Organohalogen and Organophosphorous ... - ACS Publications


Organohalogen and Organophosphorous...

1 downloads 50 Views 2MB Size

Anal. Chem. 2009, 81, 5716–5723

Organohalogen and Organophosphorous Pesticide Method for Ginseng Root s A Comparison of Gas Chromatography-Single Quadrupole Mass Spectrometry with High Resolution Time-of-Flight Mass Spectrometry Douglas G. Hayward* and Jon W. Wong U.S. Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office of Regulatory Science, 5100 Paint Branch Parkway, HFS-706, College Park, Maryland 20740-5350 A method has been developed for the analysis of 170 organohalogen and organophosphorous pesticides, isomers, and metabolites in dried ground ginseng root. Pesticides were extracted with ethyl acetate and purified with gel permeation chromatography (GPC) and primary/ secondary amine modified silica (PSA)/graphitized carbon black (GCB) combination SPE column. Each purified pesticide extract was determined by both gas chromatography single quadrupole mass spectrometry using selected ion monitoring (GC-qMS-SIM) and by gas chromatography high resolution time-of-flight mass spectrometry (GC-HR-TOFMS). The geometric mean LOQs using the qMS and TOFMS were 4 and 3 ng/g ginseng, respectively. Mean recoveries from ginseng were 83, 79, and 75% with standard deviations of 4, 5, and 3%, respectively, for 25, 100, and 500 ng/g using GC-qMS-SIM. Mean recoveries using GC-HR-TOFMS were 93, 85, and 81% with mean standard deviations of 7, 7, and 8% for 25, 100, and 500 ng/g, respectively. Seven dried ginseng root products were found to contain combinations of the following pesticides: dacthal, diazinon, DDT, hexachlorobenzene, iprodione, lindane, procymidone, and quintozene (1-460 ng/g). No significant differences were found in the concentrations measured for these pesticides on commercial ginsengs using either of the two GC/MS techniques. The dietary supplement made with powdered Ginseng is derived from the root of the plant Panax quinquefolius (American ginseng) or P. ginseng (Asian ginseng) and is a widely used supplement to promote health. The root requires 4-6 years to grow and mature for harvest. The postharvest treatment of ginseng against pests allows sufficient time for the root to accumulate chemical contaminants, such as pesticides. Several recent studies have reported finding organochlorine pesticides such as quintozene and DDT (dichloro-diphenyl-trichloroethane) and their metabolites, lindane and other hexachlorocyclohexanes, * Author to whom correspondence should be addressed. Phone: (301)4361654. Fax: (301)436-2332. E-mail: [email protected].

