This article was downloaded by:[Indexing Ind / Jour]

[Indexing Ind / Jour]

On: 18 June 2007
Access Details: [subscription number 772113004]
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Environmental Science

and Health, Part B
Pesticides, Food Contaminants, and Agricultural
Wastes
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713597269
Validation of an efficient method for the determination of
pesticide residues in fruits and vegetables using ethyl
acetate for extraction
To cite this Article: Aysal, Perihan, Ambrus, Árpád, Lehotay, Steven J. and
Cannavan, Andrew , 'Validation of an efficient method for the determination of
pesticide residues in fruits and vegetables using ethyl acetate for extraction', Journal
of Environmental Science and Health, Part B, 42:5, 481 - 490
To link to this article: DOI: 10.1080/19312450701392490
URL: http://dx.doi.org/10.1080/19312450701392490

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction,
re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly
forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents will be
complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be
independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or
arising out of the use of this material.
© Taylor and Francis 2007
 Journal of Environmental Science and Health Part B (2007) 42, 481–490
Copyright
C Taylor & Francis Group, LLC

ISSN: 0360-1234 (Print); 1532-4109 (Online)

Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007

DOI: 10.1080/19312450701392490

Validation of an efficient method for the determination of

pesticide residues in fruits and vegetables using ethyl acetate
for extraction

PERIHAN AYSAL1 , ÁRPÁD AMBRUS2 , STEVEN J. LEHOTAY3 and ANDREW CANNAVAN1

1
Food and Agricultural Organization/International Atomic Energy Agency Agriculture and Biotechnology Laboratory,
Seibersdorf, Austria
2
Hungarian Food Safety Office, Budapest, Hungary
3
US Dept. of Agriculture, Agricultural Research Service, Wyndmoor, PA

In this study, a version of the “quick, easy, cheap, effective, rugged, and safe” (QuEChERS) method was modified to use ethyl acetate
(EtOAc) rather than acetonitrile (MeCN) for extraction in the determination of multiple pesticide residues in fruits and vegetables.
EtOAc is better suited than MeCN for gas chromatographic (GC) analysis with electron capture detection (ECD) and nitrogen-
phosphorus detection (NPD). The method entailed extraction of 30 g chopped sample plus 5 g NaHCO3 and 30 g anhydrous Na2 SO4
with 60 mL EtOAc using a probe blender. After a centrifugation step, removal of residual water and cleanup were performed using
dispersive solid-phase extraction (dispersive-SPE) with MgSO4 and primary secondary amine (PSA) sorbent. 14 C-labeled chlorpyrifos
with liquid scintillation counting was used to assist in optimizing and characterizing the method, and GC-ECD and GC-NPD were
used for analysis of 24 selected pesticides. The method was validated using tomato, apple and frozen green bean matrices spiked at 0.05,
0.5, and 5 mg/kg. For 22 of the analytes, recoveries averaged 93% for all three commodities over the validation range with a relative
standard deviation of 10% (n = 1182). Lower recoveries of dichlorvos were obtained with the method and iprodione determination
was compromised in the green beans by an interfering peak. Typical limits of detection were 0.005–0.01 mg/kg with the method.
Keywords: Multi-residue pesticide analysis; gas chromatography (GC); validation.

