High-Throughput Pesticide Residue Quantitative Analysis Achieved by

Anal. Chem. 2009, 81, 3134–3142

High-Throughput Pesticide Residue Quantitative Analysis Achieved by Tandem Mass Spectrometry with Automated Flow Injection Sergio C. Nanita,* Anne M. Pentz, and Frederick Q. Bramble DuPont Crop Protection, Stine-Haskell Research Center, 1090 Elkton Road, Newark, Delaware 19714 The use of automated flow injection with MS/MS detection for fast quantitation of agrochemicals in food and water samples was demonstrated in this study. Active ingredients from the sulfonylurea herbicide and carbamate insecticide classes were selected as model systems. Samples were prepared using typical procedures from residue methods, placed in an autosampler, and injected directly into a triple quadrupole instrument without chromatographic separation. The technique allows data acquisition in 15 s per injection, with samples being injected every 65 s, representing a significant improvement from the 15-30 min needed in typical HPLC/MS/ MS methods. The availability of HPLC systems is an advantage since they can be used in flow-injection mode (bypassing the column compartment). Adequate accuracy, linearity, and precision (R2 > 0.99 and RSD < 20%) were obtained using external standards prepared in each control matrix. The limit of quantitation (LOQ) achieved for all analytes was 0.01 mg/kg in food samples and 0.1 ng/mL in water; while limits of detection (LOD) were estimated to be about 0.003 mg/ kg and 0.03 ng/mL in food and water, respectively. The advantages and limitations of flow injection MS/ MS for ultratrace-level quantitative analysis in complex matrixes are discussed. Ultratrace-level quantitation of pesticide active ingredients in food and environmental samples is an important yet challenging task. It is a common practice by regulatory authorities to ensure the safety of consumers by confirming that food produce are in compliance with maximum residue limits (MRLs) and good agricultural practices are being followed. Registrants of agrochemicals are also required to provide analytical methods that allow quantitation of the pesticides in environmental matrixes such as soil and water1,2 and food of plant and/or animal origin,2,3 depending on the intended use of the product. Analytical methods capable of quantifying pesticides at low levels are also needed to * To whom correspondence should be addressed. Phone: 1-302-451-5806. E-mail: [email protected]. (1) U.S. EPA. Ecological Effects Test Guidelines; OPPTS 850.7100, Data Reporting for Environmental Chemistry Methods, April 1996. (2) EEC Directive 91/414/EEC, Annex IIA 4.2.2 as amended by EC Directive 96/46/EC; SANCO/825/00 rev.7 (17/03/2004) Guidance Document on Residue Analytical Methods. (3) U.S. EPA. Residue Chemistry Test Guidelines; OPPTS 860.1340, Residue Analytical Method, August 1996.

