A Combined Chemical and Electrochemical Approach Using Bis

Kyle N. Knust , Matthew P. Foley , Mohammad S. Mubarak , Sonja Skljarevski , Krishnan Raghavachari ... Afsaneh Safavi , Norouz Maleki , Fariba Tajabad...
0 downloads 0 Views 214KB Size
Anal. Chem. 1997, 69, 4324-4330

A Combined Chemical and Electrochemical Approach Using Bis(trifluoroacetoxy)iodobenzene and Glucose Oxidase for the Detection of Chlorinated Phenols Coralie Saby, Keith B. Male, and John H. T. Luong*

Biotechnology Research Institute, National Research Council Canada, Montreal, Quebec, Canada H4P 2R2

A novel electrocatalytic approach using a chemical reaction and an enzymatic reaction has been developed for the measurement of 18 chlorophenol congeners, including highly chlorinated pollutants such as pentachlorophenol, 2,3,5,6-tetrachlorophenol, 2,3,4,6-tetrachlorophenol, and several trichlorophenols. Chlorophenols were oxidized to chlorobenzoquinones with very high yields using bis(trifluoroacetoxy)iodobenzene in 0.1 M trichloroacetic acid, pH 1.5, at ambient temperature. UVvisible spectrophotometry, cyclic voltammetry, and HPLC have been used to characterize the reaction products and yields. Together with glucose oxidase immobilized on a working glassy carbon electrode (+0.45 V vs Ag/AgCl), chlorinated benzoquinones have been demonstrated to be efficient mediators in a glucose oxidase/glucose system. In this approach, glucose oxidase was readily reduced by excess glucose to provide a non-rate-limiting source of electron flow toward the electrode. The oxidation products of chlorophenols then recycled the reduced glucose oxidase to its active oxidative state, i.e., mediating the rate-limiting electron transfer from the enzyme to the electrode. At pH 3.5, linear behavior of the current response was observed up to 200 nM for all chlorophenol oxidation products. The detection limit of this method for both pentachlorophenol and 2,3,5,6-tetrachlorophenol was about 4 nM, which is close to the maximum allowable contamination level of pentachlorophenol in water samples (2.7 nM). The detection limit obtained for pentachlorophenol could also be considered superior to the result obtained with the PCP immunoassay technology (13.3 nM). Chlorophenols (CPs) are a major group of chemicals used for a variety of biocidal purposes and as precursors for several pesticides.CPs have been formed as byproducts of many industrial activities, such as the production of antioxidants, dyes and drugs,1,2 chlorination of drinking water, and chlorinated bleaching of paper.3 They are also frequently present in the waste of the coal, gas, (1) Gilman, A.; Douglas, U.; Arbuckle-Sholtz, T.; Jamieson, J. Chlorophenols and Their Impurities: A Health Hazard Evaluation; Bureau of Chemical Hazards, Environmental Health Directorate, National Health and Welfare: Ottawa, ON, Canada, 1982. (2) Pavlov, B.; Terentyev, T. Organic Chemistry; Gordon and Breach Science Publishers: New York, 1965. (3) Kristiansen, N.; Froshaug, M.; Aune, K.; Becher, G. Environ. Sci. Technol. 1994, 28, 1669-1973.

4324 Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

and petroleum industries.4 Massive amounts of chlorinated phenols have been released in the environment in the form of solvents, degreasers, and industrially useful compounds. Pentachlorophenol (PCP), commonly used as a wood preservative, is the most acutely toxic of the CPs, and 2,4,6-trichlorophenol (2,4,6-TCP) has been shown to be carcinogenic in rats. Due to their widespread and toxicological properties,5 several chlorinated phenols, such as 2-chlorophenol, 2,4-dichlorophenol, 2,4,6-TCP, and PCP, have been classified by the U.S. EPA as priority pollutants.6 Recently, in the United States, PCP regulations under the Resource Conservation and Recovery Act have been created specifically for wood treatment facilities. The large-scale use of CPs has led to the contamination of terrestrial and aquatic ecosystems.Various industrial effluents contain up to 18 mg/L (ppm) of CPs. Municipal waste discharges typically contain from 1 to 21 mg/L (ppm) of CPs.1 The disposal of waste chemicals in landfill sites has been a common practice; presumably, such sites could be a source of CPs in drinking water. Therefore, monitoring and detection of CPs are of particular importance in environmental and food analysis for the investigation of human and animal exposure. Electrochemical detection is of particular interest in the determination of CPs because of its simplicity and sensitivity. However, this method requires a high oxidation potential (700900 mV), which leads to severe interferences, since contaminated samples may contain significant levels of electroactive interfering species. During electrochemical oxidation, the oxidized forms of CPs are also prone to polymerization, which results in fouling of the electrode surface.7,8 To circumvent these drawbacks, CPs can be converted into the corresponding chlorinated 1,4-benzoquinones or chlorinated hydroquinones, such that lower potentials (+450 mV or less) can be used for detection.9,10 On the basis of the mediating property of benzoquinones, amplification effected by coupling a suitable enzyme, such as glucose oxidase, can be utilized to increase the sensitivity. (4) Rogers, K.; Gerlach, C. Environ. Sci. Technol. 1996, 30, 486A-491A. (5) Soderman, J. CRC Handbook of Identified Carcinogens and Noncarcinogens: Carcinogenicity-Mutagenicity Database; CRC Press: Boca Raton, FL, 1982. (6) U.S. Environmental Protection Agency. Toxic Substance Control Act, 1979. (7) Weisshaar, D. E.; Tallman, D. E.; Anderson, J. L. Anal. Chem. 1981, 53, 1809-1813. (8) Bejerano, T.; Forgacs, C.; Gileadi, E. J. Electroanal. Chem. 1970, 27, 469476. (9) Zhao, S.; Luong, J. H. T. Anal. Chim. Acta 1996, 327, 235-242. (10) Bonakdar, M.; Vilchez, J. L.; Mottola, H. A. J. Electroanal. Chem. 1989, 266, 47-55. S0003-2700(97)00369-7 CCC: $14.00

