Structure Elucidation of an Artifact Discharging from Rubber-Based

The use of vial closures equipped with butyl rubber septa may lead to sample contamination by rubber additives discharging from the septum material...
0 downloads 0 Views 118KB Size
Download PDF
Anal. Chem. 2006, 78, 8156-8161

Structure Elucidation of an Artifact Discharging from Rubber-Based Vial Closures by Means of Gas Chromatography/Tandem Mass Spectrometry Thomas Kapp and Walter Vetter*

Institute of Food Chemistry, University of Hohenheim, Garbenstrasse 28, 70599 Stuttgart, Germany

The use of vial closures equipped with butyl rubber septa may lead to sample contamination by rubber additives discharging from the septum material. In this study, the structure elucidation of an artifact causing intense signals in gas chromatography/electron capture negative ion mass spectrometry (GC/ECNI-MS) and gas chromatographic analyses with electron capture detection is described. Tentative identification of the leached compound was achieved by employing tandem mass spectrometric techniques both in electron capture negative ion and in electron ionization modes. The artifact could thus be characterized as 2-benzothiazolyl-N,N-dimethyl dithiocarbamate, which is a known vulcanization accelerator for rubber. It is conceivable that the identified compound or related substances are also used in other applications. Therefore, two food-related matrixes were investigated for a possible migration of this compound into foods. During these analyses, the tentatively identified rubber additive was detected in an aqueous extract of a rubber seal ring for canning jars. GC/ECNI-MS provided better sensitivity and selectivity than GC/EI-MS for the determination of the rubber additive and other mercaptobenzothiazolederived substances. The contamination of analytical sample solutions by migration from storage materials can yield artifact peaks which may interfere with quantitation and identification of the analytes. Apart from the widely known problem of sample contamination by plasticizers such as phthalates or adipates originating from plastic materials,1 artifacts can also be incurred by rubber additives discharging from vial closures.1-4 During gas chromatography/electron capture negative ion mass spectrometry (GC/ECNI-MS) and gas chromatography/electron capture detection (GC/ECD) analyses of organochlorine contaminants, we often noticed an abundant peak that eluted in the same retention range as many pesticides or other contaminants (see Figure 1) but showed a mass spectrum not corresponding to any analyte. It became clear that this interfering * To whom correspondence should be addressed. Phone: +49 711 45924016. Fax: +49 711 45924377. E-mail: [email protected]. (1) Oehme, M.; Vetter, W.; Schlabach, M. Int. J. Environ. Anal. Chem. 1994, 55, 261-266. (2) Pattinson, S. J.; Wilkins, J. P. G. Analyst 1989, 114, 429-434. (3) Levine, S. P.; Puskar, M. A.; Dymerski, P. P.; Warner, B. J.; Friedman, C. S. Environ. Sci. Technol. 1983, 17, 125-127. (4) Missen, A. W.; Gwyn, S. A. Clin. Chem. 1978, 24, 2063-2064.

8156 Analytical Chemistry, Vol. 78, No. 23, December 1, 2006

Figure 1. GC/ECD chromatogram (CP-Sil 8 column) illustrating the potential of the artifact to be misinterpreted as an analyte in residual analysis. Solid line, PCB standard mixture (100 pg per substance); dotted line, sample solution contaminated by migration from vial closure.

signal originated from crimp-cap vial closures because it was observed only upon employment of these closures. Because of the artifact’s high abundance in GC/ECNI-MS analyses, its identity and the conditions affecting its presence were investigated. Generally, the occurrence of such contamination is not limited to applications in residual analysis, but has also been reported in forensic chemistry,4 pharmaceutics,5 and in manufacturing and storage of foodstuffs.6,7 In the latter fields, migration of known or unknown constituents is of special importance because it may pose a risk to patients’ and consumers’ health.5 A detailed risk assessment is not possible for unidentified contaminants. Therefore, two basic steps in risk assessment of an artifact released from the butyl rubber septa of crimp-cap vial closures used in our lab are described in this article. First, the extent of contamination was monitored under different conditions, revealing information on the leachability of the migrant. Second and more important, a complete structure elucidation for the interfering compound was attempted. MATERIALS AND METHODS Glassware, Septum Caps, and Chemicals. All vials and closures used for this work were supplied by WiCom (Heppen(5) Schmid, J. Pharmazie 1960, 15, 431-433. (6) Linssen, J. P.; Rijnen, L.; Legger-Huiysman, A.; Roozen, J. P. Food Addit. Contam. 1998, 15, 79-83. (7) Yamazaki, T.; Inoue, T.; Yamada, T.; Tanimura, A. Food Addit. Contam.1986, 3, 145-152. 10.1021/ac0611723 CCC: $33.50

