ARTICLE

pubs.acs.org/JAFC

Organochlorine Pesticide Residues in Strawberries from Integrated

Pest Management and Organic Farming
Virgínia C. Fernandes,§,# Valentina F. Domingues,*,§ Nuno Mateus,# and Cristina Delerue-Matos§
§
REQUIMTE, Instituto Superior de Engenharia do Porto, Rua Dr. Ant
onio Bernardino de Almeida 431, 4200-072 Porto, Portugal
#
CIQ-Centro de Investigac-~ao em Química, Departamento de Química, Faculdade de Ci^encias, Universidade do Porto, Porto, Portugal

ABSTRACT: A rapid, specific, and sensitive method based on the Quick Easy Cheap Effective Rugged and Safe (QuEChERS) method
and a cleanup using dispersive solid-phase extraction with MgSO4, PSA, and C18 sorbents has been developed for the routine analysis of
14 pesticides in strawberries. The analyses were performed by three different analytical methodologies: gas chromatography (GC) with
electron capture detection (ECD), mass spectrometry (MS), and tandem mass spectrometry (MS/MS). The recoveries for all the
pesticides studied were from 46 to 128%, with relative standard deviation of <15% in the concentration range of 0.005-0.250 mg/kg.
The limit of detection (LOD) for all compounds met maximum residue limits (MRL) accepted in Portugal for organochlorine pesticides
(OCP). A survey study of strawberries produced in Portugal in the years 2009-2010 obtained from organic farming (OF) and
integrated pest management (IPM) was developed. Lindane and β-endosulfan were detected above the MRL in OF and IPM. Other
OCP (aldrin, o,p0 -DDT and their metabolites, and methoxychlor) were found below the MRL. The OCP residues detected decreased
from 2009 to 2010. The QuEChERS method was successfully applied to the analysis of strawberry samples.
KEYWORDS: organochlorine pesticides, gas chromatography, mass spectrometry, tandem mass spectrometry, electron capture
detection, strawberries, organic farming

’ INTRODUCTION useful for the analysis of pesticides of varying polarities, by virtue

Despite the usefulness of pesticides in agricultural practices, of the fact that the sorbent used focuses on binding sample matrix
there are concerns about their excessive use and their presence compounds without interacting with the target analytes.11
and levels in the environment and foodstuffs. Integrated pest The determination of OCP residues is generally performed by
management (IPM) has been adopted to maximize crop yield GC using specific detectors, such as ECD12-14 and MS.11,15,16
while maintaining a quality environment with regard to pesticide The use of MS detectors17 is widely extended due to their selec-
tivity and low detection limits.18 Selected ion monitoring (SIM)
contamination, solving ecological problems and contributing to
mode allows reducing background noise but does not eliminate
the sustainability of agriculture.1 These preventive practices were
matrix interferences in all cases. Certain MS detectors can
preceded by other actions. For instance, OCP have been banned
perform in tandem mass spectrometry (MS/MS), which pro-
since the 1970s because they are extremely persistent in the
vides higher selectivity and sensitivity.19
environment and due to their ability to accumulate in sediments, The aim of this study was to set up a multiresidue methodol-
plants, and animals.2 OCP have an extensive range of both acute ogy for the determination of 14 OCP (R-, β-, γ-, and ζ-
and chronic health effects, including cancer, neurological da- hexachlorocyclohexanes (HCHs), hexachlorobenzene (HCB),
mage, and birth defects. Many OCP are also suspected to act as o,p0 -DDT ([1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane]), and
endocrine disruptors.3-5 Consequently, routine and compre- their degradation products 2,2-bis(p-chlorophenyl)-1,1-dichlor-
hensive testing of OCP is given consideration in the scientific oethylene (p,p0 -DDE) and dichlorodiphenyldichloroethane (p,p0 -
literature and enforcement actions.3,6,7 Particularly, the control DDD), aldrin, dieldrin, endrin, R- and β-endosulfan, and meth-
of OCP in fruits and vegetables is important for regulatory oxychlor) in strawberries, by the QuEChERS method and GC. The
agencies to ensure that concentrations of toxic pesticides are developed approach was, first, the use of GC-ECD due to the high
below tolerance levels,8 especially because these foods, in most sensitivity of detector to chlorinated compounds for quantification
cases, are eaten raw and often unwashed. method, followed by GC-MS operated in SIM mode and finally
Traditional methods for trace analysis of OCP are character- GC-MS/MS, the last ones as a confirmation method. Nowadays,
istically time and solvent-consuming, and their sample through- this sequence is being used by some pesticide laboratories for
put is too low.9 Development of simple and reliable methods is a multiresidue targeted screening of pesticides in food samples.20,21
challenging task. The Quick Easy Cheap Effective Rugged and The methods were applied for the detection of OCP in
Safe (QuEChERS) method has recently attracted attention for
pesticides analysis given its advantages of (i) high recovery for
Special Issue: Florida Pesticide Residue Workshop 2010
wide polarity and volatility ranges of pesticides; (ii) high sample
throughput; (iii) the use of smaller amounts of organic solvent Received: October 6, 2010
and the use of nonchlorinated solvents; (iv) the need for very Revised: November 25, 2010
little labware and increased safety for laboratory workers.10 The Accepted: November 29, 2010
main advantage of QuEChERS is that it is comprehensive, being Published: January 10, 2011

