A Simple and Efficient Method Employing Solid-Liquid Extraction with Low-Temperature Partitioning for the Determination/Monitoring of Pesticide Residues in Strawberries by GC/ECD

O uso indiscriminado de agrotóxicos em morangos pode torná-los impróprios para consumo. Neste estudo, o método extração sólido-líquido com partição em baixa temperatura (SLE/LTP) acoplado a cromatografia gasosa com detector por captura de elétrons (GC/ECD) foi otimizado. O método proposto foi validado pela análise de amostras fortificadas do branco da matriz. Os limites de detecção (LOD) variaram entre 4 e 16 μg kg, enquanto que os limites de quantificação (LOQ) variaram entre 13 e 47 μg kg. Boas recuperações (79-111%) e precisão (< 15%) foram obtidos para todos os compostos na matriz. Este método foi aplicado com sucesso na análise de resíduos de agrotóxicos encontrados em morangos coletados de sete fazendas do Vale do Jequitinhonha (MG, Brasil). Os resultados revelaram a presença de azoxistrobina, clorotalonil, difenoconazol e iprodiona nas amostras de morango.


Introduction
Strawberry is a widely appreciated fruit worldwide with great acceptability by consumers because of its attractive organoleptic characteristics, such as its typical intense red color, texture, aroma and taste. 1 However, because strawberry plants are sensitive to pests and diseases, pesticide use in this crop is very intense.Additionally, the use of unauthorized active ingredients in this culture is frequent.For these reasons, the presence of residues above the maximum acceptable limits, aside from residues of unauthorized pesticides, is frequent for this culture in Brazil. 2 Obesity is among the diseases that can be caused by chemicals present in crops. 3In large crops, pesticides are sometimes applied by aircrafts and infiltrate the soil to the waterbeds.Moreover, Meyer et al. 4 noted that many of these substances are excreted in breast milk, constituting a Vol. 25, No. 8, 2014   source of contamination in newborns.An intriguing fact is that many people consume fruits and vegetables to prevent obesity; however, the interference of certain products with hormones can further facilitate the disease.This problem has increased as the selling of pesticides has grown 93% over the last ten years worldwide.In Brazil, this growth was even higher (190%), and between 2001 and 2008, 5 pesticide sales increased from a little over US$ 2 billion to US$ 7 billion, becoming the highest world consumer with 986.5 thousand tons of applied pesticides.In 2010, the national market increased to nearly US$ 7.3 billion, representing 19% of the global pesticide market, and reached US$ 8.5 billion in 2011.
The analysis of pesticide residues in foods is not always an easy task because of the complex matrix in which the analytes are present in very small concentrations (ng g -1 ). 6In addition, several extraction stages involving large volumes of organic solvents are generally necessary, making the employed techniques long and extremely painstaking. 7or these reasons, solid-liquid extraction with low-temperature partitioning (SLE/LTP) seems to be a viable alternative for analyte extraction of complex matrices.[10][11] The main advantages of SLE/LTP include low solvent consumption and the absence of cleanup stages for obtaining relatively pure extracts that can be directly analyzed by gas chromatography (GC).
Historically, GC has been the technique selected for multiresidue pesticide analysis in products of vegetable origin, providing results with detection levels ranging from nanograms to micrograms per kg (ng kg -1 to µg kg -1 ), depending on the detector and extraction technique used. 12,13SLE/LTP in combination with GC/electron capture detection (ECD) analysis has been applied to the determination of pesticide residues in soils and foods. 14,15he present study aims to determine the residues of 11 pesticides (azoxystrobin, bifenthrin, cypermethrin, chlorothalonil, chlorpyrifos, difenoconazole, endosulfan, iprodione, λ-cyhalothrin and permethrin) in strawberries from the High Jequitinhonha Valley in Minas Gerais, Brazil using SLE/LTP and GC/ECD.
For simultaneous determination of pesticides after extraction, the following column oven temperature program was employed: initial temperature of 150 ºC (2 min), heating ramp of 40 ºC min -1 to 210 °C (2 min), followed by a ramp of 10 ºC min -1 to 250 °C (2 min), followed by a ramp of 20 ºC min -1 to 290 °C (maintained for 7 min).The injector temperature was maintained at 280 ºC and the employed carrier gas flow was 1.2 mL min -1 .The injected volume was 1.0 µL and injections were conducted with a flow split of 1:5.The total analysis time was 20.5 min and the runs were managed by Shimadzu GC-solution software.The pesticides were identified by comparing the retention time of the peak present in the extracts of the samples with the retention time of the standard and were quantified by the matrix-matched method.