© 1997 American Chemical Society

To this day, oxidations of phenolic compounds, particularly highly chlorinated compounds, in aqueous medium by photochemical, chemical, or enzymatic methods have not been very successful. Photochemical techniques11,12 involve various photosensitizers, including oxygen and hydrogen peroxide as oxidative agents. The major drawbacks of photochemical reactions are their complexity and the possible formation of a variety of compounds, including quinones, diols, triols, and polyhydroxybiphenyls.13,14 The use of enzymes, especially peroxidases, laccase, and tyrosinase, has also been attempted for the oxidation of CPs. However, the reaction using tyrosinase or laccase occurs only with 4-chlorophenol, 2-amino-4-chlorophenol, and 2,4-dichlorophenol, and the main product is an o-quinone.15,16 In the case of other enzymes, such as horseradish peroxidase, chloroperoxidase, lignin peroxidase, and manganese-dependent peroxidase, the action of the enzyme produces an intermediate phenoxy radical. Depending on the position and number of chlorines on the aromatic ring, the phenol concentration, and the nature of the enzyme, various compounds can be generated: polymers, coupling products, or quinones.17,18 Oxidation of phenolic compounds can also be carried out using chemical reagents. In this case, one of the favorite oxidative reagents is hydrogen peroxide in conjunction with various catalysts, including enzymes.19,20 However, the reaction occurs only with a few chlorinated phenols with low yields. This article describes a novel and sensitive electrochemical method for detection and determination of chlorinated phenols. In this approach, CPs were converted into the corresponding chlorinated benzoquinones with a high reaction yield by using bis(trifluoroacetoxy)iodobenzene. Cyclic voltammetry, UV-visible spectrophotometry, and HPLC have been used to characterize the reaction products and yields. Together with an amperometric electrode using immobilized glucose oxidase, chlorinated benzoquinones have been demonstrated as efficient mediators in a glucose oxidase/glucose system. EXPERIMENTAL SECTION Reagents. Glucose oxidase (EC 1.1.3.4, type X-S, 182 units/ mg) was obtained from Sigma (St. Louis, MO). Analytical grade CPs and chlorinated 1,4-benzoquinones were purchased from Aldrich (Milwaukee, WI) and ChemService (West Chester, PA). All other products were also of analytical grade and obtained from Aldrich. D-Glucose stock solutions (2 M) were allowed to mutarotate for at least 24 h before use. Chlorophenols Oxidations using Bis(trifluoroacetoxy)iodobenzene. Unless otherwise indicated, reactions were performed in 0.1 M trichloroacetic acid, pH 1.5, using freshly prepared CP (1 or 10 mM in ethanol) and bis(trifluoroacetoxy)(11) Mills, A.; Morris, S.; Davies, R. J. Photochem. Photobiol. A: Chem. 1993, 70, 183-191. (12) Silva, M.; Burrows, H.; Grac¸ a Miguel, M.; Formosinho, S. Ber. Bungenges. Phys. Chem. 1996, 100, 138-143. (13) Omura, K.; Matsuura, T. Tetrahedron Lett. 1968, 24, 3475-3487. (14) Durand, A.-P.; Brown, R. G.; Worrall, D.; Wilkinson, F. J. Photochem. Photobiol. A: Chem. 1996, 96, 35-43. (15) Aitken, M.; Massey, I.; Chen, T.; Heck, P. Wat. Res. 1994, 28, 1879-1889. (16) Marko-Varga, G.; Emme´us, J.; Gorton, L.; Ruzgas, T. Trends Anal. Chem. 1995, 14, 319-328. (17) Casella, L.; Poli, S.; Gullotti, M.; Selvaggini, C.; Beringhelli, T.; Marchesini, A. Biochemistry 1994, 33, 6377-6386. (18) Vasudevan, P.; Li, L. Appl. Biochem. Biotechnol. 1996, 60, 73-82. (19) Tang, W.; Huang, C. Chemosphere 1996, 33, 1621-1635. (20) Ito, S.; Aihara, K.; Matsumoto, M. Tetrahedron Lett. 1983, 24, 5249-5252.

