Agent orange herbicides, organophosphate and triazinic pesticides analysis in olive oil and industrial oil mill waste effluents using new organic phase immunosensors

Article abstract of DOI:10.1016/j.foodchem.2014.07.137

New immunosensors working in organic solvent mixtures (OPIEs) for the analysis of traces of different pesticides (triazinic, organophosphates and chlorurates) present in hydrophobic matrices such as olive oil were developed and tested. A Clark electrode w

Full text of DOI:10.1016/j.foodchem.2014.07.137

Food Chemistry 169 (2015) 358–365
Contents lists available at ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
Agent orange herbicides, organophosphate and triazinic pesticides
analysis in olive oil and industrial oil mill waste effluents using new
organic phase immunosensors
⇑
Elisabetta Martini, Giovanni Merola, Mauro Tomassetti , Luigi Campanella
Department of Chemistry, ‘‘Sapienza’’ University of Rome, P.le Aldo Moro 5, 00185 Rome, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
New immunosensors working in organic solvent mixtures (OPIEs) for the analysis of traces of different
pesticides (triazinic, organophosphates and chlorurates) present in hydrophobic matrices such as olive
oil were developed and tested. A Clark electrode was used as transducer and peroxidase enzyme as mar-
ker. The competitive process took place in a chloroform–hexane 50% (V/V) mixture, while the subsequent
enzymatic final measurement was performed in decane and using tert-butylhydroperoxide as substrate
Received 18 December 2013
Received in revised form 2 July 2014
Accepted 30 July 2014
Available online 7 August 2014
of the enzymatic reaction. A linear response of between about 10 nM and 5.0 lM was usually obtained in
Keywords:
Immunosensors
Organic phase
Triazinic
the presence of olive oil. Recovery tests were carried out in commercial or artisanal extra virgin olive oil.
Traces of pesticides were also checked in the oily matrix, in pomace and mill wastewaters from an indus-
trial oil mill. Immunosensors show good selectivity and satisfactory precision and recovery tests per-
formed in olive oil gave excellent results.
Organophosphate and chlorurate pesticides
Ó 2014 Elsevier Ltd. All rights reserved.
Analysis
Olive oil
Mill waste effluents
1. Introduction
Nevertheless, due to the widespread use made of these pesticides
in the past and unfortunately in certain areas also in recent times
The speedy determination of any traces of pesticides in food oil
has become an increasingly urgent need felt by both the food indus-
tries operating in this sector and consumer associations. Atrazine
and simazine (triazinic herbicides) are among the most widely used
weedkillers (Environmental Protection Agency, 2013; Ackerman,
the need to detect their presence in different environmental and
food matrixes has constantly been felt in the last few years. This
has resulted in the development of numerous analytical methods
(Font, Manes, Moltó, & Picó, 1993; Holden & Marsden, 1969;
Pylypiw, Arsenault, & Thetford, 1997). The emphasis has been laid
on methods that may be applied also ‘‘in situ’’ (Hennion & Barcelo,
2007). Although banned in the European Union (Krämer
&
Schirmer, 2007; Wackett, Sadowsky, Martinez, & Shapir, 2002), they
are the most widely used herbicides in the US, while 2,4-D and 2,4,5-
T, i.e. respectively dichloro-, or trichloro-phenoxyacetic acid are
chlorinated phytopharmaceuticals, used as synthetic defoliants
and forming the active principle of the so-called ‘‘agent orange’’
notoriously used in recent conflicts (Quastel, 1950; Freeman et al.,
2011). On the other hand, parathion, a typical organophosphate pes-
ticide, is a potent insecticide and acaricide. It was originally devel-
oped by IG Farben (Interessen-Gemeinschaft Farbenindustrie AG)
in the 1940s. According to the non-governmental organization Pes-
ticide Action Network (or PAN), parathion is one of the most danger-
ous pesticides. Its use is banned or restricted in 23 countries and its
importation is illegal in a total of 50 countries (Fee, Gard, & Yang,
2005; Metcalf, 2002; Environmental Protection Agency, 2007).
1998; Hassoon & Schechter, 2000; Henriksen, Svensmark,
Lindhardt, & Juhler, 2001). However, it should be stressed that many
phytopharmaceutical compounds, as well as the above-mentioned
pesticides, are more soluble in organic solvent or solvent mixtures
than in aqueous solutions (Conte, Milani, Morali, & Abballe, 1997).
This can cause serious problems in chemical analysis, which have
only been partially solved by techniques such as gas chromatogra-
phy (Conte et al., 1997; Hogendoorn & Van Zoonen, 2000; Eisert &
Levsen, 1996; Coulson, Cavanagh, & Stuart, 1959) or MS (Wong,
Webster, & Halverson, 2003; Kawaguchi, Inoue, Yoshimura, &
Sakui, 2004). The difficulties can increase when the low solubility
in water solution of the analyte (i.e. several pesticides) is coupled
with the very low solubility of the real matrix in which the analyte
is contained, for instance, edible oils. Enzymatic electrodes capable
of operating in organic solvents, i.e. OPEEs have made a substantial
contribution to solving this problem (Saini, Hall, Downs, & Turner,
1991; Campanella, Lelo, Martini, & Tomassetti, 2007; Campanella,
⇑
Corresponding author. Tel.: +39 0649913722; fax: +39 0649913601.
http://dx.doi.org/10.1016/j.foodchem.2014.07.137
0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.