Introduction 1960s and 1970s.[2–4] However, due to the increasing cost of

Analytical methods must be available to determine pesticide
residues in crops, feeds, and food commodities for a vari-
ety of purposes, which include regulatory monitoring and
enforcement, import/export certification, risk assessment,
field-application trials, organic food verification, and mar-
keting to consumers. For all of these purposes, the meth-
ods should be robust, give accurate results, meet detection
limit needs, and cover the desired scope of matrices and
analytes.[1]
The most common methods in current use for pesticide
residue monitoring stem from methods developed in the
Address correspondence to Steven J. Lehotay, US Dept. of Agri-
culture, Agricultural Research Service; Eastern Regional Re-
search Center; 600 East Mermaid Lane; Wyndmoor, PA 19038;
USA; E-mail: steven.lehotay@ars.usda.gov
Disclaimer: Mention of brand or firm name does not constitute
an endorsement by the FAO/IAEA or U.S. Department of Agri-
culture above others of a similar nature not mentioned.
Received March 1, 2007.
labor, solvents, equipment, and laboratory space, there is an
urgent need for residue chemists to develop and use more
cost-effective procedures.[5] Moreover, many advances have
been made in residue analysis in recent decades, even in the
traditional case of gas chromatography (GC) coupled to
selective detectors.
In light of the modern prevalence of good laboratory
practices, proficiency testing, and laboratory accreditation,
many monitoring laboratories do not need to follow the
prescribed traditional methods any longer. [6,7] The analyt-
ical method used in many applications may be developed
in-house, taken from the literature, or otherwise obtained
from a third party, provided that defined validation criteria
have been met in the laboratory and proficiency testing re-
sults are acceptable. Validation is performed to verify that
the method is fit for purpose, which means that the desired
commodities and analytes are evaluated to achieve accept-
able recoveries, reproducibilities, and detection limits.
Multi-residue pesticide analysis in fruits and vegetables
typically involves so many possible analyte/matrix pairs
and concentration levels that it is impractical for a method
 482
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007
to be fully validated for each combination. Thus, represen-
tative pesticides and commodities are typically selected to
assess the laboratory performance of the method.[8] The
chemical and physical properties of common pesticides dif-
fer considerably, and this diversity causes challenges in the
development of a universal analytical method for residues,
which usually needs to have the widest scope possible.[9]
The “quick, easy, cheap, effective, rugged, and safe”
(QuEChERS) method was first published by Anastassiades
et al. in 2003 for the monitoring of pesticide residues in fruits
and vegetables.[10] The method uses acetonitrile (MeCN) for
extraction followed by the addition of anhydrous MgSO4
and NaCl to induce partitioning of the MeCN extract from
the water in the sample. The initial extract is then mixed with
primary secondary amine (PSA) sorbent and anhydrous
MgSO4 in a simple approach termed dispersive solid-phase
extraction (dispersive-SPE) cleanup. Dispersive-SPE with
PSA effectively removes many polar matrix components
common in food matrices, such as organic acids and certain
polar pigments. However the researchers originally evalu-
ated the method only with GC-mass spectrometry (MS) for
quantitative and qualitative analysis of GC-amenable pesti-
cides. Follow-up studies demonstrated that the method also
worked well using flame photometric detection (FPD) and
electrolytic conductivity detection (ELCD) in GC and liq-
uid chromatography (LC) with post-column fluorescence
derivatization[11] and liquid chromatography-tandem mass
spectroscopy.[12–16]
Since that time, the QuEChERS method using MeCN for
extraction has been successfully validated in several labora-
tories and in interlaboratory trials. The method (with minor
modification differences) has become an Official Method of
AOAC International and the European Standard Organiza-
tion (CEN).[15,16] However, as previously stated, prescribed
methods are not required in many cases, and the QuECh-
ERS approach is flexible and easily adaptable.
In the original paper,[10] four different extraction solvent
combinations were evaluated, all of which were known to
achieve high recoveries for a wide range of pesticides. Both
MeCN and ethyl acetate (EtOAc) showed a similar degree
of matrix co-extractrives in fruits and vegetables, but MeCN
was chosen in the final method because EtOAc presented
problems with fatty matrices, was not amenable in reversed-
phase LC, and gave lower recoveries for certain pH-
dependent analytes. Also, EtOAc is a stronger partitioning
solvent in dispersive-SPE than MeCN, thus it gave slightly
dirtier extracts than MeCN after cleanup. However in the
case of GC analysis, EtOAc is a better solvent than MeCN
in that it is less polar, provides a smaller liquid-to-gas ex-
pansion volume, and greater stability to certain pesticides
(e.g. chlorothalonil). [17] Other recent papers also have com-
pared the use of different solvents, including EtOAc and/or
MeCN in pesticide residue analysis, essentially showing ad-
vantages and disadvantages in each instance.[18,19]
One critical advantage in the use of EtOAc instead of
MeCN is that the nitrogen-phosphorus detector (NPD),
Aysal et al.
which is a common GC selective detector for organophos-
phorus (OP) and GC-amenable organonitrogen pesticides,
is not usable with MeCN as the injection solvent unless
the instrument is equipped with a solvent bypass valve for
the detector. The MeCN solvent front produces a high re-
sponse in the sensitive NPD, and re-equilibration in ioniza-
tion would only be reached after many hours, depending
on the condition of the detector’s bead.
The main purpose of this study was to modify the
QuEChERS method by using EtOAc for the extraction,
thus permitting analysis by GC-NPD for common OPs
and GC with electron-capture detection (ECD) for rep-
resentatives of various types of pesticides. This combina-
tion of detectors constitutes a popular instrumental con-
figuration to monitor pesticides in laboratories that are not
equipped with GC-MS and LC-MS/MS. Also, we sought
to identify critical steps of the procedure by using a radio-
tracer technique, and to validate the method performance in
the 0.05-5 mg/kg concentration range for a set of 24 GC-
amenable pesticides in representative fruit and vegetable
commodities.
Materials and Methods
Apparatus
The samples were analyzed with Agilent 5890 and 6890 GCs
(Little Falls, DE; USA) both using HP-5 capillary columns
of 30 m, 0.25 mm i.d., and 0.25 µm film thickness. The
GC-NPD (using the 6890 GC) was equipped with a pro-
grammable temperature vaporization (PTV) injector, which
was used in the splitless mode. Helium was used as carrier
gas as at a constant flow rate of 2 mL/min. The NPD tem-
perature was 300◦ C and gas flows consisted of 3 mL/min
for H2 , 60 mL/min for air and 34 mL/min of make-up gas.
For the GC-ECD system, on-column injection was made
at 73◦ C, the He flow rate was 2 mL/min and the ECD tem-
perature was 300◦ C and gas flows of 42 mL/min of He with
constant pressure of 46 psi. Injection volume was 1 µL in
both cases, and the oven-temperature programs were the
same: 70◦ C for 1 min, ramped to 160◦ C at 20◦ C/min, fol-
lowed by a 4◦ C/min ramp to 270◦ C and held for 5 min.
Other equipment used in the study included a Stephan
UM 5 universal chopper to comminute fruit and vegetable
samples; an UltraTurrax with T25 head to homogenize
samples during extraction; a Beckman LS 6000 TA liquid
scintillation counter (LSC) to measure radioactivity in the
samples fortified with radiolabeled chlorpyrifos; a Sigma 4
K15 centrifuge to centrifuge the 250 mL extraction bottles
and 20 mL conical glass tubes; a Labinco L46 vortexer to
shake the dispersive-SPE tubes; Sartorius CP 225D ana-
lytical and top-loading balances to weigh standards, salts
and samples; and a Zymark Turbo Vap LV to concentrate
extracts when necessary.
For extraction, 250 mL high-density polyethylene cen-
trifuge bottles (Merck 525 H 2361) were employed, and
 Method for analysis of pesticide residues in fruits and vegetables 483
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007

