Fast Quantification of Chlorinated Paraffins in Environmental Samples

Feb 10, 2015 - Chlorinated paraffins (CPs) are high production volume chemicals, but data about their environmental fate are scarce. CP mixtures compo...
0 downloads 11 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Fast Quantification of Chlorinated Paraffins in Environmental Samples by Direct Injection High-Resolution Mass Spectrometry with Pattern Deconvolution Christian Bogdal,*,† Tomas Alsberg,‡ Pascal S. Diefenbacher,†,∥ Matthew MacLeod,‡ and Urs Berger‡,§ †

Institute for Chemical and Bioengineering, ETH Zurich, Vladimir-Prelog-Weg 1, CH-8093 Zurich, Switzerland Department of Environmental Science and Analytical Chemistry (ACES), Stockholm University, SE-106 91 Stockholm, Sweden ∥ Empa, Swiss Federal Laboratories for Materials Science and Technology, Laboratory for Analytical Chemistry, Ü berlandstrasse 129, CH-8600 Dübendorf, Switzerland ‡

S Supporting Information *

ABSTRACT: Chlorinated paraffins (CPs) are high production volume chemicals, but data about their environmental fate are scarce. CP mixtures composed of thousands of isomers represent a major challenge for quantification at low levels in environmental samples. Here, we present a novel analytical method for analysis of short-chain, medium-chain, and long-chain CPs in a single injection, that also yields information about congener group pattern. Our detection method is based on direct injection into an atmospheric pressure chemical ionization source operated in negative ion mode under chlorine-enhanced conditions, followed by quadrupole time-of-flight high-resolution mass spectrometry (APCI-qTOF-HRMS) operated in full-scan mode. A mathematical algorithm is applied to deconvolute the CP patterns in the analyzed samples into a linear combination of patterns of technical CP mixtures and to quantify CPs using technical mixtures as external calibration standards. For CP mixtures with known composition, the new method provided concentrations that were within a factor of 1.2 of the target value. Accuracies for CPs spiked to sediment and fish extracts were between 91% and 123%. Concentrations determined in unspiked field samples were within a factor of 5 for short-chain CPs and a factor of 16 for medium-chain CPs of results obtained with an independent method based on gas chromatography/electron capture negative ionization high-resolution mass spectrometry (GC/ECNI-HRMS). The presented APCI-qTOF-HRMS pattern deconvolution method is an interesting alternative for CP analysis in environmental samples. It is particularly sensitive for medium- and longchain CPs and has the advantage of being extremely fast (instrumental analysis time, less than 1 min).

C

CP mixtures and the associated separation difficulties represent major challenges for CP quantification, especially in environmental samples. In most studies, gas chromatography coupled to electron capture negative ionization low-resolution mass spectrometry (GC/ECNI-LRMS)6−8 was applied. Because mass interferences from different CP congener groups represent an additional challenge for LRMS-based methods (e.g., 12C141H2135Cl837Cl1 (505.8810 Da) interferes with 12 C161H2635Cl437Cl4 (505.9425 Da)), sector field high-resolution mass spectrometry (HRMS) was also applied.9,10 GC/ ECNI-MS methods, particularly with LRMS, however, require rigorous sample cleanup to avoid interferences from other halogenated compounds. Additionally, GC/ECNI-MS methods have the disadvantage that the instrument response is dependent on the chlorination degree of the CPs, making it

hlorinated paraffins (CPs), also referred to as polychlorinated n-alkanes, are a family of industrial chemicals produced by chlorination of an n-alkane feedstock. Technical CP formulations contain thousands of isomers and are grouped into chain lengths of C10−C13 (short-chain CPs, referred to as SCCPs), C14−C17 (medium-chain CPs, MCCPs), and C≥18 (long-chain CPs, LCCPs). The degree of chlorination of the alkane backbone varies but is usually between 40%Cl and 70%Cl by weight.1 CPs are mainly used as additives in cutting oils and lubricants and as secondary plasticizers and flame retardants in polymeric materials.2 Global production of CPs has increased during the last decades; it was 600 000 t/a in 20071 and exceeded 1 million t/a in 2009.3 SCCPs are currently listed as candidates for possible inclusion in the Stockholm Convention on Persistent Organic Pollutants (POPs)4 and have been listed as a substance of very high concern by the European Chemicals Agency.5 However, scientific studies on the occurrence and fate of CPs in the environment are still rare. One reason is that the complexity of © XXXX American Chemical Society

