2. HIGH-RESOLUTION GAS CHROMATOGRAPHY–LOW-
2.2. Electron-Capture Negative-Ion Mass Spectrometry
The technique of HRGC–ECNI-MS has also been applied for the analysis of PCTs and toxaphene. Although this technique is the most popular method for the analysis of toxaphene, to our knowledge, it has only been used by Canton and Grimalt [15] and Wester et al. [21] for the analysis of PCTs. In general, the usefulness of ECNI-MS in the analysis of halogenated compounds is due to its higher sensitivity and specificity compared with EI-MS. The mass spectra ob- tained with ECNI-MS for PCT and toxaphene congeners are characterized by a low fragmentation pattern using methane as a moderating gas. For example, in- tense signals of the molecular [M]⫺•cluster ions and weak signals of the [M⫺ Cl⫹H]⫺and [M⫺2Cl⫹2H]⫺cluster ions are obtained for PCT congeners, as can be seen in Figure 3, whereas for toxaphene, ECNI produces intense [M⫺ Cl]⫺ions for CHBs with 7 or more chlorine atoms, and both [M]⫺• and [M⫺ Cl]⫺ions for lower-chlorinated homologues [31,62,72]. For the PCT quantitative analysis, Wester et al. [21] studied various methods to obtain unambiguous data.
They suggested the use of the total ion current (TIC) chromatogram of [M]⫺• and [M ⫹ 2]⫺• ions of the molecular cluster for each homologue group from trichloroterphenyls to undecachloroterphenyls. However, low-chlorinated terphe- nyls (ⱕ4 chlorine atoms) show low sensitivity and, in addition, other polychlori- nated compounds, such as PCBs, polychlorinated naphthalenes (PCNs), and some organochlorine pesticides, can cause false positives. Additionally, the authors emphasize the importance of the proper selection of the standard mixture, espe- cially because of the large differences in response factors, which give higher errors in HRGC–ECNI-MS than in HRGC–EI-MS or HRGC–ECD. In unfavor- able conditions, this error can be as high as 500%. Nevertheless, the sensitivity achieved with HRGC–ECNI-MS is 5- to 10-fold higher than HRGC–ECD or HRGC–ELCD with a repeatability of 15%.
On the other hand, while HRGC–ECNI-MS is highly sensitive for the anal- ysis of toxaphene, some disadvantages have been reported, such as low reproduc- ibility and variations of the response factors with ECNI conditions (ion source temperature, pressure of the moderating gas, and instrument design) [31,62]. In addition, a wide range of response factors in congeners has been reported [36,42,64]. The low responses of some CHB congeners can produce false nega- tives, and, therefore, the reliability of the quantitative determination of toxaphene cannot be assured. For instance, Lau et al. [84] reported that the nonachlorobor- nane congener, Parlar No. 62, did not give the expected [M⫺ Cl]⫺(m/z 413)
Analysis of Chlorinated Organic Compounds 129
Figure 3 ECNI-mass spectra of (a) a pentachloroterphenyl and (b) a decachloroterphe- nyl. (From Ref. 21. Copyright 1996 American Chemical Society.)
and only a weak signal of the [M⫺HCl⫺Cl]⫺ion was observed. In order to increase the intensity of the [M⫺Cl]⫺ion of Parlar No. 62, Swackhamer et al.
[62] recommend a decrease in the ion source temperature to approximately 100°C. Some authors [85] indicate that a stereoisomeric effect is responsible for the different responses while others [31,42,72] suggest that the absence of chlo- rine atoms in C-3 and C-6 positions of the six-membering ring can explain the poor response of some congeners. Nevertheless, the lack of response of these congeners prevents their identification and determination in environmental and biological samples because low or null response is obtained using ECNI-MS. This is an important problem because Parlar No. 26, 50, and 62 have been proposed to be used to monitor CHBs residues in fish and foodstuffs [78].