20 mL glass conical Zymark tubes were used for dispersive-
SPE cleanup. For LSC, 20 mL polyethylene vials were
utilized.
Chemicals
The 24 pesticide reference standards consisted of azinphos-
methyl, chlorfenvinfos, diazinon, dichlorvos, dimethoate,
coumaphos, cyfluthrin, α-endosulfan, EPTC, fena-
rimol, fenitrothion, fenpropathrin, fenvalerate, folpet,
heptenophos, iprodione, lindane, methidathion, mevin-
phos, pirimicarb, propachlor, propiconazole, triazophos,
and vinclozolin, which were all near-100% purity and
were obtained from Dr. Ehrenstorfer GmbH (Augsburg,
Germany). Full IUPAC names for these pesticides are given
in Table 5. Stock solutions of 1 mg/mL were prepared in
85:15 (v:v) acetone:isooctane, and working solutions con-
taining the 24 pesticides at 5, 50, and 500 ng/µL for the 3
fortification levels were prepared in EtOAc. A mixture of
cold chlorpyrifos (Dr. Ehrenstorfer) and 14 C-chlorpyrifos
(Dow Chemicals, USA) to make a total chlorpyrifos con-
centration of 150 ng/µL was prepared in EtOAc. 14 C-
chlorpyrifos recoveries were measured by LSC. The spike
contained 120,000 dpm activity of 14 C-chlorpyrifos yield-
ing ≈2,000 dpm/mL activity in the 60 mL EtOAc extract.
As a quality control (QC) check of the performance of the
GCs, a 15 ng/µL bromophos-methyl solution (QC-std) was
also prepared in EtOAc and added to all final extracts prior
to injection.
All organic solvents used in the study were pesticide
or HPLC grade. High purity anhydrous Na2 SO4 and
NaHCO3 were obtained from Merck, and anhydrous
MgSO4 (≥ 98% purity) was from Fluka. The MgSO4 was
baked at 500◦ C for 5 h in a furnace to remove phthalates.
Primary secondary amine sorbent of 40 µm particle size
was obtained from Varian (Harbor City, CA; USA). Perkin
Elmer Ultima gold liquid scintillator solution was used for
the radioassay in LSC.
Blank samples of tomatoes, apples, and frozen green
beans were purchased from a local store. These were used
in fortification experiments and to prepare matrix blanks
for matrix-matched calibration standards.
Fruit and vegetable samples of 1–2 kg were comminuted
using the Stephan chopper and homogenous samples were
stored in plastic bags in a deep-freezer, or sub-samples were
taken immediately for extraction. A number of 100 and
40 mL vials containing 30 ± 0.1 g anhydrous Na2 SO4 and
5 ± 0.1 g NaHCO3 were prepared separately in advance
and were stored capped at room temperature until needed in
experiments. For dispersive-SPE, 0.25 ± 0.01 g PSA sorbent
and 1.5 ± 0.1 g anhydrous MgSO4 were weighed and mixed
into 20 mL conical glass centrifuge tubes and stored capped
at room temperature until needed.
Extraction and cleanup
In validation experiments, tomatoes, apples and frozen
green beans were used as representative matrices. Each com-
modity was spiked with the appropriate 24-pesticide solu-
tion at 0.05, 0.5 and 5 mg/kg; 6 replicates at each level (i.e.
3 matrices at 3 levels/matrix and 6 replicates/level = 54
spiked samples). For each batch of samples, a matrix blank
and reagent blank (27 mL of deionized water) were also
analyzed.
Aliquots (30 ± 0.1 g) of previously comminuted sam-
ple were weighed into each centrifuge bottle. The sam-
ples were fortified with the appropriate pesticide solution
(300 µL) to yield 5, 0.50, or 0.05 mg/kg concentration,
and 14 C-chlorpyrifos (100 µL of 150 ng/µL, equivalent to
0.5 mg/kg in matrix) was added. The fortified samples were
allowed to stand for 20 min for the pesticides to interact with
the matrix and some of the solvent to evaporate. NaHCO3
(5 g) and anhydrous Na2 SO4 (30 g) were mixed with the
samples. EtOAc (60 mL) was added, and immediately, each
sample was extracted with the probe blender for 1 min. The
tubes were centrifuged (3 min, 2,500 rpm). Three portions
of 1 mL were transferred from each extract supernatant
to scintillation vials for radio-assay. Scintillation cocktail
(12 mL) was added and the activity measured on a LSC to
determine the efficiency and repeatability of the extraction
step. Further aliquots (10 mL) of the EtOAc extract were
Table 1. 14 C-Chlorpyrifos recoveries (Q) and precision (RSD) as
determined by liquid scintillation counting in the different matri-
ces at each step in the method and overall.