5716

Analytical Chemistry, Vol. 81, No. 14, July 15, 2009

and endosulfan present in ginseng extracts and products.1-6 Pesticide and contaminant detection in dried and powdered ginseng is required to ensure safety and quality. Efficient, effective, and validated analytical methods are needed for these complex matrices. There are method reports that describe the screening of pesticides in dried botanical dietary supplements, spices, medicinal plants, herbals, and phytomedicines based on procedures primarily for fresh plant-derived foods.1-18 Multiresidue pesticide methods such as the one developed by Fillion et al.19,20 usually involve organic extraction of semivolatile pesticides from the plant matrix, a clean up procedure to remove coextractives and interferences, (1) Huggett, D. B.; Block, D. S.; Khan, I. A.; Allgood, J. C.; Benson, W. H. Human Ecol. Risk Assess. 2000, 6, 767–776. (2) Huggett, D. B.; Khan, I. A.; Allgood, J. C.; Blosck, D. S.; Schlenk, D. Bull. Environ. Contam. Toxicol. 2001, 66, 150–155. (3) Khan, I. A.; Allgood, J.; Walker, L. A.; Abourashed, E. A.; Schlenk, D.; Benson, W. H. J. AOAC Int. 2001, 84, 936–939. (4) Ahmed, M. T.; Loutfy, N.; Yousef, Y. Bull. Environ. Contam. Toxicol. 2001, 66, 421–426. (5) Durgnat, J.-M.; Heuser, J.; Andrey, D.; Perrin, C. Food Addit. Contam., Part A 2005, 22 (12), 1224–1230. (6) Leung, K. S.-Y.; Chan, K.; Chan, C.-L.; Lu, G.-H. Phytother. Res. 2005, 19, 514–518. (7) Molto´, J. C.; Lejeune, B.; Prognon, P.; Pradeau, D. Intern. J. Environ. Anal. Chem. 1994, 54, 81–91. (8) Manirakiza, P.; Covaci, A.; Schepens, P. Chromatographia 2000, 52 (1112), 787–790. (9) Zuin, V. G.; Vilegas, J. H. Y. Phytother. Res. 2000, 14, 73–88. (10) Srivastava, L. P.; Budhwar, R.; Raizada, R. B. Bull. Environ. Contam. Toxicol. 2001, 67, 856–862. (11) Bicchi, C.; Cordero, C.; Ioro, C.; Rubiolo, P.; Sandra, P.; Yariwake, J. H.; Zuin, V. G. J. Agric. Food Chem. 2003, 51, 27–33. (12) Barriada-Pereira, M.; Gonza´lez-Castro, M. J.; Muniategui-Lorenzo, S.; Lo´pezMahı´a, P. J. Chromatogr., A 2004, 1061, 133–139. (13) Hajou, R.; Afifi, F. U.; Battah, A. Pharm. Biol.(N. Y., NY, U. S.) 2005, 43 (6), 554–562. (14) Rodrigues, M. V. N.; Reyes, F. G. R.; Magalha˜es, P. M.; Rath, S. J. Braz. Chem. Soc. 2007, 18 (1), 135–142. (15) Jeon, H.-R.; Abd El-Aty, M.; Cho, S.-L.; Choi, J.-H.; Kim, K.-Y.; Park, R.-D.; Shim, J. H. J. Sep. Sci. 2007, 30, 1953–1963. (16) Kong, M. F.; Chan, S.; Wong, Y. C.; Wong, S. K.; Sin, D. W. J. AOAC Int. 2007, 90 (4), 1133–1141. (17) Chan, S.; Kong, M. F.; Wong, Y. C.; Wong, S. K.; Sin, D. W. J. Agric. Food Chem. 2007, 55 (9), 3339–3345. (18) Park, Y. S.; Abd El-Aty, A. M.; Choi, J. H.; Cho, S. K.; Shin, D. H.; Shim, J. H. Biomed. Chromatogr. 2007, 21 (1), 29–39. 10.1021/ac900494a Not subject to U.S. Copyright. Publ. 2009 Am. Chem. Soc. Published on Web 06/23/2009

