Comparison between GC-MS-SIM and GC-ECD for the Determination of Residues of Organochlorine and Organophosphorus Pesticides in Brazilian Citrus Essential Oils

Óleos essenciais cítricos brasileiros (mandarina, laranja, limão, bergamota e tangerina) foram analisados a fim de determinar resíduos de pesticidas organoclorados (aldrin, clordano, dieldrin e dicofol) e organofosforados (clorpirifós metílico, dimetoato, metidationa e paration metílico). Um estudo comparativo entre cromatografia gasosa acoplada a espectrometria de massas em modo de monitoramento seletivo de íons (CG-EM-MSI) e cromatografia gasosa com detector por captura de elétrons (CG-DCE) foi realizado usando extração em fase sólida (EFS) com fase Florisil para pré-concentração dos pesticidas. A recuperação média dos pesticidas foi de 71 a 83% por CG-EM-MSI e de 99 a 104% por CG-DCE. O limite de quantificação foi de 0,93 a 3,93 mg L por CG-EM-MSI e de 0,09 a 0,12 mg L por CG-DCE. Quinze amostras comerciais de óleos essenciais cítricos brasileiros foram analisadas e seis amostras apresentaram contaminação com valores acima do limite máximo de resíduos (LMR) permitido pelo Codex Alimentarius.


Introduction
Citrus essential oils (EOs) are extracted by coldpressing the fruit pericarp and are important raw materials for cosmetic, pharmaceutical, food and cleaning product industries, among others. 1Due to environmental issues and problems caused by pesticide poisoning, major efforts have been employed by several countries to control these contaminants.Brazil, as an important agricultural producer and exporter, has been increasingly concerned with attending the international legislation regarding the presence of contaminants in food. 2 The maximum amount of pesticide residue in food officially accepted for human or animal ingestion is called maximum residue limit (MRL).MRL is expressed by mg kg -1 or mg L -1 , and is established by Codex Alimentarius Commission (FAO/WHO, Food and Agriculture Organization/World Health Organization). 2,3There is no Vol.23, No. 2, 2012 defined value of MRL for EOs per se.The residue control of these products is based on MRL for citrus fruits.
There are few investigations of pesticide residues in citrus EOs available in the scientific literature.Di Bella et al. [4][5][6][7][8] investigated Italian citrus EOs from 1996 to 2004, and found 60 mg L -1 of chloroparaffin residue in lemon oil, above the level allowed by the Codex Alimentarius (7.1 mg L -1 ). 5 This group also observed a dicofol contamination of 1.96 mg L -1 in mandarin EO and 0.20 mg L -1 of the same organochlorine in bergamot EO. 7 In 2004, a second study noted the presence of pesticides, but now at levels below MRL. 8 Moreover, Dellacassa et al. 9 analyzed organochlorine and organophosphorus residues in lemon EO from Uruguay.Contamination of methyl chlorpyrifos, methyl parathion, fenitrothion and methidathion was found in the EOs produced before 1993, but a strong reduction in the residues was observed after that year. 9hese studies highlight the importance of the constant monitoring of pesticide residues in agricultural-based products, as well as the need for development of methodologies with higher sensitivity, precision and accuracy.Gas chromatography-mass spectrometry (GC-MS) and gas chromatography-electron capture detector (GC-ECD) are the most common and appropriate systems to investigate contaminants in food.Aramendia et al. 10 compared the GC-MS/MS and GC-ECD methods in the analysis of triazines residues in olive oils.The authors showed the high sensitivity of GC-ECD for organophosphorus and organochlorine pesticides. 10But they pointed out, as a drawback, that the identification of the compounds obtained by comparison with standards, makes the quantification of the pesticides not very reliable due to matrix interferences. 10On the other hand, gas chromatography-mass spectrometry in selective ion monitoring mode (GC-MS-SIM) provides excellent sensitivity and satisfactory quantification limits, allowing the identification and quantification of pesticides at low levels. 10he aim of this work is the validation of the GC-MS-SIM and GC-ECD methods and their comparison in the determination of residues of organochlorine and organophosphorus pesticides in Brazilian citrus EOs.

