- Email: [email protected]
Electrochemical detection of the herbicide paraquat in natural water and citric fruit juices using microelectrodes
Analytica Chimica Acta 546 (2005) 85–91
Electrochemical detection of the herbicide paraquat in natural water and citric fruit juices using microelectrodes D. De Souza ∗ , S.A.S. Machado Instituto de Qu´ımica de S˜ao Carlos, Departamento de Fisico Quimica, Universidade de S˜ao Paulo, C.P. 780, 13560-970 S˜ao Carlos, Sa Paulo, Brazil Received 29 November 2004; received in revised form 9 May 2005; accepted 10 May 2005 Available online 13 June 2005
Abstract A novel electroanalytical procedure for detecting the paraquat herbicide in natural water and citric fruit juice samples using gold microelectrodes and square wave voltammetry at high frequencies is proposed. The results obtained showed two reversible peaks for the reduction of paraquat, the first peak associated with the reduction of the paraquat molecule in solution, with subsequent adsorption of the intermediate on the electrode surface. This adsorbed species was shown to undergo electroreduction in a reaction associated to the second voltammetric peak. The variation in pH and square wave parameters indicated that the best conditions under which paraquat could be reduced were a pH of 5.0, a frequency of 1000 s−1 , a scan increment of 2 mV and a square wave amplitude of 50 mV. Under these conditions, the variation of the concentrations of paraquat from 1.00 × 10−6 to 1.66 × 10−4 mol L−1 presented, for peak 1, detection and quantification limits of 4.51 and 15.05 g L−1 respectively in pure electrolyte with a recovery factor of 99.50%. The proposed analytical procedure was also applied to natural water samples giving recovery factors of 95.00, 89.50 and 92.50% in three water samples collected from an urban stream. The recovery factor was observed to depend on the content of organic matter which was determined by the biochemical and chemical oxygen demand. In lemon and orange juice samples that were spiked with 5.70 × 10−5 mol L−1 of paraquat, the recovery factors obtained were 94.30 and 92.70% respectively. © 2005 Elsevier B.V. All rights reserved. Keywords: Square wave voltammetry; Paraquat; Microelectrodes
1. Introduction The continuous increase in agricultural production promotes an equivalent increase in the level of pesticide residues in waters, soils and foodstuff. In this light, the development of analytical methods to monitor different types of pesticides and pesticides residues in several samples of biological or environmental interest acquires a fundamental importance. It is also known that contamination by pesticides is not restricted only to soil or water, but also to the consumption of contaminated foods [1,2] or water [3]. Pesticide residues have been frequently found in blood [4], urine [5] or even in maternal milk [6] and such matrices must be considered in any ∗
Corresponding author. Tel.: +55 162739924; fax: +55 162739952. E-mail address: [email protected] (D. De Souza).
0003-2670/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2005.05.020
analytical procedure. As a consequence, several precise and sensitive analytical procedures are now under development worldwide for their application in either “in situ” or analytical laboratories for the quantification of pesticides and their residues in a large number of different matrices. Paraquat, (1,1 -dimethyl-4,4 -bipyridilium dichloride), also known as methyl-viologen (MV), is a bipyridilium pesticide of toxicological class I and is extremely hazardous for human health. However, it is one of the most largely used pesticides in over 130 countries worldwide. The paraquat include clearing, pasture renovation, inter-row weed control in vegetable crops and weed control in plantation crops. A major problem caused by the abusive and uncontrolled use of paraquat is related to its high persistence in the environment where, though only slightly absorbed by soils, it is
86
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
a potential contaminant of waters due to its high solubility (about 620 g L−1 at 25 ◦ C [7]). Literature reports show that paraquat has been analyzed using a wide range of different analytical techniques, such as spectrophotometry [8], flux analysis [9], liquid chromatography-(electrospray ionization) mass spectrometry [10], polarography [11] and immuno-assay [12]. The electrochemical detection of paraquat has been performed at different electrode surfaces, including solid electrodes [13] and Nafion® modified electrodes [14]. Walcarius and Lamberts [11] developed a fast and sensitive method to determine paraquat in aqueous media using the dropping mercury electrode and square wave voltammetry. The effects of the voltammetric parameters were evaluated, their optimization yielding a detection limit of 1.50 × 10−8 mol L−1 without any pre-concentration step. Paraquat being a cation, the detection limit can be lowered by using an electrode surface covered with a cationic ion-exchange resin. Based on this, Lu and Sun [14] used a Nafion® modified glass carbon electrode for the analytical determination of paraquat using accumulation prestep followed by detection by cathodic redissolution in a pesticide free electrolyte. Employing differential pulse voltammetry as the analytical technique, a detection limit of 2.00 × 10−9 mol L−1 was reached. A similar cationic characteristic of paraquat was used by Derouane et al. [15] to analyze the electrochemical behavior of paraquat at carbon paste electrodes modified by zeolites. The electrochemical responses were explored in several supporting electrolytes using cyclic voltammetry, chronoamperometry, square wave voltammetry and chronocoulometry. It was observed that the electrochemical responses were strongly affected by the initial conditions and by the thickness of the carbon paste used in preparing the electrode. Monk et al. [13] studied the electrochemical behavior of paraquat in aqueous media using cyclic voltammetry. Two voltammetric peaks were detected during the negative sweep, the first at around −0.70 V and the second at approximately −1.20 V versus SCE. The first peak was attributed to the radical cation formation (MV2+ +e− MV+ ) and the second to the neutral species formation (MV+ + e− MV0 ) which, according to the authors, was followed by a chemical dimerization step. Both electrochemical steps were considered as totally reversible processes. This same conclusion was reached by Rueda et al. [16] while using the electrochemical impedance spectroscopy. Finally, de Souza et al. [17] applied microelectrodes and square wave voltammetry in the electroanalytical determination of paraquat in pure electrolyte, reaching a determination limit as low as 7.10 ppb. The observed result was proof the possibility of applying electroanalytical methods in more complex samples. The aim of the present paper is study the applicability of microelectrodes and square wave voltammetry (SWV) in the electroanalytical determination of paraquat in natural waters and in citric fruit juices, without any pre-purification step.
