Journal of Chromatography A, 1218 (2011) 4790–4798

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Trace analysis of pesticides in paddy field water by direct injection using liquid
chromatography–quadrupole-linear ion trap-mass spectrometry
Lucía Pareja a,b , M.J. Martínez-Bueno a , Verónica Cesio b , Horacio Heinzen b , A.R. Fernández-Alba a,∗
a
Pesticide Residue Research Group, Department of Hydrogeology and Analytical Chemistry, University of Almeria, 04120 La Cañada de San Urbano, Almeria, Spain
b
Cátedra de Farmacognosia y Productos Naturales, Facultad de Química, UdelaR, General Flores 2124, 11800 Montevideo, Uruguay

a r t i c l e i n f o a b s t r a c t

Article history: A multiresidue method was developed for the quantification and confirmation of 70 pesticides in paddy
Available online 26 February 2011 field water. After its filtration, water was injected directly in a liquid chromatograph coupled to a
hybrid triple quadrupole-linear ion trap-mass spectrometer (QqLIT). The list of target analytes included
Keywords: organophosphates, phenylureas, sulfonylureas, carbamates, conazoles, imidazolinones and others com-
Paddy field water pounds widely used in different countries where rice is cropped. Detection and quantification limits
Pesticides
achieved were in the range from 0.4 to 80 ng L−1 and from 2 to 150 ng L−1 , respectively. Correlation
Direct injection analysis
coefficients for the calibration curves in the range 0.1–50 ␮g L−1 were higher than 0.99 except for diazi-
LC–MS/MS
non (0.1–25 ␮g L−1 ). Only 9 pesticides presented more than 20% of signal suppression/enhancement, no
matrix effect was observed in the studied conditions for the rest of the target pesticides. The method
developed was used to investigate the occurrence of pesticides in 59 water samples collected in paddy
fields located in Spain and Uruguay. The study shows the presence of bensulfuron methyl, tricyclazole, car-
bendazim, imidacloprid, tebuconazole and quinclorac in a concentration range from 0.08 to7.20 ␮g L−1 .

© 2011 Elsevier B.V. All rights reserved.

1. Introduction candidates, bentazone and glyphosate, are also under review to be

included in the list as it is presented in the Annex III of this Directive
The widespread use of pesticides, not only in the agriculture but [4]. The maximal concentrations authorized for these contaminants
also in domestic and industrial activities resulted in the presence vary from 4 ng L−1 to 20 ␮g L−1 depending on the chemical nature
of residues of these products and their metabolites in the envi- of the compound [4]. Moreover, environmental quality standards
ronment. Many of these pesticides show a strong persistence in have been proposed for a number of pesticides and other contam-
the soil–water environment and also in fatty tissue as they tend to inants in inland and other surface waters [5].
bioaccumulate [1,2]. Moreover, due to their physicochemical prop- In order to detect the contamination of water resources by
erties, pesticides can leach from agricultural fields to ground and pesticide residues that can threaten environment preservation it
surface waters being a potential risk for ecosystems as well as for is mandatory to develop simple, fast and reliable analytical tools
drinking water quality [1]. which can be used to determine a wide range of pesticides at
During the last decade, the public/government concern on envi- such low concentrations. Although there are several methodolo-
ronmental pollution caused by pesticides use has been growing. gies developed for the analysis of pesticides residues in agriculture
Many regulatory organisms, like the European Commission (EC), waters these techniques require a first step of extraction followed
have adopted strict regulations trying to hamper or minimize the by a clean-up and pre-concentration step in order to detect a
negative effects in the environment. In the water policy field, the threshold down to 0.1 ␮g L−1 .
European Union (EU) established different directives such as the The main analytical techniques reported for the analysis of pes-
Water Framework Directive 2000/60/EC whose main objective is ticide residues in water samples are solid phase extraction (SPE)
to protect and prevent water quality [3]. In 2008, the Directive and liquid–liquid extraction (LLE). SPE has been developed as an
No. 2008/105/EC was set amending the Directive 2000/60/EC and alternative for LLE, owing to its simplicity and economy in terms of
establishing a list of 33 priority substances in water to be controlled, analysis time and solvents consumption [6]. Nevertheless, still it is
where the third part of the list are pesticides [4]. Two other possible a quite laborious and expensive technique, as many work-up steps
are involved and cartridges cost is quite high. Another important
disadvantage of SPE is the high amount of co-extractives presented
∗ Corresponding author. in the final extract, as generally 50- to 750-fold pre-concentration is
E-mail address: amadeo@ual.es (A.R. Fernández-Alba). needed to reach appropriate limits of detection (LODs) [1,2,7–10].

0021-9673/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2011.02.044
 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798 4791

