Anal. Chem. 1999, 71, 212-220
Tandem-in-Time Mass Spectrometry Method for the Sub-Parts-per-Trillion Determination of 2,3,7,8-Chlorine-Substituted Dibenzo-p-dioxins and -furans in High-Fat Foods Douglas G. Hayward,*,† Kim Hooper,‡ and Denis Andrzejewski†
U.S. Food and Drug Administration, 200 C Street SW, Washington, D.C. 20204, and Hazardous Materials Laboratory, California Environmental Protection Agency, 2151 Berkeley Way, Berkeley, California 94704
Limits of quantitation (LOQs) for a quadrupole ion storage tandem-in-time mass spectrometry (QISTMS) method were evaluated through replicate analysis of unfortified peanut oil, shortening, lamb fat, and butter for all 2,3,7,8chlorine-substituted polychlorodibenzo-p-dioxins (PCDDs) and polychlorodibenzofurans (PCDFs). Ten congeners were measurable in butter (0.27-2.5 pg/g) and nine congeners were measurable in lamb fat (0.09-2.6 pg/g) with good precision. LOQs for high-fat foods were estimated by triplicate analysis of peanut oil fortified at two levels. Accurate and reproducible results were achieved at 0.5 pg/g for most PCDD/Fs (1.0 pg/g for heptachlorodibenzo-p-dioxin and heptachlorodibenzofuran and 2.0 pg/g for octachlorodibenzofuran) and at 0.2 pg/g for 2,3,7,8-tetrachlorodibenzofuran (TCDF) and 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD). QISTMS distinguished between catfish and chicken eggs with elevated TCDD levels from background samples collected from the most regions of the continental United States. QISTMS determined the extent of TCDD contamination in butter, lamb fat, and cottonseed oil collected from rural villages in Kazakhstan. Replicate analysis of catfish and chicken eggs by the QISTMS method produced comparable results to high-resolution mass spectrometry (HRMS). Lower limits of detection will be needed if QISTMS is to fully complement HRMS in the measurement of TCDD levels in food. The 2,3,7,8-substituted polychlorodibenzo-p-dioxins (PCDDs) and polychlorodibenzofurans (PCDFs) are extremely toxic to certain animal species, and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) has recently been classed with other known human carcinogens.1 They are produced in small quantities from a large number of diverse man-made and natural sources.1,2 These chemicals contaminate most areas of the world and accumulate in the food web. The major source of nonoccupational exposures †
U.S. Food and Drug Administration. California Environmental Protection Agency. (1) Polychlorinated dibenzo-p-dioxins and Polychlorinated dibenzofurans; IARC Monographs on the Evaluation of Carcinogenic Risks to Humans; IARC Press: Lyon, France, 1997; Vol. 69. (2) Travis, C. C.; Hattemer-Frey, H. A. Sci. Total Environ. 1991, 104, 97. ‡
212 Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
to humans from PCDD/Fs is animal foods.2-4 Measuring PCDD/ Fs background concentrations in foods requires methods that provide extremely high sensitivity (low-picogram to low-femtogram range). Specificity is obtained through extensive purification followed by determination with a GC/MS technique. Limits of Quantitation Appropriate for Animal Foods. PCDDs and PCDFs are measured in foods at concentrations below 1 pg/g of lipid.5-10 PCDD/F levels reported on the U.S. food samples often originate from limited sampling of a few foods10 in a single region. Daily intakes are based on a few measurements of pooled samples from a few regions.11,12 The U.S. Food and Drug Administration (FDA) collects data on the U.S. food supply.13,14 Past FDA studies measured fish and shellfish with relatively high levels (“hot spots”) of 2,3,7,8-TCDD, for example, fish from the Great Lakes.14 The United States Environmental Protection Agency (EPA) began a systematic study of PCDD/F levels in commercial animal fat using a statistically designed sampling strategy. Empirically validated limits of quantitation (LOQs) in beef back fat provided high confidence in the data set reported for background levels in American beef.15 In support of these efforts, the FDA has begun collecting and analyzing a large number of dairy, fish, and shellfish products throughout the (3) Henry, S.; Cramer, G.; Bolger, M.; Springer, J.; Scheuplein, R. Chemosphere 1992, 25, 235. (4) Gilman, A.; Newhook, R. Chemosphere 1991, 23, 1661. (5) Beck, H.; Dross, A.; Mathar, W. Environ. Health Perspect. 