Determination of Glyphosate and Aminomethylphosphonic Acid

Jul 11, 2000 - José Bernal , María T. Martin , María E. Soto , María J. Nozal , Ilaria .... on physiological responses of the pacific oyster, Cras...
1 downloads 0 Views 131KB Size
Download PDF
Anal. Chem. 2000, 72, 3826-3832

Determination of Glyphosate and Aminomethylphosphonic Acid Residues in Water by Gas Chromatography with Tandem Mass Spectrometry after Exchange Ion Resin Purification and Derivatization. Application on Vegetable Matrixes A. Royer,† S. Beguin,† J. C. Tabet,‡ S. Hulot,§ M. A. Reding,| and P. Y. Communal*,†

Groupement Interre´ gional de Recherche sur les Produits Agropharmaceutiques, Angers Technopole, 8, rue H. Becquerel, 49070 Beaucouze´ , Laboratoire de Chimie Structurale Organique et Biologique, Universite´ Pierre et Marie Curie, 4, place Jussieu, 75252 Paris Cedex 05, Institut De´ partemental d'Analyse et de Conseil, La Chantrerie, Route de Gachet, BP 80603, 44306 Nantes Cedex 3, and Monsanto Europe, Parc Scientifique Fleming, 5, rue Laid Burniat, B-1348, Louvain-La-Neuve, Belgique

An analytical method for the determination of glyphosate and its principal metabolite, aminomethylphosphonic acid (AMPA), in water of different hardnesses (5, 20, and 30 °DH, french hardness) has been developed. Samples were fortified at different levels (0.05, 0.1, 1, and 5 µg/L) and were purified by column chromatography on ion-exchange resins. After derivatization with TFAA/HFB mixture, the derivatives were quantified by using capillary gas chromatography with an ion-trap tandem mass spectrometric detector. Analytical conditions for MS/MS detection were optimized, and the quantification was carried out on the sum of areas of the three most representative ions: m/z 283, 223, and 181 for AMPA and m/z 440, 321, and 261 for glyphosate. The limit of quantification was demonstrated to be at 0.05 µg/L for each compound. The mean recovery value and the relative standard deviation (n ) 65) were 93 and 12% for AMPA and 95 and 13% for glyphosate. Glyphosate [N-(phosphonomethyl)glycine], the active ingredient of commercial herbicides, for instance, Roundup, is a verybroad-spectrum, nonselective, post-emergence weedkiller and is widely used in various applications for weed and vegetation control. Glyphosate and its main metabolite, aminomethylphosphonic acid (AMPA), are very polar (log Kow) -1.9 for glyphosate) and present a high solubility in water. These chemical properties make the use of classical organic solvent extractions practically impossible and require cleanup procedures with anion- and cationexchange columns.1 Moreover, the detection of glyphosate and * Corresponding author. Tel.: 33 241 48 75 70. Fax: 33 241 48 71 40. E-mail: [email protected] † Angers Technopole. ‡ Universite ´ Pierre et Marie Curie. § Institut De ´ partemental d'Analyse et de Conseil. | Monsato Europe.