Preliminary steps Extractiona Cleanupb Overallc

Quantification was performed using external calibration Q RSD Q RSD Q RSD

using matrix-matched standards, which entailed prepar- Matrix (%) (%) (%) (%) (%) (%)
ing blank extracts for use as the solvent in calibration Tomato 89 3.1 100 1.2 89 2.9
solutions.[20] The standard solutions were prepared at 2, Apple 94 2.2 101 1.8 95 2.8
12.5, 25, 50, 100 and 125 pg/µL concentrations for ECD Frozen 87 2.7 100 2.3 87 2.8
and 125, 250, 500, 1000, 1250 and 2500 pg/µL for NPD. green bean
The equivalent pesticide concentrations in the samples can a
be calculated from the 0.25 mg/µL and 2.5 mg/µL sample Triplicate aliquots and measurements of 18 samples.
b
Duplicate aliquots and measurements of 18 samples; recovery for
equivalent concentrations for ECD and NPD, respectively. cleanup step only.
All injected solutions also contained bromophos-methyl as b
Combination of triplicate extraction and duplicate cleanup measure-
the QC-std at 150 pg/µL. ments for 18 samples.
 484 Aysal et al.
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007

transferred to the dispersive-SPE cleanup tubes and vor-
texed (45 s). The tubes were centrifuged (1 min, 1,900 rpm)
and 2 portions of 1 mL were taken from each extract su-
pernatant for radio-assay, as described above, to determine
the efficiency and repeatability of the clean-up step. The re-
maining extracts in the dispersive-SPE tubes were taken for
GC analysis.
GC analysis
For GC/ECD, 500 µL aliquots from the low-spike
(0.05 mg/kg) and mid-spike (0.5 mg/kg) extracts were
taken and diluted with 500 µL EtOAc in the GC vials;
and 50 µL aliquots from the high-spike (5 mg/kg) ex-
tracts were diluted with 950 µL of EtOAc in the GC vials.
For GC/NPD, the low-spike extracts were concentrated
5-fold using the Turbovap, while the mid- and high-spike
extracts were injected directly. The QC-std (10 µL) contain-
ing 15 ng/µL bromophos-methyl was added to all extracts
prior to GC injection. We checked bromophos-methyl re-
sponse in each chromatogram to ensure that the injection
process had no problems. For quantitation, we used only
external standard calculation. We used 14 C-chlorpyrifos to

Fig. 1. Representative GC-NPD and GC-ECD chromatograms of pesticides for a) blank apple matrix; b) matrix-matched standards
at 250 pg/µL (0.5 mg/kg) in NPD and 12.5 pg/µL (0.05 mg/kg) in ECD; c) apple fortified at 0.5 mg/kg in NPD and 0.05 mg/kg in
ECD. Chromatograms for tomato and green bean (not shown) were similar. See Table 5 for names of the labeled peaks.
 Method for analysis of pesticide residues in fruits and vegetables
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007
check potential losses after each step of the procedure and
separately measured its quantity in the extracts using only
LSC.
Results and Discussion
Sample size and extraction
The first step in the laboratory when conducting an analy-
sis is to comminute and homogenize the sample collected in
the field so that a reasonable and representative sub-sample
Fig. 2. Individual pesticide recoveries at different levels in each matrix (n = 6 at each level).
485
can be extracted. To minimize wasted time, effort, cost, and
reagents in the analytical method, the smallest sub-sample
possible should be taken that achieves accurate results for
the original sample. For this study, a 30 g comminuted sub-
sample of a 1–2 kg original sample was chosen for anal-
ysis. This decision was based on results using the same
chopper in the same laboratory, which demonstrated that a
30 g sub-sample gave representative results within generally
≤8% relative error of the mean concentration of the orig-
inal sample.[21,22] The subsampling constant (Ks ) showed
 486 Aysal et al.
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007

Table 2. Overview of method validation characteristics for the different matrices (n = 132)

Accuracy Precision

Average Codex acceptable Repeatability of recoveries Codex

Matrix Spiking level (mg/kg) recovery (%) range RSD (% RSD) acceptable range

Tomato 0.05 91 70–120 10 20

0.5 94 70–110 7 15
5 97 70–110 8 10
Apple 0.05 95 70–120 11 20
0.5 95 70–110 5 15
5 98 70–110 4 10
Frozen Green Bean 0.05 88 70–120 13 20
0.5 86 70–110 9 15
5 90 70–110 10 10

that a 10 g sub-sample also would achieve representative
results with higher relative uncertainty, but this depends on
the matrix and care exercised during comminution.[5,21,22]
We felt that a 30 g sample would ensure that a representative
sample would be analyzed independent of these factors.
Efficiency of extraction by shaking also requires that
the sample be finely chopped to increase accessibility of the
solvent to the pesticides within the sample. Again, to remove
this issue as a factor, we decided to aid extraction by using
a probe blender to guarantee that the solvent would reach
any pesticides encased in the sample.
One of the difficulties in analyzing fruits and vegeta-
bles for pesticide residues is that the many different types
of samples have different pH, which can affect recover-
ies of pH-susceptible pesticides and their stability in the
extracts.[10,13–15] Adjustment of the pH of the different food
samples through buffering or other means serves to avoid
this problem, especially for pesticides such as chlorothalonil
and captan.[17] Thus, 5 g NaHCO3 was added during the
method for the 30 g sample to give a consistent pH during
extraction independent of the original sample pH.
For extraction, EtOAc has an advantage of partial immis-
cibility with water, which makes the addition of other non-
polar solvents to separate water from the extract unneces-
sary. To increase recoveries of polar compounds, sodium
sulfate (Na2 SO4 ) is usually added in multi-residue method
procedures.[18,23,24] Morever, EtOAc can form an emulsion
with water in the sample unless a drying agent is added to
minimize the amount of free water to interact with the sol-
vent. The drying agent also serves as a dispersant to increase
surface area for sample exposure to the solvent. In this
study, the decision was made to mix the sample 1:1 (w:w)
with anhydrous Na2 SO4 and use a 2:1 (v:w) EtOAc:sample
ratio because this had been evaluated previously to achieve
high recoveries.[18,21,22] Subsequently, it has been demon-
strated that shaking of a 10–15 g finely chopped sample
using a 1:1 (w:v) sample:solvent ratio (with MeCN and
MgSO4 ) achieved equal results as blending for a variety of
fortified and incurred pesticides over a wide concentration
range in many different matrices.[10–16]
Recovery of 14 C-chlorpyrifos
When possible, the use of isotopic tracers is an exceptional
approach to follow the pathway of a substance through a
chemical, physical, or biological system. The unique ad-
vantage of radioisotopes is that their behavior in a system
is usually identical with that of their stable counterpart,
and they can be identified easily with very high sensitivity
by their characteristic radiation, even in unclean extracts.
[25–28]
We took advantage of the IAEA laboratory capabil-
ity to use radioactive tracers and measured the recovery of

