Available online at Journal of Chromatography A, 1175 (2007) 24 37

Transcription

1 Available online at Journal of Chromatography A, 1175 (2007) Multi-residue method for the analysis of 101 pesticides and their degradates in food and water samples by liquid chromatography/time-of-flight mass spectrometry Imma Ferrer a,, E. Michael Thurman b a Department of Analytical Chemistry, University of Almería, Almería, Spain b University of Colorado, Boulder, CO, USA Received 25 July 2007; received in revised form 24 September 2007; accepted 27 September 2007 Available online 17 October 2007 Abstract A comprehensive multi-residue method for the chromatographic separation and accurate mass identification of 101 pesticides and their degradation products using liquid chromatography/time-of-flight mass spectrometry (LC/TOF-MS) is reported here. Several classes of compounds belonging to different chemical families (triazines, organophosphorous, carbamates, phenylureas, neonicotinoids, etc.) were carefully chosen to cover a wide range of applications in the environmental field. Excellent chromatographic separation was achieved by the use of narrow accurate mass windows (0.05 Da) in a 30 min interval. Accurate mass measurements were always below 2 ppm error for all the pesticides studied. A table compiling the accurate masses for 101 compounds together with the accurate mass of several fragment ions is included. At least the accurate mass for one main fragment ion for each pesticide was obtained to achieve the minimum of identification points according to the 2002/657/EC European Decision, thus fulfilling the EU point system requirement for identification of contaminants in samples. The method was validated with vegetable samples. Calibration curves were linear and covered two orders of magnitude (from 5to500 g/l) for most of the compounds studied. Instrument detection limits (LODs) ranged from 0.04 to 150 g/kg in green-pepper samples. The methodology was successfully applied to the analysis of vegetable and water samples containing pesticides and their degradation products. This paper serves as a guide for those working in the analytical field of pesticides, as well as a powerful tool for finding non-targets and unknowns in environmental samples that have not been previously included in any of the routine target multi-residue methods Elsevier B.V. All rights reserved. Keywords: Liquid chromatography/mass spectrometry; Time-of-flight; Environmental samples; Pesticides 1. Introduction The analysis of pesticides in food and water is a major environmental concern and new instrumental techniques are constantly being sought for better detection and monitoring. One of the problems for multi-residue methods by conventional LC/MS is the decision of which pesticides should be measured. With over 600 active ingredients currently in legal use in Europe [1], one must choose analytes of interest for monitoring purposes. Recent reviews [2 4] on pesticides in food and water have commented on the unique ability of accurate mass Corresponding author. Tel.: address: iferrer@ual.es (I. Ferrer). to identify both target compounds and non-targets by liquid chromatography/time-of-flight mass spectrometry (LC/TOF- MS); thus, offering a possible solution to this conundrum. Therefore, LC/TOF-MS is a relatively new and valuable technique for the control of pesticides to ensure food safety. In this sense time-of-flight techniques can record an accurate full-scan spectrum throughout the acquisition range and have resulted in an excellent tool for the unequivocal target and non-target identification and confirmation of pesticide residues in vegetable and fruits [5,6]. One of the weaknesses of LC/TOF-MS and liquid chromatography/quadrupole time-of-flight mass spectrometry (LC/Q-TOF-MS) has been the lack of quantitative results. However, recent breakthroughs in instrument design now make LC/TOF-MS a quantitative tool [7] with mass accuracies that /$ see front matter 2007 Elsevier B.V. All rights reserved. doi: /j.chroma

