1 Introduction

Viticulture has great importance in Turkey’s agricultural structure. Although vineyards are spread all over the country, production is mainly concentrated in Aegean, Mediterranean, and Central Anatolian regions. Turkey is the sixth country in the world from the view point of the area planted grapevines. However, only a small portion of the harvest is used for wine production [1].

The grapevine is subject to attack by numerous plant and animal parasites. The most frequent diseases caused by fungi are downy mildew, powdery mildew, and gray mold in grapes [2]. Pesticides are used on agricultural commodities such as grapes and wine grapes and as a result, a part of pesticides left on the grapes at harvest, particularly late-season fungicides, can be carried into the wine [3]. For pesticide residues in wine no uniform maximum residue limits (MRLs) have been established yet. However, there is a worldwide trend toward setting specific, lower MRLs for pesticides in wine, which would range from 0.01 to 2 mg kg−1 for different pesticides [4].

Chlorpyrifos (O,O-diethyl O-3,5,6-trichloro-2-pyridyl phosphorothioate) is a member of organophosphorus class insecticides and defined as an endocrine disruptor [5, 6]. It is widely used for the control of Lobesia Botrana in grapes [7]. Like the other organophosphorothioate pesticides, acute toxic effects of CP exposure are primarily due to the inhibition of acetylcholinestease [8].The potential harm effect of chlorpyrifos (CP) to ecosystems is considerable and numerous methods have been developed for its determination [9]. Analysis of Chlorpyrifos in different food matrixes is conventionally carried out by gas chromatography (GC) by using NPD [10], ECD [3], and MS detectors [11–13]. Liquid chromatography (HPLC) [14], immunoassay-based techniques, such as enzyme linked immunosorbent assay (ELISA) [15], and liquid chromatography determinations combined with ELISA have also been reported [16, 17]. Major techniques for extraction and concentration of pesticides prior to the chromatographic separation in wine are liquid–liquid extraction (LLE) and solid-phase extraction (SPE), solid-phase micro extraction (SPME). In developing countries, routinely adopting the chromatographic analysis method is, however, financially limited for these kinds of analyses. Thus, development of reliable and fit for purpose methods of analysis for mixtures of pesticides with the use of simple and relatively inexpensive instrumentation is an appropriate objective. Hence, electroanalytical techniques provide an inexpensive alternative for pesticide analysis [18–24].

In fact, electrochemical methods are not frequently used for analysis of pesticides in complex matrices. Pesticide residue analysis by voltammetric methods is carried out especially in commercial formulations [25–27], water [28–33], and soil samples [34–38]. To date, a few studies were reported describing voltammetric determination of pesticide residues in food sample including liquid–liquid extraction and clean-up stages prior to the analysis [29, 39].

Chlorpyrifos is known to be electroactive and its differential pulse polarographic (DPP) behavior was investigated [40]. Adsorptive property of the molecule was utilized in stripping voltammetric determination of CP in commercial formulations and treated waste water samples [41].

This paper deals with the development of an adsorptive catalytic stripping voltammetric method for the sensitive detection of the CP at an HMDE. The application of the method for the determination of trace amounts of CP residue in red wine samples was also elucidated. This is of particular importance since no data are available in the literature for the determination of CP in wine samples by a voltammetric method. Determination of trace amount of pesticide residues in wine is difficult task since such complex matrices hinder an accurate evaluation and therefore, a pre-separation step is required for handling the samples. Thus, the effectiveness of LLE and SPE steps on recovery percentages was also examined in this paper.

2 Experimental

2.1 Instrumentation

Voltammetric studies were carried out with a Metrohm 694 VA Processor and Stand. A three electrode system was used including Ag/AgCl (saturated KCl) reference and platinum auxiliary electrodes. The working electrode was a hanging mercury drop electrode (HMDE). All experiments were carried at pulse amplitude of 50 mV and a scan rate of 100 mV s−1.

