Analysis of Nitro Musk Compounds and Their ... - ACS Publications

A method using HRGC ion trap MS/MS for measuring simultaneously amino metabolites and the parent compounds, the nitro musks, an important group of ...
0 downloads 0 Views 100KB Size
Download PDF
Anal. Chem. 2000, 72, 2124-2131

Analysis of Nitro Musk Compounds and Their Amino Metabolites in Liquid Sewage Sludges Using NMR and Mass Spectrometry Jean-Daniel Berset*

Department of Soil Protection and Ecotoxicology, Institute of Environmental Protection and Agriculture (IUL), Schwarzenburgstrasse 155, CH-3005 Berne, Switzerland Peter Bigler

Department of Chemistry and Biochemistry, NMR Spectroscopy Group, Freiestrasse 3, CH-3012 Berne, Switzerland Daniel Herren

Interlabor Belp AG, Birkenweg 6, CH-3123 Belp, Switzerland

A method using HRGC ion trap MS/MS for measuring simultaneously amino metabolites and the parent compounds, the nitro musks, an important group of organic fragrance components found in sewage sludges, was developed. The monoamino metabolites were synthesized and characterized by 1H/13C NMR and mass spectroscopy. Among the nitro musks, musk ketone was the major compound, found at an average concentration of 5 µg/kg of dry mass (dm) whereas musk xylene was detected in only one sample (30 µg/kg dm). Three amino metabolites were identified, namely, 1-tert-butyl-3,5-dimethyl-4amino-2,6-dinitrobenzene, 1,1,3,3,5-pentamethyl-4-nitro6-aminoindane, and 4-acetyl-1-tert-butyl-3,5-dimethyl2-nitro-6-aminobenzene, the corresponding reduction products of the nitro musks xylene, moskene, and ketone. These metabolites were present in partly higher concentrations in the sludges than the corresponding nitro musk compounds. Musk xylene and musk moskene were mainly found as their monoamino metabolites, underlining the importance of anaerobic reduction processes in the sewage treatment plant. Sewage sludge application to agricultural land represents an economical way to use the high amounts of sludge produced by the wastewater treatment plants. Currently, ∼2.1 × 105 tons (dry matter) of sewage sludge are produced annually in Switzerland of which 1.07 × 105 tons (51%) are applied to agriculture.1 To monitor its quality, limit values for heavy metals have been established.2 These inorganic compounds are analyzed on a routine basis. However, the characterization and long-term observation of organic contaminants in sludge has received little * Corresponding author: (tel) +4131 323-83-98; (e-mail) [email protected]. (1) Kettler, R. Umwelt-Materialien 90: Abfall-Statistik 1996; 1998; pp 6-149. (2) Swiss Federal Government. Verordnung u ¨ ber umweltgefa¨hrdende Stoffe (Stoffverordnung StoV); 1996; pp 61-62.

2124 Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

attention so far due to the absence of such values. It is well known that sewage sludge contains many xenobiotic, organic chemicals which might have a negative impact on soil organisms and fertility.3-8 Among these, nonpolar, highly lipophilic compounds such as polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), polychlorinated dioxins (PCDDs), and dibenzofurans (PCDFs) are important contaminants as they persist for a long time in the different environmental compartments. Quite recently, two other groups of compounds exhibiting chemical and physical properties similar to those mentioned above have been described to occur as ubiquitous components: nitro musk and polycyclic musk substances. The nitro musks include a group of five synthetic alkylated nitrobenzenes with a typical musk odor: musk moskene (MM), musk tibetene (MT), musk xylene (MUXY), musk ketone (MK) and musk ambrette (MA). The polycyclic musks are chemically alkylated tetralin or indane systems. Main representatives of the polycyclic musks are galaxolide (HHCB), tonalide (AHTN), celestolide (ADBI), phantolide (AHMI), and traseolide (AITI). These compounds are widely used as fragrance ingredients, in washing and cleaning agents, air fresheners, shampoos, perfumes, and other cosmetic products, as food additives in fish baits and also in cigarettes. They have been detected in fish, surface and sewage waters,9-12 human milk and (3) Tarradellas, J.; Diercxsens, P. Schweiz. Arch. Tierheilkd. 1983, 125, 589605. (4) Drescher, U.; Matthies, M.; Bru ¨ ggemann, R. GWF 1989, 130, 613-619. (5) Demirjian, Y. A.; Joshi, A. M.; Westman, T. R. J. WPCF 1987, 59, 32-37. (6) Klo ¨pffer, W. Chemosphere 1996, 33, 1067-1081. (7) Lega, R.; Ladwig, G.; Meresz, O.; Clement, R. E.; Crawford, G.; Salemi, R.; Jones, Y. Chemosphere 1997, 34, 1705-1712. (8) Halling-Sorensen, B.; Nors Nielsen, S.; Lanzky, P. F.; Ingerslev, F.; Holten Lu ¨ tzhoft, H. C.; Jorgensen, S. E. Chemosphere 1998, 36, 357-393. (9) Eschke, H. D.; Traud, J.; Dibowski, H. J. Z. Umweltchem. Oekotox. 1994, 6, 183-189. (10) Eschke, H. D.; Traud, J.; Dibowski, H. J. Z. Umweltchem. Oekotox. 1995, 7, 131-138. (11) Eschke, H. D.; Traud, J.; Dibowski, H. J. Vom Wasser 1994, 83, 373-383. 10.1021/ac991006d CCC: $19.00

© 2000 American Chemical Society Published on Web 03/24/2000

in a few sewage sludge samples, but no efforts have been undertaken to look for nitro musk amino metabolites.24 Due to the importance of sludge use in Swiss agriculture, the aim of this study was to determine amino metabolites of the nitro musk compounds in sludges from different catchment areas and to compare their presence with the parent compounds, the nitro musks. The amino metabolites were prepared synthetically and fully characterized by 1H and 13C NMR spectroscopy. NMR has proved to be a valuable tool for structure elucidation. In this contribution, established powerful 1D and 2D NMR experiments have been applied to prove the degree and the site of hydrogenation for a series of nitro musk compounds and to spectroscopically characterize the corresponding amino products. The proven product structures and the unequivocal assignment of the corresponding 1H and 13C signals allowed us to comparatively study and interpret the structural influence on these spectroscopic parameters. Finally, mass spectral data parameters are provided for sensitive detection of these metabolic substances.