3134

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

support studies during the development of experimental agrochemicals,4 often involving the quantitation of active ingredient(s) and metabolites in thousands of soil and crop samples. The complex composition of environmental and crop samples makes it difficult to perform quantitative determinations, particularly when accuracy and precision quality standards need to be met.5 In addition, the requirement of ultratrace quantitation further complicates analytical procedures by limiting the technologies that can be used for instrumental analysis. Tandem mass spectrometry has become the preferred detection technology in pesticide residue analysis6,7 because of its versatility and high selectivity and sensitivity, allowing current multiresidue methods to quantify hundreds of active ingredients in a single analytical set.8-11 High-performance liquid chromatography (HPLC) is one of the preferred sample introduction techniques for mass spectrometric detection. It is also the sample throughput limiting step in methods for instrumental analysis to quantify pesticide residues. The recent introduction of ultraperformance liquid chromatography (UPLC)12,13 has provided significant reduction in run times with improved resolution.14-16 However, many laboratories are not equipped with UPLC systems, which are significantly more expensive than conventional HPLC units. Moreover, recent advancements in instrumentation have significantly increased the sensitivity, selectivity, and ruggedness of mass spectrometry, improving the ability for detection and quantitation of chemicals in complex matrixes. Consequently, (4) Nanita, S. C.; Pentz, A. M.; Grant, J.; Vogl, E.; Devine, T. J.; Henze, R. M. Anal. Chem. 2009, 81, 797–808. (5) Nanita, S. C.; Huber, A.; Ziegler, M.; Astor, E.; Cairns, S.; Beadle, N.; Cranwell, S.; Giammarrusti, L.; Ruhl, J.; Ortega, F.; Selley, A. Mediterranean Group of Pesticide Residue Conference; Piacenza, Italy, 2008. (6) Alder, L.; Startin, J. R.; Alonso, S.; Anspach, T.; Brewin, S.; Broekaert, C.; Christiansen, A.; Dekok, A.; Frase, U.; Fresvig, M.; Hemmerling, Ch.; Hermansson, E.; Hiemstra, M.; Hogendoorn, E.; Kolb, J.; Kombal, R.; Melk, Ch.; Polonji, B.; Quirijns, J. K.; Ross, L.; Saint-Joly, Ch.; Scherbaum, E.; Van Damme, D.; Welter, A.; Wu ¨ st, B. J. AOAC Int. 2005, 88, 1762–1776. (7) Zywitz, D.; Anastassiades, M.; Scherbaum, E. Dtsch. Lebensm.-Rundsch. 2003, 99, 188–196. (8) Greulich, K.; Alder, L. Anal. Bioanal. Chem. 2008, 391, 183–197. (9) Lehotay, S. J.; De Kok, A.; Hiemstra, M.; Van Bodegraven, P. J. AOAC Int. 2005, 88, 595–614. (10) Alder, L.; Greulich, K.; Kempe, G.; Vieth, B. Mass Spectrom. Rev. 2006, 25, 838–865. (11) Klein, J.; Alder, L. J. AOAC Int. 2003, 86, 1015–1037. (12) Swartz, M. E. LCGC North Am. 2005, 23, 8–14. (13) Swartz, M. E. J. Liq. Chromatogr. Relat. Technol. 2005, 28, 1253–1263. (14) Taylor, M. J.; Keenan, G. A.; Reid, K. B.; Fernańdez, D. U. Rapid Commun. Mass Spectrom. 2008, 22, 2731–2746. (15) Gervais, G.; Brosillon, S.; Laplanche, A.; Helen, C. J. Chromatogr., A 2008, 1202 (2), 163–172. (16) Romero-Gonza´lez, R.; Frenich, A. G.; Vidal, J. L. M. Talanta 2008, 76, 211–225. 10.1021/ac900226w CCC: $40.75  2009 American Chemical Society Published on Web 03/18/2009

the use of HPLC for analyte preconcentration, purification, and resolution may no longer be needed for tandem mass spectrometric detection as demonstrated by recently introduced ambient ionization mass spectrometry techniques, like desorption electrospray ionization (DESI),17 desorption atmospheric pressure chemical ionization (DAPCI),18 direct analysis in real time (DART),19 and atmospheric pressure solids analysis probe (ASAP),20,21 which allow analysis with minimal sample preparation. DESI and DART19 are the most used techniques within ambient ionization mass spectrometry, and they have been successfully employed for qualitative and quantitative analysis of several chemicals in complex matrixes,22-26 including pesticide active ingredients in food.27,28 The field of ambient ionization mass spectrometry compiles emerging technologies with the capability to perform high-throughput direct or in situ analysis of organic compounds at low levels in complex systems, an application of great interest. However, commercial ambient ionization instrumentation is yet to become widely available at private laboratories for quantitative analysis or in situ applications. A high-throughput technology not yet evaluated systematically for quantitative analysis of pesticide residues is flow injection mass spectrometry. Flow injection analysis has been successfully used in pesticide residue bioassays that employ fluorescence29-31 and calorimetric32 detection. Flow injection has also been used for sample introduction in mass spectrometric methods mainly for qualitative high-throughput metabolite fingerprinting33,34 and screening35 in drug discovery. The development of a flow injection approach for high-throughput pesticide residue quantitation (17) Taka´ts, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471–473. (18) Cotte-Rodrı´guez, I.; Hernańdez-Soto, H.; Chen, H.; Cooks, R. G. Anal. Chem. 2008, 80, 1512–1519. (19) Cody, R. B.; Larameé, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297– 2302. (20) McEwen, C. N.; McKay, R. G.; Larsen, B. S. Anal. Chem. 2005, 77, 7826– 7831. (21) McEwen, C.; Gutteridge, S. J. Am. Soc. Mass Spectrom. 2007, 18, 1274– 1278. (22) Ifa, D. R.; Manicke, N. E.; Rusine, A. L.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2008, 22, 503–510. (23) Nyadong, L.; Late, S.; Green, M. D.; Banga, A.; Fernańdez, F. M. J. Am. Soc. Mass Spectrom. 2008, 19, 380–388. (24) Mulligan, C. C.; MacMillan, D. K.; Noll, R. J.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2007, 21, 3729–3736. (25) Kpegba, K.; Spadaro, T.; Cody, R. B.; Nesnas, N.; Olson, J. A. Anal. Chem. 2007, 79, 5479–5483. (26) Pierce, C. Y.; Barr, J. R.; Cody, R. B.; Massung, R. F.; Woolfitt, A. R.; Moura, H.; Thompson, H. A.; Fernandez, F. M. Chem. Commun. 2007, 8, 807– 809. (27) Garcia-Reyes, J. F.; Jackson, A. U.; Molina-Diaz, A.; Cooks, R. G. Anal. Chem. 2009, 81, 820–829. (28) Schurek, J.; Vaclavik, L.; Hooijerink, H.; Lacina, O.; Poustka, J.; Sharman, M.; Caldow, M.; Nielen, M. W. F.; Hajslova, J. Anal. Chem. 2008, 80, 9567– 9575. (29) Liu, H.; Hao, Y.; Ren, J.; He, P.; Fang, Y. Luminescence 2007, 22, 302–308. (30) Flores, J. L.; Dıáz, A. M.; Fernańdez de Co´rdova, M. L. Anal. Chim. Acta 2007, 585, 185–191. (31) Su´bova´, I.; Assandas, A. K.; Icardo, M. C.; Calatayud, J. M. Anal. Sci. 2006, 22, 21–24. (32) Zheng, Y.-H.; Hua, T.-C.; Sun, D.-W.; Xiao, J.-J.; Xu, F.; Wang, F.-F. J. Food Eng. 2006, 74, 24–29. (33) Enot, D. P.; Lin, W.; Beckmann, M.; Parker, D.; Overy, D. P.; Draper, J. Nat. Protoc. 2008, 3, 446–470. (34) Beckmann, M.; Parker, D.; Enot, D. P.; Duval, E.; Draper, J. Nat. Protoc. 2008, 3, 486–504. (35) Roddy, T. P.; Horvath, C. R.; Stout, S. J.; Kenney, K. L.; Ho, P.-I.; Zhang, J.-H.; Vickers, C.; Kaushik, V.; Hubbard, B.; Wang, Y. K. Anal. Chem. 2007, 79, 8207–8213.