Solid-phase extractions were applied by using cartridges (3 mL) containing 500 mg of Florosil, (Supelco, Oakville, Canada). A mechanical vacuum pump under a 20 in. Hg vacuum (Edwards E2M2, Crawley, UK) was used in SPE experiments. High speed homogenizer (Waring Commercial Laboratory) and nitrogen evaporator (Turbovax, USA) were used during the extraction. The samples were stirred by using an IKA basic vibratory stirrer (IKA Labortechnik, HS501). pH measurements were made by using Mettler Toledo Seven Multi pH-Meter (Switzerland).

2.2 Reagents

All reagents and chemicals used were of analytical grade. Universal Britton Robinson buffer solutions (BR) were prepared by mixing equal molar (0.04 M) of phosphoric, boric, and acetic acid solutions and by drop wise addition of 0.2 M NaOH to provide a wide range (2–10) of pH. Various electrolyte ion solutions of 1.0 M were prepared by dissolving the required amount of salts in a definite volume of double-distilled water. Procymidone, Penconazole, Bromopropylate, Lambda-Cyhalotrin, Iprodione, and Chlorpyrifos pesticides were supplied by Dr. Ehrenstrofer Laboratories and stock solution of which (500 μg mL−1) was prepared by dissolving an appropriate amount of the pure solid reagents in absolute ethanol.

Analytical grade cyclohexane, ethyl acetate, acetonitrile, chloroform, dichloromethane, diethyl ether, petroleum ether, heptane (Lab Scan) ethanol, hexane, toluene, isooctane, and acetone (J.T.Beaker) reagents were used for the sample preparation. Analytical grade HClO4, H3PO4, CH3COOH, H3BO3, H2SO4, HCl, HNO3, and Titrisol buffer solutions were supplied by Merck.

The wine sample used was a red wine obtained commercially from the local market. This sample contains 12.8% (v/v) ethyl alcohol. The measurements were made at laboratory temperature and all the solutions were allowed to reach at this temperature prior to measurement.

2.3 Procedures

2.3.1 Voltammetric measurement procedure

An aliquot of buffer solution (20 mL) containing 5–25% ethanol was placed in the voltammetric cell. The solution was purged with nitrogen for 5 min and a new drop was initiated. The analyte was deposited via adsorption onto the HMDE surface by applying the potential at −0.50 V for 1 min. Then, the potential was scanned from −0.50 to −1.25 V until otherwise stated. The precision of the method was calculated from sequential eight analyses as RSD%. In the sample analysis, the standard addition method was used in order to eliminate the matrix effects. The wine samples were spiked with small volumes of CP standard solution to be 100 ng mL−1 in the cell and further additions were made. The procedure was applied to five parallel red wine samples.

2.3.2 Preparation of wine sample for analysis

20 mL of wine sample was placed in an extraction bottle, and mixed with 10 mL of n-heptane by stirring for 30 min at a rate of 350 rpm. In clean-up step, 5 mL of the organic phase taken from the above step was passed through the Florosil column and the analyte retained in the column was eluted by 3 mL of Petroleum Ether. The column was dried under vacuum and elute was dried under a nitrogen flow. The residue was dissolved in 0.5 mL EtOH by using vortex stirrer and diluted with 4.5 mL Titrisol buffer.

3 Results and discussion

3.1 Optimization of experimental and instrumental parameters

Early studies for voltammetric determination of CP were initiated with a dropping mercury electrode (DME) and the DP polarogram of CP standard solution was recorded at 1 s of drop time. In DP, polarogram of CP in pH 2.0 BR buffer solution revealed that the reduction peak of CP cannot be characterized definitely as it is located very close to hydrogen-evolution peak. Therefore, hanging mercury drop electrode (HMDE) was used and a well-separated reduction peak was obtained at −1.10 V in pH 2.0 BR medium.