Figure 1. Musk xylene (MUXY) and its primary and secondary metabolites.

adipose tissue,13-15and sediment samples16,17 as well as in the North Sea.18 The lack of biological and chemical degradation of these compounds as well as their bioaccumulation potential has raised considerable attention in the field of environmental chemistry.19 Very recently, metabolites of MUXY and MK have been detected in influent and effluent samples of a sewage treatment plant (STP) in Germany.20 These monoamino transformation products have been shown to occur also as main metabolites in mammals during metabolism21 (Figure 1). From a toxicological point of view, these metabolites might well be more toxic than their parent compounds as it is known that aromatic amines, depending on their structure, might act as carcinogenic agents.21-23 In addition, diamino and dihydroxy derivatives have been observed as minor products21 (Figure 1). These compounds have, however, not been found in samples of a STP. To our knowledge, very little is known on the contamination of musk compounds in sewage sludges. A study from The Netherlands has analyzed nitro- and polycyclic musk fragrances (12) Rimkus, G.; Wolf, M. Dtsch. Lebensm.-Rundsch. 1993, 89, 171-177. (13) Rimkus, G.; Wolf, M. Chemosphere 1996, 33, 2033-2043. (14) Rimkus, G.; Wolf, M. Dtsch Lebensm.-Rundsch. 1993, 89, 103-107. (15) Ott, M.; Failing, K.; Lang, U.; Schubring, Ch.; Gent, H. J.; Georgii, S.; Brunn, H. Chemosphere 1999, 38, 13-32. (16) Lach, G.; Steffen, D. Orientierende Untersuchungen von Gewa ¨ ssersedimenten auf Nitro-/Polymoschusverbindungen und die Flammenschutzmittel TCEP und TCPP, Niedersa ¨ chsiches Landesamt fu ¨ r Oekologie; Hildesheim, Germany, 1997; Chapter 3, pp 5-13. (17) Winkler, M.; Kopf, G.; Hauptvogel, G.; Neu, T. Chemosphere 1998, 37, 11391156. (18) Gatermann, R.; Hu ¨hnerfuss, H.; Rimkus, G.; Wolf, M.; Franke, St. Mar. Pollut. Bull. 1995, 30, 221-227. (19) Suter-Eichenberger, R.; Altorfer, H.; Lichtensteiger, W.; Schlumpf, M. Chemosphere 1998, 36, 2747-2762. (20) Gatermann, R.; Hu ¨ hnerfuss, H.; Rimkus, G.; Attar, A.; Kettrup, A. Chemosphere 1998, 36, 2535-2547. (21) Ka¨fferlein, H. U.; Go ¨en, T. Angerer, Crit. Rev. Toxicol. 1998, 28, 431-476. (22) Lutz, W. K. Arch. Toxicol. 1984, (Suppl. 7), 194-207. (23) Eisenbrand, G.; Metzler, M. Toxicology for chemists; Georg Thieme: New York, 1994; Chapter 6.2.; pp 177-188.

EXPERIMENTAL SECTION Laboratory Contamination with Musk Compounds. Due to the widespread use of musks in household products, it is important to remove all possible sources from the laboratory (soaps, hand creams, towels, etc.) that might contain these compounds in order to minimize sample contamination. Also, blank samples should be run in parallel to every series of samples extracted. Standards, Chemicals and Solvents. Nitro musks: musk xylene (MUXY, 1-tert-butyl-3,5-dimethyl-2,4,6-trinitrobenzene), musk ketone (MK, 4-tert-butyl-2,6-dimethyl-3,5-dinitroacetophenone), musk moskene (MM, 1,1,3,3,5-pentamethyl-4,6- dinitroindane), musk tibetene (MT, 1-tert-butyl-3,4,5-trimethyl-2,6-dinitrobenzene), and musk ambrette (MA, 1-tert-butyl-2,4-dimethyl-6-methoxy-3,5dinitrobenzene) were obtained from Essentia, Aetherische Oele, Winterthur, Switzerland; Amino nitro musk metabolites were prepared synthetically: AMUXY, 1-tert-butyl-3,5-dimethyl-4-amino2,6-dinitrobenzene; AMK, 4-acetyl-1-tert-butyl-3,5-dimethyl-2-nitro6-aminobenzene; AMM, 1,1,3,3,5-pentamethyl-4-nitro-6-aminoindane; AMT, 1-tert-butyl-3,4,5-trimethyl-2-amino-6-nitrobenzene; AMA, 1-tert-butyl-2,4-dimethyl-3-amino-6-methoxy-5-nitrobenzene. Cyclohexane, chloroform, dichloromethane (DCM), ethyl acetate, hexane, methanol, 2-propanol, toluene, and acetone, residue grade, were obtained from Merck, Dietikon, Switzerland, and Promochem, Wesel, Germany. Silica gel 40 for column chromatography was from Merck; palladium on activated alumina, Celite 545 AW, and ammonium formate (CH5NO2) were from Fluka, Buchs, Switzerland. Waters Sep-Pak silica cartriges (0.5 g) were from Waters, Rupperswil, Switzerland. Synthesis of the Amino Metabolites of the Nitro Musks. Amino metabolites were synthesized according to a literature method25 with some modifications. Briefly, the nitro musk (100 mg) was diluted in 3.5 mL of dry methanol. To the solution, 20 mg of palladium on activated alumina and 106 mg of CH5NO2 were (24) Blok, H. Measurement of musk fragrances in sludges of sewage treatment plants in The Netherlands. BKH Consulting Engineers, 1997. In van de Plassche, E. J.; Balk, F. Report 601503 008, Enivronmental risk assessment of the polycyclic musks AHTN and HHCB according to EU-TGD; 1997; pp 1-121. (25) Ram, S.; Ehrenkaufer, R. E. Tetrahedron Lett. 1984, 25, 3415-3418.

Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

2125

added and the reaction stirred at room temperature for 2-18 h depending on the reactivitiy of the nitro group. After completion of the reaction, the mixture was filtered over Celite, concentrated, redissolved in chloroform, washed with NaHCO3 and deionized water, dried over Na2SO4, and concentrated to dryness. Purification of the product was performed using silica gel column chromatography (100 × 10 mm) and hexane/ethyl acetate (2:1) as elution solvent. After reconcentration, the pure product was redissolved in hot hexane from which crystallization occurred upon leaving at 4 °C. Yields of the different compounds were as follows: AMUXY 100%, AMK 66%, AMM 82%, AMT 79%, and AMA 52%. Spectroscopic Characterization of the Amino Metabolites by 1H and 13C NMR Spectroscopy. NMR experiments with the amino derivatives of the nitro musk compounds were performed on a Bruker DRX 500 spectrometer at 298 ( 1 K, operating at 500.13 MHz for 1H. A 5-mm multinuclear probehead dedicated for inverse 1H detected heteronuclear experiments and equipped with a z-gradient coil for gradient-enhanced spectroscopy was used. 1H and 13C 90° pulse lengths were 7.5 and 15 µs, respectively. A total of 20-30 mg of each compound was dissolved in CDCl3, and a few drops of CDCl3 dopped with tetramethylsilane (TMS) was added to refer 1H and 13C chemical shifts to this internal standard. The standard Bruker pulse programs ZG, ZGDC, and DEPT were used to record the basic 1H, 13C, and 13C DEPT-135 spectra. The Bruker program INV4GSLPLRND was applied to perform the 2D HMBC experiment to correlate 1H and 13C chemical shifts via heteronuclear long-range (nJCH) couplings. Spectral widths were adjusted in both dimensions to encompass all 1H and 13C signals. Delays to suppress 1JCH cross-peaks and to allow nJCH correlations to evolve were set to 3.45 and 50 ms, respectively. The responses of four scans for each of 512 t1 increments were acquired. Zerofilling in t1 and squared cosine window functions in both dimensions were applied prior to 2D Fourier transformation. The NOEDIF Bruker pulse program was used to perform the selective 1D NOE experiments in a “pseudo-2D” mode. All proton signals were selected as targets for selective irradiation for each compound to monitor their complete NOE behavior. The NOE buildup delay was set to 4 s, and a decoupler power of ∼0.01 W for selective presaturation was used as the optimum with respect to selectivity and sensitivity for these samples. Eight scans were acquired for each irradation frequency, including two “offresonance” irraditions, and an overall cycle was applied eight times (giving a total of 64 scans per irradiation frequency) to minimize effects originating from eventual long-term fluctuations in spectrometer stability. Difference FIDs were calculated for each irradiation frequency, and exponential weighting was applied prior to Fourier transformation. Sample Characterization. Sewage sludges from different catchment areas were collected: six samples (A1-A6) containing mainly domestic sewage obtained from separate sewer systems (A-type); three samples (B1-B3) containing a mixture of domestic sewage, stormwater runoff, and low amounts of industrial wastewater obtained from combined sewer systems (B-type), and one sample (C) as described for the B-type but containing higher amounts of industrial wastewater (C-type). Liquid sewage sludge was sampled into glass bottles precleaned with hexane and 2126

Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

acetone, stored overnight at 4 °C, and extracted the next day. For dry mass (dm) determinations, sewage sludge samples (1.5 L) were dried at 105 °C for 10-14 h until a constant weight was achieved. Dry mass (%) of the sludge samples was as follows: A1, 3.29; A2, 2.35; A3, 2.45; A4, 3.6; A5, 5.29; A6, 3.61; B1, 4.95; B2, 4.02; B3, 5.76; C, 3.23. Sample Extraction. Liquid sewage sludge (1 L, n ) 1) was extracted with 600 mL of hexane at ambient temperature for 2 h by agitating vigorously. Phase separation was improved by adding NaCl and small amounts of 2-propanol to the two-phase system. The organic phase was decanted, dried over Na2SO4, and concentrated at 35 °C under a slight vacuum to a volume of 5 mL. Ethyl acetate (5 mL) was added and the solution (total volume 10 mL) further used for the GPC cleanup. Cleanup: Gel Permeation Chromatography (GPC) and Solid-Phase Extraction (SPE). Gel permeation chromatography was carried out on a Autoprep GPC model 1002 B from Gerber Instruments, Effretikon, Switzerland. Cyclohexane/ethyl acetate (1:1) was used as eluent at a flow rate of 5 mL/min. The column (300 × 20 mm) was filled with 30 g of Bio-Beads SX 3 (Bio-Rad Laboratories, Hercules, CA). For GPC, 5 mL of the extraction solution was used. The sampling window was between 15 and 27 min (75 and 135 mL). The GPC eluates were evaporated at 35 °C to ∼5 mL and concentrated to 1 mL by a gentle stream of nitrogen. After GPC, extracts were submitted to a final cleanup by use of a silica cartridge (SPE). The column was conditioned with 10 mL of DCM. After the cartridge was loaded with the extract, elution was carried out with 8 mL of the same solvent. The eluate was carefully concentrated nearly to dryness by a gentle stream of nitrogen and the volume adjusted to exactly 1 mL with toluene. Preparation of Standard Solutions and HRGC/Ion Trap MS Quantification Using CID. Stock solutions (100 µg/mL) of the standard compounds were prepared in toluene. A standard mixture of 500 ng/mL was obtained by appropriate dilution of a mixture of the stock solutions in toluene. Standard mixtures of lower concentrations were prepared by suitably diluting the 500 ng/mL standard mixture with toluene. HRGC was performed on a Varian 3400 GC equipped with a Varian 1078 injector, a Varian 8200 autosampler, and a Varian Saturn 2000 ion trap mass spectrometer with MS/MS option. Separation was achieved on a 60-m DB-1 capillary column, 0.25-mm i.d., 0.25-µm film (J&W Scientific, supplied by msp Friedli, Koeniz, Switzerland). Helium was used as a carrier gas. The temperature program was 100 °C, 1 min hold, 15 °C/min to 157 °C, 0.6 °C/min to 178 °C, 5 °C/min to 220 °C, 10 °C/min to 300 °C, and 6.8 min hold. The injector temperature was set at 260 °C. Injection (2 µL) was performed in the splitless mode using a splitless time of 0.7 min. The mass range for spectrum acquisition was 100-300 m/z, CID-RF m/z 90, and mass width 3 m/z for all compounds. Table 1 shows MS/ MS parameters used for the external standard quantitation. For quantification, peak heights were used and abundances of product ions were summed. Detector linearity was determined by linear regression analysis of six-level calibration curves for each compound. After achieving coefficients of regression of r ) 0.99, linear regression equations were calculated. A typical calibration curve for AMUXY is shown in Figure 2. The amount of each analyte in the final extract was obtained from the corresponding calibration curve. Sludge concentrations (Ci, µg/kg dm) were calculated

