Anal. Chem. 1997, 69, 3109-3114
A Triazine Herbicide Minisensor Based on Surface-Stabilized Bilayer Lipid Membranes Christina G. Siontorou and Dimitrios P. Nikolelis*
Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, Panepistimiopolis-Kouponia, 15771 Athens, Greece Ulrich J. Krull and Kuang-Lee Chiang
Chemical Sensors Group, Erindale Campus, University of Toronto, 3359 Mississauga Road North, Mississauga, ON L5L 1C6, Canada
This work describes an electrochemical technique that is suitable for rapid and sensitive screening of the triazine herbicides simazine, atrazine, and propazine. Egg phosphatidylcholine and dipalmitoylphosphatidic acid (DPPA) were used for the formation of self-assembled bilayer lipid membranes supported on silver wire (s-BLMs). Evidence that BLMs could form on silver wires was collected by means of ellipsometry which was done to investigate samples consisting of lipids deposited on planar reflective silver films. The interactions of triazines with s-BLMs produced electrochemical ion current increases which reproducibly appeared within ∼10 s after exposure of the lipid membranes to the herbicides. The sensitivity of the response was maximized by use of BLMs composed of 35% (w/w) DPPA and by alteration of the phase distribution within membranes by the introduction of 1.0 mM calcium ions in bulk solution. The mechanism of signal generation could be a result of rapid adsorption of the triazine on the surface of s-BLMs with a consequent rapid reorganization of the electrostatics of the membrane. The magnitude of the current signal was linearly related to the herbicide concentration, which could be determined at the nanomolar level. The present triazine minisensor exhibited good mechanical stability and longevity (routinely over 48 h), reproducible response characteristics (i.e., sensitivity and response to a given concentration of triazine in solution), fast response times, and low detection limits. The sensor can be simply and reliably fabricated at low cost. Studies have shown high selectivity for triazines in the presence of insecticides and pesticides. Triazine herbicides are of major importance in modern agriculture and have been used in large quantities. The most commonly used herbicide is atrazine, whereas other frequently used triazines are simazine and propazine.1 The triazines are extensively distributed in aquatic environments, and triazine pollution in rivers, lakes, and seas has become a serious problem.2,3 A number of studies have explored the mammalian toxicity of triazines, which were proven to be toxic at high-level (1) Gianessi,I. P. A National Pesticide Usage Data Base; Resources for the Future: Washington, DC, 1986; pp 1-14. (2) Agricultural Chemicals in Ground Water; Proposed Pesticide Strategy; U.S. Environmental Protection Agency; Washington, DC, 1987; pp 1-150. S0003-2700(97)00113-3 CCC: $14.00
© 1997 American Chemical Society
doses.4 Liquid and gas chromatographic (LC and GC, respectively) procedures currently used for the determination of atrazine and other triazine herbicides5,6 are not suitable as rapid screening methods, and are not easily employed in the field, due to limitations of discontinuous analysis using chromatographic procedures and size and cost of GC and LC instrumentation. Therefore, there is a demand for development of low-cost analytical devices for the rapid screening (in a single format) of triazine herbicides at low concentration levels in the field and in public water supplies.7,8 To overcome these problems, a number of biosensors based on molecular recognition elements such as photosynthetic systems9-11 or antibodies12 have been developed for the detection of triazine herbicides. Biosensors based on antibody/antigen interactions are more sensitive and selective7,13-15 in comparison to those based on photosynthetic systems, with detection limits of ∼1 ppb. The analytical utility of bilayer lipid membranes (BLMs) as one-shot electrochemical biosensors (bioprobes) for the rapid screening of organophosphate and carbamate insecticides such as monocrotofos and carbofuran, respectively, was previously described.16 An electrochemical investigation of interactions of atrazine with planar “freely suspended” BLMs has recently (3) Fielding, M.; Barcelo, D.; Helweg, A.; Galassi, S.; Torstenson, L.; van Zoonen, P.; Wolter, R.; G. Angeletti, G.: Pesticides in Ground and Drinking Water. In Water Pollution Report 27; Commission of the European Communities: Brussels, Belgium, 1989; pp 16-34. (4) Brusick, D. J. Mutat. Res. 1994, 317, 133-144. (5) Sherma, J. Anal. Chem. 1993, 65, 40R-54R. (6) Barcelo, D. J. Chromatogr. 