Desorption Electrospray Ionization Mass Spectrometry for Trace

Dec 17, 2008 - Juan F. Garcıa-Reyes,†,‡ Ayanna U. Jackson,† Antonio Molina-Dıaz,‡ and R. Graham Cooks*,†. Department of Chemistry, Purdue ...
0 downloads 0 Views 482KB Size
Anal. Chem. 2009, 81, 820–829

Desorption Electrospray Ionization Mass Spectrometry for Trace Analysis of Agrochemicals in Food Juan F. Garcı´a-Reyes,†,‡ Ayanna U. Jackson,† Antonio Molina-Dı´az,‡ and R. Graham Cooks*,† Department of Chemistry, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907, and Analytical Chemistry Research Group, Department of Physical and Analytical Chemistry, University of Jae´n, 23071 Jae´n, Spain Desorption electrospray ionization (DESI) is applied to the rapid, in situ, direct qualitative and quantitative (ultra)trace analysis of agrochemicals in foodstuffs. To evaluate the potential of DESI mass spectrometry (MS) in toxic residue testing in food, 16 representative multiclass agricultural chemicals (pesticides, insecticides, herbicides, and fungicides) were selected (namely, ametryn, amitraz, azoxystrobin, bitertanol, buprofezin, imazalil, imazalil metabolite, isofenphos-methyl, malathion, nitenpyram, prochloraz, spinosad, terbuthylazine, thiabendazole, and thiacloprid). The DESI-MS experiments were performed using 3 µL of solution spotted onto conventional smooth poly(tetrafluoroethylene) (PTFE) surfaces, with examination by MS and tandem mass spectrometry (MS/MS) using an ion trap mass spectrometer. Optimization of the spray solvent led to the use of acetonitrile/water (80:20) (v/v), with 1% formic acid. Most of the compounds tested showed remarkable sensitivity in the positive ion mode, approaching that attainable with conventional direct infusion electrospray mass spectrometry. To evaluate the potential of the proposed approach in real samples, different experiments were performed including the direct DESI-MS/MS analysis of fruit peels and also of fruit/vegetable extracts. The results proved that DESI allows the detection and confirmation of traces of agrochemicals in actual market-purchased samples. In addition, MS/MS confirmation of selected pesticides in spiked vegetable extracts was obtained at absolute levels as low as 1 pg for ametryn. Quantitation of imazalil residues was also undertaken using an isotopically labeled standard. The data obtained were in agreement with those from the liquid chromatography mass spectrometry (LC-MS) reference method, with relative standard deviation (RSD) values consistently below 15%. The results obtained demonstrate the sensitivity of DESI as they meet the stringent European Union pesticide regulation requirements (maximum residue levels) for a large percentage of the studied compounds. Pesticide testing in foodstuffs is a societally relevant and challenging application of mass spectrometry, which requires simultaneous trace analysis for a large range of agrochemicals * To whom correspondence should be addressed. Phone: 765-494-5262. Fax: 765-494-9421. E-mail: [email protected]. † Purdue University. ‡ University of Jae´n.

820

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

belonging to a wide variety of compound classes. Over 800 different compounds are applied to agricultural crops in order to control undesirable molds, insects, or weeds.1 Since the presence of trace amounts of agricultural chemical residues and their degradation products could represent potential health hazards (because of their toxicity or carcinogenicity), these substances are controlled. To ensure the safety of food for consumers, numerous regulations such as the European Union directives have set maximum residue limits (MRLs) for pesticides in food.2,3 At this time, liquid chromatography tandem mass spectrometry (LC-MS/MS) performed using triple quadrupole instruments, gas chromatography mass spectrometry (GC/MS), and GC/MS/ MS have become the most widely used methods for the analysis of target pesticides.4 These techniques are usually exhaustive and provide comprehensive data. However, they require considerable sample manipulation including extraction and chromatographic separation, and as a consequence take considerable time and effort.4 In addition, there is no possibility of performing pesticide testing in the field, which could be an interesting alternative, saving time and avoiding the additional effort involved in the chain of custody of samples that are analyzed in the laboratory. The direct analysis of foodstuffs without any kind of sample treatment would also be an interesting feature not provided so far by any of the techniques used for pesticides or related compounds. For these reasons, there is a great potential value in the development of new fast screening methodologies that can deliver reliable qualitative information on the presence/absence of particular chemicals and provide rapid, accurate, and relatively inexpensive analyses with high throughput as well as field portability. So far, several technologies have been proposed for rapid pesticide testing in food. Among them, inmmunoanalytical techniques have been extensively described.5 These techniques are versatile in application and can be formatted to suit different purposes such as quantitative analysis or simple “yes/no” tests that are field-portable. Other methodologies used are based on (1) Tomlin, C. The Pesticide ManualsA World Compendium, 13th ed.; British Crop Protection Council (BCPC): Hampshire, U.K., 2003. (2) Council Directive of 15 July 1991 (91/414/EEC) concerning the placing of plant protection products on the market (OJ L 230 19.8.1991).p1. (3) Plant Protection-Pesticide Residues-Regulation (EC) No. 396/2005. Available at: http://ec.europa.eu/food/plant/protection/pesticides/ regulation_ec_396_2005_en.htm (accessed December 2008). (4) Ferna´ndez-Alba, A. R., Ed. Chromatographic-Mass Spectrometric Food Analysis for Trace Determination of Pesticide Residues; Comprehensive Analytical Chemistry, Vol. XLII; Elsevier: Amsterdam, The Netherlands, 2005. (5) Rodrı´guez-Mozaz, S.; Lo´pez de Alda, M. J.; Barcelo´, D. J. Chromatogr., A 2007, 1152, 97–115. 10.1021/ac802166v CCC: $40.75  2009 American Chemical Society Published on Web 12/17/2008