Table 1. Conditions for Resonant CID MS/MS Fragmentation1a

musk compd 1 2 3 4 5 6 7 8 9 10

MA MUXY MM1 AMA1 MT AMM MK AMT AMK AMUXY

product ions for response precursor quantification CID factors tR ion (m/z) (m/z) res (V) (ng-1) (min) 253 282 263 238 251 233 279 236 264 252

219, 251 265, 280 229, 245, 246 165, 221, 223 204, 234, 249 187, 188, 216 191, 262 162, 163 148, 191 218, 235

0.30 0.35 0.30 0.60 0.30 0.35 0.35 0.35 0.30 0.35

7.52 3.92 1.32 6.54 7.44 38.8 5.84 4.53 4.10 15.02

34.9 37.1 39.7 39.7 42.4 43.2 45.2 45.6 50.6 51.4

a These substances were separated in the MS by multiple reaction monitoring.

Table 2. Recovery Data of Nitro Musks (NM) and Its Amino Metabolites (AM) from Water and Sewage Sludge compound class

name

nitro musk nitro musk nitro musk amino metabolite nitro musk amino metabolite nitro musk amino metabolite amino metabolite amino metabolite mean recovery NM mean recovery AM

MA MUXY MM AMA MT AMM MK AMT AMK AMUXY

recovery rates (%), n=2 water sewage sludge 107, 94 113, 107 104, 92 70, 67 109, 102 80, 75 104, 100 107, 110 100, 94 95, 87 103 89

109, 116 105, 102 103, 89 65, 72 91, 86 103, 92 111, 103 84, 78 60, 51 83, 71 102 76

500 ng/L analytes, (ii) samples of sewage sludge were spiked with 1 µg/L nitro musks and the corresponding amino metabolites. The proposed method allowed the detection of as little as 2 pg for most of the compounds (S/N ) 3), corresponding to 2 ng/L in the sewage sludge. However, due to the rather high background still present in the matrix, the real detection limit was estimated to be 50 ng/L (∼1 µg/kg dm), depending on the sludge under investigation (Table 2).

Figure 2. Standard calibration curve for AMUXY.

according to the equation

Ci )

(H ( b) × 2 × 100 Ri × dm

where H is the signal height of the analyte to be determined, b is the intercept, Ri is the response factor, and dm corresponds to the dry mass of the sludge (see section on sample characterization). HRGC and Electron Impact (EI), Positive Chemical Ionization (PCI), and Negative Chemical Ionization (NCI) Quadrupole MS of the Nitro Musk Amino Metabolites. HRGC was performed on a HP 6890 GC equipped with a HP 6890 Series automatic injector (enhanced parameters) and a HP 5973 mass spectrometer with PCI and NCI option. Separation was reached on a 30-m HP 35 MS capillary column, 0.25-mm i.d. and 0.25-µm film (Hewlett-Packard, Wilmington, DE) using the following temperature program: 90 °C, 1 min hold, 10 °C/min to 300 °C, and 10 min hold. Helium was used as a carrier gas at a linear velocity of 37 cm/s measured at a temperature of 90 °C. Standards (2 µL) were injected in the on-column mode. In the CI mode (PCI and NCI), CH4 was used as a reagent gas. A chemical ion gas purifier for CH4 was installed just before CH4 entered the MSD. The CH4 flow was set at 20% for PCI and 40% for NCI at a flow of 5 mL/min. MS temperature parameters: transfer line 290 °C, quadrupole 106 °C (PCI), 110 °C (NCI), ion source 250 °C (PCI), and 110 °C (NCI). Recovery Experiments and Detection Limits. Two different recovery experiments were carried out: (i) water was spiked with