E. Martini et al. / Food Chemistry 169 (2015) 358–365
359
Bonanni, Martini, Todini, & Tomassetti, 2005; Campanella, Dragone,
2.2. Reagents and materials
Lelo, Martini, & Tomassetti, 2006; Sarkar & Gupta, 1989). Some-
times, however, their LOD is not sufficiently low; in addition, since
these OPEEs for pesticide analysis are inhibition biosensors, it fol-
lows that this kind of device is relatively unselective versus pesti-
cides belonging to different phytopharmaceutical classes. It is a
known fact that immunosensors are the most selective biosensors,
and our team, as well as other authors (Garcés-García, Morais,
González-Martínez, Puchades, & Maquieira, 2004), has recently fab-
ricated several immunosensors for pesticide determination
(Tomassetti, Martini, & Campanella, 2012; Raman Suri, Boro,
Nangia, & Gandhi, 2009; Rekha, Thakur, & Karanth, 2000). However
this kind of immunosensor was able to operate only in aqueous
solution and to test pesticides in aqueous matrices (Raman Suri
et al., 2009; Tomassetti et al., 2012; Campanella, Eremin, Lelo,
Martini, & Tomassetti, 2011). Therefore, when the problem arose
of having to determine traces of pesticides in oily matrices, it was
necessary to replace OPEEs (Organic Phase Enzyme Electrodes) with
OPIEs (Organic Phase Immuno Electrodes). On the other hand, the
development of new OPIE devices for pesticide analysis in edible
oil matrices raised serious problems, both because of the scant
information concerning effective immunocomplex formation in
organic solvents available in the literature (Saini et al., 1991), and
because the organic solvent used must satisfy several different
requirements. These include the fact that the solvent can com-
pletely dissolve both the pesticide, the oily matrix and the labelled
antibody, as well as not being too volatile and having a suitable logp
value (Tomassetti et al., 2012). A series of tests were thus carried out
in previous work (Tomassetti et al., 2012) by the authors using dif-
ferent solvents, different electrochemical transducers, different
immunosensor construction and operating geometries. This series
of trials led to the development of an amperometric immunosensor
for the analysis of traces of triazinic pesticides in olive oil, working
in 50% (V/V) chloroform n-hexane mixture, using a Clark electrode
for oxygen made of PTFE as transducer and horseradish peroxidase
as marker, as illustrated in previous research. This device was cer-
tainly innovative vis-à-vis what has so far been reported in litera-
ture, but it was also very suitable, as the kaff value of the
immunological method measured using the Langmuir curve was
found to be of the order of 106 Mꢀ1 in the presence of the oily phase
and about 107 Mꢀ1 in the absence of the oily phase (Tomassetti et al.,
2012). These values show that, even when the antibody reaction
occurs in organic solvent, antigen–antibody complex formation
takes place more than satisfactorily and allows an immunological
method to be developed correctly. The presence of an oily matrix
does not affect the kaff value more significantly. Finally the devel-
oped classical competitive organic phase assay also evidenced the
need for a good solubility of the substrate of the final enzymatic
reaction (i.e. tert-butylhydroperoxide). The organic solvent found
to be best suited for the task was decane, even though the same
50% (V/V) chloroform n-hexane mixture utilized also for the com-
petitive step also works satisfactory to perform the final enzymatic
measurement (Tomassetti et al., 2012).
Anti-atrazine monoclonal antibody, anti-dichloro-phenoxyace-
tic acid (i.e. 2,4-D) and anti-trichloro-phenoxyacetic acid (i.e.
2,4,5-T) antibodies, as well as atrazine and simazine carboxyderiva-
tive, dichloro-phenoxyaceticacid(i.e. 2,4-D)andtrichloro-phenoxy-
acetic acid (i.e. 2,4,5-T), were provided by Dr. S. Eremin (Department
of Chemical Enzymology, Faculty of Chemistry, Moscow State Uni-
versity, Russia). Anti-parathion was a commercial antibody and
was obtained from Acris (Acris Antibodies, Herford, Germany).
1-Chloro-3-ethylamino-5-isopropylamino-2,4,6-triazine (i.e. Atra-
zine), 6-chloro-N,N0-diethyl-1,3,5-triazine-2,4-diamine (i.e. sima-
zine),
N-tert-butyl-6-chloro-N0-ethyl-1,3,5-triazine-2,4-diamine
(i.e. terbuthylazine), diethyl 4-nitrophenyl phosphate (i.e. Para-
thion) were supplied by Pestanal Sigma–Aldrich (Sigma Aldrich,
Milan, Italy). Potassium chloride, dibasic and monobasic anhydrous
potassium phosphate RPE, chloroform RPE, dichloromethane RPE
and diethyl ether RPE were supplied by Carlo Erba Reagents (Carlo
Erba, Milan, Italy). Ny+ Immobilon Affinity membrane (porosity
0.65 lm) was provided by Millipore (Millipore Corporation, Vimod-
rone, Milan, Italy). The biotinylation kit, supplied by Sigma Immuno-
chemicals (Sigma, Milan, Italy), was composed of biotinylation
reagent (BAC-SulfoNHS, namely biotinamido hexanoic acid 3-
sulfo-N-hydroxysuccinimide ester), 5 M sodium chloride solution,
micro-spin column (2 mL) (in practice, a small empty cylindrical
vessel prepackaged with Sephadex G-50), 0.1 M sodium phosphate
buffer pH 7.2, 0.01 M phosphate buffer saline (PBS) pH 7.4 (reconsti-
tuted with 1 L of deionised water to give 0.01 M phosphate buffer,
0.138 M NaCl, 2.7 mM KCl, pH 7.4); lastly ExtravidinÒ peroxidase
(containing 0.2 mL of Extravidin Peroxidase conjugate at
2.0 mg mLꢀ1, with 0.01% thimerosal). Phenol, dialysis membrane
(art. D-9777), 1-ethyl-3 (3-dimethylaminopropyl) carbodiimide,
albumin (from bovine serum) (BSA) and TRIS (hydroxymethyl-ami-
nomethane), tert-buthylhydroperoxide solutions in decane solvent
and TWEENÒ 20, provided by Sigma Aldrich (Sigma Aldrich, Milan,
Italy).
2.3. Samples
Pomace, olive oil, mill waste water, washing olive waters and
olive oil samples were provided by an industrial (three centrifuga-
tion type) mill located in Central Italy. Two different analyzed com-
mercial extra virgin olive oil samples, produced by the most
important industrial Italian olive oil producer firms, were pur-
chased from a local shop and stored in a sealed dark glass bottle,
while two other extra virgin olive oil products, also stored in a
sealed dark glass bottle, were supplied directly by a farmer from
an area north of Rome (Italy).
3. Methods
3.1. Immunosensor assembly
The type of electrochemical transducer used was an ampero-
metric gaseous diffusion amperometric electrode for O2 determi-
nation (see Supporting Information Fig. A). The transducer
consisted of a Clark type electrode. For the immunosensor assem-
bly, in practice, three membranes were mounted on the PTFE cap of
the Clark electrode, in the following order: the gas-permeable
membrane, the dialysis membrane and the Immobilon membrane
with antibody immobilized on it. The membranes were kept in
place by a nylon net and a PTFE O-ring. A constant potential of –
650 mV with respect to an Ag/AgCl/Clꢀ anode was applied to the
Pt cathode of the oxygen electrode. Horseradish peroxidase
enzyme was used as marker for immunocomplex detection.