In the last years, the new LC–MS/MS technology has improved to 20% in 0.1 min. The total run time for the analysis in the nega-
the instrumental detection limits of LC–MS systems from tive mode was 15 min and the re-equilibration time was 5 min. The
nanograms to sub-picograms levels, turning LC/MS–MS an invalu- injection volume was 5 ␮L and the flow rate was kept constant at
able tool for the detection of polar contaminants in aqueous 0.6 mL min−1 in both modes.
environmental matrices [11]. The turboionspray source settings were: Ionspray voltage,
Taking advantage of the high performance of new equipments, 5000–3500 V; curtain gas, 20 (arbitrary units); GS1 and GS2, 50 psi;
this work reports the development of a multiresidue method for the probe temperature, 500 ◦ C. Nitrogen served as nebulizer gas and
analysis of 70 pesticides from different chemical classes in paddy collision gas in both modes. Mass calibration and resolution adjust-
fields water by direct injection analysis avoiding some of the typical ments on the resolving quadrupoles were performed automatically
sample treatment steps employed in water analysis. The equipment by using a 10–5 mol L−1 solution of polypropylene glycol intro-
used in this study was a last generation hybrid triple quadrupole- duced via a syringe pump and connected to the interface.
linear ion trap (QqLIT) spectrometer. Applied Biosystems/MDS Sciex Analyst software was used for
The selected compounds for this study are the most commonly data acquisition and processing.
pesticides applied over rice fields worldwide and some of their
degradation products. Other substances included on the priority 2.2.1. Selected reaction monitoring (SRM) parameters
list of the EU such as diuron and isoproturon were also included in SRM experiments were carried out to obtain the maximum sen-
this methodology [4]. sitivity for the detection of the target molecules. The optimization
of MS parameters, declustering potential (DP) and entrance poten-
2. Materials and methods tial (EP) for precursor ions, and collision energy (CE) and collision
cell exit potential (CXP) for product ions, was performed by flow
2.1. Chemicals and reagents
injection analysis (FIA) of 1 mg L−1 of each compound separately.
Table 1 shows the optimized parameters and the selected SRM
Acetonitrile HPLC grade (MeCN) was supplied by Merck (Darm-
transitions.
stadt, Germany). Water used for LC–MS analysis was obtained
For all the compounds, the protonated molecule [M+H]+ was
from a Direct-Q5 Ultrapure Water System from Millipore (Bedford,
the most abundant, and so it was chosen as the precursor ion. After-
MA). Formic acid (purity, 98%) was obtained from Fluka (Buchs,
wards, in the product ion mode, two product ions for each pesticide
Germany). Analytical standards were purchased from Dr. Ehren-
were selected, along with their corresponding CE.
storfer GmbH (Augsburg, Germany) and Riedel-de-Haën (Seelze,
Some compounds yielded low mass product ions. This
Germany). The purity of all the standards was greater than 97%.
was the case for tebuconazole, triadimenol (70m/z), triflumi-
Individual stock standard solutions of the target compounds were
zole (72.9m/z), cyproconazole (70.1m/z), chlorotoluron, pirimicarb
prepared in pure MeOH or MeCN, according to the solubility prop-
(72.1m/z), difenoxuron, diuron and isoproturon (72m/z). Obtain-
erties of each compound, and stored at −20 ◦ C. Working solutions
ing such low masses represents a disadvantage as it entails
were prepared by an appropriate dilution of the stock solutions in
a decrease in specificity. Nevertheless these ions were cho-
MeCN and used for both procedures, the spiking and the calibra-
sen as product ions as no other higher mass were sensitive
tion curves preparation. Nicotine-d3 (Sigma–Aldrich, Steinheim,
enough.
Germany) was used as surrogate standard for positive electrospray
In the present work, two different algorithms: Standard and the
ionization mode (ESI).
Scheduled MRMTM mode [20] were compared in terms of limits of
detection (LODs), reproducibility (RSD) and number of data points
2.1.1. Selected analytes across a chromatographic peak.
The analytes included in this method were chosen on the basis The dwell time for the Standard mode was optimized and set
of previous experience and published literature concerning pes- as 5 and 100 ms for the positive and negative mode, respectively,
ticide used in paddy fields [12–16]. They comprise a group of whereas working in the Scheduled mode no dwell time is needed.
78 compounds belonging to different chemical classes such as Instead of that, the Scheduled mode monitors SRM transitions only
phenylureas, strobilurins, organophosphorous, carbamates, ureas, when they need to be monitored and not continuously thorough
triazoles, phenoxyacids, including some metabolites. Occurrence of the chromatographic run. Therefore the retention time of each
many of these compounds has been already reported in environ- pesticide must be accurately known and determined before data
mental waters [13,14,17–19]. acquisition [20,21].

2.2. Liquid chromatography–QqLIT-MS analysis 2.2.2. Information-dependent-acquisition (IDA) conditions

An IDA experiment in the positive mode was programmed com-
A hybrid Triple Quadrupole-Linear Ion Trap-Mass Spectrome- bining SRM as the survey scan and an enhanced production (EPI)
ter (5500 QTRAP® LC/MS/MS system, AB Sciex Instruments, Foster scan as the dependent scan, in the same injection. This experiment
city, CA) was used for the analysis of the target compounds. The sys- was developed for the confirmation of quinclorac, where further
tem was equipped with a turboionspray source operating in both, structural information was necessary for confirmatory purposes
positive and negative ionization modes. Chromatographic separa- due to the absence of its second transition (SRM2) at the concen-
tion was carried out using an HPLC system (Agilent Series 1200) tration range studied.
provided with a Zorbax Eclipse XDB-C-8 150 length × 4.5 mm i.d., SRM parameters used in the optimized survey scan are shown
5 ␮m particle size column (Agilent Technologies). The mobile phase in Table 1 for the positive amenable pesticides. IDA parameters
consisted of acetonitrile (solvent A) and 0.1% formic acid in water included the acquisition of one ion which peak height exceeded
(solvent B). For the positive mode the initial proportion of sol- 500 counts per second and without exclusion after dynamic back-
vent A was 10%, which was kept constant for 1 min, increased to ground subtraction of survey scan. EPI scan was performed with
100% within 15 min, kept constant for 10 min and reduced to 10% Q1 set at Low resolution and the linear ion trap scanning from 50
in 0.1 min. The run time analysis was 25 min and the post-run equi- to 270 amu at a scan rate of 10,000 amu s−1 . The dynamic fill-time
librium time was 10 min. The gradient program used in the negative option was selected on the ion trap. The CE, DP and EP used for
mode started with 20% A for 50 s, then linearly increased to 100% the EPI scan were 30 eV, 65 V and 10 V, respectively. The complete
in 5 min, which was maintained constant for 5 min and reduced SRM–IDA–EPI cycle time was 1.19 s.
4792 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798

Table 1
Optimized parameters for the QqLIT/MS analysis of the target pesticides (precursor ion, DP, product ions and their CE), SRM ratio and their corresponding variation coefficients
(% RSD).