© 2006 American Chemical Society Published on Web 10/27/2006

heim, Germany). Extraction experiments were performed in 10mL, snap-top vials sealed by aluminum crimp-cap closures (20mm o.d.). Subjected to these experiments were butyl rubber/ PTFE septa taken from crimp-cap closures (11-mm o.d.) designed for 2-mL snap-top vials. For contamination studies under realistic usage conditions, 2-mL vials containing solvent were sealed by 11-mm (o.d.) butyl rubber/PTFE crimp-cap closures. After the extraction process, the test solutions were transferred to screwtop vials and sealed with silicone/PTFE closures. Solvents used for extraction experiments included n-hexane (for organic residue analysis, Merck, Darmstadt, Germany), ethyl acetate (99.5%, Acros Organics, Geel, Belgium), isooctane (2,2,4trimethyl pentane, analytical reagent grade, 99.95%, Fisher Scientific, Schwerte, Germany), methanol and acetonitrile (HPLC gradient grade, Carl Roth, Karlsruhe, Germany), toluene (for residue analysis, 99.8%, Fluka, Buchs, Switzerland), cyclohexane (99%, Acros Organics), and acetone (technical grade, purified by distillation on a 50-cm column). For dissolution of 2-mercaptobenzothiazole (MBT, pro analysi, 99%, Fluka), m-xylene (for synthesis, 99%) was supplied from Schuchardt (Hohenbrunn, Germany). 2,2′-Bithiophene (purum, >97%) was obtained from Fluka. PCB mix 1 (Dr. Ehrenstorfer, Augsburg, Germany) was applied as the reference standard in GC/ECD analyses (100 pg per substance, dissolved in isooctane). Equipment and Measurement Parameters. All gas chromatography/mass spectrometry (GC/MS) measurements were carried out on a 1200 triple-quadrupole mass spectrometer connected to a CP-3800 gas chromatograph equipped with a CP8410 autosampler (all items from Varian, Darmstadt, Germany) and a Factor Four CP-Sil 8ms capillary column (30 m, 0.25-mm i.d., film thickness 0.25 µm, Varian Chrompack, Middelburg, Netherlands). Helium 5.0 (Sauerstoffwerke Friedrichshafen, Friedrichshafen, Germany) purified by a CP Gas Clean filter (Varian Chrompack) was used as the carrier gas with a constant flow of 1.0 mL/min. The GC oven temperature program started at 60 °C (held for 3 min), was then raised at 3 °C/min to 110 °C (held for 0.33 min), and was finally raised at 10 °C/min to 270 °C (held for 24 min). The total run time was 60 min. The split/ splitless injector was operated in splitless mode for 2 min and kept at 230 °C. The transfer line temperature was maintained at 280 °C. Methane 4.5 (Air Liquide, Bopfingen, Germany) was used as the reagent gas at a pressure of 8.5 Torr in the ion source, which was kept at 200 °C. The electron energy applied in both electron ionization (EI) and electron capture negative ion modes was 70 eV, and the scan time was set to 0.5 s per cycle. For tandem mass spectrometric analyses, argon 5.0 (Linde, Leuna, Germany) was employed as collision gas at a pressure of 1.6 mTorr and 1.7 mTorr in GC/ECNI-MS/MS and GC/EI-MS/MS experiments, respectively. Collision voltages were varied between 0 and 30 eV. Product ion scans in ECNI mode focused on the ions m/z 166, 167, and 168. In EI mode, m/z 88, 89, 90, 254, 255, and 256 were explored. The SIM width was set to 0.5 u for all experiments. GC/ECD analyses were performed on a HP 5890 Series II gas chromatograph equipped with a HP 7673 A autosampler (HewlettPackard/Agilent, Waldbronn, Germany). The system featured two capillary columns (CP-Sil 2 and CP-Sil 8/20% C18; both 50 m, 0.25mm i.d., film thickness 0.25 µm; Varian Chrompack) that were operated in parallel mode. The flow was divided onto the two