3826 Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

AMPA necessitates a derivatization step. In recent published methods, the determination was made by liquid chromatography with precolumn derivatization using 9-fluorenylmethylchloroformate (FMOC-Cl) or postcolumn derivatization using o-phthalaldehyde (OPA) and fluorescence detection.2-5 Because of a lack of specificity of the procedure and a difficult interpretation of chromatograms, it is very difficult to analyze these molecules in water at the 0.1 µg/L level,2,6,7 the maximum allowable concentration in drinking water set by the European Community.8 (1) Cowell, J. E.; Kunstman, J. L.; Nord, P. J.; Steinmetz, J. R.; Wilson, G. R. Validation of an analytical residue method for analysis of glyphosate and metabolite: an interlaboratory study. J. Agric. Food Chem. 1986, 34, 955960. (2) Sancho, J. V.; Hernandez, F.; Lopez, F. J.; Hogendoorn, E. A.; Dijkman, E.; van Zoonen, P. Rapid determination of glufosinate, glyphosate, and aminomethylphosphonic acid in environmental water samples using precolumn fluorogenic labeling and coupled-column liquid chromatography. J. Chromatogr., A 1996, 737, 75-83. (3) Hogendoorn, E. A.; Ossendrijver, F. M.; Dijkman, E.; Baumann, R. A. Rapid determination of glyphosate in cereal samples by means of precolumn derivatisation with 9-fluorenylmethyl chloroformate and coupled-column liquid chromatography with fluorescence detection. J. Chromatogr., A 1999, 833, 67-73. (4) Mallat, E.; Barcelo, D. Analysis and degradation study of glyphosate and of aminomethylphosphonic acid in natural waters by means of polymeric and ion-exchange solid-phase extraction columns followed by ion chromatography-post-column derivatization with fluorescence detection. J. Chromatogr., A 1998, 823, 129-136. (5) Lovdahl, M. J.; Pietrzyk, D. J. Anion-exchange separation and determination of bisphosphonates and related analytes by post-column indirect fluorescence detection. J. Chromatogr., A 1999, 850, 143-152. (6) Kataoka, H.; Ryu, S.; Sakiyama, N.; Makita, M. Simple and rapid determination of the herbicides glyphosate and glufosinate in river water, soil and carrot samples by gas chromatography with flame photometric detection. J. Chromatogr., A 1996, 726, 253-258. (7) Alferness, P. L.; Iwata, Y. Determination of glyphosate and (aminomethyl)phosphonic acid in soil, plant and animal matrices, and water by capillary gas chromatography with mass-selective detection. J. Agric. Food Chem. 1994, 42, 2751-2759. (8) Council Directive of 15 July 1980 relating to the quality of water intended for human consumption (80/778/EEC). Official Journal of the European Communities, No L229/11-29. 10.1021/ac000041d CCC: $19.00