2. Experimental
2.1. Apparatus
The amperometric measurements were performed in a 5 mL
thermostated glass cell at 23 °C under constant stirring. The Clark
electrode, supplied by Universal Sensor Inc., New Orleans (USA),
was connected to an amperometric biosensor detector provided
by the same firms and to an analog recorder Amel mod. 868. In
all experiments performed in organic phase, the plastic cap of
the electrodes was replaced by a PTFE cap.

360
E. Martini et al. / Food Chemistry 169 (2015) 358–365
3.2. Antibody immobilization on Immobilon membrane
reaction mixture was filtered and extracted with 1 M HCl-chloro-
form. The extract was dried over MgSO4, and the solvent evapo-
rated (Kim, Lee, Chung, & Lee, 2003).
A commercial Immobilon membrane was used for antibody
immobilization. It consisted of a positively charged nylon mem-
brane with polyester reinforcement optimized for reliable and
reproducible transfer, immobilization, hybridization, and subse-
quent reprobing. The Immobilon Ny+ Membrane was cut into
3.5. Albumin-pesticide conjugation
The carboxylate pesticide was dissolved in 5 mL of dichloro-
methane to which 38 mg (i.e. 0.33 mM) of N-hydroxysuccinimide,
68 mg (i.e. 0.33 mM) of N,N-dicyclohexylcarbodiimide and 3.7 mg
(i.e. 0.03 mM) of 4-dimethylaminopyridine were added. The mix-
ture was stirred for 3 h and filtered, and the solvent removed.
The format for coupling the carboxylate pesticide to the carrier
protein is shown in Fig. 1(c), where, for the sake of example, the
format used in the case of the atrazine pesticide is illustrated.
Briefly, to prepare hapten-BSA conjugates, 20 mg (i.e.
0.3 ꢁ 10ꢀ3 mM) of BSA were dissolved in 2 mL of borate buffer
(0.2 M, pH 8.7) to which 0.4 mL of DMF was added. A solution of
an active ester (0.018 mM) in 0.1 mL of DMF was then added to
the stirred protein solution, and stirring was continued for one
day at 4 °C (Kim et al., 2003; Liu et al., 2007).
1 cm2 surface area disks and 50
lL of a 0.01 M antibody solution
was directly deposited on the membrane surface. The membrane
was then dried at room temperature for about 24 h and stored at
4 °C.
3.3. Atrazine pesticide carboxylation
The atrazine carboxylation method (Goodrow, Harrison, &
Hammock, 1990) is as follows (see Fig. 1(a)): to a stirred heteroge-
neous mixture of 1.01 g of atrazine (i.e. 5.01 mM) in 100 mL of
absolute ethanol a solution of 0.574 g of 3-mercaptopropanoic acid
was added (i.e. 5.40 mM) together with 0.714 g of 85% KOH (i.e.
10.8 mM) in 10 mL of absolute ethanol, under N2. At reflux the
mixture became homogeneous and a precipitate (KCl) soon began
to form. Reflux was continued (4 h). The hot mixture was filtered,
and the filtrate concentrated to a white solid. The solid was tritu-
rated with 25 mL of 5% NaHCO3 and filtered.
3.6. Albumin-pesticide biotinylation and peroxidase conjugation
scheme
The avidin–biotin peroxidase technique (illustrated in Fig. 1(d))
is based on the use of a biotinylated antigen and of avidin horserad-
ish peroxidase conjugate as part of the labelling system (Duk,
Lisowska, Wu, & Wu, 1994; Rao, Anderson, & Bachas, 1999). The
technique exploits the high affinity binding of biotin to avidin. The
BiotioTag kit is specially designed for the small scale labelling of
antibodies using biotinamido hexanoic acid 3-sulfo-Nhydroxysuc-
cinimide ester (BAC-SulfoNHS) as the labelling reagent. This reagent
is particularly useful when mild reaction conditions are required for
the biotinylation of sensitive biomolecules such as antibodies,
enzymes and surface proteins. After the labelling reaction, the bio-
tinylated protein is separated from the unreacted or hydrolyzed
reagent by a fast gel-filtration step using G-50 microspin columns.
BAC-SulfoNHS reacts with free amino groups of proteins to form sta-
ble amide bonds. Extravidin binds to biotin with a high affinity and
3.4. Parathion carboxylation
A solution of 3.0 g of 4-nitrophenol (i.e. 21.6 mM) in 15 mL of
acetonitrile was added dropwise to a stirred solution containing
4.59 g (i.e. 27.8 mM) of methyldichlorothiophosphate, 20 g of
finely ground K2CO3 and 20 mL of acetonitrile. After stirring for
1 h at room temperature the mixture was filtered through Celite
and the solvent removed. The following format was applied to
obtain the synthesis of the carboxylated parathion (see Fig. 1(b)):
to
a
stirred solution of 500 mg of O-methyl-O-(4-nitro-
phenyl)phosphorochloridothioate (i.e. 1.87 mM) in 3 mL of MeOH
cooled in an ice-water bath, a solution of 274 mg (i.e. 4.88 mM)
of KOH and 229 mg of 4 aminobutyric acid (i.e. 2.22 mM) in
1.7 mL of MeOH was added dropwise. After stirring for 5 min, the
Fig. 1. (a) Atrazine carboxylation; (b) parathion carboxylation; (c) conjugation with BSA; (d) biotinylation and conjugation with Extravidin-peroxidase.

E. Martini et al. / Food Chemistry 169 (2015) 358–365
361
specificity. The high affinity for biotin alleviates non-specific bind-
ing interactions commonly associated with the strongly basic avidin
protein. The use of the extended spacer arm greatly improves the
interaction between Extravidin and the biotinylated macromole-
cule, thus overcoming the steric hindrance present at the biotin
binding sites of Extravidin. In brief: 0.1 mL of 4.5 ꢁ 10ꢀ4 M pesti-
cide–albumin solution in sodium phosphate buffer (pH 7.2; 0.1 M)
was prepared. BAC-SulfoNHS solution 5 mg mLꢀ1 was also prepared
separately by dissolving 5 mg of biotinamido hexanoic acid 3-sulfo-
specific adsorption on the membrane). The pesticide to be deter-
mined was added in 5 mL of chloroform–n-hexane mixture 50%
(V/V) contained in the measurement cell, together with a fixed
supply of pesticide Biotin-Avidin-peroxidase conjugated, i.e.