2. Experimental 2.1. Equipment and reagents Electrochemical measurements were carried out using a Voltalab potenciostat/galvanostat (model PGZ402, Radiometer Analytical Inc.) controlled by Voltamaster 4 software (Radiometer Analytical Inc.). A Ag/AgCl (3.00 mol L−1 KCl) system was employed as the reference/auxiliary electrode and the working microelectrodes were constructed by sealing a 25 m diameter gold micro-wire in Epoxy resin as detailed below. A Methron model 682 pH-meter with the glassAg/AgCl/KCl (3.00 mol L−1 ) combined electrode was used for adjusting pH values. All solutions were prepared with water purified in a MilliQ system from Millipore Corporation. Prior to all measurements, solutions were deaerated by passing N2 gas (SS White Martins) for approximately 15 min. Stock solutions of paraquat (99.80%, Merck) were prepared by dissolving the pesticide, without any additional purification step, in water. All the others reagents used were of analytical grade. 2.2. Microelectrodes construction and characterization The gold microelectrodes were constructed by embedding a 25 m diameter gold wire (Goodfellow) in a Pyrex® glass tube of 0.50 mm internal diameter with Epoxy resin. The tips of the gold microelectrodes were polished with emery paper until a metal micro disc was exposed at the surface. After this procedure the microelectrodes were cleaned with purified water prior to use. The voltammetric characterization consisted of analyzing the electrochemical response of a gold microelectrode in an electrolyte containing 1.00 × 10−3 mol L−1 of potassium hexacyanoferrate (III) in 0.10 mol L−1 of KCl with a pH of 3.00. The voltammograms obtained exhibited the well-known sigmoid profile which is characteristic of an electrochemical process controlled by spherical diffusion mass transport, as expected when using microelectrodes [18]. 2.3. Working procedure The initial working procedure consisted of measuring the electrochemical response of the gold microelectrode in a 0.10 mol L−1 of Na2 SO4 electrolyte at a fixed concentration of pesticide. Several supporting electrolytes were tested (Britton–Robinson buffer, Na2 SO4 , phosphate, perchlorate), the best electrochemical response measured in terms of the highest analytical signal and improved reproducibility was obtained in Na2 SO4 . This means that 0.10 mol L−1 of Na2 SO4 was used throughout the experimental program as the supporting electrolyte. The electrochemical cell was then placed in a Faraday cage in order to minimize the contribution of background noise to the analytical signal.
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
The square wave parameters: frequency of application of the pulse potential (f ), pulse amplitude (a) and scan increment (Es ) were optimized since these parameters exert a strong influence on the sensitivity of voltammetric analysis [19]. After optimization of the voltammetric conditions, calibration curves were obtained by adding aliquots of the pesticide to a pure laboratory electrolyte, the supporting electrolyte response denoted as blank. After consecutive adding aliquots of the pesticide, each of whose concentration was properly adjusted, the peak currents due to the reduction of the pesticide were plotted against its concentration in solution. The recovery experiments were then carried out in a similar way by adding a known amount of paraquat to the supporting electrolyte followed by standard additions. The suitability of the electroanalytical procedure in determining paraquat in natural matrices was tested by spiking water samples collected from three distinct points of the Monjolinho stream which crosses the district of S˜ao Carlos in the State of S˜ao Paulo, Brazil. The first sampling point (point 1), located near the stream head and before the city limits is known to be of a very low degree of pollution. The second sampling point (point 2), along the stream length is located around the urban area where the degree of pollution a high while the third sampling point (point 3), is located after crossing the city thus containing a high degree of organic matter including domestic and industrial pollution received from point 2. After contamination, the same electroanalytical procedure described above for the pure electrolyte was applied to the natural water samples. Orange and lemon juices were extracted from their respective fruits and used without any pre-treatment or separation step other than adjusting the solution pH to 5.0 with NaOH solution. The natural electrolytes were artificially contaminated with a known amount of paraquat and recovery measurements performed. All experiments were performed in triplicate.
3. Results and discussion
87
Fig. 1. Steady-state cyclic voltammograms for 4.22 × 10−3 mol L−1 of paraquat in 0.10 mol L−1 of Na2 SO4 , pH 5.0, on the gold microelectrode using different scan rates (as indicated).
This electrode process was followed by a second one, represented by peak 2 at −0.93 V and is associated to a quasireversible adsorption controlled process hence generating the neutral molecule [20]: MV+ + e− MV0
(2)
and the subsequent formation of a dimmer: MV0 + MV2+ MV2 2+
(3)
The SWV experiments yielded the same reduction and oxidation processes described in Eqs. (1)–(3). Here peak 1 was observed at −0.64 V and peak 2 at −0.94 V versus Ag/AgCl 3.00 mol L−1 , which is in close agreement with values obtained in cyclic voltammetric experiments. The main difference observed between SWV and the cyclic voltammetric experiments is that, like peak 1, the electrochemical process related to peak 2 presents a totally reversible behavior indicated by the presence of direct and reverse currents with similar values of peak current. This is probably associated to the very fast rate of pulse applications which inhibits the chemical dimer formation step. This would make the second electron transfer also reversible thus in contrast to the cyclic voltammetric experiments.
3.1. Electrochemical behavior 3.2. Optimization of experimental parameters The cyclic voltammetric responses for a gold microelectrode in 0.10 mol L−1 of Na2 SO4 + 4.22 × 10−3 mol L−1 of paraquat at different scan rates are presented in Fig. 1. In the figure, waves 1 and 5, which occur at approximately −0.70 V, represent the first process observed during the potential scan from −0.30 to −1.20 V. As the scan rate increases, these waves turn into peaks as is expected for a mass transport controlled electrochemical process occurring at microelectrodes. Waves 1 and 5 can thus be associated to the redox couple [20]: MV+2 + e− MV+
(1)
The decrease in proton concentration as pH increases from 2.0 to 9.5 is shown to exert a pronounced effect on the SWV response of paraquat. Up to a pH of 5.5, the peak currents show a five-fold increase for peak 1 and a two-fold for peak 2. The peak potentials do not however seem to be affected by the concentration of H+ , suggesting the absence of any protonation step in the reduction mechanism as indicated by Eqs. (1)–(3). A variation in the frequency of potential pulse application usually exerts a marked effect on the SWV response, thus providing a criterion for diagnosis that can be used to
88
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
Fig. 2. Square wave voltammograms for 5.00 × 10−5 mol L−1 of paraquat in 0.10 mol L−1 Na2 SO4 at a gold microelectrode with pH = 5.0 and a = 50 mV, Es = 2 mV at different frequencies. Insert: relationship between the frequency and peak current obtained for peaks 1 and 2.
Fig. 3. Calibration curves for paraquat in pure water electrolyte (pH 5.0) using f = 1000 s−1 , a = 50 mV, Es = 2 mV, considering the current responses of peaks 1 and 2.