© 2000 American Chemical Society Published on Web 07/11/2000

Recently, more-sensitive and more-selective techniques were developed: LC-ESI-MS,9 without derivatization, and LC-ESI-MSMS,10 with a FMOC-Cl derivatization step. Results obtained by these authors showed the same difficulties of interpretation without derivatization, but good results were obtained with MSMS detection. The aim of this work was to develop an analytical method for the quantification of AMPA and glyphosate in water of different hardnesses (slightly, moderately, and highly mineralized) by GC/ EI/MS/MS after extraction and purification on ion-exchange resins and derivatization. Many different derivatization methods, specific to amine or carboxylic groups, were developed. Derivatization has involved the use of diazomethane and isopropylchloroformate,6 heptafluorobutyric anhydride and 2-chloroethanol,11 N-methyl-N-(tert-butyldimethylsilyl)-trifluoroacetamide,12 trifluoroacetic anhydride and trifluoroethanol13 or trifluoroacetic anhydride (TFAA) and 2,2,3,3,4,4,4 heptafluoro-1-butanol (HFB).7,14 Among these derivatization procedures, the experiments were performed with the method studied by Alferness,7 who has determined glyphosate and AMPA residues in different matrixes. EXPERIMENTAL PROCEDURES Instrumentation. The analyses were carried out on a Varian model 3800 gas chromatograph (Walnut Creek, CA) equipped with a split/splitless injector and a 8200 autosampler. The detection was conducted with a Varian Saturn 2000 ion-trap mass spectrometer equipped with a waveboard and MS/MS software. A HewlettPackard model 6890 gas chromatograph (Palo Alto, CA) equipped with a split/splitless injector and a Hewlett-Packard 7683 autosampler was used to confirm mass spectra of compounds. It was coupled to a Hewlett-Packard 5973 MS-Engine quadrupole mass spectrometer. The separation of the analytes was achieved with a 30 m × 0.25 mm i.d. VA-5MS (Varian Instruments) fused-silica capillary column with 0.25-µm film thickness of bonded 5% phenyl/95% dimethylpolysiloxane. Gas Chromatographic Analysis. The optimized column oven temperature program was as follows: from 80 °C (hold 1.5 min.) to 260 °C at 30 °C/min (hold 1 min.) and from 260 °C to 300 °C at 30 °C/min. The total run time was 9.83 min. Two microliters (9) Bauer, K.-H.; Knepper, T. P.; Maes, A.; Schatz, V.; Voihsel, M. Analysis of polar organic micropollutants in water with ion chromatography-electrospray mass spectrometry. J. Chromatogr., A 1999, 837, 117-128. (10) Vreeken, R. J.; Speksnijder, P.; Bobeldijk-Pastorova, I.; Noij, Th. H. M. Selective analysis of the herbicides glyphosate and aminomethylphosphonic acid in water by on-line solid-phase extraction-high-performance liquid chromatography-electrospray ionization mass spectrometry. J. Chromatogr., A 1998, 794, 187-199. (11) Guivivan, R. A.; Thompson, N. P.; Wheeler, W. B. Derivatization and cleanup improvements in determination of residues of glyphosate and aminomethylphosphonic acid in blueberries. J.sAssoc. Off. Anal. Chem. 1982, 65(1), 35-39. (12) Moye, H. A.; Deyrup, C. L. A single-step derivatization method for the gas chromatographic analysis of the herbicide glyphosate and its metabolite. J. Agric. Food Chem. 1984, 32, 192-195. (13) Roy, D. N.; Konar, S. K. Development of an analytical method for the determination of glyphosate and (aminomethyl)-phosphonic acid residues in soil by nitrogen-selective gas chromatography. J. Agric. Food Chem. 1989, 37, 441-443. (14) Deyrup, C. L.; Chang, S.; Weintraub, R. A.; Moye, H. A. Simultaneous esterification and acylation of pesticides for analysis by gas chromatography. 1. Dervatization of glyphosate and (aminomethyl)phosphonic with fluorinated alcohols-perfluorinated anhydrides. J. Agric. Food Chem. 1985, 33, 944947.

Table 1. MS Analysis Conditions filament emission current AGC target AGC prescan ionization time mass range scan rate multiplier delay peak threshold background mass

10 µA software optimized value (=20 000 counts) 100 µs 50-650 u 1 scan/s 3.8 min 700 49

Table 2. MS/MS Analysis Optimization Conditions filament emission current AGC target AGC prescan ionization time mass range scan rate multiplier delay peak threshold background mass excitation mode isolation window excitation time

80 µA 5000 counts 1500 µs (m/z)c.o +1 - (m/z)p.i +2a 1 scan/s 3.8 min 700 (m/z)c.o excitation storage level nonresonant 3u 0 ms (isolation) - 20 ms (MS/MS optimization)

a (m/z) c.o is the m/z excitation storage level of each analyte placed at a stability parameter of qz)0.3. (m/z)p.i is the m/z of the target parent ion of each analyte