and subsequent analysis such as capillary gas chromatography (GC). Pesticide procedures can be costly due to the large amounts of solvents and expensive consumables used, so there is a need for the methods to be fast, robust, automated, and cost-efficient. Methods also require validation and testing on many specific products to ensure routine and reliable screening. Anastassiades et al.21 proposed a method that utilizes smaller sample sizes and fewer solvents and less laboratory glassware, simplifies procedures, and utilizes solid-phase dispersive sorbents as a costefficient alternative to solid-phase extraction cartridges. Improved procedures have combined the effectiveness of the cleanup procedures described by Fillion et al.19,20 and the efficiency and the cost-effectiveness of the method developed by Anastassiades et al.21 In previous our work, a modification of the method developed by Anastassiades et al.,21 employing dispersive SPE materials discussed by Fillion et al.19 and using capillary gas chromatography-single quadrupole mass spectrometry with selected ion monitoring (GC-qMS-SIM) and GC-FPD was successfully employed with 110 organophosphorous pesticides in powdered ginseng root.22 This study describes a method that is a substantial improvement over the existing techniques for the analysis of organohalogen and organophosphorous pesticides on ginseng. Pesticides, isomers, and metabolites in this study were measured using two different GC/MS techniques. One instrument used was a GC-qMS-SIM. This approach has been widely used for the targeted determination of pesticides in ginseng and other foods.17,19-22,27 Single quadrupole mass spectrometers have become a ubiquitous feature of many analytical laboratories due their reliability, effectiveness, and low cost. While GC-qMS-SIM provides qualitative and quantitative MS information on pesticide residues in foods, there are difficulties with this approach.28 Common pesticide identification criteria using single quadrupoles in SIM mode call for the detection, and some criteria require four ions at the proper ion ratios in combination with correct GC retention time and other QA to take regulatory action.28,29 Efficient (19) Fillion, J.; Hindle, R.; Lacroix, M.; Selwyn, J. J. AOC Int. 1995, 78, 1252– 1266. (20) Fillion, J.; Sauve´, F.; Selwyn, J. J. AOAC Int. 2000, 83, 698–713. (21) Anastassiades, M.; Lehotay, S. J.; Sˇtajnbaher, D.; Schenck, F. J. J. AOAC Int. 2003, 86, 412–431. (22) Wong, J. W.; Hennessy, M. K.; Hayward, D. G.; Krynitsky, A. J.; Cassias, I.; Schenck, F. J. J. Agric. Food Chem. 2007, 55 (4), 1117–1128. (23) Focant, J.-F.; Sjo ¨din, A.; Wayman, T. E.; Patterson, D. G., Jr. Anal. Chem. 2004, 76, 6313–6320. (24) Focant, J.-F.; Pirard, C.; Eppe, G.; De Pauw, E. J. Chromatogr., A 2005, 1067, 265–275. (25) Focant, J.-F.; Eppe, G.; Scippo, M.-L.; Massart, A.-C.; Pirard, C.; MaghuinRogister, G.; De Pauw, E. J. Chromatogr., A 2005, 1086, 45–60. ˇ ajka, T. J. Chromatogr., A 2003, 1019, 173– (26) Zrostlı´kova, J.; Hajsˇlova, J.; C 186. (27) Pang, G. F.; Fan, C. L.; Liu, Y. M.; Cao, Y. Z.; Zhang, J. J.; Li, X. M.; Li, Z. Y.; Wu, Y. P.; Guo, T. T. J. AOAC Int. 2006, 89 (3), 740–771. (28) Lehotay, S. J.; Mastovska, K.; Amirav, A.; Fialdov, A. B.; Alon, T.; Martos, P. A.; de Kok, A.; Ferna´ndez-Alba, A. R. Trends Anal. Chem. 2008, 27 (11), 1070–1080. (29) SOP No. CR-Labop-301, Identification of Residues. Bureau of Chemical Residue Lab, Florida Department of Agriculture and Consumer Services. Effective July 1, 2003. (30) Lehotay, S. J.; Mastovska´, K.; Yun, S. J. J. AOAC Int. 2005, 88 (2), 630– 638. ˇ ajka, T.; Hajsˇlova´, J. J. Chromatogr., A 2004, 1058, 251–261. (31) C (32) Leandro, C. C.; Hancock, P.; Fussell, R. J.; Keely, B. J. J. Chromatogr., A 2007, 1166, 152–162. (33) Schenck, F. J.; Lehotay, S. J. J. Chromatogr., A 2000, 868 (1), 51–61.