Pesticide tR Precursor ion SRM1 CE1 SRM2 CE2 DP SRM2/SRM1a (% RSD)

Positive mode
Azimsulfuron 12.7 425.0 182.1 29 243.7 28 100 0.03 (16)
Azoxystrobin 14.4 404.2 372.0 21 343.9 33 100 0.4 (9)
Bendiocarb 12.6 223.7 166.9 13 109.0 27 100 0.9 (12)
Bensulfuron methyl 13.3 411.2 182.2 29 148.9 30 117 0.6 (6)
Bispyribac sodium 13.5 431.2 274.9 21 413.1 28 205 0.2 (18)
Carbaryl 12.8 202.1 145.0 16 127.1 43 130 0.5 (7)
Carbendazim 6.7 192.0 160.1 27 132.0 41 200 0.2 (4)
Carbofuran 12.6 222.0 165.1 18 123.1 31 91 0.9 (4)
Chlorotoluron 10.5 214.0 72.1 32 140.8 19 87 0.04 (6)
Clomazone 13.7 240.7 125.0 30 85.0 34 80 0.01 (6)
Cyhalofop butyl 16.9 357.8 256.0 17 302.2 10 117 –
Cyproconazole 14.1 292.1 70.1 58 125.0 44 170 0.5 (7)
Diazinon 16.7 304.6 169.0 30 153.1 27 100 0.7 (10)
3,4-Dichloroaniline 13.6 162.1 126.7 29 145.1 34 213 –
Diethofencarb 14.3 268.0 226.1 14 180.1 23 76 0.8 (9)
Difenoconazole 15.8 405.9 250.9 36 337.0 24 100 0.5 (5)
Difenoxuron 12.8 287.0 72.0 23 123.3 25 150 0.6(9)
Diflubenzuron 15.0 310.8 158.1 19 141.1 47 115 0.8 (12)
Dimethoathe 10.2 230.0 199.1 15 171.2 20 60 0.5 (4)
Diuron 13.0 233.0 72.0 22 160.0 35 64 0.1 (4)
Edifenphos 15.6 311.0 283.0 17 173.0 25 160 0.4 (8)
Epoxiconazole 14.4 331.0 121.0 30 141.1 28 185 –
Ethiofencarb 13.2 226.1 107.0 28 164.1 11 60 0.4 (11)
Fenobucarb 14.4 208.1 94.9 25 152.1 12 94 0.6 (7)
Fenuron 9.8 165.1 72.0 26 120.0 24 40 0.05 (6)
Flufenoxuron 17.3 489.2 158.1 32 140.6 62 165 0.2 (4)
Fluroxypir 14.7 255.0 208.8 22 237.0 16 135 0.3 (5)
Flutolanil 15.3 323.9 261.8 23 282.2 20 120 0.8 (11)
Hexaconazole 15.2 314.3 69.9 69 245.0 23 180 –
Imazamethabenz methyl 10.1 289.1 229.0 28 257.1 24 280 0.2 (12)
Imazapic 9.5 276.1 231.0 25 163.0 36 240 0.8 (8)
Imazapyr 8.5 262.9 218.1 28 235.0 22 150 0.4 (4)
Imazaquin 11.4 312.0 267.0 28 198.9 36 245 0.4 (11)
Imazosulfuron 13.3 413.1 156.1 31 231.9 24 120 0.2 (10)
Imidacloprid 9.9 256.0 175.1 27 209.1 35 118 0.7 (9)
Iprodione 15.3 330.0 244.8 22 288.0 22 165 –
Isoprocarb 13.6 194.0 94.9 18 152.1 11 60 0.4 (9)
Isoproturon 12.9 207.1 72.0 27 165.1 19 70 0.3 (12)
Kresoxim methyl 15.9 314.0 206.1 12 266.9 8 76 0.9 (9)
Malathion 15.4 331.3 127.0 19 99.0 37 52 1(10)
Metsulfuron methyl 11.9 382.2 166.7 19 141.0 35 100 0.2 (8)
Molinate 15.1 188.0 125.9 23 97.8 48 130 –
Oxydemethon methyl 7.4 247.0 168.8 16 104.9 16 140 0.3 (2)
Picoxystrobin 15.8 368.2 205.2 13 145.0 30 80 0.8 (9)
Pirimicarb 8.3 238.7 72.1 38 182.2 23 50 0.7 (11)
Pirimiphos methyl 16.8 306.1 108 42 164.1 28 110 –
Prochloraz 13.6 377.1 309.2 23 266.7 18 100 0.1 (5)
Promecarb 14.7 208.1 151.1 14 109.0 23 80 0.7 (8)
Propaphos 15.5 304.8 221.0 19 263.0 10 90 0.3 (8)
Propaquizafop 16.9 443.7 100.0 20 371.1 19 115 0.4 (4)
Propiconazole 15.5 342.2 159.0 51 187.0 25 110 0.1 (8)
Propoxur 12.5 210.0 168.1 9 111.2 12 60 0.8 (5)
Pyraclostrobin 16.2 388.2 193.7 16 295.9 20 50 0.3 (6)
Pyrazosulfuron ethyl 14.1 415.2 182.1 19 369.1 19 44 0.2 (11)
Pyridaphenthion 14.8 341.0 189.1 36 205.0 34 100 0.5 (10)
Quinclorac 11.6 242.0 224.0 21 – – 50 –
Quinoxyfen 17.1 309.1 197.9 40 272.9 40 100 0.6 (3)
Spiroxamine 11.2 298.1 144.0 25 100.1 47 240 0.5 (7)
Tebuconazole 14.8 308.1 70.1 60 125.1 57 80 0.2 (14)
Tebufenozide 15.7 353.4 296.7 10 133.2 23 60 0.1 (9)
Temephos 17.3 467.1 404.7 25 418.8 26 190 0.6 (13)
Tetraconazole 14.7 372.0 159.0 43 70.0 65 110 0.4 (14)
Thiacloprid 11.0 253.0 125.9 34 186.0 19 110 0.2 (9)
Thiamethoxam 8.9 292.1 210.9 19 246.1 14 60 0.1 (3)
Thiodicarb 11.6 355.1 87.9 29 163.2 11 160 0.2 (12)
Thiophanathe ethyl 13.4 371.0 282.0 15 324.8 13 60 0.9 (9)
Triadimefon 14.8 294.0 197.1 20 224.9 20 180 0.1 (11)
Triadimenol 13.9 296.1 70.0 40 227.1 14 50 0.1 (16)
Triazophos 13.3 314.1 162.1 25 286.0 16 120 0.1 (10)
Tricyclazole 10.0 190.1 163.1 31 136.1 40 210 0.9 (6)
Trifloxystrobin 16.9 409.2 185.7 25 206.0 20 150 0.5 (6)
Triflumizole 15.5 346.2 278.2 15 72.9 21 100 0.4 (9)
Triflumuron 15.8 359.0 156.2 27 138.8 42 40 0.5 (10)
 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798 4793

