Multiresidue Methods for Determination of Pesticides using SPME and SPE Followed by GC-NPD System : a Comparative Study

Neste estudo, métodos multirresíduos usando extração em fase sólida (SPE) e microextração em fase sólida (SPME) são comparados na determinação de pesticidas em amostras aquosas. Parâmetros como tipo de fibras e espessura, solvente de eluição, volume da amostra, tempo de equilíbrio, coeficiente de partição e solubilidade foram investigados. Resultados satisfatórios foram obtidos para a análise de pesticidas usando cartuchos C18 e acetato de etila como eluente, com taxas de recuperação entre 75-107%. As melhores condições para SPME foram com uma fibra PDMS 100 μm de espessura, volume de amostra de 40 mL, e tempo de equilíbrio de 45 min. O método SPME é o mais adequado para a análise de pesticidas em água, devido à sua rapidez e simplicidade e ao não uso de solventes. Os limites de detecção e quantificação são inferiores aos limites de concentração máximos estabelecidos pelas autoridades brasileiras, sendo portanto aceitável.


Introduction
The protection of water quality is a real concern of government agencies that control public health.The worldwide intensive use of pesticides plays a key role in environmental contamination, especially in water resources.In the last years, Brazil has been considered the largest consumer of pesticides in the world. 1 Only in 2009, around 1 million ton of pesticides were sold in the country,mainly applied in soybean, corn, sugarcane and cotton cultures. 1According to ANDEF, there were about 475 active ingredients and 1278 products registered. 2 Pesticides that remain in the environment or reach the aquatic systems pose risks to animal species by their toxicity and ability to bio-accumulate along the food chain. 3he European Community has set the value of 0.1 mg L -1 for any individual pesticide and 0.5 mg L -1 for the total pesticide residues of pesticides.][6] The analysis of pesticides in water is commonly performed by chromatographic methods due to the high accuracies and sensitivities. 7Selective nitrogen and phosphorus (NPD) detectors are particularly suitable for the analysis of organophosphorus pesticides.However, low-limit of detection of pesticides in aqueous wastes, as required by environmental agencies, necessitates development of highly efficient and reliable extraction procedures.[10] Although efficient techniques are known, appropriate adaptation is needed in order to reduce the volume of organic solvents and the analysis time.
Solid-phase extraction is a widely used method that requires small volumes of solvent for elution.However, to achieve very low concentration detection limits in the chromatographic system, analyses in water require large sample volumes.Solid-phase microextraction is simple and fast, therefore, the technique has been applied for the determination of micro-pollutants, including non-volatile and volatile organic compounds, phenolics, polycyclic aromatic hydrocarbons, and pesticides. 11,121][22][23][24][25][26] The technique consists in the extraction of analytes using a silica fiber coated with a layer of sorbent, which is exposed directly in the sample or in the headspace.After the equilibrium time has been reached, the fiber is introduced into the chromatographic system for desorption of the analytes.Optimization of SPME involves mainly the selection of the type and the thickness of the fiber, the determination of the exposure time of the analyte in the fiber to reach equilibrium, the influence of the sample volume, the time and the temperature of desorption. 279][30][31] However, nitrogen and phosphorus detector combined with an appropriate extraction method is also able to detect different classes of pesticides with high sensitivity.Our study aimed at comparing SPE and SPME extraction procedures for pre-concentration of selected pesticides from aqueous matrices in order to achieve the detection levels required by Brazilian legislation.The pesticides were selected based on high consumption in regional cultures as well as on control measures by environmental agencies. 32

Standards and samples
Stock solutions of individual pesticides were prepared by diluting 10 mg of the analyte in 10 mL of ethyl acetate.Solutions of mixtures of pesticides (10 mg L -1 ) were prepared from solutions of individual pesticides.Calibration curve were obtained in the range 0.005-1.500mg L -1 and value of the correlation coefficient (R) of each compound.

Solid-phase extraction (SPE)
The recovery of the pesticides was achieved by solid-phase extraction (SPE) using a Manifold System (Supelco, São Paulo, Brazil), which enables extraction of up to 12 samples simultaneously.C18 and Florisil cartridges of 500 mg per 6 mL were used.Different solvents (dichloromethane, ethyl acetate and hexane:acetone) were tested for elution of the analytes.The SPE cartridge was previously conditioned with 10 mL methanol:water (80:20 v/v) and 10 mL methanol (30% v/v).Aqueous solutions containing the pesticides (250 mL) were percolated through the cartridges at a flow rate of 2 mL min -1 .The analytes were eluted with 1 mL of an adequate solvent and then injected in duplicate in a gas chromatograph coupled to a selective nitrogen-phosphorus detector (GC-NPD).Four concentrations (1.0, 2.0, 4.0 and 8.0 µg L -1 ) were used to investigate the recovery of the pesticides for each cartridge and eluting solvent selected.Calibration curves were allowed to obtain the value of the correlation coefficient (R) of each compound.