of samples were injected in the splitless mode at the operating temperature of 200 °C. Helium (purity 5.5) was used as the carrier gas, maintained at a constant flow-rate of 1 mL/min. The approximate retention times of AMPA and glyphosate derivates were 4.0 and 4.7 min. MS/MS Optimization Method. The optimization of the analytical conditions for the MS/MS detection of the two analytes was performed with a 3-step method, according to the conditions listed in tables 1 and 2.15,16 This method consisted of (1) determination of the retention time and the main diagnostic ion (parent or fragment P+°) of each analyte by GC/EI/MS, (2) isolation of the selected ions by GC/EI/MS/MS with a 3u mass isolation window, and (3) determination of the optimum CID excitation voltage of each target ion in the nonresonant excitation mode. The parent ions were placed at a stability parameter qz ) 0.3 in the Paul stability diagram, and the dissociation conditions were selected for each compound using the automated methods development editor (AMD). GC/EI/MS and GC/EI/MS/MS experiments were realized in the full-scan mode over a mass range of 50-650 u (Table 1) and with a multisegment acquisition method, one segment per compound, and a narrow mass range ((m/z)c.o+1 - (m/z)p.i + 2) to improve the sensitivity (Table 2). The experiments were performed in the electron ionization (EI) mode, under Automatic Gain Control (AGC) to control the ionization time in order to maximize the signal while holding the total-ion space charge level constant as the sample and matrix levels change. For full-scan EI analysis, a software-optimized target TIC value of approximatively 20 000 is used, but a much lower target (5000 counts) was required (15) Beguin, S.; Tabet, J. C.; Rosati, D.; Communal, P. Y. Analytical protocol for building product ions library for pesticides using tandem GC-ITMS. (16) Beguin-Georget, S.; Communal, P. Y.; Rosati, D.; Tabet, J. C.; 44th ASMS Conference Mass Spectrometry, Allied Topics, Dallas, Texas, July, 1999.

Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

3827

for the MS/MS experiments to minimize the “space charge” effects. To improve the detection limits during the MS/MS analysis, the filament emission current was set at 80 µA (with the electron energy of 70 eV), the multiplier voltage was raised from its software-optimized value according to a 300 V offset with a multiplier gain of 1 × 105, and the AGC prescan ionization time was enhanced to 1500 µs. The axial modulation voltage has been optimized at 3.5 V in order to easily eject the ions and to significantly improve mass resolution. The temperatures of the transfer line, the manifold, and the trap were set at 260, 60, and 220 °C, respectively. Reagents and Glassware. Standard samples of glyphosate of 99.9% purity and aminomethylphosphonic acid of 99.1% purity were obtained from Monsanto, F-69673 Bron Cedex. Methanol for HPLC, ethyl acetate for organic trace analysis, chlorhydric acid 37%, and iron chloride for analysis (FeCl3) were purchased from Merck, F-94736 Nogent sur Marne. Octanol was bought from Prolabo, F-45250 Briare Le Canal, and dimethyldichlorosilane, trifluoroacetic anhydre 99%, 2,2,3,3,4,4,4 heptafluoro-1-butanol 98%, and citral 95% were obtained from Sigma-Aldrich, F-38297 Saint Quentin Fallavier. Chelex-100 resin, 100-200 mesh, AG1X8 anionexchange resin, 200-400 mesh, were purchased from Biorad, F-94203 Ivry sur Seine and liquid nitrogen was obtained from Air Liquide, F-44325 Nantes. Calibration Solutions. Glyphosate and AMPA stock calibration solutions of 1000 µg/mL were separately prepared by dissolving a known amount of each analyte in water. Ten milliliters of each stock solution was taken with precision and diluted to a volummetric flask (100-mL) using ultrapure water and 3 drops of HCl, 6 N, to obtain a stock calibration solution containing both analytes at 100 µg/mL. Working calibration solutions of 10, 1, and 0.5 µg/mL were prepared by dilution of the stock solution (100 µg/mL) with ultrapure water and 3 drops of HCl, 6 N, in 50-mL volumetric flasks. Preparation of Resins. Chelex-100 Resin, 100-200 Mesh, Na Form. The resin was converted to the iron form by slurrying 0.9 kg of resin in approximately 3 L of ultrapure water. After adding about 50 mL of HCl, 6 N, and about 1 L of FeCl3 solution, 0.1 N, the resin was mixed for at least 10 min with a magnetic stirrer. After decantation, the resin was rinsed twice with a mixture of appproximately 2 L of ultrapure water and about 500 mL of FeCl3 solution, 0.1 N. The resin was transferred in a wide chromatographic tube (Φ 8 cm, 40 cm long) and was given a final rinse with about 4 L of HCl, 0.02 N. AG1X8 Anion-Exchange Resin, 200-400 Mesh, Cl Form. One bottle of about 450 g of AG1X8 resin was slurried in about 1 L of ultrapure water and mixed for at least 30 min with a magnetic stirrer. After decantation, the supernatant was discarded. This washing process was repeated two more times. Fortification. 500 ( 0.3 g of water were accurately weighed in a 500-mL flask. The sample was fortified with the appropriate working calibration solution to produce fortification levels of 0.055.0 µg/L. Water was degassed about 15 min using an ultrasonic bath, and the water was brought to pH ) 2 ( 0.05 with about 1 mL of HCl, 6 N. Purification. Cleanup on Chelex-100 Resin. Chelex-100 resin (Cl form) was prepared the day before cleanup. A sample 3828

Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

of 22 ( 0.1 mL of iron load Chelex-100 resin was transferred into a chromatographic column (Φ 2.2 cm, 13 cm long topped with a reservoir (Φ 5.5 and 9 cm long)) containing a glass wool plug. A second glass wool plug was placed on the resin layer after the resin had settled, and 8 mL of ultrapure water was added. The sample was applied on the resin and was eluted at 6-8 mL/min. The column was rinsed with 50 mL of ultrapure water and 100 mL of HCl, 0.2 N,at high flow rate. All preceding fractions were discarded. Three milliliters, then 4 mL of HCl 6 N were applied to the column and were eluted at 4 mL/min. Both fractions were rejected. Glyphosate and AMPA were eluted with 2 × 7 mL and 2 × 9 mL of HCl, 6 N, at a flow rate of about 4 mL/min. The fractions were collected in a 100-mL round-bottom flask, and 10 mL of HCl, 10 N, was added to the eluate. Cleanup over Anion-Exchange Resin AG1X8. AG1X8 resin was prepared the day before cleanup. A sample of 5.5 ( 0.1 g of AG1X8 resin was transferred into a chromatographic column (Φ 1.5 cm, 14 cm long topped with a reservoir (Φ 2.5 and 6.5 cm long)) containing a glass wool plug. A second glass wool plug was placed on the resin layer after the resin had settled, and 8 mL of ultrapure water was added. The column is prerinsed with 3 × 5 mL of HCl, 6 N. The sample extract from Chelex-100 resin was applied to the column with the stopcock wide open. The sample was followed by 1 × 2 mL and 2 × 4 mL of HCl, 6 N. The total eluate was collected in a 100-mL round-bottom flask before being silanized and 3 drops of octanol were added. The flask was evaporated to dryness (50 °C). The residue was dissolved in 0.5 mL of CAX mixture (160 mL of methanol, 40 mL of ultrapure water, and 2.7 mL of concentrated HCl). Derivatization. The derivatizing reagent mixture was freshly prepared each day in a suitably sized glass container with a PTFElined cap by adding 1 volume of 2,2,3,3,4,4,4-heptafluoro-1-butanol and 2 volumes of trifluoroacetic anhydride. The container was capped and shaken gently. Aliquots of derivatizing reagent (1.5 mL) were placed into screw-stopped autosampler vials (2 mL). The vials were capped using phenolic plastic and open-top caps with Teflon-silicone-Teflon septa. Because the derivatization reaction was extremely exothermic, the vials were cooled to a temperature of less than -60 °C in liquid nitrogen. A sample aliquot of 50 µL was added to the derivatizing reagent mixture. The pipet tip was placed under the surface of the reagent, and the contents were slowly released. The pipet tip was immediately rinsed by repeatedly sucking the reagent and sample mixture into the disposable pipet and releasing it back into the vials, always keeping the pipet tip under the surface of the reagent. After the sample aliquot was added to the reagent mixture, the vials were capped, manually shaken, and allowed to equilibrate to room temperature. Analyte derivatization was performed by transferring the reaction vials to an aluminum block with 13-mm holes. The vials were heated 1 h at 92-97 °C. After allowing the vials to cool to room temperature, the excess derivatization reagent was evaporated under a stream of nitrogen. Once apparent dryness has been achieved, the samples should remain under the stream of nitrogen for an additional 20-30 min. Residual derivatization reagents or trifluoroacetic acid can give interferences in the chromatographic run. The residuum was dissolved in 200 µL of ethyl acetate containing 2.0 µL citral per milliliter. The vials were capped and shaken by vortexing to dissolve the contents. All