pesticide methods normally attempt to measure large numbers (>100) of pesticides in a single GC/MS determination. A very large number (∼50) of overlapping mass ions must be monitored in many acquisition segments when using narrow bore capillary GC columns in a qMS SIM method. In this situation, SIM software programs are required to operate near their limits. Other MS techniques such GC/MS2 and LC/MS2 could be considered for screening.36 In the case of ginseng, most of the residues found would have little or no response in pneumatically assisted electrospray LC/MS.1-6,36 An alternative to GC-qMS SIM might be to use a full scan approach. Full scan analysis MS using a quadrupole produces lower sensitivity than SIM hindering attempts to reach the default maximum residue levels at 10 ng/g.28 Further manipulation of the pesticide extracts could be tried to help compensate for lower sensitivity in full scan analysis. Some common approaches would require any one or combination of the following: greater concentration, higher injection volumes, or more extensive cleanup. These approaches add significantly to the cost or may be ineffective for some foods.33,35,37 An alternative technique investigated in this study utilizes a capillary gas chromatograph time-of-flight mass spectrometer in electron ionization (EI) mode with accurate mass measurement (GC-HR-TOFMS). Time-of-flight instruments can have higher spectral acquisition rates than quadrupole instruments which reduces spectral skewing. The spectra contain all ions formed in the source with the possibility of extending the mass range to m/z > 1000 with nominal or accurate mass resolution.23-26,31,32,37-39 The GCT premier used in this study acquires accurate mass measurements at modest collection rates (99% purity. Pesticides are poisons and must be handled cautiously using solvent resistant gloves, fume hoods, and a dedicated preparation area with the necessary safety equipment to handle spills. Pesticide-grade ethyl acetate (EtOAc), cyclohexane, toluene, and anhydrous magnesium sulfate were purchased from Fisher Scientific (Pittsburgh, PA). Internal standards and surrogates, acenaphthalene-d10, phenanthrene-d10, and chrysene-d12, were purchased from Aldrich Chemical Corp. (Milwaukee, WI) and ChemService (West Chester, PA). PSA sorbents GCB combination SPE columns were purchased from United Chemical Technologies (United Chemical Technologies, Bristol, PA). Dried and powdered ginseng samples, Panax quinquefolius (American ginseng) and P. ginseng (Asian ginseng), were purchased in bulk packages from commercially available sources. Pesticide stock solutions were prepared by dissolving 50-100 mg of a pesticide in 25 mL of toluene. The working standards used for quantitation were prepared by mixing 1-2 mL of each standard using a 250 mL volumetric flask to prepare a 20 µg/mL working stock solution. Three mixtures were prepared each containing 50-70 pesticides. The three groups were made so that pesticides standards could be easily evaluated separately by GC/ MS eliminating incomplete separations of isomers with close retention times in the GC. Successive dilutions of the stock pesticide standards were used to prepare 6670, 3330, 1670, 833, 333, 167, 83, 33, 16.7, 8.3, and 3.3 ng/mL working solutions in toluene (each 50 mL standards). The internal standards were prepared by dissolving acenaphthalene-d10, phenanthrene-d10, and chrysene-d12 to make 20 µg/mL working solution. Fortification solutions were prepared in acetonitrile. Ginseng Extract Preparation. Test portions, each one at 2.50 g, of dry powdered ginseng root were weighed into 50 mL disposable screw capped polypropylene centrifuge bottles (Fisher Scientific). Test portions were fortified with pesticide mixture standard groups 1, 2, and 3. The standards were omitted when extracting ginseng to obtain matrix for matrix matched standards. Twenty-five milliliters of EtOAc was added followed by a stainless steel ball bearing. The mixtures were placed on a Geno/Grinder mechanical shaker (SPEX Sample Prep, LLC, Metuchen, NJ) for 5 min at 1000 strokes/minute. The polypropylene tube was centrifuged (Thermo-electron CR4i) at 4200 g × 5 min, and the extract was collected in a 170 mL Rapidvap concentration tube (Labconco). The ginseng powder was extracted and centrifuged again with another 25 mL of EtOAc. The ∼50 mL extract was reduced to ∼2 mL and brought up to 5 mL using 70:30 EtOAc: cyclohexane. 5718

Analytical Chemistry, Vol. 81, No. 14, July 15, 2009

Gel Permeation Chromatography (GPC). The ginseng extract (5 mL) was injected into a GPC system (J2 Scientific, Columbia, MO, USA) using a 70:30 EtOAc:cyclohexane mobile phase at a flow rate of 5 mL/min. The GPC column was a glass column (25 mm i.d.) packed with 22 g of stryrene divinylbenzene 40-80 µm beads (S-X3 biobeads BIO-RAD, Hercules, CA) swelled in 70:30 EtOAc:cyclohexane. The first 42.5 mL (8.5 min.) of the eluate was discarded, and the next 47.5 mL (9.5 min.) eluate was collected and reduced to approximately ∼1 mL. Inline UV detection was used to monitor the relative effectiveness of the GPC procedure. Solid-Phase Extraction (SPE). The concentrated GPC extract was loaded onto a conditioned tandem bed SPE cartridge (United Chemical Technologies, Bristol, PA) containing 0.25 g GCB (top) and 0.5 g PSA (bottom) sorbents. The cartridges were eluted with 15 mL 3:1 EtOAc:toluene, and the sample was split for GC/MS analysis (9 mL) and (LC-MS/MS) analysis (6 mL) not discussed in this paper. The GC extract was reduced (∼100 µL) and brought to 0.5 mL of toluene or 3 g ginseng /mL. Internal standards were added, 25 µL of a 20 µg/mL mixture of acenaphthalene-d10, phenanthrene-d10, and chrysene-d12. Each final fortified or incurred ginseng extract was injected into both qMS and HR-TOFMS. GC-qMS-SIM. An Agilent 6890N gas chromatograph was equipped with an Agilent 5973 mass selective detector (Agilent Technologies, Little Falls, DE) and fitted with a deactivated guard column (5 m × 0.25 mm i.d.) and HP-5MS column (30 m × 0.25 mm i.d. × 0.25 µm film thickness, Agilent Technologies) with a 1.0 mL/min He flow rate. The temperature program consisted of 105 °C (1 min hold) to 130 °C at a rate of 10 °C/min, increased to 230 at 4 °C/min, followed by a final ramp to 290 at 20 °C/min (13 min hold). The MSD was operated in EI mode at 70 eV. The injector, transfer line, qMS source, and quadrupole temperatures were 250, 290, 230, and 150 °C, respectively. The ginseng extracts, standards, and blanks were injected (1 µL) into the GC in pulsed splitless mode (pulsed pressure ) 35.0 psi; pulsed time ) 2.00 min) using an Agilent 7683 series autoinjector. The qMS system was routinely programmed in SIM mode using one target and three qualifier ions. Typically, 4-54 ions were monitored in an acquisition segment with 25 total segments, using 10-100 ms dwell times depending on the number ions. The SIM scan cycle time for each segment was between 0.36-0.54 s. Identification by mass spectrometry was established by the retention time of the target ion and the presence of two or three qualifier-to-target ion ratios. Relative response factors for each pesticide with the internal standard were calculated from a standard curve using at least two standard concentrations above and two below and one concentration near to the concentration of a fortified ginseng. Quantitation by GC-qMS SIM was based on the peak area ratios of the target ions of the analyte to that of the internal standard (the internal standard with the retention time closest to that of the pesticide) and compared to concentrations of matrix-matched calibration standards using the ChemStation software. GC-HR-TOFMS. A Waters Corporations GCT Premier interfaced with an Agilent Technologies 6890N gas chromatograph and fitted with a deactivated guard column (5 m × 0.25 mm i.d., Restek Corporation) and HP-5MS column (30 m × 0.25 mm i.d. × 0.25 µm film thickness, Agilent Technologies). The injector, transfer

