizations. UV-vis absorbance spectra were measured with a UNIC UV-4802 spectrometer. The morphologies and structures of the nanotubes were examined by an FEI Sirion-200 field emission scanning electron microscope (FESEM) and a JEOL 2010 transmission electron microscope (TEM). The infrared spectra were recorded with a Nicolet Nexus-670 FT-IR spectrometer. RESULTS AND DISCUSSION Of polymeric materials, inorganic silica with a highly rigid matrix and hydrophilic surface has a wide choice of functional precursors and structural forms, which makes it ideally suited for imprinting various organic or biological molecules.8,9 In this work, APTS was used as a functional monomer to imprint TNT molecules because of the strong acid-base pairing interaction between the electron-rich amino group of APTS and the electrondeficient aromatic ring of TNT.14,15,22,23 This interaction between

TNT and APTS has been investigated by the measurement of UVvis spectra. With the addition of APTS into a TNT solution, the UV-vis spectra show that two new absorption peaks appear at ∼525 and 630 nm, respectively, and strengthen continuously with the increase of the APTS amount (Figure 1A). Meanwhile, we can clearly see that the color of the mixture solution changes gradually from colorless into deep red, as shown in the color inset of Figure 1A. These observations of visible absorption clearly demonstrate that a strong molecular interaction occurs between the electron-deficient aromatic ring of TNT and the electron-rich amino group of APTS in the solution system. The TNT molecule is a Bronsted-Lowry acid and can be deprotonated at the methyl group by basic amine. The negative charge on the TNT anion is distributed throughout the molecule through resonance stabilization by three electron-withdrawing nitro groups.22,23 The anionic form of TNT can absorb strongly the visible light,23 leading to the change of solution color which was first reported by Janovsky in 188624 and interpreted later by others.25 As illustrated in Figure 1B, the anion-cation pair of TNT- and RNH3+ was formed in the present solution system. Figure 2 illustrates the imprinting procedure of TNT molecules in the silica matrix through the anion-cation pairs. The imprinting synthesis was simply carried out by means of the gelation reaction of APTS with TEOS in the presence of sodium acetate buffer and CTAB. TNT templates were assembled and immobilized into the matrix of silica in the form of a TNT-APTS ion pair by the crosslinking reaction between APTS and TEOS. After the templates were extracted from the silica matrix by using a mixing acidic solvent to decompose the TNT-APTS ion pair, the TNT-imprinted sites with the covalently anchored amino groups at the cavity walls were created in the silica matrix (Figure 2). These resultant imprinted sites are thus complementary to the shape, size, and chemical functionality of the TNT target analyte. Moreover, the high rigidity of the silica matrix is essentially advantageous to

(22) Walker, N. R.; Linman, M. J.; Timmers, M. M.; Dean, S. L.; Burkett, C. M.; Lloyd, J. A.; Keelor, J. D.; Baughman, B. M.; Edmiston, P. L. Anal. Chim. Acta 2007, 593, 82. (23) Bernasconi, C. F. J. Org. Chem. 1971, 36, 1671.

(24) Janovsky. J. V.; Erb, L. Ber. Dtsch. Chem. Ges. 1886, 18, 2155. (25) (a) Caldin, E. F.; Long, G. Proc. R. Soc. London, Ser. A 1955, 226, 263. (b) Blake, J. A.; Evans, M. J. B.; Russell, K. E. Can. J. Chem. 1966, 44, 119. (c) Shipp, K. G.; Kaplan, L. A. J. Org. Chem. 1966, 31, 857.

Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

439

Figure 2. Schematic illustration for the molecular imprinting mechanism of TNT in a silica matrix through the anion-cation pair and gelation reaction.

Figure 3. Fabrication and formation mechanism of TNT-imprinted silica nanotubes within alumina membrane pores.

the formation of the delicate imprinted sites and to the stabilization of these recognition sites. Here, we used the porous alumina membrane to fabricate the TNT-imprinted silica nanotubes by the gelation reaction of mixing TNT-APTS/TEOS precursors in alumina pores (Figure 3). Before the fabrication of the silica nanotubes, the porous alumina membrane was first modified with APTS, resulting in an APTS monolayer at the alumina pore walls.15 The APTS monolayer plays a critical role in the formation of highly uniform silica nanotubes. First, the APTS-modified pore walls can more easily be wetted by the sol-gel precursors at the initial stage, ensuring that the nanochannels were fully filled. Second, it is particularly important that the APTS monolayer at alumina pore walls contains affirmatively many residual ethoxyl groups (-Si-OC2H5) because the APTS molecule has three ethoxyl groups.26 As illustrated in Figure 3, the unreacted ethoxyl groups of the APTS monolayer at the pore walls will subsequently take part in the condensation with the sol-gel precursors at the early stage of imprinting polymerization. When the gelation reaction is carried out at a slow rate, the APTS monolayer strongly drives the selective and homogeneous occurrence of the gelation reaction along the pore walls during subsequent gelation. The TNT-imprinted silica nanotubes with a uniform wall thickness can thus be produced in the alumina pores. Figure 4A shows a top-view SEM image of a piece of surfacepolished alumina membrane with TNT-imprinted silica nanotubes. A large-area array of silica nanotubes was embedded inside the alumina membrane pores, and all membrane pores were success440

Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

fully filled with silica nanotubes. The nanotube array closely combined together with alumina pore walls, as shown in the inset of Figure 4A. This clearly suggests the selective occurrence of the gelation reaction at the APTS-modified pore walls of the alumina membrane. These silica nanotubes can be liberated from the alumina pores by dissolving the membranes with aqueous phosphoric acid. The TEM image of Figure 4B shows that the individual TNT-imprinted silica nanotubes are highly uniform in diameter (∼70 nm) and wall thickness (∼15 nm) along their whole length. The formation mechanism of TNT-imprinted nanotubes, as described in Figure 3, has been confirmed by the highmagnification TEM observation of individual silica nanotubes. The black arrows in the inset of Figure 4B indicate one nanotube formed in one surface-rough nanochannel of alumina membranes. The physiognomy of the inner surface of the channels was delicately replicated by the silica nanotube, and the wall thickness of the nanotube kept still constant. This detailed observation further confirms that the imprinting polymerization homogeneously occurred along the APTS-modified pore walls due to the direction of the APTS monolayer, leading to the formation of highly uniform silica nanotubes. Although template fabrication in the nanopores of alumina membranes is a common method for the production of nanotubes, the synthesis of highly uniform silica nanotubes was still difficult under normal conditions. Silica with a highly cross-linked rigid matrix is ideally suitable for the formation of delicate recognition sites, but the complete (26) Waddell, T. G.; Leyden, D. E.; DeBello, M. T. J. Am. Chem. Soc. 1981, 103, 5303.

Figure 4. SEM and TEM observations of TNT-imprinted silica nanotubes. (A) Top-view SEM image of the TNT-imprinted silica nanotube array embedded inside an alumina membrane after the membrane surface was mechanically polished. (B) TEM image of individual TNT-imprinted silica nanotubes liberated from the alumina pores by dissolving the alumina membranes. Insets are the highmagnification SEM and TEM images, respectively.

removal of the original templates is very difficult, and thus the chemo/thermo cleavage is often needed.8,9 In the case of the present imprinting procedure, the TNT-APTS ion pairs in the silica matrix can simply be decomposed by the extraction of acidic solvent at room temperature. In particular, the nanotubes with a high surface-to-volume ratio are structurally advantageous for the removal of templates due to the two open surfaces. The hollow structure with a thin wall thickness makes most of templates situated at the surfaces of tubular walls or in the proximity of the inside and outside surfaces. Moreover, like the molecular imprinting in organic polymers, CTAB surfactant may play a role of porogens in the molecular imprinting of silica nanotubes.27 Therefore, the TNT templates were very easily extracted out of the cross-linked silica network. It can clearly be seen that the color of the nanotubes changes gradually from the original pink into colorless with the removal of TNT. In the infrared spectra of silica (27) Yamaguchi, A.; Uejo, F.; Yoda, T.; Uchida, T.; Tanamura, Y.; Yamashita, T.; Teramae, N. Nat. Mater. 2004, 3, 337.

Figure 5. Comparison of molecular recognition properties. (A) The amounts of TNT molecules bound by (9) imprinted nanotubes and (b) nonimprinted nanotubes and the amounts of DNT molecules bound by (4) imprinted nanotubes and (3) nonimprinted nanotubes (the points represent the mean values of three measurements). (B) Schematic drawing for TNT binding at the walls of nonimprinted and imprinted nanotubes.

nanotubes, the characteristic peaks of TNT molecules including the aromatic rings at 1543 cm-1 and nitro groups at 1353 cm-1 disappeared after the solvent extraction. To evaluate the molecular recognition properties of imprinted materials, the nonimprinted silica nanotubes as a control sample were also synthesized under the same chemical conditions as the imprinted nanotubes, but only in the absence of TNT templates. The uptake amounts of TNT by the imprinted/nonimprinted nanotubes were determined by measuring the difference between the total TNT amount and the residual amount in the solution by a UV spectrometer. Figure 5A shows that the amount of binding TNT by imprinted nanotubes increases significantly with the concentrations of TNT in solution. An amount of 1 mg of the TNTimprinted and nonimprinted nanotubes can most uptake ∼250 and 30 nmol of TNT at equilibrium condition, respectively. Namely, the maximum of weight percentage of uptaking TNT is about 5.6% (TNT/nanotube) by imprinted nanotubes and ∼0.7% by nonimprinted nanotubes. Thus, the rebinding capacity of imprinted nanotubes is about 8.0-fold that of nonimprinted nanotubes. From the imprinting procedure used, we should note the fact that both Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