electrochemical enzyme inhibition assays using screen-printed electrodes.6 Techniques based on fluorescence measurements have been proposed, for example, using gold nanoparticles.7 Novel mass spectrometric techniques have also been proposed. Atmospheric pressure glow discharge desorption mass spectrometry and silicon-nanoparticle-assisted laser desorption/ionization mass spectrometry have been reported for pesticide testing in food and in the environment.8,9 Recently, a new family of techniques has emerged that allows ions to be created from condensed phase samples under ambient conditions and then collected and analyzed by MS. This innovation in mass spectrometry, so-called ambient ionization mass spectrometry,10-12 allows the acquisition of mass spectra on ordinary samples, in their native environment, without sample preparation or preseparation by creating ions from surfaces outside the instrument. Of these methods, desorption electrospray ionization (DESI)13,14 has proven to be a widely applicable ionization technique which allows rapid (under 10 s per sample) in situ analysis with either minimal or without any sample pretreatment, while retaining high molecular specificity and broad applicability.13,14 In DESI, a charged, high-velocity spray of microdroplets is directed toward a surface of interest, and secondary droplets which include the species of interest are transferred through air to the atmospheric pressure interface of a mass spectrometer where solvent evaporation yields the ionized compound(s). Due to its speed, ease of use, and high salt tolerance, DESI is becoming a useful tool in a wide variety of applications such as the analysis of pharmaceuticals, intact bacteria, tissues, lipids, urine, drugs of abuse, steroids, explosives, chemical warfare agents, and agricultural chemicals.10-15 To evaluate the potential of the proposed approach in food analysis, 16 compounds belonging to different classes of agrochemicals (insecticides, herbicides, and fungicides) have been selected as representative of the agrochemicals used worldwide (Table 1). Previous initial experiments on DESI-MS of pesticides revealed the potential usefulness of this approach for atrazine, alachlor, and N,N-diethyl-toluamide (DEET) tested in water and on leaf material.16,17 In this work, a much larger set of compounds and a more thorough and systematic study were completed. Different experiments were performed including direct DESI-MS/ MS analysis of fruit peels and also of foodstuff extracts, the latter (6) Del Carlo, M.; Mascini, M.; Pepe, A.; Diletti, G.; Compagnone, D. Food Chem. 2004, 84, 651–656. (7) Dasary, S. S. R.; Rai, U. S.; Yu, H.; Anjaneyulu, Y.; Dubey, M.; Ray, P. C. Chem. Phys. Lett. 2008, 460, 187–190. (8) Jeckin, M. C.; Gamez, G.; Touboul, D.; Zenobi, R. Rapid Commun. Mass Spectrom. 2008, 22, 2791–2798. (9) Wen, X.; Dagan, S.; Wysocki, V. H. Anal. Chem. 2007, 79, 434–444. (10) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566–1570. (11) Venter, A.; Nefliu, M.; Cooks, R. G. Trends Anal. Chem. 2008, 27, 284– 290. (12) Harris, G. A.; Nyadong, L.; Fernandez, F. M. Analyst 2008, 133, 1297– 1301. (13) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471–473. (14) Takats, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 1261– 1275. (15) Van Berkel, G. J.; Pasilis, S. P.; Ovchinnikova, O. J. Mass Spectrom. 2008, 43, 1161–1180. (16) Mulligan, C. C.; Talaty, N.; Cooks, R. G. Chem. Commun. 2006, 1709– 1711. (17) Mulligan, C. C.; MacMillan, D. K.; Noll, R. J.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2007, 21, 3729–3736.