Figure 1. UV absorption trace from gel permeation chromatography cleanup of a ginseng extract prepared using a Geno/Grinder for extraction with EtOAc. The vertical scale is in absorbance. Collection time was indicated with an arrow at 8.5 min (42.5 mL).

line, and source temperatures were 280, 250, and 220 °C, respectively. Flow rate for He was set to 1.0 mL/min with constant flow enabled. Ginseng test portions, fortified ginsengs, matrix matched standards, and blanks were injected (1 µL) into the GC in splitless mode at 280 °C using a Gerstel MPS2 autosampler. The 6890N Agilent GC was not capable of pressure pulsed splitless injection under Waters Corp. software control like the Agilent 6890N interfaced to the mass selective detector. The temperature program rate for the GCT 6890N GC was increased over that of the Agilent GC interfaced to the mass selective detector to help compensate. The temperature program consisted of 90 °C (1 min hold) to 150 °C at a rate of 10 °C/min, increased to 280 at 5 °C/ min (11 min hold). Ginseng extracts that had been fortified at 500 ng/g were injected in split mode (10/1) along with matrix matched standards (833 to 6670 ng/mL). The GCT was operated in EI mode at 70 eV. Ion pulses were averaged at a rate of 4/s with dynamic range enhancement (DRE) turned on in lock mass mode using m/z 265.9964 lock mass from tris 2,4,6-trifluoromethyl 1,3,5-triazine. Resolution was >6000 with the lock mass peak 600 Da) are expected to Analytical Chemistry, Vol. 81, No. 14, July 15, 2009

5719

Figure 2. Standard curve for delta-BHC generated using TOFMS with the standards in toluene, 1 µL splitless injection -16 ng/mL to 6670 ng/mL. The plotted points that are crossed out were not used to calculate the r2 ) 0.986.

