Confirmation and Quantitation of Selected Sulfonylurea

degree of scientific certainty. Confirmation required that. (1) LC/MS retention times of the analytes be within 1% of the retention times of the stand...
0 downloads 0 Views 296KB Size
Download PDF
Anal. Chem. 1998, 70, 2710-2717

Confirmation and Quantitation of Selected Sulfonylurea, Imidazolinone, and Sulfonamide Herbicides in Surface Water Using Electrospray LC/MS Marisol Rodriguez* and David B. Orescan

DuPont Agricultural Products, E. I. du Pont de Nemours and Company, Experimental Station, Wilmington, Delaware 19880-0402

A multianalyte method has been developed for the confirmation and quantitation of 16 selected sulfonylurea, imidazolinone, and sulfonamide herbicides in surface water. Samples are acidified, the analytes are extracted using RP-102 extraction cartridges, and the extracts are cleaned-up using an anion-exchange cartridge (SAX) stacked on top of an alumina cartridge. The final extracts are evaporated to dryness, redissolved in 10:90 or 20:80 acetonitrile/water, and analyzed using electrospray LC/MS. Confirmation criteria were defined so that identification of the analytes could be made with a reasonable degree of scientific certainty. Confirmation required that (1) LC/MS retention times of the analytes be within 1% of the retention times of the standards, (2) the molecular ion and two characteristic fragment ions per analyte be present, and (3) ion abundance ratios of the fragment ions relative to the molecular ion be within 20% of the ion ratios obtained for the standards. Each of the analytes in all of the samples used to validate this method met these confirmation criteria. Quantitation of the analytes at 0.1 and 1.0 ppb was demonstrated, with average recoveries between 70% and 114% and relative standard deviations that were less than 13%. The limit of quantitation (LOQ) for this method is 0.1 ppb (ng/mL) for all analytes. Six water types were used for the validation of this method.

been analyzed by high-performance liquid chromatography (HPLC),1-3 HPLC with mass spectrometric (MS) detection,4-10 immunoassay,11-15 bioassay,16-19 capillary electrophoresis,20,21 radio immunoassay,3 and, if they are first derivatized, gas chromatography (GC).22-28 Analytical methodologies that have been applied

* To whom correspondence should be addressed. Current address: DuPont Agricultural Caribe Industries, Ltd., P.O. Box 30000, Hwy. 686, Km. 2.3, Manati 00674, Puerto Rico. Fax: (787) 884-1795. E-mail: Marisol.Rodrigues@ pri.dupont.com.

(1) Molins, D.; Wong, C. K.; Cohen, D. M.; Munelly, K. P. J. Pharm. Sci. 1975, 64, 123-124. (2) Zahnow, E. W. J. Agric. Food Chem. 1982, 30, 854-857. (3) Zahnow, E. W. J. Agric. Food Chem. 1985, 33, 479-483. (4) Reiser R. W.; Barefoot, A. C.; Dietrich, R. F.; Fogiel, A. J.; Johnson W. R.; Scott, M. T. J. Chromatogr. 1991, 554, 91-101. (5) Reiser R. W.; Fogiel, A. J. Rapid Commun. Mass. Spectrom. 1994, 8, 252257. (6) Dietrich, R. F.; Reiser, R. W.; Stieglitz, B. J. Agric. Food Chem. 1995, 43, 531-536. (7) Shalaby, L. M.; Reiser, R. W. In Mass Spectrometry of Biological Materials; McEwen, C. N., Larsen, B. S., Eds.; Practical Spectroscopy Series 8; Marcel Dekker: New York, 1990; pp 379-402. (8) Shalaby, L. M. In Pesticide Chemistry: Chemical Analysis; Rosen, J. B., Ed.; Wiley-Interscience: New York, 1987; p 161. (9) Shalaby, L. M.; George, S. W. In Liquid Chromatography/Mass Spectrometry: Applications in Agricultural, Pharmaceutical, and Environmental Chemistry; Brown, M. A., Ed., ACS Symposium Series 420, American Chemical Society: Washington, DC, 1990; pp 75-91. (10) Shalaby, L. M.; Bramble, F. Q.; Lee, P. W. J. Agric. Food Chem. 1992, 40, 513-517. (11) Kelley, N. M.; Zahnow, E. W.; Petersen, W. C.; Toy S. T. J. Agric. Food Chem. 1985, 33, 962-965. (12) Schlaeppi, J.-M. A.; Meyer, W.; Ramsteiner, K. A. J. Agric. Food Chem. 1992, 40, 1093-1098. (13) Schlaeppi, J.-M. A.; Kessler, A.; Fory, W. J. Agric. Food Chem. 1994, 42, 1914-1919. (14) Nord-Christerson, M.; Bergstrom, L. In Proceedings of Brighton Crop Protection Conference on Weeds; The British Crop Protection Council: Surrey, UK, 1989; Vol. 3, p 1127. (15) Brady, J. F.; Turner, J.; Skinner, D. H. J. Agric. Food Chem. 1995, 43, 25422547. (16) Parker, C.; Fryer, Y. O. FAO Plant Prot. Bull. 1975, 23, 83. (17) Sunderland, S. L.; Santelmann, P. W.; Baughmann, T. A. Weed Sci. 1991, 39, 296-298. (18) Gunther, P.; Rahman, A.; Pestemer, W. Weed Res. 1989, 29, 141-146. (19) Rahman, A. In Environmental Bioassay Techniques and Their Application; Munawar, M., Dixon, G., Mayfield, C. I., Reynoldson, T., Sador, M. H., Eds.; Developments in Hydrobiology 54 (reprinted from Hydrobiologia 1989, 188/ 189); Kluwer: Dordrecht, The Netherlands, 1989; pp 1365-1375. (20) Dinelli, G.; Vicari, A.; Catizone, P. J. Agric. Food Chem. 1993, 41, 742746. (21) Dinelli, G.; Vicari, A.; Brandolini, V. J. J. Chromatogr. A 1995, 700, 201207. (22) Ahmad, J. J. Assoc. Off. Anal. Chem. 1987, 70, 745-748. (23) Ahmad, I.; Crawford, G. J. Agric. Food Chem. 1990, 38, 138-141. (24) Klaffenbach, P.; Holland, P. T. J. Agric. Food Chem. 1993, 41, 388-395.