Ep −2, 3RT = log f αnF
indicate any adsorption or reaction in solution, reversibility of the electrochemical process, etc., [19,22]. Fig. 2 presents the effect of frequency on the SWV response for the reduction of paraquat at a gold microelectrode in 0.10 mol L−1 of Na2 SO4 + 4.20 × 10−4 mol L-1 of paraquat with a = 50 mV and Es = 2 mV. As can be observed, an increase in the pulse frequency is accompanied by an increase in current for both peaks 1 and 2. Moreover, the peak potentials of peak 2 show a shift towards more negative values while those of peak 1 remains practically constant, a behavior which indicates that peak 1 refers to a reaction in solution while peak 2 involves adsorption as the rate determining step. This is also confirmed by the dependence of the peak current on the square root of the frequency for peak 1 and a direct dependence on the frequency for peak 2 as demonstrated in the insert of the Fig. 2 [19]. The advantages of using microelectrodes with SWV for the determination of paraquat must be highlighted. Previous studies on the electroanalytical determination of paraquat: Zen and co-workers [21] with a Nafion® coated solid electrode or Walcarius and Lamberts [15] with HMDE, were restricted to a frequency range below 200 s−1 due to the severe signal distortion observed above this limit value. However, in the present study, the use of a gold microelectrode is shown to permit the application of frequencies as high as 1000 s−1 without any loss of quality in the electrochemical signal. This is probably associated with the fast diffusion of species to or from the electrode surface. The high frequency used means an increase in sensitvity is obtained when the proposed methodology is employed. The influence of frequency on the peak potential permits the determination of the number of electrons transferred during the electrode process when the following equations, developed respectively for totally reversible reactions in solution and with adsorption, [19] are used:
DL =
3Sb b
(6)
QL =
(7)
Ep −2, 3RT = log f 2αnF
10Sb b
where Sb is the standard deviation of current measured for the blank solution at the same potentials as those of
(4)
(5)
where R is the gas constant, T the temperature, α the transfer coefficient, n the number of electrons, and F is the Faraday constant. From the equation, the slope of the straight line obtained for the relationships Ep versus log f1/2 for peak 1 and Ep versus log f for peak 2 result in the value of n = 1 (considering α = 0.50), thus confirming the reaction pathway presented in Eqs. (1)–(3). The other two parameters employed in SWV, the amplitude (a) and the potential step (Es ), both exerts a much less effect on the peak current or potential. Following this, the respective values for these parameters were taken as 50 mV for a, as suggested by SWV theory [22], and 2 mV for Es . 3.3. Analytical methodology for paraquat in pure water Using the obtained optimized parameters mentioned above, a calibration curve was obtained for paraquat in an electrolyte prepared with pure water. For this, aliquots from the stock solution were consecutively added to the electrochemical cell. The SWV responses were recorded for a range of concentration between 1.00 × 10−6 and 1.66 × 10−4 mol L−1 . The SWV responses together with the linear relationships obtained between peak currents and concentrations (as the insert) for both peak 1 and 2 are presented in Fig. 3. From Fig. 3, the detection (DL) and quantification limits (QL) were obtained for the experimental conditions employed using the following equations from IUPAC [23]:
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
the peak currents (in this case for an average of eight values of current) and b is the slope of the analytical curve. The observed values of DL and QL for peak 1 were respectively 1.76 × 10−8 mol L−1 (4.51 g L−1 ) and 5.86 × 10−8 mol L−1 (15.05 g L−1 ). The same calculations for peak 2 yielded 3.82 × 10−8 mol L−1 (9.82 g L−1 ) and 1.27 × 10−7 mol L−1 (32.63 g L−1 ) mol L−1 , respectively. Thus, peak 1 showed a greater sensibility for the determination of paraquat than peak 2. The reproducibility of the proposed methodology was determined from five different measurements in the same solution containing 8.00 × 10−5 mol L−1 of paraquat and 0.10 mol L−1 of Na2 SO4 obtaining coefficients of variation of 1.70 and 1.20% for peaks 1 and 2, respectively. The repeatability of this method was also determined from measurements performed for 10 different solutions of the same composition as stated above, obtaining coefficients of variation of 1.60 and 1.20% for peaks 1 and 2, respectively. 3.4. Application of this electroanalytical methodology to natural waters The analytical procedure presented above was applied to natural water samples collected from different points of the Monjolinho stream as described in the Section 2. The sampling points were selected based on different levels of organic matter including industrial and domestic pollution. The water samples were used as received in preparing the supporting electrolyte (adding 0.10 mol L−1 of Na2 SO4 ) and the analytical curves were again obtained by SWV experiments. In this way, the influences of sample contamination were evaluated. The analytical curves obtained under these new experimental conditions are presented in Fig. 4. The results of linear regression from the straight lines obtained presented in Fig. 4 are shown in Table 1. The results obtained indicate that peak 1 is practically unaffected by the presence of contaminants in the natural samples. The analytical sensitivities, defined by the slope of the analytical curves, are constant when compared to those obtained for a pure water electrolyte independent of the origin of the sample. This can be associated with the fact that the electrochemical process for peak 1 occurs with the species
89
Fig. 4. Calibration curves for paraquat in electrolytes prepared with different natural water samples and spiked with paraquat under the experimental conditions presented in Fig. 3.
in solution, thus being independent of adsorption at the electrode surface. On the other hand, peak values of current for peak 2 displayed some influence due to organic matter present in the water samples. Only the analytical curve obtained with water collected from point 1 (free from urban pollution) was comparable to that obtained for purified water. This behavior is expected since peak 2 reflects an electrochemical process involving adsorption of reagents and/or products which can be inhibited by the presence of other organic molecules. In order to evaluate the amount of organic molecules in the stream water samples, analyses of biochemical oxygen demand (BOD) and chemical oxygen demand (COD) were performed for the natural water samples. The results are also presented in Table 1 where it can be observed that the values of s (the sensitivity of the analytical method) decreases with increasing values of BOD and COD. This reinforces the hypothesis that the presence of organic matter interferes in the methodology developed for peak 2. Recovery experiments were also performed in order to evaluate the interference of organic and inorganic components in natural water matrices. Samples were spiked with 2.00 × 10−5 mol L−1 of paraquat and the recovery curves obtained by the standard addition method. The presence of
Table 1 Results of linear regression obtained from Fig. 4 for the square wave voltammograms at the gold microelectrode in natural waters samples, recovered concentration obtained by samples spiked with 2.00 × 10−5 mol L−1 of paraquat and determination of BOD and COD for dissolved O2 Parameters
r Sb × 10−11 (A) s × 10−3 (A/mol L−1 ) Recovery (%) R.S.D. (%) BOD (mg L-1 ) COD (mg L-1 )
Point 1
Point 2
Point 3
Peak 1
Peak 2
Peak 1
Peak 2
Peak 1
Peak 2
0.9994 7.32 5.78 95.00 1.50
0.9966 36.90 8.28 95.00 1.75
0.9985 30.00 5.27 89.50 1.60
0.9967 43.60 4.08 91.00 1.80
0.9993 15.80 5.55 92.50 1.60
0.9982 41.02 5.50 94.50 1.80
6.00 19.00
12.00 33.00
9.00 29.00
90
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91 Table 2 Results obtained from the linear regression curves (Fig. 5) for the determination of paraquat in lemon and orange juices at a gold microelectrode Parameters
r Sb × 10−11 (A) s × 10−3 (A/mol L−1 ) Repeatability (%) Reproducibility (%) Recovery (%)
Fig. 5. Square wave voltammograms for paraquat in electrolyte prepared with lemon juice at a gold microelectrode with f = 1000 s−1 , a = 50 mV, Es = 2 mV and different concentrations of paraquat.