Polychlorinated biphenyls are a potential interference in the toxaphene analysis using HRGC–ECNI-LRMS in SIM mode [60,84–87]. The presence of a small amount of oxygen in the ion source produces [M⫺Cl⫹O]⫺•and [M⫺ H ⫹ O]⫺• fragment ions from hexa- and heptachlorobiphenyls with the same m/z as hepta- and nonachlorobornanes. This interference can be solved by using HRMS. Although HRGC–ECNI-MS has some drawbacks, it has been exten-
130GalceranandSantos Table 3 Analysis of Toxaphene Using HRGC–ECNI-MS
Sample GC–MS conditions Quanfication Comments Concentration Reference
Fish (salmon and HRGC–ECNI-MS-SIM Congener-specific LOD: 0.3–7.0 pg in- (1) 62–5,020àg kg⫺1 41 cod liver) oil (1) MS: Hewlett-Packard method: (chlorobor- jected of pure conge- wet weight
Red fish (2) 5988A MS nane derivatives stan- ner standard (2) 53àg kg⫺1wet
Trout (3) ECNI: methane 2⫻ dards) Linear over 4 orders of weight
Halibut (4) 10⫺4Torr magnitude (3) 14àg kg⫺1wet
Caviar (5) 100°C ion source tem- weight
perature (4) 21àg kg⫺1wet
200àA emission cur- weight
rent (5) 932–1,070àg kg⫺1
wet weight
Fish and fish prod- HRGC–ECNI-MS-SIM Congener-specific LOD: 0.3–7 pg in- 4.65–1,394 pg kg⫺1as 42 ucts (caviar, sea MS: Hewlett-Packard method:⌺(Parlar jected ⌺(Parlar No. 26,50
eel, cod fish, cod 5988A MS No. 26, 50 and 62) and 62)
liver, salmon oil) ECNI: methane 100°C ion source tem-
perature
200àA emission cur- rent
Fish (1) HRGC–ECNI-MS-SIM Homologue specific CV(%): 22% (1) 1–19,000àg kg⫺1 53
Milk (2) ECNI: methane, 1 Torr method: wet weight as total
100°C ion source tem- Internal standard: toxaphene
perature 1,2,3,4-tetrachlo- (2) 0.3–7.6 ng kg⫺1
ronaphthalene lipid weight as total
toxaphene
AnalysisofChlorinatedOrganicCompounds131 Air HRGC–ECNI-MS-SIM Congener-specific High variability of the 0.9–10.1 pg/m3 54
MS: Hewlett-Packard method: response factor 5989A MS-Engine Standards: 22 conge- (CV(%): 177%) ECNI: methane 1.2 ners, 11 congeners
Torr and 5 congeners
150°C ion source tem- Internal standards:
perature 1,3,5-tribromoben- 230 eV, 300 mA emis- zene and octachloro-
sion current naphthalene
Lake trout (1) HRGC–ECNI-MS-SIM Homologue specific CV(%): 28% (1) 4,400–15,000 ng 57 Whitefish (2) MS: Hewlett-Packard method: LOD: 1ng g⫺1lipid g⫺1lipid
5985B MS (2) 4,500–10,000 ng
ECNI: methane, 0.35 g⫺1lipid
Torr
100°C ion source tem- perature
Water (1) HRGC–ECNI-MS-SIM Homologue specific CV(%): 19–22% (1) 0.17–1.12 ng L⫺1 58
Sediment (2) MS: : Hewlett-Packard method: (2) 15 ng g⫺1dry
Phytoplankton, net 5988A MS Internal standard: PCB- weight
zooplankton (3) ECNI: methane 1.0 204 (3) 50–250 ng g⫺1dry
Lake trout (4) Torr weight
100°C ion source tem- (4) 4.3–10.3àg g⫺1
perature lipid
Fish (troup, white- HRGC–ECNI-MS-SIM Homologue specific LOD: 75 pg injected 220–510 ng g⫺1of 62 fish, carp) MS: Hewlett-Packard method:⌺hepta-, Linear response over 4 whole fish
5985B MS octa-, nona, deca- orders of magnitude ECNI: methane, 0.35 CHBs
Torr Internal standard: PCB- 100°C ion source tem- 204
perature
200àA emission cur- rent, 200 eV
132GalceranandSantos
Table 3 Continued
Sample GC–MS conditions Quanfication Comments Concentration Reference
Fish: HRGC–ECNI-MS-SIM Congener-specific (1) 1–33.6àg kg⫺1wet 63
(1) herring, MS: Finnigan SSQ70 method: Parlar No. weight
(2) mackrel, halibut MS 26,50 and 62 (2) 2.3–13.5àg
redfish, sardines ECNI: methane, 0.5 kg⫺1wet weight
(3) salmon Torr (3) 6.4–19.2àg kg⫺1
100°C ion source tem- wet weight
perature 90 eV
Fish and fish HRGC–ECNI-MS-SIM Congener-specific 12–755àg kg⫺1as⌺ 64
products MS: Hewlett-Packard method: Parlar No. (Parlar No. 26,
5988A MS 26, 32, 50, 62 and 50,62)
ECNI: methane 69
100°C ion source tem- perature
200àA emission cur- rent
Lake Superior food HRGC–ECNI-MS-SIM Homologue specific LOD: 92 ng of total (1) 21–1,360 ng g⫺1 59
web (planktonic method:⌺(hepta-, toxaphene wet weight
organisms) (1) octa-, nona-CHBs) (2) 12–24 ng g⫺1wet
Lake trout (2) Internal standard: PCB- weight
204
AnalysisofChlorinatedOrganicCompounds133 Soil HRGC–ECNI-MS-SIM Homologue specific Precision CV%⫽10% 16–69àg kg⫺1of soil 60
MS: Finigan-MAT method: Reproducibility CV%
4021 Internal standard: PCB- ⫽10%
ECNI: methane, 0.4 204 Linearity range: 0.1–1
Torr mg kg⫺1
170°C ion source tem- perature
70 eV, 500àA emis- sion current
Sediment HRGC–ECNI-MS-SIM Homologue specific LOD: 3–6 ng injected 33–15 ng g⫺1of sedi- 61 MS: Hewlett-Packard method:⌺(hexa-, Precision, CV%⫽ ment
5988A MS hepta-, octa-, nona- 22%
ECNI: methane, 1 Torr CHBs).