SLE/LTP preparation of samples
Samples of pesticide-free strawberries were obtained from an organic cultivation system and ground in a mixer.Then, 4.0000 g of pulp measured on an analytical balance (Sartorius BP 2215, Göttingen, Germany) with an accuracy of 1 × 10 -4 g were added to a clear glass with a capacity of 22.0 mL, which was then spiked with 80 µL of a working solution containing pesticides.
Spiked samples of strawberry (0.40 µg g -1 azoxystrobin, cypermethrin, chlorothalonil, chlorpyrifos and endosulfan; 0.80 µg g -1 bifenthrin, iprodione, λ-cyhalothrin and permethrin and 2.0 µg g -1 difenoconazole) were allowed to stand for approximately 3 h after the contained mixture was opened.The spiked samples were then subjected to SLE/LTP analysis for the optimization process.
Two full factorial designs were used to optimize the parameters of the SLE/LTP for the analysis of select pesticides in the strawberry samples.
First, a full factorial design, 2 2 , was employed to review the behavior of concurrent factors: solvent extractor and the ratio of sample:extracting solvent on two levels.The analyses were performed in duplicate (Table 1).
Next, the influence of ionic strength on the extraction efficiency of the analytes in the strawberry samples was assessed by adding an aqueous solution of sodium chloride at a concentration of 2.0% (m/v) to the extractor solvent (acetonitrile), which increased the ionic strength of the medium.The effect of acidification of the medium was also evaluated by acidifying the extractor solvent (acetonitrile) with acetic acid.To assess the behavior of these two factors simultaneously, a second factorial design, 2 2 , was employed.The analyses were performed in triplicate at the central point (Table 2).
The best conditions were assessed for chromatographic responses (areas) obtained in each test at a 95% level of confidence.These data were analyzed using Statistica (version 8) software.
The sample:solvent extractor mixture was stirred for 15 min in an orbital shaker (Tecnal, model TE-141, Piracicaba, Brazil) at 25 °C with a rotation of 200 rpm, centrifuged (Centrifuge Excelsa II, model 206BL, FANEM, São Paulo, Brazil) for 4 min at a rotation of 617.28 g and allowed to cool in a freezer at approximately -20 °C for 9 h, based on the proposed method from Heleno et al.. 16 From the supernatant fraction (acetonitrile), 1.5 mL of solution was removed and placed directly into a vial for chromatographic analysis.

Method validation
To evaluate the selectivity of the method, the method was applied to strawberry samples without pesticides and those spiked with pesticides.Subsequently, the chromatograms from these samples were compared.The linearity of the method was evaluated using the analytical curve obtained by analyzing the extracts of the samples spiked with concentrations of pesticides ranging from 0.05 to 0.60 mg kg -1 (n = 7 points) by a linear square regression analysis.For endosulfan, all of the concentrations followed the same isomeric ratios of the standard.The mathematical relationship between the signal (response) and the concentration of the analytes of interest is expressed by the equation of the analytical curve and its determination coefficient (R 2 ).The performance parameters for the method were evaluated, which included selectivity, precision (repeatability), accuracy/recovery, limit of detection (LOD), and limit of quantification (LOQ) of the analytes.Limits of detection were calculated according to the expression 3.3 σ/S, where σ is the standard deviation of the blank and S is the slope of the calibration curve.Limits of quantification (LOQ) were established by using the expression 10 σ/S. 17 To determine the repeatability, samples were spiked in replicates of six at three different concentrations (0.10, 0.15 and 0.60 mg kg -1 ).The results are expressed as the coefficient of variation.The accuracy was determined from the recovery assays in which known quantities of the analyte had been added to the sample in replicates of six at three different concentrations (0.10, 0.15 and 0.60 mg kg -1 ).The results are expressed by the recovery percentage.