r 2011 American Chemical Society 7582 dx.doi.org/10.1021/jf103899r | J. Agric. Food Chem. 2011, 59, 7582–7591
Journal of Agricultural and Food Chemistry ARTICLE

strawberries sampled from organic farming (OF) and IPM and Sigma-Aldrich Co. n-Hexane and acetonitrile were of HPLC
cultivated in central Portugal. Adequate sampling of strawberry grade from Merck. All solutions were prepared in n-hexane, and the
collection was done in May of 2009 and 2010. To compare internal standard (IS) solution of 4,40 -dichlorobenzophenone was
samples produced by different cultivation methods, a careful 350 μg/L. Mixed fortification standards, each containing 14 OCP
sampling was carried out taking into account (i) the maturity (R-, β-, γ- and ζ-HCHs, HCB, o,p0 -DDT, p,p0 -DDE, and p,p0 -DDD,
(75% of the strawberry should be red); (ii) the type of farming aldrin, dieldrin, endrin, R- and β-endosulfan, and methoxychlor) were
(at least 5 years in IPM and OF); (iii) the variety of strawberry; prepared by diluting stock standard solutions. All working solutions were
(iv) similar geographic location; and (v) type of soil, in order to stored in glass vials at -20 °C prior utilization. The selected QuEChERS
perform a statistical correlation that may affect the studied (EUMIV50CT-VP) and cleanup (CUMPS15C18CT) were obtained
factors.22,23 from UCT (Bristol, PA).
Sampling. The sampling was scheduled with producers during the
two years of harvest and was performed by the authors. Strawberry
’ MATERIALS AND METHODS samples were collected in two types of cultivation: OF (N 41° 12.0400 ;
The pesticide standards in this study were obtained from Chemser- W 8° 39.9430 ) and IPM (N 38° 35.1000 ; W 8° 41.4460 ), both from
vice (West Chester, PA), Dr. Ehrenstorfer GmbH (Augsburg, Germany), Palmela in central Portugal. We collected different varieties of strawber-
ries including Siba, Camarosa, Festival, and Albion. The sampling period
was the first week of May in 2009 and 2010. The 2009 and 2010 fields of
Table 1. Quantitation and Identification Ions for the GC-MS
OF were close due to crop rotation.
and GC-MS/MS Analyses of Selected OCP Procedure. The extraction and cleanup steps were not optimized
because the original method24 provided adequate recoveries and preci-
SIM MS/MS
sion. Samples were crushed and homogenized using a blender jar.
identification quantification precursor Strawberry samples were frozen and stored at -20 °C until required
OCP ions (m/z) ion (m/z) ions (m/z) for analysis. An amount (10 g) of strawberries was weighed into a 50 mL
centrifuge tube, and 50 μL of IS solution was added. The strawberry
(R, β, δ) HCH 109, 181, 219 181 219
sample was left during 30 min at room temperature to let the n-hexane
HCB 142, 284, 286 284 142 evaporate before the addition of 10 mL of acetonitrile.
lindane 181, 183, 219 181 183 The resulting solution was shaken for 1 min, followed by the addition
aldrin 66, 263, 293 263 263 of 6 g of anhydrous magnesium sulfate, 1.5 g of sodium chloride, 1.5 g of
R-endosulfan 195, 207, 241 241 195 trisodium citrate dehydrate, and 0.75 g of disodium hydrogencitrate
dieldrin 79, 263, 277 263 243 sesquihydrate. The centrifuge tube was capped and shaken vigorously
p,p0 -DDE 176, 246, 318 246 318 for 1 min to prevent salt agglomeration before centrifugation at 3000
rpm for 5 min. An aliquot of 1.5 mL was sampled from the upper layer
endrin 81, 263, 281 263 245
into a 2 mL cleanup vial containing 150 mg of primary secondary amine
β-endosulfan 195, 207, 241 195 195
(PSA), 150 mg of MgSO4, and 50 mg of C18 and again vortexed for 1
p,p0 -DDD 165, 235, 318 235 235 min and then centrifuged for 5 min at 4000 rpm. From the upper layer of
o,p0 -DDT 165, 235, 352 235 235 the prepared sample, an aliquot of 1.0 mL was transferred into a vial and
methoxychlor 227, 237 227 227 put under a mild stream of nitrogen to dryness. Finally, 1 mL of hexane

Figure 1. GC-ECD chromatogram of mixture OCP (70 μg/L). Peak identification in order of increasing retention time: 1, HCB; 2, lindane (γ-HCH);
3, aldrin; 4, IS; 5, R-endosulfan; 6, dieldrin; 7, p,p0 -DDE; 8, endrin; 9, β-endosulfan; 10, p,p0 -DDD; 11, o,p0 -DDT; 12, methoxychlor.

7583 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591

Journal of Agricultural and Food Chemistry ARTICLE

Figure 2. GC-MS/SIM chromatogram of mixture OCP (150 μg/L). Peak identification in order of increasing retention time: 1, HCB; 2, lindane
(γ-HCH); 3, aldrin; 4, IS; 5, R-endosulfan; 6, dieldrin; 7, p,p0 -DDE; 8, endrin; 9, β-endosulfan; 10, o,p0 -DDT; 11, methoxychlor.

