3. HIGH-RESOLUTION GAS CHROMATOGRAPHY–HIGH-
3.1. Electron-Ionization High-Resolution Mass
High-resolution gas chromatography–electron-ionization high-resolution mass spectrometry (HRGC–EI-HRMS) operated in SIM mode has been used for the analysis of PCTs and toxaphene [23,28,80]. This technique allows high selectiv- ity in the analysis of PCTs by eliminating the internal interferences in the mea- surement of the molecular cluster ions of each PCT homologue group. For toxa- phene analysis, a higher specificity on the CHB compounds than LRMS and unambiguous determination of the individual congeners in environmental and biota samples have been achieved.
The analysis of PCTs using EI-LRMS presents some interferences due to the [M⫺Cl2]⫹ fragment ions of the homologues with two additional chlorine atoms, which can be overcome using HRMS [28]. A scheme of the contribution of the nonachloroterphenyl molecular ion to the selected heptachloroterphenyl molecular ion is given in Figure 4, as example of the high resolution needed.
The loss of two chlorine atoms gives the same [M⫺2Cl]⫹cluster fragment that interferes with the molecular cluster ion of octachloroterphenyls. In order to avoid this interference, a resolution of 25,300 is required. In general, for the complete elimination of all interferences, a resolution higher than 34,600 is needed [28].
Figure 4 Interferent mass of nonachloroterphenyls on the heptachloroterphenyls moni- tored molecular mass for Aroclor 5460. (From Ref. 28.)
138 Galceran and Santos
The HRGC–EI-LRMS SIM profile for the heptachloroterphenyls and the corre- sponding HRGC–EI-HRMS SIM profiles for the hepta- and nonachloroterphe- nyls after the loss of two chlorine atoms (m/z 469.8152) are shown in Figure 5.
The contribution of the [M⫺2Cl]⫹ion of nonachloroterphenyls on the molecular ion of heptachloroterphenyls is completely eliminated at a resolving power of 35,000. Under these conditions, the homologue distributions of PCT mixtures can be determined. For example, the values obtained for Aroclor 5460 and Lero- moll 141 [28] indicate that Aroclor 5460 is mainly composed of terphenyls with 7 (10.4%), 8 (45.7%), 9 (29.4%), and 10 (11.9%) chlorine atoms with the largest group being octacloroterphenyls, whereas Leromoll 141 is composed by penta- (12.0%), hexa- (30.2%), hepta- (42.6%), and octachloroterphenyls (14.1%).
These homologue distributions could be considered as reference values for the determination of the composition of these commercial formulations. The tech-
Figure 5 HRGC–EI-selected ion monitoring using (a) LRMS for heptachloroterphe- nyls, m/z 469.8, (b) HRMS for heptachloroterphenyls, m/z 469.8338, and (c) nonachloro- terphenyls after loss of two chlorine atoms, m/z 469.8152. (From Ref. 28.)
Analysis of Chlorinated Organic Compounds 139
nique of HRGC–HRMS has been used in the analysis of PCTs by only a small number of authors [23,28].
Electron-ionization high-resolution mass spectrometry has also been ap- plied for the analysis of toxaphene to prevent the potential interferences of other organochlorine compounds. Three different approaches have been used for the analysis of these compounds. The first is based on the monitoring of the m/z 158.9768 and the isotope peaks at m/z 160.9739 and 162.9709, which correspond to the dichlorotropylium ion structure (C7H5Cl2⫹) formed through successive de- chlorinations and rearrangements of CHBs. The use of these ions in combination with HRMS at a resolving power higher than 10,000 was firstly proposed by Saleh [37] in 1983. It has been used by several laboratories [95–97] with relatively good results. The correct isotope ratios of 100 : 65 : 11% for m/z 158.9768, 160.9739, and 162.9709, respectively, give an unambiguous identification of CHBs. This method has been applied in the determination of toxaphene in biota samples [95,96] and limits of detection lower than 10àg g⫺1for total toxaphene [97] and 0.2 ng g⫺1for specific single congeners (2 pg injected) [98,99] have been reported.