441

nonimprinted and imprinted nanotubes would have the exposed amino groups at their surfaces, and their amounts are nearly equal. As illustrated in Figure 5B, the residual amino groups can also bind TNT molecules through the acid-base pairing interactions. However, the amount of residual amines at the inside and outside surfaces of the nanotubes is very small because the total amount of APTS is much less than that of TEOS in this synthesis system. The low binding capacity of nonimprinted nanotubes to TNT molecules further suggests this assumption reasonable. Therefore, the higher capacity of imprinted nanotubes mainly attributes to the specific binding of imprinted sites to target analytes (Figure 5B). The above comparison and analysis clearly confirm the effectiveness of this imprinting procedure that TNT templates were efficiently combined into the silica matrix through the TNTAPTS ion pairs and were then removed from the walls of silica tubes. A high density of imprinted sites was created at the walls of silica tubes, and a high rebinding capacity was thus achieved. Meanwhile, the effect of molecular imprinting at silica nanotubes was further investigated by testing the molecular recognition selectivity of imprinted nanotubes through comparing the capacities of rebinding TNT and DNT. Figure 5A (9, 4) shows that the rebinding capacity of imprinted nanotubes to TNT are about 5-fold that to DNT at the same condition. However, the nonimprinted nanotubes did not exhibit the obvious difference in the rebinding capacities of TNT and DNT (Figure 5A, b, 3). These results indicate that the imprinted sites show a higher molecular selectivity for TNT than for DNT. On the other hand, Figure 5A also shows that the affinity of imprinted nanotubes to DNT is slightly larger than that of nonimprinted ones. In general, DNT should also be able to fit into the imprinted cavities of TNT, but the aminopropyl groups anchored at the cavity walls have a much larger affinity to TNT than to DNT. Detailed experiments have revealed that there is not any visible absorption band detectable by UV-vis measurement, when adding APTS into DNT solution. This suggests that the DNT molecule with two nitro groups is a much weaker Bronsted-Lowry acid than TNT molecules and likely cannot be deprotonated by the relatively weakly basic amine groups. Thus, the imprinted sites with amine ligands exhibit a very low affinity to DNT molecules.14,15 On the other hand, the better match of imprinted cavities with the shape of TNT molecules may also enhance the affinity of imprinted sites, which is similar to the interaction between an enzyme and substrate. A more exact evaluation on the imprinting effectiveness is to obtain the partition coefficients (K) and selectivity coefficient (R) of imprinted materials. The partition coefficient (K) represents the ratio of the bound amount of target molecules in nanotubes to the residual amount in solution phase under equilibrium conditions, and the selectivity coefficient (R) is the ratio of the partition coefficient of imprinted materials to that of nonimprinted ones.15,28 Here, the analyte concentration of 5 × 10-5 M was chosen to get the partition coefficients because it is in the linear part of the binding curve, as shown in Figure 5A. Table 1 lists the partition coefficients of the imprinted and nonimprinted nanotubes. The K value of imprinted nanotubes for TNT is much greater than that of nonimprinted ones and is also greater than the K value for DNT. The R values of imprinted nanotubes for TNT and DNT are 6.98 and 1.30, respectively. The selectivity coefficient for TNT is about 442 Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

Table 1. Partition Coefficients and Selectivity Coefficients for TNT-Imprinted and Nonimprinted Silica Nanotubes partition coefficient (K)a TNT-imprinted nonimprinted TNT DNT

1988 ( 98 332 ( 20

285 ( 17 255 ( 16

selectivity coefficient (R) R ) Kimp/Knonimp 6.98 1.30

a The solution concentration of TNT/DNT used to obtain the partition coefficients is 5.0 × 10-5 mol L-1.