1993, 643, 117-143. (7) Franek, M.; Kolar, V.; Eremin, S. A. Anal. Chim. Acta 1995, 311, 349356. (8) National Survey of Pesticides in Drinking Water Wells, Phase II Report, EPA 570/9-91-020, U.S. Environmental Protection Agency, National Technical Information Service, Springfield, VA, 1992. (9) McArdle, F. A.; Persaud, K. C. Analyst 1993, 118, 419-423. (10) Jockers, R.; Bier, F. F.; Schmid, R. D.; Wachtveitl, J.; Oesterhelt, D. Anal. Chim. Acta 1993, 274, 185-190. (11) Brewster, J. D.; Lightfield, A. R.; Bermel, P. L. Anal. Chem. 1995, 67, 12961299. (12) Wust, S.; Hock, B. Anal. Lett. 1992, 25 (6), 1025-1037. (13) Bier, F. F.; Jockers, R.; Schmid, R. D. Analyst 1994, 119, 437-441. (14) Wortberg, M.; Middendorf, C.; Katerkamp, A.; Rump, T.; Krause, J.; Cammann, K. Anal. Chim. Acta 1994, 289, 177-186. (15) Tom-Moy, M.; Baer, R. L.; Spira-Solomon, D.; Doherty, T. P. Anal. Chem. 1995, 67, 1510-1516. (16) Nikolelis, D. P.; Krull, U. J. Anal. Chim. Acta 1994, 288, 187-192.
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997 3109
appeared in the literature.17 These interactions were transduced by BLMs to provide a transient current signal as a single event within 1 min after exposure of the membrane to atrazine. The magnitude of the transient current signal was linearly related to the concentration of atrazine in bulk solution and provided submicromolar detection limits. However, planar freely suspended BLMs suffer from physical instability and are prone to failure after even a minor electrical or mechanical shock. Experimental work has been extended to investigate flow injection monitoring and analysis of the triazine herbicides simazine, atrazine, and propazine in mixtures using filter-supported BLMs.18 The interactions of triazines with filter-supported BLMs produced a transient current with a duration of seconds. The peak height of the transients was linearly related to the concentration of herbicide, which could be determined at the submicromolar level. A possible mechanism of response may involve three steps, adsorption, nucleation and aggregation, as was suggested in our previous studies.17,18 Triazine may adsorb to the surface of a BLM, and the resulting assembly of lipid and triazine may then associate as aggregates to provide electrostatic perturbation of the lipid membrane.19,20 Aggregation appears to be relatively independent of the rate of the first step (adsorption). It seems that a threshold concentration of aggregative material in the BLM must be present for nucleation to occur; and this nucleation event sets the time for a transient to appear. After nucleation has occurred, the rate of aggregation should be relatively constant (at any one temperaturesour data have only been collected at room temperature) and is expected to be slow as based on lateral rates of movement across the surface of a membrane. Such slow aggregative events have been observed by fluorescence and turbidity measurements for atrazine interactions with lipid vesicles.19,20 The extent (degree) of aggregation could moderate rapid reorganization of electrostatic and phase structure at the surface of a membrane. At higher concentrations of triazine, aggregates may be larger or more numerous and, therefore, could cause a greater magnitude of ion current transient. The use of filter-supported BLMs has allowed repetitive cycles of injection of herbicides to be performed without signal degradation during each cycle. The time of appearance of the transient response was different for each triazine and increased in the order of simazine, atrazine, and propazine. This time dependence has been used to selectively detect these triazines in mixtures in less than 2 min. However, the detection limits of the technique do not allow applications in real samples, and the system and the support equipment is not suitable for rapid screening of herbicides in field. A technique for the preparation of stabilized metal-supported BLMs (s-BLMs) has recently been reported by Tien.21,22 The present work investigates the use of s-BLMs as bioprobes for the rapid electrochemical screening