2 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) are in the 1 2 ppm range, for several types of instruments when used in environmental analyses [8 14]. These changes relate to extending the linear dynamic range of the instrument by using analog-to-digital converter (ADC) rather than time-to-digital converter (TDC) [15]. Furthermore, innovations in chromatographic particle chemistry (from 5 to 3.5 or 1.8 m packings, as well as new bonding chemistries) have improved the separation of pesticides [16]. In general, official routine laboratories analyze a certain number of target compounds (ranging from 1 up to less than 50 different compounds) [16,17] depending on the legal requirements for positive identifications and the scope of the methodology used in the respective labs. The literature has hundreds of papers reporting diverse LC MS methodologies for the analysis of all the different classes of pesticide compounds. Several review papers have tried to compile all the existent information regarding mass spectrometric data (including fragment ions) using different instrumentation (ion-trap, triple quad, TOF, Q-TOF), but unfortunately, in every case, singular information is obtained depending on the method of detection used [18]. For example when using tandem mass spectrometric techniques the instrument parameters (especially the fragmentor voltage and collision energy) play an important role on the number of fragments and relative intensities obtained. For this reason, many attempts to exploit MS MS fragmentation mass libraries have failed due to the differences in instrumentation and operating conditions. However, this is not the case of time-of-flight techniques, since accurate mass measurements are specific and universal for every target analyte and do not depend on the instrumentation used. In this way, a number of publications regarding the use of accurate mass databases of pesticides have been reported recently [19,20]. Accurate mass determination allows obtaining specific information for a given molecule plus an additional confirmation if more fragments are present in the spectra. For this reason, a study containing an extensive number of compounds has been carried out in this work. This paper describes a multi-residue method for 101 commonly used pesticides, including complete information on accurate masses for the protonated molecules and fragment ions, retention times on a C 8 reversed-phase column, limits of detection and calibration curves. We have evaluated the potential of LC/TOF-MS for the quantitative analyses of pesticides in food and water samples at concentrations in the low g/l range. The proposed method for vegetable and fruit samples consists of a sample treatment step using an extraction with acetonitrile followed by quantitative analyses by LC/TOF-MS. The sample treatment applied to water samples is based on solid-phase extraction (SPE) using Sep-Pak C 18 cartridges. The method developed is sensitive for the detection of the 101 pesticides in food samples, which meets the current 0.01 mg/kg standard of the EU 91/414/EC food directive. This method will work well for accurate mass instruments since it is not instrument specific. Thus, it is highly useful for identification of at least 101 pesticides in food and water matrices. Finally, the proposed method has been successfully applied to real environmental samples including food commodities and surface water samples. 2. Experimental 2.1. Chemicals and reagents Pesticide analytical standards were purchased from both Sigma (St. Louis, MO, USA) and Chem Service (West Chester, PA, USA). Individual pesticide stock solutions (1000 g/ml) were prepared in pure acetonitrile and stored at 18 C. HPLC grade acetonitrile and methanol were obtained from Merck (Darmstadt, Germany). Formic acid was obtained from Fluka (Buchs, Switzerland). A Milli-Q-Plus ultra-pure water system from Millipore (Milford, MA, USA) was used throughout the study to obtain the HPLC-grade water used during the analyses. Anhydrous magnesium sulfate and sodium acetate were from Sigma Aldrich (Madrid, Spain). For the SPE procedure, Seppak C 18 cartridges (500 mg, 6 ml) obtained from Waters (Milford, MA, USA) were used Sample preparation Vegetable and fruit samples The QuEchERS method (acronym for quick, easy, cheap, effective, rugged and safe) was used for the extraction of food samples [21]. According to this method, a 15-g portion of food sample previously homogenized was weighted in a 200 ml PTFE centrifuge tube. Then, 15 ml of acetonitrile were added and the tube was vigorously shaken for 1 min. After this time, 1.5 g of NaCl and 4 g of MgSO 4 were added repeating then the shaking process again for 1 min to prevent coagulation of MgSO 4. The extract then was centrifuged (3700 rpm) for 1 min. A 5 ml aliquot of the supernatant (acetonitrile phase) was then taken with a pipette and transfer to a 15 ml graduated centrifuge tube, containing 250 mg of PSA (propylamino SPE cartridge; Supelco, Bellefonte, PA, USA) and 750 mg of MgSO 4, being then energetically shaken for 20 s. After this, the extract was centrifuged again (3700 rpm) for 1 min. Finally, an extract containing 1 g of sample per ml in 100% acetonitrile was obtained. The extract was then evaporated near to dryness and reconstituted to initial mobile phase composition up to 1 ml. Prior to analysis, the extract was filtered through a 0.45 m PTFE filter and transferred into a vial. Matrix extracts were used for validation of the method by appropriate spiking with the pesticide mix. The scope of this work was simply to develop a method for the screening, quantitation and confirmation of 101 pesticides in vegetable and fruit matrices, so recovery of the compounds from raw samples was not taken into account here. Vegetables and fruit samples included green-peppers, tomatoes, cucumbers and oranges Water samples An off-line SPE was used for the pre-concentration of the water samples. All the extraction experiments were performed using an automated sample preparation with extraction column system (ASPEC XL, Gilson, Villiers-le-Bel, France) fitted with an external 306 LC pump for dispensing the water samples through the SPE cartridges and with 817 switching valve for the selection of each sample. Disposable cartridge columns packed