Determination of pesticides in strawberries from the High Jequitinhonha Valley in Minas Gerais, Brazil
Strawberry samples were obtained directly from farms in the High Jequitinhonha Valley at three different harvest periods during the same season in 2012: T1, T2 and T3.
The fruits used for the analysis were selected among those collected for sale.The sample collection method followed the Food and Agriculture Organization of the United Nations. 18The fruits were wrapped in foil and plastic bags, transported in isothermal boxes to the laboratory where they were then ground in a mixer.The matrix extraction of pesticides was performed by SLE/LTP, followed by detection/quantification via GC/EDC.
The fruits originated from seven different farms totaling 21 samples (triplicate), and the results are expressed as triplicate means.

Optimization of pesticide extraction
The effects and significance of the variables in the SLE/LTP system were evaluated using Pareto charts (Figure 1).
From the analysis of the effects of these factors on the pesticides, the solvent extractor factor showed a significant effect at the 95% confidence level (p < 0.05) over two pesticides (endosulfan-α and iprodione).However, the sample:solvent extractor ratio affected the response for all of the pesticides investigated in this study.Furthermore, reducing the proportion of the extracting solvent to the sample increased the chromatographic response for all of the pesticides.Thus, we chose acetonitrile as the solvent extractor in a 1:1 sample:extractor ratio.
A second full factorial design, 2 2 , was employed to evaluate the effect of adding salt or acid to the extraction of the analytes.The addition of salt to the aqueous phase in the liquid-liquid extraction method using acetonitrile allows for efficient separation of the homogeneous system, improving the extraction of polar compounds. 19The increase in ionic strength also causes a reduction of the solubility of the analytes in the matrix, facilitating their extraction by the organic phase.Acidification of the sample may change the ionized form of certain analytes, affecting their solubility and improving the efficiency of the extraction. 20Maštovská and Lehotay 21 have shown in their studies that the addition of 0.1% (v/v) acetic acid to acetonitrile increases the stability of pesticides prior to analysis.
The analysis of the effect of these factors on pesticides has shown that both salt and acid addition exhibit a significant effect at the 95% confidence level (p < 0.05) on iprodione.Thus, we chose not to add salt or acid in the extraction process.
Taking into consideration the results of the optimization method and the practical aspects of the SLE/LTP technique, a 4.0000 g sample of strawberries in 4.0 mL of acetonitrile was chosen.The mixture was stirred for 15 min on an orbital shaker at 25 °C at a rotation of 200 rpm, centrifuged for 4 min at a rotation of 617.28 g and allowed to cool in a freezer at approximately -20 °C for 9 h.From the supernatant fraction (acetonitrile), 1.5 mL of solution was removed and placed directly into a vial for chromatographic analysis.
The SLE/LTP technique requires low consumption of the sample and solvent extractor because it is not necessary to implement steps of evaporation and solvent exchange, which reduces the risk of contamination and loss of samples.