Table 2. Calibration, LOD, and S/N for the Studied OCP in Spiked Strawberries by GC-ECD, GC-MS/SIM, and GC-MS/MS
GC-MS/SIM GC-MS/MS GC-ECD

equationa equationa equationa

correl LOD correl LOD correl LOD

OCP m b coeff (μg/kg) S/Nb m b coeff (μg/kg) S/Nb m b coeff (μg/kg)

R-HCHc 7.78  104 -6.18  102 0.9984 7.34 2.92 5.78  104 -2.12  102 0.9989 7.02 31.30 4.24  108 4.92  105 0.9963 4.49
β-HCHc 8.11  104 3.52  101 0.9987 6.92 2.38 6.78  104 4.56  101 0.9990 6.52 28.53 3.85  108 5.29  105 0.9980 3.84
δ-HCHc 7.47  104 6.45  102 0.9991 7.56 2.87 5.34  104 5.88  102 0.9991 7.11 26.02 2.96  108 3.60  105 0.9974 4.10
HCBc 1.18  106 -5.13  103 0.9982 9.20 9.15 6.66  105 5.16  103 0.9990 7.65 36.40 6.36  108 -1.46  106 0.9971 7.44
lindanec 2.92  105 5.36  102 0.9982 9.10 2.60 1.91  105 4.28  103 0.9991 7.03 29.27 2.71  108 -6.36  103 0.9974 7.08
aldrinc 2.04  105 2.08  102 0.9988 8.20 20.09 2.03  105 -8.12  102 0.9994 5.99 28.65 3.68  108 7.76  105 0.9987 5.05
R-endosulfand 8.69  104 1.43  103 0.9989 9.90 1.94 1.11  105 1.13  103 0.9981 8.76 300.66 1.74  108 1.17  106 0.9962 8.58
dieldrinc 1.92  105 9.06  102 0.9982 9.60 2.02 1.82  105 -7.39  101 0.9990 8.02 108.31 2.45  108 1.13  106 0.9984 4.48
p,p0 -DDEd 3.17  105 -1.13  103 0.9990 9.50 1.20 2.04  105 -6.55  102 0.9995 7.50 56.21 2.65  108 6.29  105 0.9989 5.61
endrinc 7.33  104 9.33  102 0.9987 8.70 2.85 7.84  104 3.67  102 0.9990 7.80 162.90 2.52  108 1.66  106 0.9973 7.14
β-endosulfand 4.07  105 5.53  102 0.9985 8.60 15.61 3.33  105 -1.37  103 0.9988 7.64 148.41 1.87  108 8.52  105 0.9984 5.61
p,p0 -DDDd 4.77  102 -2.26  104 0.9978 8.91 4.41 5.01  102 -3.18  104 0.9985 6.12 180.88 1.89  108 5.54  104 0.9994 3.43
o,p0 -DDTd 1.51  106 -9.12  103 0.9976 9.90 3.04 4.96  105 -3.05  103 0.9991 6.02 192.95 3.17  108 -6.66  105 0.9980 6.18
methoxychlorc 1.94  105 -6.86  102 0.9983 9.32 5.35 1.54  105 -3.62  102 0.9988 7.86 48.64 2.40  107 -4.96  104 0.9961 8.68
a
m, slope; b, intercept. b S/N, concentration 10 μg/L. c MRL for each pesticide in Europe in strawberries = 10 μg/kg. d MRL for each pesticide in Europe
in strawberries = 50 μg/kg.

was added to dissolve the residue, and then 2 μL of this solution was Linearity. The calibration curves for all of the compounds in matrix
injected onto the gas chromatograph. were obtained by plotting the peak area against the concentration of the
Method Validation. The standard working solutions were used for corresponding calibration standards at seven calibration levels ranging
spiking strawberries for recovery assays and for the matrix-matched calibration between 0.005 and 0.250 mg/kg.
standards in strawberry blank used for quantification of the OCP. The devel- Sensitivity. The limit of detection (LOD) and the limit of quantifica-
oped method was completely validated in a strawberry matrix. A validation tion (LOQ) were determined by considering the slope of the calibration
procedure was carried out to guarantee the feasibility of the method. line and the residual standard deviation of the regression line.25
The performance of the method was evaluated considering the Fortification. Strawberries with no pesticides detected previously
following validation parameters. were used for the fortification experiments. Ten grams of homogenized

7584 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591

Journal of Agricultural and Food Chemistry ARTICLE

Figure 3. GC-MS/SIM (a), GC-MS/MS (b), and GC-ECD (c) chromatograms of OF strawberry sample contaminated with lindane (13 μg/kg) with
spectra comparison of the sample by GC-MS and GC-MS/MS.