Instrumentation and chromatographic conditions
GC-MS analyses were performed in a GC6850 gas chromatography coupled to a 5975C MSD Instrument (Agilent Technologies) with electron ionization at 70 eV.A HP5-MS capillary column (Agilent 19091S 433E) with a 0.25 mm i.d., 30 m and 0.25 μm phase film diameter was used.The carrier gas was helium at a flow of 1.0 mL min -1 .The temperature program was an isothermal period of 5 min at 50 o C, then increased to 150 o C at a rate of 20 o C min -1 and finally to 290 o C at a rate of 10 o C min -1 with a final isothermal period of 5 min.The run time was of 29 min and splitless mode was used.The injector and interface temperatures were held at 290 o C. All samples were analyzed in SIM mode for quantification measurements and SCAN mode in the mass range of 50-400 u for confirmation of the spectral data.
GC-ECD analyses were performed in a GC2014 gas chromatography with electron-capture detector (model ECD2014 AOC20) radioisotope Ni 63 (Shimadzu).The same HP5-MS capillary column with a 0.25 mm i.d., 30 m and 0.25 μm phase film diameter was used.The carrier gas was nitrogen at a flow of 1.0 mL min -1 .The temperature program was the same reported for GC-MS-SIM.The injector temperature was held at 290 o C and splitless mode was used.

Standard stock solution
For GC-MS analyses, a standard stock solution of each pesticide was prepared by dissolving 1.0 mg of each pesticide in 1.0 mL of ethyl acetate (concentration 1000 mg L -1 ), in a volumetric flask.For GC-ECD analyses, 1.0 μL of the GC-MS stock solution was taken using automatic pipette and diluted in 1.0 mL of ethyl acetate (concentration 1 mg L -1 ).The stock solutions were stored at −20 °C.

GC-MS pesticide fortified solution
A fortified solution was prepared by weighing 8 mg of the EO, adding 40 mL of the 1000 mg L -1 stock solution of each pesticide and diluting (in a volumetric flask) to 4.0 mL with ethyl acetate.This procedure was repeated for each one of the five EOs analyzed (mandarin, orange, lemon, bergamot and tangerine).The concentration of these solutions was 2000 mg L -1 of EO and 10 mg L -1 of each pesticide.
Comparison between GC-MS-SIM and GC-ECD for the Determination of Residues of Organochlorine J. Braz.Chem.Soc.308

GC-ECD pesticide fortified solution
For GC-ECD analyses, the fortified solutions were prepared by mixing 10 mL of the 2000 mg L -1 EO solution and 10 mL of each diluted pesticide stock solution (1 mg L -1 ), the volume was then completed to 1.0 mL with ethyl acetate.The concentration of these solutions was 20 mg L -1 of EO and 0.01 mg L -1 of each pesticide.

Internal standard solution
n-Pentacosane was used as internal standard (IS).The stock solution was prepared by dissolving 1.0 mg of IS in 1.0 mL ethyl acetate (concentration 1000 mg L -1 ).The solution was stored at −20 °C.

GC-MS calibration solutions
Five standard EO solutions for GC-MS-SIM analyses were prepared by mixing 1, 5, 10, 15 and 20 mL of each 1000 mg L -1 pesticide stock solution, 500 mL of orange EO (EO 9) and the volume was completed to 1.0 mL with ethyl acetate, obtaining solutions with 1, 5, 10, 15 and 20 mg L -1 , respectively.A volume of 200 mL of the standard EO solution was transferred to the Florisil ® cartridge.The cartridge was pre-conditioned with 3 mL of acetone and the elution of 5 mL of pentane yielded the first fraction (F1).The second fraction (F2) was collected after the elution of 5 mL of ethyl acetate.The elution was processed at 1 drop min -1 without pressure apparatus.F1 and F2 fractions were dried under nitrogen flow and resuspended with 5 mL of the 1000 mg L -1 IS stock solution and completed to 50 mL with ethyl acetate.According to this procedure, the calibration graphs were built up with 4, 20, 40, 60 and 80 mg L -1 of pesticide and IS at 100 mg L -1 .Pesticide concentrations correspond to 2, 10, 20, 30 and 40 mg L -1 of pesticide in the original EO, respectively.