Table 3. Overview of method validation characteristics of the method at different levels independent of the matrix. Analysis uncertainty
is expressed as CVA (n = 396)

Accuracy Precision
Average Codex Reproducibility of Codex
Spiking level recovery (%) acceptable range recoveries CVA (%)∗ acceptable ranges

0.05 mg/kg 92 70–120 12 32

0.5 mg/kg 92 70–110 8 23
5 mg/kg 95 70–110 8 16
Overall 93 10
∗
coefficient of variation for the analysis (excluding contribution from sample heterogeneity).
 Method for analysis of pesticide residues in fruits and vegetables 487
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007

the 14 C-chlorpyrifos after the two main steps in the method Table 4. Calculated LOD of the 22 pesticides yielding accept-
(extraction and dispersive-SPE). As described in Materials able recoveries in all matrices as determined from the weighted
and Methods, 14 C-chlorpyrifos was applied at a consistent regression calibrations of the analytes
concentration to all samples. Table 1 shows the recoveries of LOD (mg/kg)
14
C-chlorpyrifos during the extraction and dispersive-SPE
clean-up steps, which gave overall results of 89, 95 and 87% Spike Level Green
in tomato, apple and green bean, respectively, with very Detector Pesticide (mg/kg) Tomato Apple Bean
good precision of 1–3% relative standard deviation (RSD). NPD EPTC 0.05 0.020 0.006 0.015
The dispersive-SPE step had no losses of the chlorpyrifos in 0.5–5 0.010 0.016 0.012
any of the matrices. Also, there is no reason that tomato, ap- Mevinphos 0.05 0.020 0.015 0.010
ple, or green bean would cause real differences in the overall 0.5–5 0.010 0.012 0.018
method or LSC measurements, thus it is safe to conclude Heptenophos 0.05 0.018 0.013 0.011
that the chlorpyrifos recovery averaged 90% among the dif- 0.5–5 0.013 0.013 0.003
ferent matrices. Folpet 0.05 0.013 0.009 0.004
0.5–5 0.008 0.008 0.007
Dimethoate 0.05 0.022 0.020 0.015
0.5–5 0.010 0.005 0.004
Recoveries of fortified pesticides Diazinon 0.05 0.015 0.014 0.012
0.5–5 0.010 0.003 0.004
Figure 1 presents the GC-ECD and GC-NPD chro-
Pirimicarb 0.05 0.014 0.010 0.006
matograms of the analyses for calibration standards of the 0.5–5 0.006 0.011 0.005
stated concentrations in matrix and matrix blanks for ap- Fenitrothion 0.05 0.017 0.018 0.015
ple. The individual peaks, with their retention times, are 0.5–5 0.008 0.004 0.002
identified in Table 5. Chromatograms for tomato and green Chlorfenvinphos 0.05 0.011 0.015 0.013
bean (not shown) were similar. As shown in the apple chro- 0.5–5 0.008 0.004 0.007
matograms, peak shapes for the pesticides were good and Methidathion 0.05 0.014 0.022 0.012
the matrices were quite clean. Only green bean gave an inter- 0.5–5 0.009 0.003 0.006
ference for iprodione, and otherwise, all analytes could be Triazophos 0.05 0.014 0.015 0.011
integrated in all matrices at concentrations <0.05 mg/kg. 0.5–5 0.008 0.005 0.006
The consistency of the QC-std, 0.5 mg/kg equivalent of Fenpropathrin 0.05 0.024 0.016 0.009
0.5–5 0.004 0.012 0.008
bromophos-methyl, in the injection sequences ranged from
Azinphos-methyl 0.05 0.023 0.018 0.014
6% to 18% RSD among the replicates for ECD and NPD 0.5–5 0.009 0.014 0.023
and different matrices. Fenarimol 0.05 0.024 0.017 0.002
Many of the pesticides contained nitrogen, phosphorus, 0.5–5 0.005 0.011 0.007
and chlorine, thus both ECD and NPD could be used Coumaphos 0.05 0.017 0.010 0.008
for their analysis. In the final results, we used ECD for 0.5–5 0.006 0.014 0.010
propachlor, lindane, vinclozolin, α-endosulfan, propicona- ECD Fenvalerate 0.05 0.001 0.001 0.003
zole, cyfluthrin, and fenvalerate and NPD for all of the other 0.5–5 0.001 0.001 0.016
analytes. Cyfluthrin 0.05 0.016 0.014 0.003
Individual pesticide recoveries of the replicates for the 0.5–5 0.016 0.014 0.028
different levels and matrices were calculated using weighted Propiconazole 0.05 0.004 0.002 0.005
0.5–5 0.004 0.002 0.020
linear regression templates. To check for suspected outliers,
α-Endosulfan 0.05 0.003 0.002 0.001
the Dixon Test was performed and any data rejected as 0.5–5 0.003 0.002 0.028
outliers were excluded from calculations. In all, only six Vinclozolin 0.05 0.002 0.004 0.002
outliers were removed from the data set of ≈1,200 results. 0.5–5 0.002 0.004 0.020
Figure 2 shows the recoveries for all spiked pesticides Lindane 0.05 0.003 0.003 0.001
at each level in each matrix. As shown in Figure 2A, all 0.5–5 0.003 0.003 0.032
pesticide recoveries fell within the 70-110% recovery range Propachlor 0.05 0.003 0.002 0.003
at all spiking levels in tomato, with nearly all recoveries at 0.5–5 0.003 0.002 0.032
100%. In the case of apple (Fig. 2B), most recoveries again
LOD: limit of detection.
were ≈100%, but dichlorvos at the 0.05 and 0.5 mg/kg
concentrations dipped just below 70% average recovery. which is the most likely explanation for this lower result.
The recoveries for dichlorvos were even lower in the frozen In the case of dichlorvos, the low recovery may have been
green bean at all fortification levels, and iprodione at 0.5 connected with the use of the evaporation step prior to GC-
and 5 mg/kg levels gave average recovery just below 70% NPD analysis, since dichlorvos is the most volatile pesticide
(Fig. 2C). The aforementioned interferant affected the in- analyzed and is susceptible to volatilization losses during
tegration of iprodione in the green bean chromatograms, solvent evaporation. Dichlorvos recoveries were also lower
 488 Aysal et al.
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007