2710 Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

S0003-2700(97)01128-1 CCC: $15.00

Sulfonylureas, imidazolinones, and sulfonamides are potent herbicides that share a common mode of action. They all inhibit acetolactate synthase (ALS), which is an essential enzyme in the production of the branched-chain amino acids, valine, leucine, and isoleucine. These ALS-inhibiting herbicides are characterized by relatively low application rates, typically less than 100 g of active ingredient per hectare. Consequently, the expected environmental concentrations at which these analytes might be found makes their detection and analysis difficult compared to that of traditional herbicides. The analytical techniques applied to the analysis of these compounds have been varied. Sulfonylurea herbicides have

© 1998 American Chemical Society Published on Web 05/13/1998

to the analysis of imidazolinones include HPLC/UV,29-33 LC/MS,34 capillary zone electrophoresis (CZE)/UV,35 and, if the compounds are methylated, GC with phosphorus detection36 and GC/MS.37 Sulfonamides have been analyzed using HPLC/UV,38-39 LC/ MS,40-41 bioassay,42 and, when derivatized with methyl iodide, GC/ MS.43 For confirmatory and quantitative trace-level analyses of thermally labile or nonvolatile compounds, such as the compounds of interest in this work, the technique of choice is LC/MS, because it offers chromatographic separation without thermal decomposition, high sensitivity, mass selectivity, and structural information. Three multiresidue methods using LC/MS and one using LC/ MS/MS have been developed for the determination of twelve,44 eight,45 seven,46 and eight47 sulfonylureas, respectively, in water. The use of LC/MS and LC/MS/MS has also been reported for the analysis of six imidazolinones in water.34 One LC/MS method has been published in the literature for the analysis of flumetsulam (a sulfonamide herbicide) in water.44 To identify an analyte with the greatest degree of certainty, a full-scan mass spectrum of the compound should be compared and matched to that of a standard. However, for trace analysis when a full-scan mass spectrum cannot be obtained, it is acceptable that selected ions from the analyte be monitored for confirmation purposes. To date, no universally accepted criteria have been established for the unequivocal identification of a compound using selected ion monitoring (SIM) with GC/MS, LC/ MS, and MS/MS. However, a number of factors have been commonly used in the literature as proof for confirmation since (25) Klaffenbach, P.; Holland, P. T. J. Agric. Food Chem. 1993, 41, 396-401. (26) Klaffenbach, P.; Holland, P. T. Biol. Mass. Spectrom. 1993, 22, 565-578. (27) Colterill, E. G. Pestic. Sci. 1992, 34, 291-295. (28) Meyer, M. T.; Mills, M. S.; Thurman, E. M. J. Chromatogr. 1993, 629, 5559. (29) Devine, J. M. In Residue Analysis: The Imidazolinone Herbicides; Shaner, D. L., O’Connor, S. L., Eds.; CRC: Boca Raton, FL, 1991. (30) Wells, M. J. M.; Michael, J. L. J. Chromatogr. Sci. 1987, 25, 345-350. (31) Reddy, K. N.; Locke, M. A. Weed Sci. 1994, 42, 249-253. (32) Curran, W. S.; Liebl, R. A.; Simmons, F. W. Weed Sci. 1992, 40, 482-489. (33) Loux, M. M.; Reese, K. D. Weed Sci. 1992, 40, 490-496. (34) Stout, S. J.; daCunha, A. R.; Picard, G. L.; Safapour, M. M. J. Agric. Food Chem. 1996, 44, 2182-2186. (35) Safapour, M. M.; Picard, G. L.; Cavalier, T. C.; Nejad, H.; Zeng, M.; Safarpour, D.; Souza, M.; Krynitsky, A. J. Presented at the 211th ACS National Meeting, New Orleans, LA March 24-29, 1996. (36) Mortimer, R. D.; Weber, D. F. J. AOAC Int. 1993, 76, 377-381. (37) Stout, S. J.; DaCunha, A. R.; Allardice, D. G Anal. Chem. 1996, 68, 653658. (38) Lehmann, R. G.; Miller, J. R.; Fontaine, D. D.; Laskowski, D. A.; Hunter, J. H.; Cordes, R. C. Weed Res. 1992, 32, 197-205. (39) Baskaran, S.; Lauren, D. R.; Holland, P. T. J. Chromatogr. A. 1996, 746, 25-30. (40) Henion, J. D.; Thomson, B. A.; Dawson, P. H. Anal. Chem. 1982, 54, 451. (41) Perkins, J. R.; Parker, C. E.; Tomer, K. B. J. Am. Soc. Mass. Spectrom. 1992, 3, 139-149. (42) Rahman, A.; James, T. K.; Baskaran, S.; Holland, P. T,; Lauren, D. R. Proceedings of the 2nd International Weed Control Congress, Copenhagen, June 1996. (43) Lehman, R. G.; Fontaine, D. D.; Olberding, E. L. Weed Res. 1993, 33, 187195. (44) Krynitsky, A. J. J. AOAC. Int. 1997, 80, 392-400. (45) Volmer, D.; Wilkes, J. G.; Levsen, K. Rapid Commun. Mass Spectrom. 1995, 9, 767-771. (46) DiCorcia, A.; Crescenzi, C.; Samperi, R.; Scappaticcio, L. Anal. Chem. 1997, 69, 2819-2826. (47) Demers, R.; Mendonca, M.; Fenwick, J.; Matassa, L. A Validated APCI/ MS/MS Method for the Quantitative Determination of Trace Levels of Sulfonylurea Herbicides in Soil and Water; Internal Document, Novamann International.