iodobenzene (BTFAIB, 100 mM in ethanol) stock solutions. The reaction mixture consisted of 10 µM CP and 500 µM BTFAIB, and all reactions were carried out at room temperature, with light protection for 30 min. At the end of the reaction, hydrogen peroxide (500 µM) was added to neutralize any unreacted BTFAIB. Electrochemical Measurements. Amperometric detection was performed using a CV1B voltammograph (BioAnalytical Systems, West Lafayette, IN). Digitization and data acquisition were accomplished using a DAS-8 A/D card (MetraByte, Taunton, MA) connected to an IBM-AT computer with custom software. Cyclic voltammograms were performed using a standard threeelectrode cell consisting of a counter platinum wire, an Ag/AgCl (3 M NaCl) reference electrode, and a glassy carbon working electrode (3 mm diameter, BioAnalytical Systems). HPLC Experiments. The HPLC system consisted of a manual injector with a 20 µL sample loop, a Model 590 pump (Waters, Milford, MA), and a 3.9 mm i.d. × 15 cm stainless steel column (µBondapack C18, Waters). Electrochemical detection of chlorinated hydroquinones was performed at +0.45 V using a LC-44 thin-layer electrochemical cell with a dual glassy carbon working electrode, an Ag/AgCl reference electrode, and a stainless steel auxiliary electrode (BioAnalytical Systems). The mobile phase was a mixture of 0.1 M tartrate, pH 3 (25% v/v), water (60 or 40% v/v), and acetonitrile (15 or 35% v/v) with a flow rate of 0.5 mL/min. Standard chlorinated benzoquinones and CPs were diluted in 0.1 M trichloroacetic acid, pH 1.5, containing 50 mM tartrate. The samples were reacted for 30 min with 500 µM BTFAIB, and then the reaction was stopped by the addition of 500 µM hydrogen peroxide. Zinc powder was added to reduce the chlorobenzoquinones to the corresponding hydroquinone forms, and the resulting sample was filtered before injection to the LC system. Enzyme Electrode Preparation. A glassy carbon disk electrode was polished with 1 µm diamond paste and 0.05 µm alumina slurry (Buehler) and washed with deionized water. After drying of the electrode surface, 20 µL of a 100 mg/mL glucose oxidase solution in phosphate buffer (0.1 M, pH 7) was placed on the surface and covered by a dialysis membrane (MWCO 14 000) held in place by an O-ring. Finally, the electrode was rinsed with 0.1 M phosphate buffer (pH 7) and stored in this buffer at 4 °C. Responses of the enzyme electrode were measured using a standard three-electrode cell system with a counter platinum wire and an Ag/AgCl (3 M NaCl) reference electrode. Measurements were performed at +0.45 V in a stirred 0.1 M tartrate buffer, pH 3.5, containing 40 mM glucose at room temperature with humidified nitrogen sparging. RESULTS AND DISCUSSION Principles of Electrocatalytic Detection. Hypervalent iodine agents21-23 have been used very recently for phenol and phenol ether oxidations. Although BTFAIB has been used to oxidize various naphthols and phenols22 to the corresponding quinones, no attempt has been made to oxidize chlorinated phenols. Preliminary results obtained in this study indicated that the conversions of CPs into chlorinated 1,4-benzoquinones using BTFAIB were more selective and complete in acidic media (pH (21) Pelter, A.; Elgendy, S. J. Chem. Soc., Perkin Trans. 1 1993, 1891-1896. (22) Barret, R.; Daudon, M. Tetrahedron Lett. 1990, 31, 4871-4872. (23) McKillop, A.; McLaran, L.; Taylor, R. J. Chem. Soc., Perkin Trans. 1 1994, 2047-2048.

Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

4325

range 1.4-3.0). At higher pH (pH 4.5), the kinetics became sluggish, and several byproducts were formed. Interestingly, BTFAIB is an irreversible oxidant and can be easily destroyed by either zinc powder or hydrogen peroxide. Its reaction byproducts, iodobenzene and trifluoroacetate ion, were not electroactive and caused no interference with the electrode response. Based on this reaction, it was reasoned that CPs can be first converted into their corresponding chlorinated benzoquinone derivatives using BTFAIB. The resulting products are then used as mediators in a glucose oxidase/glucose system since the chlorinated benzoquinones (ClBQ) may behave similar to 1,4benzoquinone, i.e., they possess the capability of oxidizing the reduced form of glucose oxidase (FADH2) to its original active state. Consequently, chlorinated hydroquinones (ClBQH2) formed in the reaction are reoxidized to chlorinated benzoquinones by the electrode surface poised at a suitable potential (+ 450 mV). An amplification of the current is obtained due to the cycling of the mediator between the enzyme and the electrode surface. In this amplification scheme, excess glucose is necessary so that the electrode response will be dependent only on the chlorinated benzoquinone concentration and the kinetics of the reaction between the enzyme and the mediator.

Characterization Using UV-Visible Spectrophotometry and Cyclic Voltammetry. A preliminary characterization of reaction products obtained from the oxidation of 18 CP congeners with BTFAIB was carried out using UV-visible spectrophotometry (Beckman DU-640, Fullerton, CA) and cyclic voltammetry. In this particular case, reactions were performed in 0.1 M acetate buffer (pH 3) to minimize pH variations that affected cyclic voltammetric behavior of p-benzoquinones. Results showed that the major products were chlorinated p-benzoquinones. Structural elucidation based on the CP substrates revealed that it was only possible for each of the groups to yield the same quinone product when the identical product was 1,4-benzoquinone, and dechlorination had to occur at position 4 if it was chlorinated. In other words, the 1,4-benzoquinone structure exhibited a unique symmetry, and only benzoquinone and six chlorobenzoquinones could be produced from the 18 CP congeners. As will be discussed later, the HPLC data confirmed that only benzoquinone and six different chlorobenzoquinone peaks were observed from the chromatograms. Four commercially available chlorinated benzoquinones, 2-chloro1,4-benzoquinone (2-ClBQ), 2,5-dichloro-1,4-benzoquinone (2,5diClBQ), 2,6-dichloro-1,4-benzoquinone (2,6-diClBQ), and tetrachloro-1,4-benzoquinone (tetraClBQ), were used to facilitate the identification of the reaction products. Stable behavior of aqueous solutions of p-benzoquinones is observed at pH less than 5. Table 1 shows the electrochemical behavior and the UV-visible peak position of the different p-benzoquinones. As the number of chlorines on the aromatic ring increased, the location of the UV peak shifted to a longer 4326 Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

Table 1. UV-Visible Peak Positions and Electrochemical Behavior of Benzoquinone and Four Commercially Available Chlorinated Benzoquinones

product

UV-visible peak position (nm)

1,4-benzoquinone 2-chloro-1,4-benzoquinone 2,5-dichloro-1,4-benzoquinone 2,6-dichloro-1,4-benzoquinone tetrachloro-1,4-benzoquinone

246 255 272 273 290

cyclic voltammetric resultsa (V vs Ag/AgCl) E1/2

∆E

0.26 0.28 0.30 0.32 0.29

0.41 0.37 0.14 0.10 0.07

a ∆E ) (E - E ) and E a c 1/2 ) (Ea + Ec)/2, where Ea and Ec are the anodic and cathodic peak potentials of cyclic voltammograms (0.1 M acetic acid, pH 3, 100 mV s-1), respectively.

Figure 1. UV-visible spectra of pentachlorophenol (10 µM, curve a), the oxidation product of PCP (curve b), standard tetrachloro 1,4benzoquinone (10 µM, curve c), and bis(trifluoroacetoxy)iodobenzene (100 µM after hydrogen peroxide addition, curve d). The reaction was carried out in 0.1 M acetate buffer, pH 3, for 2 h with 10 µM pentachlorophenol and 100 µM bis(trifluoroacetoxy)iodobenzene.

wavelength. Moreover, even if the half-wave potential was not influenced by the number of chlorines, the electrochemical behavior showed better reversibility as the chlorine content increased. However, it was not possible to distinguish between 2,6-diClBQ and 2,5-diClBQ. Therefore, the reactions of the 18 CPs should produce only six compounds distinguishable by looking at either cyclic voltammetric waves or UV-visible spectra. Generally, cyclic voltammetric behavior of CPs is not quite reversible. Electrochemical oxidation of chlorinated phenols results in electrode fouling caused by the formation of a nonconducting polymer on the electrode surface. However, after reaction with BTFAIB ([BTFAIB]/[CP] ratio of 10), drastic changes in cyclic voltammogram and UV-visible spectrum were observed. New and more reversible cyclic voltammetric waves appeared, and the intensity of the CP oxidation wave decreased or disappeared. In acidic media, CPs exhibited a double peak in the range 275-320 nm, but during reaction this double peak decreased, and a new peak emerged and intensified at a shorter wavelength. For example, the UV-visible spectral and cyclic voltammetric data obtained from the reacted solution demonstrated that the reaction product of pentachlorophenol with BTFAIB is tetrachloro-1,4benzoquinone. As the reaction progressed, the pentachlorophenol peaks at 294 and 303 nm declined (Figure 1, curve a), and a more intense absorption peak at 290 nm emerged and intensified (Figure 1, curve b). Both the absorbency profile and the