Table 5. Identification and retention times (tR ) of peaks in the chromatograms presented in Figure 1

Peak # tR, min Pesticide IUPAC Name

1 5.59 Dichlorvos 2,2-dichlorovinyl dimethyl phosphate

2 6.45 EPTC S-ethyl dipropylthiocarbamate
3 7.10 Mevinphos 2-methoxycarbonyl-1-methylvinyl dimethyl phosphate
4 8.80 Heptenophos 7-chlorobicyclo[3.2.0]hepta-2,6-dien-6-yl dimethyl phosphate
5 9.28 Folpet N-(trichloromethylthio)phthalimide
6 11.00 Dimethoate O, O-dimethyl S-methylcarbamoylmethyl phosphorodithioate
7 12.24 Diazinon O, O-diethyl O-2-isopropyl-6-methylpyrimidin-4-yl phosphorothioate
8 13.13 Pirimicarb 2-dimethylamino-5,6-dimethylpyrimidin-4-yl dimethylcarbamate
9 14.88 Fenitrothion O, O-dimethyl O-4-nitro-m-tolyl phosphorothioate
14
C-Chl. 15.71-npd Chlorpyrifos O, O-diethyl O−3,5,6-trichloro-2-pyridyl phosphorothioate
12.66-ecd
QCstd 16.34-npd Bromophos-methyl O-4-bromo-2,5-dichlorophenyl O, O-dimethyl phosphorothioate
13.41-ecd
10 17.36 Chlorfenvinphos 2-chloro-1-(2,4-dichlorophenyl)vinyl diethyl phosphate
11 17.93 Methidathion S-2,3-dihydro-5-methoxy-2-oxo-1,3,4-thiadiazol-3-ylmethyl O, O-dimethyl
phosphorodithioate
12 21.97 Triazophos O, O-diethyl O-1-phenyl-1H-1,2,4-triazol-3-yl phosphorothioate
13 24.52 Iprodione 3-(3,5-dichlorophenyl)-N-isopropyl-2,4-dioxoimidazolidine-1-carboxamide
14 25.36 Fenpropathrin (RS)-α-cyano-3-phenoxybenzyl 2,2,3,3-tetramethylcyclopropane carboxylate
15 26.25 Azinphos methyl S-(3,4-dihydro-4-oxobenzo[d]-[1,2,3]-triazi-3-ylmethyl) O, O-dimethyl
phosphorodithioate
16 27.40 Fenarimol (±)-2,4’-dichloro-α-(pyrimidin-5-yl)benzhydryl alcohol
17 29.45 Coumaphos O-3-chloro-4-methyl-2-oxo-2H-chromen-7-yl O, O-diethylphosphorothioate
18 8.14 Propachlor 2-chloro-N-isopropylacetanilide
19 9.93 Lindane 1,2,3,4,5,6-hexachlorocyclohexane
20 12.23 Vinclozolin (RS)-3-(3,5-dichlorophenyl)-5-methyl-5-vinyl1,3-oxazolidine-2,4-dione
21 14.43 α-Endosulfan (1,4,5,6,7,7-hexachloro-8,9,10-trinorborn-5-en2,3-ylene bismethylene)sulfite
22 22.40 Propiconazole (±)-1-[2-(2,4-dichlorophenyl)-4-propyl-1,3-dioxolan-2-ylmethyl]-1H-1,2,4-
triazole
23 25.23 Cyfluthrin (RS)-α-cyano-4-fluoro-3-phenoxybenzyl (1RS, 3RS;1RS,3SR)-3-=(2,2-
dichlorovinyl)-2,2-dimethylcyclopropanecarboxylate
24 30.58 Fenvalerate (RS)-α-cyano-3-phenoxybenzyl (RS)-2-)4-chlorophenyl)-3-methylbutyrate