sample was spiked prior to the determination procedure by the addition at a flow of 30 mL/min. The system was operated by GCsolution
of a mixture of standard pesticides solution to give 0.030, 0.090, and Shimadzu software.
0.180 mg/kg of each compound. The fortification tests were done in Gas Chromatography-Mass Spectrometry (GC-MS). In
triplicate. Spiking samples were left to stand for 3 h to allow pesticide addition, the real samples were analyzed using a Thermo Trace-Ultra
absorption onto the sample and evaporate the solvent. They were then gas chromatograph coupled to an ion trap mass detector Thermo
prepared according to the determination procedure described above. Polaris, operated in the electron impact ionization (EI) at 70 eV. The
Quantification was performed by using the peak area of the analyte and ion source temperature was 250 °C and the MS transfer temperature,
calculating the concentration by preparing a calibration curve. 250 °C. The system was operated by Xcalibur v 1.3 software. Confirma-
Gas Chromatography-Electron Capture Detector (GC- tion of residues was carried out by GC-MS/SIM and MS/MS using a
ECD). OCP were analyzed using a Shimadzu GC-2010 with an ECD Supelco column fitted with an SLB-5MS (30 m  0.25 mm, 0.25 μm film
apparatus, equipped with a capillary column of 30 m, TRB-5MS (0.25 thickness) column operating in the splitless mode; helium was used as
mm i.d., 0.25 μm film thickness, Teknokroma). The oven temperature carrier gas at a constant flow rate of 1.3 mL/min. The injector was
was programmed starting at 65 °C and held for 2 min, followed by maintained at 240 °C. The oven temperature was programmed starting
increases of 8 °C/min to 160 °C, then 2 °C/min to 235 °C, and then at 40 °C and held for 2 min, followed by increases of 30 °C/min to
15 °C/min to 250 °C. The injection port was at 250 °C in splitless mode, 220 °C, held for 5 min, then 10 °C/min to 270 °C, and held for 1 min.
and the detection was carried out at 300 °C. Helium (Linde Sog
as) was For the identification of pesticides, the retention time, and three ions, the
used as carrier gas at a constant flow rate of 1.3 mL/min, whereas NIST and Wiley pesticide libraries were used. Selected ions (m/z) used
nitrogen (Linde Sog
as, purity g 99.999%) was employed as makeup gas for confirmation are summarized in Table 1. The MS/MS conditions

7585 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591

Journal of Agricultural and Food Chemistry ARTICLE

Figure 4. GC-MS/SIM (a), GC-MS/MS (b), and GC-ECD (c) chromatograms of OF strawberry sample contaminated with β-endosulfan (7 μg/kg)
with spectra comparison of the sample by GC-MS and GC-MS/MS.

were fixed for each compound, trying to select as precursor ion the one that acetonitrile resulted in extracts that were heavily pigmented,
with the highest m/z ratio and abundance (Table 1). containing large amounts of matrix coextractants. At this point,
acetonitrile extracts are unsuitable for further analysis due to the
’ RESULTS AND DISCUSSION high levels of endogenous interferences coextracted with the
pesticides during solvent extraction. The sample matrix coex-
Sample Extraction and Cleanup. Matrix effects can gener- tractants may have a deleterious effect on the capillary columns,
ate serious problems in pesticide residue analysis using GC may interfere with the detection of pesticides at trace levels, and/
coupled with MS due a possible over- or underestimation of or may result in a sample matrix-induced enhancement effect,
analyte concentration when compared with the same analyte in dispersive solid phase extraction is necessary for further cleanup
organic solvent. Acetonitrile was preselected for sample prepara- prior to chromatographic analysis. The cleanup procedure in this
tion in this study as it is a common solvent for the extraction of work employs 150 mg of PSA, which is higher compared with the
OCP from fruit and vegetable samples and was also used in the amount commonly used,10,26,27 and was shown to be effective
QuEChERS procedure. However, a preliminary assay showed for the removal of a variety of matrix interferents such as sugars,
7586 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591
Journal of Agricultural and Food Chemistry ARTICLE

Table 3. Percent Recoveries ( RSD Based on Three Replicate Experiments for the Analysis of OCP by GC-ECD in Strawberries
Produced by OF and IPM

recoveriesa (%)

OF IPM

name 0.030 mg/kg 0.090 mg/kg 0.180 mg/kg 0.030 mg/kg 0.090 mg/kg 0.180 mg/kg

R-HCH 107 ( 8 101 ( 10 98 ( 7 105 ( 5 100 ( 7 97 ( 9

β-HCH 103 ( 10 100 ( 9 99 ( 8 103 ( 7 99 ( 8 99 ( 10
δ-HCH 109 ( 7 99 ( 10 100 ( 10 106 ( 8 101 ( 9 100 ( 7
HCB 120 ( 3 85 ( 1 91 ( 1 112 ( 5 88 ( 6 92 ( 5
lindane 121 ( 1 97 ( 8 110 ( 12 119 ( 10 99 ( 7 107 ( 10
aldrin 101 ( 5 80 ( 2 110 ( 3 99 ( 3 85 ( 10 107 ( 5
R-endosulfan 116 ( 6 97 ( 1 91 ( 2 118 ( 11 100 ( 2 90 ( 10
dieldrin 111 ( 1 111 ( 10 123 ( 7 113 ( 10 108 ( 11 117 ( 9
p,p0 -DDE 51 ( 13 46 ( 11 54 ( 12 59 ( 13 49 ( 9 57 ( 11
endrin 105 ( 3 128 ( 4 124 ( 8 100 ( 8 121 ( 7 118 ( 8
β-endosulfan 122 ( 7 111 ( 8 108 ( 12 115 ( 1 107 ( 10 105 ( 10
p,p0 -DDD 100 ( 4 112 ( 12 128 ( 5 105 ( 3 115 ( 10 121 ( 10
o,p0 -DDT 111 ( 10 90 ( 10 118 ( 7 104 ( 7 97 ( 8 117 ( 11
methoxychlor 94 ( 12 89 ( 12 121 ( 11 99 ( 9 92 ( 9 114 ( 8
a
Mean percent recovery ( RSD of OCP in strawberries at 0.030, 0.090, and 0.180 mg/kg fortification levels (n = 3).