this pesticide was not detected in natural waters used as collected. The results obtained for peak 1 and peak 2 (Table 1) presented the same inverse dependence on the amount of organic matter presented in different samples. Following Ib´anez [24] and Hesketh [25], paraquat strongly interacts with humic substances as well as with surfactants present in natural water samples, thus promoting a negative effect on the recovery values. 3.5. Applications in orange and lemon juices “in natura” Paraquat is one of the most widely used herbicides in citric fruit cultures. Due to its toxicity and persistence in the environment, there is a very high probability of its contaminating crops, thus presenting a risk to human health. Because of this, the determination of paraquat in “in natura” citric fruits juice serves as a means of minimizing the hazardous effect of such pesticide. The methodology developed in the preceding sections was hence applied to “in natura” orange and lemon juices. The juices were extracted from their respective fruits without any pre-separation or pre-concentration. The juices had their pH adjusted to 5.00 with an appropriate volume of a 3.00 mol L−1 of NaOH solution and then added to the electrochemical cell. The SWV response curves were later obtained in the range concentration of 1.00 × 10−5 to 25.00 × 10−5 mol L−1 of paraquat. The observed SWV responses are shown in Fig. 5 for lemon juice with f = 1000 s−1 , a = 50 and 2 mV. Similar results obtained for orange juice are also shown. As it can be observed, peak 1 is free from interference. However, peak 2 overlays a background current, maybe related to other matrix components which, in the case of citric fruits, are basically ascorbic acid (Vitamin C) and citric acid. From the relationship between peak current and the concentration of paraquat, calibration curves were obtained, which, for lemon juice, are presented in Fig. 5 (insert) for
Lemon
Orange
Peak 1
Peak 2
Peak 1
Peak 2
0.9993 4.30 1.03 3.40 1.00 94.30
0.9976 126.00 1.74 2.70 1.01 84.40
0.9983 5.54 1.14 3.85 1.15 92.70
0.9972 215.00 2.12 2.90 2.05 77.90
peaks 1 and 2. In order to obtain the correct values for peak currents, the base line was subtracted from each curve. Reproducibility, for n = 5, repeatability, for n = 10 and recovery experiments (by the standard addition method, as described above) were also performed on citric fruit juice electrolytes using a concentration of paraquat of 5.70 × 10−5 mol L−1 . The obtained results are all included in Table 2. A comparison of the analytical sensibility of the methodology (defined by the slopes of the calibration curves) obtained for pure water and citric fruit juice shows that the analytical sensibility of citric fruits is 20-fold that of pure water. This can be attributed to the adsorption of the matrix components at the electrode surface or to interactions between the pesticide molecule and components of the juice. In order to evaluate any possible interference of citric and ascorbic acids in the reduction of paraquat at a gold microelectrode, electrolytes containing a known concentration of each of these acids in pure water were recorded and plotted together with that of paraquat. The results obtained demonstrated that the second reduction peak of paraquat is more likely distorted due to the presence of those contaminants. The fact that their peaks occurs at approximately −0.85 V indicates that an interference only in the second peak.
4. Conclusions A gold microelectrode in a square wave voltammetric procedure was successfully applied in quantifying paraquat in natural matrices. In the present work, the the primary goal, on using microelectrodes, was to verify the possibility of obtaining responses in SWV frequencies as high as 1.000 s−1 which are much more convenient since at this level, the analytical signal is proportional to the frequency. In fact, no such electroanalytical analysis has yet been reported. The novelty of the proposed methodology is rooted in the use of microelectrodes in citric fruits juice without any pre-treatment or extraction procedure. This, in the authors’ view, is the first time such a measurement which serves as an important new tool for the direct determination of pesticides in “in vivo” fruit is being reported. To the best of the authors’
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
knowledge, no other methodology that permits such a direct measurement in growing plants or maturing fruits is available. The first step for such a procedure is however reported here. Of course, had microelectrodes not been capable of detecting paraquat without any change in the fruit juices, the direct determination of paraquat would not have been possible. However, as the present work shows, microelectrodes do actually respond very well in pure fruits juices, resulting in detection and quantification limits lower than permitted by Brazilian environmental council. Initially, the electrochemical behavior of paraquat was evaluated using the voltammetric results. The SWV response showed two reduction waves associated to the partial reduction of paraquat molecule followed by the formation of dimer species in solution. These processes were determined as being reversible, the first being controlled by diffusion and the second by electron transfer to an adsorbed species. This is the most suitable situation for SWV analysis since the resulting current is the sum of forward and reverse contributions, the latter only being present in reversible processes, a situation which is far from common in pesticide analysis. Analysis on a purified laboratory electrolyte allows the determination of a very low detection limit of 4.51 g L−1 associated with a high level of repeatability and reproducibility. Because of this, the electroanalytical methodology was applied in determining paraquat in spiked stream water samples which presented different degrees of pollution. The influence of organic matter (pollution or natural) content on the sensitivity of the analytical procedure was clearly established by determining BOD and COD of the samples. Finally, the same analytical procedure was applied in determining paraquat in citric fruit juices where it was demonstrated that either ascorbic acid or Vitamin C exerts detectable effects on the recovery experiments of spiked samples that were obtained without any pre-treatment. In this way, the use of microelectrodes in conjunction with SWV in determining paraquat seems to present a very suitable analytical methodology which permits its application in a variety of different natural matrices.
91
Acknowledgements The authors acknowledge the financial support from FAPESP (proc. 03/12926-3 and 04/00839-1) and CNPq.
References [1] A. Saba, F. Messina, Food Qual. Prefer. 14 (2003) 637. [2] H.G.J. Mol, R.C.J. Van Dam, O.M. Steijer, Chromatogr. A 1015 (2003) 119. [3] E. Ballesteros, M.J. Parrado, J. Chromatogr. A 1029 (2004) 267. [4] U. Heudorf, J. Angerer, H. Drexler, Int. J. Hyg. Environ. Health 206 (2003) 485. [5] J.M. Pozzebon, W. Vilegas, I.C.F. Jardim, J. Chromatogr. A 987 (2003) 375. [6] M. Jimenez, A. Jerez, J.M. Diaz, M.F. Olea, R. Bayes, C. Campoy, Pediat. Res. 54 (2003) 613. [7] E. Halfon, S. Galassi, R. Br¨uggemann, A. Provini, Chemosphere 33 (1996) 1543. [8] M.K. Ra´ı, J.V. Das, V.K. Gupta, Talanta 45 (1997) 343. [9] B. Saad, M. Ariffin, M.I. Saleh, Talanta 47 (1998) 1231. [10] V.Y. Taguchi, S.W.D. Jenkins, P.W. Crozier, D.T. Wang, J. Am. Soc., Mass Spectrom. 9 (1998) 830. [11] A. Walcarius, L. Lamberts, J. Electroanal. Chem. 406 (1996) 59. [12] M.J. Nuwer, J.J. O’Dea, J. Osteryoung, Anal. Chim. Acta 362 (1998) 03. [13] P.M.S. Monk, C. Turner, S.P. Akhtar, Electrochim. Acta 44 (1999) 4817. [14] T.H. Lu, I.W. Sun, Talanta 53 (2000) 443. [15] A. Walcarius, L. Lamberts, G. Derouane, Electrochim. Acta 38 (1993) 2257. [16] M. Rueda, R.G. Compton, J.A. Alden, F. Prieto, J. Electroanal. Chem. 443 (1998) 227. [17] D. de Souza, S.A.S. Machado, Qu´ım. Nova 26 (2003) 644. [18] M.B. Moressi, H. Fern´andez, J. Electroanal. Chem. 369 (1994) 153. [19] M. Lovric, S. Komorsky-Lovric, J. Electroanal. Chem. 248 (1988) 239. [20] R.M. Elofson, R.L. Edsberg, Can. J. Chem. 35 (1957) 646. [21] J.M. Zen, S.H. Jeng, H.J. Chen, Anal. Chem. 68 (1996) 498. [22] J.H. Christie, J.A. Turner, R.A. Osteryoung, Anal. Chem. 49 (1977) 1899. [23] L.A. Currie, Anal. Chim. Acta 391 (1999) 103. [24] M. Ib´anez, M. Pico, J. Manes, J. Chromatogr. A 727 (1996) 245. [25] N. Hesketh, M.N. Jones, E. Tipping, Anal. Chim. Acta 327 (1996) 191.