Sphon in 197848 defined specific requirements for the confirmation of drug residues by GC/electron impact MS. These factors include the matching of the compound retention time with a standard and the presence of three or more characteristic ions with the appropriate relative abundance compared to a standard.49-57 A number of reviews describing the development of criteria for confirmation using GC/MS, LC/MS, and MS/MS have been published.58-61 The original guideline of a minimum of three ions for confirmation has been widely adopted. Today, the majority of confirmations reported in the literature have been performed with three or more ions. In this paper, we describe a multianalyte method for the confirmation and quantitation of 16 selected sulfonylurea, imidazolinone, and sulfonamide herbicides (Figure 1) in surface water using electrospray LC/MS. To our knowledge, this is the most comprehensive multianalyte method developed to quantitate and confirm, with a reasonable degree of scientific certainty, this number of low use rate herbicides (twelve sulfonylureas, three imidazolinones, and one sulfonamide) in water. The major advantage of this method over other methods developed strictly for an individual compound or type of compounds (e.g., sulfonylureas or imidazolinones) is that simultaneous information is provided about a much greater number and type of compounds that could potentially be found in the same agricultural area; other advantages are time and money savings for the laboratory performing the analysis. Additionally, this method, unlike other water methods developed for the analysis of these compounds, provides a set of criteria to confirm with certainty the presence of the analytes. This method is intended for use on agricultural runoff waters or waters from other rural areas, not for wastewaters in industrial and municipal areas or other sources. This method was developed as part of an EPA-industry joint effort to develop water methods for selected ALS-inhibiting herbicides.44 EXPERIMENTAL SECTION Reagents and Chemicals. Analytical and technical grade standards of imazapyr (99.7%, 98.6%), imazethapyr (99.7%, 99.3%), and imazaquin (99.6%, 97.4%) were provided by American Cyanamid Co., Agricultural Products Research Division, Princeton, NJ. Flumetsulam (99.6%, 98.0%) analytical and technical grade standard was provided by DowElanco, Indianapolis, IN. Analytical and (48) Sphon, J. A. J. Assoc. Off. Anal. Chem. 1978, 61, 1247-1250. (49) USDA. Analytical Chemistry Guidebook-Residue Chemistry; Food Safety and Inspection Service: Washington, DC, 1991; ORP2-7. (50) Edlund, P. O.; Lee, E. D.; Henion, J. D.; Budde, W. L. Biomed. Environ. Mass. Spectrom. 1989, 18, 233-240. (51) Schneider, R. P.; Ericson, J. F.; Lynch, M. J.; Fouda, H. G. Biol. Mass. Spectrom. 1993, 22, 595-599. (52) Hurtaud, D.; Delepine, B.; Sanders, P. Analyst 1994, 119, 2731-2736. (53) Wilson, R. T.; Wong, J.; Johnston, J.; Epstein, R.; Heller, D. N. J. AOAC Int. 1994, 77, 1137-1142. (54) Hornish, R. E.; Cazers, A. R.; Chester, S. T.; Roof, R. D. J. Chromatogr. B 1995, 674, 219-235. (55) Ioerger, B. P.; Smith, J. S. J. Agric. Food Chem. 1994, 42, 2619-2624. (56) Schilling, J. B.; Cepa, S. P. Menacherry, S. D.; Barda, L. T.; Heard, B. M.; Stockwell, B. L. Anal. Chem. 1996, 68, 1905-1909. (57) Li, L. Y. T.; Campbell, D. A.; Bennett, P. K.; Henion, J. D. Anal. Chem. 1996, 68, 3397-3404. (58) Gilbert, J.; Startin, J. R.; Crews, C. Pestic. Sci. 1987, 18, 273-290. (59) Cairns, T.; Siegmund, E. G.; Stamp, J. J. Mass Spectrom. Rev. 1989, 8, 93117. (60) Cairns, T.; Siegmund, E. G.; Stamp, J. J. Mass Spectrom. Rev. 1989, 8, 127145. (61) Cairns, T.; Siegmund, E. G. Crit. Rev. Food Sci. Nutrit. 1991, 30, 397-402.

Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

2711

Figure 1. Structures, common names, and trade names of the 16 target analytes. 1, Imazapyr (Arsenal); 2, flumetsulam (Broadstrike); 3, imazethapyr (Pursuit); 4, nicosulfuron (Accent); 5, imazaquin (Scepter); 6, thifensulfuron methyl (Pinnacle); 7, metsulfuron methyl (Ally); 8, sulfometuron methyl (Oust); 9, chlorsulfuron (Glean); 10, triasulfuron (Amber); 11, bensulfuron methyl (Londax); 12, prosulfuron (Exceed); 13, halosulfuron methyl (Permit); 14, chlorimuron ethyl (Classic); 15, triflusulfuron methyl (Upbeet); and 16, primisulfuron methyl (Beacon).

technical grade standards of triasulfuron (95.0%, 95.0%), primisulfuron methyl (95.9%, 95.0%), and prosulfuron (99.5%, 96.0%) were provided by Ciba Geigy Corp., Greensboro, NC. Analytical and technical grade standard of halosulfuron methyl (99.5%, 99.0%) was provided by Monsanto Agricultural Co., St. Louis, MO. Analytical standards of nicosulfuron (98.3%), thifensulfuron methyl (99.7%), metsulfuron methyl (99.0%), sulfometuron methyl (99.2%), chlorsulfuron (99.2%), bensulfuron methyl (99.4%), chlorimuron ethyl (98.8%), and triflusulfuron methyl (98.8%) were obtained from DuPont Agricultural Products, Experimental Station, Wilmington, DE. Acetonitrile, water, methanol, and dichloromethane were Omni Solv HPLC grade solvents from EM Science, Gibbstown, NJ. Deionized water used for sample preparation was obtained from a Milli-Q UV-Plus water purification system (Millipore Corp. Bedford, MA). Glacial acetic acid was purchased from EM Science. HPLC/MS. A Hewlett-Packard HP 5989B MS engine singlequadrupole mass spectrometer equipped with a Hewlett-Packard HP 59987A API-electrospray MS interface was used for the analyses. The HPLC system was a Hewlett-Packard 1090 L Series II with a 79883A diode array detector, a 79847A autosampler, a 79846A autoinjector, a 79835A (DR5 SDS) solvent delivery system with a static mixer installed, and a chilled autosampler compartment. The compounds were separated by HPLC using a Zorbax RXC8 column, 25 cm × 4.6 mm, 5-µm particle size (MAC-MOD Analytical, Inc., Chadds Ford, PA), protected by a Zorbax RX-C8 guard column, 4.6 mm × 12.5 mm, 5-µm particle size (MAC-MOD 2712

Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

Analytical, Inc.). The initial mobile phase composition was 20:80 acetonitrile/0.15% acetic acid in water, which was held constant from 0.00 to 10.00 min. A gradient was then programmed to increase the amount of acetonitrile from 20% to 36.8% in 21 min and from 36.8% to 50% in 1 min. The acetonitrile composition was kept constant at 50% for 12 min. To clean the column from strongly retained substances, the amount of acetonitrile was increased to 90% and was left at that concentration for 3 min. The initial mobile phase composition was restored, and the column was equilibrated for 7 min. The column temperature was 38 °C, and the injection volume was 100 µL. These LC conditions were developed by Roby and Chou62 from the Science Applications International Corp. (SAIC) in San Diego, CA. The flow rate in the HPLC was 1.0 mL/min, but the effluent from the HPLC was split 20:80 using a PEEK tee (Upchurch Scientific, Oak Harbor, WA, P-727) to allow only a flow of 0.2 mL/min into the MS. The mass spectrometer was operated in the positive ion mode using time-programmed capillary exit voltages (CapEx) and timescheduled (programmed) selected ion monitoring (SIM). To generate fragment ions in addition to the molecular ion [MH]+ from each compound, collision-induced dissociation (CID) was optimized by changing the CapEx throughout the chromatographic run. The voltage difference between the CapEx and the (62) Roby, M.; Chau, N. Evaluation of Improved Analytical Methods for the Determination of Sulfonylurea and Imidazolinone Herbicides in Aqueous Samples; Science Applications International Corporation Document submitted to U.S. EPA, Office of Pesticide Programs, Contract No. 68-D2-0183, June 19, 1996.

Table 1. Positive Ions Selected for Monitoring and Their Relative Abundance compound imazapyr (1) flumetsulam (2) imazethapyr (3) nicosulfuron (4) imazaquin (5) thifensulfuron methyl (6) metsulfuron methyl (7) sulfometuron methyl (8) chlorsulfuron (9) triasulfuron (10) bensulfuron methyl (11) prosulfuron (12) halosulfuron methyl (13) chlorimuron ethyl (14) triflusulfuron methyl (15) primisulfuron methyl (16)

MW 261 325 289 410 311 387 381 364 357 401 410 419 434 414 492 468

ions(relative abundance) 217 (29), 234 (33), 262 (100) 129 (45), 262 (30), 326 (100) 177 (18), 245 (22), 290 (100) 182 (46), 213 (32), 411 (100) 199 (15), 267 (21), 312 (100) 141 (36), 167 (34), 388 (100) 141 (34), 167 (43), 382 (100) 150 (91), 199 (19), 365 (100) 141 (65), 167 (21), 358 (100) 141 (31), 167 (11), 402 (100) 149 (49), 182 (34), 411 (100) 141 (25), 167 (7), 420 (100) 182 (50), 403 (19), 435 (100) 186 (37), 213 (6), 415 (100) 264 (26), 461 (6), 493 (100) 199 (17), 254 (40), 469 (100)

