Anal. Chem. 2010, 82, 9835–9842
Gas Chromatography/Electron Ionization-Mass Spectrometry-Selected Ion Monitoring Screening Method for a Thorough Investigation of Polyhalogenated Compounds in Passive Sampler Extracts with Quadrupole Systems Natalie Rosenfelder,† Nathan J. Van Zee,†,‡ Jochen F. Mueller,§ Caroline Gaus,§ and Walter Vetter*,†,§ Institute of Food Chemistry (170b), University of Hohenheim, Garbenstrasse 28, D-70599 Stuttgart, Germany, Baker Laboratory, Cornell University, Ithaca, New York 14853-1301, United States, and National Research Centre for Environmental Toxicology (EnTox), The University of Queensland, 39 Kessels Road, Coopers Plains 4108, Australia Nontarget analysis and identification of unknown polyhalogenated compounds is important in acquiring a thorough picture of the present pollution status as well as for identifying emerging environmental problems. Such analyses usually require the application of electron ionization mass spectrometry because the resulting mass spectra frequently allow for compound identification. When quadrupoles are used as mass separators, the full scan technique often suffers from low sensitivity along with nonspecificity for polyhalogenated trace compounds which often result in interference by matrix compounds. We have developed a novel nontarget gas chromatography/electron ionization-mass spectrometry-selected ion monitoring (GC/ EI-MS-SIM) method that overcomes these sensitivity and selectivity issues. Our method is based on the fact that the molecular ions and isotope patterns of polyhalogenated compounds involve the most relevant primary information with regard to the structure of polyhalogenated compounds. Additionally, the retention times of polyhalogenated compounds generally increase with increasing molecular weight. The retention time range of polyhalogenated compounds was divided in three partly overlapping segments of 112 u (segment A: m/z 300-412; segment B: m/z 350-462; segment C: m/z 450-562) that were screened in eight GC runs consisting of 15 consecutive SIM ions. This method was tested with a passive water sampler extract known to contain over 30 polyhalogenated compounds according to the sensitive analysis by GC/electron capture negative ion (ECNI)-MS. While none of these polyhalogenated compounds could be detected by GC/EI-MS in full scan mode, our nontarget GC/EI-MS-SIM method allowed for the detection of 38 polyhalogenated compounds. Only seven could be identified by means of reference standards while more than 15 * Corresponding author. E-mail:
[email protected]. Phone: +49 711 45924016. Fax: +49 711 45924377. † University of Hohenheim. ‡ Cornell University. § The University of Queensland. 10.1021/ac102134x 2010 American Chemical Society Published on Web 11/04/2010
of the unknowns could be traced back to at least the class of compounds based on the mass spectrometric data from the nontarget SIM runs. All compounds identified originated from halogenated natural products. The nontarget GC/EI-MS-SIM method combines the high sensitivity obtainable with quadrupole systems for trace analysis with the structural information essential for the identification of unknown pollutants. Polyhalogenated organic compounds are widely distributed in the environment, and their environmental properties cause them to be classified as persistent organic pollutants (POPs). Their analysis in different matrices is, thus, an important task in environmental and food chemistry. Polyhalogenated compounds are composed of different classes of anthropogenic and naturally produced polybrominated and polychlorinated compounds.1,2 While a range of methods have been developed for the quantitation of known organohalogens, the detection and identification of “unknowns” is often difficult due to the lack of reference standards and limited information available from standard gas chromatography coupled to mass spectrometry (GC/MS) measurements. Routine analysis of polyhalogenated compounds can be performed in electron ionization (GC/electron ionization (EI)MS) as well as in electron capture negative ion (GC/electron capture negative ion (ECNI)-MS) mode, both of which are usually operated in the selected ion monitoring (SIM) mode.3 GC/ECNIMS is usually more sensitive and enables detection of virtually all polybrominated compounds by the screening of m/z 79 and m/z 81, i.e., the characteristic bromide ion isotopes.4,5 However, GC/ECNI-MS full scan analysis of polybrominated compounds often suffers from low abundant or even nondetectable molecular (1) Vetter, W. Rev. Environ. Contam. Toxicol. 2006, 188, 1–57. (2) Covaci, A.; Voorspoels, S.; Ramos, L.; Neels, H.; Blust, R. J. Chromatogr., A 2007, 1153, 145–171. (3) Ong, V. S.; Hites, R. A. Mass Spectrom. Rev. 1994, 13, 259–283. (4) Buser, H.-R. Anal. Chem. 1986, 58, 2913–2919. (5) de Boer, J.; de Boer, K.; Boon, J. P. Handb. Environ. Chem. 2000, 3, 61– 95.