of triazine herbicides. The structure of the s-BLMs was tailored to contain defect sites so that direct membrane/triazine interactions could provide enhanced transient signals. The transduction was likely a result of perturbation of the electrostatics of the membrane caused by (17) Nikolelis, D. P.; Andreou, V. G. Electroanalysis 1996, 8, 643-647. (18) Nikolelis, D. P.; Siontorou, C. G. Electroanalysis 1996, 8, 907-912. (19) Tanfani, F.; Ambrosini, A.; Albertini, G.; Bertoli, E.; Curatola, G.; Zolese, G. Chem. Phys. Lipids 1990, 55, 179-189. (20) Zolese, G.; Ambrosini, A.; Bertoli, E.; Curatola, G.; Tanfani, F. Chem. Phys. Lipids 1990, 56, 101-108. (21) Tien, H. T.; Salamon, Z. Bioelectrochem. Bioenerg. 1989, 22, 211-218. (22) Zviman, M.; Tien, H. T. Biosens. Bioelectron. 1991, 6, 37-42.
3110
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
reorganization of hydrogen-bonding networks due to triazine aggregation events at the surface of BLMs, with subsequent alteration of the internal structure of the BLMs. EXPERIMENTAL SECTION Reagents and Apparatus. The materials and apparatus used throughout this study were essentially identical to those described previously.17,18,23 The lipids that were used in this work were lyophilized egg phosphatidylcholine (PC; Avanti Polar Lipids, Birmingham, AL) and dipalmitoylphosphatidic acid (DPPA; Sigma Chemical Co., St. Louis, MO). 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) used for the preparation of buffer solutions and gramicidin D were also purchased from Sigma. Silver wires (diameters 0.2-1.0 mm) with and without Teflon coating were obtained from Aldrich Chemical Co., Inc. (Milwaukee, WI). Stock solutions of atrazine in ethyl acetate (100 mg L-1) and simazine in acetone (10 mg L-1) and solid propazine were kindly donated by Benaki Phytopathological Institute (Athens, Greece). Water was purified by passage through a Milli-Q cartridge filtering system (Milli-Q, Millipore, El Paso, TX) and had a minimum resistivity of 18 MΩ‚cm. All other chemicals were of analytical-reagent grade. Ellipsometry of lipid films on reflective silver surfaces was done using an Auto-EL II nulling ellipsometer (Rudolph Research). This instrument used a 1 mW 632 nm helium-neon laser source, at an incidence angle of 70°. A two-electrode configuration was used for all electrochemistry experiments and consisted of a sensing electrode (i.e., silver wire with a self-assembled BLM) and a Ag/AgCl electrode acting as a reference. A dc potential of 25 mV was applied between the electrodes, and the ionic current through the s-BLM was measured with a digital electrometer (Model 614, Keithley Instruments, Cleveland, OH). The sensing electrode was connected to the power supply source whereas the reference electrode was connected to the electrometer; the applied potential at the sensing electrode was positive to ground (i.e., electrometer). The same electrometer was used as a current-to-voltage converter. The electrochemical cell and sensitive electronic equipment were placed in a grounded Faraday cage. Procedures. The lipid solution used for the formation of the metal-supported BLMs contained 2.5 mg mL-1 total lipid and was composed of 0, 15, and 35% (w/w) DPPA. These solutions were prepared daily from stock solutions of PC (2.5 mg mL-1) and DPPA (2.5 mg mL-1) in a solvent system of n-hexane and absolute ethanol (80 + 20, v/v). The stock lipid solutions were stored in the dark in a nitrogen atmosphere at -4 °C. The BLMs were supported in a 0.1 M KCl electrolyte solution buffered with 10 mM HEPES. Stock aqueous solutions of atrazine (28 mg L-1), simazine (0.5 mg L-1), and propazine (8 mg L-1) were prepared either by evaporation of the organic solvent by a nitrogen stream and then addition of the proper volume of water (for atrazine and simazine) or by directly dissolving propazine powder in water. s-BLMs were constructed according to established techniques.21,22,24 A lipid layer was deposited onto a nascent metallic surface cut with a scalpel just before or while immersing the metal wire into the lipid solution. The wire coated with the lipid solution was subsequently immersed into a 0.1 M KCl solution, and the (23) Nikolelis, D. P.; Siontorou, C. G.; Krull, U. J.; Katrivanos, P. L. Anal. Chem. 1996, 68, 1735-1741. (24) Otto, M.; Snejdarkova, M.; Rehak, M. Anal. Lett. 1992, 25, 653-662.