Table 1 (Continued)

Pesticide tR Precursor ion SRM1 CE1 SRM2 CE2 DP SRM2/SRM1a (% RSD)

Negative mode
2,4-D 7.5 220.7–218.7 162.8 15 160.8 34 110 0.7 (16)
Bentazone 7.6 238.8 175.2 28 197.1 29 250 0.6 (6)
Fipronil 8.4 437.1 331.0 24 319.1 32 100 0.1 (10)
Propanil 7.9 216.9 160.8 23 125.0 33 110 0.04 (12)
Teflubenzuron 8.7 379.0 339.0 16 359.0 10 110 0.4 (12)

DP, declustering potential (V); CE, collision energy (eV); EP, entrance potential (10 V); CXP, collision cell exit potential (5 V).
a
The SRM ratio is calculated from mean values obtained from the matrix-matched calibration curves.

Previously, the spectra generated in matrix solutions at 1 and results are presented in Table 1 together with their correspond-
10 ␮g L−1 concentrations, acquired in EPI mode, were stored in a ing coefficients of variation. As it is shown in Table 1 differences
mass spectral library at the CE selected (30 eV), which enables the in intensity of up to ten times between the two monitored transi-
confirmation of quinclorac in real positive samples. In this case, tions were observed for almost 18 pesticides. This is a disadvantage
confirmation criteria applied to quinclorac in the water samples when identifying these analytes, especially at low concentrations,
were: the presence of the characteristic SRM transition at the cor- were the signal-to-noise ratio for the confirmation must be higher
rect retention time, and a library search fit value higher than 70%. than three.

2.3. Sampling and sample preparation 3.2. Sample treatment and direct injection analysis

Paddy field water samples used in this study were taken Direct injection of water samples is becoming an attractive
from two different regions were different pesticides are currently procedure to traditional analytical techniques which in general
applied, namely – South America, Uruguay and Europe, Spain. A include a preliminary pre-concentration step either with LLE or
total of 59 samples were analyzed. SPE [23–25]. This technique presents many advantages, such as no
Water samples were collected in clean amber glass bottles and pre-concentration step and, as a consequence, minimum sample
stored in the dark at −20 ◦ C. After collection, the samples were manipulation, less co-extractives compounds in the final extract,
adjusted to pH 3 and filtered through a 0.7 ␮m glass fibre filter low consume of solvents, low cost and better reproducibility.
(Teknokroma, Spain), in order to remove particles that may inter- The excellent sensitivity of the equipment allowed us to study
fere with the analysis. Before analysis, 100 ␮L of a 10 ␮g L−1 labeled the performance of the direct injection of water samples for the
standard of nicotine-d3 in MeCN (surrogate standard) was added determination of pesticide residues in paddy fields water.
to 900 ␮L of water and the mixture was filtered directly into a vial Sample preparation was performed as it is described in Section
using a 0.45 ␮m PTFE syringe filter (Millipore, USA). 2.3. Based in our experience, the addition of MeCN to the water
sample before the filtration step, improves the efficiency during
3. Results and discussion the filtration process; otherwise some pesticides are partially lost
during the filtration process [26–28].
3.1. Optimization of SRM conditions For the selection of the working conditions, the effect of the pH
of the samples was studied by comparing the response of each pes-
For the SRM method two transitions per compound were ticide obtained during the analysis of a spiked water sample at pH
selected in order to comply with EU requirements for confirma- 3, 5, 7 and 8. In general no significant differences were observed
tory analysis (Commission Decision 2002/657/EC) [22]. The less for almost all the pesticides. The responses at different pH were
intense transition (SRM2) was used for the confirmation of each in the same order of magnitude except for some compounds such
analyte, while for quantitative purposes the peak area of the most as malathion, temephos, propoxur, propaphos and metsulfuron
intense transition (SRM1) was considered. The ratio between the methyl where the difference in the response between acidic or
two SRM transitions (SRM2/SRM1) was calculated in order to be basic pH was around 20%, therefore pH 3 was selected for method
used as the identification criterion along with the retention time validation [7].
and the presence of both transitions according to Ref. [22]. The
3.3. Validation study

