556 YOUNG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001 RESIDUES AND TRACE ELEMENTS Mixed-Mode Solid-Phase Extraction and Cleanup Procedures for the Liquid Chromatographic Determination of Thiabendazole and Carbendazim in Fruit Juices MICHAEL S. YOUNG, DOROTHY J. PHILLIPS, PAMELA C. IRANETA, and JIM KROL Waters Corp., 34 Maple St, Milford, MA 01757 Solid-phase extraction (SPE) procedures were developed for rapid cleanup and determination of thiabendazole and carbendazim in orange, apple, and grape juices. Samples were prepared by using an SPE cartridge containing a mixed-mode sorbent with both reversed-phase and strong cation-exchange chemistries. Analysis was by liquid chromatography with photodiode-array UV detection. Orange juice was analyzed by mixed-mode cation-exchange extraction with reversed-phase cleanup; the other juices were analyzed by reversed-phase extraction with cation-exchange cleanup. Recoveries >80% for carbendazim and >90% for thiabendazole. Quantitation limits were 20 g/l for both analytes. Thiabendazole (TBZ) and benomyl are benzamidazolic compounds commonly used worldwide as fungicides for protection of numerous fruit and vegetable crops. Benomyl residue is not stable in fruit matrix and, therefore, benomyl is usually determined as a degradation product, methyl-2-benzimidazole carbamate (carbendazim, MBC). Structures of these compounds are shown in Figure 1. Many sensitive methods have been published for the determination of these fungicides in foods and fruit juices. A standard U.S. Food and Drug Administration method has been used for the determination of carbendazim and thiabendazole (1, 2). This method involves sequential liquid liquid partition steps for extraction and cleanup, followed by liquid chromatography (LC) with UV and fluorescence detection. Thiabendazole has been determined in fruit juices by gas chromatography with nitrogen phosphorus selective detection (GC NPD) after isolation of the residue by liquid liquid extraction (3). A significant advance in the determination of benzamidizole fungicides has been the incorporation of solid-phase extraction (SPE) procedures both for isolation of the residues and for subsequent cleanup of sample extracts. Thiabendazole has been determined in milk by reversed-phase Received March 6, 2000. Accepted by JS June 7, 2000. Presented in part at the 113th AOAC INTERNATIONAL Annual Meeting and Exposition, Houston, TX. SPE on octadecylsilica (4). Extracts from tissue samples have been cleaned up by using cyanopropylsilica SPE (5), and fruit extracts have been cleaned up by using diol-silica SPE (6). However, the most common and effective SPE procedures for extraction and cleanup for fungicide determination have used strong cation-exchange (SCX) resins, often in conjuction with other extraction techniques. Thus, improved procedures for LC determination of carbendazim and thiabendazole use SCX cartridges for cleanup of liquid extracts of fruit (2, 7 9) and milk (10). The intentional use of 2 (or more) different functional groups on one sorbent defines the resulting separation as a mixed-mode separation. A review of this technique and its applications has been presented (11). Recently, a new mixed-mode polymeric sorbent, Oasis MCX resin, was introduced for SPE. This new sorbent cartridge incorporates both reversed-phase and SCX chemistries and has been shown to be highly effective for the extraction of basic drugs and related compounds from biological fluids (12). In this paper, we report the application of this new SPE sorbent to the LC determination of thiabendazole and carbendazim in apple, orange, and grape juices. Although the mixed-mode approaches described in this paper are not conceptually different from sequential extraction/cleanup approaches previously documented, the overall process is improved. The new mixed-mode method allows the analyst to conduct the sequential steps, using only one cartridge. The SPE strategies were developed to ensure that the fungicides are retained on the cartridge while succesive wash steps remove interferences. In these wash steps solvent strength and ph are varied while the analytes are cycled between cationic and neutral forms. Because the sorbent is a mixed-mode material, the fungicide analytes are effectively retained in either the cationic or the neutral form. The LC analysis documented in this paper was accomplished with photodiode array UV detection (PDA). The recovery results were determined by using a single wavelength, 288 nm, for quantitation of each analyte. Incurred residues were identified in real samples by using PDA; UV spectra were obtained from 195 to 350 nm and were compared with spectra of standard compounds.