Electrochemical detection of the herbicide paraquat in natural water and citric fruit juices using microelectrodes D. De Souza ∗ , S.A.S. Machado Instituto de Qu´ımica de S˜ao Carlos, Departamento de Fisico Quimica, Universidade de S˜ao Paulo, C.P. 780, 13560-970 S˜ao Carlos, Sa Paulo, Brazil Received 29 November 2004; received in revised form 9 May 2005; accepted 10 May 2005 Available online 13 June 2005
Abstract A novel electroanalytical procedure for detecting the paraquat herbicide in natural water and citric fruit juice samples using gold microelectrodes and square wave voltammetry at high frequencies is proposed. The results obtained showed two reversible peaks for the reduction of paraquat, the first peak associated with the reduction of the paraquat molecule in solution, with subsequent adsorption of the intermediate on the electrode surface. This adsorbed species was shown to undergo electroreduction in a reaction associated to the second voltammetric peak. The variation in pH and square wave parameters indicated that the best conditions under which paraquat could be reduced were a pH of 5.0, a frequency of 1000 s−1 , a scan increment of 2 mV and a square wave amplitude of 50 mV. Under these conditions, the variation of the concentrations of paraquat from 1.00 × 10−6 to 1.66 × 10−4 mol L−1 presented, for peak 1, detection and quantification limits of 4.51 and 15.05 g L−1 respectively in pure electrolyte with a recovery factor of 99.50%. The proposed analytical procedure was also applied to natural water samples giving recovery factors of 95.00, 89.50 and 92.50% in three water samples collected from an urban stream. The recovery factor was observed to depend on the content of organic matter which was determined by the biochemical and chemical oxygen demand. In lemon and orange juice samples that were spiked with 5.70 × 10−5 mol L−1 of paraquat, the recovery factors obtained were 94.30 and 92.70% respectively. © 2005 Elsevier B.V. All rights reserved. Keywords: Square wave voltammetry; Paraquat; Microelectrodes
1. Introduction The continuous increase in agricultural production promotes an equivalent increase in the level of pesticide residues in waters, soils and foodstuff. In this light, the development of analytical methods to monitor different types of pesticides and pesticides residues in several samples of biological or environmental interest acquires a fundamental importance. It is also known that contamination by pesticides is not restricted only to soil or water, but also to the consumption of contaminated foods [1,2] or water [3]. Pesticide residues have been frequently found in blood [4], urine [5] or even in maternal milk [6] and such matrices must be considered in any ∗
Corresponding author. Tel.: +55 162739924; fax: +55 162739952. E-mail address: [email protected] (D. De Souza).
0003-2670/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2005.05.020
analytical procedure. As a consequence, several precise and sensitive analytical procedures are now under development worldwide for their application in either “in situ” or analytical laboratories for the quantification of pesticides and their residues in a large number of different matrices. Paraquat, (1,1 -dimethyl-4,4 -bipyridilium dichloride), also known as methyl-viologen (MV), is a bipyridilium pesticide of toxicological class I and is extremely hazardous for human health. However, it is one of the most largely used pesticides in over 130 countries worldwide. The paraquat include clearing, pasture renovation, inter-row weed control in vegetable crops and weed control in plantation crops. A major problem caused by the abusive and uncontrolled use of paraquat is related to its high persistence in the environment where, though only slightly absorbed by soils, it is
86
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
a potential contaminant of waters due to its high solubility (about 620 g L−1 at 25 ◦ C [7]). Literature reports show that paraquat has been analyzed using a wide range of different analytical techniques, such as spectrophotometry [8], flux analysis [9], liquid chromatography-(electrospray ionization) mass spectrometry [10], polarography [11] and immuno-assay [12]. The electrochemical detection of paraquat has been performed at different electrode surfaces, including solid electrodes [13] and Nafion® modified electrodes [14]. Walcarius and Lamberts [11] developed a fast and sensitive method to determine paraquat in aqueous media using the dropping mercury electrode and square wave voltammetry. The effects of the voltammetric parameters were evaluated, their optimization yielding a detection limit of 1.50 × 10−8 mol L−1 without any pre-concentration step. Paraquat being a cation, the detection limit can be lowered by using an electrode surface covered with a cationic ion-exchange resin. Based on this, Lu and Sun [14] used a Nafion® modified glass carbon electrode for the analytical determination of paraquat using accumulation prestep followed by detection by cathodic redissolution in a pesticide free electrolyte. Employing differential pulse voltammetry as the analytical technique, a detection limit of 2.00 × 10−9 mol L−1 was reached. A similar cationic characteristic of paraquat was used by Derouane et al. [15] to analyze the electrochemical behavior of paraquat at carbon paste electrodes modified by zeolites. The electrochemical responses were explored in several supporting electrolytes using cyclic voltammetry, chronoamperometry, square wave voltammetry and chronocoulometry. It was observed that the electrochemical responses were strongly affected by the initial conditions and by the thickness of the carbon paste used in preparing the electrode. Monk et al. [13] studied the electrochemical behavior of paraquat in aqueous media using cyclic voltammetry. Two voltammetric peaks were detected during the negative sweep, the first at around −0.70 V and the second at approximately −1.20 V versus SCE. The first peak was attributed to the radical cation formation (MV2+ +e− MV+ ) and the second to the neutral species formation (MV+ + e− MV0 ) which, according to the authors, was followed by a chemical dimerization step. Both electrochemical steps were considered as totally reversible processes. This same conclusion was reached by Rueda et al. [16] while using the electrochemical impedance spectroscopy. Finally, de Souza et al. [17] applied microelectrodes and square wave voltammetry in the electroanalytical determination of paraquat in pure electrolyte, reaching a determination limit as low as 7.10 ppb. The observed result was proof the possibility of applying electroanalytical methods in more complex samples. The aim of the present paper is study the applicability of microelectrodes and square wave voltammetry (SWV) in the electroanalytical determination of paraquat in natural waters and in citric fruit juices, without any pre-purification step.
2. Experimental 2.1. Equipment and reagents Electrochemical measurements were carried out using a Voltalab potenciostat/galvanostat (model PGZ402, Radiometer Analytical Inc.) controlled by Voltamaster 4 software (Radiometer Analytical Inc.). A Ag/AgCl (3.00 mol L−1 KCl) system was employed as the reference/auxiliary electrode and the working microelectrodes were constructed by sealing a 25 m diameter gold micro-wire in Epoxy resin as detailed below. A Methron model 682 pH-meter with the glassAg/AgCl/KCl (3.00 mol L−1 ) combined electrode was used for adjusting pH values. All solutions were prepared with water purified in a MilliQ system from Millipore Corporation. Prior to all measurements, solutions were deaerated by passing N2 gas (SS White Martins) for approximately 15 min. Stock solutions of paraquat (99.80%, Merck) were prepared by dissolving the pesticide, without any additional purification step, in water. All the others reagents used were of analytical grade. 2.2. Microelectrodes construction and characterization The gold microelectrodes were constructed by embedding a 25 m diameter gold wire (Goodfellow) in a Pyrex® glass tube of 0.50 mm internal diameter with Epoxy resin. The tips of the gold microelectrodes were polished with emery paper until a metal micro disc was exposed at the surface. After this procedure the microelectrodes were cleaned with purified water prior to use. The voltammetric characterization consisted of analyzing the electrochemical response of a gold microelectrode in an electrolyte containing 1.00 × 10−3 mol L−1 of potassium hexacyanoferrate (III) in 0.10 mol L−1 of KCl with a pH of 3.00. The voltammograms obtained exhibited the well-known sigmoid profile which is characteristic of an electrochemical process controlled by spherical diffusion mass transport, as expected when using microelectrodes [18]. 2.3. Working procedure The initial working procedure consisted of measuring the electrochemical response of the gold microelectrode in a 0.10 mol L−1 of Na2 SO4 electrolyte at a fixed concentration of pesticide. Several supporting electrolytes were tested (Britton–Robinson buffer, Na2 SO4 , phosphate, perchlorate), the best electrochemical response measured in terms of the highest analytical signal and improved reproducibility was obtained in Na2 SO4 . This means that 0.10 mol L−1 of Na2 SO4 was used throughout the experimental program as the supporting electrolyte. The electrochemical cell was then placed in a Faraday cage in order to minimize the contribution of background noise to the analytical signal.