obtained using the conventional sample treatment protocol used worldwide for pesticide analysis. Identification and confirmation capabilities and sensitivity were examined in detail using different spray solvents, and the optimized method was applied to a wide range of actual market-purchased samples. Finally, quantitative performance including accuracy were also evaluated. EXPERIMENTAL SECTION Chemicals and Reagents. Pesticide analytical standards were purchased from Dr. Ehrenstorfer GmbH. (Ausburg, Germany) and from Riedel de Hae¨n, Pestanal quality (Seelze, Germany). Individual pesticide stock solutions (200-300 µg mL-1) were prepared in methanol and stored at -20 °C. Working solutions were prepared by appropriate dilution with acetonitrile. Isotopically labeled imazalil-d5 (2-propenyl-d5) 100 ng µL-1 acetone solution was obtained from Dr. Ehrenstorfer GmbH. A MilliQ-Plus ultrapure water system from Millipore (Bedford, MA) was used throughout the study to obtain the HPLC-grade water used during the analyses. HPLC-grade acetonitrile and methanol were obtained from Mallinckrodt Baker Inc. (Phillipsburg, NJ). Formic acid was obtained from Fluka (Buchs, Switzerland). Primary-secondary amine (PSA) Bond Elut was purchased from Varian, Inc. (Palo Alto, CA). Acetic acid was from Panreac (Barcelona, Spain). Anhydrous magnesium sulfate and sodium acetate were purchased from Sigma-Aldrich (Madrid, Spain). The poly(tetrafluoroethylene) (PTFE) surfaces (PTFE sheetss1/16 in., 12 in.× 12 in.) used throughout the study were acquired from Small Parts, Inc. (Miami Lakes, FL). DESI Source and Mass Spectrometer. Experiments were performed using a Thermo LTQ linear ion trap mass spectrometer (Thermo Finnigan San Jose, CA) tuned for optimum detection of the precursor ion of interest. Data were acquired via the Xcalibur software. DESI analyses were performed in the positive ion mode for all the compounds studied. The instrument was set to collect spectra in the automatic gain control mode for a maximum ion trap injection time of 200 ms and 2 microscans per spectrum. The capillary temperature was set at 200 °C. All DESI experiments were carried out using an OmniSpray ion source from Prosolia, Inc. (Indianapolis, IN). This DESI source was fitted with a sample platform, X-Y-Z positioners, and a charge-coupled device (CCD) camera to allow precise positioning and to maintain positional accuracy. The main experimental parameters used were as follows: solvent flow rate, 5 µL min-1; m/z range 150-600; ion spray voltage, 4.5 kV; capillary temperature, 200 °C; tube lens (V), -65 V; capillary voltage, -15 V. DESI-MS parameters were as follows: spray angle, 55°; collection angle, 10°; nitrogen gas pressure, 150 psi; distance from sample to tip, 5 mm; distance from sample to analyzer, 1.5 mm. Different spray solvent compositions were used in the course of optimization of the analysis including methanol/H2O (50:50, v/v), methanol/H2O (80:20, v/v), acetonitrile/H2O (50:50, v/v), acetonitrile/H2O (80: 20, v/v), acetonitrile/H2O (95:5, v/v), and acetonitrile/H2O (80: 20, v/v) 1% formic acid (selected). Different surfaces for sample deposition were tested and included paper, glass, and PTFE. An aliquot of 3.0 µL of solution was pipetted onto the surface. Tandem mass spectrometry experiments (MS/MS or MS3) were performed using collision-induced dissociation (CID) experiments in order to confirm the presence of the agrochemicals in the studied samples. These experiments were Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

821

Table 1. DESI-MS Analysis of Agrochemicals: Mass Spectral Features

a

Molecular mass (Mr) calculated using isotope-averaged atomic masses for the constituent elements.

performed using an isolation window of 1.5 (m/z units) and 25-35% collision energy (manufacturer’s unit). The MS/MS spectral features of the studied agrochemicals are shown in Table 1; they are consistent with previous literature on the same compounds using electrospray mass spectrometry.18 Electrospray Mass Spectrometry. Electrospray mass spectrometry was performed by direct infusion using the LTQ ion trap equipped with a Finnigan Ion Max API source to acquire spectra in the full-scan mode. Analyte standard solutions were tested at different concentration levels in MeOH/H2O 1:1 (v/v), with the following source parameters: sheath gas flow rate, 10 (manufacturer’s unit); ionspray voltage, 5 kV; capillary temperature, 275 °C; flow rate, 5 µL min-1. Samples. Fruit and vegetable samples (orange, lemon, apple, green pepper, persimmon, grapefruit, tomato, pear, and grape samples) were purchased from different local markets. Two main (18) Garcı´a-Reyes, J. F.; Hernando, M. D.; Ferrer, C.; Molina-Dı´az, A.; Ferna´ndezAlba, A. R. Anal. Chem. 2007, 79, 7308–7323.