without chromatographic separation prior to MS/MS detection was addressed in this study, and the capabilities of the technique were demonstrated. In addition, the first methods for quantitation of sulfonylurea herbicides and carbamate insecticides using this technology are reported. EXPERIMENTAL SECTION Details about the preparation of samples and matrix-matched standards are available as Supporting Information, together with mass spectra of the compounds studied. Summarized experimental procedures are provided below. Reagents and Standards. All reagents and solvents used were commercially available at the time of this study. Acetonitrile, methanol, and water (HPLC grade) and hexane, ammonium hydroxide (28-30%), and formic acid (98%, ACS grade) were obtained from EMD Chemicals (Gibbstown, NJ). Ammonium acetate was purchased from Mallinckrodt (Phillipsburg, NJ), while potassium phosphate (dibasic) and phosphoric acid were obtained from J.T. Baker (Phillipsburg, NJ). The sulfonylurea herbicides triflusulfuronmethyl, flupyrsulfuron-methyl, rimsulfuron, and sulfometuron-methyl as well as the carbamate insecticides oxamyl and methomyl were selected as representative agrochemical active ingredients for this study. The structures of these compounds appear in Figure 1. Analytical standards used in this study were synthesized by DuPont Crop Protection, Global Technology Division, E. I. du Pont de Nemours and Company. Analytical standards of these active ingredients are also available through the U.S. E.P.A National Pesticide Repository and several commercial sources. Analytical standard stock solutions were individually prepared in acetonitrile, each at 100 µg/mL. A 10 µg/mL multianalyte standard was prepared by 1:10 dilution of the individual stock standard solutions into a common volumetric flask. Additional mixed standards at lower concentrations were prepared by serial dilution of the 10 µg/mL multianalyte standard. The mixed standards were used for sample fortification during method development and validation. Control Samples. Control drinking (tap) water was collected at DuPont Stine-Haskell Research Center on December 15, 2008. Prehomogenized corn, lemon, and pecan untreated control samples were available from previous DuPont studies and were maintained frozen at a target temperature of -20 °C prior to analysis. These samples were used for method development and validation in this study. Food Sample Extraction and Purification. Corn, lemon, and pecan control samples were fortified with sulfonylureas and carbamates, extracted, and purified following typical analytical procedures used by DuPont Crop Protection for residue analysis by HPLC/tandem mass spectrometry (MS/MS). In this case, flow injection tandem mass spectrometry substituted HPLC/MS/MS for instrumental analysis. Briefly, homogeneous crop subsamples (10 g) were extracted twice with 90 mL of acetonitrile/aqueous pH 7 dibasic potassium phosphate buffer solution (75/25, v/v) using a tissue-grinding homogenizer, and the final volume of the combined extracts was adjusted to 200 mL with acetonitrile. Fractions (20 mL) of the extracts were purified by hexane partition, and the hexane phase was discarded. A 10 mL aliquot of each resulting extract was purified by solid phase extraction (SPE) using 6 cm3/500 mg ENV Bond-Elut cartridges (part no. 952493, Varian, Inc., Harbor City, CA). Aliquots of purified Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3135