1994, 102 (Suppl 1), 173. (6) Fu ¨ rst, P.; Fu ¨ rst, C.; Groebel, W. Chemosphere 1990, 20, 787. (7) Fu ¨ rst, P.; Fu ¨ rst, C.; Wilmers K. Chemosphere 1992, 25, 1039. (8) Startin, J. R.; Rose, M.; Wright, C.; Parker, I.; Gilbert, J. Chemosphere 1990, 20, 793. (9) Theelen, R. M. C.; Liem, A. K. D.; Slob, W.; Van Wijnen, J. H. Chemosphere 1993, 27, 1625. (10) Schecter, A.; Startin, J.; Wright, C.; Kelly, M.; Papke, O.; Lis, A.; Ball, M.; Olson, J. Chemosphere 1994, 29, 2261. (11) Schecter, A.; Cramer, P.; Boggess, K.; Stanley, J.; Olson, J. R. Chemosphere 1997, 34, 1437. (12) Fiedler, H.; Cooper, K. R.; Bergek, S.; Hjelt, M.; Rappe, C. Chemosphere 1997, 34, 1411. (13) Firestone, D.; Fehringer, N. V.; Walters, S. M.; Kozara, R. J.; Ayres, R. J.; Ogger, J. D.; Schneider, L. F.; Glidden, R. M.; Ahlrep, J. R.; Brown, P. J.; Ford, S. E.; Davy R. A.; Gulick, D. J.; McCullough, B. H.; Sittig, R. A.; Smith, P. V.; Syvertson, C. N.; Barber, M. R. JAOAC Int. 1996, 79, 1174. (14) Fehringer, N. V.; Walters, S. M.; Kozara, R. J.; Schneider, L. F. J. Agric. Food Chem. 1985, 33, 626. (15) Ferrario, J.; Byrne, C.; McDaniel, D.; Dupuy, A., Jr. Anal. Chem. 1996, 68, 647. 10.1021/ac980282+ CCC: $18.00
© 1998 American Chemical Society Published on Web 11/20/1998
United States. The goal is to produce a data set on food levels that can be used to estimate human exposures. Comparability between data sets is essential. The FDA and EPA must demonstrate that their methodologies will routinely provide similar LOQs with the expected precision and accuracy. Tandem Mass Spectrometry. Low-resolution electron impact multiple ion detection mass spectrometry (EI-LRMS) is not specific and sensitive enough for food or other low-level PCDD/F analysis.16 Methane-enhanced electron capture negative chemical ionization (NCI) with multiple ion detection (MID) can provide very high sensitivities for PCDFs, but 10-fold lower sensitivities for PCDDs, and no useful response for 2,3,7,8-TCDD (∼102 lower response).17 During the past decade, high-resolution mass spectrometry (HRMS) has become the method of choice for obtaining reliable determinations of 2,3,7,8-TCDD at levels appropriate for food analyses. Tandem mass spectrometry has produced results comparable to that of high-resolution mass spectrometry for PCDDs and PCDFs in soils,18,20 fish,19,21,22,24,26 fly ash,23 sediments,19,25 and human milk.26 Contaminated soils were rapidly screened with minimal sample preparation and GC separation using a triplequadrupole tandem mass spectrometry.20 Hybrid GC/MS/MS systems afforded greater specificity in pulp effluent27,28 than the conventional HRMS approach. Fraisse et al. reported pentachlorodibenzo-p-dioxin (PeCDD) interferences in fly ash that were not eliminated by a triple quadrupole in MS/MS mode or by HRMS.23 McCurvin et al. reported greater sensitivity and specificity by MS/ MS over EI-LRMS using a triple-quadrupole system21 but found chlorinated diphenyl ether interferences for TCDFs with both EILRMS and MS/MS. TCDD sensitivity comparable to HRMS has been reported using MS/MS.19,24,29 Often MS/MS provides lower sensitivity than HRMS by a factor of 5 or more.18,23,25,26 Until recently, tandem mass spectrometry systems, either hybrid HRMS systems or a triple-quadrupole MS, have been as costly to purchase and operate as HRMS. Increasing demand for PCDD/F measurements in foods prompted the FDA to investigate an alternative MS/MS method using a relatively inexpensive quadrupole ion storage mass (16) Clement, R. E.; Bobbie, B.; Taguchi, V. Chemosphere 1986, 15, 1147. (17) Oheme, M.; Kirschmer, P. Anal. Chem. 1984, 56, 2754. (18) Tondeur, Y.; Niederhut, W. N.; Campana, J. E. Biomed. Environ. Mass Spectrom. 