20 lL (0.01 M solution) of conjugated pesticide in the same solvent
mixture. The peroxidase-conjugated pesticide was allowed to com-
pete with the non-conjugated pesticide, both free in solution, in
binding with the antibody immobilized on the Immobilon mem-
brane. After washing with the same solvent mixture to remove
all the unbound labelled pesticide, the specific substrate of the
N-hydroxysuccinimide ester in 30
and adding sodium phosphate buffer (pH 7.2; 0.1 M) for a final vol-
ume of 1 mL. 10 L of BAC-SulfoNHS solution were immediately
lL DMSO (dimethylsulphoxide)
enzyme, i.e. 20 lL of tert-buthylhydroperoxide solution 1% V/V
l
was added to 5 mL of the selected organic solvent, i.e. decane, in
which the immunosensor was dipped, under stirring. The signal
measured (as nA) of the transducer correlated directly with the
pesticide concentration to be measured. The higher the pesticide
concentration to be measured, the lower the oxygen consumed
in the enzymatic reaction reported in Fig. (A) of Supporting Infor-
mation, the higher the amperometric signal of the O2 reduced at
the Clark electrode. The sequence for measuring the pesticide by
the above format is schematized in Fig. 2. Using this format a cal-
ibration curve was constructed (by plotting the current variation
added to the pesticide–albumin solution with gentle stirring and
the mixture incubated under stirring for 30 min at 2–8 °C. Then
the resin contained in a micro-spin G-50 column was re-suspended
in the column by vortexing and equilibrated with 0.2 mL of PBS, (pH
7.40; 0.01 M); this buffer was needed also for the elution of the
labelled protein from the column. The biotinylation reaction mix-
ture was applied to the top-centre of the resin and the column
was centrifuged for 5 min at 700ꢁg. The purified sample was then
collected at the bottom of an Eppendorf test tube. This step was
repeated twice more and a total of three fractions were collected.
As Extravidin binds to biotin with a high affinity and specificity,
D
I (nA) as a function of the log of pesticide concentration (as M)
and employed to determine the unknown concentration of pesti-
cide contained in the sample. In this way the original data of the
immunosensor response to increasing antigen concentration,
which always display logarithmic trends, take on a sufficiently lin-
ear trend. Moreover, the application of other, even more complex
methods, described in the literature to obtain the same result, for
instance the so called logit-log method, or equations with three-
four parameters (Flynn, 2004) do not provide better results than
those described above. The enzymatic reaction response took
about 15–20 min. Individual measurements were performed, each
time using a new membrane.
an Extravidin peroxidase solution (20 l
L, 2.0 mg mLꢀ1), diluted
1:100 in PBS containing 1% BSA solution, was added to the collected
fractions and incubated for 1 h at room temperature, then gently
rinsed with PBS, (pH 7.4; 0.01 M).
3.7. Determination of pesticide by immunosensor. (Competitive format
between pesticide Biotin-Avidin-peroxidase conjugated and non-
conjugated, both free in chloroform–hexane mixture, for the antibody
immobilized in membrane)
For this purpose, the Immobilon membrane, on which the anti-
body was immobilized, was fixed to the head of the amperometric
Clark type electrode as described in Section 3.1. Before measure-
ment, the immunosensor was dipped into a Tris–HCl buffer solu-
tion, (pH 8.0; 0.1 M), containing 0.05% Tween-20 by weight and
2.5% BSA by weight (bovine albumin was used to minimize non
3.8. Membrane regeneration
A membrane already used in a previous test can be regenerated
by suitably rinsing it in a solution of glycine 0.1 M and MgCl2
2.5 M, at pH = 2. This solution is apparently capable of splitting
Fig. 2. Measurement: competition between pesticides and a fixed concentration of peroxidase pesticide conjugated, both free in organic phase solution, for the antibody
immobilized in membrane.

362
E. Martini et al. / Food Chemistry 169 (2015) 358–365
the antibody complex. However, it was found that the immunosen-
sor response when using a regenerated membrane may be up to
15% lower than in the previous measurement. This would of course
affect the repeatability of the measurement and thus the method’s
precision.
factors into consideration. An ad hoc table, as reported in a previ-
ous paper (Tomassetti et al., 2012), showed how the chloroform–n-
hexane 50% (V/V) mixture, used also in the present work, repre-
sents a good compromise between all the various requirements
set out in the ‘‘Section 1’’. Indeed, both the checked analyte and
the oily samples are completely soluble in this solvent mixture,
as is also the marked antigen. The hydrophobicity value of this
mixture is still satisfactory (logp < 3, but > 2), and the dielectric
constant (DEC = 3.7) is neither too high nor too low. In other words
the choice fell on an organic but non alcoholic mixture that had
previously proved to be particularly suitable also when enzymatic
OPEEs were being developed (Campanella, De Luca, Sammartino, &
Tomassetti, 1999). Other tests, performed in the previous research
and in the same conditions (Tomassetti et al., 2012), evidenced that
also the signal obtained from the final peroxidase enzymatic reac-
tion was higher when performed in organic solvent, particularly in
decane, than in aqueous solution, since, as reported in the litera-
ture (Saini et al., 1991; Tomassetti et al., 2012), the enzymatic sen-
sor’s response is several times better when operating in organic
solvent (or solvent mixture) than in aqueous buffer (owing to the
lower interference from hydrophilic ionic species, lower microbial
contamination, enhanced thermostability, and so on).