YOUNG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001 557 Figure 1. Structures of carbendaim and thiabendazole. Experimental Materials (a) Syringe filters. Millex-LCR13, 0.5 µm, 13 mm (Millipore Corp., Bedford, MA). (b) Centrifuge tubes. Polypropylene, 29 104 mm (Beckman Instruments, Inc., Palo Alto, CA). (c) SPE cartridges. Oasis MCX cartridges; mixed-mode (reversed-phase and SCX) sorbent, 6 ml, 150 mg, 30 µm particle size (Waters Corp., Milford, MA; p/n 186000256). Reagents (a) Thiabendazole and carbendazim. 100 µg/ml in methanol (AccuStandard, Inc., New Haven, CT). (b) Solvents. Acetonitrile and methanol, LC grade (J.T. Baker, Phillipsburg, NJ). (c) High-purity water. Obtained in-house with a Milli-Q system (Millipore Corp.). (d) Sodium phosphate monobasic monohydrate. ACS grade (J.T. Baker). (e) Sodium phosphate dibasic. ACS grade (J.T. Baker). (f) Ammonium hydroxide (NH 4 OH), 30% (ca 7.5M). Instra-analyzed grade (J.T. Baker). (g) Phosphate buffer, 20mM, ph 6.8. Dissolve 1.38 g sodium phosphate monobasic monohydrate and 1.41 g sodium phosphate dibasic in ca 900 ml water, and adjust to ph 6.8 with dilute NaOH or H 3 PO 4 as necessary. Then adjust volume to exactly 1.0 L. (h) NH 4 OH, 150mM. Dilute 2 ml 7.5M NH 4 OH to 100 ml with water. (i) NH 4 OH, 150mM, in methanol water (30 + 70, v/v). Mix 30 ml methanol and 2 ml 7.5M NH 4 OH with water to make 100 ml solution. (j) NH 4 OH, 300mM. Dilute 4 ml 7.5M NH 4 OH to 100 ml with water. (k) NH 4 OH, 300mM, in methanol. Dilute 4 ml 7.5M NH 4 OH to 100 ml with methanol. (l) Mobile phase. Phosphate buffer, ph 6.8 acetonitrile (72.5 + 27.5); used for LC analysis of all samples. Figure 2. Oasis MCX cartridge SPE procedures for determination of fungicides in fruit juices.
558 YOUNG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001 Table 1. Recovery of carbendazim from various fruit juices Mean recovery, % (RSD, %) a Sample Spiking level, µg/l No. of replicates SPE method Carbendazim Thiabendazole Orange juice 1 20 4 A 88.3 (1.3) 95.2 (1.9) Orange juice 1 50 4 A 83.8 (7.7) 92.8 (3.7) Orange juice 1 200 4 A 91.2 (1.9) 91.3 (1.5) Apple juice 1 20 4 B 81.8 (1.7) 95.7 (1.6) Apple juice 1 200 4 B 80.6 (1.6) 95.8 (3.7) Grape juice 1 20 5 B 90.7 (11) 97.7 (11) Grape juice 1 50 5 B 85.6 (1.0) 90.9 (2.6) Grape juice 1 200 4 B 83.7 (3.1) 96.1 (2.3) a RSD = relative standard deviation. Juice Samples All juice samples were commercial, bottled, whole juices except for orange juice 2, which was squeezed in the laboratory from Florida juice oranges. The commercial orange juices were pasteurized Florida juices obtained at local markets. Apple juice 1 was pasteurized whole juice from Mexico marketed as organic. Apple juice 2 was fresh-squeezed cider obtained from a local apple orchard. The other apple juices were pasteurized ciders obtained at a local supermarket. Orange juice 1, apple juice 1, and grape juice 1 were used for recovery experiments. Spiking Solutions Spiking solutions of 20 and 50 µg/ml were prepared by dilution of the commercially prepared methanolic standards. The 20 µg/ml spiking solution was used for preparation of samples fortified at 20 µg/l, and the 50 µg/ml spiking solution was used for preparation of samples fortified at higher levels. Instrumentation The liquid chromatograph was a Waters Alliance System comprising of a Model 2690 Separations Module and a Model 996 photodiode array detector. Results were calculated from chromatograms obtained at 288 nm. The Waters Millenium Chromatography Manager (version 3.15.5) was used for system control, data acquisition, and data analysis. A Waters XTerra RP 18 column (3.5 µm particle size, 100 4.6 mm id) was used for analysis of all samples. A Waters Sentry guard column (20 3.9 mm), packed with 3.5 µm XTerra RP 18 sorbent, was used to protect the analytical column. Injection volume was 25 µl. Flow rate was 1.0 ml/min. were combined with the supernatant. The prepared juice samples were then processed according to the scheme shown in Figure 2 (Method A). A 12-position vacuum manifold (Waters Corp.) was used for all SPE. Flow rates for each step were maintained at ca 4 ml/min. The acidification of the sample ensures that the fungicides are in the cationic form. The fungicides are therefore retained on the Oasis MCX sorbent by both reversed-phase and cation-exchange mechanisms (mixed-mode interaction). Wash 1, with 0.1N HCl, ensures that that the fungicides remain in the cationic form. At this point, the cartridge is washed with methanol (wash 2) to remove any neutral or acidic compounds that are retained by the reversed-phase mechanism. After the analytes are converted back to the nonionic form, wash 3, 150mM NH 4 OH in methanol water (30 + 70), is used to remove interferences that are retained by only the cation-exchange mechanism or interferences that are retained by weak mixed-mode interaction. The analytes are finally eluted from the SPE cartridge with 300mM NH 4 OH in SPE of Orange Juice Samples The juice samples (10 ml) were diluted 1:1 with reagent water, acidified to ph 2 with 1M HCl, and centrifuged for 10 min at 8000 g. The supernatant was removed, the pellet was carefully rinsed with 2 ml reagent water, and the rinses Figure 3. Liquid chromatogram showing recovery of fungicides from orange juice (spiking level, 20 g/l).