Analytical Chemistry, Vol. 82, No. 23, December 1, 2010
9835
ions and other fragment ions in the high-mass range.6 Thus, the identification of organobromines detected by means of their bromide ion isotopes in SIM mode is difficult without authentic reference standards. In contrast, GC/EI-MS usually provides fragment ions in the high-mass range and consequently more structural information. However, the lower sensitivity of GC/EIMS in full scan mode compared to GC/ECNI-MS does not allow for the detection of POPs in the low picogram range, especially when widely distributed quadrupole instruments are used. Furthermore, background noise caused by sample matrix residues mostly in the low-mass range is detected as well and may affect the quality of the chromatograms and mass spectra. To achieve the required low limits of detection for POPs, quadrupole instruments must be operated in the SIM mode. Such GC/EI-MS-SIM methods are designed to detect carefully selected POPs (targeted analysis) while unknown compounds that do not form the ions selectively monitored are overlooked. Recently developed techniques such as GCxGC time-of-flight (TOF)-MS may be used to solve such problems.7 However, these instruments are very expensive and not widely distributed while quadrupole GC/MS systems are virtually found in all laboratories focusing on pollutants. Hence, we wished to develop a method for quadrupole instruments that can be used for the thorough screening of samples. To compensate for the low sensitivity of quadrupole systems operated in scan mode, we developed a novel nontarget GC/EIMS-SIM method for the structural investigation of low concentrated polyhalogenated compounds in environmental samples. This method was based on the fact that the masses of polyhalogenated compounds correlate with the retention times on nonpolar GC columns. For this reason, time windows with a small mass range were established and screened in subsequent SIM runs. This nontarget screening method was tested with passive water sampler extracts known to contain a range of known halogenated natural products (HNPs) and different unknown organohalogens.8 MATERIALS AND METHODS Gas Chromatography Coupled to Electron Ionization Mass Spectrometry (GC/EI-MS). An HP 5890 Series II Plus GC coupled with an HP 5972 mass selective detector (MSD) was used in combination with an HP-5 ms column (30 m × 0.25 mm i.d. × 0.25 µm film thickness). The GC oven was programmed as follows: after 2 min isothermal at 60 °C, the temperature was raised at 10 °C/min to 300 °C and held for 14 min. Injections (1 µL) were performed in splitless mode at 250 °C. Helium 5.0 (Sauerstoffwerke, Friedrichshafen, Germany) was used as the carrier gas. In the full scan mode, either m/z 200-700 or m/z 300-600 was measured throughout the run. Alternatively, we scanned the following ranges in three time windows: m/z 384-398 (segment A), m/z 434-448 (segment B), and m/z 534-548 (segment C). In SIM mode, we measured 15 consecutive masses per run (30 (6) Sellstro ¨m, U. Determination of some polybrominated flame retardants in biota, sediment and sewage sludge. Ph.D. thesis, Department of Environmental Chemistry and Institute of Applied Environmental Research (ITM) Laboratory for Analytical Environmental Chemistry, Stockholm University, Stockholm, Sweden, 1999. (7) Hilton, D. C.; Jones, R. S.; Sjoedin, A. J. Chromatogr., A 2010, 1217, 6851– 6856. (8) Vetter, W.; Haase-Aschoff, P.; Rosenfelder, N.; Komarova, T.; Mueller, J. F. Environ. Sci. Technol. 2009, 43, 6131–6137.
9836
Analytical Chemistry, Vol. 82, No. 23, December 1, 2010
Table 1. Setup of the Nontarget GC/EI-MS SIM Method for the Determination of Polyhalogenated Compounds in Environmental Samples GC run
segment A (10.0-17.5 min) (m/z)
segment B (17.5-22.5 min) (m/z)
segment C (22.5-40.0 min) (m/z)
1 2 3 4 5 6 7 8 Σa
300-314 (A1) 314-328 (A2) 328-342 (A3) 342-356 (A4) 356-370 (A5) 370-384 (A6) 384-398 (A7) 398-412 (A8) 300-412
350-364 (B1) 364-378 (B2) 378-392 (B3) 392-406 (B4) 406-420 (B5) 420-434 (B6) 434-448 (B7) 448-462 (B8) 350-462
450-464 (C1) 464-478 (C2) 478-492 (C3) 492-506 (C4) 506-520 (C5) 520-534 (C6) 534-548 (C7) 548-562 (C8) 450-562
a Sum of the masses determined in SIM mode (and initially also in the full scan mode).