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
The square wave parameters: frequency of application of the pulse potential (f ), pulse amplitude (a) and scan increment (Es ) were optimized since these parameters exert a strong influence on the sensitivity of voltammetric analysis [19]. After optimization of the voltammetric conditions, calibration curves were obtained by adding aliquots of the pesticide to a pure laboratory electrolyte, the supporting electrolyte response denoted as blank. After consecutive adding aliquots of the pesticide, each of whose concentration was properly adjusted, the peak currents due to the reduction of the pesticide were plotted against its concentration in solution. The recovery experiments were then carried out in a similar way by adding a known amount of paraquat to the supporting electrolyte followed by standard additions. The suitability of the electroanalytical procedure in determining paraquat in natural matrices was tested by spiking water samples collected from three distinct points of the Monjolinho stream which crosses the district of S˜ao Carlos in the State of S˜ao Paulo, Brazil. The first sampling point (point 1), located near the stream head and before the city limits is known to be of a very low degree of pollution. The second sampling point (point 2), along the stream length is located around the urban area where the degree of pollution a high while the third sampling point (point 3), is located after crossing the city thus containing a high degree of organic matter including domestic and industrial pollution received from point 2. After contamination, the same electroanalytical procedure described above for the pure electrolyte was applied to the natural water samples. Orange and lemon juices were extracted from their respective fruits and used without any pre-treatment or separation step other than adjusting the solution pH to 5.0 with NaOH solution. The natural electrolytes were artificially contaminated with a known amount of paraquat and recovery measurements performed. All experiments were performed in triplicate.
3. Results and discussion
87
Fig. 1. Steady-state cyclic voltammograms for 4.22 × 10−3 mol L−1 of paraquat in 0.10 mol L−1 of Na2 SO4 , pH 5.0, on the gold microelectrode using different scan rates (as indicated).
This electrode process was followed by a second one, represented by peak 2 at −0.93 V and is associated to a quasireversible adsorption controlled process hence generating the neutral molecule [20]: MV+ + e− MV0
(2)
and the subsequent formation of a dimmer: MV0 + MV2+ MV2 2+
(3)
The SWV experiments yielded the same reduction and oxidation processes described in Eqs. (1)–(3). Here peak 1 was observed at −0.64 V and peak 2 at −0.94 V versus Ag/AgCl 3.00 mol L−1 , which is in close agreement with values obtained in cyclic voltammetric experiments. The main difference observed between SWV and the cyclic voltammetric experiments is that, like peak 1, the electrochemical process related to peak 2 presents a totally reversible behavior indicated by the presence of direct and reverse currents with similar values of peak current. This is probably associated to the very fast rate of pulse applications which inhibits the chemical dimer formation step. This would make the second electron transfer also reversible thus in contrast to the cyclic voltammetric experiments.
3.1. Electrochemical behavior 3.2. Optimization of experimental parameters The cyclic voltammetric responses for a gold microelectrode in 0.10 mol L−1 of Na2 SO4 + 4.22 × 10−3 mol L−1 of paraquat at different scan rates are presented in Fig. 1. In the figure, waves 1 and 5, which occur at approximately −0.70 V, represent the first process observed during the potential scan from −0.30 to −1.20 V. As the scan rate increases, these waves turn into peaks as is expected for a mass transport controlled electrochemical process occurring at microelectrodes. Waves 1 and 5 can thus be associated to the redox couple [20]: MV+2 + e− MV+
(1)
The decrease in proton concentration as pH increases from 2.0 to 9.5 is shown to exert a pronounced effect on the SWV response of paraquat. Up to a pH of 5.5, the peak currents show a five-fold increase for peak 1 and a two-fold for peak 2. The peak potentials do not however seem to be affected by the concentration of H+ , suggesting the absence of any protonation step in the reduction mechanism as indicated by Eqs. (1)–(3). A variation in the frequency of potential pulse application usually exerts a marked effect on the SWV response, thus providing a criterion for diagnosis that can be used to
88
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
Fig. 2. Square wave voltammograms for 5.00 × 10−5 mol L−1 of paraquat in 0.10 mol L−1 Na2 SO4 at a gold microelectrode with pH = 5.0 and a = 50 mV, Es = 2 mV at different frequencies. Insert: relationship between the frequency and peak current obtained for peaks 1 and 2.
Fig. 3. Calibration curves for paraquat in pure water electrolyte (pH 5.0) using f = 1000 s−1 , a = 50 mV, Es = 2 mV, considering the current responses of peaks 1 and 2.