first skimmer determines the amount of fragmentation. The difference in voltages generates enough energy to break bonds by accelerating ions into collisions with molecules of the drying gas. Since the skimmer voltage was a fixed value during tuning, the amount of fragmentation was controlled by the CapEx. The CapEx was chosen so that each compound would produce at least three ions, with the molecular ion having the highest abundance and the fragment ions having abundances g15% relative to the molecular ion. For three of the 16 compounds, prosulfuron, chlorimuron ethyl, and triflusulfuron methyl, the relative abundance of the least abundant ion (m/z 167, 213, and 461, respectively) was between 5% and 10%. Although lower than ideal, the relative abundances for these ions were reproducible; thus, these ions were deemed acceptable for confirmation. To increase sensitivity, selected ion monitoring (SIM) was used in this method. Programmed SIM was used, where specific ions were monitored for each compound throughout the chromatographic run. In some cases, the ions for two compounds were monitored simultaneously because the compounds eluted very closely to each other. Table 1 shows the positive ions selected for monitoring each of the 16 compounds. The SIM and CapEx time programs are shown in Tables 2 and 3, respectively. The mass spectrometer was tuned using a 5 µg/mL solution of imazapyr, metsulfuron methyl, and triflusulfuron methyl directly infused from a syringe at a flow rate of 20 µL/min, combined with a stream of 20:80 acetonitrile/0.15% acetic acid in water from the HPLC at a flow rate of 0.2 mL/min. Since the instrument cannot be tuned to optimize the signals for ions of all mass/charge ratios (m/z), three ions with low, medium, and high m/z were selected for the tuning. The ions used for the tuning were the protonated molecular ions for imazapyr (m/z 262), metsulfuron methyl (m/z 382), and triflusulfuron methyl (m/z 493). The dwell time per ion was set at 100 ms, the quadrupole temperature was 150 °C, and the electron multiplier voltage was 2504 V. At the electrospray cabinet, the nitrogen flow was set to 40 (unitless) and the drying gas heater to 300 °C. At the spray chamber, the nebulizing gas pressure was set to 80 psi. Preparation of Standard Solutions. Individual stock solutions (500 µg/mL) of each analytical standard were prepared in acetonitrile. Then, 1.0 mL of each individual stock solution was pipetted into a 100-mL volumetric flask and diluted with acetoni-

Table 2. Selected Ion Monitoring (SIM) Time Program time (min) 0.01 18.50 20.20 23.00 26.00 29.40 31.00 32.50 33.65 37.00 40.00 41.90 43.50

ions monitored [MH]+

217, 234, 262 129, 262, 326 [MH]+ 177, 245, 290 [MH]+ 182, 213, 411 [MH]+ 199, 267, 312 [MH]+ 141, 167, 388 [MH]+ 141, 167, 382 [MH]+ 150, 199, 365 [MH]+ 141, 167, 358 [MH]+ 141, 167, 402 [MH]+ 149, 182, 411 [MH]+ 141, 167, 420 [MH]+ 182, 403, 435 [MH]+ 186, 213, 415 [MH]+ 264, 461, 493 [MH]+ 199, 254, 469 [MH]+

compound(s) monitored imazapyr (1) flumetsulam (2) imazethapyr (3) nicosulfuron (4) imazaquin (5) thifensulfuron methyl (6) metsulfuron methyl (7) sulfometuron methyl (8) chlorsulfuron (9) triasulfuron (10) bensulfuron methyl (11) prosulfuron (12) halosulfuron methyl (13) chlorimuron ethyl (14) triflusulfuron methyl (15) primisulfuron methyl (16)

Table 3. Capillary Exit (CapEx) Voltage Time Program time (min)

CapEx (V)

time (min)

CapEx (V)

time (min)

CapEx (V)

0.01 20.20 23.00

150 160 130

26.00 29.40 32.50

160 125 135

33.65 37.00 40.00

125 140 125

trile to obtain a 5 µg/mL combined standard stock solution. The 5 µg/mL standard stock was appropriately diluted to prepare two sets of four working standards; 20.0, 50.0, 75.0, and 100.0 ng/mL and 150.0, 200.0, 250.0, and 300.0 ng/mL to analyze the 0.1 and 1.0 ppb fortified samples, respectively. The final composition of the working standard solutions was 10:90 or 20:80 acetonitrile/ water. To prepare the fortification standard (500.0 ng/mL), 1.0 mL of the 5 µg/mL standard stock solution was pipetted into a 10-mL volumetric flask and diluted with acetonitrile. Source of Samples. Water from six sources was used for this method validation. River water was obtained from the Brandywine River, Wilmington, DE. Pond water was obtained from Lums Pond State Park, Kirkwood, DE. Lake water was obtained from an artificial lake, Hoopes Reservoir, Wilmington, DE. Creek water was obtained from White Clay Creek, Newark, DE. Marsh water was obtained from Knowles Marsh, Beltsville, MD, and tap water was obtained from a laboratory at the DuPont Experimental Station (Building 402), Wilmington, DE. Water samples were stored in a refrigerator if the samples were not going to be analyzed the day they were collected. The tap water was degassed with helium before fortification to eliminate chlorine, which can react with and degrade some of the compounds. Validation samples were filtered after they were fortified. Water Samples. Two hundred and fifty grams ((0.1 g) of water was weighed into a beaker. Each water sample was fortified according to the sample fortification procedure described below and filtered using a Costar 250-mL bottle filter system (Corning Costar Corp., Cambridge, MA). Two and a half milliliters of glacial acetic acid was added to the water sample immediately prior to the extraction procedure. Sample Fortification Procedure. To validate this method, water samples were fortified in triplicate at two different levels: at the LOQ (0.1 ppb) and at 10 times the LOQ (1.0 ppb). A typical Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