ms dwell time), respectively, in eight GC runs (Table 1). Additional SIM runs were programmed to further investigate the molecular ions of specific compounds, e.g., 5.0-12.0 min (m/z 230, 232, 274, 276, 309, 311, 344, 346, 388, 390), 12.0-14.5 min (m/z 323, 324, 325, 327, 328, 329, 331, 333, 335, 337), 14.5-15.4 min (m/z 323, 325, 327, 329, 331, 386, 388, 390, 392, 394), 15.4-16.2 min (m/z 348, 350, 352, 354, 356, 357, 358, 360, 361, 362), 16.2-17.0 min (m/z 323, 325, 326, 327, 328, 329, 331, 333, 335, 337), 17.0-19.5 min (m/z 426, 428, 430, 431, 432, 433, 434, 436, 438, 440), 19.5-20.5 min (m/z 416, 418, 420, 422, 424, 426, 472, 474, 476, 478), 20.5-21.7 min (m/z 447, 448, 449, 450, 451, 452, 453, 454, 456, 457), 21.7-24.5 min (m/z 444, 446, 448, 450, 451, 452, 454, 456, 458, 460), and 24.5-40.0 min (m/z 481, 483, 485, 487, 518, 520, 522, 562, 564, 566). Gas Chromatography Coupled to Electron-Capture Negative Ion Mass Spectrometry (GC/ECNI-MS). GC/ECNI-MS analyses were carried out with a Varian CP-3800/1200 system (Darmstadt, Germany). The electron energy was set to 70 eV, and the ion source temperature was set to 150 °C. Nitrogen (7 Torr, purity 5.0; Sauerstoffwerke Friedrichshafen, Germany) was used as the reagent gas.9 The transfer line temperature was set to 280 °C. The same type of GC column and oven program as for GC/EI-MS analyses was used in this instrument. In full scan mode, the mass range m/z 30-800 was measured. Standards and Samples. 2,3,3′,4,4′,5,5′-Heptachloro-1′-methyl1,2′-bipyrrole (Q1) was synthesized according to Wu et al.10 while 4,6-dibromo-2-(2′,4′-dibromo)phenoxyanisole (2′-MeO-BDE 68, BC2) was synthesized according to Vetter and Jun.11 2,2′-Dimethoxy3,3′,5,5′-tetrabromobiphenyl (2,2′-diMeO-BB 80, BC-1), 3,5-dibromo2-(2′,4′-dibromo)phenoxyanisole (6-MeO-BDE 47, BC-3), and 3,5dibromo-2-(3′,5′-dibromo,2′-methoxy)phenoxyanisole (2′,6-diMeOBDE 68, BC-11) were synthesized according to Marsh et al.12,13 5,5′-Dichloro-1,1′-dimethyl-3,3′,4,4′-tetrabromo-2,2′-bipyrrole (DBPBr4Cl2, BC-10) was synthesized according to Gribble et al.14 2,4,6-Tribromoanisole (TBA) was purchased from Aldrich (9) Rosenfelder, N.; Vetter, W. Rapid Commun. Mass Spectrom. 2009, 23, 3807– 3812. (10) Wu, J.; Vetter, W.; Gribble, G. W.; Schneekloth, J. S. J.; Blank, D. H.; Go ¨rls, H. Angew. Chem., Int. Ed. 2002, 41, 1740–1743. (11) Vetter, W.; Jun, W. Chemosphere 2003, 52, 423–431. (12) Marsh, G.; Athanasiadou, M.; Athanassiadis, I.; Bergman, A.; Endo, T.; Haraguchi, K. Environ. Sci. Technol. 2005, 39, 8684–8690. (13) Marsh, G.; Stenutz, R.; Bergman, A. Eur. J. Org. Chem. 2003, 2566–2576.