3 26 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Table 1 LC/TOF-MS accurate masses for the protonated molecules and the main fragment ions for all the compounds studied (fragmentor voltage 190 V) Compound Retention time (min) Elemental composition a Accurate mass [M +H] + Frag ion 1 Frag ion 2 Frag ion 3 Acetamiprid 16.8 C 10 H 11 N 4 Cl Acetochlor 26.1 C 14 H 20 NO 2 Cl Alachlor 26.1 C 14 H 20 NO 2 Cl Aldicarb 18.7 C 7 H 14 N 2 O 2 S b Aldicarb sulfone 11.8 C 7 H 14 N 2 O 4 S Aldicarb sulfoxide 6.5 C 7 H 14 N 2 O 3 S Atrazine 21.4 C 8 H 14 N 5 Cl Azoxystrobin 24.3 C 22 H 17 N 3 O Benalaxyl 26.8 C 20 H 23 NO Bendiocarb 20.8 C 11 H 13 NO Bensultap 21.4 C 17 H 21 NO 4 S Bromoxynil 21.7 C 7 H 3 NOBr Bromuconazole C 13 H 12 N 3 OCl 2 Br Buprofezin 27.4 C 16 H 23 N 3 OS Butylate 29.7 C 11 H 23 NOS Captan 24.4 C 9 H 8 NO 2 SCl Carbaryl 21.3 C 12 H 11 NO Carbendazim 7 C 9 H 9 N 3 O Carbofuran 20.8 C 12 H 15 NO Cartap 3.1 C 7 H 15 N 3 O 2 S Chlorfenvinphos 26.5 C 12 H 14 O 4 PCl Chlorpyrifos methyl 28.2 C 7 H 7 NO 3 PSCl Cyanazine 19.6 C 9 H 13 N 6 Cl Cyproconazole 23.6 C 15 H 18 N 3 OCl Cyromazine 2.9 C 6 H 10 N DEET 21.3 C 12 H 17 NO Deethylatrazine 15.9 C 6 H 10 N 5 Cl Deethylterbuthylazine 19.6 C 7 H 12 N 5 Cl Deisopropylatrazine 13 C 5 H 8 N 5 Cl Diazinon 27.8 C 12 H 21 N 2 O 3 PS Dichlorvos 20 C 4 H 7 O 4 PCl Difeconazole C 19 H 17 N 3 O 3 Cl Difenoxuron 21.6 C 16 H 18 N 2 O Diflubenzuron 25.2 C 14 H 9 N 2 O 2 F 2 Cl Dimethenamide 24.3 C 12 H 18 NO 2 SCl Dimethoate 16.6 C 5 H 12 NO 3 PS Dimethomorph C 21 H 22 NO 4 Cl Diuron 21.7 C 9 H 10 N 2 OCl Ethiofencarb 21.8 C 11 H 15 NO 2 S Fenamiphos 24.1 C 13 H 22 NO 3 PS Fenuron 15.7 C 9 H 12 N 2 O Flufenacet 26.1 C 14 H 13 N 3 O 2 F 4 S Flufenoxuron 29.5 C 21 H 11 N 2 O 3 F 6 Cl Fluoroacetamide 3.1 C 2 H 4 NOF Fluroxypyr 19.2 C 7 H 5 N 2 O 3 FCl Hexaflumuron 27.5 C 16 H 8 N 2 O 3 F 6 Cl Hydroxyatrazine 12.1 C 8 H 15 N 5 O Imazalil 18.1 C 14 H 14 N 2 OCl Imazapyr 13.7 C 13 H 15 N 3 O Imazaquin 19 C 17 H 17 N 3 O Imidacloprid 16 C 9 H 10 N 5 O 2 Cl Ioxynil 23 C 7 H 3 NOI Iprodione 25.6 C 13 H 13 N 3 O 3 Cl Irgarol C 11 H 19 N 5 S Irgarol metabolite 17 C 8 H 15 N 5 S Isoproturon 21.6 C 12 H 18 N 2 O Lenacil 19.6 C 13 H 18 N 2 O Lufenuron 28.9 C 17 H 8 N 2 O 3 F 8 Cl Malathion 26 C 10 H 19 O 6 PS Mebendazole 18.4 C 16 H 13 N 3 O Metalaxyl 21.5 C 15 H 21 NO Metamitron 15.2 C 10 H 10 N 4 O Methidathion 24.1 C 6 H 12 N 2 O 4 PS Methiocarb 23.7 C 11 H 15 NO 2 S