2713

set of validation samples for each matrix consisted of two control samples, six fortified samples (three at each of two levels), and three or more solvent blanks (10:90 or 20:80 acetonitrile/water). For the last four water matrixes used for the validation of this method, half of the sample set (one control, three fortifications at the same level, and three solvent blanks) was extracted and analyzed on one day and the other half on a different day. For marsh water, only duplicate fortifications were made at each fortification level because of the limited amount of water available. Fortified samples were prepared by adding 0.05 or 0.5 mL of the 500.0 ng/mL fortification standard solution to 250 mL of water, resulting in fortification levels of 0.1 and 1.0 ppb (ng/mL), respectively. Sample Extraction and Cleanup Procedures. The extraction and cleanup procedures used in this method were developed for capillary electrophoresis determination of sulfonylureas by A. J. Krynitsky from the U.S. Environmental Protection Agency (EPA) Office of Pesticide Programs (OPP), Beltsville Laboratories.44 This procedure was first modified to be used for HPLC/ UV analysis by Roby and Chau62 from the Science Applications International Corporation (SAIC) in San Diego, CA. Further modifications to the extraction and cleanup procedure were made in our laboratory. Sample Extraction. The RP-102 cartridges, prepacked with a polystyrene divinylbenzene resin (Spe-ed RP-102 Cartridges, 500 mg/mL, Applied Separations, Allentown, PA), were attached to a vacuum manifold apparatus (Visiprep DL, Supelco, Inc., Belfonte, PA), the apparatus was connected to a vacuum pump, and the cartridges were conditioned first with three 6-mL portions of methanol and then with three 6-mL portions of 1% (v/v) acetic acid in water. The packing was not allowed to go dry until the end of the sample extraction step. After conditioning, the cartridge was filled with 4 mL of 1% (v/v) acetic acid. The water samples were acidified with 2.5 mL of glacial acetic acid, mixed well, and then immediately passed through the RP-102 cartridges using Visiprep large-volume samplers (Supelco Inc.). Sufficient vacuum was applied to draw the samples through the cartridges at a rate of 1-2 drops/s. The beakers were rinsed with 5-10 mL of 1% (v/v) acetic acid solution, and the rinses were allowed to flow through the RP-102 cartridges. The samplers were removed and the cartridges washed with 6 mL of 1% (v/v) acetic acid at a rate of 1-2 drops/s. Air was then pulled through the cartridges for 30 min. The analytes were eluted from the RP-102 cartridges with 10 mL of methanol at a rate of 1 drop/s. The eluates were collected in disposable culture tubes. The extracts were evaporated just to dryness using an N-Evap with a water bath temperature of 38 °C. The walls of the tubes were rinsed with 5 mL of acetonitrile, and the extracts were evaporated to dryness. (This is a critical step. The residue must not contain even a trace amount of water for the subsequent cleanup procedure to be successful.) Sample Cleanup. (We found that cleanup of the extracts was necessary to avoid buildup of material in the MS capillary and, therefore, the need for constant cleaning.) The SAX cartridges (Isolute SAX cartridges, 1 g/6 mL, International Sorbent Technology, available from Jones Chromatography, Lakewood, CO) were conditioned with three 6-mL portions of methanol. The packing was not allowed to go dry once conditioning started. Separately, 2714 Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

the alumina cartridges (Bakerbond SPE alumina cartridges, Al2O3 neutral, 6-mL HC, available from VWR, San Francisco, CA) were also conditioned with three 6-mL portions of methanol as described for the SAX cartridges. Following the last addition of methanol, the cartridges were allowed to remain about half full. The SAX cartridges were stacked on top of the alumina cartridges using adapters. The extract residues from the RP-102 extraction were dissolved in 10 mL of methanol. These samples were sonicated for 2 min and gently vortexed for 5 s. The 10-mL sample extracts were added to the SAX/alumina stacks, and the extracts were allowed to elute at a rate of 1 drop/s. The packing was not allowed to go dry. The disposable culture tubes were rinsed with four 5-mL portions of methanol, and the rinses were added to the SAX/ alumina stack. The rinses were allowed to elute at a rate of 1 drop/s. After the last rinse, the column stacks were allowed to go to dryness, the SAX cartridges were discarded, and air was pulled through the remaining alumina cartridges for 3 min. Three milliliters of 0.5% (v/v) acetic acid in dichloromethane was added. The dichloromethane solutions were allowed to thoroughly wet the packing. (The dichloromethane solution wet the packing without vacuum.) After the cartridges were soaked for at least 3 min, the dichloromethane solution was allowed to drip down just to the top of the column head at a rate of 2 drops/s. The eluants were discarded. The analytes were then eluted with 17 mL of dichloromethane solution and collected in disposable culture tubes at a rate of 2 drops/s. At the end of the elution, vacuum was applied to make sure all the dichloromethane was eluted. The extracts were evaporated just to dryness using an N-Evap with a bath at of 38 °C. The walls of the tubes were rinsed with 5 mL of dichloromethane, and the extracts were evaporated to dryness. The residues were dissolved in 1 mL of 10:90 or 20:80 acetonitrile/ water and sonicated for 2 min. (Initially during validation, standards and samples were dissolved in 10:90 acetonitrile/water, but we decided to increase the acetonitrile content and dissolve the samples and standards with 20:80 acetonitrile/water because we started to observe undissolved white residues in the final culture tubes with the extracts.) A small magnetic stir bar was inserted into the tubes, and the samples were agitated on a stir plate for 5 min at ambient temperature. The samples and standard solutions were filtered through a 0.45-µm filter unit (e.g., Acrodisk 13 CR PTFE, Gelman Corp, Ann Arbor, MI, or equivalent) before analysis. Calibration Procedure. The external standard method of calibration was used for this method. At least four standard solutions containing all 16 compounds, at levels which bracketed the concentration of the fortifications, were analyzed at the beginning and end of an automated sequence. Determination of the amount of each compound in a sample can be done using calibration curves prepared from the analysis of standard solutions or response factors of the standards. Calibration curves are prepared for each compound by plotting the average total ion peak area versus the compound’s concentration in nanograms per milliliter. Using linear regression, the equation for the line (slope and intercept) that best fits the experimental calibration data is determined. The correlation coefficient for each curve should be equal to or greater than 0.9900.63 The correlation coefficient should not be the only factor