than those of most of the other analytes in the tomato and
apple. However, recoveries of most of the compounds were
lower in green beans than in the other matrices, so it is prob-
able that the extraction conditions were more optimal for
the tomato and apple matrices than for green beans, and a
combination of these factors resulted in the observed low
recovery for dichlorvos in green beans.
Otherwise, the differences between recoveries for the 22
other analytes in all matrices and levels were not found to
be statistically significant. This can be observed from the
averaged results in Tables 2 and 3 for those 22 pesticides
from the combined results for each matrix and/or level.
For all 22 pesticides, three matrices, and three levels, overall
recovery of the method was 93% with RSD of 10% (n =
1,182). As would be expected, this matches the recovery
determined for 14 C-chlorpyrifos by the LSC method.
Limit of detection (LOD)
LOD is another important parameter to be determined in
method validation experiments. One simple way to estimate
LOD is use of the calibration curves in matrix. The standard
deviations of relative y (response) residuals (Srr), which is
a decisive parameter in internal quality control, should be
≤0.1.[26–29] This was the case for all analytes in the study.
LOD values for analytes at different fortification levels are
summarized in Table 4. Typical LOD were 0.01 mg/kg us-
ing GC-NPD and 0.005 mg/kg using GC-ECD. All LOD
were <0.05 mg/kg.
Conclusions
The IAEA-modified QuEChERS method using EtOAc
for extraction and GC-NPD and GC-ECD for multi-
residue analysis was successfully validated for 22 of the
24 pesticides in all three matrices and all three spiking
levels. Only dichlorvos in green beans gave unacceptable
recoveries, even though acceptable results were obtained
in apple and tomato matrices. Iprodione also gave in-
termittent difficulties in the green beans due to an in-
terferant. LODs were <0.05 mg/kg, and were typically
 Method for analysis of pesticide residues in fruits and vegetables
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007
0.005-0.01 mg/kg for the pesticides in the different
matrices.
This method could be very useful for the analysis of pesti-
cide residues in non-fatty foods, especially in laboratories in
developing countries which are not equipped with expen-
sive MS instrumentation. There is also a cost advantage
in the use of EtOAc rather than MeCN as the extractant
because EtOAc is typically 30-50% cheaper than MeCN,
which may be an important factor especially in developing
countries. The method can also be used for any other GC
selective detector and for GC-MS, making it applicable for
many purposes in a wide array of laboratories.
Acknowledgments
Steven J. Lehotay’s visit to the IAEA laboratory when
this work was initiated was supported by the Co-operative
Research Program: Biological Resource Management for
Sustainable Agricultural Systems, administered by the Or-
ganization of Economic Cooperation and Development
(OECD). The generous provision of 14 C-labeled chlorpyri-
fos by Dow Chemicals is gratefully acknowledged. Thanks
also to Ms. Perihan Yolci for her assistance during the
experiments.
References
[1] Krynitsky, A.J.; Lehotay, S.J. Overview of analytical technologies
available to regulatory laboratories for the determination of pesti-
cide residues. In Handbook of Residue Analytical Methods for Agro-
chemicals; Lee, P.W., Ed.; Wiley & Sons: Chichester, England, 2002;
753–786.
[2] Mills, P.A.; Onley, J.H.; Guither, R.A. Rapid method for chlorinated.
pesticide residues in nonfatty foods. J. Assoc. Off. Anal. Chem. 1963,
46, 186–191.
[3] Luke, M.A.; Froberg, J.E.; Masumoto, H.T. Extraction and cleanup
of organochlorine, organophosphate, organonitrogen, and hydro-
carbon pesticides in produce for determination by gas-liquid chro-
matography. J. Assoc. Off. Anal. Chem. 1975, 58(5), 1020–1026.
[4] Specht, W.; Tilkes, M. Gas-chromatographic determination of pes-
ticide residues after clean-up by gel-permeation chromatography
and mini-silica gel-column chromatography. III. Communication.
Clean-up of foods and feeds of vegetable and animal origin for mul-
tiresidue analysis of fat-soluble and water-soluble pesticides. Frese-
nius Z. Anal. Chem. 1980, 301(4), 300–307.
[5] Hemingway, R.J.; Aharonson, N; Greve, P.A.; Roberts, T.R.; Their,
H.P. Improved cost-effective approaches to pesticide residues anal-
ysis. Pure Appl. Chem. 1984, 56(8), 1131–1152.
[6] Hill, A.R.; Reynolds, S.L. Guidelines for in-house validation of an-
alytical methods for pesticide residues in food and animal feeds.
Analyst 1999, 124(6), 953–958.
[7] Fajgelj, A., Ambrus, Á., Eds. Principles and Practices of Method
Validation; Royal Society of Chemistry: Cambridge, England, 2000.
[8] Food and Agriculture Organization/World Health Organization
FAO/WHO Guidelines on Good Laboratory Practice in Residue
Analysis, CAC/GL 40-1993; Rome, Italy, 2003.