Table 2
Validation studies were carried out using a real water sample.
Number of data points per peak and the S/N at 0.5 ␮g L−1 obtained for Scheduled
and Standard mode in water extract. As no certified pesticide-free water sample could be obtained to be
used as blank, paddy field water was used, which was previously
Pesticide name No. data points No. data points
analyzed and the presence of the target compounds considered.
per peak per peak
Standard mode Scheduled Nicotine-d3 was used as surrogate standard in order to check the
mode entire procedure.
Bensulfuron methyl 11 12
Diazinon 9 14 3.3.1. LODs, LOQs and selection of the working method
Edifenphos 9 12 In order to compare the sensitivity of Standard and MRM
Isoproturon 8 15 ScheduledTM modes, the limits of detection (LODs) were cal-
Malathion 8 13
Picoxystrobin 8 14
culated using standard solutions prepared in pure solvent
Tebuconazole 9 15 and in spiked paddy field water. The LODs were determined
Teflubenzuron 10 11 as the lowest pesticide concentration whose qualified tran-
Thiacloprid 10 15 sition (SRM2) presented a signal-to-noise ratio (S/N) ≥ 3. The
Tricyclazole 9 17
quantification limits (LOQ) were determined also in pure sol-
4794 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798

Table 3
Main validation parameters: LOD, LOQ, coefficient of determination (R2 ), reproducibility, repeatability and slope in matrix/slope in solvent ratio obtained for the developed
method.

Pesticide LOD (␮g L−1 ) LOQ (␮g L−1 ) Reproducibility (% RSDwR ) Repeatability (% RSDr ) R2 Slope in
matrix/slope in
solvent

1 ␮g L−1 50 ␮g L−1 1 ␮g L−1 50 ␮g L−1

Azimsulfuron 0.04 0.045 10.6 9.6 7.4 4.7 0.9978 1.11

Azoxystrobin 0.0004 0.003 10.1 11.8 8.4 5.0 0.9941 1.07
Bensulfuron methyl 0.003 0.006 10.1 3.2 8.6 13.9 0.9907 0.93
Bentazone 0.008 0.03 16.5 16.8 11.5 10.3 0.9974 0.70
Bispyribac sodium 0.06 0.060 18.1 15.7 7.9 7.9 0.9965 1.16
Carbaryl 0.03 0.050 16.3 12.2 5.8 5.6 0.9916 0.58
Carbendazim 0.009 0.009 17.8 13.1 5.2 2.4 0.9950 0.94
Carbofuran 0.002 0.004 16.5 14.9 6.1 4.9 0.9940 0.84
Chlorotoluron 0.07 0.075 12.2 10.5 6.8 4.7 0.9962 0.93
Clomazone 0.5 0.5000 18.3 14.6 11.6 3.3 0.9911 0.89
Cyproconazole 0.004 0.011 15.9 12.5 5.7 4.0 0.9922 1.00
Diazinon 0.001 0.002 11.0 3.2 1.5 2.5 0.9994 0.24
Diethofencarb 0.008 0.026 12.3 11.9 3.6 4.1 0.9949 0.89
Difenoconazole 0.006 0.010 16.8 9.5 1.2 3.6 0.9954 0.89
Difenoxuron 0.001 0.002 18.4 16.4 8.2 4.9 0.9940 1.01
Diflubenzuron 0.02 0.032 12.6 11.8 11.1 5.8 0.9987 0.82
Dimethoathe 0.008 0.008 17.0 14.7 3.7 3.4 0.9969 0.94
Diuron 0.01 0.015 13.5 10.8 6.4 6.8 0.9967 0.95
Edifenphos 0.005 0.011 16.6 10.1 7.4 5.2 0.9974 0.74
Ethiofencarb 0.002 0.007 11.1 10.5 5.6 4.4 0.9980 0.83
Fenobucarb 0.02 0.066 15.3 11.9 4.5 5.6 0.9972 0.90
Fenuron 0.03 0.030 13.3 9.5 5.8 1.8 0.9928 0.94
Fipronil 0.05 0.2 10.4 11.8 8.1 7.0 0.9947 0.95
Flufenoxuron 0.06 0.07 12.7 5.0 5.3 4.9 0.9938 0.70
Fluroxypir 0.02 0.04 17.6 15.0 4.2 4.7 0.9916 0.93
Flutolanil 0.004 0.005 10.0 9.1 11.6 5.5 0.9975 0.87
Imazamethabenz methyl 0.01 0.013 19.3 17.4 6.0 7.9 0.9901 0.94
Imazapic 0.01 0.090 7.5 1.9 4.1 3.4 0.9968 1.10
Imazapyr 0.007 0.017 16.2 12.1 3.5 3.0 0.9978 1.12
Imazaquin 0.005 0.005 9.3 9.0 13.2 6.4 0.9951 1.26
Imazosulfuron 0.01 0.012 14.2 12.0 7.0 7.3 0.9973 0.86
Imidacloprid 0.003 0.004 17.1 15.0 1.8 4.6 0.9959 1.10
Isoprocarb 0.03 0.032 13.9 13.7 3.7 8.8 0.9958 0.95
Isoproturon 0.06 0.060 16.3 14.7 6.2 3.6 0.9912 0.91
Kresoxim methyl 0.03 0.035 7.9 9.0 4.0 6.8 0.9973 0.92
Malathion 0.01 0.014 9.6 8.2 5.4 0.8 0.9945 0.95
Metsulfuron methyl 0.003 0.003 18.7 13.7 3.7 3.0 0.9941 1.00
Oxydemethon methyl 0.001 0.003 3.7 5.2 3.3 0.9 0.9954 0.91
Picoxystrobin 0.001 0.001 18.5 14.2 4.7 4.7 0.9951 1.00
Pirimicarb 0.001 0.001 16.3 4.9 2.1 1.6 0.9979 0.91
Prochloraz 0.04 0.055 6.4 9.0 6.9 7.8 0.9941 1.04
Promecarb 0.04 0.05 10.3 14.1 3.7 6.7 0.9935 0.92
Propanil 0.08 0.08 9.5 12.6 12.3 10.2 0.9990 0.87
Propaphos 0.003 0.005 18.4 14.2 7.4 6.3 0.998 0.92
Propaquizafop 0.03 0.03 12.0 10.3 6.2 0.6 0.9978 0.26
Propiconazole 0.04 0.04 10.1 11.0 5.9 3.1 0.9924 0.91
Propoxur 0.006 0.01 12.4 15.1 4.6 5.5 0.9973 0.82
Pyraclostrobin 0.003 0.004 10.1 12.0 3.8 0.8 0.9986 0.67
Pyrazosulfuron ethyl 0.03 0.005 14.9 7.6 5.9 10.7 0.9994 0.85
Pyridaphenthion 0.001 0.002 14.9 5.6 4.5 3.8 0.9974 0.95
Quinclorac 0.05 0.15 12.0 9.5 8.7 6.4 0.9985 1.03
Quinoxyfen 0.02 0.02 18.2 3.7 2.1 1.0 0.9980 0.83
Spiroxamine 0.02 0.04 17.3 2.6 3.5 2.9 0.9994 0.92
Tebuconazole 0.002 0.005 11.4 7.2 2.9 2.9 0.9982 0.97
Tebufenozide 0.001 0.002 14.3 8.9 7.8 4.0 0.9977 0.87
Teflubenzuron 0.01 0.03 14.3 12.7 7.3 9.2 0.9971 0.18
Temephos 0.03 0.03 11.4 12.7 2.5 3.9 0.9915 0.76
Tetraconazole 0.02 0.008 4.1 7.8 2.9 6.3 0.9978 1.03
Thiacloprid 0.002 0.001 16.6 14.2 2.7 2.1 0.9934 0.93
Thiamethoxam 0.03 0.03 3.9 4.0 3.3 1.8 0.9964 0.98
Thiodicarb 0.01 0.15 15.9 13.3 4.9 7.5 0.9968 0.80
Thiophanathe ethyl 0.008 0.02 15.7 10.1 12.6 9.3 0.9981 0.87
Triadimefon 0.02 0.02 13.1 10.0 9.1 11.8 0.9952 0.89
Triadimenol 0.03 0.03 15.2 10.9 4.2 3.0 0.9905 0.91
Triazophos 0.001 0.003 13.8 11.1 5.3 5.4 0.9954 0.89
Tricyclazole 0.001 0.002 15.4 13.8 3.7 3.8 0.9961 0.90
Trifloxystrobin 0.001 0.003 13.7 11.5 7.3 6.5 0.9985 0.96
Triflumizole 0.04 0.04 18.3 18.5 7.2 6.3 0.9987 0.90
Triflumuron 0.5 0.50 16.1 13.5 5.5 6.2 0.9985 0.86
2.4 D 0.01 0.08 6.7 8.7 10.2 8.1 0.9997 0.91