Figure 1. Electronic impact mass spectra and strucures of (A) glyphosate derivative (MW 811) and (B) AMPA derivative (MW 572).

derivatizations should be performed in a well-ventilated hood, and the analyst should wear protective gloves. RESULTS AND DISCUSSION Full-Scan and MS/MS Optimization of AMPA and Glyphosate Derivatives. The full-scan EI mass spectrum of a standard mixture of the two analytes at 1 mg/L allowed determination of their retention time and verification of their identity by studying the molecular and the fragment ions with the use of their molecular weight. The retention time of AMPA and the glyphosate derivatives were 4.0 and 4.7 min, respectively, and the mass spectra (Figure 1) contained specific peaks, already observed in

the literature.7 First, the isolation step was performed according to the conditions listed in Table 2, with the highest mass-to-charge ratio ion with the highest intensity selected as the parent ion, to remove the most matrix chemical background interference and to obtain the most intense signals. The aim of this step was to ensure that the target ion was correctly isolated in the ion trap without interference in the isolation window so that the product ions formed could be detected. After investigating the m/z 572 ion in the case of AMPA, the m/z 446 ion was found to be more suitable for the MS/MS quantitative analysis because the m/z 572 ion did Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

3829

Figure 2. Mechanism of fragmentation of parent ion for glyphosate derivative. Table 3. MS/MS Conditions of Each Analyte compounds

parent ion (m/z)

excitation storage level (m/z)

nonresonant excitation voltage (V)

product ions for quantification (m/z)

AMPA derivative glyphosate derivative

446 612

147.6 202.6

82 95

283+223+181 440+321+261

not give MS/MS information during the CID process. For glyphosate derivative, the parent ion chosen was the m/z 612 ion, coming from the fragmentation of the M+° molecular ion (MW ) 811), not observed in the full-scan EI mass spectrum due to mass-range of the ion-trap apparatus. The optimization of the MS/MS conditions of the two analytes was stepped using the AMD method (Table 3). After examining the evolution of the product ions’ intensity as a function of the nonresonant CID excitation voltage, the excitation voltage which produced the most intense 2 or 3 product ions and achieved maximum sensitivity was chosen for the quantitative aspect of this study. For the AMPA and glyphosate, the highest intensity of products ions occurred at an excitation voltage of 82 and 95 V. A fragmentation mechanism for the molecular ion of glyphosate, for determining the structure of the quantification ions, was proposed in Figure 2. Unfortunately, for AMPA, the fragmentation mechanism could not be explained. Recovery of AMPA and Glyphosate. The quantification was realized on the sum of the three most representative ions for each 3830

Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

derivative: the m/z 283, 223, and 181 ions for AMPA and the m/z 440, 321, and 261 ions for glyphosate (Figure 3). Fortified water samples were quantified using the external standard method. Calibration curves of each compound were linear in the range of 0.0025-0.200 µg/mL and the squared correlation coefficients r2 were 0.997 (RSD ) 0.3%) and 0.998 (RSD ) 0.2%) for glyphosate and AMPA derivatives, respectively (n ) 14). Two different mineralized waters (5 and 30 °DH) and water coming from Angers (hardness about 20 °DH) were analyzed. These waters were spiked with a mixture of glyphosate and AMPA at four different levels (5, 1, 0.1, and 0.05 µg/L). As shown in Table 4, the recoveries of AMPA and glyphosate added in water were 87-97% for AMPA, with relative standard deviation 9-16%, and were 90-100% for glyphosate, with relative standard deviation 11-15%. The overall recoveries (n ) 65) were 93% (RSD ) 12%) and 95% (RSD ) 13%) for AMPA and glyphosate, respectively. The quantification limit was 0.05 µg/L and the detection limit was 0.025 µg/L. For each type of water, untreated samples were extracted and no interferences were observed. After the develop-