electrochemical current was stabilized over a period of ∼25 min. The formation of s-BLMs was verified by the magnitude of the current through the membranes and by electrochemical characterization using gramicidin D.25 The chemical characterization of the bimolecular thickness of s-BLMs by addition of gramicidin (this peptide does not induce conductance alterations if the membrane is thicker than one bilayer25,26) is based on the selective transport of monovalent cations through the membrane by formation of ion channels, and the conductance increases many fold if gramicidin can span the membrane.26 Experiments using s-BLMs were done using 0.1 M NaCl as electrolyte solution, which provided background current values ∼100 pA, whereas in the presence of 4 µM gramicidin, the ion current increased to 166 nA. Similar experiments using KCl as electrolyte have provided larger ion current increases. A study of the thinning behavior of lipid solutions on reflective silver surfaces was done by ellipsometry. It was necessary to simulate an electrode surface based on a silver wire surrounded by Teflon (as used to support BLMs), as the wires used in the electrochemical experiments had insufficient diameter for ellipsometric analyses with the equipment that was available for these studies. A base of poly(vinylidene difluoride) was metalized in a vacuum evaporator by first depositing ∼10 nm of chromium, followed by 150 nm of silver. The resulting systems had metal pads of 3.5, 5.5, or 7.5 mm diameter on a polymer similar to Teflon. The samples were mounted in a closed windowed cell (glass windows) so that they could be immersed in solution and studied by ellipsometry. Volumes of stock solution of PC were placed directly onto the metal pads in sufficient quantity to completely coat the metal, and electrolyte solution was then immediately added to cover the samples and fill the cell. Data from the ellipsometer were collected over a period of 2 h for each sample. Calibration of the concentration response of s-BLMs was done by stepwise addition of a standard triazine solution while continuously stirring. Once the calibration graph (or its equation) was established, the concentration of an unknown triazine solution could be independently determined using a freshly prepared sensor. All experiments were performed at 25 ( 1 °C. RESULTS A simple and inexpensive method for preparation of stabilized BLMs involves the physical support of membranes on metal surfaces.21,22,24 It has been proposed that BLMs can spontaneously form on silver wire that is surrounded by an annulus of plastic such as Teflon (E. I. du Pont de Nemour and Co., Inc.), by means of a thinning process similar to formation of planar “freely suspended” BLMs. The characterization of s-BLMs has been done electrochemically, and it is clear that substantial differences exist between the electrochemical properties of s-BLMs and planar BLMs. We used ellipsometry to investigate whether BLMs actually formed on silver surfaces that were surrounded by an annulus of hydrophobic plastic. Presuming a refractive index of ∼1.5 for the organic films, the ellipsometric data indicated formation of films of thickness that varied from monolayer to bilayer dimensions. The films thinned over a period of 20-30 min, and the final film thickness and the rate of thinning showed no dependence on the diameter of the metal pads. In combination (25) Goodall, M. C. Arch. Biochem. Biophys. 1971, 147, 129-135. (26) Yeagle, P. The Structure of Biological Membranes; CRC Press: Boca Raton, FL, 1992; Chapters 15 and 16.