To investigate the effect of pH on the peak shape, the analysis was carried out in BR buffer solutions over a pH range of 2.0 to 10.0 each containing 25% ethanol as described earlier [40]. Figure 1 illustrates the dependence of the peak height of the CP on the pH. As the medium was made more alkaline, the peak potential has shifted from −1,110 to −1,206 mV indicating a pH dependent reduction process. Considering the structural effects on the reduction mechanism, even the pyridine moiety is electroinactive in nature, the mesomeric effect of the three Cl atoms in CP molecule increases the localization of the electrons in the pyridine system and results in a reduction of the C=N centre [41]. The reduction reaction was attributed to the reduction of this centre of the pyridine ring via 2e− and 2H+. This explains the decline in the peak current at high pHs. Best results were obtained in acidic media and pH 2.0 BR buffer solutions. More acidic regions were avoided as the reduction peak of CP is located very close to hydrogen-evolution peak.

Fig. 1
figure 1

The effect of pH on the peak current of 0.6 μg mL−1 CP recorded at HMDE in BR buffer solutions containing 25% ethanol. Inset figure represents the dependence of the peak current on the ethanol percentage

Full size image

A small volume of ethanol is usually added to the buffer solution to maintain appropriate solubility of the pesticide molecule. In this study, the effect of the optimum ethanol percentage was found to be 5% (Fig. 1 inset). The decrease in the peak height of the CP in high ethanol percentages could be explained by the less affinity of CP to mercury rather than the solution. Since adsorption of the CP on mercury drop strongly affects the reduction peak characteristics, further studies were focused on the adsorption parameters, namely deposition time and the potential.

The effect of deposition potential on the peak height of 1.0 μg mL−1 CP was investigated at pH 2.0 BR buffer solution containing 5% ethanol. Since a strong adsorption was observed at −0.50 V in the plot of the peak current versus the deposition potential (Fig. 2), this potential was selected for further studies.

Fig. 2
figure 2

Effect of deposition potential on the peak height of 1.0 μg mL−1 CP in BR buffer solutions containing 5% ethanol (Deposition time = 60 s)

Full size image

Deposition of CP on the HMDE should be carefully controlled for avoiding saturation and maintain linearity with increased analyte loading. Figure 3 shows the effect of deposition time on the adsorptive cathodic stripping voltammetric (AdCSV) peak heights of CP deposited on an HMDE. Since the concentration of CP is high (1.0 μg mL−1), a deposition time of 60 s would be adequate. However, a longer deposition period can be used for the more sensitive regions.

Fig. 3
figure 3

Effect of deposition time on the AdCSV peak height of 1.0 μg mL−1 CP in BR buffer solutions containing 5% ethanol (Deposition potential = −0.5 V)

Full size image

The composition of the electrolyte solution is another significant parameter which affects the adsorption as well as the reduction of CP. Replacing the BR buffer solution with 0.01 M HCl, HNO3, H2SO4, HClO4, and H3PO4 solutions to obtain a more sensitive peak, it was observed that the most sensitive one was recorded in 0.01 M HClO4 solution, and further studies were carried out with this electrolyte.

The effect of solution temperature on the CP peaks was also studied over a range of 20 to 40 °C and maximum peak current was obtained at 25 °C (data not shown). It is concluded that the decreasing current observed at higher temperatures indicates the thermal desorption or volatilization of CP molecule.

3.2 Cyclic voltammetric behavior of CP molecule at HMDE

Cyclic voltammetry (CV) is a useful tool for mechanistic studies. As can be seen from Fig. 4, 1.2 μg mL−1 of CP molecule gives a reduction peak at −1,200 mV in pH 2.0 medium at a scan rate of 100 mV s−1. No anodic peak was observed as an irreversible reduction occurred.

Fig. 4
figure 4

The CV voltammograms of 1.2 μg mL−1 CP at pH 2.0 medium recorded at an HMDE for a scan rate of 1,000 mV s−1

Full size image

The reduction peak has shifted to more negative potentials with increasing scan rate (ν). The plot of peak potential versus log (ν) which exhibits a linear relationship with a good correlation for CP corresponds to the fact that the analyte is transferred to the mercury drop by diffusion (Fig. 5).