RESULTS AND DISCUSSION Structural Elucidation of the Nitro Musk Amino Metabolites by 1H and 13C NMR Spectroscopy. To prove the degree and the site of hydrogenation for the investigated series of nitro musk compounds and to spectroscopically characterize the corresponding amino products, i.e., to unequivocally assign all proton and carbon signals, standard 1D methods to measure 1H, 13C, 13C DEPT, and 1H{1H} NOE spectra were applied together with the powerful heteronuclear 2D shift correlation experiment HMBC,26 to selectively detect heteronuclear long-range couplings (nJCH). The results of this NMR investigation may be summarized as follows: The number of proton and carbon signals, the relative intensities of the proton signals, and the multiplicities of the carbon signals obtained with the DEPT spectra confirmed the high purity of the measured samples and allowed first tentative signal assignments to be made. For each of the compounds, a broadened signal in the range of 3.6-4.1 ppm, corresponding to an equivalent of two protons and attributed to the amino protons, could be observed, which proves the hydrogenation of only one nitro group (Figure 3). Whereas for compounds AMK and AMT with two “equivalent” nitro groups obviously no differentiation between the two identical hydrogenation products can be made, the site of hydrogenation for AMUXY could immediately be determined on the basis of symmetry arguments and the number of proton and carbon signals measured in the corresponding spectra. The number of only three proton and seven carbon signals is in agreement only with a hydrogenation of the former nitro group at C-4. The site of hydrogenation for compounds AMA and AMM could be determined by exploiting homonuclear 1H{1H} NOE (26) Wilker, W.; Leibfritz, D.; Kerssebaum, R.; Bermel, W. Magn. Reson. Chem. 1993, 31, 287.

Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

2127

Table 3. 1H and 13C Chemical Shifts Together with 1H{1H} NOE and 1H/13C nJ CH Connectivities of the Monoamino Metabolites Obtained from Corresponding NMR Experiments C( )

shift 1Ha

shift 13Ca

1H{1H}

NOEb

nJ

c CH

AMUXY 1 2, 6 3, 5 4 Cq(t-Bu) CH3(t-Bu) CH3(C-3,5) NH2

1.3911 1.9942 4.0040

120.536 143.619 115.784 151.210 37.222 31.433 13.853

CH3(t-Bu) CH3(C-3,5) CH3(C-3,5) CH3(C-3,5) CH3(t-Bu) CH3(t-Bu)

AMM 1

Figure 3. Structures of the amino nitro musk metabolites as determined by 1H and 13C NMR.

effects.27 NOE effects (Table 3) measured for AMA and induced by the saturation of the NH2 protons for protons H(C-2) and CH3(C-4), but not for the protons of the methoxy group, prove that the former nitro group at C-3 had been hydrogenated. NOE effects measured for AMM and induced by the saturaration of the NH2 protons for H(C-7) and CH3(C-5), but not for the CH3(C-3) protons, proves that the former nitro group at C-6 had been hydrogenated. NOE effects together with heteronuclear one-bond (1JCH) and longrange (nJCH) interactions served to unambiguously assign all proton and carbon signals (Table 3) for all compounds. The basic 2D HMBC26 experiment was applied to measure connectivities through nJCH and 1JCH couplings taking advantage of the usually nonperfect 1JCH suppression filter incorporated in this pulse sequence giving rise to small residual 1JCH satellites. They could be easily identified and differentiated from the nJCH cross-peaks by their large splittings (130-165 Hz) in the 1H domain. To demonstrate how this complementary information has been used for signal assignments, NOE traces and the 2D HMBC spectrum (expansion) measured for AMK are shown in Figure 4. Starting with the NOE responses measured with the saturation of the NH2 protons, the CH3(C-6) and CH3(t-Bu) signals can immediately be assigned. The latter show a strong cross-peak in the HMBC spectrum which must be attributed to C-4 (3JCH). The CH3(C-6) protons show three cross-peaks with aromatic carbons which must be attributed to C-1 (3JCH), C-5 (3JCH), and C-6 (2JCH). Whereas C-6 can immediately be assigned; taking advantage of carbon chemical shift arguments the differentiation between C-1 and C-5 is based on the fact that C-1 (but not C-5) must show additional 3JCH cross-peaks to CH3(C-2) and CH3(CO). For CH3(C-2), one expects additional cross-peaks to C-2 (2JCH) and C-3 (3JCH), which may easily be assigned due to carbon chemical shift arguments. These assignments and the site of hydrogenation are further confirmed by weak 3JCH cross-peaks in the HMBC spectrum between the NH2 protons and C-4 and C-6. Liquid/Liquid Extraction and Cleanup Procedures of the Sewage Sludge Samples. Due to the rather nonpolar character of the nitro musk compounds, extraction was carried out similarly to that of organochlorine pesticides and PCBs, for which nonpolar organic solvents such as hexane, cyclohexane, or petroleum ether (27) Neuhaus, D.; Williamson, M. The Nuclear Overhauser Effect; VCH: Weinheim, 1989.

2128 Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

2 3 3a

42.490 1.9011

4 5 6 7 7a CH3(C-1)

1.2677

CH3(C-3) CH3(C-5) NH2

1.2944 1.9811 3.7311

6.5465

59.393 43.971 130.302 145.686 112.496 149.813 111.583 153.881 31.901 30.310 12.471

CH3(C-1,C-3)

H(C-7), CH2, CH3(C-1) CH3(C-1,C-3) CH2, CH3(C-3) H(C-7), CH2, CH3(C-3) CH3(C-5) H(C-7), CH3(C-5) H(C-7),d CH3(C-5)

NH2, CH3(C-1) CH2, H(C-7) CH2

CH2, CH3(C-1) CH2, CH3(C-1) CH2, CH3(C-3)

H(C-7), CH3(C-5) AMT

1 2 3 4 5 6 CH3(C-3) CH3(C-4) CH3(C-5) Cq(t-Bu) CH3(t-Bu) NH2

2.1256 2.2107 2.0098 1.4953 3.9768

120.337 143.029 124.517 135.338 117.626 151.250 15.125 17.759 15.637 36.233 30.242

CH3(t-Bu) CH3(C-3) CH3(C-3,C-4) CH3(C-3,C-4,C-5) CH3(C-4,C-5) CH3(C-5) CH3(C-4) CH3(C-3,C-5) NH2, CH3(C-4) NH2 CH3(C-3, t-Bu)