3.9. Pesticides measurement in real oily samples or mill effluents and
recovery tests
The immunosensor described in this paper was used first of all
to measure the triazinic (atrazine, simazine, tert-buthylazine),
organophosphate (i.e. parathion) and chlorurate (i.e. 2,4-D and
2,4,5-T) pesticides possibly present in commercial and artisanal
olive oils using the respective calibration curves, then to measure
the traces of triazinic pesticides found during the olive oil produc-
tion process in different waste effluent samples (e.g. olive washing
and vegetable waters), as well as in olive oil and pomace, firstly
several tests were performed to check for traces of triazinic pesti-
cides in olive oil samples examined ‘as is’. To this end, 0.5 mL of the
oily sample were added in 4.5 mL of chloroform–n-hexane 50% (V/
V) mixture and the measure was performed as described in para-
graph 3.7 above using the simazine calibration curve to obtain
the final concentration of triazinic pesticides, expressed as sima-
zine concentration. In addition, recovery tests were carried out
on the oily samples spiked with known concentrations of each pes-
ticide in order to obtain a final pesticide concentration of about
4.2. Carboxylation and BSA conjugation formats
With regard to the fabrication of the labelled pesticide required
in the functioning of the immunosensors constructed, it should be
noted that, in the case of atrazine, as described in the preceding
Section 3.3, it had first to be carboxylated (see Fig. 1(a)). The latter
could thus be conjugated with the BSA (Fig. 1(c)). As far as the
enzymatic labelling is concerned, the classical method was used
involving avidin and Extravidin (Duk et al., 1994; Rao et al.,
1999) (Fig. 1(d)) as described in Section 3.6. Also in the case of
the parathion it was necessary to follow the same reaction format,
described for atrazine in Fig. 1(d), to obtain the labelled hapten. Of
course the reactions to achieve carboxylation (see Fig. 1(b)) were
slightly different from those used to carboxylate the atrazine
(Fig. 1(a)). Finally, in the case of the 2,4-D and the 2,4,5-T, it was
naturally not necessary to perform pesticide carboxylation as their
molecule already possesses a carboxyl group. It was therefore suf-
ficient directly to perform the conjugation with BSA, and thus the
customary enzymatic labelling with avidin and Extravidin illus-
trated respectively in Fig. 1(c) and (d) for the atrazine. Fig. 2 shows
the scheme of the measurement and the competitive format
between the pesticide and a fixed concentration of peroxidase pes-
ticide conjugated, both free in the organic phase mixture, for the
antibody immobilized in the membrane.
10ꢀ8 M. For this purpose 200 L of a solution of pesticide 10ꢀ5
l M
was dissolved in 4.5 mL of chloroform–n-hexane 50% (V/V) mix-
ture, to which 0.5 mL of a commercial or artisanal extra virgin olive
oil sample were added. Also in this case the measurements were
performed as described in paragraph 3.7 and using the respective
calibration curves to obtain the concentrations of spiked solutions.
A second kind of test was run to determine triazinic pesticides
concentration in real samples taken during olive oil production
process in industrial mills (olive washing waters, olive mill waste-
waters (vegetable waters), oil and olive oil). When the olive oil
sample was being tested, 0.5 mL of the sample were added directly
to the measurement cell, as previously explained for the measures
carried out on commercial and artisanal oils, while, in the case of
the sample of ‘‘olive washing waters’’ and ‘‘olive mill wastewaters’’
(vegetable waters) 2.0 mL and 1.0 mL samples, respectively, were
taken and added to the measurement cell making up to 5.0 mL
with chloroform–n-hexane 50% (V/V) mixture. The measurement
was then performed as above described in Section 3.7. In the case
of the pomace sample, 50 g of the sample were mixed with the
chloroform–n-hexane 50% (V/V) mixture, stirred for about
10 min, then centrifuged at 1500 rpm for 8 min. Then 2.5 mL of
the supernatant were removed and added to 2.5 mL of chloro-
form–n-hexane 50% (V/V) mixture in the measurement cell and
the measurement performed in the normal way.
4.3. Analytical data
The behaviour of new OPIEs’ response as a function of growing
2,4-D, 2,4,5-T, or parathion pesticide concentration, the corre-
sponding calibration curves and the confidence intervals for three
pesticide determinations, obtained using a semilogarithmic scale,
are shown in Fig. 3, while the analytical data obtained for the
respective calibration curves, for the analysis of these pesticides
in extra virgin olive oil, are summarized in Table 1. On the other
hand, the same data, for the analysis of atrazine and simazine,
already found in the previous research (Tomassetti et al., 2012),
are, for the reader’s convenience, also displayed in the same Table 1
in the present paper. Significantly, the linear range is very similar
4. Results and discussion
4.1. Parameter optimization
The factors such as the concentration of the labelled antigen
used free in solution, or the concentration of the substrate used
in the final enzymatic reaction, the competitive step time and so
on, were optimized in a previous investigation (Tomassetti et al.,
2012). The Clark type electrode was used as in practice the gas-
permeable membrane protects the working and reference inner
electrodes. This ensures that the system is more repeatable and
the method ‘robust’ (Sassolas, Prieto-Simón, & Marty, 2012). As
explained in the introduction, the choice of the organic solvent in
which to perform the competitive format involved taking multiple
in all cases (between about 10 nM and 5.0 lM) as is the LOD
(between about 4 and 8 nM). The highest calibration sensitivity
was found to be that of the immunosensor for atrazine towards
atrazine itself, i.e. almost double that of the other immunosensors
towards the respective pesticides. However, also in the case of the

E. Martini et al. / Food Chemistry 169 (2015) 358–365
363
Fig. 3. (a) Behaviour of immunosensors response, respectively to 2,4,5-T, 2,4-D and parathion, as a function of growing pesticide concentration, using the Immobilon
membrane for antibody immobilization, an amperometric electrode for O2 as transducer and the competition in chloroform – n-hexane (50% V/V) in the presence of olive oil,
lastly final enzymatic measurement in decane; (b) corresponding calibration curves and confidence intervals for respective pesticides determination obtained using a
semilogarithmic scale.
Table 1
Analytical data of new OPIEs, obtain for different pesticides measurement: of 2,4-D, 2,4,5-T, parathion, atrazine and simazine calibration curves in presence of extra virgin olive
oil, using Clark electrode for oxygen as transducer. The solvent used for competitive step was chloroform–n-hexane (50% V/V), while in the final enzymatic measurement decane
was used as solvent and tert-butylhydroperoxide as substrate.