Figure 1. Structures and common names of sulfonylurea herbicides and carbamate insecticides tested as representative agrochemical active ingredients in flow injection MS/MS methods for quantitative analysis. Table 1. Optimized Sciex API-5000 Parameters for Tandem Mass Spectrometric Detection of Sulfonylurea Herbicides and Carbamate Insecticides analyte

precursor ion typea

Q1 m/z

Q3 m/z

DP (V)

CE (V)

triflusulfuron-methyl rimsulfuron flupyrsulfuron-methyl sulfometuron-methyl oxamyl methomyl

(M + H)+ (M + H)+ (M + H)+ (M + H)+ (M + NH4)+ (M + H)+

493.0 432.1 466.0 365.1 236.9 163.1

264.0 182.0 182.0 150.0 71.9 88.1

75 100 75 100 75 50

40 25 40 25 40 10

a MS/MS spectra recorded for these precursor ions are available as Supporting Information. The spectra show additional fragment ions which could be monitored for analyte quantitation or confirmation.

extract solutions were filtered through 0.45 µm PTFE syringe filters prior to quantitative instrumental analysis by flow injection MS/MS. This sample purification procedure was originally designed for analysis of sulfonylurea herbicides. Initial tests revealed that the carbamate insecticides oxamyl and methomyl were retained and eluted in the SPE cleanup together with the sulfonylureas. Therefore, the same procedure was used for both classes of pesticides to quantify all six analytes concurrently by flow injection MS/MS. Water Sample Purification. Water aliquots (50 mL) were measured into propylene centrifuge tubes and fortified with known amounts of the analytes. The samples were acidified by adding 10 µL of concentrated formic acid prior to purification, which was achieved together with 10-fold preconcentration of the analytes by solid phase extraction (SPE) using 6 cm3/500 mg ENV BondElut cartridges (part no. 952493, Varian, Inc., Harbor City, CA). Quantitative analysis was performed by flow injection tandem mass spectrometry. Instrumental Conditions. An Agilent 1100 series HPLC (Agilent Technologies, Wilmington, DE) coupled to an Applied Biosystems API-5000 triple quadrupole mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA) equipped with an electrospray ionization source was used for instrumental analysis. The HPLC system consisted of a vacuum degasser, a binary pump, a temperature-controlled column compartment, and a refrigerated autosampler. Red PEEK capillary 1/16 in. outer diameter and 0.13 3136


mm inner diameter (part no. 0890-1915, Agilent Technologies, Wilmington, DE) with the length adjusted to approximately 1 m was used to connect the flow from the autosampler to the electrospray ion source. The PEEK tubing connection bypassed the column compartment to allow direct injection of samples into the mass spectrometer. The small inner diameter tubing was chosen to minimize the void volume. Methanol was used as both the carrier and sample solvent. A carrier flow of 400 µL/min and a 1 µL sample injection volume were used in all experiments and method validation, unless specified otherwise. The flow was introduced directly and continuously into the ion source (no solvent splitting or divert valve before the ion source). The entire system and data acquisition were controlled by Analyst 1.4.1 software. The Applied Biosystems API-5000 was operated in MS/MS positive ion mode with multiple reaction monitoring (MRM) detector output for quantitative analysis. A summary of optimized mass spectrometric conditions is provided in Table 1. Note that the selection of the fragment ions used for quantitation was based on the observed response in the MS/MS spectra (MS/MS spectra available as Supporting Information). Additional parameters were set as follows: resolution Q1 ) unit, resolution Q3 ) unit, ESI source voltage 4.5 kV, dwell time ) 20 ms for each ion transition, CUR ) 10 psi, GS1 ) 70 psi, GS2 ) 70 psi, ion source temperature ) 450 °C, CAD pressure ) 7 psi, entrance potential (EP) ) 10 V, and CXP ) 25 V. The Analyst 1.4.1 acquisition method consisted of a 15 s MRM data collection for all ion transitions

Figure 2. TIC peak area obtained during optimization of analyte ionization efficiency. The sample solvents tested were (1) water, (2) 0.1% formic acid (aq), (3) 0.1 M ammonium acetate (aq), (4) methanol, (5) 0.1% formic acid in methanol, (6) 0.1 M ammonium acetate in methanol, (7) methanol/water 50/50, (8) 0.1% formic acid in methanol/water 50/50, (9), 0.1 M ammonium acetate in methanol/water 50/50, (10) 0.1 M ammonium acetate and 0.1% formic acid in methanol/water 50/50, and (11) 0.1 M ammonium acetate and 0.1% formic acid in methanol.