The characteristic ion at m/z 159 can be considered as a universal ion for all toxaphene congeners regardless of their degree of unsaturation and chlorination.
Therefore, chlorobornenes and chlorobornadienes can be determined in a single injection. Total toxaphene concentration can be calculated by comparing the total areas of samples and technical toxaphene standard. One of the advantages of EI- HRMS is that these ions (159/161) are not affected by the presence of common environmental pollutants such as PCBs and organochlorine pesticides, although other compounds, such as chlorinated pinenes, camphenes, and compounds with a similar structure to toxaphene, can interfere [46]. In addition, since the m/z 159⁄161 ions are end products from a series of fragmentation pathways, the iden- tity of the original compound is lost and different homologue groups cannot be distinguished.
The second approach proposed in the literature for the analysis of toxa- phene by EI-HRMS is based on the determination of the homologue composition.
Santos et al. [80], using SIM mode, studied the two most intense ions of the [M⫺ Cl]⫹, [M⫺HCl]⫹•, and [M⫺Cl⫺ HCl]⫹clusters for each homologue group (between Cl6and Cl10). From these profiles, they concluded that the [M⫺ Cl]⫹cluster ions had the maximum intensity. The homologue profiles for toxa- phene at different resolving powers are shown in Figure 6 as an example The interferences of homologues with an additional chlorine atom are not observed at a resolving power of 20,000, and at 10,000 most of the interferences are elimi- nated. By integrating the EI-HRMS SIM profiles for each homologue group us- ing [M ⫺ Cl]⫹ions (hexa-CHBs, m/z 308.9352; hepta-CHBs, m/z 342.8963;
octa-CHBs, m/z 376.8573, nona-CHBs, m/z 410.8173; and deca-CHBs, m/z 444.7793), the homologue distribution of commercial toxaphene can be calcu- lated. Using this approach, it is deduced that toxaphene was mainly composed
Figure 6 HRGC–EI-MS-selected ion monitoring of [M⫺Cl]⫹ions for the homologues of toxaphene at resolving power of (a) 1,000, (b) 10,000, and (c) 20,000. (From Ref. 80.)
Analysis of Chlorinated Organic Compounds 141
of CHBs with 7 chlorine atoms, with smaller amounts of 6 and 8, and very low amounts of 9 and 10. This homologue distribution of toxaphene was distinctly different from the one obtained by HRGC–ECNI-MS [62,91], and these differ- ences might be due to the MS techniques used. The ECNI-MS gives variable responses for CHBs, even for congeners of the same homologue group, and it is highly affected by ECNI conditions. In contrast, the EI-HRMS technique gives very similar responses for each homologue group and eliminates interferences among them. Therefore, this technique can be considered as reference for the determination of toxaphene homologue composition. The HRGC–EI-HRMS in SIM mode has been also applied to the quantification of CHB congeners in com- mercial toxaphene mixtures, and the detection limits obtained with this technique were between 5 and 9 pg injected [80].
Finally, a third method has been proposed for the analysis of toxaphene using EI-HRMS in SIM mode [98,99]. This method is based on the congener- specific analysis for Parlar No. 26, 50, and 62, which are considered as indicators of toxaphene contamination. Selective ion monitoring at a resolving power of 10,000 of mass fragments at m/z 340.8806 ([M⫺Cl⫺HCl]⫹cluster ion) for Parlar No. 26 and at m/z 338.8649 and 340.8620 ([M⫺ Cl⫺ 2HCl]⫹ cluster
142 Galceran and Santos
ions) for Parlar No. 50 and 62 were used. Good results were obtained with this method for the analysis of these three specific congeners in herring allowing low detection limits (0.2 ng g⫺1wet weight) [98].