fatty acids, organic acids, and anthocyanin pigments, yielding a Table 4. Pesticides Residues in Portuguese Strawberries from
clear supernatant. OF and IPM in Several Varieties (Siba, Camarosa) and Their
Method Validation: Calibration, Linearity, S/N, and LOD. Concentrations in 2009
Standard mixtures of OCP in n-hexane were initially used to
determine their characteristic parameters such as retention time OF (μg/kg) IPM (μg/kg)
(GC-ECD), mass spectra, and acquisition time segments for Siba Camarosa Camarosa
both GC-MS/SIM and GC-MS/MS. All analyses were per-
lindane 13.00 27.30 14.60
formed by GC-ECD, GC-MS/SIM, and MS/MS mode. Pesti-
β-endosulfan 7.02 <LODa <LOD
cides were identified according to their retention times and a
selected and precursor ions. First, the strawberry extracts were <LOD, below the LOD.
analyzed by GC-ECD. The GC-ECD demonstrated a high per- Considering the previously referred to matrix effects, external
formance for the detection and quantification of OCP (Figure 1) calibrations were performed with pesticide-fortified samples.
compared with GC-MS/SIM (Figure 2). In SIM, the identifica- These samples were previously confirmed for the absence of
tion of pesticide peaks was confirmed by matching retention OCP. Table 2 summarizes the matrix-matched calibration re-
times of standards and by the presence of major ions. In MS sults, along with LOD values and signal-to-noise ratios (S/N) for
analysis, the MS/MS conditions were fixed for each compound, the pesticides studied. Calibration curves were constructed for
trying to select as precursor ion the one with the highest m/z each compound using seven different concentration levels be-
ratio (greater selectivity) and abundance (greater sensitivity). tween 0.005 and 0.250 mg/kg and finding the minimum LOD
Analysis of the targeted pesticides for quantification and identi- expected for the method. An internal standard was employed at
fication purposes from the strawberry extracts indicates the the final sample preparation stage to help control the significant
desired specificity by selecting the precursor-to-product ion losses of analytes during extraction. The calibration plots exhibit
transitions used in GC-MS/MS over the nonspecific qualifier good linearity for pesticides ranging from 0.005 to 0.250 mg/kg.
and target ions used in GC-MS/SIM. Although a cleanup Average coefficients of determination were >0.996. Because the
procedure was used, many of the pesticides were still very difficult LOD is matrix dependent, it is recommended that matrix-
to detect and identify in low concentrations by GC-MS/SIM. matched calibration be performed for quantitative analysis for
Matrix interferences can contribute to the pesticide target and unknown samples in complex matrices such as, in this case,
qualifier ion abundances that affect the qualifier-to-target per- strawberries. Matrix interferences may be noted for some qua-
centage ratios used for identification in GC-MS/SIM. GC-MS/ lifier ions, especially those of low masses, at low concentrations.
SIM can lead to false negatives due to the difficulty of distinguish- Despite these restrictions, the results indicate that the LOD are
ing the peak from the baseline noise. However, demand for the sufficient for strawberry safety purposes.
precursor ion in MS/MS allows confirmation due to better Calibration data were generated from three replicate samples
resolution of the peak. In the analysis by GC-MS/SIM, noise of seven points at linearity range and LOD was determined as in
baseline is high, which does not allow proper identification of the eq 1, where m is the slope of the calibration line and Sx/y the
pesticide found. On the other hand, the specificity of GC-MS/ residual standard deviation of a regression line.25
MS provided transition ions that are only specific to the pesticide 3:3Sx=y
LOD ¼ ð1Þ
even in the presence of the strawberry matrix. m
7587 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591
Journal of Agricultural and Food Chemistry ARTICLE

Figure 5. Percentage of 12 pesticide residues in Portuguese strawberries produced in 2009 and 2010 above and below the MRL (a); percentage of all
OCP studied in Portuguese strawberry samples that were below the MRL, above the MRL, and not detected in two years of study (b).

For all of the compounds assayed, the LOD are below their showed higher peak resolution and low background noise
MRL, as shown in Table 2. The lowest LOD for all pesticides and compared to GC-MS/SIM (Figures 3b and 4b). This could
the chromatogram with better resolution and sensibility were possibly be due to the higher selectivity in the analysis by MS/MS
obtained through ECD. compared to SIM. The confirmation performed in SIM and MS/
Enhancement of S/N Ratio by GC-MS/MS. The specificity MS results in similar mass spectra of sample and standard. Both
of MS/MS results in an improvement of the S/N ratio, allowing GC-MS/SIM and GC-MS/MS were able to detect, quantify, and
for improved sensitivity over MS/SIM and resulting in a higher identify all of the compounds assayed, but the biggest advantage
number of pesticides detected and identified at low levels. In of GC-MS/MS over GC-MS/SIM is its instrument specificity
Table 2, S/N ratios are calculated for both GC-MS (SIM and and sensitivity.
MS/MS) analyses for all OCP through the Xcalibur v 1.3 Recovery Studies. The recovery study was based on the
software, in a strawberry sample contaminated at 10 μg/kg. SANCO guideline28 and European directive 96/46/EC.29
The S/N ratios obtained by GC-SIM/MS mode are considerably Satisfactory recoveries (from 70 to 120%) were obtained with
lower than those obtained by GC-MS/MS. These results are 43% of recovery data ranging from 90 to 105%, 6% of the data
especially important for pesticides with poor detectability and ranging from 70 to 89%, and 33% of the data ranging from 105 to
with the same quantification ions as the interferences of matrix by 120%, with relative standard deviations (RSDs) of <13%. These
GC-SIM/MS. For example, lindane and β-endosulfan show results were obtained from two strawberry samples spiked (n = 3)
similar quantification in the same strawberry sample (Figures 3 at 0.030, 0.090, and 0.180 mg/kg. Table 3 lists the mean
and 4). When the chromatograms obtained from ECD are recoveries (n = 3) and RSD for 14 OCP fortified at 3 levels of
compared, the peak that identifies the pesticide found offers OF and IPM measured by GC-ECD. For each pesticide, the
the best resolution. The baseline GC-ECD has less noise, being absolute recovery value was obtained from the ratio between the
thus considered a good technique for detecting OCP. In the case peak area corresponding to the spiked strawberries and the peak
of GC-MS/MS analysis, the chromatogram for most analytes area of the standard solution corresponding to the same pesticide
7588 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591
Journal of Agricultural and Food Chemistry ARTICLE