columns by an inlet splitter (Gerstel, Mu¨lheim/Ruhr, Germany). Each column ended in a 63Ni electron capture detector maintained at 300 °C with nitrogen 5.0 (Sauerstoffwerke Friedrichshafen) as the makeup gas. Helium 5.0 (Sauerstoffwerke Friedrichshafen), purified by a CP Gas Clean filter, was used as the carrier gas with 1.25-bar column head pressure. The injector was kept at 250 °C and operated in splitless mode for 2 min. The oven temperature program started at 60 °C (held for 1.5 min); was then raised at 40 °C/min to 180 °C (held for 2 min), followed by another step at 2 °C/min to 230 °C (held for 9 min); and was finally raised at 10 °C/min to 270 °C (held for 0.5 min). The total run time was 45 min. Procedures. Extraction experiments of the 11-mm (o.d.) butyl rubber/PTFE septa were performed with 10 different solvents or solvent mixtures common for residual analysis of organic compounds: n-hexane, isooctane, ethyl acetate, toluene, methanol, acetone, n-hexane/toluene (65/35, v/v), ethyl acetate/cyclohexane (1/1, v/v), toluene/acetone (95/5, v/v), and toluene/acetone (8/ 2, v/v). For each of these extraction agents, five septa were placed into a 10-mL, snap-top vial and covered with 5 mL of solvent or solvent mixture, respectively. The test vessels were then shut tight and stored at room temperature. After 24 h and 7 d, the contents of the crimp-sealed vials were analyzed by GC/MS and GC/ECD. For negative controls, crimp-sealed 10-mL vials containing 5 mL of ethyl acetate were stored upside down over the test period. To check for possible sample contamination during the measurement, closed screw-top vials filled with pure ethyl acetate were stored upside down for 24 h. Their contents were subsequently analyzed as described above. Contamination studies under realistic usage conditions were performed in crimp-sealed 2-mL vials filled with 1 mL of the aforementioned solvents or solvent mixtures. After storage for 24 h and 3 d both at room temperature and in a refrigerator (kept at ∼10 °C), the solutions were transferred to screw-top vials sealed by silicone/PTFE septa in which they were analyzed by GC/ECNIMS. The effect on contamination of sample solutions by piercing the septum during gas chromatographic injections was studied for three selected solvents (n-hexane, ethyl acetate, toluene). For this purpose, four injections within 2 days occurred from each vial containing 1 mL of solvent. The extent of contamination was monitored by GC/ECNI-MS analyses in full-scan mode (m/z 50300), with special attention to the ion trace m/z 166. To determine whether the artifact can also be found in foodrelated rubber products, four batches containing five rubber seals taken from common toggle-clasped beer bottles were put into 20 mL of water and stored in closed glass vessels. Every two weeks, the batches were put into an ultrasonic bath for 15 min to accelerate possible migration processes. After 3 months, the solutions were decanted and extracted twice with 10 mL of ethyl acetate. The combined extracts were concentrated to 1 mL by a gentle stream of nitrogen and analyzed by GC/MS. Furthermore, a standard canning jar rubber ring was refluxed in 25 mL of boiling water for 90 min. After cooling to room temperature, the liquid was decanted and extracted twice with 5 mL of ethyl acetate. The extract was then analyzed both directly and after 10-fold concentration by GC/ECNI-MS. Analytical Chemistry, Vol. 78, No. 23, December 1, 2006

8157

Figure 2. (a) Poorly fragmented GC/ECNI mass spectrum of the artifact. (b) GC/ECNI-MS/MS product ion scan of the ions at m/z 166, 167, and 168. (c) GC/EI mass spectrum of the artifact. (d) GC/EI-MS/MS product ion scan gained by further ionization of the two most abundant signals of the EI mass spectrum (m/z 88 and 254). For screening purposes, the GC/EI-MS/MS experiment occurred with two signals in one data acquisition. However, no signals above the m/z 88 ion were observed.