Figure 3. Flow injection MS/MS ion chronogram36 for a 1 µL injection of a 0.7 ng/mL standard containing 4 sulfonylureas herbicides and 2 carbamate insecticides prepared in corn matrix and recorded using the optimized conditions.

immediately after injection. Following the acquisition, the system was allowed to flush for approximately 20 s prior to completing the run. The autosampler motion/sample injection required ∼30 s, for a total elapsed time between 1 µL injections of approximately 65 s. The time required for sample injection was directly proportional to the sample injection volume. RESULTS AND DISCUSSION Optimization of Ionization Efficiency. The MS/MS conditions (ion source and MRM transitions) were tuned for quantitation of the analytes following typical optimization procedures. After

the mass spectrometer was tuned, the electrospray ionization efficiency was optimized for detection of sulfonylureas and carbamates by preparing a set of eleven 10 ng/mL mixed standards in solutions containing different concentrations of formic acid and/or ammonium acetate. Water, methanol, and water/ methanol mixtures were used as solvents for the standards. The set was analyzed using four different carrier solvents and 5 µL injection volume, and the total ion current peak area obtained for each sample solvent/carrier solvent combination appears in Figure 2. The highest sensitivity was achieved when methanol was used as both the sample and carrier solvent. The peak shape (analyte Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3137

Figure 4. Flow injection MS/MS ion chronograms36 (TIC) demonstrating the effect of the injection volume. A 1.0 ng/mL mixed standard was prepared in methanol and analyzed with (a) 1, (b) 5, (c) 10, (d) 25, and (e) 50 µL injection volumes.

band) was acceptable and reproducible for all six analytes using methanol as the solvent. This should be evaluated and solvents selected to minimize analyte diffusion. In principle, this straightforward approach can be used to optimize the ionization efficiency of any analyte for quantitation by flow injection MS/MS. Optimum conditions for carbamates and sulfonylureas were achieved in ∼2-3 h of laboratory work, which compares to days of testing often needed to develop a reliable HPLC method. Stationary phase, column dimensions, and analyte retention time represent additional parameters that require optimization during HPLC method development. Reproducibility and repeatability of HPLC methods often depend on careful preparation of mobile phases (e.g., buffers). HPLC often limits the choice of solvents used to prepare samples and mobile phases since high aqueous content may be required to achieve adequate peak shapes, and this prevents the use of the best possible conditions for electrospray ionization efficiency. On the basis of the experiment and results discussed 3138


above, solvent limitations are less strict in flow injection, at least for the analytes tested. A lower response was observed for samples containing matrix, thus calibration standards were prepared in the control matrix to correct for ion suppression in flow injection quantitation experiments (discussed under the Method Validation section). The sensitivity obtained for sulfonylureas and carbamates using flow injection ranged from 500-700 fg injected (1 µL injection of 0.5-0.7 ng/mL standards). A representative ion chronogram36 for a 0.7 ng/mL corn matrix-matched standard is provided in Figure 3. The lowest standard concentration detectable depended on the magnitude of matrix effects. The typical sensitivity observed for sulfonylurea standards prepared in neat solvents (no matrix) with (36) Note that chromatography was not used for instrumental analysis. Consequently, the term “ion chronogram” is used here to describe the mass spectrometer response recorded over time, instead of “chromatogram”, which typically refers to a chromatographic separation chart.

Table 2. Flow Injection MS/MS Method Tryout Results for Food and Drinking Water analyte

average % recovery (±RSD)a no. of samplesb

triflusulfuron-methyl rimsulfuron flupyrsulfuron-methyl sulfometuron-methyl methomyl oxamyl

Corn 119 ± 3 110 ± 9 115 ± 8 115 ± 6 103 ± 2 107 ± 9

5 5 5 5 5 5


Lemon 124 ± 12 96 ± 8 108 ± 13 117 ± 13 110 ± 10 106 ± 12

4 4 4 4 4 4


Pecan 97 ± 5 100 ± 4 96 ± 9 97 ± 4 86 ± 10 85 ± 12

5 5 5 5 5 5


All Food Matrixes Tested 113 ± 13 103 ± 9 106 ± 12 109 ± 12 99 ± 13 99 ± 15

14 14 14 14 14 14

triflusulfuron-methyl rimsulfuron flupyrsulfuron-methyl sulfometuron-methyl oxamyl methomyl

Drinking (Tap) Water 92 ± 15 102 ± 19 98 ± 17 108 ± 17 109 ± 13 100 ± 18

8 8 8 8 8 8

a Percent recovery is defined as (analyte added/analyte measured) × 100. b Three samples fortified at 0.01 mg/kg (LOQ of the method) and two at 0.10 mg/kg were analyzed as part of the method tryout for food. One lemon sample fortified at the LOQ was an outlier for all analytes, thus was excluded from this table. Four water samples fortified at 0.1 ng/mL (LOQ of the method) and four at 1.0 ng/mL were analyzed as part of the method tryout for drinking water.