Ep −2, 3RT = log f αnF
indicate any adsorption or reaction in solution, reversibility of the electrochemical process, etc., [19,22]. Fig. 2 presents the effect of frequency on the SWV response for the reduction of paraquat at a gold microelectrode in 0.10 mol L−1 of Na2 SO4 + 4.20 × 10−4 mol L-1 of paraquat with a = 50 mV and Es = 2 mV. As can be observed, an increase in the pulse frequency is accompanied by an increase in current for both peaks 1 and 2. Moreover, the peak potentials of peak 2 show a shift towards more negative values while those of peak 1 remains practically constant, a behavior which indicates that peak 1 refers to a reaction in solution while peak 2 involves adsorption as the rate determining step. This is also confirmed by the dependence of the peak current on the square root of the frequency for peak 1 and a direct dependence on the frequency for peak 2 as demonstrated in the insert of the Fig. 2 [19]. The advantages of using microelectrodes with SWV for the determination of paraquat must be highlighted. Previous studies on the electroanalytical determination of paraquat: Zen and co-workers [21] with a Nafion® coated solid electrode or Walcarius and Lamberts [15] with HMDE, were restricted to a frequency range below 200 s−1 due to the severe signal distortion observed above this limit value. However, in the present study, the use of a gold microelectrode is shown to permit the application of frequencies as high as 1000 s−1 without any loss of quality in the electrochemical signal. This is probably associated with the fast diffusion of species to or from the electrode surface. The high frequency used means an increase in sensitvity is obtained when the proposed methodology is employed. The influence of frequency on the peak potential permits the determination of the number of electrons transferred during the electrode process when the following equations, developed respectively for totally reversible reactions in solution and with adsorption, [19] are used:
DL =
3Sb b
(6)
QL =
(7)
Ep −2, 3RT = log f 2αnF
10Sb b
where Sb is the standard deviation of current measured for the blank solution at the same potentials as those of
(4)
(5)
where R is the gas constant, T the temperature, α the transfer coefficient, n the number of electrons, and F is the Faraday constant. From the equation, the slope of the straight line obtained for the relationships Ep versus log f1/2 for peak 1 and Ep versus log f for peak 2 result in the value of n = 1 (considering α = 0.50), thus confirming the reaction pathway presented in Eqs. (1)–(3). The other two parameters employed in SWV, the amplitude (a) and the potential step (Es ), both exerts a much less effect on the peak current or potential. Following this, the respective values for these parameters were taken as 50 mV for a, as suggested by SWV theory [22], and 2 mV for Es . 3.3. Analytical methodology for paraquat in pure water Using the obtained optimized parameters mentioned above, a calibration curve was obtained for paraquat in an electrolyte prepared with pure water. For this, aliquots from the stock solution were consecutively added to the electrochemical cell. The SWV responses were recorded for a range of concentration between 1.00 × 10−6 and 1.66 × 10−4 mol L−1 . The SWV responses together with the linear relationships obtained between peak currents and concentrations (as the insert) for both peak 1 and 2 are presented in Fig. 3. From Fig. 3, the detection (DL) and quantification limits (QL) were obtained for the experimental conditions employed using the following equations from IUPAC [23]:
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
the peak currents (in this case for an average of eight values of current) and b is the slope of the analytical curve. The observed values of DL and QL for peak 1 were respectively 1.76 × 10−8 mol L−1 (4.51 g L−1 ) and 5.86 × 10−8 mol L−1 (15.05 g L−1 ). The same calculations for peak 2 yielded 3.82 × 10−8 mol L−1 (9.82 g L−1 ) and 1.27 × 10−7 mol L−1 (32.63 g L−1 ) mol L−1 , respectively. Thus, peak 1 showed a greater sensibility for the determination of paraquat than peak 2. The reproducibility of the proposed methodology was determined from five different measurements in the same solution containing 8.00 × 10−5 mol L−1 of paraquat and 0.10 mol L−1 of Na2 SO4 obtaining coefficients of variation of 1.70 and 1.20% for peaks 1 and 2, respectively. The repeatability of this method was also determined from measurements performed for 10 different solutions of the same composition as stated above, obtaining coefficients of variation of 1.60 and 1.20% for peaks 1 and 2, respectively. 3.4. Application of this electroanalytical methodology to natural waters The analytical procedure presented above was applied to natural water samples collected from different points of the Monjolinho stream as described in the Section 2. The sampling points were selected based on different levels of organic matter including industrial and domestic pollution. The water samples were used as received in preparing the supporting electrolyte (adding 0.10 mol L−1 of Na2 SO4 ) and the analytical curves were again obtained by SWV experiments. In this way, the influences of sample contamination were evaluated. The analytical curves obtained under these new experimental conditions are presented in Fig. 4. The results of linear regression from the straight lines obtained presented in Fig. 4 are shown in Table 1. The results obtained indicate that peak 1 is practically unaffected by the presence of contaminants in the natural samples. The analytical sensitivities, defined by the slope of the analytical curves, are constant when compared to those obtained for a pure water electrolyte independent of the origin of the sample. This can be associated with the fact that the electrochemical process for peak 1 occurs with the species
89
Fig. 4. Calibration curves for paraquat in electrolytes prepared with different natural water samples and spiked with paraquat under the experimental conditions presented in Fig. 3.
in solution, thus being independent of adsorption at the electrode surface. On the other hand, peak values of current for peak 2 displayed some influence due to organic matter present in the water samples. Only the analytical curve obtained with water collected from point 1 (free from urban pollution) was comparable to that obtained for purified water. This behavior is expected since peak 2 reflects an electrochemical process involving adsorption of reagents and/or products which can be inhibited by the presence of other organic molecules. In order to evaluate the amount of organic molecules in the stream water samples, analyses of biochemical oxygen demand (BOD) and chemical oxygen demand (COD) were performed for the natural water samples. The results are also presented in Table 1 where it can be observed that the values of s (the sensitivity of the analytical method) decreases with increasing values of BOD and COD. This reinforces the hypothesis that the presence of organic matter interferes in the methodology developed for peak 2. Recovery experiments were also performed in order to evaluate the interference of organic and inorganic components in natural water matrices. Samples were spiked with 2.00 × 10−5 mol L−1 of paraquat and the recovery curves obtained by the standard addition method. The presence of
Table 1 Results of linear regression obtained from Fig. 4 for the square wave voltammograms at the gold microelectrode in natural waters samples, recovered concentration obtained by samples spiked with 2.00 × 10−5 mol L−1 of paraquat and determination of BOD and COD for dissolved O2 Parameters
r Sb × 10−11 (A) s × 10−3 (A/mol L−1 ) Recovery (%) R.S.D. (%) BOD (mg L-1 ) COD (mg L-1 )
Point 1
Point 2
Point 3
Peak 1
Peak 2
Peak 1
Peak 2
Peak 1
Peak 2
0.9994 7.32 5.78 95.00 1.50
0.9966 36.90 8.28 95.00 1.75
0.9985 30.00 5.27 89.50 1.60
0.9967 43.60 4.08 91.00 1.80
0.9993 15.80 5.55 92.50 1.60
0.9982 41.02 5.50 94.50 1.80
6.00 19.00
12.00 33.00
9.00 29.00
90
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91 Table 2 Results obtained from the linear regression curves (Fig. 5) for the determination of paraquat in lemon and orange juices at a gold microelectrode Parameters
r Sb × 10−11 (A) s × 10−3 (A/mol L−1 ) Repeatability (%) Reproducibility (%) Recovery (%)
Fig. 5. Square wave voltammograms for paraquat in electrolyte prepared with lemon juice at a gold microelectrode with f = 1000 s−1 , a = 50 mV, Es = 2 mV and different concentrations of paraquat.