RESULTS AND DISCUSSION The retention time of the artifact peak in GC/ECNI-MS and GC/ECD analyses (CP-Sil 8 column) was 34.8 and 26.0 min, respectively (Figure 1). With both oven temperature programs, the artifact eluted in a time range within which many analytes observed in residual analysis of environmental contaminants or pesticides also elute. In GC/ECD chromatograms obtained using a CP-Sil 8 column (similar to a DB-5 column), the artifact peak can easily be mistaken for a polychlorinated biphenyl (PCB) because it elutes in the time range between PCB 52 and PCB 101. In GC/ECNI-MS analyses, confusion with analytes can be avoided more easily by surveillance of the corresponding mass spectra (see below). The direct extraction experiments of the 11-mm (o.d.) butyl rubber septa revealed that a fraction of the interfering compound could be extracted from the septa by all of the tested solvents within 24 h. Toluene extraction yielded the highest artifact signal, followed by ethyl acetate and acetone extractions. Protic or highly nonpolar solvents, such as methanol, n-hexane, or isooctane, extracted lesser amounts of the compound. The negative control (extraction of the septum used for sealing the 10-mL testing vessels with ethyl acetate) showed no peaks. A second extraction of septa that had already been subjected to a 24-h ethyl acetate extraction provided no artifact signal. This shows that the total amount of the compound that is accessible from the surface of the septum can be extracted even within 1 day. Investigations of the migration tendency under realistic usage conditions of crimp-sealed vials revealed that contamination normally occurred after the septum was pierced. During appropriate usage, the contents of the vial are usually in no or only in very short contact with the inner surface of the septum. Hence, the vapor pressure and the polarity of the solvent have to be the decisive factors for migration under these conditions. In concordance with this, repetitive GC/ECNI-MS measurements of three solvents (n-hexane, ethyl acetate, toluene) stored in crimp-sealed vials exhibited the most abundant peaks not in toluene, but in ethyl acetate. Between these injections, the septa were not changed, and thus, were pierced four times within 2 days. Using ethyl acetate as solvent, a clearly visible artifact peak could be observed even when two injections occurred within 1 h. 8158 Analytical Chemistry, Vol. 78, No. 23, December 1, 2006

The scarcely fragmented GC/ECNI mass spectrum of the interfering compound always exhibited only one intensive signal at m/z 166. The signal was characteristically accompanied by the presence of two less abundant signals at m/z 167 and 168 with relative intensities of 8% each, relating to the base peak at m/z 166 (Figure 2a). The intensity of the isotopic peak at m/z 168 was on one hand too high to be attributed to carbon isotopes and on the other hand too low for halogen isotopes. Thus, the presence of sulfur in the molecule was assumed. For further fragmentation of the ion cluster, ethyl acetate extracts of the septum caps were investigated by means of tandem mass spectrometry. By specific product ion scans in ECNI mode that were conducted using a collision voltage of 30 eV, fragments could be observed whose isotope pattern confirmed the presence of two sulfur atoms in the precursur ion m/z 166 (Figure 2b). The fragment ion at m/z 134 showed a loss of 32 u when compared to the parent ion at m/z 166. This difference corresponds not only with the mass of a 32S isotope, it was also observed that the isotopic peak at m/z 136 (caused by 34S) now showed only half the intensity of the 13C isotopic peak at m/z 135. Because the parent ion (m/z 166) exhibits two equally abundant isotopic peaks and because according to the mass difference of 32 u, only sulfur has been eliminated from the molecule, two sulfur atoms must have been present in the parent ion. This conclusion is corroborated by the detection of the fragment ion at m/z 102, which could be formed from m/z 134 by subsequent elimination of another sulfur atom. The formation of the fourth fragment ion at m/z 58 is discussed below. With the above information, it was first assumed that the artifact peak may be caused by a bithiophene. This assumption was based on the facts that bithiophenes contain two sulfur atoms per molecule with a molecular weight of 166 u. However, GC/ ECNI-MS analysis of 2,2′-bithiophene revealed a mass spectrum more complex and, thus, significantly different from the artifact (Figure 3a). Although the EI spectrum displayed relative isotopic abundances for the two isotopic peaks relative to the monoisotopic m/z 166 ion similar to those for the artifact (Figure 3b), the retention time of 2,2′-bithiophene (tR ) 23.2 min) did not match that of the artifact (tR ) 34.8 min). Furthermore, the analyzed bithiophene showed a higher response in EI mode than in ECNI

Figure 4. N,N-dimethyl carbamoyl moiety as possible partial structure of the artifact. The fragment ion m/z 88 observed in GC/EI mass spectra is easily accessible by R-cleavage. Note that here, “R” does not denote an alkyl group but the remainder of the molecule.