Received: November 18, 2014 Accepted: January 26, 2015

A

DOI: 10.1021/ac504444d Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

CP formulations. These extracts were measured before and after spiking, to subtract the amount of CPs detected prior to spiking from the spiked extracts. Details about CP formulations and the preparation of standard mixtures are provided in the Supporting Information. To test the performance of the deconvolution procedure with real field samples and to perform an intercomparison of analytical methods, a set of field samples was analyzed for CPs. For the deconvolution of SCCP and MCCP patterns in field samples, three commercially available SCCP standard mixtures and three MCCP standard mixtures supplied by Dr. Ehrenstorfer GmbH (Augsburg, Germany) were used, including: (1) SCCPs with 51.5%Cl (denoted hereafter as Standard SCCP 51.5%Cl), (2) SCCPs with 55.5%Cl (Standard SCCP 55.5%Cl), (3) SCCPs with 63.0%Cl (Standard SCCP 63%Cl), (4) MCCPs with 42.0%Cl (Standard MCCP 42%Cl), (5) MCCPs with 52.0%Cl (Standard MCCP 52%Cl), and (6) MCCPs with 57.0%Cl (Standard MCCP 57%Cl). For the deconvolution of LCCP patterns in field samples, no standard mixtures were available and, therefore, the technical LCCP formulations LCCP 40%Cl, LCCP 49%Cl, and LCCP 70%Cl were used. The same field samples were also analyzed for SCCPs and MCCPs in an independent laboratory using GC/ ECNI-HRMS, as presented in detail in the Supporting Information. The set of analyzed field samples included seven sewage sludge samples from municipal wastewater treatment plants around Zurich, Switzerland, from 2007 and two urban air samples from Zurich, Switzerland, from summer 2012. The preparation of the field sample extracts including Soxhlet extraction and chromatographic cleanup is described elsewhere23,24 and is also summarized in the Supporting Information. Particular care was taken to avoid background contamination of the field samples with CPs during the sample preparation procedure. However, sample preparation was the same for both analytical methods, so contamination of the samples during sample preparation would not impede their comparison. Instrumental blanks were checked by injecting pure acetonitrile into the APCI-qTOF-HRMS system; only noise was observed for pure solvent injections. Chemical Analytical Method. Samples were directly injected (5 μL in acetonitrile) into the qTOF-HRMS via flow injections without chromatographic separation using an autoinjector for an ultra-performance liquid chromatography (UPLC) system (Waters, Milford, USA). The eluent consisted of acetonitrile with a flow rate of 150 μL/min and the total acquisition time per sample was 1 min. Dichloromethane was added to the eluent between the injector and the ion source with a syringe pump at a flow of 15 μL/min using a Tconnection. The addition of dichloromethane has earlier been shown to form a plasma of Cl− ions in the ion source that significantly enhances the formation of [M + Cl]− adduct ions.25 These so-called chlorine-enhanced ionization conditions suppress the generation of multiple fragment ions other than [M + Cl]− and consequently improve the selectivity and sensitivity of CP detection with negative ion chemical ionization.20,26 The settings of the APCI source operated in negative ion mode are provided in the Supporting Information. The qTOFHRMS (QTOF Premier, Waters, Manchester, UK) was operated in V reflection mode at a resolution of 10 000. For each congener group, the two most abundant m/z signals of the [M + Cl]− isotope cluster were extracted from the full-scan mass spectra (scan range m/z 250−1420) using the TargetLynx