Figure 2. Cyclic voltammograms (0.1 M acetate buffer, pH 3, 100 mV s-1) of pentachlorophenol (10 µM, 4), the oxidation product of PCP (O), and the standard solution of tetrachloro 1,4-benzoquinone (10 µM, straight line). Oxidation of 10 µM pentachlorophenol was carried out in 0.1 M acetate buffer, pH 3, for 1 h using 500 µM bis(trifluoroacetoxy)iodobenzene.

maximum wavelength of the PCP oxidation product were confirmed to be identical to those of standard tetrachloro-1,4benzoquinone (Figure 1, curve c). The cyclic voltammogram of the product obtained from the oxidation of PCP by BTFAIB was identical to that of the standard tetrachloro-1,4-benzoquinone with respect to the oxidation and reduction peaks and the peak-to-peak separation (Figure 2). As expected, the 18 CPs generated six products distinguishable by cyclic voltammetry and spectrophotometry. Moreover, in 13 cases, the product behavior was similar to that of benzoquinone or one of the four commercially available chlorinated benzoquinones. Such results clearly indicated that substitutions occurred only in the para position and that chlorines in other positions were not attacked. In all reactions, CP oxidations were completed in less than 10 min (ratio [BTFAIB]/[CP] ) 10), except in the case of 3,5-DCP, 3,4,5-TCP, and PCP, which required more reaction time. Kinetics could be improved by increasing the [BTFAIB]/ [CP] ratio to 50; in that case, all reactions were complete in less than 30 min. It should be noted that, due to the tetraClBQ solubility problem, concentrations of either PCP or 2,3,5,6 -tetraCP should be less than 10 µM. In view of this, all subsequent experiments were carried out using 10 µM CP, [BTFAIB]/[CP] ratio of 50, and 30 min reaction time. Identification and Quantitation of the Reaction Product and Yield Using HPLC. HPLC separation techniques were developed to identify as well as quantify the reaction products of the chlorophenol conversions with BTFAIB. Preliminary results indicated that the oxidation of chlorohydroquinones at +450 mV exhibited a more sensitive signal than the reduction of chlorobenzoquinones at +0 mV. As a consequence, the CP reaction products or chlorobenzoquinones were reduced to chlorohydroquinones by zinc powder. It was noticed that the addition of zinc powder could increase the pH of the reaction solution to above 5, causing irreproducible signals. Therefore, the pH was maintained below 3 by adding 50 mM tartrate to the reaction mixture. It was confirmed that the addition of tartrate had no effect on the reaction yield. Good efficiency of the conversion of chlorobenzoquinones to the corresponding chlorinated hydroquinone by

Figure 3. HPLC chromatograms of (a) 2-ClBQ, 2,5-diClBQ, and tetraClBQ and (b) converted 3-CP, 2,5-DCP, 2,3,6-TCP, and PCP. Potential was +0.45 V vs Ag/AgCl; mobile phase was a mixture of 0.1 M tartrate, pH 3 (25% v/v), water (40% v/v), and acetonitrile (35% v/v) with a 0.5 mL/min flow rate. Conversions were performed in a 0.1 M trichloroacetate-0.05 M tartrate buffer, pH 1.5, for 30 min, with 500 µM BTFAIB and 10 µM for each chlorophenol or standard. The small peak observed at 4 min comes from the zinc powder.