Figure 5. Complete structure of the artifact, tentatively identified as 2-benzothiazolyl-N,N-dimethyl dithiocarbamate.

Figure 3. (a) GC/ECNI-MS spectrum of 2,2′-bithiophene (528 ng). (b) GC/EI-MS spectrum of 2,2′-bithiophene (52,8 ng).

mode, contrary to the responses observed for the artifact. Hence, it could be ruled out that the artifact arose from any bithiophene isomer. Still, the results of mass spectrometric investigations in ECNI mode strongly suggested that the ion at m/z 166 is the molecular ion of the unknown compound. However, the GC/EI mass spectrum revealed that obviously not m/z 166 but the ion at m/z 254 represented the molecular ion (Figure 2c). In ECNI mode, no higher masses than m/z 166 and its isotopic peaks were detectable, not even in parent ion scans that were conducted using a mass range from m/z 150 to 800. The base peak of the EI mass spectrum (m/z 88) showed a mass difference of 166 u, as compared to the molecular ion (Figure 2c). The fragment ion at m/z 166, which is typically observed under ECNI conditions, can also be found in EI mass spectra, but with very low abundance. In combination with the findings from ECNI experiments it could thus be concluded that this fragment is easily formed, shows high stability once it is formed and has a high potential to stabilize negative charge. To gather further structure information from the EI mass spectrum, product ion scans of the most abundant fragments (m/z 88 and 254) were performed. Using a collision energy of 15 eV, informative spectra for the mass range below m/z 88 were obtained. Accordingly, two fragmentations of the ion m/z 88 were revealed by EI-MS/MS (Figure 2d). The transition m/z 88 f m/z 73 (mass loss 15 u) is typical for the elimination of a methyl moiety, whereas the loss of 32 u (m/z 88 f m/z 56) is consistent with the presence of a sulfur atom. Both the fragments m/z 88 and m/z 73 show an isotope peak with an offset of two mass units and with relative intensity of ∼5%. The ion at m/z 56, however, shows no such satellite. Therefore, it was assumed that the molecule comprises a third sulfur atom in the fragment with m/z 88 apart from the two sulfur atoms already determined in the fragment with m/z 166. The comparatively high relative abundance

of the molecular ion’s isotopic peak at m/z 256 (∼14%), as witnessed in the EI spectrum of the artifact, provides further evidence for the presence of three sulfur atoms in the molecule. Of all empirical formulas that are arithmetically possible for the m/z 88 fragment ion, C3H6NS appeared to be the most plausible one. It can lead to a structure that provides the opportunity to easily eliminate sulfur from the fragment because the sulfur atom can be bound neither as a thioether nor in an alicyclic way. A structure that allows fragmentation, as shown in Figure 2d, and that has the ability to readily stabilize positive charge is the twicemethyl-substituted thiocarbamoyl moiety (Figure 4). Because the artifact shows a molecular ion with even mass (m/z 254), the molecule must contain an even number of nitrogen atoms according to the nitrogen rule. In addition, assuming the fragment ion m/z 166 is an [M - 88]- ion that predominates under ECNI conditions, it must contain at least one nitrogen atom. The only reasonable empirical formula for an ion of the mass 166 u containing two sulfur atoms and an uneven number of nitrogen atoms is C7H4NS2, which has 6.5 rings and double bonds, suggesting the presence of an aromatic ring. Further evidence for this hypothesis offers the signal m/z 58 in ECNI product ion scans (Figure 2b), which has not yet been discussed. On the basis of the elemental limits of the C7H4NS2- ion, this fragment ion’s mass corresponds to the thiocyanate anion [SCN]- or to [C2H2S]-; however, the latter possibility appears to be rather unlikely. In contrast, the thiocyanate anion can be easily formed from a benzothiazole skeleton. By consideration of all discussed fragment ions and their stabilities under the respective ionization conditions, the artifact was tentatively identified as 2-benzothiazolyl-N,Ndimethyl dithiocarbamate (Figure 5). Although this compound is not commercially available, additional evidence for the identification was provided by the mass spectrometric investigation of 2-mercaptobenzothiazole (MBT), which is a hydrolysis product of the artifact. The GC/ECNI mass spectra obtained in full scan mode showed only one intensive signal at m/z 166, as did the spectra of the contaminating compound. Furthermore, identical ionization conditions brought forth identical fragment ions in GC/ECNI-MS/MS product ion scans of both the artifact and MBT (Figures 2b, 6). Thus, the Analytical Chemistry, Vol. 78, No. 23, December 1, 2006

