TANDEM MASS SPECTROMETRY (HRGC–MS–MS)
As has been pointed out previously, HRMS can be used for the characterization and quantification of PCTs and toxaphene, but its application to routine analysis is difficult because of the high resolving power required, and the high cost of the instrumentation. EI-MS–MS can be an excellent alternative to HRMS because of its high sensitivity and its lower cost.
Tandem MS monitoring selected fragment ions produced in the collision cell of a sector or multiquadrupole instrument or in an ion-trap chamber offers advantages over single-stage MS in terms of selectivity of analysis and specificity in molecular structure determination [104]. Tandem MS techniques have been widely used with good results for the analysis of polyhalogenated compounds, such as isomers of the polychlorinated dibenzo-p-dioxines and dibenzofurans [105], PCBs [106], PCNs, and related compounds [107]. However, the applica- tion of MS–MS to the analysis of PCTs and toxaphene has been very limited.
For PCT analysis, Santos et al. [29] reported good results in the determina- tion of the homologue distribution of these compounds using HGC–MS–MS.
Experiments have been performed using the precursor-ion and product-ion scan modes, which give information about interferences and typical fragmentation re- actions, as well as selective reaction monitoring (SRM), which allows one to obtain information about the distribution of PCT homologues. The precursor- ion and product-ion mass spectra of a heptachloroterphenyl from Aroclor 5460 obtained by collision-activated dissociation (CID) with xenon as collision gas are given in Figure 7 as an example [29]. In the precursor-ion spectrum of m/z 470 for heptachloroterphenyls (see Fig. 7a), the ions at m/z 540, 542 and 544, corresponding to the nonacloroterphenyls, and the ions at m/z 505 and 507, corre- sponding to the nonachloroterphenyls after the loss of 1 chlorine atom, are ob- served. In this spectrum, it can be seen that the major interferences in the hepta- chlototerphenyls are the nonachloroterphenyls [28]. In the product-ion mass spectrum of a heptachloroterphenyl (precursor-ion m/z 470) (see Fig. 7b), intense fragment ions are observed corresponding to the loss of 2 and 4 chlorine atoms, [M⫺2Cl]⫹and [M⫺4Cl]⫹. Using the information of the product-ion and pre- cursor-ion mass spectra, the transitions ([M]⫹•→[M⫺235Cl]⫹) for each homo- logue group can be selected and used in SRM experiments. From these experi- ments, it can be deduced that some interferences of the homologues with two additional chlorine atoms occurred. The inability of the HRGC–MS–MS experi- ments to completely overcome interferences is mainly due to the low resolution of the first MS stage, which allows some interfering ions access to the CID cell when chromatographic coelution occurs. Although the interferences are not com- pletely eliminated, the distribution of PCT mixtures can be calculated taking the
144 Galceran and Santos
Analysis of Chlorinated Organic Compounds 145
coelution and the effect of the contribution of the [M⫺2Cl]⫹fragment ions into account, following the same strategy as proposed with LRMS [20]. In order to compare the capabilities of LRMS, HRMS, and MS–MS for the study of homo- logue distribution of PCTs, the results obtained for two commercial formulations, Aroclor 5460 and Leromoll 141, are given in Table 4. The percentages of the homologue distribution obtained using HRMS and MS–MS are very similar, which ensures that the interference of the [M⫺2Cl]⫹fragment ion with [M]⫹• of each homologue is minimized by MS–MS. Consequently, HRGC–MS–MS in SRM mode can be considered as a useful and relatively inexpensive technique for the determination of the homologue distribution of PCTs. This technique has also been successfully applied to the identification and quantification of PCTs in fish samples, in which relatively low levels of PCTs have been detected [28].
Sensitivity greater by a factor of about 3 than that obtained under the best HRMS conditions was achieved.
Electron-ionization tandem mass spectrometry (EI-MS–MS) has also been applied for the analysis of toxaphene using SRM [107,108]. This technique has been used for the identification, characterization, and structural elucidation of polychlorobornanes in technical toxaphene and marine mammals. Buser et al.
[107] reported detailed fragmentation pathways for two toxaphene congeners iso- lated from marine mammals, Parlar No. 26 and 50. These authors concluded that the most important fragmentation pathway of these two CHB congeners involves the elimination of C2H2Cl2and C2HCl3from [M⫺HCl]⫹•ions through a retro–
Diels-Alder mechanism. The transition corresponding to the loss of C2H2Cl2can be considered specific for these compounds. In Figure 8, the SRM chromatograms of technical toxaphene corresponding to the retro–Diels-Alder transitions (a) 374⫹ → 244⫹ and (b) 374⫹ → 278⫹ for octachlorobornanes, and (c) 408⫹ → 278⫹ and (d) 408⫹→ 312⫹for nonachlorobornanes are shown as an example.