Fig. 5
figure 5

The plot of potential versus log ν

Full size image

On the other hand, the peak currents for very low concentrations were much larger than expected from Randles–Sevcik equation, indicating that another process is involved beside the diffusion. As pointed out in a former study [30], a relatively high peak current indicates a possible hydrogen-evolution catalytic process over pyridine moiety. Further experiments were designed to cover diagnostic tests depending on the scan rate. Figure 6 shows that the ratio of “peak current ν −½”decreases at higher scan rates and the peak current reaches a stable value at lower scan rates (Fig. 6 inset). Both findings confirm that a catalytic process is involved.

Fig. 6
figure 6

The plot of current/ν ½ versus ν. Inset figure shows the dependence of peak current on ν for lower scan rates

Full size image

Moreover it was observed that, the peak current of the CP is strongly dependent on the drop size of HMDE. The peak current measured experimentally increases 26 times, although the drop size changes 1 to 9 with the surface area changing 0.15 to 0.60 mm2 as specified in the instrumentation manual. This means that as greater the surface area where the CP molecules are adsorbed is greater the catalytic effect.

As a result, the reduction mechanism of CP can be evaluated as a catalytic process. Pyridine moiety of the CP molecule undergoes protonation leading to pyridinium ion and then it is reduced with the regeneration of pyridine as hydrogen gas is liberated (Fig. 7). Finally, the overall reaction is the reduction of the hydrogen ion to hydrogen molecule. However, this occurs at a potential significantly lower than that for the normal reduction of the hydrogen ion and forms an independent peak whose height is controlled by the concentration of the CP. Although catalytic currents are usually of poor reproducibility, they do offer greater sensitivity as the magnitude of the current is two orders larger than that of diffusion currents. This mechanism differs from that proposed earlier [40, 41] and implies the high sensitivity of the method. Lower ethanol percentages utilized in this study would also contribute to the sensitivity as it affects the adsorption characteristics. Medium pH is another effect which indicates the presence of the catalytic process since hydrogen ion is reduced.

Fig. 7
figure 7

The mechanism proposed for CP reduction at an HMDE

Full size image

3.3 Analytical characteristics

A calibration curve of the peak current versus concentration was plotted in pH 2.0 HClO4 solutions containing 5% EtOH with a drop size of 6 of hanging mercury electrode. Figure 8 shows the voltammograms obtained for a range of 20–100 ng mL−1 CP. Similar calibration curve was also plotted for the more sensitive concentration region of 0.2–1.2 ng mL−1. The slope and the intercept of the curve were found to be 66.21 ± 7.28 nA mL ng−1 and 0.69 ± 0.08, respectively. Each point of the two calibration curves is the average mean value of three measurements. LOD is the lowest concentration that can be distinguished from the noise level. The LOD and LOQ values were calculated as 0.14 and 0.45 ng mL−1, respectively, and the RSD value at this level was found to be 7.4% (n = 8).

Fig. 8
figure 8

The DP voltammograms and corresponding calibration graph for CP in higher concentration range at pH 2.0 HClO4 solution containing 5% ethanol (Deposition potential = −0.5 V, Deposition time = 30 s)

Full size image

3.4 Interference studies

Metal ions tend to do interference by forming complexes with the organic analyte or simply being reduced with a potential close to the analyte. Therefore, the effect of several metal ions was investigated in the presence of 3 × 10−7 M CP in pH 2.0 HClO4 solution containing 5% ethanol. Copper, lead, cadmium, and zinc ions were added into the cell to give a tenth, the same or ten fold of the CP concentration. Ten-fold copper, lead, and cadmium ions have led a decrease in the peak current less than 8%. In contrast to those ions above, zinc ion severely hinders the signal since it is reduced at a very close potential to that of CP molecule. Nevertheless, the wine samples require a pre-separation step for removal of organic residues as well as the inorganic constituents in the sample matrix.