1987, 14, 449. (19) Bobbie, B. A.; Clement, R. E.; Taguchi, V. Y. Chemosphere 1989, 18, 155. (20) Smith, J. S.; Ben Hur, D.; Urban, M. J.; Kleopfer, R. D.; Kirchmer, C. J.; Smith, W. A.; Viswanathan, T. S. In Chlorinated Dioxins and Dibenzofurans in Perspective, Rappe, C., Choudhary, G., Keith, L. H., Eds.; Lewis Publishers: Chelsea, MI, 1986; pp 367-380. (21) McCurvin, D. M. A.; Schellenberg, D. H.; Clement R. E.; Taguchi. V. Y. Chemosphere 1989, 19, 201. (22) Huang, L. Q.; Eitzer, B.; Moore, C.; McGown, S.; Tomer, K. B. Biol. Mass Spectrom. 1991, 20, 161. (23) Fraisse, D.; Gonnord M. F.; Becchi, M. Rapid Commun. Mass Spectrom. 1989, 3, 79. (24) Reiner, E. J.; Schellenberg D. H.; Taguchi, V. Y. Environ. Sci. Technol. 1991, 25, 110. (25) Charles, M. J.; Green W. C.; Marbury, G. D. Environ. Sci. Technol. 1995, 29, 1741. (26) de Jong, A. P. J. M.; Liem, A. K. D.; den Boer, A. C.; van der Heeft, E.; Marsman, J. A.; van de Werken, G.; Wegmen, R. C. C. Chemosphere 1989, 19, 59. (27) Charles, M. J.; Green, B.; Tondeur, Y.; Hass, J. R. Chemosphere 1989, 19, 51. (28) Charles M. J.; Tondeur, Y.; Environ. Sci. Technol. 1990, 24, 1856. (29) McCurvin, D. M. A.; Clement, R. E.; Taguchi, V. Y.; Reiner, E. J.; Schellenberg, D. H.; Bobbie, B. A. Chemosphere 1989, 19, 205.
spectrometer.30 The current study utilizes an optimized resonant collision activated dissociation technique discussed by Plomley et al.31 with some additional modifications to improve sensitivity and specificity. The purpose of this study is to determine the practical limits to measuring 2,3,7,8-TCDD and related congeners in high-fat foods. Quadrupole ion storage tandem-in-time mass spectrometry (QISTMS) and HRMS were recently employed by the FDA to measure TCDD contamination in catfish and chicken eggs. QISTMS and HRMS results for all 17 2,3,7,8-chlorinesubstituted dibenzo-p-dioxins and dibenzofurans are compared for repeat analyses of catfish and chickens eggs. Interferences encountered using QISTMS and HRMS are discussed. MATERIALS AND METHODS Reagents. Methylene chloride, hexane, methanol, acetone, and toluene were from Burdick and Jackson (high-purity HPLC and pesticide residue grade solvents for GC/MS) and 95% n-hexane was from EM-Sciences. Tetradecane (98+%) and nonane (99+%) were from Aldrich Chemical Co. Silica gel 60 (70-230 mesh; Aldrich Chemical Co.) was column-extracted with methanol, followed by dichloromethane, and stored at 130 °C. Other reagents were as follows: acid silica gel, sulfuric acid (EM Sciences), 98% combined 40:60 (by weight) with silica gel 60, Woelm activity I neutral alumina (ICN chemicals) stored at 130 °C; anhydrous sodium sulfate reagent grade, 10-60 mesh (Fisher Scientific), baked 2 h in a muffle furnace at 550 °C and stored at 130 °C; purified (99.9995%) nitrogen gas UHP grade (MG industries) and chromatographic grade helium gas (99.9999%) (Air Products). The stock standard concentrates (200 ng/mL 12C native mixed standard 17 congeners EDF-7999 and 13C12 mixed standards 100 ng/mL EDF-8999) were stored in the dark at -10 °C, in 1-mL minivials. Aliquots of EDF-8999 and -7999 were combined in different amounts in nonane to provide a four-point calibration curve. The 13C12-labeled standards were maintained at 10 ng/mL in all calibration solutions. 13C12-labeled and native analytes are from Cambridge Isotope Laboratory, Woburn, MA. Safety. PCDD/Fs are highly toxic, notably causing chloracne, and are presumed to be carcinogenic and teratogenic. Products containing PCDD/Fs are likely to also have PCBs or other toxic compounds and should be handled with extreme care to avoid skin contact and inhalation of dust or aerosols. After skin contact with PCDD/Fs or other toxic substances, the affected area(s) should be washed at once with cold running water and soap for several minutes. Preparation of samples or chemicals emitting potentially toxic or infectious dust, aerosols, or vapors was conducted entirely in fume hoods. In this study, the quantities of PCDD/Fs used in analysis and fortification were below 500 pg international toxic equivalents (ITEQs). Sampling and Pool Fortification. Lamb fat, butter, peanut oil, and shortening were collected from supermarkets in the Washington, DC, area in 1997. Large quantities of each matrix were collected to ensure sufficient test portions for multiple determinations. Approximately 10-30-g portions of butter, cottonseed oil, and lamb fat were collected from the homes of persons from southern Kazakhstan with elevated breast milk 2,3,7,8-TCDD levels. (30) Hayward, D. G. Chemosphere 1997, 34, 929. (31) Plomley, J. B.; March, R. E.; Mercer, R. S. Anal. Chem. 1996, 68, 2345.