4 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Table 1 (Continued ) Compound Retention time (min) Elemental composition a Accurate mass [M +H] + Frag ion 1 Frag ion 2 Frag ion 3 Methiocarb sulfone 17.7 C 11 H 15 NO 4 S Methomyl 12.6 C 5 H 10 N 2 O 2 S Metolachlor 25.9 C 15 H 22 NO 2 Cl Metolcarb 19.7 C 9 H 11 NO Metribuzin 20.1 C 8 H 14 N 4 OS Molinate 24.8 C 9 H 17 NOS Monuron 19.2 C 9 H 11 N 2 OCl Nicosulfuron 18.1 C 15 H 18 N 6 O 6 S Nitenpyram 12.1 C 11 H 15 N 4 O 2 Cl Oxadixyl 19.1 C 14 H 18 N 2 O Parathion ethyl 27.3 C 10 H 14 NO 5 PS Pendimethalin 30.2 C 13 H 19 N 3 O Phosmet 24.3 C 11 H 12 NO 4 PS Prochloraz 23 C 15 H 16 N 3 O 2 Cl Profenofos 28.6 C 11 H 15 O 3 PSClBr Promecarb 24.4 C 12 H 17 NO Prometon 16.6 C 10 H 19 N 5 O Prometryn 20.3 C 10 H 19 N 5 S Propachlor 22.8 C 11 H 14 NOCl Propanil 23.3 C 9 H 9 NOCl Propiconazole C 15 H 17 N 3 O 2 Cl Prosulfocarb 29 C 14 H 21 NOS Simazine 19.1 C 7 H 12 N 5 Cl Spinosad A 20.7 C 41 H 65 NO Spinosad D 21.4 C 42 H 67 NO Spiromesifen 30.7 C 23 H 30 O Spiroxamine 19.7 C 18 H 35 NO Teflubenzuron 27.9 C 14 H 6 N 2 O 2 F 4 Cl Terbuthylazine 23.8 C 9 H 16 N 5 Cl Terbutryn 20.4 C 10 H 19 N 5 S Thiabendazole 8.8 C 10 H 7 N 3 S Thiacloprid 18.3 C 10 H 9 N 4 SCl Thiocyclam 4.5 C 5 H 11 NS Thiosultap 3.2 C 5 H 13 NO 6 S Triclocarban 27.5 C 13 H 9 N 2 OCl Triflumizole 25.9 C 15 H 15 N 3 OF 3 Cl Trifluralin 30.6 C 13 H 17 N 3 O 4 F In bold the base peak ion observed in the spectrum at 190 V. a Elemental compositions correspond to the neutral molecule. b Ion corresponding to the sodium adduct [M + Na] +. with 500 mg of Seppak C 18 sorbent were used. The cartridges were conditioned with 6 ml of methanol followed by 6 ml of HPLC water at a flow rate of 1 ml/min. The water samples (100 ml) were loaded at a flow rate of 10 ml/min. Elution of the analytes from the cartridge was carried out with 3 ml of ethyl acetate. The solvent was evaporated with a stream of nitrogen to near dryness and re-dissolved in 0.3 ml of mobile phase for LC/TOF-MS analysis LC/TOF-MS analyses The separation of the selected herbicides was carried out using an HPLC system (consisting of vacuum degasser, autosampler and a binary pump) (Agilent Series 1100, Agilent Technologies, Santa Clara, CA, USA) equipped with a reversed phase C 8 analytical column of 150 mm 4.6 mm and 5 m particle size (Zorbax Eclipse XDB-C8). Column temperature was maintained at 25 C. The injected sample volume was 50 L. Mobile phases A and B were acetonitrile and water with 0.1% formic acid, respectively. The optimized chromatographic method held the initial mobile phase composition (10% A) constant for 5 min, followed by a linear gradient to 100% A after 30 min. The flow-rate used was 0.6 ml/min. A 10 min post-run time was used after each analysis. This HPLC system was connected to a time-of-flight mass spectrometer Agilent MSD TOF equipped with an electrospray interface operating in positive ion mode, using the following operation parameters: capillary voltage, 4000 V; nebulizer pressure, 40 psig; drying gas, 9 L/min; gas temperature, 300 C; fragmentor voltage, 190 V; skimmer voltage, 60 V; octopole d.c. 1, 37.5 V; octopole RF, 250 V. LC/MS accurate mass spectra were recorded across the range m/z. The data recorded was processed with Applied Biosystems/MDS-SCIEX Analyst QS software (Frankfurt, Germany) with accurate mass application-specific additions from Agilent MSD TOF software. Accurate mass measurements of each peak from the total ion chromatograms were obtained by means of an automated calibrant delivery system using a dual-nebulizer ESI source that introduces the flow

5 28 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) from the outlet of the chromatograph together with a low flow of a calibrating solution (calibrant solution A, Agilent Technologies), which contains the internal reference masses (purine (C 5 H 4 N 4 at m/z and HP-921 [hexakis-(1h,1h,3htetrafluoro-pentoxy)phosphazene] (C 18 H 18 O 6 N 3 P 3 F 24 )atm/z The instrument worked providing a typical resolution of 9700 ± 500 (m/z 922). 3. Results and discussion 3.1. LC/TOF-MS separation and detection of 101 pesticides The pesticides included in this study were selected among different classes of compounds (triazines, organophosphates, carbamates, phenylureas, etc.) and several chemical uses (insecticides, herbicides and fungicides). Most of these compounds are currently analyzed by hundreds of laboratories performing target analysis of pesticides in both food and water samples, and for this reason they were included in this study. Table 1 compiles the chemical formulae and exact accurate masses obtained by TOF-MS, as well as the retention times for all the pesticides analyzed in this study. Of the 101 pesticides, 76 presented an [M +H] + peak as a base peak in the spectrum (base peak ions are marked in bold in Table 1). Surprisingly, 25 pesticides did not present the protonated molecule as a main base peak in the spectrum in spite of the low fragmentor voltage used; in all these cases the larger ion was a fragment ion. Only one compound (aldicarb) presented a sodium adduct as a base peak and in only one case (cartap) both the protonated molecule and the sodium adduct were absent, only two fragments showed up in the spectrum in this particular case. Some of the most usual detected degradation products in environmental samples were also included in this study (e.g. degradation products for atrazine, aldicarb, etc.) for more complete and detailed information. A linear gradient starting with 10% acetonitrile up to 100% in 30 min was applied, which was first developed by our group [8] and had proven to be successful for the separation of a wide variety of pesticide compounds. Fig. 1 shows the total ion chromatogram for the 101 pesticides analyzed. As it can be observed in Table 1 from the retention times, the majority of compounds elute in a 10 min time window comprised between 16 and 26 min, mainly due to the similarity in polarity among the pesticides studied. Nevertheless, good chromatographic separation was obtained for all the compounds by using extracted narrow mass windows of 0.05 Da Structural characterization of the analytes Accurate mass of fragment ions The fragmentor voltage role in LC MS is critical to obtain structural information of the target analytes, as well as a way to get the best balance between sensitivity and fragmentation. For this reason, the fragmentor voltage was increased to obtain additional information from characteristic fragments of the compounds. Every compound was studied separately (single Fig. 1. Total ion chromatogram (TIC) corresponding to the analysis of a mix of 101 pesticides (0.1 g/ml) by LC/TOF-MS.