inevitably suffer using these GPC conditions and would require a different approach. After the GPC cleanup the purification of the concentrated extract was completed by elution through an SPE column containing 0.5 g of PSA and 0.25 g of GCB. GC/MS analysis was now possible with improved chromatography and greatly reduced interference for nearly all pesticides investigated. Dynamic Range. Table S-1 lists the dynamic ranges determined for HR-TOFMS and GC-qMS-SIM. The GC-qMS-SIM dynamic ranges listed encompassed most of the entire range of standard concentrations used (3.3-6670 ng/mL). The major exceptions were cypermethrin, cyfluthrin, methidation, isoazophos, tribufos, and demeton-S-methyl which all had higher LOQs of 83 ng/mL. HR-TOFMS matrix standard responses produced a linear response with a mean r2 ) 0.990 achieved over a 10-fold lower concentration range than for qMS (3.3-833 ng/mL). TOFMS instruments have more limited dynamic ranges than do quadrupoles.24,30-32 The GCT premier attempts to compensate for this limitation with a feature called dynamic range enhancement (DRE). The DRE feature alternately produces low and high sensitivity sets of pushes that are used to form a single recorded spectrum. During the low sensitivity set of pushes the ion beam was deflected such that ∼95% of the ions do not reach the MCP detector. Typically, only the high sensitivity spectrum was stored unless saturation of the MCP was detected. Initially pesticides standard mixes in solvent were analyzed over the entire concentration range (3.3-6670 ng/mL) using the DRE. Nearly all pesticides behaved in a similar manner, with the HR-TOFMS estimating a larger response than expected above 333-833 ng/ mL (Figure 2). There were several exceptions including tecnazene and DDE-op′ which both achieved r2 >0.992 over the entire range tested (3.3-6670 ng/mL). Quintocence gave an r2 of 0.996 over most of the range (3.3-1670 ng/mL). We concluded that 833 ng/mL would be the highest useful concentration for most pesticides to produce an r2 > 0.99 with the DRE feature turned on. A reduced linear dynamic range for the TOFMS may at first glance be considered a disadvantage for accomplishing screening and quantitation in a single determination. The view neglects to 5720

Analytical Chemistry, Vol. 81, No. 14, July 15, 2009

consider the role these methods play in a pesticide monitoring program. Ginseng extracts screened in this study were concentrated 3-fold which provides a useful quantitative range for the HR-TOFMS up to 833 ng/mL extract or 280 ng/g of ginseng. This range was sufficient to encompass nearly all of the pesticide residues in the ginsengs (Table 1). The DDE in the incurred ginseng (460 ng/g) was injected using a 10/1 split injection resulting in much less analyte reaching the MCP. Method Validation. Standard mixes in solvent were used to determine retention times, mass spectra, and acquisition windows (GC-qMS-SIM) for 158 pesticides, isomers, and metabolites. Matrix matched standards containing 3.3-6670 ng/mL established the linear range for each pesticide and the minimum LOQs expected for the method (Table S-1). The pesticide target and 3 qualifier ions used by qMS were preselected for the SIM method constructed in the Agilent ChemStation software. The accurate mass ions in Table S-1 ion were deduced from the theoretical elemental compositions for each from the pesticide structure and the NIST05 library spectrum. The high resolution mass spectrum obtained from the pesticide standard with the GCT Premier was used to confirm accurate masses calculated from the elemental compositions. For the purposes of contrasting the performance of the qMS with the HR-TOFMS the LOQ was defined as the amount of pesticide that would provide 3/1 signal/noise in the matrix matched standards for all qualifying ions and 10/1 for the target ion used quantitation divided by the concentration factor for ginseng. The pesticides, isomers, and metabolites fortified in ginseng are listed in Table S-1 along with the GC retention times, LOQs for the MSD and TOF instruments, target and qualifying ions, calibration ranges, and r2 for HR-TOFMS. The geometric mean LOQ for 170 pesticides, isomers, and metabolites using TOFMS was somewhat lower (3 ng/mL) than the mean LOQ using qMS (4 ng/mL). Most LOQs were found to be within a factor of 2 between the instruments. The sensitivity of the HR-TOFMS in high resolution lock mass mode was roughly equivalent to this quadrupole operated in SIM mode with average LOQs in the low ng/g range. Several pesticides were more sensitive by HR-TOFMS than qMS, while a few were found to be less sensitive (Table S-1). The LOQ needed for pesticide determinations is assumed to be 10 ng/g wet weight for foods.27,28 Based on the matrix matched standards the LOQs for most pesticides were expected to be mainly below 10 ng/g in dry ginseng. The main exceptions being some of the pyrethroid esters pesticides (acrinathrin, cypermethrin, cyfluthrin, and fenvalerate) and a few others such as demeton-S-methyl, isoazophos, methidation, tribufos, and temephos that have LOQs of 33-83 ng/mL by either MS technique. Table S-2 presents the mean recoveries and standard deviations for 170 pesticides fortified at 25, 100, and 500 ng/g dry weight of ginseng. Each fortified ginseng extract was measured using GCqMS-SIM and GC-HR-TOFMS. The recoveries for both detection systems are listed in Table S-2. The grand mean recoveries for the GC-qMS-SIM were 83, 79, and 75% with mean standard deviations of 4, 5, and 3% at 0.025, 0.1, and 500 ng/g fortifying concentrations, respectively. The HR-TOFMS gave similar mean recovery values of 93, 85, and 81% with mean standard deviations of 7, 7, and 8% for 25, 100, and 500 ng/g fortifying concentrations,