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
213
Washington, DC, area butter, shortening, and peanut oil were analyzed four to seven times before any fortifications were attempted. Subpools (100 g) of the peanut oil were fortified at 0.5 and 1.0 pg/g for PeCDD/Fs, hexachlorodibenzo-p-dioxins and -furans (HxCDD/Fs), and HpCDD/Fs. TCDD and TCDF were fortified at 0.1 and 0.2 pg/g. OCDD/F were fortified at 1.0 and 2.0 pg/g. Native standards dissolved in 1 mL of hexane were added to the subpool, mixed thoroughly, and allowed to stand overnight. Each fortified pool was then analyzed in triplicate. A fortified congener was judged to be at or above the LOQ, if the average measurement of two product ions using the internal labeled surrogate standard were within 20% of the theoretical fortification level with a 20% or less relative standard deviation (RSD) for repeat analyses. Sample Preparation. All test portions of Washington, DC, peanut oil, shortening, and butter were analyzed identically using 30-g portions. Test portions were fortified before extraction with 100 pg each of 14 13C12-labeled 2,3,7,8-substituted PCDD/Fs and 200 pg of [13C12]OCDD (labeled 1,2,3,7,8,9-HxCDD and OCDF were not added). Food samples from southern Kazakhstan, cottonseed oil, butter, and lamb fat, were prepared likewise, except the entire sample was used for analysis (10-30 g). Catfish were filleted, composited, and ground by the U.S. FDA Baltimore District Laboratory and returned immediately for 2,3,7,8-TCDD analysis. Egg yolks were separated from whole eggs and six yolks were pooled. A 25-g aliquot of yolk or catfish fillet was used for analysis. Four test portions were processed in parallel. The test portions were dissolved in 250 mL of 50:50 hexane/methylene chloride. The extract is filtered over a 2-cm layer of sodium sulfate, 30 g of activated silica gel 60 (70-230 mesh), a 2-cm layer of sodium sulfate, and then directly into a column of 50 mg of AX21 carbon30,32 on 500-mg glass fibers. The carbon column was eluted with 30 mL of dichloromethane and 40 mL of warm toluene (40 °C) in the reverse direction. The carbon column was regenerated for the next sample by washing with 50 mL of toluene, 50 mL of methanol, 50 mL of toluene, and 20 mL of dichloromethane. The purification was completed by passing the extract resuspended in n-hexane over acid/silica (1 g) and (0.5 g) neutral Woelm alumina. PCDD/Fs were recovered from the alumina with 2 mL of dichloromethane. Octachloronaphthalene (recently replaced with [13C12]1,2,3,7,8,9-HxCDD and [13C12]1,2,3,4-TCDD) and 2 µL of tetradecane were added, and the extract was concentrated and resuspended in 10 µL of nonane.30 Extraction and cleanup of egg yolk and catfish samples are described elsewhere.33 Note that the sample purification procedure used here is a standard approach for PCDD/Fs. An important aspect of the MS/ MS method was its ability to substitute as the determinant step with a standard cleanup. An alternate cleanup approach similar to EPA 1613 was tried with eggs and catfish but produced problems in GC resolution with the 40-m minibore column and sometimes high noise levels as well.34 This procedure was used to prepare the samples of eggs and catfish analyzed by HRMS (32) Smith, L. M.; Stalling, D. L.; Johnson, J. L. Anal. Chem. 1984, 56, 1830. (33) Hayward, D. G.; Nortrup, D.; Gardner, A.; Clower, M. Environ. Res., in press. (34) Gardner, A.; Andrzejewski, D.; Method for extraction, cleanup and determination of 17 2,3,7,8-substituted chlorodibenzo-p-dioxins/furans in fish. Laboratory Information Bulletin 3981; Office of Regional Operations, U.S. Food and Drug Administration, Rockville, MD, 1995.