crucial to select an adequate CP mixture as reference standard for quantification.7−11 Alternative quantification procedures based on external standards of CP mixtures have been explored to correct for the relationship observed between ECNI response factor and chlorine content of injected CPs.8 Two dimensional gas-chromatography (GCxGC) methods have been developed, but complete separation of CPs has so far not been achieved.12,13 Nilsson et al.14 applied a principal component analysis method to identify different CP formulations in samples. Geiß et al.15 developed a multiple linear regression procedure for the quantification of SCCPs. Faster methods with lower complexity have also been presented. These analysis techniques include GC coupled to electron ionization tandem mass spectrometry (GC/EI-MS/ MS),16 short GC-column ECNI-LRMS,17 and dechlorination− hydrogenation reaction GC.18,19 Each of these methods is an interesting alternative for rapid determination of the sum of SCCPs and MCCPs; however, each involves some trade-offs, particularly the loss of information about the chlorination degree. Finally, Zencak and Oehme20 presented high-performance liquid chromatography coupled to atmospheric pressure chemical ionization low-resolution mass spectrometry (HPLC/ APCI-LRMS) for CP analysis. The sensitivity of the chlorineenhanced APCI method was less dependent on the chlorination degree than established ECNI-MS methods. Furthermore, the APCI-LRMS method was also suitable for LCCPs. However, it was not applied to environmental samples, probably due to the limited sensitivity of HPLC-LRMS systems available at that time. Until now, this method has not been taken up in later studies. Here, we describe a fast analytical method to quantify SCCPs, MCCPs, and LCCPs that also provides information about the congener group patterns. We focus on CPs starting with five chlorine substituents, but the same method is also applicable to lower-chlorinated CPs. Our detection method is adopted from Zencak and Oehme20 and is based on chlorineenhanced direct-injection negative-ion APCI quadrupole timeof-flight high-resolution mass spectrometry (APCI-qTOFHRMS) operated in full-scan mode. SCCPs, MCCPs, and LCCPs can be analyzed in a single injection. A mathematical algorithm is used to deconvolute the CP patterns in the analyzed samples into a linear combination of patterns of technical CP formulations. The deconvoluted CP patterns in the samples can then be quantified using selected technical CP products as external calibration standards.

■

MATERIALS AND METHODS Technical CP Products and Environmental Sample Extracts. A set of 21 technical CP formulations was initially analyzed, and a subset of nine formulations was selected for method development in this study. The nine selected technical CP formulations included three SCCP, three MCCP, and three LCCP formulations that covered the whole range of chlorination degrees within each group. Hereafter, these technical CP formulations are denoted: (1) SCCP 49%Cl, (2) SCCP 60%Cl, (3) SCCP 70%Cl, (4) MCCP 45%Cl, (5) MCCP 50%Cl, (6) MCCP 56%Cl, (7) LCCP 40%Cl, (8) LCCP 49%Cl, and (9) LCCP 70%Cl. Synthetic binary and ternary SCCP, MCCP, and LCCP mixtures of these technical formulations were prepared to test the pattern deconvolution procedure. In a second stage, potential matrix interferences or ion suppression effects were investigated by spiking sediment21 and fish extracts22 with known amounts of these nine technical B

DOI: 10.1021/ac504444d Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 1. Principle and workflow of the pattern deconvolution and quantification procedure developed in this study. The individual steps numbered in the figure are explained in the text.

deconvolution could be performed separately for SCCPs, MCCPs, and LCCPs. Only congener groups with a signal intensity above their respective LOD were considered in the deconvolution procedure. In the illustrative example of Figure 1, the pattern of sample S is assumed to consist of four congener groups (CmClw, CnClx, CoCly, CpClz) in a ratio of 0.1:0.4:0.4:0.1, whereas real CP patterns consist of considerably more congener groups (30 congener groups consisting of 60 m/z values for SCCPs from C10Cl5 to C13Cl13; 46 congener groups consisting of 92 m/z values for MCCPs from C13Cl5 to C17Cl17; 185 congener groups consisting of 370 m/z values for LCCPs from C18Cl5 to C27Cl27). Analogously, the congener group pattern of technical CP formulations to be applied in the deconvolution procedure is derived (box 2 in Figure 1). In the example, the formulations are assumed to have a congener group composition of 0.1, 0.5, 0.2, and 0.2 for Y1 and of 0.2, 0.2, 0.5, and 0.1 for Y2. We set up a system of linear equations, which assumes that the CP congener pattern of the sample is a linear combination of the congener patterns of the technical products (box 3 in Figure 1). The contributions x1 and x2 of the technical formulation Y1 and Y2 to the pattern of S are the unknowns of the system. The equation system is over-