Pesticides in bold letter are those optimized in the negative mode.

 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798 4795

Fig. 1. Determination of carbendazim in paddy field water at 0.1 ␮g L−1 by both modes (Standard mode: A and Scheduled mode: B).

vent and in paddy field water, as the minimum detectable and presenting a fit value higher than 70% in the IDA experi-
amount of analyte with a S/N ≥ 10 for the SRM1 transition. ments.
The LOD of quinclorac was 0.05 ␮g L−1 calculated as the low- The criterion followed to select the working method (Scheduled
est pesticide concentration recognizable by library searching or Standard) was the method with the highest number of com-

Fig. 2. Matrix effect discriminated by % of pesticides presenting strong, medium or no matrix effect.
4796 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798

Fig. 3. Identification of tricyclazole in a paddy field water sample (A) at 1.90 ␮g L−1 and comparison with the standard in matrix (B).

pounds presenting a S/N ratio ≥ 3 at a concentration of 0.1 ␮g L−1 According with these results the best sensitivity in multiple
in matrix. The Scheduled method presented the best performance reaction monitoring mode was achieved through the acquisition
as 67 of the target pesticides presented a S/N of 3 or higher at of the selected reaction monitoring transitions (SRM) with Sched-
0.1 ␮g L−1 while the Standard method presented 41compounds uled mode, therefore the MRM ScheduledTM method was chosen
which satisfied with these criteria. Fig. 1 shows the total ion for the validation study.
chromatogram (TIC) and two extracted ion chromatograms for car- From the 78 pesticides included in the method, 7 compounds
bendazim. As it is shown in Fig. 1, the Scheduled method allows the (bendiocarb, cyhalofop butyl, 3,4-dichloroaniline, epoxiconazole,
identification and the quantitation of carbendazim, as both transi- hexaconazole, iprodione and molinate) presented LODs higher
tions presented a S/N higher than 3 and 10 corresponding to the LOD than 0.1 ␮g L−1 in the range 0.5–5 ␮g L−1 ,and pirimiphos-methyl
and LOQ, respectively, whereas in the chromatogram obtained with presented linearity problems, therefore these 8 pesticides were
Standard method the second transition (SRM2) of carbendazim is excluded from the method and the validation studies were carried
missing and therefore this analyte cannot be confirmed using this out with the other 70 pesticides (see Table 3).
method.
Overall, the LOQ values for the Scheduled method were in the 3.3.2. Linearity and matrix effect
range 2–150 ng L−1 . The linearity of the method was studied preparing a seven-point
Moreover the number of data points across a chromatographic calibration curve in paddy field water in the range from 0.1 to 50 ␮g
peak was compared for both methods. The differences between L−1 except for diazinon 0.1 to 25 ␮g L−1 . The linearity along the
the numbers of data points for both methods were not so pro- studied range was good, with correlation coefficients higher than
nounced as the LODs nevertheless for most of the pesticides the 0.99 for all target compounds as it is shown in Table 3.
Scheduled method provided a higher number of data points per Matrix effects, either as signal suppression or enhancement,
peak, 10–27 and the Standard method 7–14. In Table 2 are sum- are a major drawback for quantitative trace analysis by LC–ESI/MS
marized the number of data points per peak for some of the target systems. Matrix co-extractives can compromise the quantita-
pesticides. tive analysis of the compounds at trace levels, as well as it
 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798 4797