822

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

experiments were performed: (1) direct DESI-MS analysis on peels of fruits and vegetables, which were analyzed without any treatment and (2) DESI-MS analysis of acetonitrile extracts from a detailed extraction protocol which is usually used in pesticide testing and follows guidelines.19 Sample Treatment and Extraction of Agrochemicals from Fruits and Vegetables. To obtain the acetonitrile extracts used for DESI-MS and LC-MS analyses a common procedure described elsewhere20 (so-called “QuEChERS”, from quick, easy, cheap, effective, rugged, and safe) was used. It is based on an acetonitrile extraction/partitioning from the foodstuff, followed by a cleanup stage using dispersive solid-phase extraction with PSA. The detailed protocol is described in the Supporting Information. (19) European Commission. Method validation and Quality Control Procedures for Pesticide Residues Analysis in Food and Feed. Document No. SANCO/ 2007/3131, October 31, 2007. Available at: http://ec.europa.eu/food/plant/ protection/resources/qualcontrol_en.pdf (accessed December 2008). (20) Anastassiades, M.; Lehotay, S. J.; Stajnbaher, D.; Schenck, F. J. J. AOAC Int. 2003, 86, 412–431.

Direct DESI-MS on Fruit and Vegetable Skins. No sample treatment was required. A piece of ca. 1 cm2 of peel was cut and secured using double-sided tape on a microscope glass slide (beveled microslides, size 75 mm × 25 mm, thickness 1 mm, Gold Seal, Bencton and Dinckinson Company, Franklin Lakes, NJ). This glass slide was fitted in the sample holder of the DESI source. DESI-MS Analyses of Fruit and Vegetable Extracts. Sample extracts from the QuEChERS method were diluted 1:3 (v/v) with acetonitrile (the final extract containing 0.33 g of commodity/mL of extract). Aliquots (3 µL of each solution) were deposited in a straight line about 4 mm apart (to avoid sample cross-contamination) on the PTFE surface (size ca. 0.25 in. × 1.5 in.) using a micropipet. Then, the samples/solutions spotted on the PTFE surface were allowed to dry (for ca. 2 min) at room temperature prior to analysis. The surface area of the spotted sample was 3-5 mm2. Optimal positioning of the sample stage and DESI spray source was achieved before sample analysis by maximizing the DESI ion signal using a standard solution of deuterated imazalil (200 µg L-1). Samples were analyzed by manually rastering across each spotted aliquot with the aid of the moving stage. The surface containing five samples was examined in ∼3 min using the XYZ moving stage. The angle of the spray source and position of the sample stage relative to the mass spectrometer inlet were kept constant during the analyses, and only the translational axes of the stage were moved to locate the analyte and obtain the signal. Safety Considerations. The DESI emitter floats at a high voltage, and necessary safety precautions, including grounding, should be taken. Liquid Chromatography Mass Spectrometry Reference Method. Liquid chromatography electrospray ion trap mass spectrometry (LC-ESI-IT-MS), in the positive ion mode was used as a reference method to roughly evaluate the results obtained with the DESI-MS approach. The LC-MS analyses were carried out using a high-performance liquid chromatography (HPLC) system (Agilent Series 1100, Agilent Technologies, Palo Alto, CA) equipped with a reversed-phase C8 analytical column. This HPLC system was connected to a quadrupole ion trap mass spectrometer, Bruker Esquire 6000 (Bruker Daltonics, Bremen, Germany) equipped with an electrospray interface operating in the positive ion and full-scan modes. The detailed method is described in the Supporting Information. Validation Studies and Accurate Quantitation Using Isotopically Labeled Standards. In both the DESI-MS and LC-MS methods, quantitation of sample extracts was accomplished with calibration curves using matrix-matched standards (containing 0.33 g of commodity/mL of extract, that is, 1:3 dilution). These matrix-matched standards were prepared from “blank” acetonitrile extracts of the commodities (the extract was analyzed to make sure it was free of the studied pesticide(s)) spiked with known amounts of the selected pesticide (imazalil) and a fixed concentration (500 ng mL-1) of imazalil-d5. For quantitation of imazalil residues in real samples, the ratio between the ion abundances of MS/MS transitions of imazalil (m/z 297 f m/z 255) and imazalil 37Cl35Cl isotope signal d5 (m/z 304 f m/z 257) was used for quantitation purposes in the DESI-MS method. Five repetitions were performed for each determination using five