Figure 1. Recordings obtained at pH 8.0 (0.1 M KCl, 10 mM HEPES, 1.0 mM Ca2+) with BLMs composed of 35% (w/w) DPPA to stepwise increases or decrease of atrazine concentration in bulk electrolyte solution (silver wire of 0.5 mm diameter). (A)-(G) indicates the forward ion current response of increasing atrazine levels in solution from 0 to 411 ppb, whereas (H) shows recording of current response during dilution of an atrazine standard solution with the background electrolyte. Total atrazine concentration (ppb) in solution: (A) 0; (B) 15.8; (C) 40.0; (D) 82.0; (E) 122; (F) 206; (G) 411; (H) 114.
with the electrochemical results, it would appear that BLMs likely do form on the silver wires used for preparation of s-BLMs. The use of s-BLMs represents an efficient approach for preparation of disposable single-sample format bioprobes having a small size for field screening applications. Figure 1 shows recordings of responses from s-BLMs (at pH 8.0) for different concentrations of atrazine, using BLMs composed of 35% (w/w) DPPA in the presence of 1.0 mM Ca2+ (silver wire of 0.5 mm diameter). The process of formation of s-BLMs requires variable times (25-40 min, depending on the silver wire diameter), and atrazine additions have been made at different times after membrane formation and stabilization. Only after addition of atrazine were the increases of electrochemical current shown in Figure 1 observed. The response times to establish 99% of steadystate current were on the order of 10 s. The ion current values can be used to quantify the concentration of atrazine as current values are linearly related to the concentration of atrazine in bulk solution as shown in Figure 2 (using 0.5 mm diameter silver wire). The results of Figure 2 indicate that the analytically useful concentration range for atrazine determination is between 15 and 400 ppb and that the ion current values are linearly related to atrazine concentration [I (nA) ) 1.12C (ppb) + 52.6, r2 ) 0.99]. The detection limit of atrazine (for S/N ) 3) was found to be 15 ppb. This was determined from the concentration of atrazine that provided a current difference of ∼20 nA with respect to the current obtained from s-BLMs containing no atrazine in electrolyte solution (using experimental noise levels of ∼7 nA). The diameter of silver wire did not significantly affect the sensitivity or detection limit of atrazine but shifted the calibration graph to larger or smaller current values (i.e., larger diameter wire provided larger current values). A number of s-BLMs were prepared to determine the reproducibility of independently fabricated devices. Reproducibility of response was found to be on the order of (5 to 7% (N ) 5, 95% confidence limit). For example, the RSD was found to be 5.9% for atrazine concentrations of 100 ppb. The results demonstrate that a constant background ion current, sensitivity, and Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
3111
Figure 2. Optimization and calibration of the analytical signal from the experimental results shown in Figure 1. (A) pH 8.0 (0.1 M KCl, 10 mM HEPES, 1.0 mM Ca2+) with BLMs composed of 35% (w/w) DPPA; (B) pH 5.5 (0.1 M KCl, 10 mM HEPES) with BLMs composed of 15% (w/w) DPPA; (C) pH 8.0 (0.1 M KCl, 10 mM HEPES, 1.0 mM Ca2+) with BLMs composed of 0% (w/w) DPPA.