214
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
for comparison with QISTMS. Improvements to the purification of biological samples, particularly the lipid removal step, have been described by other investigators.35-37 Alternative cleanups may prove useful to the optimization and overall performance of this MS/MS method. QISTMS. A Varian Saturn 4D ion trap gas chromatography/ mass spectrometer equipped with CI and MS/MS capability was used during all PCDD/F data acquisition. The ion trap manifold was held at 250 °C. The ion trap was connected by a heated (240 °C) transfer line to a Varian 3400 model gas chromatograph equipped with split/splitless injection and septum programmable on-column injector. A 40-m J&W DB-5 ms minibore capillary column 0.18-mm i.d. and 0.18-µm film thickness was connected to the septum programmable on-column injector. A DOS-based version 5.2 GC/MS software was supplemented with a windows based “toolkit” software that allowed multiple reaction monitoring. A 42-min temperature program was used to separate the 17 2,3,7,8substituted PCDD/Fs from non-2,3,7,8-substituted congeners (140 °C for 2 min, 20 °C/min to 200 °C, 5 °C/min to 240 °C for 12 min, hold, 10 °C/min to 280 °C for 10 min). This temperature program provided sufficient separation of 2,3,7,8-substituted congeners to permit the use of separate PCDD and PCDF acquisition segments (see Table 1). A four-point standard curve was determined using 2-µL injections from four mixed standards. Relative response factors were measured between fixed amounts of labeled product ions and the following amounts of unlabeled standard: 4, 20, 50, or 200 pg of PeCDD/Fs, HxCDD/Fs, and HpCDD/Fs; 0.8, 4, 10, or 40 pg of TCDD/F; and 8, 40, 100, or 400 pg of OCDD/F. Relative response factors were calculated for each product ion relative to a single product ion for the corresponding [13C12]-2,3,7,8-substituted congener. The RSD calculated for a product ion across the four concentrations was usually 20% or less (Table 2). The grand mean RSD for 40 product ions in the four point calibration was 11%. A modification of the multiple frequency irradiation conditions described by Plomley31 was used to produce product ion spectra.30 Table 1 lists the collision-induced dissociation (CID) bandwidths used for each PCDDs or PCDFs along with the excitation amplitudes and radio frequency (rf) storage m/z (qz ) 0.4). The qz ) 4 eV/mr2Ω2, where m is the ion mass, e is the charge, V is the amplitude of the trapping rf and Ω is its frequency, and r is the ring electrode radius. The qz was set to provide maximum ion stability during CID. Emission current was set at 100 µA, and the electron multiplier was set 150 V above 105 gain. During the quantitation limit studies, five precursor ions were monitored simultaneously for pentachlorinated through heptachlorinated congeners (e.g., PeCDD and PeCDF; m/z 356, 368, 338, 340, 352). PCDDs (native and labeled) were monitored using two of the five precursor ions in a segment while the other three were used to measure the homologous PCDF. These precursor ion groupings allowed faster GC runs (30 min) and therefore were more convenient during method development. TCDD and TCDF were monitored in separate time segments using three precursor ions for each congener during fortification (35) Liem, A. K. D.; de Long, A. P. J. M.; Marsman, J. A.; den Boer, A. C.; Groenemeijer, G. S.; den Hartog, R. S.; de Korte, G. A. L.; Hoogerbrugge, R.; Kootstra, P. R.; van’t Klooster, H. A. Chemosphere 1990, 20, 843. (36) van Bavel, B. Ph.D. Thesis, University of Umea, Umea, Sweden, 1995. (37) Strandberg, B.; Bergqvist, P.-A.; Rappe, C. Anal. Chem. 1998, 70, 526.