zinc powder was confirmed by comparing the electrochemical signal obtained from commercially available 2-chlorohydroquinone with the signal of reduced 2-ClBQ. A series of experiments was carried out using benzoquinone and the four available standard chlorobenzoquinones. As shown in Figure 3a, retention times increased with an increase in the number of chlorines on the benzoquinone. The response for tetraClBQ was linear up to 10 µM, with an integrated area of 0.207 ( 0.002 µA min µM-1 and a detection limit of 250 nM (S/N ) 3). However, at this high level of acetonitrile (35%), the different forms of dichlorobenzoquinone could not be separated. Peak separation was obtained by lowering the acetonitrile content to 15% such that the two standard dichlorobenzoquinones could be distinguished as a double peak. In this case, retention times of 2,5-diClBQ and 2,6-diClBQ were 17.6 and 18.6 min, respectively. The responses determined by the integrated area were 0.163 ( 0.002 and 0.158 ( 0.003 µA min µM-1 for 2,5-diClBQ and 2,6-diClBQ, respectively. In both cases, the response was linear up to 20 µM, and the detection limit was 250 nM. The response for 2-ClBQ (retention time, 8.4 min) was also linear up to 20 µM, with an integrated area of 0.072 ( 0.003 µA min µM-1 and a detection limit of 500 nM. It is interesting to note that, if acetonitrile was replaced by methanol, the separation could not be achieved, and the baseline exhibited a high level of noise. Reaction products of each of the 18 CPs with BTFAIB were analyzed by HPLC, and, when possible, their results were compared with the available chlorinated benzoquinones. As shown in Table 2, for 13 of the chlorophenols, the retention time of the single reaction product was similar to that of the expected benzoquinone. In the case of the other five chlorophenols, the expected chlorobenzoquinone products, 2,3-diClBQ and triClBQ, are not commercially available. However, their formation is strongly suggested. For example, 2,3,5-TCP, 2,3,6-TCP, and 2,3,4,6-tetraCP all exhibit the same single peak after conversion with BTFAIB, and the retention time of this peak falls between those of 2,5-dichlorobenzoquinone and tetrachlorobenzoquinone (Figure 3a). When different concentrations of 2,3,6-TCP were used, the response was linear to 20 µM, with an integrated area of 0.142 ( 0.005 µA min µM-1 and a detection limit of 250 nM. Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

4327

Table 2. HPLC Results for 10 µM Benzoquinone, Various Chlorinated Benzoquinones, and Converted Chlorophenols benzoquinone 1,4-benzoquinone (BQ) 2-chloro-1,4-benzoquinone (2-ClBQ)

standard integration (µA min)

chlorophenol

integration peak

yield %

0.264 ( 0.014 0.753 ( 0.011

4-CP 2-CP 3-CP 2,4-DCP 3,4-DCP 2,3-DCP 2,3,4-TCP 2,5-DCP 2,4,5-TCP 2,6-DCP 3,5-DCP 2,4,6-TCP 3,4,5-TCP 2,3,5-TCP 2,3,6-TCP 2,3,4,6-tetraCP 2,3,5,6-tetraCP PCP

0.273 ( 0.015 0.813 ( 0.015 0.753 ( 0.021 0.703 ( 0.015 0.513 ( 0.006 1.49 ( 0.14 1.49 ( 0.16 1.12 ( 0.05 1.05 ( 0.04 1.47 ( 0.10 0.39 ( 0.11 1.09 ( 0.07 0.12 ( 0.02 1.54 ( 0.08 1.55 ( 0.13 1.41 ( 0.04 2.00 ( 0.26 1.90 ( 0.09

103 ( 11 108 ( 3 100 ( 4 93 ( 4 68 ( 3

2,3-dichloro-1,4-benzoquinone (2,3-diClBQ) 2,5-dichloro-1,4-benzoquinone (2,5-diClBQ)

1.62 ( 0.04

2,6-dichloro-1,4-benzoquinone (2,6-diClBQ)

1.45 ( 0.15

trichloro-1,4-benzoquinone (triClBQ) tetrachloro-1,4-benzoquinone (tetraClBQ)

1.93 ( 0.07

Figure 3b shows the chromatogram obtained from a mixture of four CPs, which generated four different chlorinated benzoquinones. In another HPLC run, where acetonitrile was reduced from 35 to 15%, the retention time of the reaction products of 2,3-DCP and 2,3,4-TCP was significantly longer, as expected (figure not shown). The reaction products of 2,3-DCP and 2,3,4-TCP exhibited only one single peak, just after those of the other dichlorobenzoquinones (18.0 and 19.1 min), with a retention time of 22.2 min. Calibration of 2,3-DCP resulted in a response that was linear up to 20 µM, with an integrated area of 0.152 ( 0.010 µA min µM-1, and a detection limit of 250 nM. When CPs were converted separately, only one electroactive compound was detected, and the seven groups, each of which yielded products with the same retention time in the chromatogram, were 4-CP, (2-CP/3-CP/2,4-DCP/3,4-DCP), (2,3-DCP/2,3,4TCP), (2,5-DCP/2,4,5-TCP), (2,6-DCP/3,5-DCP/2,4,6-TCP/3,4,5TCP), (2,3,5-TCP/2,3,6-TCP/2,3,4,6-tetraCP), and (2,3,5,6-tetraCP/ PCP). When standard chlorinated benzoquinones were available, conversion yields were calculated (see Table 2). High conversion (100%) is observed in the case of pentachlorophenol and 2,3,5,6tetraCP. Based on the results obtained and from the literature,21-23 a plausible reaction mechanism between phenolic compounds and BTFAIB can be suggested as below:

As shown in this reaction, aryloxenium ion formation (2) was 4328

Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

69 ( 7 64 ( 6 101 ( 17 27 ( 39 75 ( 17 8 ( 27

104 ( 17 98 ( 8

suggested21 and could explain the low yield observed with 3,5DCP and 3,4,5-TCP. Chlorine in the ortho or para position would increase the aryloxenium stability (predominant resonance effect), while chlorine in the meta position would decrease stability (electron-withdrawing induction effect). However, presence of chlorine in the para position strongly affected kinetics by decreasing the p-benzoquinone formation rate. For example, the reaction with 2,3,5,6-tetraCP (10 µM) was 80% complete in the first minute, and the reaction was finished after only 2.5 min, whereas the PCP reaction required 25-30 min to complete and was only 20-25% complete in the first 2.5 min. Glucose Oxidase Electrode Response. Previous studies have shown that several CPs could be oxidized to quinoid compounds using ceric sulfate or hydrogen peroxide/chloroperoxidase systems, and the resulting products could be used as mediators between glucose oxidase and the electrode surface.9 In fact, high responses have been obtained in the case of the reaction products of 2,4,6-TCP and 2,3,4,6-tetraCP. However, reaction products of other highly chlorinated phenols, including PCP, were not able to generate any electrode response, either because of lack of CP reactivity or lack of product electroactivity. One interesting result reported here is the capability of chlorobenzoquinones to mediate glucose oxidase in the presence of excess glucose. For example, glucose oxidase immobilized on a glassy carbon electrode was readily reduced by its substrate, glucose. Tetrachloro-1,4-benzoquinone then recycled the reduced enzyme to its original active form, i.e., mediating the rate-limiting electron transfer from the enzyme to the electrode. This resulted in a change in the plot, i.e., the peaks completely disappeared, and, instead, a large catalytic current flows at the oxidizing potential (Figure 4A, curves a and b). The product obtained from the PCP-BTFAIB reaction was also able to mediate the glucose oxidase-glucose reaction, and a similar catalytic wave was obtained (Figure 4B). Conditions of measurement were then optimized to obtain the maximal response of a glucose oxidase electrode with p-benzoquinones from highly chlorinated phenols. Figure 5 shows the effect of the pH on the current responses with the four commercially available chlorinated 1,4-benzoquinones. Half-wave potentials of benzoquinones were pH-dependent; however, a sufficiently high potential (+0.45 V) was used to obtain an efficient

Figure 6. Steady state sensitivities of a GOX electrode to (a) various benzoquinones and (b) converted chlorophenols. Conversions were made in 0.1 M trichloroacetic acid, pH 1.5, during 30 min, with 10 µM chlorophenol and 500 µM BTFAIB concentrations. Standards and reaction products were diluted 50× in deoxygenated 0.1 M tartrate buffer, pH 3.5, containing 40 mM glucose. Potential was +0.45 V vs Ag/AgCl.

Figure 4. (A) Cyclic voltammograms of the GOX electrode in 1 µM tetrachloro-1,4-benzoquinone standard solution (in 9 mL of 0.3 M phosphate buffer, pH 5, and 1 mL of 0.1M acetate buffer, pH 3, at a scan rate of 2 mV s-1) before (a) and after (b) addition of 40 mM glucose. (B) Cyclic voltammograms of the GOX electrode in 1 µM PCP reaction solution, i.e., anticipated tetrachloro-1,4-benzoquinone (in 9 mL of 0.3 M phosphate buffer, pH 5, and 1 mL of 0.1 M acetate buffer, pH 3, at a scan rate of 2 mV s-1) before (a) and after (b) addition of 40 mM glucose.

Figure 5. pH dependence of the GOX electrode sensitivity to (b) 2-ClBQ, (9) 2,5-diClBQ, (4) 2,6-diClBQ, and (O) tetraClBQ. Glucose concentration was 40 mM, and applied potential was +0.45 V vs Ag/ AgCl. Tartrate buffer, pH 3.5, was prepared by mixing 0.1 M tartaric acid and 0.1 M sodium tartrate. Benzoquinone concentrations were 200 nM.

electron transfer between the mediator and the electrode surface. Remarkably, as the number of chlorines on the aromatic ring of the mediator increases, the electronic charge density of the aromatic ring decreases, and the optimum response of the glucose oxidase electrode is obtained at more acidic pH conditions. In the case of tetrachloro-1,4-benzoquinone, maximal current re-