this pesticide was not detected in natural waters used as collected. The results obtained for peak 1 and peak 2 (Table 1) presented the same inverse dependence on the amount of organic matter presented in different samples. Following Ib´anez [24] and Hesketh [25], paraquat strongly interacts with humic substances as well as with surfactants present in natural water samples, thus promoting a negative effect on the recovery values. 3.5. Applications in orange and lemon juices “in natura” Paraquat is one of the most widely used herbicides in citric fruit cultures. Due to its toxicity and persistence in the environment, there is a very high probability of its contaminating crops, thus presenting a risk to human health. Because of this, the determination of paraquat in “in natura” citric fruits juice serves as a means of minimizing the hazardous effect of such pesticide. The methodology developed in the preceding sections was hence applied to “in natura” orange and lemon juices. The juices were extracted from their respective fruits without any pre-separation or pre-concentration. The juices had their pH adjusted to 5.00 with an appropriate volume of a 3.00 mol L−1 of NaOH solution and then added to the electrochemical cell. The SWV response curves were later obtained in the range concentration of 1.00 × 10−5 to 25.00 × 10−5 mol L−1 of paraquat. The observed SWV responses are shown in Fig. 5 for lemon juice with f = 1000 s−1 , a = 50 and 2 mV. Similar results obtained for orange juice are also shown. As it can be observed, peak 1 is free from interference. However, peak 2 overlays a background current, maybe related to other matrix components which, in the case of citric fruits, are basically ascorbic acid (Vitamin C) and citric acid. From the relationship between peak current and the concentration of paraquat, calibration curves were obtained, which, for lemon juice, are presented in Fig. 5 (insert) for
Lemon
Orange
Peak 1
Peak 2
Peak 1
Peak 2
0.9993 4.30 1.03 3.40 1.00 94.30
0.9976 126.00 1.74 2.70 1.01 84.40
0.9983 5.54 1.14 3.85 1.15 92.70
0.9972 215.00 2.12 2.90 2.05 77.90
peaks 1 and 2. In order to obtain the correct values for peak currents, the base line was subtracted from each curve. Reproducibility, for n = 5, repeatability, for n = 10 and recovery experiments (by the standard addition method, as described above) were also performed on citric fruit juice electrolytes using a concentration of paraquat of 5.70 × 10−5 mol L−1 . The obtained results are all included in Table 2. A comparison of the analytical sensibility of the methodology (defined by the slopes of the calibration curves) obtained for pure water and citric fruit juice shows that the analytical sensibility of citric fruits is 20-fold that of pure water. This can be attributed to the adsorption of the matrix components at the electrode surface or to interactions between the pesticide molecule and components of the juice. In order to evaluate any possible interference of citric and ascorbic acids in the reduction of paraquat at a gold microelectrode, electrolytes containing a known concentration of each of these acids in pure water were recorded and plotted together with that of paraquat. The results obtained demonstrated that the second reduction peak of paraquat is more likely distorted due to the presence of those contaminants. The fact that their peaks occurs at approximately −0.85 V indicates that an interference only in the second peak.
4. Conclusions A gold microelectrode in a square wave voltammetric procedure was successfully applied in quantifying paraquat in natural matrices. In the present work, the the primary goal, on using microelectrodes, was to verify the possibility of obtaining responses in SWV frequencies as high as 1.000 s−1 which are much more convenient since at this level, the analytical signal is proportional to the frequency. In fact, no such electroanalytical analysis has yet been reported. The novelty of the proposed methodology is rooted in the use of microelectrodes in citric fruits juice without any pre-treatment or extraction procedure. This, in the authors’ view, is the first time such a measurement which serves as an important new tool for the direct determination of pesticides in “in vivo” fruit is being reported. To the best of the authors’
D. De Souza, S.A.S. Machado / Analytica Chimica Acta 546 (2005) 85–91
knowledge, no other methodology that permits such a direct measurement in growing plants or maturing fruits is available. The first step for such a procedure is however reported here. Of course, had microelectrodes not been capable of detecting paraquat without any change in the fruit juices, the direct determination of paraquat would not have been possible. However, as the present work shows, microelectrodes do actually respond very well in pure fruits juices, resulting in detection and quantification limits lower than permitted by Brazilian environmental council. Initially, the electrochemical behavior of paraquat was evaluated using the voltammetric results. The SWV response showed two reduction waves associated to the partial reduction of paraquat molecule followed by the formation of dimer species in solution. These processes were determined as being reversible, the first being controlled by diffusion and the second by electron transfer to an adsorbed species. This is the most suitable situation for SWV analysis since the resulting current is the sum of forward and reverse contributions, the latter only being present in reversible processes, a situation which is far from common in pesticide analysis. Analysis on a purified laboratory electrolyte allows the determination of a very low detection limit of 4.51 g L−1 associated with a high level of repeatability and reproducibility. Because of this, the electroanalytical methodology was applied in determining paraquat in spiked stream water samples which presented different degrees of pollution. The influence of organic matter (pollution or natural) content on the sensitivity of the analytical procedure was clearly established by determining BOD and COD of the samples. Finally, the same analytical procedure was applied in determining paraquat in citric fruit juices where it was demonstrated that either ascorbic acid or Vitamin C exerts detectable effects on the recovery experiments of spiked samples that were obtained without any pre-treatment. In this way, the use of microelectrodes in conjunction with SWV in determining paraquat seems to present a very suitable analytical methodology which permits its application in a variety of different natural matrices.
91
Acknowledgements The authors acknowledge the financial support from FAPESP (proc. 03/12926-3 and 04/00839-1) and CNPq.
References [1] A. Saba, F. Messina, Food Qual. Prefer. 14 (2003) 637. [2] H.G.J. Mol, R.C.J. Van Dam, O.M. Steijer, Chromatogr. A 1015 (2003) 119. [3] E. Ballesteros, M.J. Parrado, J. Chromatogr. A 1029 (2004) 267. [4] U. Heudorf, J. Angerer, H. Drexler, Int. J. Hyg. Environ. Health 206 (2003) 485. [5] J.M. Pozzebon, W. Vilegas, I.C.F. Jardim, J. Chromatogr. A 987 (2003) 375. [6] M. Jimenez, A. Jerez, J.M. Diaz, M.F. Olea, R. Bayes, C. Campoy, Pediat. Res. 54 (2003) 613. [7] E. Halfon, S. Galassi, R. Br¨uggemann, A. Provini, Chemosphere 33 (1996) 1543. [8] M.K. Ra´ı, J.V. Das, V.K. Gupta, Talanta 45 (1997) 343. [9] B. Saad, M. Ariffin, M.I. Saleh, Talanta 47 (1998) 1231. [10] V.Y. Taguchi, S.W.D. Jenkins, P.W. Crozier, D.T. Wang, J. Am. Soc., Mass Spectrom. 9 (1998) 830. [11] A. Walcarius, L. Lamberts, J. Electroanal. Chem. 406 (1996) 59. [12] M.J. Nuwer, J.J. O’Dea, J. Osteryoung, Anal. Chim. Acta 362 (1998) 03. [13] P.M.S. Monk, C. Turner, S.P. Akhtar, Electrochim. Acta 44 (1999) 4817. [14] T.H. Lu, I.W. Sun, Talanta 53 (2000) 443. [15] A. Walcarius, L. Lamberts, G. Derouane, Electrochim. Acta 38 (1993) 2257. [16] M. Rueda, R.G. Compton, J.A. Alden, F. Prieto, J. Electroanal. Chem. 443 (1998) 227. [17] D. de Souza, S.A.S. Machado, Qu´ım. Nova 26 (2003) 644. [18] M.B. Moressi, H. Fern´andez, J. Electroanal. Chem. 369 (1994) 153. [19] M. Lovric, S. Komorsky-Lovric, J. Electroanal. Chem. 248 (1988) 239. [20] R.M. Elofson, R.L. Edsberg, Can. J. Chem. 35 (1957) 646. [21] J.M. Zen, S.H. Jeng, H.J. Chen, Anal. Chem. 68 (1996) 498. [22] J.H. Christie, J.A. Turner, R.A. Osteryoung, Anal. Chem. 49 (1977) 1899. [23] L.A. Currie, Anal. Chim. Acta 391 (1999) 103. [24] M. Ib´anez, M. Pico, J. Manes, J. Chromatogr. A 727 (1996) 245. [25] N. Hesketh, M.N. Jones, E. Tipping, Anal. Chim. Acta 327 (1996) 191.