Table 1. Ginseng Pesticide Concentrations Found in American and Asian Ginsengs from Commercial Sources with BHC Isomers, Diazinon, DDT, Procymidone, Quintocence, and Transformation Productsa Asian Panax ginseng diazinon tetrachloroaniline hexachlorobenzene* pentachloroaniline pentachloroanisole pentachlorobenzene pentachlorobenzonitrile pentachlorothioanisole quintozene* tecnazene R-BHC β and γ-BHC δ-BHC p,p′-DDE p,p′-DDD* p,p′-DDT procymidone a

Asian Red Panax ginseng

white ginseng Panax quinquefolius

American Panax quinquefolius

#1

#2

#3

qMS

TOF

qMS

TOF

qMS

TOF

qMS

TOF

qMS

TOF

qMS

TOF

nd 8.6 ± 0.3 26 ± 0.5 230 ± 10 nd 25 ± 3 nd 42 ± 2 104 ± 1 nd 12 ± 0.2 8.2 ± 0.9 61 ± 1 nd nd nd nd

nd 8.3 ± 0.9 23 ± 0.3 236 ± 30 nd 21 ± 2 nd 42 ± 3 124 ± 6 2.5 ± 1 11 ± 1 8.1 ± 0.1 66 ± 5 nd nd nd nd

nd 27 ± 0.7 55 ± 1 290 ± 30 nd 18 ± 1 nd 67 ± 3 430 ± 10 21 ± 1 na na na nd nd nd nd

nd 28 ± 0.5 58 ± 0.8 310 ± 7 nd 17 ± 0.6 nd 73 ± 1 421 ± 22 20 ± 0.5 58 ± 1 190 ± 3 310 ± 3 nd nd nd nd

nd nd nd nd nd nd nd nd nd nd nd nd nd 460 ± 11 16 ± 0.2 95 ± 4 nd

nd nd nd nd nd nd nd nd nd nd nd nd nd 450 ± 20 13 ± 0.5 86 ± 4 nd

19 ± 1 nd nd 1.9 ± 0.1 0.95 ± 0.1 0.9 ± 0.1 2.8 ± 0.4 1.2 ± 0.1 9±1 nd nd na nd nd nd nd nd

17 ± 1 nd nd 1.6 ± 0.3 nd 1.6 ± 0.3 2.4 ± 0.4 0.7 ± 0.1 7.3 ± 0.5 nd nd 46 ± 4 nd nd nd nd nd

nd nd 1.7 ± 0.1 8.0 ± 0.9 nd 2.0 ± 0.2 nd 2.2 ± 0.2 7.4 ± 1.1 nd 5.4 ± 0.3 na na nd nd nd nd

nd nd 1.6 ± 0.3 7.1 ± 1.2 nd 1.8 ± 0.3 nd 2.2 ± 0.4 6.4 ± 1.6 nd 4.9 ± 0.7 8.3 ± 0.7 16 ± 3 nd nd nd 2.1 ± 0.3

nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd 16 ± 0.8

nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd 17 ± 0.7

Values are ng/g, means ± standard deviations, n ) 3. * ) significant T-test result p < 0.025, critical point; p < 0.025. nd ) not detected.

Figure 3. Accurate mass chromatograms for quintozene in Panax ginseng (A) at 124 ng/g (note that incurred m/z 262.86 and 266.85 coincide nearly completely); Panax quinquefolius Asian white ginseng with quintozene at 7.3 ng/g (B); diazinon, 17 ng/g (C); and Panax ginseng Asian red with hexachlorobenzene at 58 ng/g (D). All incurred traces are shown with the appropriately scaled matrix matched standards superimposed to the right except B (coincides). Standard concentrations were (A) 333 ng/mL, (B) 17 ng/mL, (C) 83 ng/mL, and 167 ng/mL (D). All standards and incurred ginsengs were acquired on the same day.

respectively. The mean recoveries for 20% of 151 pesticides were significantly different in a two sided t test for at the highest fortification level (t test, p < 0.025) (Table S-2). More differences were observed at the lower fortification levels with most often higher calculated recoveries found for HR-TOFMS than qMS. Most often the HR-TOFMS results demonstrating significant differences were higher than the qMS except in a few cases such as tecnazene, aldrin, endrin, and cis-chlordane.