different sample spots. The integrated peak area of the protonated molecule of imazalil ([M + H]+ (m/z 297)) was used to construct the calibration curves and perform the quantitation in the LC-MS method. RESULTS AND DISCUSSION DESI-MS of Agrochemicals: Mass Spectral Features. Sixteen agrochemicals belonging to different chemical classes (insecticides, fungicides, and herbicides) and representative of hundreds of individual chemicals were selected for this study. DESI full mass spectra of the selected agrochemicals deposited on PTFE surfaces were recorded. The DESI-MS spectra obtained for all of the pure compounds (Table 1) compared well with the typical electrospray mass spectra recorded in the course of LC-MS experiments.18 All the compounds gave high-quality spectra in the positive ion mode, as usual for most pesticides.18 The detected ions varied in mass from m/z 202 to m/z 732. In most cases, the protonated molecule ([M + H]+) was the most abundant ion observed. Sodium adduct formation was noticed only for amitraz, malathion, and isofenphos-methyl. This is consistent with the literature on electrospray mass spectrometry as well.18 Selection of the DESI Substrate and Optimization of the Solvent Spray Composition. Several parameters are important in the design of a DESI-MS method. Among them, the choice of a substrate/surface and the spray solvent composition are critical. Three different surfaces were tested as substrates: glass, paper, and PTFE. We obtained the highest and most stable signals with PTFE. Moreover, we avoided using paper as a substrate in order to avoid chromatographic effects in the course of solution deposition that could affect the performance of the method in terms of quantitation and precision. With concentrated samples, due to the matrix or analyte, the PTFE substrate sometimes gives a “coffee ring” drying effect where the sample is more concentrated in the outer circumference. This effect was not observed in this study as the matrix samples were diluted and the extract samples were relatively clean. A unique feature of DESI-MS compared to other ambient mass spectrometry ionization methods is the ability to tailor the spray solvent for specific analytes. For trace analyses of the selected agrochemicals, several spray solvent systems were investigated. They included different ratios of methanol, water, and acetonitrile including the addition of formic acid to promote the protonation of the target compounds, all used in the positive ion mode. The results are shown in Table S1 (Supporting Information). The composition of the solvent sprayed dramatically affected the signal of some of the compounds. Note, for instance, the remarkable increase in signal shown by the organophosphorus insecticides (i.e., malathion and isofenphos-methyl) when using acetonitrilebased solvent sprays instead of methanol. It should be kept in mind that the target of this study is not the analysis of one pesticide but the multiresidue analysis of several agrochemicals. For this reason, our goal was to find conditions that give good sensitivity for the wide range of classes included in this study. These conditions should also be satisfactory for other compounds added later or made the subject of separate targeted applications, if they show similar physicochemical properties to the chemicals studied. From the results obtained acetonitrile/water (80:20) with 1% formic acid was found to provide overall the best sensitivity Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

823

Figure 1. (a) DESI-MS full-scan mass spectrum of 75 µg L-1 (225 pg) of ametryn ([M + H]+, m/z 228); (b) DESI-MS full-scan mass spectrum of 200 µg L-1 (600 pg) of amitraz ([M + H]+, m/z 294); (c) DESI-MS full-scan spectrum of 333 µg L-1 (1000 pg) of imazalil ([M + H]+, m/z 297). All the spectra were recorded in the positive ion mode using acetonitrile/water (80:20, v/v), with 1% formic acid as the spray solvent and PTFE as substrate.

for all the compounds tested. We also tested matrix-matched extracts (orange and tomato), obtaining satisfactory results as well. Interestingly, the selected solvent is similar to the mobile phase composition used for the elution of these compounds in the reversed-phase liquid chromatography gradient method for electrospray ionization mass spectrometry detection (see the Supporting Information). As an example, the DESI-MS spectra obtained for some of the studied compounds with this spray solvent composition are shown in Figure 1. The signal-to-noise (S/N) ratio in full-scan acquisition mode is satisfactory even with subnanogram amounts (50 and 300, respectively. This kind of compound tends to remain in the citrus peel because it is used postharvest, in order to avoid mold during transportation and storage. This application was also tested successfully in many other fruit samples (apples, pears, and citrus). Therefore, the direct identification and confirmation of trace agrochemicals from the skin of market-purchased samples demonstrate the usefulness of the DESI-MS technique for direct pesticide testing, without any prior sample treatment, using the sample itself as the substrate. Another interesting example is illustrated in Figure S1 (see the Supporting Information), where both imazalil (m/z 297 f m/z 255) and thiabendazole (m/z 202 > m/z 175) were identified in the skin of grapefruit. As in the former example, the S/N ratio obtained in the DESI-MS/MS experiments illustrates the sensitivity of the proposed approach, which can be used for qualitative screening of pesticides in food. An equally interesting feature that can be exploited with the

