Anal. Chem. 2002, 74, 458-467
Development and Characterization of Molecularly Imprinted Sol-Gel Materials for the Selective Detection of DDT Amy L. Graham, Catherine A. Carlson, and Paul L. Edmiston*
Department of Chemistry, The College of Wooster, Wooster, Ohio 44691
Molecularly imprinting sol-gel materials for DDT using both a noncovalent and a covalent approach was examined. A nonpolar porous sol-gel network was created through the use of the bridged polysilsesquioxane, bis(trimethoxysilylethyl)benzene (BTEB), as the principal sol-gel component. Noncovalent molecular imprinting was deemed unsuccessful, presumably because of the lack of strong intermolecular interactions that can be established between the DDT and the sol-gel precursor. A covalent imprinting strategy was employed by generating a sacrificial spacer through the reaction of two 3-isocyanatopropyltriethoxysilanes with one of two different template molecules: 4,4′-ethylenedianiline (EDA) or 4,4′ethylidenebisphenol (EBP). After formation of the solgel, the bonds linking the spacer template to the matrix were cleaved in a manner that generated a pocket of the appropriate size bordered by amine groups that could aid in the binding of DDT through weak hydrogen bonding interactions. Experiments indicated that DDT could be bound selectively by such an approach. To generate a sensor, an environmentally sensitive fluorescent probe, 7-nitrobenz-2-oxa-1,3-diazole, (NBD) located adjacent to the DDT binding site was used to transduce the binding of analyte. EDA-imprinted sol-gels, deposited as films on glass microscope slides, were shown to quantitatively detect DDT in water to a limit-of-detection of 50 ppt with a response time of 90% was observed. ESCA measurements again showed a decrease in the total amount of carbon in the films after spacer removal (Table 2), some of which again corresponded to a decrease in an oxidized species of carbon. Further confirmation of spacer removal was demonstrated by preparing a fluorescent urea-linked spacer silane using acriflavin prepared by the same chemistry used to prepare the EDA-spacer. After LiAlH4 treatment, a complete loss of fluorescence was noted, corresponding to the liberation of the fluorescent acriflavin sacrificial spacer. DDT Sensing Experiments. Figure 8 shows the change in NBD fluorescence intensity for an EDA molecularly imprinted sensing film in the presence of varying concentrations of DDT in
water. The sensing film demonstrated a measurable response down to 10 ppt DDT and a limit of detection that is roughly 50 ppt. The dynamic range is limited, because the fluorescence intensity increase begins to level off at concentrations >1 ppb. Solutions with a concentration >10 ppb could not be tested, because the solubility of DDT in water is only 17 ppb at 25 °C.23 The fluorescence intensity consistently returned to the baseline level following the acetone and water rinses, which demonstrates that response was reversible and that a film can be used repeatedly. The response time was rapid (2000% increase is observed in the fluorescence intensity in acetone, as compared to that in water (data not shown). A second reason for the minimal fluorescence change on binding is likely due to the NBD’s not being located near enough to the binding pocket. IR measurements indicate that only a fraction of the NBD-APTS condenses with the sacrificial spacer silanes. Moreover, the position and orientation of NBD is not controlled. This leaves only a potentially small number of binding sites that have a NBD located ideally to transduce the binding event. A method of directing the location of every NBD fluorophore to be in close proximity to the binding pocket should significantly improve the amount of fluorescence increase that can be achieved. A covalent attachment scheme that replaces the sol-gel condensation method described within that would link the NBD to the sol-gel matrix in an ideal position is currently being investigated in our laboratory. Another limitation is photobleaching, which is inherent in almost all fluorescence-based measurements. In this case, the amount of photobleaching observed was typical of most fluorophores. Using one film as an example, a 12% decrease in (25) Shea, K. J.; Loy, D. A.; Webster, O. W. J. Am. Chem. Soc. 1992, 114, 67006710.
fluorescence intensity was noted after 320 experiments (1600 s total light exposure time). As a result, a calibration scheme must be developed in order to allow for accurate long-term use of a single sensing film. The dynamic range of the sensor is quite limited, which is presumably due to the small number of high affinity binding sites contained within the sol-gel matrix. Preparation of materials with a much greater number of imprinted sites having an adjacent sensing fluorophore may be one method of overcoming this problem. However, too many imprint sites may lead to a loss in structural stability after removal of the sacrificial spacer. Adjusting the size and the chemical functionality of the binding pocket by using different spacers may be an approach for reducing the binding affinity of the imprinted material. This could be another method in extending the dynamic range if the problem of the minimal change in fluorescence could be overcome. The use of LiAlH4 to remove the urea-linked sacrificial spacer proved to be somewhat problematic. Approximately one-half of the EDA imprinted sensing films deposited on Ge ATR crystals were partially stripped from the substrate in the LiAlH4 reflux step, presumably as a result of the abrasive action of alumina particulate matter formed during the process. These films were not used for any measurements. Films deposited on glass substrates were also affected by LiAlH4 treatment. A measurable reduction in fluorescence was observed after template removal, presumably as a result of the chemical degradation of the NBD molecules.
CONCLUSIONS Molecularly imprinting a sol-gel using a sacrificial spacer has been shown to be successful in creating a material that can bind an analyte with a moderate degree of selectivity. Use of an environmentally sensitive fluorescent probe can transduce the presence of a bound molecule. The films produced in this work were imprinted using a sacrificial spacer, which following its removal, generates a site that can selectively bind DDT. The detection limit for DDT was ∼50 ppt DDT in aqueous solution. The greatest limitation with this initial sensing design is the small fluorescence intensity changes observed upon binding of an analyte. It is anticipated that this can be improved by better control of the placement of the NBD near the binding pocket and is an area of research we are actively exploring. ACKNOWLEDGMENT This work was supported by a Camille and Henry Dreyfus Foundation Faculty Start-Up Grant for Undergraduate Institutions. The ESCA measurements were performed by Dr. Andrew Back of Evans PHI Inc.
Received for review June 1, 2001. Accepted October 9, 2001. AC0106142
Analytical Chemistry, Vol. 74, No. 2, January 15, 2002
467