software of MassLynx 4.1 (Waters, Manchester, UK). A total of 522 m/z ratios corresponding to CP congener groups with chain lengths of C10−C27 and with chlorine substituents from five up to the number of carbons, i.e., from C10Cl5 to C27Cl27, were considered (Table S1 in the Supporting Information). The instrumental sensitivity for SCCPs, MCCPs, and LCCPs was assessed by seven replicate analyses of technical CP formulations (Supporting Information). The limit of detection (LOD) was derived from the standard deviation of the seven injections and the Student’s t-value at a 99% confidence level. Deconvolution and Quantification Procedure. The general principle and workflow of the novel deconvolution and quantification procedure is presented in Figure 1. The detailed mathematical algorithm is described in the Supporting Information. The congener group patterns of SCCPs, MCCPs, and LCCPs are derived on the basis of the high-resolution mass spectra recorded in full-scan mode for a sample (box 1 in Figure 1). Since the qTOF-HRMS resolved all of the selected 522 m/z values (minimum resolution of 7000 required for the separation of the two closest m/z values for C22H24Cl22 (1104.4537 Da) and C27H36Cl20 (1104.6129 Da)), the C

DOI: 10.1021/ac504444d Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 1. Calculated Chlorination Degree (Average of Ten Measurements ± Standard Deviation), Linear Range, and Limit of Detection for the Different Technical CP Formulations calculated chlorination degree SCCP 49%Cl SCCP 60%Cl SCCP 70%Cl MCCP 45%Cl MCCP 52%Cl MCCP 56%Cl LCCP 40%Cl LCCP 49%Cl LCCP 70%Cl a

59%Cl 62%Cl 67%Cl 53%Cl 56%Cl 57%Cl 45%Cl 52%Cl 69%Cl

(±1.0) (±0.4) (±0.1) (±0.5) (±0.1) (±0.1) (±0.4) (±0.1) (±0.2)

linear range [ng/μL]

limit of detection (LOD) [ng/μL]

relative response factora

1.2−500 0.50−500 0.10−500 0.20−25 0.10−25 0.10−25 0.02−2.0 0.02−2.0 0.03−2.0

1.2 0.50 0.10 0.20 0.10 0.08 0.02 0.02 0.03

1 19 51 35 51 125 396 822 593

Average response factors normalized to the average response factor of SCCP 49%Cl based on ten measurements.