HPLC/MS/MS residue methods at DuPont using the same model triple quadrupole mass spectrometer ranged from 625-1250 fg (2.5 µL injection of 0.25 ng/mL standard and 25 µL injection of 0.05 ng/mL standard, respectively). The flexibility of using any carrier and sample solvent allows the use of conditions for optimum ionization efficiency, resulting in improved sensitivity with flow injection MS/MS, enough to compensate for the matrix suppression expected due to the lack of the analyte purification offered by HPLC. Injection Volume Effect. The effect of the injection volume was assessed as part of method optimization. A 1.0 ng/mL mixed standard prepared in methanol was sequentially analyzed using injection volumes ranging from 1 to 50 µL, and the obtained flow injection MS/MS ion chronograms36 appear in Figure 4. Acceptable signal-to-noise ratios were obtained with 1-5 µL injection volumes. The use of larger injection volumes resulted in a wider analyte band that reduced the signal-to-noise response and relative sensitivity for pesticide analysis by flow injection. Moreover, larger injection volumes translate to an unnecessary amount of matrix and analyte(s) introduced in the system and the potential for

analyte carryover between injections. On the other hand, the experiment highlights a disadvantage of flow injection, which is the lack of the on-column preconcentration capability that enables high-performance liquid chromatography methods to use larger injection volumes. Consequently, sample preparation procedures used with flow injection MS/MS for pesticide residue quantitation must account for the lower injection volumes required for instrumental analysis. Sulfonylureas and carbamates had approximately the same MS/MS response on a weight basis for quantitation by flow injection and high-performance liquid chromatography methods. Selectivity in Flow Injection MS/MS. High-performance liquid chromatography contributes to selectivity by resolving multiple analytes in time prior to tandem mass spectrometric detection. Tandem mass spectrometry contributes to analyte selectivity by precursor ion isolation, dissociation, and measurement of the abundance of fragment ions. The formation of many fragment ions in the collision cell can provide additional selectivity for the respective analyte. A potential issue for flow injection MS/ MS methods can be the in-source (ESI) formation of the same ionic species from different analytes or matrix that are not resolved in time, thus producing an interfering MS/MS response. Note that the recently introduced ambient ionization techniques (e.g., DESI and DART) can also be affected by interferences generated from in-source fragmentation. Highly specific fragmentation pathways should be evaluated and used for analyte quantitation and confirmation by MS/MS, and this is particularly important when chromatographic separation is not used. Individual standards of triflusulfuron-methyl, flupyrsulfuronmethyl, rimsulfuron, sulfometuron-methyl, oxamyl, and methomyl were prepared in methanol at a concentration of 10 ng/mL. Each of the six standard solutions was analyzed separately, and ion transitions for the six analytes were monitored. The selected fragmentation reactions were found to be specific since the monitored fragments were only detected for the appropriate analyte in all cases. All six analytes studied herein are active ingredients, and in-source fragmentation interferences are unlikely to occur between parent compounds (except if isomers are present). This type of selectivity test may be particularly important when metabolites are analyzed in samples that also contain the active ingredient. Molecular ions of active compounds may form the molecular ion of a metabolite through in-source fragmentation, and this should be ruled out prior to developing flow injection MS/MS methods. If this is observed, ion transitions exclusive to each analyte should be selected for adequate specificity. Another selectivity test involved the analysis of control samples for the matrixes of interest. Control samples of tap water, corn, lemon, and pecan were analyzed, and matrix components did not interfere with the ion transition monitored for each analyte, yielding baseline responses comparable to those obtained in solvent injections. The absence of interferences from control matrixes is necessary for the use of matrix-matched standards in flow injection MS/MS methods. The data acquisition of a single ion transition per analyte (as done in this study) seems appropriate for pesticide residue analysis in product development studies if the tests described above are performed and demonstrate adequate selectivity, since the identity of agrochemicals applied to the field is known. Multiple fragment Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3139

Figure 5. Multianalyte flow injection MS/MS ion chronograms36 obtained for corn samples during method validation: (a) control, (b) 0.01 mg/kg fortification, and (c) 0.10 mg/kg fortification. The relative response observed for the analytes was as follows: sulfometuron-methyl ≈ triflusulfuronmethyl > methomyl > rimsulfuron > oxamyl > flupyrsulfuron.