100°C Ion source tem- Internal standard: PCB-
perature 204
Cod liver oil NIST HRGC–ECNI-MS Congener-specific CV(%): 20% Parlar No. 26: 375àg 78
SRM 1588 MS: Finnigan 4500- method: Parlar No. kg⫺1of fat
MS 26, 50, 62 and 32 Parlar No. 50: 434àg
ECNI: methane, 0.5 kg⫺1of fat
Torr Parlar No. 62: 220àg
100°C ion source tem- kg⫺1of fat
perature, 90 eV
Whole blood HRGC–ECNI-MS-SIM Homologue specific Total toxaphene: 162– 86
MS: Hewlett-Packard method 174 ng L⫺1
5989A MS-Engine Congener-specific T2 and T12 is the 90%
ECNI: methane, 1.9 method: T2 (Parlar of the total toxa-
bar No. 26) and T12 phene content
120°C ion source tem- (Parlar No. 50) perature, 200 eV Internal standard:13C-
PCB-153 and PCB- 209
134GalceranandSantos
Table 3 Continued
Sample GC–MS conditions Quanfication Comments Concentration Reference
Fish HRGC–ECNI-MS-SIM Homologue specific 0.1–7 mg kg⫺1lipid 87
Herring oil MS: Hewlett-Packard method 5988A MS
ECNI: methane, 1 Torr 120°C ion source tem-
perature
Beluga whale HRGC–ECNI-MS-SIM Congener-specific (1a) 602–1300 ng g⫺1 88
blubber (1) MS: VG-7070E-HF method: T2 (a) and of lipid
Arctic char, whole double-focusing MS T12 (b) congeners (1b) 1050–2350 ng g⫺1
fish (2) ECNI: methane of lipid
Milk (3) 175°C ion source tem- (2a) 43.6–73.9 ng g⫺1
perature of lipid
100 eV, 200àA emis- (2b) 83.2–137.9 ng g⫺1
sion current of lipid
(3a) 69.5 ng g⫺1of lipid
(3b) 151 ng g⫺1of lipid
AnalysisofChlorinatedOrganicCompounds135
Seal blubber HRGC–ECNI-MS-SIM Congener-specific TOX8: 5.3àg kg⫺1 89
MS: Hewlett-Packard method:⌺(Parlar TOX9: 3.0àg kg⫺1
5989B MS No. 26 (TOX8) and Total toxaphene as⌺
ECNI: methane, 1.2 62 (TOX9)) (TOX8, TOX9): 8.3
bar àg kg⫺1
200°C ion source tem- perature
Fish HRGC–ECNI-MS-SIM Homologue specific LOD:⬍0.3 ng g⫺1as 1.6 ng g⫺1neuston 91 MS: Finigan TSQ-70 method total toxaphene 12.8 ng g⫺1pollack
MS
ECNI: Methane 9,400 mTorr
Soil HRGC–ECNI-MS-SIM Homologue specific 668–1950 mg kg⫺1 92
MS: Hewlett-Packard method:
5989A MS-Engine Internal standard: PCB- ECNI: methane, 0.4 107
Torr
150°C ion source tem- perature
136 Galceran and Santos
sively used for the analysis of toxaphene using SIM mode. The [M]⫺•ions of hexachlorobornane/bornene (m/z 340 and 342, respectively) and the [M⫺Cl]⫺ from hepta- to decachlorobornanes/bornenes (m/z 374 and 376, m/z 408 and 410, m/z 442 and 444, and m/z 476 and 478, respectively) have been commonly used for quantitative analysis. This method was proposed by Swackhamer et al.
[57,62] and it has been used by several researchers for the determination of total toxaphene as well as the specific CHB congeners in environmental and biota samples [31,42,86–92]. In addition, some other ions have also been used in order to correct the above-mentioned PCBs interferences. Two approaches have been proposed for quantitative analysis. The first is based on the use of technical toxa- phene as standard and provides a rough estimation of the total toxaphene concen- tration [68] and the homologue distribution of the CHB. The uncertainty in this determination may be higher than using individual compounds, particularly in samples from biota with a high biotransformation rate for toxaphene, such as marine mammals [93] and human milk [53].
The second approach for toxaphene analysis is the congener-specific method. Nevertheless, HRGC, which offers high selectivity, or multidimensional GC [68] is required when coelution with other organochlorine compounds occurs.
A summary of some selected literature data is provided in Table 3, where some HRGC–ECNI-MS methods in SIM mode are given. Detection limits for the individual congeners (Parlar No. 26, 50, and 62] between 0.3 and 7 pg injected are achieved and a wide range of toxaphene concentrations in environmental and biological samples is found. Some authors used the technical toxaphene as stan- dard to estimate total concentration [58,90–92]. Recently, the individual determi- nations of Parlar No. 26 and 50 for marine mammals [63,88] and Parlar No. 26, 50, and 62 for fish and fish products [42,63,64,70] have been used as indicators of the level of toxaphene contamination.