YOUNG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001 559 Figure 4. Liquid chromatograms showing recovery of fungicides from apple juice (spiking level, 20 g/l). The upper chromatogram shows the result when SPE method B was used, and the lower chromatogram shows the result when SPE method A was used. methanol. The eluate is evaporated to dryness at 45 o C with a gentle nitrogen stream, and the residue is dissolved in 1.0 ml mobile phase. The final extract is filtered through the 0.5 µm membrane filter, and the filtrate is analyzed by LC PDA. SPE of Apple and Grape Juice Samples The juice samples (10 ml) were diluted 1:1 with reagent water, adjusted to ph 10 with 2M NaOH, and centrifuged for 10 min at 8000 g. The supernatant was removed, the pellet was carefully rinsed with 2 ml reagent water, and the rinses were combined with the supernatant. The prepared juice samples were then processed according to the scheme shown in Figure 2 (Method B). Before the sample is loaded on the SPE cartridge, the fungicides are converted to their free-base form by treatment with aqueous base. The fungicides are therefore retained on the Oasis MCX sorbent by the reversed-phase mechanism only. Table 2. Determination of fungicide residues in various juices Sample SPE method Carbendazim, µg/l Thiabendazole, µg/l Orange juice 1 A ND a ND Orange juice 2 A ND ND Orange juice 3 A ND ND Apple juice 1 B ND ND Apple juice 2 B 2 b 1 b Apple juice 3 B 68 130 Apple juice 4 B 52 120 Apple juice 5 B 19 93 Apple juice 6 B 7 b 27 Apple juice 7 B 4 b 22 Grape juice 1 B ND ND Grape juice 2 B ND ND a b ND = not detected. Detected at level below the limit of quantitation (20 µg/l).
560 YOUNG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001 Figure 5. Liquid chromatogram from the LC PDA analysis of a commercial apple juice (apple juice 3 ); UV spectra showing the UV confirmation of the identity of carbendazim. Wash 1 (150mM NH 4 OH) ensures that that the fungicides remain in the neutral form. At this point, the cartridge is washed with 150mM NH 4 OH in methanol water (30 + 70; wash 2) to remove any neutral or acidic compounds that are weakly retained by the reversed-phase mechanism. After the analytes are converted back to the ionic form with 0.1N HCl, wash 3 (methanol) is used to remove interferences that are strongly retained by reversed-phase interaction. The analytes are finally eluted from the SPE cartridge with 300mM NH 4 OH in methanol. The eluate is evaporated to dryness at 45 o C with a gentle nitrogen stream, and the residue is dissolved in 1.0 ml mobile phase. The final extract is filtered through the 0.5 µm membrane filter, and the filtrate analyzed by LC PDA. Results and Discussion Recovery of Carbendazim and Thiabendazole from Various Juices Initial experiments were conducted with the intention of using a high-capacity reversed-phase cartridge with Oasis HLB as the primary extraction sorbent, followed by cleanup on Oasis MCX sorbent. Experiments with apple juice indicated that this approach was effective (recovery was >90%). Experiments with orange juice, however, indicated that the cation-exchange sorbent should be used for the primary extraction followed by cleanup on the reversed-phase sorbent. Because the Oasis MCX resin consists of cation-exchange sites chemically bonded to a reversed-phase matrix, it seemed reasonable that both the extraction of the fungicides and subsequent cleanup could be accomplished with only 1 cartridge. Refinements of the SPE procedures using a single Oasis MCX cartridge resulted in the protocols summarized in Figure 2; SPE method A (mixed-mode extraction, reversed-phase cleanup) for orange juice and SPE method B (reversed-phase extraction, cation-exchange cleanup) for the other juices. Results obtained by using the single-cartridge methods were comparable to results obtained by using separate ion-exhange and reversed-phase cartridges. Results for the recovery experiments are presented in Table 1. Figure 3 shows a chromatogram obtained from analysis of an orange juice sample, spiked with each fungicide at 20 µg/l. In analyses of orange juice, recoveries were more consistent and interferences were fewer with method A than with method B. Figure 4 shows chromatograms obtained for similar spiked samples of apple juice extracted by both SPE methods; the interference level is clearly decreased by using SPE method B. Similar behavior was observed for grape juice samples. Visually, for both apple and grape juices, the residual color of the final sample extracts was much lower for samples processed with SPE method B than for those processed with method A. Recoveries were consistently >80% for carbendazim and >90% for thiabendazole, with excellent reproducibility. Analysis of Nonspiked Juice Samples Results for the analysis of various juice samples are presented in Table 2. No detectable fungicide residues were found in any of the orange or grape juices analyzed. However, most of the apple juices had measurable levels of both carbendazim and thiabendazole. The results for some of the