8159

Figure 6. (a) GC/ECNI mass spectrum of 2-mercaptobenzothiazole (7 ng MBT). (b) GC/ECNI-MS/MS product ion scan (m/z 30-170) of the cluster around m/z 166 (330 ng MBT). The fragmentation pattern is identical to that of the artifact (see Figure 2).

Figure 7. (a) The GC/ECNI-MS total ion current (m/z 50-300) of the decocted canning jar rubber ring shows that several compounds can be discharged from the rubber. (b) Monitoring of the base peak ion for the investigated vulcanization accelerator (m/z 166, extracted from the TIC) revealed a clear peak at 34.8 min. (c) A peak at the same retention time can also be found in the positive control (butyl rubber septa decocted under the same conditions).

part of the compound’s structure that is responsible for the abundant GC/ECNI-MS signal was unequivocally elucidated. A compound similar to the structure in Figure 5, providing the same empirical formula (C10H10N2S3) and molecular weight (254 u), is 2-benzothiazolyl-N-ethyl dithiocarbamate (CAS # 152085-18-2). Nevertheless, it is improbable that the artifact is caused by this isomer because the GC/EI-MS/MS product scan (Figure 2d) clearly shows the elimination of a methyl group and not that of an ethyl group. Furthermore, the ethyl derivative has been reported in literature only for its use in photography.8 2-Benzothiazolyl-N,N-dimethyl dithiocarbamate and related compounds are used as vulcanization accelerators in the manufacturing of elastomers such as butyl rubber.9-11 Since elastomers can be utilized in many different applications, including the manufacturing and storage of foodstuffs, the question arose whether the identified vulcanization accelerator is also of relevance as a source of contamination in food. Preliminary extraction tests found that the substance was extracted from the septum caps by boiling water. For this reason, rubber seals of beer bottles with toggle clasp closures and rubber rings of household canning jars were analyzed for the migration of the described vulcanization accelerator. In the extract of the canning jar rubber ring, 2-benzothiazolyl-N,N-dimethyl dithiocarbamate could, indeed, be

detected by GC/ECNI-MS, although with low intensity (Figure 7). In addition to this, several other components were present in the extract of the heated aqueous solution; however, their identities were not explored further in this work. In contrast to the canning jar rubber ring, the additive could not be detected in extracts of beer bottle seals. Because the extraction of the canning jar rubber ring occurred under conditions similar to those employed during the preservation of foodstuffs, it is possible that additives such as 2-benzothiazolyl-N,N-dimethyl dithiocarbamate can also be present in food. The presence of MBT in the positive control (Figure 7c), that is after extraction of butyl rubber septa in boiling water, can be explained by heat-induced hydrolysis of the vulcanization accelerator. Similar breakdown processes for MBT derivatives have been reported previously.12,13 For a group analysis of MBT-derived compounds, it may, therefore, be useful to determine not only the actual derivatives, but also MBT itself. Literature research found several references indicating the usage of related substances also bearing the benzothiazolyl moiety as similarly acting rubber additives, that is, vulcanization accelerators.9-11,14 Sample contamination by possible migration of these compounds was reported in food and drinks12 or upon usage of disposable syringes equipped with rubber plunger seals.15 Vulcanization accelerators based on MBT are also used in the

(8) (9) (10) (11)

(12) Barnes, K. A.; Castle, L.; Damant, A. P.; Read, W. A.; Speck, D. R. Food Addit. Contam. 2003, 20, 196-205. (13) Nawrocki, S. T.; Drake, K. D.; Watson, C. F.; Foster, G. D.; Maier, K. J. Arch. Environ. Contam. Toxicol. 2005, 48, 344-350.