Pesticide
2,4,5-T
2,4-D
Parathion
Atrazine
Simazine
Regression equation
y = 72.1 ( 1.2) logx
y = 98.1 ( 1.9) logx
y = 80.5 ( 2.1) logx
y = 177.9 ( 4.2) logx
y = 98.1 ( 1.9) logx
(y =
D
i (nA), x = M)
+ 143.6 ( 10.9)
+ 468.6 ( 14.9)
+ 395.2 ( 17.5)
+ 881.8 ( 34.4)
+ 468.6 ( 14.9)
Linear range (M)
1.2 ꢁ 10ꢀ8 ꢀ 5.0 ꢁ 10ꢀ6 1.5 ꢁ 10ꢀ8 ꢀ 5.0 ꢁ 10ꢀ6 1.0 ꢁ 10ꢀ8 ꢀ 2.5 ꢁ 10ꢀ6 1.2 ꢁ 10ꢀ8 ꢀ 5.0 ꢁ 10ꢀ6 1.2 ꢁ 10ꢀ8 ꢀ 5.0 ꢁ 10ꢀ6
Correlation coefficient
Repeatability of the measurement
(as pooled SD%)
Low limit of detection (LOD) (M)
(RSD% 6 5.0)
0.9869
7.2
0.9827
7.3
0.9882
7.0
0.9895
6.1
0.9869
7.2
0.4 ꢁ 10ꢀ8
0.8 ꢁ 10ꢀ8
0.5 ꢁ 10ꢀ8
0.8 ꢁ 10ꢀ8
0.8 ꢁ 10ꢀ8
latter, the sensitivity towards the respective pesticides was found
to be very good (between about 72 and 98 nA/logM). Generally
speaking, the precision was of the same order and always accept-
able (the pooled SD was found to be 67%, in any case). The Sup-
porting Information (Table (a)) illustrates the selectivity of the
immunosensor devices developed for the different pesticides (i.e.
triazinic, chlorurate and organophosphate) by comparing the %
response of each respective immunosensor to different pesticides.
It may be observed that of the immunosensors constructed, the
one for parathion analysis is certainly the most selective. Also
noteworthy is that, except for the cases in which the structural for-
mula of the pesticide being tested is very similar to that of the spe-
cific antigen of the antibody used (for instance, this is the case of
simazine versus atrazine, or of 2,4,5-T versus 2.4-D), the antibody
response to antigens belonging to different classes of pesticide,
(i.e. paraoxon and carbaryl) is either very low or even negligible.
This means that each of the immunosensors constructed possesses
a good selectivity versus other classes of pesticide that differ from
the class of pesticides for which it was constructed.
4.4. kaff value determination
It was decided to test the validity also from the point of view of
the immunological methods underlying the immunosensors con-
structed. For this purpose, a comparison is made in Supporting
Information (Table (b)) between the kaff and IC50 values obtained
for the different pesticides tested using the Langmuir curve for
the respective immunosensors. It is noteworthy that the kaff values
all lie between about (4 ꢁ 106 and 6 ꢁ 106) Mꢀ1; in the case of
simazine, the value is slightly lower (about 2.6 ꢁ 106), which is
only to be expected as it was calculated using the immunosensor
constructed for atrazine. Moreover, this is in agreement with what
may be observed in the table of selectivities (see S.I. Table (a)).
4.5. Tests on real samples
A series of tests were run on real samples of extra virgin olive
oil. Table 2 shows the triazinic pesticide concentrations (expressed
as simazine concentration) found in commercial or artisanal extra

364
E. Martini et al. / Food Chemistry 169 (2015) 358–365
Table 2
Triazinic pesticides concentration (expressed as simazine concentration, both as M and as mg kgꢀ1 of oil) found in commercial or artisanal extra virgin olive oil samples, using the
atrazine OPIE.
*
Sample
Found concentration of simazine RSD% 6 6.5
(M)
Value found in literature (mg kgꢀ1
)
(mg kgꢀ1
)
Commercial oil n.1
Commercial oil n.2
Artisanal oil n.1
Artisanal oil n.2
Lower of the LOD value (0.8 ꢁ10ꢀ8
)
Lower of the LOD value (0.002)
0.007
0.016
0.005–0.5
0.005–0.5
0.005–0.5
0.005–0.5
2.8 ꢁ 10ꢀ8
6.5 ꢁ 10ꢀ8
Lower of the LOD value (0.8 ꢁ 10ꢀ8
)
Lower of the LOD value (0.002)
*
Ferrer et al., 2005.
pesticide were detected in all types of sample, the larger traces
being found in pomace if operating in organic solvent. It should
also be noted that the values obtained using the OPIE are always
slightly higher than those obtained using the immunosensor oper-
ating in aqueous medium. As expected, the greater differences are
those obtained for olive oil and for pomace since, as these products
are highly hydrophobic, the determination using the OPIE gives
much better results than when an immunodevice operating in
aqueous medium is used. In the other two cases in which the sam-
ples were either water–oil emulsions (vegetable waters), or an
almost wholly aqueous sample (olive oil mill washing waters),
the results obtained were much more similar but in any case
always slightly higher when an OPIE was used than those obtained
using the immunosensor operating in aqueous solution. In our
view, since the pesticide is much more soluble in organic solvent
than in aqueous medium (Conte et al., 1997), this fact favours
the distribution of the pesticide in the organic solvent, thus facili-
tating its determination and providing, also in this case, values that
are only slightly higher than those obtained using the immunosen-
sor operating in aqueous medium.
Table 3
Traces of triazinic pesticides (expressed as simazine concentration) found during olive
oil production process in different fluid samples.
Samples
Enzymatic measurement Enzymatic measurement in
in phosphate buffer
Concentration (M)
organic solvent (decane)
Concentration (M)
In all cases RSD% 6 5.5
In all cases RSD% 6 5.5
Olive washing waters 1.71 ꢁ 10ꢀ8
2.43 ꢁ 10ꢀ8
Olive mill
4.00 ꢁ 10ꢀ8
4.05 ꢁ 10ꢀ8
wastewaters
(vegetable waters)
Pomace
Olive oil
2.16 ꢁ 10ꢀ8
2.82 ꢁ 10ꢀ8
8.25 ꢁ 10ꢀ8
6.52 ꢁ 10ꢀ8
virgin olive oil samples using the new OPIE for atrazine. Values are
expressed both in M and in (mg kgꢀ1 of oil) and referred to sima-
zine, since the latter is the triazinic pesticide most frequently
found in olive oil samples (Amvrazi & Albanis, 2006; Garcia-
Reyes, Ferrer, Michael Thurman, Fernández-Alba, & Ferrer, 2006;
Ferrer et al., 2005). In practice, in only one sample of the various
artisanal oil samples tested were pesticide concentrations found
that, although very small, proved not to be completely negligible,
while the traces found in one of the commercial oils, although
detectable, were so low as to be deemed practically negligible.