6 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Fig. 2. (a) Extracted ion chromatogram (XIC) corresponding to the analysis of a blank green-pepper sample where the banned pesticide nitenpyram was detected. (b) Spectrum of nitenpyram showing the characteristic isotopic chlorine signature. Fig. 3. (a) Total ion chromatogram (TIC) corresponding to the analysis of a spiked tomato sample with the studied pesticides (0.05 mg/kg) by LC/TOF-MS. (b) Extracted ion chromatograms (XICs) corresponding to some protonated molecules (mass window 0.05 Da).

7 30 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) compound injections were carried out at different fragmentor voltages) in order to obtain specific information on the fragments obtained, and accurate mass was used to determine the particular fragment for each compound. Among the 101 pesticides, 96 of them clearly showed at least one fragment ion at a medium fragmentor voltage of 190 V (Table 1). About half of the pesticides (49) presented at least two fragment ions, and a smaller number of pesticides (12) showed as much as three fragment ions. Some of these fragments are the base peak ions in the corresponding spectra, so it is important to account for them when carrying out quantitation in order to achieve maximum sensitivity. Most of these fragments have been reported by other studies using tandem mass spectrometry techniques, so all the results obtained here match in every case and they demonstrate that time-of-flight without MS MS can be used as an identification tool using fragments from the in-source collision induced dissociation. In addition to fragmentation we obtain accurate mass information for every specific fragment that is highly useful for unequivocal identification. For example, in some cases the exact formula and hence the accurate mass of the protonated molecule was identical for four pairs of compounds: alachlor/acetochlor, deethylterbuthylazine/simazine, ethiofencarb/methiocarb, and prometryn/terbutryn. In all these cases, the specific information on their fragment ions, which are different, is essential to tell both compounds apart and to make a correct identification, especially if retention time is close. Thus, if necessary, the fragmentor voltage may be increased to get enhanced sensitivity for the fragment ion for a positive confirmation of the analyte Accurate mass of isotopes Additional information can be obtained for those compounds containing elemental isotopes, such as chlorine, bromine or sulfur. In these cases, the accurate masses for these isotopic signals are obtained and offer an extra added identification point for confirmatory purposes [13]. An example is shown in Fig. 2 for the identification of nitenpyram, a non-authorized pesticide, in a blank of green-pepper sample that was found to contain this insecticide. As it can be seen in this figure, the chlorine isotopic signal is obtained and the accurate mass of the chlorine 37 isotopes can be measured with a very small error. It is important to note that only 30 out of the 101 pesticides studied did not present an A + 2 isotopic signal, these were the compounds that contained mainly C, H, O and N. For the rest of pesticides (70%), the accurate mass value of the A + 2 ion is highly useful for the correct identification of the analyte as shown in Fig. 2 and it should be used as a tool for identification. In summary, the accurate mass analysis of the protonated molecule together with that of additional characteristic fragment ion(s) (including characteristic isotopic signals and retention times) enables the unambiguous identification and confirmation of the studied pesticides at low concentration levels. This fits the requirements of the EU according to the identification point system [22] Analytical performance To evaluate the usefulness of LC/TOF-MS for quantitative analyses in vegetable matrices, the analytical performance of the Fig. 4. Quantitation window showing some extracted ion chromatograms (XICs) corresponding to the base peak ion for 12 selected compounds (mass window 0.1 Da).

8 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Table 2 Calibration data, correlation coefficients and instrument LODs for all the analytes studied in a green-pepper matrix sample Compound Calibration curve R 2 LODs ( g/kg) Acetamiprid y = C Acetochlor y = C Alachlor y = C Aldicarb y = C Aldicarb sulfone y = C Aldicarb sulfoxide y = C Atrazine y = C Azoxystrobin y = C Benalaxyl y = C Bendiocarb y = C Bensultap y = C a Bromoxynil y = C Bromuconazole y = C Buprofezin y = C Butylate y = C Captan y = C Carbaryl y = C Carbendazim y = C Carbofuran y = C Cartap y = C Chlorfenvinphos y = C Chlorpyrifos methyl y = C Cyanazine y = C Cyproconazole y = C Cyromazine y = C DEET y = C Deethylatrazine y = C Deethylterbuthylazine y = C Deisopropylatrazine y = C Diazinon y = C Dichlorvos y = C Difeconazole y = C Difenoxuron y = C Diflubenzuron y = C Dimethenamide y = C Dimethoate y = C Dimethomorph y = C Diuron y = C Ethiofencarb y = C Fenamiphos y = C Fenuron y = C Flufenacet y = C Flufenoxuron y = C Fluoroacetamide y = C Fluoroxypyr y = C Hexaflumuron y = C Hydroxyatrazine y = C Imazalil y = C Imazapyr y = C Imazaquin y = C Imidacloprid y = C Ioxynil y = C Iprodione y = C Irgarol 1051 y = C Irgarol metabolite y = C Isoproturon y = C Lenacil y = C Lufenuron y = C Malathion y = C Mebendazole y = C Metalaxyl y = C Metamitron y = C Methidathion y = C Methiocarb y = C