5.2-fold that for DNT. These above results suggest clearly that the imprinted sites have a high binding capacity and molecular selectivity for TNT target analyte. Moreover, the molecular selectivity is also higher than that of organic polymer nanotubes reported in our previous work.15 This should be attributed to the high rigidity of the silica matrix for the formation of delicate imprinted sites and the stabilization of imprinted sites. The imprinting effects were further investigated by the comparison with traditional imprinted bulk particles. The imprinted bulk particles were synthesized using the same sol-gel precursors as the imprinted nanotubes. Figure 6A (O) shows that 1 mg of the TNT-imprinted bulk particles can only uptake ∼75 nmol of TNT at equilibrium condition. The maximum of uptake capacity of the nanotubes is nearly 3.6 times that of bulk particles. This measurement shows that the density of effective imprinted sites in nanotubes is much higher than that in normal bulk particles. This can roughly be understood as drawn in Figure 6B. In the highly cross-linked rigid matrix of silica, those original templates situated at the central area of bulk particles cannot completely be removed. Moreover, the deep imprinted sites are difficultly accessed by target analyte due to the high mass-transfer resistance. Only these imprinted sites in the proximity of the particle’s surface are effective for rebinding target analyte. These make traditional bulk particles exhibit both lower binding capacity and slower kinetics than the imprinted nanotubes with an ultrathin wall thickness. Figure 7A shows the time-dependent evolution of the TNT amount bound by imprinted nanotubes and bulk particles. Before equilibrium adsorption was reached, the imprinted nanotubes took up TNT molecules from solution phase in a much faster rate than the imprinted bulk particles. The nanotubes took up 50% of the equilibrium amount during only ∼20 min and spent the equilibrium time period shorter than 120 min (Figure 7A, 0). Meanwhile, the imprinted bulk particles needed about 80 min to take up half of the equilibrium amount, and the equilibrium period is longer than 300 min (Figure 7A, O). The fitting lines of binding kinetics in Figure 7B reveal a rapid absorption of TNT molecules into the walls of the silica nanotubes. The binding rates obtained from the fitting results are 3.10 and 0.55 nmol mg-1 min-1 for nanotubes and bulk particles, respectively. Thus, the binding rate of imprinted nanotubes to TNT is 5.6-fold that of imprinted bulk particles. Usually, rapid binding kinetics requires more recognition sites located at the surface or in the proximity of the surface for easy diffusion of target analyte into the recognition sites. In the case of the thin wall thickness of the nanotubes, all imprinted sites (28) Graham, A. L.; Carlson, C. A.; Edmiston, P. L. Anal. Chem. 2002, 74, 458.

Figure 6. Comparison of imprinted nanotubes and traditional bulk particles. (A) The amounts of TNT molecules bound by (0) imprinted nanotubes and (O) imprinted bulk particles (the points represent the mean values of three measurements). (B) Schematic illustration of imprinted sites in traditional bulk particles.

Figure 7. Kinetic uptake to TNT molecules. (A) Temporal evolution of the TNT amounts bound by (0) imprinted nanotubes and (O) imprinted bulk particles. (B) The fitting lines of TNT binding kinetics by (0) imprinted nanotubes and (O) imprinted bulk particles before equilibrium adsorption is reached. The points represent the mean values of three measurements obtained by suspending 1 mg of imprinted materials in 5 mL of 2.0 × 10-4 mol L-1 TNT solution.

are almost situated in the proximity of the inside and outside surfaces of the hollow tubular structure, providing excellent site accessibility and low mass-transfer resistance. Moreover, the target species can also diffuse into these recognition sites through the hollow cores of the nanotubes.29 Therefore, the imprinted silica nanotubes display very fast binding kinetics to target species. CONCLUSIONS In summary, we have developed the molecular imprinting at the walls of highly uniform silica nanotubes for the recognition of TNT molecules. The molecular imprinting through acid-base pairing interaction is simple and effective and creates a high density of delicate recognition sites at the walls of the silica nanotubes, whereby a high binding capacity and molecular selectivity and fast binding kinetics to target species were obtained. The imprinted silica nanotubes as the recognition elements of nanosensors could potentially be exploited in detecting the highly explosive and environmentally deleterious substances. Although the current work is mainly focused on the imprinting of TNT molecules, the experimental results have demonstrated (29) Hou, S.; Wang, J.; Martin, C. R. Nano Lett. 2005, 5, 231.

that the silica nanotube with a rigid matrix and hollow structure is an ideal form of material for the imprinting of various organic or biological molecules by a wide choice of functional precursors. Moreover, the silica nanotubes are highly hydrophilic and can further be functionalized with fluorescent marking molecules, greatly expanding the usefulness of molecularly imprinted materials in the immunoassay and sensors in which a biological antibody might be used. The studies on the fluorophore tagging of imprinted silica nanotubes are in progress at our laboratory. ACKNOWLEDGMENT This work was supported by the National Natural Science Foundation of China (No. 60571038), the National 863 high technology project of China (2007AA10Z434), the National Basic Research Program of China (2006CB300407), the Innovation Project of the Chinese Academy of Sciences (KJCX2-SW-W31), and the Basic Research Program of Anhui (07041420). We also thank the Hundreds Talent Program of the Chinese Academy of Sciences for financial support. Received for review August 21, 2007. Accepted October 31, 2007. AC701767H

Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

443