GC-ECD calibration solutions
Organochlorine and organophosphorus pesticide standard solutions were prepared in the range of 50-150 μg L -1 for GC-ECD.The calibration curves were constructed from five solutions of 50, 75, 100, 125 and 150 µg L -1 of each pesticide.A solution of 50 µg L -1 was prepared with 50 mL of the 1 mg L -1 stock solution of each pesticide and 10 mL of the 2000 mg L -1 EO solution, the volume was then completed to 1.0 mL with ethyl acetate.

Commercial samples
Fifteen commercial samples of Brazilian citrus oils were analyzed.Six of the samples were orange EOs (EO 2, 4, 7, 9, 14 and 15), five were lemon (EO 1, 3, 5, 10 and 12), one was tangerine (EO 8), two samples were mandarin EOs (EO 6 and 11) and one bergamot (EO 13).For GC-MS-SIM analyses, solutions of 500 μL of each EO in 500 μL of ethyl acetate were prepared, and then a volume of 200 mL was extracted through a SPE Florisil ® cartridge.The F1 fraction was dried under nitrogen and resuspended with 5 mL of the 1000 mg L -1 IS stock solution and completed to 50 μL with ethyl acetate.For the GC-ECD analysis, the same procedure was repeated but the F1 fraction was resuspended and completed to 100 μL with ethyl acetate.Then, 10 μL of this solution was completed to 1.0 mL with ethyl acetate.

Results and Discussion
Qualitative analysis of pesticides GC-MS-SIM and GC-ECD chromatograms of the pesticide fortified solutions, in each type of EO (orange, lemon, tangerine, mandarin and bergamot), were obtained and the peak identifications were confirmed by injecting different concentrations of the pesticides.Table 1 presents the retention times (under the chromatographic conditions used) and the characteristic fragments selected for qualitative and quantitative determinations by GC-MS.The pesticide chlordane can present 26 isomers, but only areas of two peaks increased by the fortification of the EO.Figures 1 and 2 illustrate the GC-MS-SIM and GC-ECD chromatograms, respectively, of lemon EO (EO 1) fortified with standard pesticides.The pesticide Aldrin was not analyzed by GC-ECD.
After peak identification, pure essential oils were analyzed by GC-MS-SIM and GC-ECD.The purpose of this test was to detect compounds which could have the same retention time or the same ions chosen in the GC-MS-SIM method for the quantification of the pesticides and n-pentacosane IS in the matrix.The analyses showed that the EOs did not have compounds close to the retention time of the analytes or the internal standard.Thus, it was possible to affirm that the peaks in these areas only correspond to the pesticides or the internal standard.