(Milwaukee, USA), and 2,4,6-tribromophenol (TBP) was from Sigma-Aldrich (Steinheim, Germany). The sources of further reference standards are reported elsewhere.15 The standard mixture for method development contained the following in elution order on a DB-5 like HP-5 column: TBA, TBP, allyl2,4,6-tribromophenyl ether (ATE), 2-bromoallyl-2,4,6-tribromophenyl ether (BATE), (1R,2S,4R,5R,1′E)-2-bromo-1-bromomethyl-1,4-dichloro-5-(2′-chloroethenyl)-5-methylcyclohexane(MHC1), Q1, 2,3-dibromopropyl-2,4,6-tribromophenyl ether (DPTE), 2′-MeO-BDE 68 (BC-2), 2,2′-diMeO-BB 80 (BC-1), DBP-Br4Cl2 (BC-10), 6-MeO-BDE 47 (BC-3), DBP-Br4Cl2 (BC-10), 2′,6diMeO-BDE 68 (BC-11), 2,7-dibromo-4a-bromo-methyl-1,1-dimethyl-2,3,4,4a,9,9a-hexahydro-1H-xanthene (triBHD), and 2,5,7tribromo-4a-bromomethyl-1,1-dimethyl-2,3,4,4a,9,9a-hexahydro1H-xanthene (tetraBHD). Passive Water Sampler Extracts. Experimental details on locations, deployment, passive water sampling method, and sample cleanup are reported in Vetter et al.8 In short, semipermeable membrane devices (lay-flat, low density polyethylene tubes filled with triolein) were deployed at a range of sites along the Great Barrier Reef at a depth of about 1 m for 1 to 2 months. Analysis consisted of accelerated solvent extraction and gel-permeation chromatography.8,16 Five replicate samples collected near Low Island at five time periods (December 2006, February 2007, April 2007, July 2007, and September 2007) were purified and pooled for analysis. RESULTS AND DISCUSSION The GC/ECNI-MS chromatogram of a purified passive water sampler extract from the Great Barrier Reef, Australia, showed over 30 peaks originating from polybrominated compounds as verified by the presence of m/z 79 and m/z 81 in the characteristic ratio of the bromide ion isotopes (Figure 1a). The retention times of the vast majority of the detected compounds were not documented in our recently established database of over 120 organobromine compounds.15 Additionally, analysis in the GC/ ECNI-MS full scan mode added no further information to the structure of the unknowns because the mass spectra showed only low abundant or interfered ions in the high mass range. For this reason, we switched to GC/EI-MS measurements for compound identification. However, the corresponding GC/EI-MS full scan chromatogram (m/z 200-700) was overloaded by a range of nonhalogenated compounds in the time range of 18-22 min which gave no response in the more selective GC/ECNI-MS mode (Figure 1a,b). Moreover, GC/EI-MS measurements in the full scan mode did not provide the required sensitivity for the detection of any polyhalogenated compound including the abundant halogenated natural products known to be present in the sample according to GC/ECNI-MS (Figure 1b). Narrowing the mass range to m/z 300-600 partly eliminated the background noise problem, which hence was mainly caused by low-mass matrix residues (Figure S-1a, Supporting Information). Still, the sensitivity was insufficient for a direct identification of the polyhalogenated compounds. (14) Gribble, G. W.; Blank, D. H.; Jasinski, J. P. Chem. Commun. 1999, 2195– 2196. (15) Vetter, W.; Rosenfelder, N. Anal. Bioanal. Chem. 2008, 392, 489–504. (16) Komarova, T. V.; Bartkow, M. E.; Rutishauser, S.; Carter, S.; Mueller, J. F. Environ. Pollut. 2009, 157, 731–736.
Figure 1. GC/ECNI-MS (a) and GC/EI-MS (b,c) chromatograms of a passive water sampler extract in (a) full scan mode (m/z 79 extracted), (b) full scan mode, m/z 200-700, and (c) nontarget SIM mode, m/z 384-398 (segment A), m/z 434-448 (segment B), and m/z 534-548 (segment C) (*: compounds only detected by GC/ECNIMS; O: nonhalogenated compounds, differences in retention time ranges are due to different instruments used for ECNI- and EI-MS.)
Figure 2. Retention times of different polyhalogenated compounds, divided into three time intervals (segments A-C) and mass ranges (shaded): segment A: 10-17.5 min (m/z 300-412), segment B: 17.5-22.5 min (m/z 350-462), and segment C: 22.5-40 min (m/z 450-562).
In order to obtain the required sensitivity in GC/EI-MS, we focused on the fact that the most important structural information of polyhalogenated compounds is derived from the mass of the molecular ion along with its characteristic isotope pattern. Noteworthy, polyhalogenated compounds elute in order with increasing molecular weights from DB-5 columns (Figure 2). Thus, we split the retention time range into three segments of increasing mass ranges of 112 u each (m/z 300-412 (segment A), m/z 350-462 (segment B), and m/z 450-562 (segment C)). Due to the wide distribution of isomers, the selected mass ranges were adjusted with some overlap on the basis of the following observations: Under the GC conditions used (see Materials and Methods), the known organobromine compounds eluting prior to 17.5 min (segment A, Figure 2) had molecular weights of 450 u (pentaBDEs (M+, m/z 560), tri- and tetrabrominated hexahydroxanthene derivatives (PBHDs, M+, m/z 464 and m/z 542)17). The retention time range between 17.5 and 22.5 min (segment B) featured compounds like hexachlorinated biphenyls (m/z 358), BATE (M+, m/z 446), the halogenated natural products Q1 (M+, m/z 384), and MHC-1 (M+, m/z 396), as well as tetraBDEs (M+, m/z 482).18-20 The proposed range of 112 u (m/z 350-462) covered all compounds except tetraBDEs and DPTE (Figure 2). However, since these compounds also form [M - Br]+ fragment ions (abundance 3-50% relative to M+), we suspected that they would still be detected in this segment (see below for examples). GC/EI full scan screening of the three segments spanning over 112 u, respectively, demonstrated that the matrix background from aliphatic compounds was practically eliminated because these artifacts predominantly formed masses