response to a given atrazine concentration solution is obtained with the use of the present s-BLM minisensor. However, a drift of the ion current with time (dependent on atrazine concentration) was noticed and had a maximum value of ∼2.4 nA min-1. Optimization of the magnitude of the analytical signal obtained from s-BLMs for atrazine can be achieved when using mixtures of egg PC and DPPA for BLM preparation (Figure 2). A phase separation of these lipids in membranes is important for signal generation and can be controlled by altering the amount and degree of ionization of DPPA in BLMs by adjustment of pH and concentration of calcium ions in the bulk electrolyte.17 Recent studies using differential scanning calorimetry have indicated that at pH 8.0 and in the presence of calcium ions there is a coexistence of solid (gel) and fluid (liquid-crystalline) phases in vesicular BLMs composed of PC and 35% (w/w) DPPA at 25 °C.27 Such a coexistence results in an increased adsorption (and therefore increased partition coefficient values) of foreign molecules such as anaesthetics, insecticides, herbicides, and bitter or odorous substances which are allowed to interact with BLMs.28,29 More, atrazine seemed to adsorb to BLMs at Tm, as compared to the relatively homogeneous surfaces of solid (below the Tm) or liquidcrystalline matrixes (above Tm). Similar results were obtained from s-BLMs when simazine and propazine at pH 8.0 were used with BLMs composed of 35% (w/ w) DPPA in the presence of 1.0 mM Ca2+. The response times of s-BLMs to simazine and propazine sensing were also on the order of 10 s. The ion current values were used to quantify the concentration of these herbicides, as the maximum current level was linearly related to the concentration of simazine and propazine in bulk solution (Figure 3, using 0.5 mm diameter silver wire). The results of Figure 3 indicate that the analytically useful concentration range for simazine and propazine determination is between 1 and 10 and 20 and 200 ppb, respectively, and that the ion current values are linearly related to herbicide concentration [I (nA) )17.7C (ppb) + 21.3, r2 ) 0.993, for simazine and I (nA)) (27) Nikolelis, D. P.; Siontorou, C. G.; Andreou, V. G.; Krull, U. J. Electroanalysis 1995, 7, 531-536. (28) Jorgensen, K.; Ipsen, J. H.; Mouritsen, O. G.; Bennett, D.; Zuckermann, M. J. Biochim. Biophys. Acta 1991, 1067, 241-253. (29) Okahata, Y.; Enna, G.-I.; Ebato, H. Anal. Chem. 1990, 62, 1431-1438.
3112 Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
Figure 3. Calibration of the analytical signal for simazine and propazine at pH 8.0 (0.1 M KCl, 10 mM HEPES, 1.0 mM Ca2+) using metal-supported BLMs composed of 35% DPPA: (A) simazine; (B) propazine.
2.06C (ppb) + 3.92, r2 ) 0.99, for propazine]. The detection limit of simazine and propazine (for S/N ) 3) were found to be 1 and 20 ppb, respectively. The reproducibility of independently fabricated devices was found to be similar to that presented for atrazine. The results demonstrate that a relatively constant and reproducible background ion current, sensitivity, and response to each different herbicide is obtained with the use of the present s-BLM minisensor. Discrimination between the three herbicides cannot be achieved with a single s-BLM as the bioprobe shows an integrated response for all triazine herbicides that are present. DISCUSSION Increases of electrochemical current when s-BLMs were used for detection of the three herbicides were obtained within seconds and were indicative of rapid transient alterations of the electrostatic fields and perhaps the internal structural organization of BLMs. Previous studies have examined the interactions of atrazine with dipalmitoylphosphatidyl (DPPC) vesicles using Fourier transform infrared spectroscopy (FT-IR), differential scanning calorimetry (DSC), and fluorescence polarization.19,20 The results indicated that atrazine localizes near the glycerol backbone of the lipid without perturbing the hydrophobic core of the lipid bilayer. DSC studies have shown that alterations of the transition temperature of the vesicular BLMs did occur.17 These studies have suggested that atrazine is involved in hydrogen bonding with the carbonyl moiety of the lipid, which is similar to the interactions of other