9 32 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Table 2 (Continued ) Compound Calibration curve R 2 LODs ( g/kg) Methiocarb sulfone y = C Methomyl y = C Metolachlor y = C Metolcarb y = C Metribuzin y = C Molinate y = C Monuron y = C Nicosulfuron y = C Nitenpyram y = C Oxadixyl y = C Parathion ethyl y = C Pendimethalin y = C Phosmet y = C Prochloraz y = C Profenofos y = C Promecarb y = C Prometon y = C Prometryn y = C Propachlor y = C Propanil y = C Propiconazole y = C Prosulfocarb y = C Simazine y = C Spinosad A y = C Spinosad D y = C Spiromesifen y = C Spiroxamine y = C Teflubenzuron y = C Terbuthylazine y = C Terbutryn y = C Thiabendazole y = C Thiacloprid y = C Thiocyclam y = C Thiosultap y = C a Triclocarban y = C Triflumizole y = C Trifluralin y = C a Bensultap and thiosultap degraded in the standard solutions due to hydrolysis of the molecule. proposed method was studied and validated in terms of linearity, limits of detection and reproducibility of the technique for food commodities Quantitation by LC/TOF-MS Quantitation of the sample extracts was accomplished using a calibration curve based on matrix-matched standards: blank sample extracts from vegetable and fruits were evaporated until near dryness under a nitrogen flow and then reconstituted with the 101 pesticide mix standard solution at different concentrations ranging from to 0.5 mg/kg in order to have a wide range of concentrations. The use of matrix-matched standards provides reliable quantitation capabilities for food pesticide analysis [8]. Fig. 3 shows the total ion chromatogram of a tomato-matched standard spiked at 0.05 mg/kg with the mixture of the 101 pesticides as well as ion extracted chromatograms for some selected pestides. Analytes can be easily distinguished among the matrix by the use of narrow accurate mass windows as shown in this figure. Quantitation was performed by measuring the peak area of the base peak ion of each analyte (numbers in bold in Table 1). Fig. 4 shows an example of the quantitation wizard for 12 compounds. As it can be seen in this figure, the extracted ion chromatograms are automatically obtained and data are compared to the retention time of each analyte, based on a previous standard injection. If two or more peaks are present in the chromatogram, the software assigns as positive the closest peak to the correct retention time for the analyte under study. For example, for aldicarb, two peaks are obtained: one at 6.7 min and the other one at 18.8 min for the extracted ion at m/z The software assigns (in blue matching in the figure) aldicarb to the peak at 18.8 min based on a previous standard injection. Aldicarb sulfoxide, a metabolite of aldicarb, which presents obviously the same fragment ion at m/z , gets assigned at 6.7 min. Another interesting case is the chromatographic coelution of acetochlor and alachlor at 26.2 min, both having the same protonated molecule at m/z as mentioned before. Fig. 4 shows again how acetochlor and alachlor can be differenciated by their respective fragment ions at m/z and (which are base peak ions in their respective spectra in both cases). Interestingly, the peak next to acetochlor at 26.9 corresponds to one of the fragments of benalaxyl (m/z = ) which elutes at

10 Table 3 Typical diagnostic ions and accurate mass of several pesticide families I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Pesticide family Compound Diagnostic ion Accurate mass of diagnostic ion Phenylureas Diuron Fenuron Isoproturon Monuron Organophosphates Chlorpyrifos methyl Dimethoate Malathion Triazines Deethylatrazine Deethylterbuthylazine Deisopropylatrazine Atrazine and Terbuthylazine Fluorobenzoylureas Diflubenzuron Flufenoxuron Hexaflumuron Lufenuron Teflubenzuron Neonicotinoids Acetamiprid Thiacloprid Conazole fungicides Bromuconazole Propiconazole Imazalil this retention time and happened to be extracted with the 148 ion Calibration curves Linearity was studied in both solvent and matrix-matched standard solutions of green pepper at five different concentration levels. Quantitation was carried out using the peak area from the extracted ion chromatograms (XIC) of the base peak ion (in bold in Table 1) using a mass window of 0.05 Da. Table 2 shows the calibration equations obtained for the 101 pesticides in greenpepper matrices and their correlation coefficients. As it can be observed, the linearity of the analytical response within the studied range of two orders of magnitude is good, with correlation coefficients equal or higher than 0.99 in all cases Limits of detection and reproducibilty The instrument limits of detection (LODs) were estimated from the injection of matrix-matched standard solutions with low concentration levels giving a signal-to-noise ratio of 3. The results are summarized in Table 2 as well. It should be pointed out that the LODs were as low as 0.04 g/kg in the case of benalaxyl. The average values are about 3 g/kg, which is enough to meet the 10 g/kg standard (Directive 91/414/EC) established for pesticides in fruits and vegetables [23]. Only few compounds showed higher LODs due to their low response under electrospray conditions (bromoxynil, captan, chlorpyrifos-methyl, fluoroacetamide, fluoroxypyr, spiromesifen, teflubenzuron and trifluralin). Bensultap and thiosultap were found to be degraded in water solutions due to a possible hydrolysis of the standards