Figure 6. TIC peak area obtained from 100 sequential injections of a 1.0 ng/mL standard in corn matrix. The automated analysis of 100 standard injections required less than 2 h of instrument time.

ions can be monitored for each analyte if needed for enhanced specificity, and this could be particularly important for analyte identity confirmation in multiresidue methods for pesticide screening and enforcement of maximum residue limits. A triple quadrupole mass spectrometer allows a limited number of ion transitions to be monitored simultaneously, and this limitation becomes particularly important when flow injection is used, since all analytes are quickly monitored in a short time. For example, the mass spectrometric acquisition performed in this study used a 20 ms dwell time for each of the 6 monitored MS/MS transitions and the typical analyte flow injection bandwidth was 4 s. These conditions meet the requirement of at least 20 data points across the flow injection band for reliable quantitation. Thus, the dwell time setting and number of analytes are important considerations in flow injection MS/MS quantitative analysis. Modern triple quadrupole mass 3140


spectrometers allow dwell times as short as 2-5 ms for each ion transition monitored; therefore, dozens of fragmentation reactions may still be simultaneously recorded with flow injection MS/MS. This limitation is intrinsic to tandem mass spectrometry, thus is shared with HPLC/MS/MS methods (see a recent article by Fernańdez-Alba and Garcıá-Reyes37 for an objective review of advantages and disadvantages of tandem mass spectrometry in large scale multiresidue methods, LSMRMs). Method Validation. The flow injection MS/MS method developed for sulfonylureas and carbamates was tested using typical extraction and sample purification procedures used at DuPont Crop Protection for residue analysis by HPLC/MS/MS (see Experimental Section for details). Method validation tryouts were performed in three food matrixes: corn, lemon, and pecan, representing crops that may receive applications of sulfonylurea

Figure 7. Pesticide residue quantitative results obtained for lemon samples fortified at (a) 0.01 and (b) 0.10 mg/kg in 10 reinjections of the tryout set. The set reanalyses required a total of 160 injections performed in approximately 3 h of instrument time.

herbicides and/or carbamate insecticides. Validation of a water method was also performed to demonstrate the capability of flow injection MS/MS to meet the lowest limit of quantitation (LOQ) regulatory requirement which is encountered when environmental methods are developed. A 10-fold preconcentration of the analytes was achieved with SPE, which is comparable to typical environmental methods developed for residue analysis. The analyte concentration in the analyzed extracts (nanograms per milliliter) was calculated for each validation set using average mean response factors (peak area/standard concentration) obtained from calibration standards. This concentration was then used to calculate the analyte present in the original food or water sample. Validation tryouts included the analysis of food samples fortified at 0.01 mg/kg (ppm) and 0.10 mg/kg, and water samples fortified at 0.1 ng/mL (ppb) and 1.0 ng/mL. (37) Fernańdez-Alba, A. R.; Garcıá-Reyes, J. F. Trends Anal. Chem. 2008, 27, 973–990.

Results from method validation tryouts are summarized in Table 2 for food matrixes. Suppression of ionization efficiency was observed for all analytes and this is expected since analytes and matrix present in each sample travel together in the injection band entering the electrospray ion source at the same time. Analytical standards ranged from 0.7-20 ng/mL for most validation sets and were prepared in the control matrix to correct for ion suppresion. Acceptable linearity and relative standard deviation (RSD) for response factors were obtained for all analytes (representative calibration curves are available as Supporting Information). A wider calibration range, i.e. 0.5-100 ng/mL was tested during the last validation tryout set (pecan), and calibration results were similar to those obtained for other matrixes. Representative flow injection MS/MS ion chronograms36 for a control corn extract, 0.01 (LOQ) and 0.10 mg/kg (10 × LOQ) fortifications, appear in Figure 5. Results obtained during validation of the water flow injection MS/MS method are also Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3141