The instrumental LOD was lowest for LCCPs, followed by MCCPs and SCCPs (Table 1). For SCCPs, the linear ranges of the detector covered at least 3 orders of magnitude (Table 1, Figure S7, Supporting Information). For MCCPs and LCCPs, the linear ranges covered approximatively 2 orders of magnitude. Deconvolution and Quantification of Synthetic CP Mixtures. The CP concentrations determined in the synthetic mixtures are close to the expected concentrations with an accuracy between 88% and 107% (Table S2, Supporting Information). The deviation between expected and calculated compositions is the largest for mixtures including lower chlorinated CPs. The coefficient of determination (R2) indicating the goodness of the fit of the linear regression model (i.e., a perfect fit would result in R2 = 1) is higher than 0.90 for all mixtures. The uncertainty of the calculated concentration of CPs in each synthetic mixture derived from the error propagation procedure (Supporting Information, equations 15, 18, and 19) is high for the lower chlorinated SCCP formulations; e.g., for the concentration of SCCP 49%Cl, the relative error is 200% in test 1 (18.7 ± 38.3 ng/μL). For the concentration of SCCP 60%Cl, the relative error is 10% (8.86 ± 0.872 ng/μL in test 1), and for the concentration of SCCP 70%Cl, the relative error is 1% (7.78 ± 0.099 ng/μL). For calculated concentrations of MCCPs, the uncertainty is also the highest for the lower chlorinated formulations (relative error 14% for MCCP 45%Cl) and lower for the higher chlorinated formulations (relative error 8% for MCCP 50%Cl and around 1% for MCCP 56%Cl). For LCCPs, the uncertainty of the calculated concentration is very small (relative error 0.92). The uncertainty of the calculated concentration of CPs is the highest for SCCP and MCCP formulations with a lower chlorine content (highest relative error is 47% for MCCP 45%Cl in the spiked fish sample) and low for CP formulations with a high chlorine content (relative error 0.91). For SCCPs, five sludge samples (samples SL1 to SL5) had a congener group pattern that could be deconvoluted with R2 > 0.88, whereas for SL6 and SL7, the deconvolution procedure yielded R2 of only 0.48 and 0.66. In these two samples, the measured SCCP pattern contained some noticeable congener group contributions that deviated from SCCP patterns observed in technical CP formulations. However, the individual mass trace signals of SCCPs did not provide indications for strong matrix effects. Also, Zencak and Oehme20 showed that the chlorine-enhanced APCI ionization method is relatively specific for CPs and has the important advantage of being hardly susceptible to interferences from matrix or other common environmental contaminants. Biologically mediated transformation of SCCPs in sewage sludge could be the reason for an altered SCCP pattern.29 Higher stability of MCCPs toward biodegradation compared to SCCPs could also explain why the measured MCCP patterns in sewage sludge samples were all similar to the CP patterns of technical CP formulations in contrast to some SCCP patterns. The absolute concentrations of SCCPs (135−581 ng/g dw) and MCCPs (1070−8960 ng/g dw) quantified in the sewage sludge samples are comparable to levels reported in sewage sludge from municipal wastewater treatment plants sampled in 2010 in Sweden (median of SCCPs: 1100 ng/g; median of MCCPs: 3800 ng/g)31 and in the early 2000s in Germany (range of SCCPs: 200−900 ng/g).32 For LCCPs, the measured congener group patterns in sewage sludge were dominated by the C18 group and could not be deconvoluted into the available technical CP formulations (Figure S12, Supporting Information). To successfully deconvolute the measured LCCP patterns, technical CP formulations with a clear dominance of the C18 group would be required, which are not known to exist. A more likely explanation for the observed LCCP patterns in sewage sludge is the presence of C18 congener groups within technical MCCP formulations (Supporting Information). In urban air, the congener group patterns of SCCPs and MCCPs appeared altered from technical CP patterns. The deconvolution procedure resulted in reconstructed patterns with R2 of 0.56 and 0.70 for SCCPs and 0.77 and 0.79 for MCCPs, which indicated that the pattern has been modified by environmental fate processes. The absolute concentrations of SCCPs (1.5−3.3 ng/m3) and MCCPs (1.3−26 ng/m3) quantified in the urban air samples are comparable to levels