11 34 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Fig. 5. Extracted ion chromatogram for m/z (from m/z to 72.07, mass window 0.05 Da) corresponding to the analysis of an orange matrix-matched standard spiked with the mixture of 101 pesticides. Peaks: (1) fenuron, (2) monuron, (3) isoproturon and (4) diuron. made in water. For the rest of compounds, the signal-to-noise ratios were good, thus illustrating the high sensitivity and suitability of LC/TOF-MS for trace analysis of pesticides in environmental matrices. The LODs for water samples were similar to the ones obtained in food commodities (results not shown here). The reproducibility, repeatability and accuracy of the method were also evaluated on matrix-matched solutions at two different concentration levels: 0.01 and 0.1 mg/kg. The RSD (n = 5) values for intra-day analyses were in the range 0.9 4% and the RSD for inter-day (n = 5) values were between 3.5 and 9% Potential application to non-target pesticides From data compiled in Table 1 one can extrapolate some useful information referent to fragment ions. Depending on the family of pesticides (triazines, phenylureas, organophosphates, etc.) a trend is observed for fragmentation ions present in their Fig. 6. (a) Total ion chromatogram corresponding to the LC/TOF-MS analysis of an orange sample where imazalil was detected. (b) Extracted ion chromatogram of imazalil at m/z 297 (inset: accurate mass spectrum).

12 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) Table 4 LC/TOF-MS accurate mass measurements for the protonated molecules and main fragment ions of positive findings in a surface water sample Compound Ion Elemental composition m/z theoretical m/z experimental Error mda ppm Atrazine [M +H] + C 8 H 14 ClN Frag. Ion C 5 H 9 ClN DEET [M + H] + C 12 H 17 NO Frag. Ion C 12 H 17 NO Deethylatrazine [M +H] + C 6 H 10 ClN Frag. Ion C 3 H 5 ClN Deisopropylatrazine [M +H] + C 5 H 8 ClN Frag. Ion C 3 H 5 ClN Diazinon [M +H] + C 12 H 21 N 2 O 3 PS Frag. Ion C 8 H 13 N 2 S Dimethenamide [M +H] + C 12 H 18 ClNO 2 S Frag. Ion C 11 H 15 ClNOS Diuron [M +H] + C 9 H 10 Cl 2 N 2 O Frag. Ion C 3 H 6 NO Metolachlor [M +H] + C 15 H 22 ClNO Frag. Ion C 14 H 19 ClNO Prometon [M +H] + C 10 H 19 N 5 O Frag. Ion C 7 H 14 N 5 O Simazine [M +H] + C 7 H 12 ClN Frag. Ion C 5 H 9 N 35 5 Cl Fig. 7. (a) Total ion chromatogram corresponding to the LC/TOF-MS analysis of a surface water sample where DEET was detected. (b) Extracted ion chromatogram of DEET (inset: accurate mass spectrum).