To study the selectivity of the method, possible interference of other pesticides, which might be used in grapevine, was studied. Procymidone (PRO), Iprodione (IP), Penconazole (PNC), Bromopropylate (BRP), and Lambda-Chyhalothrin (LB) pesticides were usually detected in Turkish wine samples as reported in a former study [42]. Voltammetric measurements were carried out in Titrisol buffer solution containing 5% ethanol and spiked with CP to be 0.5 μg mL−1 in the cell. Results have revealed that PRO, IP, PNC, and LB pesticides have no electro-active properties under the experimental conditions. On the other hand, BRP standard has given a single peak close to a large reduction signal of hydrogen at −1.2 V. The peak current of CP at 0.5 μg mL−1 concentration level has decreased to the half of the initial one in the presence of equal concentrations of BRP.

In fact, most organic residues tend to be surface active and result in a certain loss in the signal. Yet, pre-separation step was applied in wine analyses where the analytes are separated mostly by their polarities. By applying an appropriate separation technique, this interference can be eliminated. Moreover, this finding indicates that both pesticide residues can be determined simultaneously in wine sample.

3.5 Application of the method for wine samples

The complex matrix of the wine samples requires a pre-separation step for the determination of the pesticides. Pesticide residue analysis in food is usually carried out by means of multi-residue-methods for extraction and clean-up stages followed by a final determination step. In this study, various separation techniques including liquid–liquid extraction (LLE) with different solvents and solid-phase extraction (SPE) with different cartridges were tested in the clean-up step prior to the voltammetric measurement of CP. LLE step was optimized by determining recovery ratios for different solvents, namely n-heptane, petroleum ether, cyclohexane, toluene, diethyl ether, dichloromethane, chloroform, and ethyl acetate. For this purpose, a synthetic wine sample containing 12% EtOH and 0.5 μg mL−1 CP was placed in an extraction bottle and LLE step was applied according to the procedure given in Sect. 2. Recovery percentages were found to decrease accordingly with the polarity index of the extraction solvent and then n-heptane was chosen as it gives the best result. However, this step alone does not fulfill the requirements in eliminating the matrix effects and additional clean-up process is needed. Therefore, as a further step, solid-phase extraction (SPE) cartridges including Florosil sorbents were utilized. Elution was made by Petroleum Ether and elute was dried under nitrogen stream. The residue was dissolved in small amount of ethanol and diluted with pH 2.0 Titrisol buffer. The procedure was applied to five parallel red wine samples. Voltammograms were recorded for the samples prepared in this way (Fig. 9) and by using standard addition method. Thus, the CP content of sample was calculated as 0.060 ± 0.006 μg mL−1 (n = 5). The accuracy of the method was tested upon recovery studies. Red wine samples were spiked with 0.05 μg mL−1 of CP and recovery ratios were found in a range of 84–110%.

Fig. 9
figure 9

The standard addition DP voltammogram of a wine sample and after addition of CP standard to be b 0.1, c 0.2, and d 0.3 μg mL−1

Full size image

4 Conclusions

In this paper, electrochemical behavior of chlorpyrifos pesticide was investigated by differential pulse adsorptive stripping voltammetry on an HMDE. The reduction mechanism of the CP was elucidated as a catalytic process via pyridine moiety of the molecule. Excellent sensitivity of the method inherited from the catalytic process was utilized in the determination of the CP in sub ppb levels. Based on this, a convenient procedure for the determination of the CP residue in wine samples is proposed. The complex matrix effect was eliminated by using both LLE and SPE steps with very satisfactory recovery values. Hence, this electroanalytical technique provides an inexpensive, simple, sensitive, and rapid alternative for electroactive pesticide analysis in wine samples with a sufficient precision and accuracy.