proposed approach is the ability to perform field analysis, for instance, with the use of a hand-held portable mass spectrometer,22 which would save time and money and may help to speed up the monitoring programs undertaken to enforce the appropriate usage of agrochemicals in agriculture. In this context, the recent alerts concerning the use of banned/nonauthorized pesticides (from the “black market”) in agricultural production makes necessary the development of rapid in situ pesticide screening methods.23 For instance, the use of isofenphos-methyl, a nonauthorized organophosphorus insecticide, in pepper has been recently reported.23 In this kind of analysis, no quantitation is required, just a reliable yes/no response based on the use of MS/MS spectra to figure out whether a pesticide has been used in the greenhouse. DESIMS has the ability to perform ambient sampling and ionization of solid surfaces, and this was tested in a laboratory experiment, spiking a piece of green pepper peel with 600 pg cm-2 of isofenphos-methyl, with the sample being incubated for 1 h at room temperature before DESI-MS analysis (Figure 3). The full-scan DESI-MS spectrum is shown in Figure 3a. The insecticide main ions (m/z 332 ([M + H]+) and m/z 354 ([M + Na]+)) were observed near the detection limit. DESI-MS/ MS spectra (Figure 3, parts b and c) on both ions (m/z 332 f m/z 240; m/z 354 f m/z 313 and m/z 253) give further (22) Gao, L.; Sugiarto, A.; Harper, J. D.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2008, 80, 7198–7205. (23) Mezcua, M.; Ferrer, C.; Garcı´a-Reyes, J. F.; Martı´nez-Bueno, M. J.; Albarracin, M.; Claret, M.; Ferna´ndez-Alba, A. R. Rapid Commun. Mass Spectrom. 2008, 22, 1284–1392.

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

825

Figure 2. Identification of the fungicide imazalil in a market-purchased lemon by DESI: (a) full-scan mass spectrum of the studied lemon peel; (b) DESI-MS/MS spectrum of imazalil standard (1 ng/3 µL of 330 µg L-1 standard solution) deposited on a PTFE surface; (c) DESI-MS/MS spectrum of imazalil (m/z 297) detected in lemon peel; (d) DESI-MS/MS spectrum of imazalil (m/z 299s37Cl signalsdetected in lemon peel). For details, see text. All the spectra were recorded in the positive ion mode using acetonitrile/water (80:20, v/v), with 1% formic acid as the spray solvent and PTFE as substrate.

confirmation of the presence of the studied insecticide. The S/N ratio attained was enough for the confirmation of subnanogram amounts of the banned chemical in a complex substrate like a green pepper peel. For further comparison, similar analysis of a standard solution of isofenphos-methyl is presented within the Supporting Information, Figure S2. In the same sample, traces of malathion (a common organophosphorus insecticide) which were already present in the sample were also detected (data not shown). These results highlight the convenience and efficiency of DESI-MS as a sampling and ionization technique for pesticide use and enforcement in the field (i.e., greenhouses). Future work may include the optimization of nongeometry-dependent DESI sources24 and the coupling of DESI and portable MS instruments,22 but so far, the data in hands is highly promising. DESI-MS Analysis of Fruit and Vegetable Extracts. After testing the sensitivity of the method in fruit and vegetable skins, the following step was implemented to evaluate the approach with extracts. Twenty market-purchased fruit and vegetable samples were extracted and analyzed by DESI-MS/MS and LC-MS. Qualitative screening and confirmation of trace amounts of pesticides in fruit and vegetables were accomplished satisfactorily using DESI-MS with MS/MS or MS3 experiments. The proposed DESI-MS method was found to provide qualitative results for most of the compounds detected by the full (24) Venter, A.; Cooks, R. G. Anal. Chem. 2007, 79, 6398–6403.

826

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

LC-MS method (Table 3). From the data collected, DESIMS/MS experiments enabled the confirmation of agrochemicals in real samples at low concentration levels, as low as 15 pg of azoxystrobin in a studied grape extract (15 µg kg-1 in the sample, based on LC-MS data). The DESI-MS/MS spectra (MS/MS and MS3) of azoxystrobin in this extract are shown in Figure 4a. MS3 (m/z 404 f m/z 372 f m/z 344; 316) provided unambiguous confirmation on the presence of this fungicide in the grape extract, based on consecutive neutral losses of methanol during MS/MS and MS3 stages. Another example is the DESI-MS/MS of an apple extract, where low concentration levels of imazalil were detected (90 pg or 90 µg kg-1 in the sample based on LC-MS data), shown in Figure S3 (see the Supporting Information). Even at such a low concentration, where the pesticide is revealed by LC-MS analysis, good quality MS/ MS spectra with relatively high S/N ratios are recorded. Note, for instance, that the MRLs of imazalil in citrus fruits established in Europe, the United States, or Japan are in the range of 5-10 mg kg-1. Therefore, in some of these samples, we performed the analysis of trace agrochemicals at concentration levels as much as 50-100 times lower than the MRL tolerated, obtaining satisfactory results and unambiguous confirmation of the presence of this species. Figure 4b illustrates another example of the DESI-MS/MS method with the analysis of a pear extract. According to the