Figure 6. Correlation between (a) GC-ECD and GC-MS/MS, (b) GC-ECD and GC-MS/SIM, and (c) GC-MS/SIM and GC-MS/MS methods for
determination of OCP in strawberries.

standards injected directly onto the analytical column. Unsatis- following the sample preparation method described above. The
factory recoveries (>120%) were obtained for the remainder of residues were found in all strawberry samples produced by OF
the pesticides due to the phenomenon commonly known as the and IPM. Pesticide levels encountered in the analyzed samples
“matrix-induced chromatographic response enhancement ef- are shown in Table 4. Lindane was detected in all samples
fect”, contracted to matrix-induced response enhancement. analyzed in 2009 and 2010 at concentrations ranging from 0.06
Some factors may affect sample matrix enhancement such as to 27.30 μg/kg (Table 4). In 2009, β-endosulfan and methoxy-
the nature of the pesticide, the nature of the matrix, the pesticide- chlor were detected in all samples. In only one of these samples
to-matrix ratio, and the GC system.17,30 The performances of (Siba OF 2009) was β-endosulfan detected above the LOD. o,p0 -
recoveries were similar for most pesticides at all three fortifica- DDT and their metabolites were also identified in lower con-
tion concentrations when analyzed by GC-ECD. centrations in all samples below the LOD. Endrin was detected in
Determination in Samples. The strawberry samples from two samples and aldrin in three samples at concentrations
local farmers in central Portugal were sampled and analyzed ranging from 0.13 to 4.94 μg/kg and from 0.65 to 1.38 μg/kg,
7589 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591
Journal of Agricultural and Food Chemistry ARTICLE