CH3(t-Bu) CH3(t-Bu)

AMK 1 2 3 4 5 6 Cq(t-Bu) CH3(t-Bu) CH3(C-2) CH3(C-6) CH3(CO) NH2 1 2 3 4 5 6 Cq(t-Bu) CH3(t-Bu) CH3(C-4) OCH3 NH2

1.4958 1.9454 2.0310 2.4524 4.1062

6.7217

1.3551 2.0005 3.7394 3.6166

142.726 113.129 151.604 123.289 143.883 119.412 36.440 29.913 14.590 15.7445 33.4178

NH2 CH3(CO) CH3(CO), NH2 CH3(C-2, C-6) CH3(C-6, t-Bu)

AMA 141.260** 115.870 CH3(t-Bu) 140.528 113.401 143.672 141.260** 35.962 31.529 H(C-2), OCH3 12.552 63.502 CH3 (t-Bu) H(C-2), CH3(C-4)

CH3(C-2, C-6) CH3(C-2) CH3(C-2) CH3(t-Bu), NH2d CH3(C-6) CH3(C-6), NH2d CH3(t-Bu) CH3(t-Bu)

H(C-2), CH3(t-Bu) CH3(C-4), H(C-2)d H(C-2), CH3(C-4) CH3(C-4) OCH3 CH3(t-Bu) CH3(t-Bu)

a Chemical shifts measured in CDCl relative to internal TMS. b NOE 3 effects measured for indicated protons after preirradiation of H-C( ) c (column 2). Heteronuclear long-range coupling interactions of indicated protons to carbon C( ) (column 2). d Weak cross-peak.

are used. Recovery rates of the different compounds are summarized in Table 2. Nitro musks were recovered quantitatively whereas rates of the amino metabolites were slightly lower except

Figure 5. MS/MS run of a standard mixture containing nitro musks and amino metabolites. For compound identification, see Table 1.

Figure 4. 1H{1H} NOE traces and 1H/13C HMBC spectrum measured for AMK. Arrows in the NOE spectra mark the target signals for selective irradiation.

for AMK of which only 50% was found. The reason for this might be the more polar character of this compound. For extraction, liquid sewage sludge was preferred over dried material in order to avoid analyte loss during the pretreatment step. Frequently sewage sludge contains significant amounts of lipid-like material which is coextracted with the analytes. Therefore, a cleanup of the extracts is essential. GPC allowed us to partly remove the lipid matrix and SPE to eliminate further interfering compounds. HRGC/Ion Trap MS Quantification and Mass Spectral Characterization of the Amino Metabolites of the Nitro Musks. Separation of most of the compounds could be achieved by the chromatographic conditions described. Coelution was observed for MM (3) and AMA (4) as shown in Figure 5. Solution to this problem was offered by the multiple-reaction-monitoring tool (MRM) of Saturn 2000, allowing change in the precursor ion from scan to scan. Quantification of the analytes was based on product ions after collision-induced dissociation (CID) in the ion trap (Table 1). Basically, two groups of mass spectra can be distinguished: one having the molecular ion (M+) as base peak, as for AMA, AMT, and AMK; the other group having the M+ - CH3 fragment ion as base peak, as for AMM and AMUXY. CID experiments on the precursor ion (M+ or M+ - 15; see Table 1) resulted in mass spectra with characteristic product ions. Their presence can be rationalized as follows: The MS/MS spectra of AMA, AMT, and AMK show a strong, common M+ - 73 fragment probably due to the cleavage of a methyl and a nitro group (M+ - CH3 - NO

- CO). A second set of experiments was performed on the M+ 73 fragment ions of AMA, AMT, and AMK. For AMA, for example, loss of m/z 30 (NO) and 58 (NO + CO) again demonstrate that another nitro group is involved. Two model compounds were found in the NIST mass spectral database for comparison of spectral features: 4-(1,1-dimethylethyl)-2,6-dinitrobenzenamine (C10H13N3O4) and 4,5-dimethyl-2-nitrobenzenamine (C8H10N2O2). The mass spectra of the first compound is dominated by the M+ - 15 (M+ - CH3) fragment, followed by M+ - 61 (M+ - CH3 NO2) and M+ - 107 (M+ - CH3 - (NO2)2). The second compound shows fragment ions at M+ - 30 (M+ - NO), M+ - 46 (M+ NO2), and M+ - 58 (M+ - NO - CO) and a major fragment ion at M+ - 73 (M+ - CH3 - NO - CO). Clearly, cleavage of a methyl and a nitro group are most important in these model compounds and are also found in the mass spectra of the amino musk metabolites. Furthermore, AMA reveals a strong M+ - 17 fragment (m/z 221) which corresponds to the cleavage of an OH radical ion (OH•). This fragmentation is frequently observed for o-nitrotoluenes.28-29 The same type of cleavage can be deduced from the MS/MS spectrum of AMUXY, which shows fragment ions at m/z 235 and 218 corresponding to M+ - 15 - 17 (M+ - CH3 OH•) and M+ - 15 - 2 × 17 (M+ - CH3 - (OH•)2). In AMUXY, this OH• cleavage is very pronounced as there are two nitro groups ortho to a methyl group. Finally, the MS/MS spectrum of AMM shows strong fragmentation ions at m/z 216, 187, and 173 which can be explained as follows: m/z 216 (M+ - 32) corresponds to the cleavage of a CH3 group and an OH• (M+ - 15 - 17) similarly to AMUXY; m/z 187 and 173 are product ions by additional cleavage of a C2H5 (M+ - 15 - 17 - 29) and C3H7 (M+ - 15 17 - 43) originating from ring-opening and cleavage of the tetramethylated cyclopentane ring. EI, PCI, and NCI Quadrupole MS of the Amino Metabolites. In addition to ion trap MS, a comparison study applying quadrupole EI, PCI, and NCI MS was performed in order to get information on fragmentation and response sensitivity of the amino metabolites (Table 4). Besides the M+ peak, numerous fragment ions can be observed in EI MS, whereas M+ - CH3 and m/z 57 (C4H9) are dominant ones. PCI MS shows less fragmentation; (28) Williams D. H.; Fleming I. Spectroscopic Methods in Organic Chemistry, 3rd ed; McGraw-Hill: Maidenhead, U.K. 1975; Chapter 5-7, pp 180-186. (29) Hesse M.; Meier H.; Zeeh B. Spektroskopische Methoden; Thieme Verlag: Stuttart, 1984; Chapter 8.13, pp 345-350.

Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

2129

Table 4. Main Mass Ions of the Amino Metabolites Obtained from Full Scan Spectra Using EI, PCI and NCI Quadrupole MSa EI MS

PCI MS

53(22), 57(29), 65(25), 67(23), 74(27), 91(23), 107(40), 120(39), 132(21), 134(21), 136(21), 146(20), 147(26), 165(33), 177(35), 191(8), 208(4), 221(13), 223(23), 238(100)

41(88), 183(28), 193(13), 209(7), 222(14), 239(100), 267(10), 279(4)

51(15), 65(6), 71(7), 77(7), 91(5), 115(5), 130(6), 144(5), 157(11), 172(17), 187(10), 216(9), 233(100), 248(20)

41(40), 193(56), 202(27), 233(17), 249(100), 277(17), 289(9)

57(100), 65(27), 77(36), 91(26), 107(15), 119(29), 135(68), 148(48), 163(93), 175(29), 187(49), 204(7), 221(40), 236(91)

57(8), 181(100), 190(24), 221(6), 236(24), 237(16), 265(2), 277(2)

57(100), 65(15), 77(18), 91(15), 103(6), 115(10), 134(9), 144(17), 160(15), 176(17), 191(27), 202(6), 215(18), 249(20), 264(53)

57(9), 164(9), 178(7), 209(43), 218(11), 219(8), 233(3), 249(4), 265(22), 293(5), 305(2)

57(33), 65(20), 77(22), 91(19), 106(10), 118(10), 130(11), 144(17), 159(9), 175(6), 187(8), 218(21), 235(12), 252(100), 267(42)

57(55), 177(5), 189(5), 212(12), 252(3), 268(3), 296(9), 308(2)

a

NCI MS

tentative interpretation of some important mass ions

AMA 46(2), 238(100), 239(14)

AMM 248(100), 249(16)

EI-MS: 238 (M+), 223 (M+ - 15, M+ - CH3), 221 (M+ - 17, M+ - OH), 165 (M+ - 73, M+ CH3 -NO - CO), 57 (C4H9+) PCI-MS: 239 (MH+), 267 (M+ + 29, (M+ C2H5)+), 279 (M+ + 41, (M + C3H5)+) 222 (MH+ - 17, MH+ - OH), 209 (MH+ - 30, MH+ - NO), 193 (MH+ 46, MH+ - NO2) NCI-MS: 238 (M-), 46 (NO2-) EI-MS: 248 (M+), 233 (M+ - 15, M+ - CH3), 216 (M+ - 32, M+ - CH3 - OH) PCI-MS: 249 (MH+), 277 (M+ + 29, (M + C2H5)+), 289 (M+ + 41, (M + C3H5)+), 233 (MH+ - 16, MH+ - NH2 or O), 219 (MH+ - 30, MH+ - NO) NCI-MS: 248 (M-)

AMT 46(6), 180(3), 220(2), 236(100), 237(15) AMK 46(16), 264(100), 265(16)

AMUXY 46(100), 204(3), 211(3), 251(2), 267(77), 268(11)

EI-MS: 236 (M+), 221 (M+ - 15, M+ - CH3), 163 (M+ 73 M+ - CH3 - NO - CO), 57 (C4H9+) PCI-MS: 236 ((M -H) +H+)), 237 (MH+), 265 (M+ + 29, (M+ C2H5)+), 277 (M+ + 41, (M + C3H5)+), 190 ((M - H) + H+ - 46, ((M -H) + H)+ - NO2) NCI-MS: 236 (M-), 46 (NO2-), EI-MS: 264 (M+), 249 (M+ - 15, M+ - CH3), 232 (M+ - 32, M+ - CH3 - OH), 191 (M+ - 73, M+ CH3 - NO - CO), 57 (C4H9+) PCI-MS: 265 (MH+), 293 (M+ + 29, (M + C2H5)+), 305 (M+ + 41), (M + C3H5)+), 249 (MH+ - 16, MH+ - NH2 or O), 219 (MH+ - 46, MH+ - NO2) NCI-MS: 264 (M-), 46 (NO2-) EI-MS: 267 (M+), 252 (M+ - 15, M+ - CH3), 235 (M+ - 32, M+ - CH3 - OH), 218 (M+ - 49, M+ -CH3 - 2 (OH)) PCI-MS: 267 ((M -H) +H+)), 268 (MH+), 296 (M+ + 29), (M+ + C2H5)+), 308 (M+ + 41), (M + C3H5)+) NCI-MS: 267 (M-), 46 (NO2-)

Values in parentheses are percent abundance of the base peak (boldface numbers).

typically a molecular ion MH+ due to proton-transfer reactions and adduct ions M+ + 29 and M+ + 41 can be identified, well known from PCI with methane.30 NCI MS exhibits virtually no fragmentation. The molecular ion (M-) turned out to be the base peak besides m/z 46, characteristic for the presence of a nitro group. Regarding the sensitivity of the different ionization methods, NCI was the most sensitive one, allowing the detection of femtogram quantities, followed by EI and finally PCI. As a conclusion, quadrupole NCI MS might be a valuable alternative to ion trap MS/MS for low-level quantification of these amino metabolites. Concentrations of Nitro Musk Compounds and the Corresponding Amino Metabolites in Sewage Sludges of Different Catchment Areas. Levels of nitro musks as well as the corresponding amino metabolites were determined in some selected sewage sludges containing no (A-type), low (B-type), or higher (C-type) amounts of industrial wastewater. A typical chromatogram of an extract is shown in Figure 6 and quantitative results are presented in Table 5. Among the nitro musks, MK was identified in 7 out of 10 sludges originating from domestic as well as low-level industrial wastewater (A/B-type) whereas MUXY (30) Chapman J. R. Practical Organic Mass Spectrometry; John Wiley and Sons: New York, 1985, Chapter 3, pp 46-64.