A statistically detected difference did not translate into a useful difference in results from an analytical or regulatory perspective. For example, diazinon found in an incurred white ginseng #1 at 18 ng/g ± 1.5 (95% CI, n ) 6) (Table 1) is clearly below the tolerance for ginseng of 75 ng/g and above the method LOQ of 1 ng/g using TOFMS results in Table S-1. The qMS and TOFMS results in the incurred sample were not significantly different. However, recoveries at 25 ng/g for diazinon were statistically Analytical Chemistry, Vol. 81, No. 14, July 15, 2009

5721

Figure 4. Accurate mass chromatograms and nominal mass chromatograms for alachlor in American ginseng fortified at 100 ng/g. Retention times are 18.06 min for HR-TOFMS and 22.57 min qMS. (A) qMS SIM mass chromatograms m/z 160, 188, 146, 237 and (B) HR-TOFMS m/z 160.113, 188.108, 237.092, 160.113, 146.097, 269.11.

significantly different using qMS or HR-TOFMS (t test p < 0.01; critical point p < 0.025) but not at 100 or 500 ng/g. Recoveries were 60-104% for 151 of the pesticides by GC-qMSSIM and 40-140% for 164 pesticides by HR-TOFMS (Table S-2). In the cases of chlorothalonil and dichlofluanid some loss was due to chemical degradation on the PSA adsorbent. Overall, the precision and accuracy of both instrumental techniques were similar at either the low or higher fortification levels. The fluridone recoveries by HR-TOFMS were systematically high at all fortification levels, 128-147, while recoveries based on qMS SIM were 93- 96%. The matrix matched standard curve was quite linear (r2 )0.993) for fluridone and had no major coeluting matrix peaks that might have affected the response. Table S-2 lists eleven pesticides, isomers, or metabolites with no result (na) for GCqMS SIM but with a recovery result for HR-TOFMS. These 5722

Analytical Chemistry, Vol. 81, No. 14, July 15, 2009

pesticides were not found due to the fact that their characteristic ions were not included in the SIM acquisition program. Incurred Pesticides. Table 1 provides the results for triplicate determinations of 6 commercial ginsengs products collected in 2005. The ginsengs extracts were measured using both GC-qMSSIM and GC-HR-TOFMS. All contained residues above the LOQs and were no tolerance violations. Residue mean values measured by GC-HR-TOFMS or GC-qMS-SIM were not significantly different with the exception of DDD-pp′ (t test, p < 0.025), quintocence, and hexachlorobenzene in one ginseng (t test, p < 0.025). This result was not surprising, because few significant differences were found between the fortified ginseng recoveries for the pesticides found in commercial ginsengs (Table S-2) with exceptions of tecnazene at all fortified levels, DDD-p,p′ and DDT-p,p′ at the 25 and 100 ng/g, and diazinon at 25 ng/g.

Figure 3 provides the accurate masses found for incurred ginsengs along with a matrix matched standard to the right of the incurred mass chromatograms with the correct RRTs and ion ratios. Quintozene shown in Asian white ginseng at 7.3 ng/g and diazinon (17 ng/g) were below and just above the target LOQ, respectively. Accurate masses were found within 1-3 mDa of the accurate mass recorded with the matrix matched standards and usually 1-3 mDa from the calculated accurate mass (Table S-1) for the fragment or molecular ion trace based on the assumed empirical forumla as has been reported previously.31 The two American ginsengs with no quintozene contained only procymidone or DDT. These pesticides have been reported previously in ginseng root.1-3,34 The ginseng product with DDT contained 5-fold lower amounts than its main degradation product DDE-pp′ suggesting that the contamination was due to a persistent residue on soil from past applications.40,4 Ginseng has no tolerance for DDT and 90 ng/g found was above the LOQ (3 ng/g). Tolerances or action levels have been established for some pesticides on ginseng, for example, tolerances for iprodione at 2000 ng/g and diazinon at 75 ng/g or the action level for total BHC isomers, 50 ng/g. Iprodione as well as dacthal were found on some ginseng batches used in the fortification studies. Iprodione was below the 2000 ng/g tolerance (∼50 ng/g). Dacthal was estimated to be