However, also several recovery tests were carried out on spiked
extra virgin olive oil samples in which the absence of any traces
of triazinic pesticides had been ensured. In this case the tests were
performed alternatively using both atrazine and simazine, or else
tert-butylazine and, in the Supporting Information (Table (c)), it
can be seen how recoveries always lie between about 96% and
104%, and are thus satisfactory. Other recovery tests were carried
out also for the other classes of non triazinic pesticides using the
respective OPIEs. As shown in the same Supporting Information
(Table (c)), also in the case of the recovery tests involving 2,4-D
or 2,4,5-T, or the Parathion pesticide, carried out in spiked extra
virgin olive oil samples lacking any trace of pesticides, the percent-
age recoveries were found to be comparable to those found for the
triazinic pesticides, and were thus also more than satisfactory. One
final test was performed to check for the possible presence of
traces of triazinic pesticides (expressed as simazine concentration)
found during the olive oil production process in an industrial mill
in Central Italy. For this purpose the atrazine OPIE was used to test
both the oily products and the wastewater from a ‘‘centrifugation
type’’ industrial mill. Samples of olive oil, pomace, olive mill waste-
waters (vegetable waters) and olive washing waters were therefore
analyzed. Since in this case the various samples were of different
types, i.e. oily, mixture of solid-oil phase, water–oil emulsion
and, in one case, also essentially aqueous phase, the tests were per-
formed using the atrazine immunosensor both operating entirely
in organic solvent mixture, following the format described in the
Sections 3.9 and 3.7, or else following the same format but operat-
ing in aqueous medium. Table 3 shows how very small traces of
5. Conclusions
The results obtained so far have shown that several immuno-
sensors operating in a 50% (V/V) chloroform–hexane solvent mix-
ture can be developed for the purpose of determining atrazine,
simazine, parathion, 2,4-D and 2,4,5-T in olive oil. It was also
observed that good results were obtained using a Clark type trans-
ducer, performing not only the competitive assay but also the
actual final enzymatic reaction in organic solvent instead of in
aqueous solution, i.e. in practice carrying out the final electroenzy-
matic measurement using a classical OPEE. Best results were
obtained using as substrate for the peroxidase catalyzed enzymatic
reaction a solution of tert-butylhydroperoxide in decane. It was
also demonstrated that when using the atrazine OPIE good results
were obtained not only in the analysis of extra virgin olive oil or
pomace oil, but also in the analysis of partially aqueous effluents
which normally originate in ordinary industrial mills used to pro-
duce olive oil. The simplicity of the method, the relatively low cost
and the ease of using these new immunodevices to detect traces of
different types of pesticide in such an important food oil as olive oil
means that the method is of great practical interest to the modern
food industry. Fresh research is therefore already under way to
extend the method also to other types of food oil, such as seed
oil, in which for instance the detection of several of the pesticides
tested, such as 2,4-D and 2,4,5-T, is generally much more frequent.
Acknowledgements
Authors are indebted to Prof. S. Eremin (Moscov State Univer-
sity), who supplied antibodies of atrazine, 2,4-D and 2,4,5-T. This
work was funded by ‘‘La Sapienza’’ University of Rome, ‘‘Ateneo
and University Projects.

E. Martini et al. / Food Chemistry 169 (2015) 358–365
365
Garcia-Reyes, J. F., Ferrer, C., Michael Thurman, E., Fernández-Alba, A. R., & Ferrer, I.
(2006). Analysis of herbicides in olive oil by liquid chromatography time-of-
flight mass spectrometry. Journal of Agricultural and Food Chemistry, 54(18),
6493–6500.
Goodrow, M. H., Harrison, R. O., & Hammock, B. D. (1990). Hapten synthesis,
antibody development, and competitive inhibit ion enzyme immunoassay for s-
Triazine herbicides. Journal of Agricultural and Food Chemistry, 38, 990–996.
Hassoon, S., & Schechter, I. (2000). In situ fluorimetric determination of pesticides
on vegetables. Analytica Chimica Acta, 405, 9–15.
Hennion, M. C., & Barcelo, D. (1998). Strengths and limitations of immunoassays for
effective and efficient use for pesticide analysis in water samples: A review.
Analytica Chimica Acta, 362, 3–34.
Henriksen, T., Svensmark, B., Lindhardt, B., & Juhler, R. K. (2001). Analysis of acidic
pesticides using in situ derivatization with alkylchloroformate and solid-phase
microextraction. Chemosphere, 44(7), 1531–1539.
Hogendoorn, E., & Van Zoonen, P. (2000). Recent and future developments of liquid
chromatography in pesticide trace analysis. Journal of Chromatography A, 892(1–
2), 435–453.
Holden, A. V., & Marsden, K. (1969). Single-stage clean-up of animal tissue extracts
for organochlorine residue analysis. Journal of Chromatography A, 44(3),
481–492.
Kawaguchi, M., Inoue, K., Yoshimura, M., & Sakui, N. (2004). Trace analysis of
phenolic xenoestrogens in water samples by stir bar sorptive extraction with in
situ derivatization and thermal desorption-gas chromatography–mass
spectrometry. Journal of Chromatography A, 1041(1–2), 19–26.
Kim, M. J., Lee, H. S., Chung, D. H., & Lee, Y. T. (2003). Synthesis of haptens of
organophosphorus pesticides and developed of enzyme-linked immunosorbent
assays for parathion-methyl. Analitica Chimica Acta, 493, 47–62.
Krämer, W., & Schirmer, U. (2007). In W. Krämer & U. Schirmer (Eds.). Modern crop
protection compounds (Vol. 1). Weinheim: WILEY-VCH Verlag gmbh & Co. kgaa.
ISBN: 978-3-527-31496-6.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.foodchem.
2014.07.137.
References
Ackerman, F. (2007). The economics of atrazine. International Journal of Occupational
and Environmental Health, 13(4), 437–445.
Amvrazi, E. G., & Albanis, T. A. (2006). Multiresidue method for determination of 35
pesticides in virgin olive oil by using liquid–liquid extraction techniques
coupled with solid-phase extraction clean up and gas chromatography with
nitrogen phosphorus detection and electron capture detection. Journal
Agricultural and Food Chemistry, 54(26), 9642–9651.
Campanella, L., Bonanni, A., Martini, E., Todini, N.,
& Tomassetti, M. (2005).
Determination of triazine pesticides using a new enzyme inhibition tyrosinase
OPEE operating in chloroform. Sensors and Actuators B, 111–112, 505–514.
Campanella, L., De Luca, S., Sammartino, M. P., & Tomassetti, M. (1999). A new
organic phase enzyme electrode for the analysis of organophosphorus
pesticides and carbamates. Analytica Chimica Acta, 385(1–3), 59–71.