Figure 3. Chromatograms and MS/MS spectra of AMPA and glyphosate derivatives of a water sample fortified at 0.05 µg/L.

ment of this method, an optimization was achieved in order to reduce the volume of the water sample. One hundred milliliters of water with 30 °DH hardness was extracted and purified on the ion-exchange resins, and the procedure was not changed. However, the volume of an aliquot of sample for the derivatization reaction was 100 µL instead of 50 µL and the volume injected was 4 µL. Five experiments were realized, and the average recoveries were 87.6% (RSD ) 15.6%) for glyphosate on water spiked at 0.05 µg/L and 74.0% (RSD ) 6.6%) for AMPA, but the limit of quantification was 0.1 µg/L. This optimization has allowed a

shortening of the analysis time, and eight samples can be processed per day. Additionally, to broaden the scope of this procedure, this method was tried on blackcurrant and hazelnut samples. Each sample was only fortified with glyphosate at 0.020 mg/kg. Homogenized samples (15 g) were mixed with 75 mL of HCl, 0.01 N, and 25 mL of dichloromethane and extracted by Ultraturrax for about 1 min. The extracts were centrifuged for 20 min at about 2000 rpm. The supernatant liquid (aqueous layer) was filtered on glass wool plug, and ultrapure water was added to bring the total volume to 400 mL. The pH of mixed sample Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

3831

Table 4. Recoveries of Glyphosate and AMPA in Water spiking level (µg/L)

no. of repetitions

AMPA avg recovery (%)

RSD (%)

5 1 0.1 0.05 overall

15 15 21 14 65

93.5 92.5 96.6 86.8 92.8

9.2 15.8 11.1 11.1 12.3

glyphosate avg recovery RSD (%) (%) 92.8 90.1 96.7 99.5 94.9

10.9 10.7 14.8 12.3 12.9

was adjusted to 2 ( 0.05. Before the derivatization step, the cleanup was performed using Chelex-100 resin (15 ( 0.1 mL) and AG1X8 resin (11 g(0.1 g). The average recovery was 81% (RSD ) 12%) for blackcurrant (n ) 3) and, for hazelnuts, the average recovery was 92.2% (n ) 5) and relative standard deviation was 12.6% (Table 5). CONCLUSION This study illustrates a new approach to the determination of glyphosate and AMPA in water. The cleanup on ion-exchange resins has permitted a high purification of the residues. Analysis by GC/EI/MS/MS after derivatization of compounds has demonstrated a very good selectivity and a high sensitivity. We have demonstrated the utility of MS/MS in order to obtain specific mass spectra, excluding false positives samples that could be observed

3832 Analytical Chemistry, Vol. 72, No. 16, August 15, 2000

Table 5. Recoveries of Glyphosate in Blackcurrant and Hazelnut

blackcurrant hazelnut

spiking level (mg/kg)

no. of repetitions

avg recovery (%)

RSD (%)

20 20

3 5

81 92

12 13

by other methods. The method was validated over the range 0.05-5 µg/L, responding to the European Community requirement for drinking water. The limit of detection was 0.025 µg/L, and the limit of determination was 0.05 µg/L (Figure 3). These results were confirmed with the quadrupole mass spectrometer. The measurement was carried out in the selected ion monitoring (SIM) mode using the three most specific ions for each derivative (m/z 502, 446, and 372 for the AMPA derivative and m/z 612, 584, and 486 for the glyphosate derivative). Moreover, this method has been applied to waters of different hardness (over the range 5-30 °DH). The application on vegetable matrixes for glyphosate determination has also given good results at a fortification level of 0.020 mg/kg.

Received for review January 10, 2000. Accepted May 18, 2000. AC000041D