molecules containing hydrogen-bonding donor groups. Localization of the herbicide near the glycerol backbone can result in greater hydration of the CdO moieties of the lipid,20 which can substantially alter the dipolar potential values at the membrane/ electrolyte interface.30 The significance of the sn-1 carbonyl group in PC for the interactions with atrazine and generation of an electrochemical current signal was previously investigated by the addition of PAF (an 1-alkyl-2-acetyl ether analog of PC) in BLMs.17 The present rapid response of s-BLMs to triazine suggests a fast alteration of ion transport through s-BLMs. Lipid membranes composed from 35% DPPA in the presence of calcium ions show phase separation of the two lipids, and localized domains enriched in DPPA exist as revealed by fluorescence microscopy31 and conductivity measurements.32 The formation of intermolecular hydrogen bonds in localized domains enriched in charged lipid will result in increased concentrations of the herbicide in these domains. This would disrupt the lipid/lipid interactions and modulate internal structure of membranes. Our previous DSC experiments have shown that atrazine can destabilize the phase structure of BLMs by disrupting the hydrogen-bonding network of DPPA molecules.17 This is reflected by a structural transition to a more fluid phase and leads to increases in ionic current flow as noticed in our present electrochemical experiments using s-BLMs. There were no discernable threshold transients observed during our present experiments using s-BLMs, suggesting that s-BLMs may differ significantly in structure from planar BLMs.33,34 It is possible that the transient signal, which is based on a membrane-charging current, is masked or eliminated by the large ion current through s-BLMs. The pathways for transmembrane ion current could originate from structural defects between phase domains of a BLM, which would provide transient water-containing pathways capable of salt transport. Other commonly proposed mechanisms found in the literature include formation of hexagonal or micellar phase structures that may act as carriers of salts, or lipid flip-flop across BLMs, where the lipid headgroup would be associated with a salt. The short time delay (seconds) for atrazine detection when using s-BLMs in comparison to that observed when using planar freely suspended and filter-supported BLMs (minutes) is probably due to structure and mechanism of ion current formation associated with the silver metal electrode that supports s-BLMs.23 One side of the s-BLMs (surface of one monolayer) is in contact with the aqueous electrolyte solution. The second monolayer contacts the metal surface, with some regions (depending on the metal roughness) potentially trapping electrolyte solution. The very limited trapped volume of solution makes it possible to reach equilibrium in terms of ion concentration and diffusion very rapidly. In our experiments, the metal-supported BLMs remained stable for periods of over 48 h. The currents through BLMs stabilized within 25, 40, and 90 min (if the wire diameter was 1.0, 0.5, and 0.25 mm, respectively) after immersion of the metal wire with the (30) Hianik, T.; Passechnik, V. I.; Paltauf, F.; Hermetter, A. Bioelectrochem. Bioenerg. 1994, 34, 61-68. (31) Krull, U. J.; Brennan, J. D.; Brown, R. S.; Hosein, S.; Hougham, B. D.; Vandenberg, E. T. Analyst 1990, 115, 147-153. (32) Nikolelis, D. P.; Brennan, J. D.; Brown, R. S.; Krull, U. J. Anal. Chim. Acta 1992, 257, 49-57. (33) Nikolelis, D. P.; Tzanelis, M. G.; Krull, U. J. Anal. Chim. Acta 1993, 281, 569-576. (34) Hianik, T.; Dlugopolsky, J.; Gyepessova, M. Bioelectrochem. Bioenerg. 1993, 31, 99-111. Mol. Electron. Biocomput. 1993, 7, 37-40. Hianik, T.; Passechnik, V. I.; Sargent, D. F.; Dlugopolsky, J.; Sokolikova, L. Bioelectrochem. Bioenerg. 1995, 37, 61-68.