Campanella, L., Dragone, R., Lelo, D., Martini, E., & Tomassetti, M. (2006). Tyrosinase
inhibition organic phase biosensor for triazinic and benzotriazinic pesticide
analysis (part two). Analytical and Bioanalytical Chemistry, 384, 915–921.
Campanella, L., Eremin, S., Lelo, D., Martini, E., & Tomassetti, M. (2011). Reliable new
immunosensor for atrazine pesticide analysis. Sensors and Actuators B, 147,
78–86.
Campanella, L., Lelo, D., Martini, E., & Tomassetti, M. (2007). Organophosphorus and
carbamate pesticide analysis using an inhibition tyrosinase organic phase
enzyme sensor; comparison by butyrylcholinesterase + choline oxidase OPEE
and application to natural waters. Analytica Chimica Acta, 587(1), 22–32.
Liu, Y. H., Jin, M. J., Gui, W. J., Chen, J. L., Guo, Y. R., & Zhu, G. N. (2007). Hapten design
and indirect competitive immunoassay for parathion determination:
Correlation with molecular modelling and principal component analysis.
Analitica Chimica Acta, 591, 173–182.
Metcalf, R. L. (2002). Insect control. Ullman’s encyclopedia of industrial chemistry. New
York: Wiley-VCH Verlag GmbH & Co. KGaA.
Pylypiw, H. M., Arsenault, T. L., & Thetford, C. M. (1997). Suitability of microwave-
assisted extraction for multiresidue pesticide analysis of produce. Journal of
Agricultural and Food Chemistry, 45(9), 3522–3528.
Conte, E., Milani, R., Morali, G.,
& Abballe, F. (1997). Comparison between
accelerated solvent extraction and traditional extraction methods for the
analysis of the herbicide diflufenican in soil. Journal of Chromatography A,
765(1), 121–126.
Coulson, D. M., Cavanagh, L. A.,
& Stuart, J. (1959). Pesticide analysis, gas
chromatography of pesticides. Journal of Agricultural and Food Chemistry, 7,
250–260.
Duk, M., Lisowska, E., Wu, J. H., & Wu, A. M. (1994). The biotin/avidin-mediated
microtiter plate lectin assay with the use of chemically modified glycoprotein
ligand. Analytical Biochemistry, 221, 266–272.
Quastel, J. H. (1950). 2,4-Dichlorophenoxyacetic Acid (2,4-D) as
herbicide. Agricultural control chemicals. Advances in Chemistry, 1, 244–249.
Raman Suri, C., Boro, R., Nangia, Y., Gandhi, S. (2009). Immunoanalytical
a selective
&
techniques for analyzing pesticides in the environment. TrAC Trends in
Analytical Chemistry, 28(1), 30–40.
Eisert, R.,
& Levsen, K. (1996). Solid-phase microextraction coupled to gas
chromatography: A new method for the analysis of organics in water. Journal
of Chromatography A, 733(1–2), 143–157.
Environmental Protection Agency (2007). Parathion. Integrated Risk Information
System. U.S. January 26, 2007.
Environmental Protection Agency (2013). Atrazine Updates. Current as of January
2013, August 24, 2013.
Fee, D. C., Gard, D. R., & Yang, C. (2005). Phosphorus compounds. Kirk-Othmer
encyclopedia of chemical technology. New York: John Wiley & Sons.
Ferrer, C., Gomez, M. J., Garcıa-Reyes, J. F., Ferrer, I., Thurman, E. M., & Fernandez-
Alba, A. R. (2005). Determination of pesticide residues in olives and olive oil by
matrix solid-phase dispersion followed by gas chromatography/mass
spectrometry and liquid chromatography/tandem mass spectrometry. Journal
of Chromatography A, 1069, 183–194.
Rao, S. V., Anderson, K. W.,
& Bachas, L. G. (1999). Controlled layer-by-layer
immobilization of horseradish peroxidase. Biotechnology and Bioengineering, 65,
389–396.
Rekha, K., Thakur, M. S., & Karanth, N. G. (2000). Biosensors for organophosphorus
pesticide monitoring. CRC Critical Reviews in Biotechnology, 20, 213–235.
Saini, S., Hall, G. F., Downs, M. E. A., & Turner, A. P. F. (1991). Organic phase enzyme
electrodes. Analytica Chimica Acta, 249, 1–15.
Sarkar, A., & Gupta, R. S. (1989). Determination of organochlorine pesticides in
indian coastal water using
937–1071.
a
moored in-situ sampler. Water Research, 23,
Sassolas, A., Prieto-Simón, B.,
&
Marty, J. L. (2012). Biosensors for pesticide
detection: New Trends. American Journal of Analytical Chemistry, 3, 210–232.
Tomassetti, M., Martini, E., & Campanella, L. (2012). New immunosensors operating
in organic phase (OPIEs) for analysis of triazinic pesticides in olive oil.
Electroanalysis, 24(4), 842–856.
Wackett, L. P., Sadowsky, M. J., Martinez, B., & Shapir, N. (2002). Biodegradation of
atrazine and related s-triazine compounds: From enzymes to field studies.
Applied Microbiology and Biotechnology, 58(1), 39–45.
Wong, J. W., Webster, M. G., & Halverson, C. A. (2003). Multiresidue pesticide
analysis in wines by solid-phase extraction and capillary gas chromatography–
mass spectrometric detection with selective ion monitoring. Journal of
Agricultural and Food Chemistry, 51(5), 1148–1161.
Flynn, M. R. (2004). The 4-parameter lognormal (SB) model of human exposure.
Annual Occupational Hygiene, 48(7), 617–622.
Font, G., Manes, J., Moltó, J. C., & Picó, Y. (1993). Solid-phase extraction in multi-
residue pesticide analysis of water. Journal of Chromatography A, 642, 89–105.
Freeman, L. E. B., Rusiecki, J. A., Hoppin, J. A., Lubin, J. H., Koutros, S., Andreotti, G.,
et al. (2011). Atrazine and cancer incidence among pesticide applicators in the
agricultural health study (1994–2007). Environmental Health Perspectives,
119(9), 1253–1259.
Garcés-García, M., Morais, M., González-Martínez, M. A., Puchades, R., & Maquieira,
A. (2004). Rapid immunoanalytical method for the determination of atrazine
residues in olive oil. Analytical and Bioanalytical Chemistry, 378, 484–489.