can greatly affect the method accuracy and reproducibility. Sev-

eral proposals have been published to overcome this problem
[29], but the most common one is the use of matrix-matched
calibration standards for the quantification of the target ana-
lytes.
The analysis of water by solid phase extraction implies a pre-
concentration step, thus matrix effect is generally observed. In this
procedure, as there is not pre-concentration step we expected lit-
tle or none matrix effect. However, in order to evaluate the extent
of the matrix effects with the presented method, matrix-matched
and solvent-based calibrations curves were prepared and the cor-
responding slope in matrix/slope in solvent ratio was calculated for
each of the studied pesticides (see Table 3). Depending on the value
(in percentage) different matrix effects could be observed. A per-
centage between −20% and 20% was considered as no matrix effect.
Signal enhancement occurs if the % of the difference between the
slopes is positive whereas a negative value is indicative of signal
suppression. Fig. 4. Occurrence and concentrations mean in ␮g L−1 of the pesticides detected in
paddy fields waters.
A medium matrix effect was observed when the values ranged
between −50% and −20% or 20–50% and a strong matrix effect
would be below −50% or above 50%. with the three criteria used for the identification of the pesti-
Eighty-seven percent of the pesticides under study did not cide.
present relevant matrix effect whereas 12% of the analytes (benta-
zone, carbaryl, diazinon, edifenphos, flufenoxuron, propaquizafop, 4. Application to real samples
pyraclostrobin and teflubenzuron) showed signal suppression.
Imazaquin was the only pesticide presenting medium signal The proposed method was applied to the analysis of 59 paddy
enhancement. Both strong and medium signal suppression were field water samples collected from different regions, 33 from Spain
observed as it represented in Fig. 2. The high proportion of ana- and 26 from Uruguay. Fig. 4 shows the occurrence and the mean
lytes showing little if any matrix effect, points out the advantage concentrations found in the samples. From 59 analyzed samples, 31
of working with high sensitivity equipment, which allows the presented pesticides. 10 samples presented 3 pesticides, 9 samples
analysis of the sample without performing a pre-concentration presented 2 pesticides and 12 other samples presented 1 pesti-
step. cide.
Based on these results the quantification could be performed Their occurrence follows the technological package applied in
using solvent-based calibration curves for 61 of the studied ana- each country, as expected. The most frequently found pesticide in
lytes, avoiding the use of matrix-matched standards curves and Uruguayan samples was the fungicide tebuconazole whereas the
hence reducing the uncertainty of the methodology. herbicide quinclorac was presented in the highest concentration,
less often the antifungal carbendazim and the insecticide imida-
cloprid. In Uruguay, carbendazim is forbidden to use in rice, so its
3.3.3. Reproducibility and repeatability presence could be due to a high persistence in soils from where it
The precision of the instrumental method, was estimated by can leach to the water. Tebuconazole, tricyclazole and bensulfuron
determining the intra- and interday, % RSD, by the repeated anal- methyl were found in the samples from Spain.
ysis (n = 5) of a spiked paddy field water at 0.1 and 50 ␮g L−1 The compounds found in the samples are pesticides widely used
level, from run-to-run, over 1 and 5 consecutive days. As it in rice crops which are included in the Annex I of the EU legislation
is shown in Table 3, the RSD ranged between 1–20% and [30], but considering the standard of 0.1 ␮g L−1 as the maximum
below 14% for the reproducibility and repeatability, respec- concentration level allowed for individual pesticides in water, the
tively. concentration of the individual pesticides found in the analyzed
samples were 1–70 times higher.
3.3.4. Specificity Quinclorac was present in four of the analyzed samples; this
In order to achieve an unambiguous identification of the stud- pesticide has been out of the Annex I since 2008, but is commonly
ied pesticides, the specificity of the method was evaluated via used in Uruguay for rice crops. As it was discussed in Section 2.2.2,
the SRM ratio of the target compounds. This ratio was calcu- its confirmation was carried out by using linear ion trap in EPI mode,
lated as the quotient between the qualifier and the quantifier because it was not possible to obtained the two transitions required
transition. The maximum permitted tolerances for relative ion for it confirmation.
intensities is from 20 to 50% [22]. In order to validate this
SRM ratio in water matrix, the SRM ratios in standard solution 5. Conclusions
were also calculated. Afterwards the SRM ratios in matrix and
in solvent were evaluated for each of the concentration levels A quick, highly sensitive and “green” liquid chromatography
of the calibration curves (0.1–50 ␮g L−1 ) and the average was ion trap spectrometric method for the separation and quantifi-
obtained. The identification criteria set for each of the presents cation of 70 pesticides in rice paddy field water was developed
compounds were very stable throughout the linearity range, with and optimized. The direct injection analysis of paddy fields water
values of RSD below 20% (Table 1). Two other identification cri- in the LC–MS/MS system is an attractive methodology as mini-
teria were taken for the identification of the pesticides in the mum sample manipulation is needed and non further clean-up
real matrices: the retention time and the presence of the two is performed as in other conventional reported techniques for
monitored transitions. In Fig. 3, the TICs and the extracted ion water analysis such as LLE and SPE. The LODs achieved with this
chromatograms (XIC) of the two monitored transitions of tri- methodology are within the range of ␮g L−1 or ng L−1 which are
cyclazole in real water and in standard are represented along enough to fulfil the most restricted regulations. As most of the pes-
4798 L. Pareja et al. / J. Chromatogr. A 1218 (2011) 4790–4798