Method validation
To evaluate the selectivity of the method, the method was applied to strawberry samples without the application of pesticides.Subsequently, these samples were spiked with pesticides and again subjected to the extraction and analysis method.The chromatograms from these samples were then compared.It was observed that there was no interference of the response and the retention time for the analytes of interest.To assess the linearity of the method, seven concentration levels were analyzed, resulting in a determination coefficient (R 2 ) greater than 0.95.The calculated limits of detection (LOD) values for the investigated active ingredients, using GC/EDC, varied from 4 to 16 µg kg -1 , and the limits of quantification (LOQ) varied between 13 and 47 µg kg -1 (Table 3).
Among the studied pesticides, iprodione, λ-cyhalothrin, difenoconazole and azoxystrobin are allowed in strawberry culture.The maximum residue limits (MRLs) for these pesticides in fruits, established by the Brazilian Health Surveillance Agency (ANVISA), 22 are 2.0, 0.5, 0.5 and 0.3 mg kg -1 , respectively, which are above the LOQ values obtained using the technique in this study.Thus, the LODs obtained in the study allow for the detection of residues in samples with values below the MRL, which characterizes them as unsatisfactory or satisfactory.For the other evaluated pesticides, simple detection characterizes the samples as unsatisfactory because they are not allowed in this culture.
The precision of this method was determined to range from 1 to 15% for the lowest concentration evaluated (0.10 mg kg -1 ), 2 to 15% for the middle concentration (0.15 mg kg -1 ), and 1 to 14% for the highest concentration (0.60 mg kg -1 ).The recoveries of the pesticides ranged from 79 to 111% when employing the proposed method.These results are summarized in Table 4.

Determination of pesticides in strawberries from the High Jequitinhonha Valley in Minas Gerais, Brazil
The results obtained for the 21 samples from different farms at three collection periods are presented in Table 5.The pesticides chlorpyrifos, α-endosulfan, β-endosulfan, iprodione, λ-cyhalothrin, permethrin and cypermethrin were not detected and were therefore omitted from the table.the established MRL.Additionally, 38.1% of the samples were considered unsatisfactory for containing residues of unauthorized products and/or presenting levels above the MRL.
Among the unsatisfactory samples, chlorothalonil was detected in 62.5% of the samples (corresponding to 23.8% of the total analyzed samples), which is an unauthorized active principle for the strawberry crop; 25.0% of the samples (9.5% of the total analyzed samples) presented authorized product concentrations above the MRL and two irregularities were observed in 12.5% of the samples (4.8% total).These results are in agreement with the latest Food Pesticide Residues Analysis Program data, 22,23 which revealed that the major irregularities are associated with the application of unauthorized active ingredients.
The use of unauthorized pesticides results in two types of irregularities: applying unauthorized pesticides to the crop whose active ingredient is registered in Brazil and is allowed for other crops, or applying pesticides that have been banned in Brazil or have no registration in the country, which are consequently not allowed for use in any crop. 22nother irregularity is pesticide concentration in fruits above the MRL.This occurrence is disturbing because in some cases, it has been determined that the simple washing of fruits and vegetables is not effective for completely removing the residues of these products.Pesticides can have two action mechanisms in the vegetable: systemic and contact action.Contact action occurs externally to the plant, whereas systemic action occurs when the applied pesticide circulates through the sap in all vegetable tissues and therefore cannot be removed by washing.
The established MRL for azoxystrobin in the strawberry crop is 0.3 mg kg -1 .This active principle was detected For endosulfan, all concentrations followed the same isomeric ratios of the standard.%CV: percent coefficient of variation; %RE: percent relative error.
According to the data in Table 5, residues of the investigated pesticides were not detected in 28.6% of analyzed samples and 33.3% presented residues below

Figure 1 .
Figure 1.Pareto charts of the full factorial experimental design for the analysis of the two variables: solvent extractor and the proportion of sample:extractor.

Table 1 .
2actorial designs, 22, performed in duplicate to establish the best conditions for the extraction of pesticides from strawberry samples

Table 2 .
Factorial design, 2 2 , performed in triplicate at the center point to establish the best conditions for the extraction of pesticides from strawberry samples

Table 3 .
Linear equation, determination coefficient, limit of detection, limit of quantification and maximum residue limits (MRL) for the pesticides studied aActive ingredient not allowed in strawberry culture.

Table 4 .
Precision and accuracy for each analyte obtained from strawberry samples spiked at three different concentration levels a

Table 5 .
Pesticide residues in mg kg -1 found in strawberry samples from the High Jequitinhonha Valley in 2012 Active ingredient not allowed in strawberry culture; b n.d.= not detected; c < LOQ = below quantification limit; d above legal MRL for strawberry crop. a