Takiguchi, H.; Nakayama, T. Japanese Patent No. 05011380, 1993. Wolf, G. M.; de Hilster, C. C. Rubber Age 1950, 67, 193-200. Ritter, E. J.; Robinson, C. N. U.S. Patent 2524082, 1950. Hardman, A. F. U.S. Patent 2615893, 1952.

8160 Analytical Chemistry, Vol. 78, No. 23, December 1, 2006

production of rubber tires. Therefore, MBT derivatives can also be found in the environment, for example, in road dust, indicating anthropogenic contamination incurred by tire abrasion.16 On a smaller scale, MBT and its derivatives are also used as antifouling agents and fungicides in the manufacturing of leather and paper.13 On the basis of these applications, the substances may also enter aquatic ecosystems. In addition, the employment of MBT as a corrosion inhibitor for copper in aqueous systems has caused MBT to be detectable even in drinking water.13,17 Although MBT did not show genotoxic activity in the Ames test, it has been reported that MBT may cause transitional cell tumors in the renal pelvis and bladder of rodents.17 Furthermore, MBT is known to provoke dermatitis in sensitized persons.18 Sadly, no toxicological data are available about the compound tentatively identified in this work (2-benzothiazolyl-N,N-dimethyl dithiocarbamate). The comparison of aquatic toxicities of another MBT derivative, 2-(thiocyanomethylthio)benzothiazole, and its degradation products, however, showed that the toxicity of the derivative can be several orders of magnitude higher than that of MBT.16 This clearly shows the importance of a sensitive detection method covering this class of substances. (14) Rodgers, B.; Solis, S.; Tambe, N.; Sharma, B. B.; Waddell, W. H. 167th Spring Technical Meeting, 2005; American Chemical Society, Rubber Division. (15) Salmona, G.; Assaf, A.; Gayte-Sorbier, A.; Airaudo, C. B. Biomed. Mass Spectrom. 1984, 11, 450-454. (16) Kumata, H.; Sanada, Y.; Takada, H.; Ueno, T. Environ. Sci. Technol. 2000, 34, 246-253. (17) Whittaker, M. H.; Gebhart, A. M.; Miller, T. C.; Hammer, F. Toxicol. Ind. Health 2004, 20, 149-163. (18) Williams, T. M.; Hickey, J. L. S.; Bishop, C. C. Health Hazard Evaluation Report No. HETA-83-196-1492; Goodyear Tire and Rubber Company, Gadsden, AL; Hazard Evaluations and Technical Assistance Branch, NIOSH, U.S. Department of Health and Human Services: Cincinnati, OH, 1984.

CONCLUSIONS Analyses of the migrating vulcanization accelerator identified in this work (2-benzothiazolyl-N,N-dimethyl dithiocarbamate) showed a significantly higher sensitivity in ECNI mode than in EI mode. Mass spectra obtained by GC/ECNI-MS exhibited only one characteristic signal at m/z 166 arising from the benzothiazolyl moiety. Therefore, we expect that other mercaptobenzothiazole derivatives will also produce this ion in high abundance. Consequently, monitoring of this fragment ion allows a sensitive and simultaneous detection of both the identified additive and related substances that are also derived from MBT. However, an unequivocal structural assignment requires either measurements in both ECNI and EI modes or the use of appropriate reference standards. It is noteworthy that the structure elucidation presented in this work was based on the employment of GC/ECNI-MS/MS, a technique that has been used only rarely but that has proved to be a powerful tool to gather information not accessible otherwise, for example, by tandem mass spectrometry in EI mode. Because the identified compound was detected not only in vial closures but also in food-related rubber materials, further work is needed to exclude a possible risk for consumers’ health. This includes the development of a reliable quantitation method, preferably using GC/ECNI-MS, GC/ECNI-MS/MS, or both techniques, as well as the gathering of toxicological information on the identified additive and its related compounds.

Received for review June 28, 2006. Accepted September 15, 2006. AC0611723

Analytical Chemistry, Vol. 78, No. 23, December 1, 2006

8161