Table 3. Application of the Proposed DESI-MS/MS Method for the Screening and Identification of Agrochemicals in Fruit and Vegetable Extracts

no.

sample

1

pear

2 3 4 5 6

pear lemon apple apple tomato

7 8

apple pear

9 10 11 12 13

persimmon grapes lemon orange orange

14 15 16 17 18

orange tomato apple persimmon lemon

19 20

apple lemon

agrochemical(s) detected and concentration found by LC-MS (µg g-1) carbendazim (0.040); imazalil (0.255) imazalil (0.630) imazalil (0.775) imazalil (0.170) imazalil (0.210) azoxystrobin (0.080); buprofezin (0.040) thiabendazole (0.280) imazalil (0.420); thiabendazole (0.765) malathion (0.399) azoxystrobin (0.015) imazalil (0.910) imazalil (0.127) imazalil (0.470); thiabendazole (0.466) imazalil (0.110) buprofezin (0.017) thiabendazole (0.448) imazalil (0.230) imazalil (0.230); buprofezin (0.006) imazalil (0.090) imazalil (0.244)

identified/confirmed by DESI-MS/MS proposed methoda carbendazim (nontarget) (D/C);b imazalil (D/C) imazalil (D/C) imazalil (D/C) imazalil (D/C) imazalil (D/C) azoxystrobin (D/C); buprofezin (D-LOD)c thiabendazole (D/C) imazalil (D/C); thiabendazole (D/C) malathion (D/C) azoxystrobin (D-LOD) imazalil (D/C) imazalil (D/C) imazalil (D/C); thiabendazole (D/C) imazalil (D/C) buprofezin (ND)d thiabendazole (D/C) imazalil (D/C) imazalil (D/C); buprofezin (ND) imazalil (D/C) imazalil (D/C)

a Acetonitrile extracts from QuEChERS procedurescontaining 1 g of matrix (fruit/vegetable)/mLsare diluted 1:3 (v/v) prior to analysis. b D/C: detected and confirmed by MS/MS spectrum. c D-LOD: detected (approaching limit of detection). d ND: nondetected.

Figure 3. (a) DESI-MS full-scan spectrum of a green pepper spiked with nonauthorized insecticide isofenphos-methyl (600 pg cm-2). Characteristic ions: [M + H]+ m/z 332 and [M + Na]+ m/z 354 observed; (b) DESI-MS/MS spectrum of m/z 332; (c) DESI-MS/MS spectrum of m/z 354. For details, see text. All the spectra were recorded in the positive ion mode using acetonitrile/water (80:20, v/v), with 1% formic acid as the spray solvent and PTFE as substrate.

LC-MS analysis, 0.25 mg kg-1 of imazalil and 0.040 mg kg-1 of carbendazim (a benzimidazole fungicide) were detected. Despite the fact that carbendazim was not included in our list of target chemicals, using the substrate and spray solvent conditions selected for the DESI-MS/MS method, the identification and confirmation of this fungicide was performed, and MS/ MS gave data consistent with the available literature,18 even at such low concentration. This example illustrates the potential nontarget capabilities of the DESI-MS approach. The selected conditions might also be convenient for a large number of chemicals obtaining remarkable sensitivity as occurred with carbendazim in the studied pear extract. For instance, 13 pg µL-1 (40 µg kg-1 sample) of carbendazim was detected in the diluted extract (where the amounts are based on the LC-MS data).