lipid coating into the electrolyte solution without triazine. The 90 min period required before commencement of measurements when using wire with 0.25 mm diameter is shorter than previously reported times.35 This is due to the use of hexane instead of decane as a solvent for the preparation of lipid solution. The use of hexane also alters the specific capacitance of s-BLMs from that observed when decane is used.36 The use of a wire of 0.25 mm diameter should also be avoided due to the increased background noise level effects. The formation of BLMs on solid supports allowed the electrochemical investigation of the reversibility of signal response of BLMs to triazines (i.e., observation of atrazine concentration decreases in solution). The present minisensor responds to increases or decreases of triazine concentration in solution. Increments of electrolyte could be added so as to lower the atrazine concentration (Figure 1). The response times (to establish 99% of steady-state current) were on the order of 30 s for decreasing concentrations, and the current values obtained corresponded to the calibration graph for increases of atrazine concentration in bulk solution (vide supra). For practical applications to sensing, a fresh sensing membrane should be prepared for determination of an unknown sample. This ensures elimination of carryover effects. This also avoids BLM destabilization, which can occur in dry cycling (i.e., transfer of the s-BLM to another electrolyte solution without triazine) due to dehydration of a BLM, or repetitive washing out of triazine after use. The use of freshly prepared s-BLMs is not a serious limitation owing to the simple, inexpensive, and relatively fast method of preparation of this type of sensor. In addition, a number of sensors can be concurrently prepared to compensate for the delay time for s-BLM formation. In conclusion, the results indicate that triazine herbicides can be rapidly screened using the present metal-supported BLM-based minisensor. The approach provides response times of seconds and detection sensitivity and limits for triazine herbicides (i.e., simazine, atrazine and propazine) that are suitable for direct analysis of some field samples without preconcentration (although sample preparation to eliminate other adsorbents to BLMs may be necessary). The detection limits in the present studies are similar in magnitude to those obtained from systems that use antibodies or photosynthetic “receptors”. Note that photosynthetics9-11 also provide an integrated response to triazines present in a sample, but there were reports in literature in which antibodies have provided the necessary specificity and selectivity in closely related herbicides of the trazine class (i.e., atrazine and ametryn).7 The fastest reported response time for a biosensor that can detect herbicides has been obtained using a surface transverse wave acoustic device (i.e., less than 3 min),15 and most recently a device based on filter-supported BLMs reported response time of less than 2 min.18 The present method offers response times on the order of seconds. While the work presented here represents an attractive configuration and application of electrochemistry of BLM-based sensors, the practical use of such a sensor for real world applications needs to be further researched for robustness, lifetime, manufacturability, and other performance requirements that will further allow commercialization of the present device. (35) Rehak, M.; Snejdarkova M.; Otto, M. Electroanalysis 1993, 5, 691-694. (36) Nikolelis, D. P.; Krull, U. J. Talanta 1992, 39, 1045-1049.
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
3113
The direct screening of triazine herbicides from environmental samples could be done in protein-free water and skimmed milk samples, a restriction not found in other antibody-based sensor systems. Proteins should be eliminated from the samples prior to analysis, as these macromolecules can cause a nonselective interference with BLMs.23 Interference from proteins must be considered, as demonstrated by our investigations of the effect of bovine albumin and casein on BLMs. The results indicate that concentrations of albumin and casein in bulk solution larger than about 0.35 and 0.06 mg mL-1, respectively, provide ion current values corresponding to S/N ratio larger than 3. A number of pesticides and insecticides were tested as potential interferents. The compounds tested were diuron, alachlor, chloropyrifos, carbofuran, monocrotofos, aldicarb, methylparathion, and lindane.
3114
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
These compounds did not produce any electrochemical current increases, even at concentration levels of ∼1 × 10-5 M. ACKNOWLEDGMENT This work was carried out in the framework of “Copernicus” Contract CIPA CT94-0231 and “Inco-Copernicus” Contract IC15CT96-0804 with the financial contribution from the European Commission. Received for review January 28, 1997. Accepted May 2, 1997.X AC970113+ X
Abstract published in Advance ACS Abstracts, June 15, 1997.