Solid-phase microextraction (SPME)
Two commercially available fibers of polydimethylsiloxane (PDMS; 100 µm thickness) and Carboxen-PDMS (75 µm thickness) were evaluated for their extraction efficiencies towards the pesticides.The fibers were initially conditioned for at least 1 h at 250-300 °C, according to the manufacturer's instructions.The conditioned fiber was immersed into the aqueous solution and maintained at the equilibrium time under stirring (150 rpm) at ambient temperature (28 ± 2 °C).After extraction, the fiber was thermally desorbed during 5-10 min into the glass liner of the GC injection port.The equilibrium time was determined by varying the exposure time of the fiber in 30 mL of a spiked sample (4 mg L -1 ) for 15, 30, 45 and 60 min.The effects of the sample volumes on the retention of the pesticides were evaluated using volumes of 20, 40 and 80 mL.A five-level calibration curve was constructed using aqueous standards (1.0 to 8.0 µg L -1 ) that were extracted in the same conditions as those applied for the samples.The values of the limits of detection and of quantification were calculated by equations 1 and 2, respectively, where s is the standard deviation of the lower level of detection (7 injections) and S is the slope of the curve.

Chromatographic conditions
The analyses of the pesticides were performed using a gas chromatograph (GC-Trace, Thermo Finnigan) equipped with a split/splitless injector and a selective nitrogen-phosphorus detector (NPD).All separations were accomplished on a OV-5 column (30 m × 0.25 mm I.D. × 0.25 mm thickness).The initial temperature of the oven was 100 °C and the temperature was increased to 150 °C at a rate of 25 °C min -1 , then to 290 ºC at a rate of 30 °C min -1 .The temperatures of the injector and the detector were set at 250 °C and 300 °C, respectively.The splitless injection mode was used (at 2 min) and 2 µL of the sample volume was injected.Helium was used as the carrier gas at a flow rate of 1 mL min -1 .

Chromatographic separation and calibration curves
Initially, solutions of individual pesticides were studied in order to determine the retention time in GC system, afterwards the mixture of the pesticides was separated using a suitable temperature program.The retention times and the calibration curves are given in Table 2.The calibration curves for all five analytes were linear over the range 0.005-1.500mg L -1 , as shown in Table 2.The coefficient of correlation, R, exceeded 0.99 for all the compounds studied.

Solid-phase extraction (SPE) Selection of adsorbent and eluent
Experiments were conducted to evaluate the recovery of the pesticides using C18 and florisil cartridges.The best results were obtained for C18 cartridges with recoveries between 75% and 107%, while low recoveries were observed using florisil cartridges.Ethyl acetate (ACET) and dichloromethane (DCM) were suitable eluents on the C18 cartridges (Table 3).Ethyl acetate was selected in view of the superior recoveries (85-107%) with respect to dichloromethane (75-106%).

Accuracy, precision and linearity
The recoveries obtained for each concentration (Table 4) proved to be reproducible, as represented by the low values of the standard deviations.Neto and Siqueira, 9 also used C18 cartridges and ethyl acetate as eluent to monitor organophosporus pesticides.

Solid-phase microextraction (SPME) Effect of film thickness and nature fiber on the extraction of pesticides
Figure 1 shows the results of the extraction of pesticides by using two different SPME fibers.Clearly, the 100 µm PDMS fiber shows a higher efficiency than the 75 µm carboxen-PDMS.Besides the type of the fiber, the thickness may also influence the extraction efficiency, because thicker fibers tend to retain higher amounts of analytes. 11DMS fiber values mass calculated for parathion, malathion, chlorpyrifos, pendimethalin and triazophos were higher than carboxen-PDMS fiber, except for molinate.Thereby the values mass on PDMS fiber for chlorpyrifos and pendimethalin are particularly noteworthy because it is around 5 times higher than the  1.The pesticides whith lower polarity, chlorpyrifos (log K ow = 4.7) and pendimentalina (log K ow = 5.2), showed a higher amount of mass retained due to higher affinity for the PDMS fiber (nonpolar).Similar results were found by Beltran et al. 33 and Silva et al. 34 for extraction of organophosphorus pesticides using PDMS and PA fibers.According to Dugay et al., 12 carboxen-PDMS fibers are particularly suitable for lower-molecular-weight and more volatile analytes.Thus, the PDMS fiber of 100 µm thickness was used in the subsequent studies.