13 36 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) respective spectra. For example, almost all the organophosphate pesticides show the diagnostic ion at m/z corresponding to the elemental formula of C 2 H 6 O 2 PS. Phenylureas diagnostic ion is primarily m/z corresponding to the C 3 H 6 NO moiety. Other examples are shown in Table 3 for triazines, fluorobenzoylureas, neonicotinoids and conazole fungicides. This information is highly useful for the identification of non-target compounds that have not been included in a multi-residue method but that they may present a common fragment ion for identification. For example in Fig. 5 an extracted chromatogram for the ion at m/z is shown corresponding to an orange matrix-matched standard spiked with the 101 mixture. All the phenylurea compounds listed in Table 3 are thus clearly identified with this ion, showing the potential of this tool for positive identification of compounds belonging to the same family or degradates. If a new peak shows up in the chromatogram then a new identification could be made by taking a look at the spectrum obtained for such peak and assigning elemental formula information to the corresponding protonated molecule. Moreover, if the mass of a fragment ion shows up at a different retention time than the protonated molecule this may indicate the presence of a possible degradation product that either presents the same mass or contains that fragment ion [24] Analysis of food and water samples To evaluate the application of the proposed methodology, it was applied to the analysis of real samples including food commodities and surface water. All the samples were extracted and analyzed as described in the experimental part and some examples are described as follows Food samples The methodology was applied to the analysis of several fruits and vegetables (green-peppers, tomatoes, cucumbers and oranges). About 82% of the samples analyzed (a total of 20) showed at least two pesticides. Only 15% of the positive samples presented pesticides at concentrations higher than the MRL. An example is shown in Fig. 6 where the identification and confirmation of imazalil is carried out in an orange sample. The accurate mass of the chlorine isotope is used for additional information and it matches with the theoretical exact mass. Imazalil has been reported by other authors [25] pointing out the importance of carrying out the identification of fungicides in citrus fruits Surface water samples A total of four surface water samples from Kansas (USA) were analyzed with the methodology described in this paper and the comprehensive screening for the 101 pesticides was carried out. All the samples gave positive findings for several pesticides. As an example, Fig. 7 shows the total ion chromatogram and the extracted ion chromatogram for DEET (meta-n,ndiethyltoluamide) at 21.3 min; the corresponding spectrum is also shown. As it can be observed in the spectrum, the accurate mass of the protonated molecule (m/z ) presents an error of only 0.5 ppm as well as its main fragment ion at m/z does. DEET is used as a repellent for mosquitoes and other related insects and it is one of the banned insecticides in some countries in Europe, although it is an authorized pesticide in the USA. Positive findings for other nine pesticides were confirmed in the same surface water sample with excellent mass accuracies (<3 ppm) (see Table 4). Triazine compounds such as atrazine and simazine were identified in this sample along with the main degradation products, deethylatrazine and deisopropylatrazine. The other compounds detected were also herbicides which are widely used in corn and soybean fields in this geographic area, thus further illustrating the usefulness and reliability of LC/TOF-MS for the analysis of pesticides in environmental samples [26]. 4. Conclusions A study to evaluate the potential of LC/TOF-MS for identification and confirmation of pesticides in environmental samples was carried out. The developed method allows the screening of 101 pesticides in vegetables and water samples. The LODs obtained with this method are in compliance with the regulations on food established by the EU. The applicability of the method was demonstrated by analysis of real samples (food and water) showing excellent selectivity and sensitivity, thus making possible the unambiguous identification of the selected pesticides. Acknowledgements This work was supported by the Ministerio de Ciencia y Tecnología (Spain, Contract: AGL /ALI). I. F. acknowledges the research contract (Contrato de retorno de Investigadores) from the Consejería de Educación y Ciencia de la Junta de Andalucía, Spain. Dr. Mike Meyer from the US Geological Survey in Lawrence, Kansas (USA) is acknowledged for surface water samples supply. References [1] The Pesticide Manual, 14th ed., British Crop Production Council (BCPC) Publications, Hampshire, UK, [2] I. Ferrer, E.M. Thurman, TrAC-Trends Anal. Chem. 22 (2003) 750. [3] Y. Picó, Mass Spectrom. Rev. 25 (2006) 837. [4] S.D. Richardson, Anal. Chem. 79 (2007) [5] I. Ferrer, E.M. Thurman, A.R. Fernandez-Alba, Anal. Chem. 77 (2005) [6] E.M. Thurman, I. Ferrer, A. Fernandez-Alba, J. Chromatogr. A 1067 (2005) 127. [7] N.L. Williamson, M.G. Bartlett, Biomed. Chromatogr. 21 (2007) 567. [8] I. Ferrer, J.F. García Reyes, M. Mezcua, E.M. Thurman, A.R. Fernandez- Alba, J. Chromatogr. A 1082 (2005) 81. [9] S. Grimalt, O.J. Pozo, J.V. Sancho, F. Hernández, Anal. Chem. 79 (2007) [10] A. Kaufmann, P. Butcher, Rapid Commun. Mass Spectrom. 20 (2006) [11] J.V. Sancho, O.J. Pozo, M. Ibanez, F. Hernández, Anal. Bioanal. Chem. 386 (2006) 987. [12] O. Nuñez, E. Moyano, M.T. Galceran, J. Mass Spectrom. 39 (2004) 873. [13] E.M. Thurman, I. Ferrer, J.A. Zweigenbaum, Anal. Chem. 78 (2006) [14] S. Ojanpera, A. Pelander, M. Pelzing, I. Krebs, E. Vuori, I. Ojanpera, Rapid Commun. Mass Spectrom. 20 (2006) 1161.

14 I. Ferrer, E.M. Thurman / J. Chromatogr. A 1175 (2007) [15] J. Fjeldsted, in: Time-of-flight mass spectrometry, technical overview, agilent application note, EN, Agilent, USA, [16] C.C. Leandro, P. Hancock, R.J. Fussell, B.J. Keely, J. Chromatogr. A 1103 (2006) 94. [17] M. Hiemstra, A. Toonen, A. De Kok, J. AOAC Int. 82 (1999) [18] C. Soler, Y. Pico, Trends Anal. Chem. 26 (2007) 103. [19] E.M. Thurman, I. Ferrer, O. Malato, A.R. Fernandez-Alba, Food Addit. Contam. 23 (2006) [20] I. Ferrer, A. Fernandez-Alba, J.A. Zweigenbaum, E.M. Thurman, Rapid Commun. Mass Spectrom. 20 (2006) [21] S.J. Lehotay, K. Maštovská, A.R. Lightfield, J. AOAC Int. 88 (2005) 615. [22] M. Ibañez, J.V. Sancho, O.J. Pozo, F. Hernández, Anal. Chem. 76 (2004) [23] Regulation (EC) No. 396/2005 of the European Parliament and of the council of 23 February 2005 on maximum residue levels of pesticides in or on food and feed of plant and animal origin, amending Council Directive 91/414/EC. [24] E.M. Thurman, I. Ferrer, J.A. Zweigenbaum, J.F. Garcia-Reyes, M. Woodman, A.R. Fernandez-Alba, J. Chromatogr. A 1082 (2005) 71. [25] T. Zamora, O.J. Pozo, F. Lopez, F. Hernandez, J. Chromatogr. A 1045 (2004) 137. [26] E.M. Thurman, I. Ferrer, R. Parry, J. Chromatogr. A 957 (2002) 3.