Table 1. QISTMS Conditions Using Resonant Multiple Frequency Irradiationa segment (no.)/congener (1) TCDF (2) TCDD (3) PeCDF (4) PeCDD (5) HxCDF (6) HxCDD (7) HxCDF (8) HpCDF (9) HpCDD (10) HpCDF (11) OCDF OCDD
precursor ion (m/z)
storage (m/z)
excitation amplitude (V)
CID bandwidth (kHz)
304 306 318 320 322 334 338 340 352 354 356 368 372 374 386 388 390 402 372 374 386 408 410 422 424 426 436 408 410 422 442 444 458 460 472
134 135 140 141 142 148 149 150 156 156 157 163 164 165 171 171 172 178 164 165 167 180 181 187 187 188 193 180 181 187 195 196 203 203 209
1.3 1.3 1.3 1 1 1 1.6 1.6 1.6 1.2 1.2 1.2 1.5 1.5 1.5 1.3 1.3 1.3 1.5 1.5 1.5 1.6 1.6 1.6 1.3 1.3 1.3 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6
2 2 1 2 2 1 2 2 2 2 2 2 2 2 2 1 1 1 2 2 2 1 1 1 2 2 2 1 1 1 2 2 1 1 1
a 2,3,7,8-Substituted PCDD and PCDF acquistion segments in elution order on a DB-5 ms capillary column. CID time, 30 ms; mass isolation window, 3 m/z.
studies with peanut oil and Washington, DC, butter analyses. Thereafter, chlorodibenzo-p-dioxins were monitored separately whenever possible from chlorodibenzofurans in all subsequent food analyses (Table 1). OCDD and OCDF do not separate sufficiently and therefore were always acquired in the same segment. HRMS. A VG-Autospec Q high-resolution mass spectrometer EBEQQ was operated at a minimum of 10 000 resolution (10% valley definition) acquiring selective molecular and fragment ions for all PCDD/Fs in five time segments using voltage scans. Perfluorokerosene was used for lock mass mode MID. Total scan time was ∼0.8 s including dwell, delay, and reset times. Two ions from the molecular ion cluster (M+ and (M + 2)+ typically for TCDDs and TCDFs) monitored along with their corresponding M+ - COCl fragment ions. Extracts manually injected, splitless mode (2 µL of 5 µL) onto a 60 M DB-5 ms 0.25-mm-i.d. capillary column. Confirmation of a PCDD/F congener required at least 3:1 S/N ratio on all monitored ions and area ratios for molecular and fragment ions for the native analyte must be within 10% of the value from authentic standards. The values obtained for labeled standards must also be within 10% of authentic standards.
Table 2. Mean Relative Response Factors (RRFs) and RSDs for QISTMS Using Resonant Multiple Frequency Irradiationa unlabeled std amt (pg)
congener TCDD/F PeCDD/F-HpCDD/F OCDD/F
2,3,7,8-TCDF 2,3,7,8-TCDF 2,3,7,8-TCDD 2,3,7,8-TCDD 2,3,7,8-TCDD 1,2,3,7,8-PeCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,7,8-PeCDD 1,2,3,7,8-PeCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDFb 2,3,4,6,7,8-HxCDFb 1,2,3,7,8,9-HxCDFb 1,2,3,7,8,9-HxCDFb 1,2,3,4,7,8-HxCDD 1,2,3,4,7,8-HxCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,6,7,8-HpCDD 1,2,3,4,6,7,8-HpCDD 1,2,3,4,7,8,9-HpCDF 1,2,3,4,7,8,9-HpCDF OCDD OCDD OCDD OCDF OCDF
0.8 4 8
4 20 40
10 50 100
40 200 400
m/z
mean
RSD (%)
241 243 257 259 287 275 277 275 277 291 293 321 309 311 309 311 309 311 309 311 327 329 264 327 329 264 327 329 264 345 347 361 363 345 347 395 397 332 379 381
1.5 1.0 1.3 1.3 0.25 0.98 0.95 0.97 1.0 1.3 1.0 0.34 1.0 0.93 1.0 0.9 0.73 0.87 0.85 1.3 1.0 0.57 0.48 0.93 0.6 0.48 1.0 0.57 0.56 0.49 0.63 1.4 0.78 0.51 0.63 1.4 1.1 0.81 0.75 0.46
3 5 9 19 10 3 9 11 5 19 8 9 21 2 22 3 7 10 9 12 5 12 19 18 9 13 12 12 12 16 8 12 5 6 8 16 16 10 28 4
a Grand mean RSD 11%. b RRFs measured on 4/97, all other RRFs measured on 9/98.