Evaluation of repeatability and recovery of solid phase extraction
Florisil ® SPE cartridge was used for the concentration of the analytes and clean-up of the matrix.Preliminary analyses showed that all the pesticides were in the F1 fraction, and the F2 fraction was discarded.
For the evaluation of repeatability, the standard EO solution with 10 mg L -1 of each pesticide was used.The F1 fraction was dried under nitrogen flow and resuspended with 200 mL of ethyl acetate.This procedure was done in triplicate and each F1 solution was also injected in triplicate in GC-MS-SIM.In order to check the recovery of pesticides on Florisil ® SPE, the standard solution containing 50 mg L -1 of each pesticide and IS (without the EO) was first analyzed.The effect of the drying step in F1 was investigated analyzing nine repetitions with and without the nitrogen drying step.The response factor of each pesticide was calculated using concentrations and area data from three consecutive direct injections of the standard solution.The results are given in Table 2.
The results indicate the good repeatability of the SPE procedure since all the relative standard deviation (RSD) values were below 5%. 11,12The recovery values were near 100% and were very similar, showing that the use of adsorbent and nitrogen drying step did not affect the pesticide analysis.The average recovery values of 96% for organochlorine and 95% for organophosphorus pesticides using Florisil ® SPE are in agreement with values described in the literature. 13he recovery of pesticides in the presence of EO matrix was also evaluated for the SPE procedure.The quantitative   analysis was done for the five types of EO used in this work.
The results are summarized in Table 3, where low recovery was found.This low recovery was also obtained by González et al. 14 using various matrices and this effect is caused by the increasing level of noise in GC-MS, resulting in a decrease of the integrated area of the analytes. 14So, it is noteworthy that the matrix has to be used in the construction of the calibration curve in order to ensure accuracy of the results.

Quantitative analysis of pesticides
Only the orange EO (EO 9) was used to obtain a calibration curve.At first, because the recovery of pesticides in the matrices of the five EOs was very similar, as can be seen in Table 3, and secondly, it presented a smaller number of compounds, simplifying the measurement of different levels of the pesticides used.
In GC-MS-SIM, the calibration graphs were built up with five points in triplicate at 4, 20, 40, 60 and 80 mg L -1 of pesticide and IS at 100 mg L -1 .In GC-ECD, the calibration graphs were constructed with five points in duplicate at 50, 75, 100, 125 and 150 μg L -1 .Grubbs and Cochran 15 statistical tests were applied and none of the points was discarded.The standard calibration parameters are summarized in Table 4.The curves showed a good linear relationship between the instrumental response and the analyte concentration for both methods.
The GC-ECD showed higher sensitivity than GC-MS, even in SIM mode.In GC-MS-SIM and GC-ECD, the limits of detection (LOD) and quantification (LOQ) were calculated on the basis of a signal-to-noise ratio of 3:1 and 10:1, respectively. 16,17In GC-ECD, the LOD and LOQ values are below or close to the values of MRLs, making easier the identification and determination of possible contaminations.
The precision and accuracy of the methods were evaluated using repeatability and recovery values, respectively.Both were calculated for the five concentrations of the calibration  graphs.The average RSD (%) value and the average recovery (%) for each pesticide are shown in Table 5.The RSD values were below 5% for all concentrations. 11,12The recovery values obtained in the GC-MS-SIM varied from 71 to 83%.These results showed an acceptable recovery considering the effect of the matrix, as previously discussed.
In GC-ECD, the recoveries were from 99 to 104%, which can be attributed to the high selectivity of the electron capture detector for organochlorine and organophosphorus compounds.The recovery values obtained in both GC-MS-SIM and GC-ECD are according to the values reported in the literature.Barrek et al. 18 using GC-MS-SIM obtained recoveries from 79 to 84% for organochlorine and from 88 to 93% for organophosphorus in pesticide analyses of citrus EO using Florisil ® SPE.Milhome et al. 19 studied the determination of pesticides (multiresidue method) using SPE C18 in aqueous samples and obtained a recovery of 86.8% for chlorpyrifos and 92.1% for methyl parathion.Zuin et al. 20 using GC-ECD obtained recovery values of 72 to 111% for organochlorine and organophosphorus pesticides.The limits of detection and quantification, repeatability and recovery showed that GC-ECD is better than GC-MS-SIM for the analysis of all pesticides tested.These results suggest that GC-MS-SIM could be used as a screening method for the identification of the pesticide and GC-ECD as quantitative method.
Determination of pesticide residues in commercial EO Fifteen Brazilian commercial citrus EOs were analyzed in order to investigate possible contamination of pesticides.The detected pesticides are shown in Table 6.
Figure 3 shows the GC-ECD and GC-MS-SIM chromatograms with the residues of dicofol in orange EO (EO 2) and Figure 4 presents lemon EO (EO 10) chromatograms in which residues of methyl parathion could be detected.The MRL established by the Codex Alimentarius 3 for dicofol in citrus culture is 5.0 mg L -1 and the value of this pesticide residue found by GC-ECD was 35.36 mg L -1 .The Codex Alimentarius does not establish any MRL reference value for the methyl parathion 3 and the use of this pesticide is not authorized for citrus culture.The residue was quantified at 43.45 mg L -1 by GC-ECD.
Chlorpyrifos residue was detected in three samples of lemon EO (EO 1, EO 5 and EO 10), one sample of orange     (EO 9) and one of bergamot (EO 13).Figure 5 shows the GC-ECD chromatograms of the methyl chlorpyrifos contamination.The residues of methyl chlorpyrifos were much higher than that permitted by the Codex Alimentarius. 3MRL for this pesticide in citrus culture is 2.00 mg L -1 .The dimethoate residue in EO 2 was 37.60 mg L -1 and methidathion residue in EO 9 was 26.27 mg L -1 .Both are orange EO and the value of MRL for these pesticides is 2.00 mg L -1 .Figure 6 shows the GC-ECD chromatograms of dimethoate and methidathion residues.
In GC-MS-SIM and GC-ECD, the LOD values for chlordane were 0.30 and 0.07 mg L -1 , respectively.MRL for chlordane in citrus culture is 0.02 mg L -1 .This pesticide was not detected above the LOD of both methods.The methods were not able to detect the presence of this pesticide above the MRL, from 0.02 mg L -1 to 0.07 mg L -1 by GC-ECD for example.