ticides did not present matrix effect, the quantification step can [6] Y. Picó, M. Fernández, M.J. Ruiz, G. Font, J. Biochem. Biophys. Methods 70 (2007)
be performed using standard curves in solvent, avoiding the use of 117.
[7] R. Rohers, R. Zanella, I. Pizzuti, M.B. Adaime, L. Pareja, S. Niell, M.V. Cesio, H.
matrix-matched calibration curves. In this way, the precision and Heinzen, J. AOAC Int. 92 (2009) 1190.
the analysis time are dramatically improved, which is an interesting [8] N. Dujakovíc, S. Grujić, M. Radišić, T. Vasiljević, M. Laušević, Anal. Chim. Acta
issue for pesticide residue routine analysis. 678 (2010) 63.
[9] X. Ouyang, W. Zhang, J. Xu, N. Chang, C. Pang, J. Zhang, W. Niu, J. Anal. Chem.
The method usefulness was established through the con- 64 (2009) 935.
firmation of 6 different pesticides during the analysis of 59 [10] M. Kuster, M. López de Alda, D. Barceló, Mass Spectrom. Rev. 25 (2006) 900.
real samples. 31 samples presented pesticides; being tricycla- [11] L. Díaz, J. Llorca-Pórcel, I. Valor, Anal. Chim. Acta 624 (2008) 90.
[12] CVUA Stuttgart, 2007. Pesticides Online website: http://www.pesticides-
zole and tebuconazole the compounds most often found. Most
online.com (accessed January 2010).
of these samples presented a concentration of pesticides higher [13] T.C.R. Santos, J.C. Rocha, D. Barceló, J. Chromatogr. A 879 (2000) 3.
than 0.5 ␮g L−1 which is the maximum concentration allowed [14] M. Kuster, M.J. López de Alda, C. Barata, D. Raldúa, D. Barceló, Talanta 75 (2007)
390.
by the EU legislation. These results highlight the consistent
[15] J.D. Mattice, B.W. Skulman, R.J. Norman, E.E. Gbur, J. Soil Water Conserv. 65
and good sensitivity that can be achieved using the developed (2010) 12.
method. [16] A. Ferrero, A. Tinarelli, in: E. Capri, D.G. Karpouzas (Eds.), Pesticide Risk Assess-
ment in Rice Paddies: Theory and Practice, Elsevier, Amsterdam, p. 18.
[17] T. Iwafune, K. Inao, T. Horio, N. Iwasaki, A. Yokoyama, T. Nagai, J. Pestic. Sci. 35
Acknowledgements (2010) 114.
[18] T.K. Phong, K. Yoshino, K. Hiramatsu, M. Harada, T. Inoue, Paddy Water Environ.
The authors acknowledge ANII (Agencia Nacional de Investi- 8 (2010) 361.
[19] N. Mazzella, F. Delmas, B. Delest, B. Méchin, C. Madigou, J.-P. Allenou, R. Gabel-
gación e Innovación) and PEDECIBA (Programa de Desarrollo de las lec, T. Caquet, J. Environ. Monit. 11 (2009) 108.
Ciencias Básicas), AUIP (Asociación Universitaria Iberoamericana [20] Product bulletin Analyst® 1.5 Software. Available from:
de Postgrado) and AECID (Agencia Española de Cooperación Inter- http://positivespace.org/portfolio/enews/MassSpec images/Analyst1.5
ProductBulletin 114PB23-01.pdf. (accessed October 2010).
nacional para el Desarrollo), for the economical support. Instituto [21] A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 673
Nacional de Investigación Agropecuaria (INIA), Uruguay, Labora- (2010) 60.
torio Agroalimentario de la Generalitat Valenciana and Bernardo [22] Commission Decision 2002/657/EC, Off. J. Eur. Commun. L 221 (2002) 8.
[23] U. Ochsenbein, M. Zeh, J.D. Berset, Chemosphere 72 (2008) 974.
Böcking from Donistar S.C, Col. Rubio, Salto, Uruguay are greatly [24] S. Huang, T.J. Mayer, R.A. Yokley, J. Agric. Food Chem. 56 (2008) 2595.
acknowledge for the water sampling. [25] P. Speksnijder, J. van Ravestijn, P. de Voot, J. Chromatogr. A 1217 (2010) 5184.
[26] B. Kmellár, L. Pareja, C. Ferrer, P. Fodor, A.R. Fernández-Alba, Talanta (2010),
doi:10.1016/j.talanta.2010.12.006.
References [27] B. Kméllar, P. Fodor, L. Pareja, C. Ferrer, M.A. Martínez-Uroz, A. Valverde, A.R.
Fernández-Alba, J. Chromatogr. A 1215 (2008) 37.
[1] S. Barrek, C. Cren-Olivé, L. Wiest, R. Baudot, C. Arnaudguilhem, M.F. Grenier- [28] C. Ferrer, M.J. Martínez-Bueno, A. Lozano, A.R. Fernández-Alba, Talanta (2010),
Loustalot, Talanta 79 (2009) 712. doi:10.1016/j.talanta.2010.11.061.
[2] R. Araújo, V. Homem, L. Santos, A. Alves, Int. J. Environ. Anal. Chem. 90 (2010) [29] M.J. Martínez Bueno, A. Agüera, M.D. Hernando, M.J. Gómez, A.R. Fernández-
205. Alba, J. Chromatogr. A 1216 (2009) 5995.
[3] Directive 2000/60/EC, Off. J. Eur. Commun. L 327 (2000) 1. [30] Directive 91/414/EEC, Off. J. Eur. Commun. L 230 (1) (1991). Available from:
[4] Decision No. 2008/105/EC, Off. J. Eur. Commun. L 348 (2008) 84, amending http://ec.europa.eu/food/plant/protection/pesticides/community legislation
Directive 2000/60/EC. en.htm (accessed January 2010).
[5] Directive 2006/118/EC, Off. J. Eur. Commun. L 372 (2006) 19.