summarized in Table 2. Satisfactory linearity and precision (R2 > 0.99 and RSD < 20%) were also obtained for all analytes during validation of the water method. As shown in Table 2, acceptable quantitation results3 (average recovery between 70-120% and RSD < 20%) were obtained for all analytes and matrixes tested using external standards prepared in the control matrix, with the exception of one specific case (triflusulfuron methyl in lemon), demonstrating the capability of flow injection MS/MS for pesticide residue analysis. The limit of quantitation (LOQ) achieved for all analytes was 0.01 mg/kg in food samples and 0.1 ng/mL in water; while limits of detection (LOD) were ∼0.003 mg/kg and 0.03 ng/mL in food and water samples, respectively. The sensitivity achieved allows quantitative measurements to be made at or below the maximum residue limits established for each analyte. For example, the strictest maximum residue limit of any pesticide active ingredient in water is 0.1 ng/mL (groundwater in Europe). In the United States, the maximum residue of rimsulfuron permitted in lemon, pecan, and corn grain is 0.01 mg/kg and for methomyl, 0.1 and 2 mg/kg, in pecan and lemon, respectively. System Stability, Ruggedness, and Repeatability. An experiment was performed to test the stability of the detection system when used with the flow injection sample introduction technique. A total of 100 1 µL injections of a 1.0 ng/mL corn matrix-matched standard were performed, and the TIC peak areas in the ion chronograms were measured (see Figure 6). The system response was found to be stable, with a relative standard deviation (RSD) of 5.7% for peak areas from all injections. A drop of ∼8% in sensitivity was observed by the end of the experiment, and this was likely due to matrix buildup in the transfer capillary line and/ or ion source, thus increasing suppression effects. Note that the flow injection MS/MS analysis of 100 injections performed in 1.8 h is equivalent to samples containing matrix being injected sequentially for 25-50 h using a conventional HPLC/MS/MS method without injection of solvent blanks. Solvent blanks are often injected for system equilibration and flushing within an analytical set (typical set size 15-25 samples) when pesticide residue analysis is performed by HPLC/MS/MS with matrix-matched standards. This practice can be employed in flow injection MS/ MS to reduce the loss of sensitivity overtime. The ruggedness and repeatability of the technique was further tested by reinjecting the entire lemon validation set 10 times to simulate the sequential analysis of multiple analytical sets (lemon seemed to be the most difficult matrix based on the accuracy and precision obtained in initial method validation tryouts (Table 2)). This experiment evaluated the precision of the instrumental analysis technique for quantitation of each sample independently from sample extraction and preparation procedures. The experiment included a total of 160 sequential sample injections (16 per analytical set reinjection) including solvent blanks, control samples, fortifications, and standards prepared in control matrix, simulating the typical configuration used for pesticide residue analysis by HPLC/MS/MS. Briefly, each analytical set started with two equilibration solvent blanks, followed by the lowest calibration standard, control sample,

3142


and the remaining calibration standards with ascending concentration injected every two or three fortified samples. Each set ended with three solvent blanks for system cleaning. An injection volume of 5 µL was used for solvent blanks for better needle rinsing. Representative quantitation results for a sample fortified at 0.01 mg/kg (LOQ) and another at 0.10 mg/kg (10 × LOQ) appear in Figure 7. The comparison of results obtained in reinjected sets show consistent quantitative measurements for all analytes in each sample with better precision obtained for samples at higher concentration (10 × LOQ, Figure 7b). The lower precision obtained for samples fortified at the LOQ can be attributed to concentration measurements based on peaks with lower signal-to-noise. Acceptable linearity and response factor RSDs were also obtained for each reinjection set. The total instrument time needed for the 160 injections was approximately 3 h. Increased sample throughput translates to a more efficient use of mass spectrometers but may also require an increased frequency of maintenance (e.g., ion source cleaning and PEEK capillary replacement). The additional analyte purification provided by HPLC may be appropriate for the toughest samples, like pasture hay and other very dry matrixes which were not tested in this study. CONCLUSIONS The capability of flow injection tandem mass spectrometry for high-throughput multiresidue analysis of pesticides in food and water has been demonstrated. HPLC systems can be used in flow-injection mode, thus the instrumentation needed to employ this technique is already widely available. The methods introduced in this paper allow satisfactory quantitation of sulfonylureas and carbamates with external standards. The use of isotopically labeled internal standards was not evaluated; however, internal standards may improve the accuracy and precision of quantitative analysis and could be appropriate in flow injection MS/MS methods, depending on the analyte and matrix of interest. Current efforts are aimed at improving the initial flow injection MS/MS methods introduced in this paper and expanding the application of this technique to quantitation of other compounds in complex matrixes. In addition, improving sample extraction and cleanup procedures is of great interest to increase sample preparation throughput and therefore to maximize the benefit from the fast instrumental analysis achieved with flow injection MS/MS. ACKNOWLEDGMENT Discussions with James J. Stry, Joseph P. McClory, Teri Quinn Gray, and John H. May (DuPont Co.) prior to publication of this manuscript are acknowledged. The authors also thank J.J.S. for providing typical mass spectrometric parameters for the analysis of sulfonylureas. 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 January 30, 2009. Accepted February 27, 2009. AC900226W

High-Throughput Pesticide Residue Quantitative Analysis Achieved by

Recommend Documents