2130

Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

was detected in just one B-type sludge. Concentrations were rather low with an average concentration of 5 µg/kg dm for MK and one value of 32 µg/kg dm for MUXY. These results are in a similar range as described in recent studies which found values of 1060 µg/kg dm (mean 30 µg/kg dm) for MK and 5 µg/kg dm for MUXY.31-32 The concentrations found in our study reflect the industrial production volume of these compounds, MUXY and MK being the major nitro musks still used in different products. Also, efforts undertaken by the fragrance industry in Switzerland since 1994 to reduce the amount of MUXY in detergents and cleaning agents tend to be confirmed by our results. MA, MT, and MM could not be detected in the sewage sludge samples. The use of MA was banned in 1995 in the EU and included in the list of banned components in cosmetics. MA was also forbidden in Switzerland in 1998. MT and MM play a minor role as fragrances and are produced only in low quantities. Considerable research is focused on transformation products of the nitro musks that have been described to occur in the environment and biota.20,21 These metabolites are formed by (31) Sauer, J.; Antusch, E.; Ripp, Ch. Vom Wasser 1997, 88, 49-69. (32) Blok, J. Measurements of polycyclic and nitro musks in sludge of sewage treatment plants in The Netherlands; Report to RIVM, BKH Consulting Engineers, 1998.

Table 5. Mean Concentrations of Nitro Musks (MA, MUXY, MM, MT, MK) and Nitro Musk Amino Metabolites (AMA, AMUXY, AMM, AMT, AMK) in Sewage Sludges of Different Catchment Areas (µg/kg dm) sludge typea

MA

MUXY

MM

MT

MK

AMA

AMUXY

AMM

AMT

AMK

A1 A2 A3 A4 A5 A6 B1 B2 B3 C

ndb nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd 32.5 nd nd nd

nd nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd nd nd nd

6.1 6.8 6.9 1.5 nd 5.1 7.1 nd 3.1 nd

nd nd nd nd nd nd nd nd nd nd

10.9 15.3 49.1 nd nd 24.9 31.5 nd 15.6 nd

7.9 nd 1.6 nd 0.7 nd nd 7.2 nd 36.2

nd nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd 13.1 nd nd nd

a A1-A6, domestic bnd, not detected.

sewage; B1-B3, sewage with low amounts of industrial wastewater; C, sewage with high amounts of industrial wastewater.

Figure 6. Total ion chromatogram of sewage sludge extract A1 and extracted ion chromatograms (EIC) of 6 (AMM,), 7 (MK), and 10 (AMUXY). Ions used for reconstructed ion current profiles of 6, 7, and 10: 6, m/z 187 + 188 + 216; 7, m/z 191 + 262; 10, m/z 218 + 235; selected MS/MS spectra obtained from sample A1 and the corresponding standards.

reduction of a nitro group to the corresponding amine. Very recently, published data have shown that the 4-amino-musk-xylene (AMUXY) was much more toxic for Daphnia magna than the parent compound,33 emphasizing the importance of these compounds for risk assessment considerations. In our sludge study, three amino metabolites could be identified and quantified: AMUXY, AMM, and AMK. Concentrations were between 1 and (33) Behechti, A.; Schramm, K. W.; Attar, A.; Niederfellner, J.; Kettrup, A. Water Res. 1998, 32, 1704-1707.

50 µg/kg dm and therefore reached levels in the same order as the parent compounds or even higher. AMUXY turned out to be the major metabolite, followed by AMM and AMK. Clearly, the two nitro musks MUXY and MM are found mainly as their amino metabolites. The decrease of the concentrations of the parent compounds and the increase of the amino metabolites strongly suggest that, besides adsorption to sludge particles, transformation reactions of nitro musks under anaerobic conditions are a relevant process in the STP. The amino metabolite of MK was found in only one sludge sample. Based on its structure, other reactions involving the acetyl group present in the molecule seem to dominate the metabolism of this compound, and therefore, the reduction of the nitro group plays a minor role. As expected, no amino metabolites of MA and MT were detected. From these results, the following conclusions can be drawn: sewage sludge represents an excellent environmental indicator to study the input and fate of nonpolar and poorly degradable organics such as fragrance compounds released into the wastewater. Information regarding the efforts undertaken by industrial companies to reduce the amount of such products into the environment might be obtained by monitoring their concentrations on a long-term basis. Finally, the occurrence of metabolites clearly demonstrates the need to look not only for the parent compounds but also for transformation products. Future investigations will therefore concentrate on the analysis of sewage sludges for polycyclic musks and metabolites thereof. ACKNOWLEDGMENT T. Kupper, T. Candinas, G. Chassot are highly acknowledged for their cooperation in this project, W. Stauffer and co-workers for sewage sludge sampling, and G. Sabbioni (University of Munich, Institute of Pharmacology) for recommendations of the synthesis of the amino metabolites. Received for review August 31, 1999. Accepted January 28, 2000. AC991006D

Analytical Chemistry, Vol. 72, No. 9, May 1, 2000

2131