Conclusion
The Florisil ® SPE cartridge associated to GC-MS-SIM or GC-ECD is efficient in the extraction and pre-concentration of pesticide residues in citrus EOs, allowing also their quantification.The quantitative results showed that GC-ECD is better than GC-MS-SIM for analysis of all pesticides tested.From fifteen investigated samples, six (40%) contain pesticide residue.In all of them, the pesticides were quantified in concentrations above their MRL.The lemon EO 10, as well as the orange EOs 2 and 9, presented residues of two pesticides.The use of methyl parathion is not allowed in citrus culture, however this pesticide was found in EO 10.These results emphasize the need for monitoring cultivation of citrus fruits and contribute to the standardization of the quality of Brazilian citrus essential oil production.

Figure 2 .
Figure 2. GC-ECD chromatograms of the fortified lemon EO (upper line) (a) organochlorine pesticides: dicofol, chlordane and dieldrin and (b) organophosphorus pesticides: dimethoate, methyl parathion, methyl chlorpyrifos and methidathion.The botton lines show the chromatogram of pesticide standard solutions for peak identification.
coefficient; b not analyzed by GC-ECD; c Codex Alimentarius; d no reference value; LOD, LOQ and MRL in mg L -1 .

Figure 3 .
Figure 3. EO 2 chromatograms showing the residue of dicofol: (a) GC-MS-SIM and (b) GC-ECD, upper line commercial EO and bottom line standard solutions.

Figure 4 .
Figure 4. Chromatograms of EO 10 that showed residue of methyl parathion: (a) GC-MS-SIM and (b) GC-ECD ECD, upper line commercial EO and bottom lines standard solutions.

Table 1 .
Retention times (t R ) and selected characteristic ions for GC-MS

Table 2 .
Repeatabilities (n = 9) and recoveries of pesticides using SPE Florisil ® by GC-MS-SIM a Rate between pesticide and IS areas; b relative standard deviation.

Table 3 .
Recoveries of pesticides in each EO matrix

Table 4 .
Parameters of calibration equations

Table 5 .
Repeatabilities and recoveries of the pesticides in GC-MS-SIM and GC-ECD

Table 6 .
Results of commercial EOs analyzed by GC-ECD