Figure 2. Total ion chromatogram (TIC) of a 20 ng/mL standard solution containing the 16 target analytes.

used to determine the linearity of the calibration curve. Some calibration curves have correlation coefficients of 0.9990 and may not be linear. Response factors are a more sensitive measure of linearity than correlation coefficients.64 For a linear calibration curve with intercept ) 0, the response factors should remain constant with concentration. Response factors can also be used to determine the analyte concentration in the samples. The use of response factors is valid if linearity has been established for the calibration curves (correlation coefficients equal to or greater than 0.9900) and the intercepts are not statistically different from zero. To use response factors, first, the total ion chromatogram (TIC) peak areas for both standard injections at a given concentration are averaged. Then, the response factor (RF), defined as the average peak area divided by the analyte concentration, is calculated for each standard, and finally an average response factor is calculated by averaging the response factors of all standards. The average response factor relative standard deviation should be less than 15% to assume linear response. For the reason described above, determination of the amount of each compound in a sample was done using response factors of the standards, except when the calibration line intercepts were statistically different from zero, in which case the calibration curves (slope and intercept) were used. Standard and Sample Analysis. Standards and samples were analyzed in order of increasing concentration. The standards were analyzed at the beginning and end of an automated sequence to confirm their stability as well as that of the instrument. The standards showed less than a 20% change in peak area response over the course of the analysis of the sample set. (63) Miller, J. C.; Miller, J. N. Statistics for Analytical Chemistry, 3rd ed.; Ellis Horwood PTR Prentice Hall: New York, 1993; Chapter 5. (64) Coopersmith, B. I.; Hamilton, B. H.; Abolin, C. R. (Sandoz Pharmaceutical). Presented at The 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, May 21-26, 1995.

Confirmation Criteria/Qualitative Determination. For identification of an analyte in surface water samples using this method, three factors must be considered and each criterion met: (1) retention times of the analytes are within (1% of the retention times of the corresponding standards, (2) the presence of three characteristic SIM ions (molecular ion and two fragment ions) is monitored per analyte, and (3) the ion abundance ratios, relative to the most abundant ion (the molecular ion, [MH]+), are within (20% relative difference of the ion ratios monitored for the corresponding standards. These criteria used for confirmation were derived, in part, from guidelines for GC/electron impact MS48,49 and from recently published suggested criteria for LC/thermospray MS55 and ion spray tandem mass spectrometry (LC/MS/MS)56,57 experiments. To verify the presence of the monitored ions for each compound, an average SIM mass spectrum was obtained across the peak of each compound in the TIC, and an average SIM spectrum of the background (near the peak) was subtracted from it. This background-subtracted SIM spectrum of the compound was used to calculate the relative abundance of the fragment ions by dividing the abundance of each fragment ion by the abundance of the molecular ion and multiplying by 100. To obtain an accurate mass spectrum of the analytes and, therefore, an accurate value of the relative abundance of the fragment ions, it was crucial that the background chosen to be subtracted from each peak spectrum was as clean and stable as possible. This was particularly important when analytes were present at low concentrations. Also, it is important to know when the background region for each peak starts and ends (see SIM and CapEx time programs, Tables 2 and 3), since different ions are monitored throughout the chromatographic run. To avoid this background selection issue, individual ion chromatograms can be used for the calculation of the relative abundance of fragment ions. During the early stages of this work, subtracted SIM spectra and individual ion chromatograms were used to determine the relative abundance of fragment Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