Ion profiles for native PCDD/Fs and labeled standards must maximize simultaneously ((2 s). Quality Assurance. Daily calibration standards were used to verify sensitivity, retention times, and product ion relative intensities. Daily calibration results for the product ions (loss of COCl) were expected to be within at least 30% of the initial standard curve before further analysis proceeded. Empirically derived quantitation limits were used to estimate minimum quantification levels. During the food analyses, blanks were almost always free of quantifiable levels for most congeners. However, OCDD was always found in blanks. Blank background was not subtracted, but results were flagged (L) if the total picogram amount of analyte in a sample was not at least 3 times greater than the total picogram amount of that analyte in the blank analyzed along with the sample. Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
215
Table 3. PCDD and PCDF Mean Levels and RSDs for Repeat Analyses of Unfortified High-Fat Foods from Washington, DC (pg/g; 30-g Test Portions)a peanut oil (n ) 7)
shortening (n ) 6)
butter (n ) 4)
matrix congener
mean
RSD (%)
mean
RSD (%)
mean
RSD (%)
2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD
0.04* 0.23* 0.11* 0.09* 0.11* 1.0 L 32 L
30 32 53 51 68 60 72
0.04* 0.18* 0.20* 0.13* 0.14* 0.93 L 25 L
41 61 17 19 17 40 58
0.08 0.35 I 0.61 1.3 0.40 2.5 5
44 17 24 22 19 7 14
2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF
0.04* 0.05* 0.04* 0.09* 0.07* 0.07* 0.14* 0.83 L 0.12* 6L
35 57 73 36 35 63 59 88 70 112
0.06* 0.07* 0.07* 0.11* 0.09* 0.22* 0.14* 0.51 L 0.11* 3L
43 71 47 55 59 40 58 55 84 67
0.09 I 0.08* 0.30 0.46 0.27 0.15* 0.28 0.89 0.18 L 0.81 L
51 44 14 30 19 nd 19 9 54 15
a Key: *, analyte not identified; values are the mean and RSD (%) of the chemical noise relative to the 13C12-labeled standard response. L, upper limit, level less than 3 times the blank level; I, interference, daughter ion ratios incorrect; nd, not determined.
Identification of 2,3,7,8-TCDD by QISTMS required at least two product ions that maximized at the correct retention time for [13C12]2,3,7,8-TCDD ((2 s). Product ion responses were measured against the [13C12]2,3,7,8-TCDD product ion response observed in each sample and are automatically corrected for the labeled standard recovery during sample preparation. Quantitation was based on an average result from at least two product ions. The values obtained from the product ions must agree to within 30%. Theoretical chlorine isotope ratios were not used as part of the identification criteria. Often the chlorine isotope ratios were the same as theoretical, but frequently they were not. Product ion intensities were not adjusted to conform to theoretical isotope ratios. Identification was based on producing a product ion spectrum that closely resembled the spectra from authentic standards. Qualitative supporting information such as the presence of other product ions and the absence of interferences and excessive noise was also utilized. Adequate labeled surrogate recoveries required a minimum of 40% and not greater than 120%. One or two other product ions usually (M + 2)+ - Cl, (M)+, or (M + 2)+ - 2COCl)) were useful near or at the limit of quantitation for all dioxins (except HpCDD). One of these ions was used for quantitative and qualitative verification of PCDD congeners. Product ion intensities, other than loss of COCl, for chlorodibenzofurans were always too weak to be useful near the desired LOQs (Table 2). RESULTS AND DISCUSSION PCDD/F LOQs for QISTMS. High-fat food pools were analyzed repetitively to assess the matrix/laboratory background and its variability (Table 3). No congeners were measured in peanut oil or shortening that were not already present in laboratory reagent blanks (OCDD, OCDF, 1,2,3,4,6,7,8-HpCDD, 1,2,3,4,6,7,8216 Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
HpCDF). As the study proceeded, blank background levels fell to the point where only OCDD was detected in blanks (