[9] Alder, L.; Greulich, K.; Kempe, G.; Vieth, B. Residue analysis of
500 high priority pesticides: better by GC-MS or LC-MS? Mass
Spectrom. Rev. 2006, 25(6), 838–865.
[10] Anastassiades, M.; Lehotay, S.J.; Štajnbaher, D.; Schenck, F.J.
Fast and easy multiresidue method employing acetonitrile extrac-
489
tion/partitioning and “dispersive solid phase extraction” for the
determination of pesticide residues in produce. J. AOAC Int. 2003,
86(2), 412–431.
[11] Schenck, F.J.; Hobbs, J.E. Evaluation of the quick, easy, cheap,
effective, rugged, and safe (QuEChERS) approach to pesticide
residue analysis. Bull. Environ. Contam. Toxcol. 2004, 73(1), 24–
30.
[12] Lehotay, S.J.; de Kok, A.; Hiemstra, M.; van Bodegraven, P. Vali-
dation of a fast and easy method for the determination of 229 pesti-
cides in fruits and vegetables using gas and liquid chromatography
and mass spectrometric detection. J. AOAC Int. 2005, 88(2), 595–
614.
[13] Lehotay, S.J.; Mastovska, K.; Lightfield, A.R. Use of buffering and
other means to improve results of problematic pesticides in a fast and
easy method for residue analysis of fruits and vegetables. J. AOAC
Int. 2005, 88(2), 615–629.
[14] Lehotay, S.J.; Mastovska, K.; Yun, S.-J. Evaluation of two fast and
easy methods for pesticide residue analysis in fatty food matrices. J.
AOAC Int.2005, 88(2), 630–638.
[15] Lehotay, S.J. Determination of Pesticide Residues in Foods by Ace-
tonitrile Extraction and Partitioning with Magnesium Sulfate: Col-
laborative Study. J. AOAC Int. 2007, 90, 485–520.
[16] Anastassiades, M. http://www.quechers.com/, viewed 21 March
2007 Chemisches und VeterinäruntersuchungSant Stuttgart,
Germany.
[17] Mastovska, K.; Lehotay, S.J. Evaluation of common organic sol-
vents for gas chromatographic analysis and stability of multiclass
pesticide residues. J. Chromatogr. A 2004, 1040(2), 259–272.
[18] Mol, H.G.J.; van Dam, R.C.J.; Steijger, O.M. Determination of
polar organophosphorus pesticides in vegetables and fruits us-
ing liquid chromatography with tandem mass spectrometry: se-
lection of extraction solvent. J. Chromatogr. A 2003, 1015, 119–
127.
[19] Diez, C.; Traag, W.A.; Zommer, P.; Marinero, P.; Atienza, J. Com-
parison of an acetonitrile extraction/partitioning and “dispersive
solid-phase extraction” method with classical multi-residue meth-
ods for the extraction of herbicide residues in barley samples. J.
Chromatogr. A 2006, 1131, 11–23.
[20] Soboleva, E.; Rathor, N.; Mageto, A.; Ámbrus, A. Estimation of sig-
nificance of matrix-induced chromatographic effects. In Principles
and Practices of Method Validation; Fajgelj, A., Ambrus, Á., Eds.;
Royal Society of Chemistry: Cambridge, England, 2000; 138–156.
[21] Maestroni, B.; Ghods, A.; El-Bidaoui, M.; Rathor, N.; Ton, T.; Am-
brus, Á. Testing the efficiency and uncertainty of sample processing
using 14C-labelled chlorpyrifos: part I. description of the method-
ology. In Principles and Practices of Method Validation; Fajgelj, A.,
Ambrus, Á., Eds.; The Royal Society of Chemistry: Cambridge, Eng-
land, 2000; 49–58.
[22] Maestroni, B.; Ghods, A.; El-Bidaoui, M.; Rathor, N.; Jarju, O.P.;
Ton, T.; Ambrus, Á. Testing the efficiency and uncertainty of sample
processing using 14C-labelled chlorpyrifos: part II. In Principles and
Practices of Method Validation; Fajgelj, A., Ambrus, Á., Eds.; The
Royal Society of Chemistry: Cambridge, England, 2000; 59–74.
[23] Holland, P.T.; Boyd, A.J.; Malcolm, C.P. Performance validation
of a multi-residue method for 170 pesticides in kiwifruit. In Prin-
ciples and Practices of Method Validation; Fajgelj, A., Ambrus, Á.,
Eds.; The Royal Society of Chemistry: Cambridge, England, 2000;
29–40.
[24] Kadenczki, L.; Zoltan, A.; Gardi, I.; Ambrus, Á.; Gyorfi, L.; Reese,
G.; Ebing, W. Column extraction of residues of several pesticides
from fruits and vegetables: a simple multiresidue analysis method.
J. AOAC Int. 1992, 75(1), 53–61.
[25] Food and Agriculture Organization/International Atomic Energy
Agency Laboratory Training Manual on the Use of Nuclear and Asso-
ciated Techniques in Pesticide Research, STI/DOC/10/329; IAEA:
Vienna, Australia, 1991; 87–96.
 490
Downloaded By: [Indexing Ind / Jour] At: 17:58 18 June 2007
[26] Gozek, K.; Yucel, U.; Ilim, M.; Cayci-Aysal, P.; Tuncbilek, A.S. C14-
dimethoate residues in olive oil during oil processing. Isotope Aided
Studies of Pesticide Residues during Food Processing, FAO/IAEA
TECDOC 818; IAEA: Vienna, Australia, 1995; 33–40.
[27] Aysal, P.; Gozek, K.; Artik, N.; Tuncbilek, A.S. C14-chlorpyrifos
residues in tomatoes and tomato products. Bull. Environ. Contam.
Toxicol. 1999, 62, 377–382.
Aysal et al.
[28] Tiryaki, O.; Baysoyu, D. Estimation of sample processing uncer-
tainty for chlorpyrifos residue in cucumber. Accred. Qual. Assur.
2006, 10, 550–553.
[29] Miller, J.N.; Ambrus, Á. Chapter 9 – Statistics in calibration analysis
(1). Manual on Basic Statistics, FAO/IAEA Training and Reference
Centre for Food and Pesticide Control: Vienna, Australia 2000; 1–
18.