2715

ions, and no differences were found in the results. Subtracted SIM spectra were selected at that time for the determination of relative abundance because it was faster to process them with the data analysis software available. RESULTS AND DISCUSSION Each of the analytes in all of the samples analyzed using this method met the confirmation criteria. The three ions monitored for each compound were present in all samples. The average ion abundance ratio for each ion of each compound was within (20% of that of the standard. An average relative abundance for each of the fragment ions has been calculated on the basis of the relative abundance of these ions in the standards and samples in all six matrixes (Table 1). The retention times of the compounds did not deviate by more than 0.3% from the retention times of the corresponding standards. During method validation, we found that, in general, at the 0.1 ppb level, the relative abundance of the fragment ions was more variable and more dependent on background selection than at the 1.0 ppb level. For this reason, if a substance in a sample meets the criteria for the retention time and has the three characteristic ions, but does not meet the criterion for the relative abundance of the fragment ions, we recommend that the same sample and one of the standards (closest concentration to sample) originally analyzed be concentrated and reinjected to verify the ratio of the ions, or that individual ion chromatograms be used instead for the determination of the relative abundance of fragment ions. The linearity of the detector’s response was evaluated using the average response factor relative standard deviations as well as linear regression for the 10 sets of standards used for the validation of this method. The average response factor relative standard deviation for each compound in each set of standards was below 15%. Using linear regression, all 160 calibration curves (16 compounds × 10 sets of standards) had correlation coefficients g0.9900, with the exception of two calibration curves, one for bensulfuron methyl and one for sulfometuron methyl, which had correlation coefficients of 0.98. A typical total ion chromatogram (TIC) of a standard solution is shown in Figure 2. Total ion chromatograms of control samples from the sources used to validate this method were free of interference in the region where compounds elute. A typical total ion chromatogram (TIC) of an unfortified pond water sample is shown in Figure 3A. Based on the control matrixes used during the validation of this method, matrix interferences are not expected to be present. Due to the specificity of the mass spectrometer, other sulfonylurea, imidazolinone, and sulfonamide herbicides should not interfere. The following compounds were tested as possible interferences to demonstrate the specificity of the method: ethametsulfuron methyl, rimsulfuron, tribenuron methyl, flupyrsulfuron methyl, azimsulfuron, imazameth, and imazamox. The first five compounds are sulfonylureas, and the last two are imidazolinones. Ethametsulfuron methyl, azimsulfuron, imazameth, and imazamox did not coelute with any of the 16 target analytes. Rimsulfuron coeluted with triasulfuron, but since rimsulfuron did not generate the same characteristic ions that were monitored for triasulfuron, we do not expect it to affect the detection and quantitation of triasulfuron. Tribenuron methyl eluted close to bensulfuron methyl, and flupyrsulfuron methyl eluted close to prosulfuron; 2716 Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

Figure 3. Total ion chromatograms (TICs) of (A) unfortified, (B) 0.1 ppb fortified, and (C) 1.0 ppb fortified pond water samples.

however, we do not expect them to pose any interference problems because they did not generate the same characteristic ions that were monitored for bensulfuron methyl and prosulfuron, respectively. This method was validated using river water, pond water, lake water, creek water, marsh water, and tap water. Water samples were fortified with 16 selected sulfonylurea, imidazolinone, and sulfonamide herbicides at both 0.1 and 1.0 ppb. Typical total ion chromatograms of fortified pond water samples are shown in Figure 3B,C. Chromatograms for other fortified water samples are not shown because they are similar to those obtained for the pond water samples. Average recoveries and relative standard deviations (RSDs) are summarized in Table 4. These results are based on 34 recovery determinations in six types of water fortified at 0.1 and 1.0 ppb. The method generated acceptable recoveries at the LOQ and 10 times the LOQ, with averages ranging from 70% to 114% and RSDs ranging from 3% to 13%. Recoveries for imazapyr, imazethapyr, and imazaquin were lower than those for the other compounds. The limit of quantitation (LOQ) is defined as the lowest fortification level evaluated at which acceptable average recoveries

Table 4. Average Recoveries and Relative Standard Deviations (Precision) for the 16 Target Analytes average recovery (RSD) (n ) 17) compound

0.1 ppb

1.0 ppb

imazapyr (1) flumetsulam (2) imazethapyr (3) nicosulfuron (4) imazaquin (5) thifensulfuron methyl (6) metsulfuron methyl (7) sulfometuron methyl (8) chlorsulfuron (9) triasulfuron (10) bensulfuron methyl (11) prosulfuron (12) halosulfuron methyl (13) chlorimuron ethyl (14) triflusulfuron methyl (15) primisulfuron methyl (16)

75 (13) 114 (9) 71 (10) 97 (7) 78 (5) 104 (12) 104 (9) 104 (7) 103 (11) 109 (8) 106 (5) 95 (8) 89 (10) 96 (9) 95 (7) 89 (10)

72 (8) 102 (4) 70 (7) 92 (5) 77 (4) 98 (3) 97 (3) 97 (4) 97 (6) 99 (5) 98 (4) 93 (5) 85 (12) 95 (5) 92 (6) 87 (7)

and precision (70-120% and RSD e 20%, respectively) are demonstrated. The average recoveries at 0.1 ppb in the six types of water used to validate this method were between 71% and 114%, with RSDs e 13%. In this method, the limit of quantitation also represents the fortification level at which the smallest analyte peak of the 16 compounds (flumetsulam) consistently generated a signal

about 10 times the noise background of the control matrix. The limit of quantitation of this method is 0.1 ppb. The limit of detection was not determined for this method. The limit of detection can be estimated to be 2-3 times the noise background of the control matrix, which would be approximately 0.02-0.03 ppb in the absence of interfering peaks. CONCLUSION This method has clearly demonstrated good recoveries (70114%), good precision (3% e RSD e 13%), good sensitivity (LOQ ) 0.1 ppb), ruggedness (validated in six matrixes), and the ability to confirm with certainty the presence of 16 analytes. This method has been submitted for AOAC interlaboratory validation. ACKNOWLEDGMENT Special thanks to Drs. Aldos Barefoot, Mary Ellen McNally, Lamaat Shalaby, and Michael Duffy for their helpful discussions and Dr. Alex Krynitsky, from the U.S. EPA Office of Pesticide Programs, for supplying the marsh water sample for the validation of this method.

Received for review October 10, 1997. Accepted March 26, 1998. AC971128A

Analytical Chemistry, Vol. 70, No. 13, July 1, 1998

2717