As can be seen, TOX9 (Parlar No. 50) can be identified as a major component in technical toxaphene, whereas TOX8 (Parlar No. 26) is only present as a minor component. In addition, low responses for these congeners were obtained when the loss of C2HCl3 was registered. Different fragmentation pathways for toxa- phene have been studied [108]: the loss of Cl and HCl, the retro–Diels-Alder mechanism, and the loss of CHCl2 and CH2Cl. From these fragmentation reac- tions, it can be deduced that higher selectivity and sensitivity can be obtained using both the retro–Diels-Alder and the loss of CHCl2and CH2Cl for the analy-
Figure 7 MS–MS spectra of a heptachloroterphenyl of Aroclor 5460. (a) Precursor- ion spectrum of the molecular ion m/z 470. (b) Product-ion spectrum of the ion m/z 470.
(From Ref. 29.)
146GalceranandSantos
Table 4 Homologue Distribution for Aroclor 5460 and Leromoll 141 Obtained by HRGC–EI-LRMS-SIM (Resolving Power 1,500), HRGC–EI-HRMS (Resolving Power 35,000) and HRGC–EI-MS–MS-MRM
Percentage of Homologues
Aroclor 5460 Leromoll 141
Homologues LRMS-SIM HRMS-SIM MS–MS-MRM LRMS-SIM HRMS-SIM MS–MS-MRM
Penta-CTs NDa ND ND ND 12.0 11
Hexa-CTs 9 0.9 1 20 30.2 32
Hepta-CTs 11 10.4 11 33 42.6 44
Octa-CTs 25 45.7 44 39 14.1 12
Nona-CTs 34 29.4 29 8 0.7 0.6
Deca-CTs 18 11.9 13 0.2 0.4 ND
Undeca-CTs 3 1.2 2 ND ND ND
Dodeca-CTs ND ND ⬍0.5 ND ND ND
aND: not detected.
Source: Adapted from Refs. 28 and 29.
Figure 8 Selected reaction monitoring chromatograms of technical toxaphene. Retro–
Diels-Alder transitions from octachlorobornanes (a) 374⫹→244⫹(loss of C2HCl3), (b) 374⫹→278⫹(loss of C2H2Cl2), and from nonachlorobornanes (c) 408⫹→278⫹(loss of C2HCl3), and (d) 408⫹→312⫹(loss of C2H2Cl2). Absence of signal at retention times of TOX 8 (Parlar No. 26) and TOX 9 (Parlar No. 50) is marked by asterisk in chromato- grams a and c, respectively. Retention times indicated in min.). (From Ref. 107.)
148 Galceran and Santos
sis of homologue groups in technical toxaphene, while the specificity of the loss of Cl and HCl was very small.
Recently, ion-trap–based EI-MS–MS has been used for the determination of toxaphene in biological samples [109]. Ion trap MS offers high selectivity and sensitivity at relatively low cost. In contrast with hybrid or triple-stage quadrupole instruments, where the isolation of precursor ions and further dissociation takes place in space, ion-trap MS–MS takes place in time. Consequently, there are no transmission losses and, hence, it provides better sensitivity. For use in ion-trap MS–MS, Chan et al. [109] proposed the ion at m/z 125 as a precursor ion for toxaphene analysis. This corresponds to a chlorinated monochlorotropylium structure. Polychlorinated biphenyls and other organochlorine compounds also produce the fragment ion at m/z 125 in the ion-trap MS–MS mode, but these compounds do not produce the product ion m/z 89, which corresponds to totally dechlorinated toxaphene. Therefore, the transition m/z 125 to m/z 89 can be used for quantification of total toxaphene and individual toxaphene congeners [109].
Using this method, instrumental detection limits between 1 and 3 pgàl⫺1for the individual congeners Parlar No. 26, 32, 50, and 62 and 280 pgàl⫺1for the toxa- phene technical mixture have been achieved. Coefficients of variation of 14%
were obtained. Although ion-trap MS–MS allows good results in terms of accu- racy and precision, some interferences from organochlorine pesticides have been detected. Nevertheless, these interferences seem not to be important in environ- mental and biological samples.