sponses are obtained at pH 3.5. Glucose oxidase usually exhibits a less acidic pH optimum with mediators such as oxygen, quinones24 (pH 5.5), or biphenol25 (pH 7.5). However, with dichlorophenolindophenol, a low charge density compound, glucose oxidase also exhibits an acidic optimum.24 At pH 3.5, current responses (Figure 5) increased as the chlorine content of the benzoquinone mediator increased. Differences observed between Figures 5 and 6a were caused by the use of different glucose oxidase electrodes. The response time of enzymatic electrodes also increased with chlorine number. In fact, only 1.1 min is necessary to reach 95% of the steady state current with 2-ClBQ as the mediator, while 4.3 min is required with tetraClBQ. For the purpose of detection and determination of PCP and other highly chlorinated phenols, all subsequent experiments were carried out at pH 3.5. At this condition, linear behavior of the current response was observed up to 200 nM for all CP oxidation products. Figure 6b shows current responses of the glucose oxidase electrode to the CP reaction products with BTFAIB. For each CP (10 µM), three reactions were carried out at the same time. CPs generating the same chlorinated pbenzoquinone and the commercial standard were studied on the same day. As expected, the glucose oxidase electrode was more sensitive to the oxidation compounds of highly chlorinated phenols. In the case of tetraClBQ generating CP, the relative standard deviation observed between three different reaction solutions is less than 3% and comparable to that for the standard p-quinone. The detection limit (S/N ) 3) in the case of pentachlorophenol is 4 nM, close to the maximum allowable contamination level of pentachlorophenol in water samples (2.7 nM). This sensitivity is favorably compared to the 13.3 nM detection limit obtained with the PCP immunoassay technique26 and could be improved by optimizing the electrochemical detection system (reduction of the electrode area for example). The glucose oxidase-based electrode also possessed very good sensitivity toward 2,3,4,6-tetraCP, 2,3,5,6-tetraCP, 2,3,5-TCP, and 2,3,6TCP (Table 3). However, as can be observed from Figure 6b, the detection limits of the other CPs would be higher due to the (24) Wilson, R.; Turner, A. Biosens. Bioelectron. 1992, 7, 165-185. (25) Fraser, D. Anal. Lett. 1994, 27, 2039-2053. (26) Van Emon, J.; Rock, S. PCP Immunoassay Technology; U.S. Environmental Protection Agency: Washington, DC, 1995.

Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

4329

Table 3. Detection Limits of the Glucose Oxidase-Based Amperometric Electrode for 18 Chlorophenols chlorophenol

limit of detection (nM)

chlorophenol

limit of detection (nM)

4-CP 2-CP 3-CP 2,4-DCP 3,4-DCP 2,3-DCP 2,3,4-TCP 2,5-DCP 2,4,5-TCP

65 17 20 16 12 10 7 12 11

2,6-DCP 3,5-DCP 2,4,6-TCP 3,4,5-TCP 2,3,5-TCP 2,3,6-TCP 2,3,4,6-tetraCP 2,3,5,6-tetraCP PCP

10 26 9 38 5 5 4 4 4

lower current response for the corresponding chlorinated benzoquinones (Table 3). Glucose oxidase has been known as a very stable enzyme; therefore, no attempt was made to study its stability. However, it was observed that each glucose oxidasebased electrode could be used for at least 60 repeated analyses without losing its activity. CONCLUSIONS This article demonstrated for the first time that highly chlorinated phenols can be oxidized at ambient temperature under easy-to-control conditions to quinones with a very high yield. The reaction products or chlorinated quinones have proven to be very effective mediators for glucose oxidase in the presence of excess glucose. This combined chemical and enzymatic reaction can be (27) Hilmi, A.; Luong, J. H. T.; Nguyen, A. L. J. Chromatogr. A 1997, 761, 259268. (28) Hilmi, A.; Luong, J. H. T.; Nguyen, A. L. Environ. Sci. Technol. 1997, 31, 1794-1800.

4330 Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

easily exploited in electrochemistry using glucose oxidase immobilized on the surface of a working electrode for sensitive detection and determination of such chlorinated pollutants. The electrocatalytic configuration can be easily incorporated into a fully automated flow injection analysis system and could be applied for screening various highly chlorinated phenols, especially for PCP, tetrachlorophenols, 2,3,4-trichlorophenol, and 2,4,6-trichlorophenol in contaminated biological and environmental samples. Indeed, this type of flow injection analysis (FIA) biosensor has been demonstrated in our laboratory to analyze several chlorophenols with a sample throughput as high as one per minute.9 With a capillary electrophoresis instrument upstream, the biosensor reported here could be extended for analysis of chlorophenols in mixtures. As a matter of fact, a simple endcolumn amperometric detector, without a porous junction, has been designed and attached to a capillary electrophoresis instrument to analyze eight chlorophenols with concentrations as low as 0.10 µM.27 The CE system equipped with such a detection cell was also effective in separating and quantifying several chlorophenols, including pentachlorophenol, in real aqueous samples and soil extracts.28 Work is in progress to immobilize glucose oxidase onto the electrode surface to improve the detection limit of the amperometric detector, and this combined separation and detection system will be applied for detection and determination of all 18 chlorinated phenols, including pentachlorophenol, in contaminated soil and water and in other samples.

Received for review April 7, 1997. Accepted August 23, 1997.X AC970369F X

Abstract published in Advance ACS Abstracts, October 1, 1997.