respectively. HCB, dieldrin, and R-endosulfan were not detected identification and quantification of pesticides in the sample. The
in any samples. The residues obtained for 2009 are higher than applicability of the method was demonstrated by analysis of real
for 2010. At least one of the OCP was detected in 100% of the samples (strawberries produced by different agricultural meth-
analyzed samples (Figure 5). ods), resulting in good quality control data and thus making
The OCP concentration in strawberry samples was higher possible reliable determination of the targeted pesticides.
than the MRL in 5% of samples in 2009, whereas none of them
were detected in 2010. About 55% of the samples were positive ’ AUTHOR INFORMATION
(<MRL) in 2009 versus 26% in 2010 (Figure 5).
Because all of these pesticides have not been used over the past Corresponding Author
30 years, the data reported indicate that OCP persist in the *Phone þ351-228340500; fax þ351-228321159; e-mail vfd@
environment. This study also shows that the proposed method to isep.ipp.pt).
determine residues of OCP is rapid, simple, and sensitive and Funding Sources
uses smaller amounts of organic solvents, reducing the risk for This research was supported by a Ph.D. grant from FCT
workers and the environment. The results showed that the (Fundac-~ao para a Ci^encia e a Tecnologia - BD/47200/2008).
amounts of pesticide residues detected were similar in samples
produced by IPM and OF. These results also showed that
strawberries were able to accumulate OCP from soils or as a
result of atmospheric deposition of those pesticides. The OF and ’ ACKNOWLEDGMENT
IPM fields have the same type of soil, and they are close enough We acknowledge the Portuguese farmers for providing straw-
to have the same type of environmental exposure. berry samples.
The fields where we collected the OF and IPM strawberries are
located in a large area of cultivation. In both cases, the fields are
surrounded by other fields that may or not be cultivated. The OF ’ REFERENCES
and IPM fields are located near a small village, away from the city. (1) Singh, S. B.; Mukherjee, I.; Maisnam, J.; Kumar, P.; Gopal, M.;
Correlation Studies. The variables from different groups Kulshrestha, G. Determination of pesticide residues in IPM and non-
were compared with the Student t test. The significance of the IPM samples of mango (Mangifera indica). J. Environ. Sci. Health B 2008,
43 (4), 300–306.
differences was tested at the level of p < 0.05. Figure 6 shows plots
(2) Nollet, L. M. L.; Rathore, H. S. Handbook of Pesticides: Methods
of correlation between OCP concentrations detected in straw- of Pesticide Residues Analysis; CRC Press, Taylor & Francis Group:
berry samples. Correlation coefficients were 0.9893 and 0.9892 New York, 2010.
for all pesticides in a and b, respectively, which indicates a good (3) Gonzalez, M.; Miglioranza, K. S. B.; de Moreno, J. E. A.; Moreno,
agreement in results from GC-ECD versus GC-MS/MS and V. J. Organochlorine pesticide residues in leek (Allium porrum) crops
GC-MS/SIM versus GC-MS/MS. A correlation coefficient with grown on untreated soils from an agricultural environment. J. Agric. Food
lower value was obtained for GC-ECD versus GC-MS/SIM with Chem. 2003, 51 (17), 5024–5029.
R2 = 0.9714 (Figure 6c). The slopes of correlation plots are (4) Gonzalez, M.; Miglioranza, K. S. B.; de Moreno, J. E. A.; Moreno,
slightly lower than 1 in a and b and slightly higher (1.1021) in c, V. J. Occurrence and distribution of organochlorine pesticides (OCPs)
indicating that the OCP concentrations detected by GC-ECD in tomato (Lycopersicon esculentum) crops from organic production.
J. Agric. Food Chem. 2003, 51 (5), 1353–1359.
were higher than OCP concentrations obtained by GC-MS/SIM
(5) Hoh, E.; Lehotay, S. J.; Mastovska, K.; Ngo, H. L.; Vetter, W.;
and that the pesticide concentrations obtained by GC-MS/ Pangallo, K. C.; Reddy, C. M. Capabilities of direct sample introduction-
MS were also higher than those obtained by GC-MS/SIM, comprehensive two-dimensional gas chromatography-time-of-flight
respectively. mass spectrometry to analyze organic chemicals of interest in fish oils.
Moreover, statistical analysis (paired t test) shows that there Environ. Sci. Technol. 2009, 43 (9), 3240–3247.
is no significant difference between GC-ECD and GC-MS/MS (6) Saitta, M.; Di Bella, G.; Salvo, F.; Lo Curto, S.; Dugo, G.
(in most results p > 0.05), and in the case of GC-ECD versus GC- Organochlorine pesticide residues in Italian citrus essential oils 1991-
MS/SIM methods a statistical difference has been observed 1996. J. Agric. Food Chem. 2000, 48 (3), 797–801.
(p < 0.05). The concentrations of most OCP did not differ (7) Fernandez-Alvarez, M.; Llompart, M.; Lamas, J. P.; Lores, M.;
between the GC-ECD and GC-MS/MS results, whereas the Garcia-Jares, C.; Cela, R.; Dagnac, T. Development of a matrix solid-
phase dispersion method for the simultaneous determination of pyre-
OCP levels were significantly higher in the GC-ECD results than
throid and organochlorinated pesticides in cattle feed. J. Chromatogr., A
in the GC-MS/SIM. 2009, 1216 (14), 2832–2842.
In conclusion, these pesticides tend to be very persistent, (8) Guan, H.; Brewer, W. E.; Garris, S. T.; Morgan, S. L. Disposable
and this method using QuEChERS sample preparation and GC- pipette extraction for the analysis of pesticides in fruit and vegetables
ECD, GC-MS/SIM, or MS/MS analysis showed a high sensitiv- using gas chromatography/mass spectrometry. J. Chromatogr., A 2010,
ity and confirmatory capacity necessary for the determination of 1217 (12), 1867–1874.
pesticide residues at the levels required in EU MRL for straw- (9) Fernandez-Alvarez, M.; Llompart, M.; Lamas, J. P.; Lores, M.;
berries. The proposed method not only allowed the simulta- Garcia-Jares, C.; Cela, R.; Dagnac, T. Development of a matrix solid-
neous determination and confirmation of a large number of OCP phase dispersion method for the simultaneous determination of pyre-
with good recoveries and low detection limits but was also useful throid and organochlorinated pesticides in cattle feed. J. Chromatogr., A
2009, 1216 (14), 2832–2842.
in routine analysis due to its speed and ease of performance. This
(10) Nguyen, T. D.; Yu, J. E.; Lee, D. M.; Lee, G. H. A multiresidue
work has demonstrated that OCP quantification by GC-ECD method for the determination of 107 pesticides in cabbage and radish
can be a powerful tool for a complex matrix. The results using QuEChERS sample preparation method and gas chromatography
showed that GC-ECD can be used for screening these OCP. mass spectrometry. Food Chem. 2008, 110 (1), 207–213.
However, GC-MS/MS requires only one injection, rather than (11) Guan, H.; Brewer, W. E.; Garris, S. T.; Morgan, S. L. Disposable
multiple injections for multiple instruments for the screening pipette extraction for the analysis of pesticides in fruit and vegetables