YOUNG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001 561 commercial whole apple juices showed levels of thiabendazole in excess of 100 µg/l and levels of carbendazim in excess of 40 µg/l. Figure 5 shows the liquid chromatogram from the LC PDA analysis of apple juice 3, the UV spectrum of carbendazim obtained for this sample, and the reference UV spectrum of the standard carbendazim compound. A similar perfect match was observed for thiabendazole. Confirmation of fungicide residues by electrospray LC/mass spectrometry is currently under investigation by the authors. The entire study was accomplished by using only one analytical column for LC. The guard column was replaced only once, before the optimization of SPE methods A and B was completed. Using the optimized methods, we observed that well over 100 samples can be processed and analyzed without the need to change the guard column. Batches of 6 8 samples were processed by using simple vacuum manifold technology. The analyst can easily process a batch of samples in <2 h. Conclusions The most effective extraction and cleanup were obtained for orange juice by means of mixed-mode extraction at acidic ph, followed by sequential SCX and reversed-phase cleanups. Conversely, for apple and grape juices, the most effective approach was reversed-phase extraction at basic ph, followed by sequential reversed-phase and SCX cleanup steps. A possible explanation for these observations is that apple and grape juice contain significant amounts of polyphenolic (acidic) interferences which display strong reversed-phase retention under acidic conditions. These highly colored substances are not retained at basic ph. Orange juice may have fewer of these highly colored polyphenolic interferences. The extraction and cleanup procedures discussed in this paper were developed from literature sources indicating that both reversed-phase and cation-exchange processes are useful for the isolation of carbendazim and thiabendazole residues in fruit juices. The results of this study improve the established procedures by providing a protocol that incorporates both mixed-mode cation-exchange and reversed-phase extraction and cleanup through the use of only one cartridge. Recoveries were >80% for carbendazim and >90% for thiabendazole in all juices tested. The results were highly reproducible, with relative standard deviations of <5% for replicate samples in most cases. Based on results obtained from the spiking experiments, the quantitation limit for each analyte is 20 µg/l. However, the consistent recoveries and high reproducibility observed at this level suggest that lower quantitation limits are possible with little or no modification of the method. References (1) Pesticide Analytical Manual (1990) Vol. I, U.S. Food and Drug Administration, Rockville, MD, sec. 404 (2) Levine, R.A., Luchtefeld, R.G., Hopper, M.L., & Salmon, G.D. (1998) J. AOAC Int. 81, 1217 1223 (3) Oishi, M., Onishi, K., Kano, I., Nakazawa, H., & Tanabe, S. (1994) J. AOAC Int. 77, 1293 1296 (4) Long, A.R., Hsieh, L.C., Malbrough, M.S., Short, C.R., & Barker, S.A. (1989) J. Assoc. Off. Anal. Chem. 72, 739 741 (5) Cannavan, A.C., Haggan, A.S., & Kennedy, D.G. (1998) J. Chromatogr. B 718, 103 113 (6) Hiemstra, M., Joosten, J.A., & dekok, A. (1995) J. AOAC Int. 78, 1267 1274 (7) Ito, Y., Ikai, Y., Hayakawa, J., & Kagimi, T. (1998) J. Chromatogr. A 810, 81 87 (8) DiMuccio, A., Girolimetti, S., Attard Barbini, D., Pelosi, P., Generali, T., Vegori, L., De Merulis, G., Leonelli, A., & Stefanelli, P. (1999) J. Chromatogr. A 833, 61 65 (9) Arenas, R.V., Rahman, H., & Johnson, N.A. (1996) J. AOAC Int. 79, 579 582 (10) Arenas, R.V., & Johnson, N.A. (1995) J. AOAC Int. 78, 642 646 (11) Thurman, E.M., & Mills, M.S. (1998) Solid-Phase Extraction, Principles and Practice, John Wiley and Sons, New York, NY, pp 44 45 and pp 200 210 (12) Cheng Y., Neue U.D., & Woods L.L. (1999) J. Chromatogr. B 729, 19 31
562 JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 2, 2001