Experimental determination of the partition coefficient (K)
It's well known that the equilibration time depends on the partition coefficient (K) and for higher the K value, more the larger the amount extracted at equilibrium. 35The results may be interpreted via the partition coefficients (K) of the analyte between the stationary and the aqueous phase.According to equation 3, higher values of K give higher extraction efficiencies: where n s is the number of moles of analyte, V aq is the sample volume, V s is the volume of the stationary phase and C aq is the initial concentration of the analyte.The value of V s can be calculate as where e is the film thickness and a is the diameter of the fiber of the silica rod.
Taking into account the precision on both V s and n, and the fact that equilibrium should be reached before n values can be determined, the partition coefficient values can only be approximated.
The values of K have been determined using two types of fibers and the values are shown in Table 5.The high values of K were obtained for the PDMS fiber.Although an analyte may have been more efficiently retained, the K values depend also the fiber thickness.The results obtained in this study show that K values can vary with film thickness, depending on the analyte.

Determination of the equilibrium time
The study of the equilibrium time was carried out by determining the extracted amount as a function of the exposure time obtained with the PDMS fiber at a concentration of 4.0 mg L -1 .According to the results the equilibrium is reached in about 30 min for most compounds, except for chlorpyrifos and pendimethalin.Thus, we adopted a time of 45 min for further studies to ensure total retention of the pesticides.Other authors also found times between 15 and 45 min suitable for extraction of pesticides from water. 16,36,37learly, water solubility was important, but not the only, factor in determining the overall partition ratio for a given analyte between the carboxen-PDMS fiber coating and water.

Variation of the sample volume
Different volumes were investigated for the extraction of pesticides by SPME.As shown in Table 6, it appears that a volume of 40 mL is sufficient for recovery of the pesticides under the conditions used, and, therefore, this volume was used for subsequent experiments.

Influence of pH of the sample
The Figure 2 shows the effect of pH of the sample in the retention of pesticides using SPME.According to the results, at pH 3, 7 and 9 did not significantly influence the extraction of pesticides molinate, atrazine, methyl parathion, fenitrothion, pendimentalina and triazophos.The pesticide malathion and chlorpyrifos were extracted with higher efficiency at pH 7.

Efficiency of extraction by SPME
The efficiency of extraction of the pesticides was evaluated using four concentrations.The response was linear in the concentration range studied (1.0 to 8.0 µg L -1 ), as represented by the correlation coefficients of about 1.The calibration curves and recoverie rates of the pesticides after extraction by SPME are show in Table 7.

Comparison of SPE and SPME extraction efficiencies
Table 8 summarizes the limits of detection (LOD) and limits of quantification (LOQ) values calculated for  all analytes studied, beyond maximum limit permitted in the Brazil.The LOD values using SPME for parathion, malathion, fenitrothion, chlorpyrifos, pendimethalin and triazophos were lower than LOD using SPE, except for molinate.Thereby the LOD value using SPME for malathion is particularly noteworthy because it is around 50 times lower than the regulatory limit cited in CONAMA (Brazil).Tomkins and Ilgner, 38 found similar results for malathion using a 65-mm thickness polydimethylsiloxanedivinylbenzene (PDMS-DVB) fiber.Table 9 gives the detection limit of some pesticides found by other authors 33,39 by using different SPME fibers with detection by GC-NPD or GC-MS system.The values are comparable to those found in the present study.
The extraction efficiencies for SPME and SPE are comparable and concentration levels below those required in current legislation were attained (Table 8).For routine analysis, SPME should be preferred because of the rapidity, sensibility, and free-solvent.

Conclusions
The present study shows that both solid-phase extraction (SPE) and solid-phase microextraction (SPME) are highly efficient for the analysis of pesticides using C18 cartridges and ethyl acetate as the eluent, as represented by the high recovery rates.The best conditions for SPME were obtained with a PDMS fiber of 100 mm thickness, a sample of 40 mL, and an equilibration time of 45 min.Both procedures showed a linear response in the concentration range between 1.0 and 8.0 µg L -1 .SPME is most suitable for the analysis of pesticides in water based on the rapidity of the method and the restricted use of eluent.The limits of detection and the limits of quantification are well below the maximum limits set by the Brazilian authorities.

Table 1 .
Main physicochemical characteristics of the pesticides studied

Table 3 .
Selection of eluent using C18 cartridges of 500 mg per 6 mL

Table 2 .
Retention times and calibration curves of the pesticides in water

Table 5 .
K values as a function of the film thickness of fibers at equilibrium

Table 4 .
Results of percentages of recovery of the pesticides in water sample using SPE

Table 6 .
Sample volume used for extraction of pesticides by SPME

Table 7 .
Calibration curves and percentage of recovery of the pesticides after extraction by SPME

Table 8 .
Limits of detection (LOD) and limits of quantification (LOQ) of pesticides analized whith SPE and SPME, and the maximum limit permitted in Brazil

Table 9 .
Limits of detection (LOD) of the pesticides using different analytical techniques and comparison with our results