Quantitation of Imazalil Residues in Citrus Extracts Using Isotopically Labeled Standards. Accurate quantitation of agrochemicals was explored for the first time in this study although several previous papers have reported on quantitation by DESI, in a few cases using isotopic labeling.25-27 The use of an isotopically labeled standard improves the precision of the method providing accurate and reliable results. Selected citrus samples (acetonitrile extracts obtained after QuEChERS protocol) containing imazalil residues were analyzed with the proposed DESI-MS/MS method, and the results were compared with those obtained using the LC-MS reference method (Table 4). Note that the precision using isotopically labeled internal standards is consistently below 15%, which is difficult to achieve under regular DESI-MS without using an internal standard. In this sense, quantitation errors due to matrix-induced signalsuppression effects (a common phenomenon also in electrospray mass spectrometry) were negligible when using DESIMS/MS with isotopically labeled standards. We used the main fragmentation of imazalil (m/z 297 f 255) and the transition of 37Cl of imazalil-d5 (m/z 304f m/z 257) simply to avoid the mass spectral overlap between the 13C signal of the “A + 4” signal of imazalil (C14H15N2O37Cl2) (m/z 302) with the (25) Ifa, D. R.; Manicke, N.; Rusine, A. L.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2008, 22, 503–510. (26) Nyadong, L.; Late, S.; Green, M. D.; Banga, A.; Ferna´ndez, F. M. J. Am. Soc. Mass Spectrom. 2008, 19, 380–388. (27) Talaty, N.; Mulligan, C. C.; Justes, D. R.; Jackson, A. U.; Noll, R. J.; Cooks, R. G. Analyst 2008, 133, 1532–1540.

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

827

Figure 4. (a) DESI-MS/MS analysis of a grape extract where azoxystrobin was detected: (a.1) MS/MS (m/z 404 f 372); (a.2) MS3 (m/z 404 f m/z 372 f m/z 344; 316). (b) DESI-MS/MS analysis of a pear extract where carbendazim and imazalil were detected: (b.1) MS/MS for the identification of carbendazim (m/z 192 f m/z 160); (b.2) MS3 (m/z 192 f m/z 160 f m/z 132). For details, see text. All the spectra were recorded in the positive ion mode using acetonitrile/water (80:20, v/v), with 1% formic acid as the spray solvent and PTFE as substrate.

Table 4. Quantitation of Imazalil Residues in Citrus Fruits by DESI-MS/MS and Comparison with the LC-MS Reference Method imazalil concentration (µg g-1) no.

sample

LC-MS (µg g-1)

DESI-MS (µg g-1)

RSD (%) (n ) 5)

1 2 3 4 5 6

orange lemon lemon orange grapefruit orange

0.19 0.35 0.44 0.67 0.33 0.15

0.16 0.38 0.48 0.58 0.30 0.12

12.8 11.2 9.6 8.1 12.1 14.5

main signal of imazalil-d5 (m/z 302). The intensity ratios of both transitions were used to construct a calibration curve using a fixed internal standard concentration (500 pg µL-1) and increasing concentrations of imazalil (30-3000 pg µL-1) spiked into an orange extract so that matrix effects were avoided. Both the precision and linearity of the method were satisfactory with relative standard deviation (RSD) (%) values consistently below 15%, and the calibration curve obtained with a regression coefficient being better than 0.99. These RSD values are satisfactory as they approach the typical precision values obtained with triple quadrupole LC-MS (i.e., 5%). In addition, the absolute concentration levels detected were in agreement with those obtained by LC-MS. 828

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

This is the first time quantitative (ultra)trace analysis in complex matrixes has been reported with DESI mass spectrometry, and it establishes the potential of the proposed approach for quantitative pesticide testing in foodstuffs. CONCLUSIONS Food safety is a research field whose requirements map well against the features of ambient ionization mass spectrometric techniques. In this work, DESI-MS was successfully used as a rapid method for trace analysis of agrochemicals in ordinary fruit and vegetable samples. The results obtained correlate well with those obtained by LC-MS, and the sensitivity achieved using DESI was adequate for the analysis of these 16 representative agrochemicals in a variety of types of samples. Ion suppression due to matrix effects, which are common in electrospray mass spectrometry, were observed but can be circumvented with appropriate dilutions, since the sensitivity and LODs obtained (in the micrograms per kilogram range in fruit and vegetable matrixes) were satisfactory even after dilution of the food extracts. Direct analysis of market samples without any further treatment was also demonstrated. Accurate trace quantitation in complex matrixes is also reported for the first time using DESI-MS with isotopically labeled standards, obtaining satisfactory results. Future work may include the use of DESI-MS and other ambient mass spectrometry techniques

for field-oriented pesticide testing. Research in order to perform pesticide testing with DESI-MS implemented in portable instruments in the field is being carried out in our laboratory. ACKNOWLEDGMENT Funding at Purdue University was provided by Thermo-Fisher (Grant No. 1320036658) and the Andrews fellowship (A.U.J.). J.F.G.-R. also acknowledges the postdoctoral Fulbright scholarship from the “Secretarı´a de Estado de Universidades e Investigacio´n” of the Spanish Ministry of Education and Science.

SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs. org.

Received for review November 26, 2008.

October

13,

2008.

Accepted

AC802166V

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

829