7590 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591

Journal of Agricultural and Food Chemistry ARTICLE

using gas chromatography/mass spectrometry. J. Chromatogr., A 2010, (28) Sanco/10684/2009. Method validation and quality control
1217 (12), 1867–1874. procedures for pesticide residues analysis in food and feed; 2009.
(12) Barriada-Pereira, M.; Gonzalez-Castro, M. J.; Muniategui- (29) Commission of the European Communities. Directive 96/46/
Lorenzo, S.; Lopez-Mahia, P.; Prada-Rodriguez, D.; Fernandez- EC amending Council Directive 91/414/EEC concerning the placing of
Fernandez, E. Determination of organochlorine pesticides in horticul- plant protection products on the market; Brussels, Belgium, 1996.
tural samples by microwave assisted extraction followed by GC-ECD. (30) Anastassiades, M.; Mastovska, K.; Lehotay, S. J. Evaluation of
Int. J. Environ. Anal. Chem. 2005, 85 (4-5), 325–333. analyte protectants to improve gas chromatographic analysis of pesti-
(13) Fytianos, K.; Raikos, N.; Theodoridis, G.; Velinova, Z.; Tsoukali, cides. J. Chromatogr., A 2003, 1015 (1-2), 163–184.
H. Solid phase microextraction applied to the analysis of organopho-
sphorus insecticides in fruits. Chemosphere 2006, 65 (11), 2090–2095.
(14) Domingues, V.; Cabral, M.; Alves, A.; Delerue-Matos, C. Use
and reuse of SPE disks for the determination of pyrethroids in water by
GC-ECD. Anal. Lett. 2009, 42, 706-726.
(15) Albero, B.; Sanchez-Brunete, C.; Tadeo, J. L. Multiresidue
determination of pesticides in juice by solid-phase extraction and gas
chromatography-mass spectrometry. Talanta 2005, 66 (4), 917–924.
(16) Cortes-Aguado, S.; Sanchez-Morito, N.; Arrebola, F. J.; Fre-
nich, A. G.; Vidal, J. L. M. Fast screening of pesticide residues in fruit
juice by solid-phase microextraction and gas chromatography-mass
spectrometry. Food Chem. 2008, 107 (3), 1314–1325.
(17) Mansilha, C.; Melo, A.; Rebelo, H.; Ferreira, I. M. P. L. V. O.;
Pinho, O.; Domingues, V.; Pinho, C.; Gameiro, P. Quantification of
endocrine disruptors and pesticides in water by gas chromatography-
tandem mass spectrometry. Method validation using weighted linear
regression schemes. J. Chromatogr., A 2010, 1217, 6681-6691.
(18) Zrostlikova, J.; Lehotay, S. J.; Hajslova, J. Simultaneous analysis
of organophosphorus and organochlorine pesticides in animal fat by gas
chromatography with pulsed flame photometric and micro-electron
capture detectors. J. Sep. Sci. 2002, 25 (8), 527–537.
(19) Frenich, A. G.; Plaza-Bola~nos, P.; Vidal, J. L. M. Comparison of
tandem-in-space and tandem-in-time mass spectrometry in gas chro-
matography determination of pesticides: application to simple and
complex food samples. J. Chromatogr., A 2008, 1203 (2), 229–238.
(20) Wong, J. W.; Zhang, K.; Tech, K.; Hayward, D. G.; Krynitsky,
A. J.; Cassias, I.; Schenck, F. J.; Banerjee, K.; Dasgupta, S.; Brown, D.
Multiresidue pesticide analysis of ginseng powders using acetonitrile- or
acetone-based extraction, solid-phase extraction cleanup, and gas chro-
matography-mass spectrometry/selective ion monitoring (GC-MS/
SIM) or -tandem mass spectrometry (GC-MS/MS). J. Agric. Food
Chem. 2010, 58 (10), 5884–5896.
(21) Vidal, J. L. M.; Arrebola, F. J.; Mateu-Sanchez, M. Multi-residue
method for determination of pesticides in vegetable samples by GC-MS-
MS. Chromatographia 2002, 56 (7-8), 475–481.
(22) Deike, S.; Pallutt, B.; Christen, O. Investigations on the energy
efficiency of organic and integrated farming with specific emphasis on
pesticide use intensity. Eur. J. Agron. 2008, 28 (3), 461–470.
(23) Beltran-Gonzalez, F.; Perez-Lopez, A. J.; Lopez-Nicolas, J. M.;
Carbonell-Barrachina, A. A. Effects of agricultural practices on instru-
mental colour, mineral content, carotenoid composition, and sensory
quality of mandarin orange juice, cv. Hernandina. J. Sci. Food Agric. 2008,
88 (10), 1731–1738.
(24) Anastassiedes, M. QuEChERS, A Mini-Multiresidue Method for
the Analysis of Pesticide Residues in Low-Fat Products; URL: http://
www.quechers.com/docs/quechers_en_oct2005.pdf (accessed June
2010).
(25) The European Agency for the evaluation of Medicinal Products
Human Medicines Evaluation Unit, Validation of analytical procedures:
methodology. In ICH Topic Q 2B; Nov 1996.
(26) Walorczyk, S. Development of a multi-residue method for the
determination of pesticides in cereals and dry animal feed using gas
chromatography-tandem quadrupole mass spectrometry: II. Improve-
ment and extension to new analytes. J. Chromatogr., A 2008, 1208 ((1-2)),
202–214.
(27) Lehotay, S. J.; Son, K. A.; Kwon, H.; Koesukwiwat, U.; Fu,
W. S.; Mastovska, K.; Hoh, E.; Leepipatpiboon, N. Comparison of
QuEChERS sample preparation methods for the analysis of pesticide
residues in fruits and vegetables. J. Chromatogr., A 2010, 1217 (16),
2548–2560.

7591 dx.doi.org/10.1021/jf103899r |J. Agric. Food Chem. 2011, 59, 7582–7591