graphic separation of CPs, which often require multiple injections for separate detection of SCCPs, MCCPs, and LCCPs.8−10,12,14,15,17,18 Response Factors and Patterns of Technical CP Formulations. The sensitivity of the APCI-qTOF-HRMS method depends on the chlorine content of the CPs, particularly for SCCPs. For SCCPs, the difference in sensitivity by a factor of 50 between the 49%Cl and the 70%Cl formulation is comparable to ECNI based methods.8 For MCCPs, the difference in response by a factor of 4 between the 45%Cl formulation and the 56%Cl formulation in our study is smaller than for ECNI based methods. The lower response factors for technical CP formulations with a low chlorine content is also reflected in the chlorination degree calculated on the basis of the APCI-qTOF-HRMS measurements (Table 1). For the SCCP 49%Cl, MCCP 45%Cl, and LCCP 40%Cl formulations, the chlorination degree determined by our method was higher than the manufacturer’s specifications. Limit of Detection and Linear Range. For SCCPs, our LOD (approximately 0.1−1.0 ng/μL) is comparable to GC/ ECNI-LRMS methods30 and also to the previous HPLC-APCILRMS method,20 whereas GC/ECNI-HRMS allows a detection in the low pg/μL range.10 For MCCPs, our LOD of 0.1 ng/μL is ten times lower than for HPLC-APCI-LRMS20 and in the same range as for GC/ECNI-HRMS.9 For LCCPs, our method is particularly sensitive with an LOD ten times lower than with the previously published HPLC-APCI-LRMS method.20 The linear range of more than 3 orders of magnitude obtained for SCCPs is larger than in the previous HPLC-APCILRMS study by Zencak and Oehme,20 where the linear range covered only 2 orders of magnitude. For MCCPs and LCCPs, both methods provided similar linear ranges of 2 orders of magnitude. Deconvolution and Quantification of Synthetic CP Mixtures. The APCI-qTOF-HRMS pattern deconvolution procedure provided results that are in good agreement with the known concentrations of technical CP formulations in nine synthetic mixtures of SCCPs, MCCPs, and LCCPs that we prepared (Table S2, Supporting Information). The pattern of the synthetic mixtures can also be adequately reconstructed with the deconvolution procedure, as shown by the good agreement between expected and calculated compositions, as well as by the coefficients of determination R2 > 0.90. The error propagation procedure (Supporting Information, equations 15, 18, and 19) shows that the uncertainty in the calculated CP concentrations is not influenced by the number of technical CP formulations considered in the deconvolution procedure but rather by the response factors of the formulations. Although the APCI-qTOF-HRMS analysis and pattern deconvolution procedure performed well for SCCPs (see tests 1, 4, and 5 in Table S2, Supporting Information), the uncertainty in the calculated concentrations is appreciable when SCCPs with a low degree of chlorination and, hence, a low response factor are included. The uncertainty of the calculated amount of CPs is dominated by the contribution of variability of the response factor (i.e., the second term including Rj2 in equations 15, 18, and 19, Supporting Information). Multiple injections are required to obtain a good estimate of the average response factor. Thus, for CPs with a low response factor, the APCIqTOF-HRMS analysis and deconvolution procedure is accurate (the calculated amounts are close to the expected values) but of limited precision (the uncertainty of the calculated amounts are large). Therefore, when samples are measured in replicates and G

DOI: 10.1021/ac504444d Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

already provide an estimation of the magnitude of CP concentrations in a sample. In regard of the urgent need for more reliable data about the occurrence and fate of CPs in the environment, the method we present here is of high relevance.

measured in ambient air in the Northern UK in spring 2003 (SCCPs 0.18−3.4 ng/m3; MCCPs 0.81−14 ng/m3).33 Comparison with Independent Measurements. For CPs, no analytical method can be considered as a reference standard method providing “correct” concentrations. For field samples with a complex matrix, such as sewage sludge, the difference of up to a factor of 4.4 for SCCP concentrations and 16 for MCCP concentrations between the APCI-qTOF-HRMS and the GC/ECNI-HRMS based concentrations is within deviations observed in previous interlaboratory comparison studies. Concentrations differing by more than an order of magnitude between laboratories have been reported for complex samples such as soil34 or fish.35,36 For the urban air samples with less matrix interferences but also with considerably lower concentrations of CPs in the injected extracts, the difference in concentrations of CPs by a factor of 0.7−2.8 between the two analytical methods is small. In the interlaboratory comparison study performed by Tomy et al.,35 the seven participating laboratories overestimated SCCPs in pure standard mixtures by a factor of 1.3−2.5. In the comparative study by Stevenson et al.37 including two independent laboratories, a similar difference by a factor of 1.4−2.5 was observed for SCCPs in indoor air samples. Also in the recent study by van der Veen et al.,38 most of the 15 participating laboratories overestimated SCCPs by up to a factor of 4.1. No interlaboratory study has been performed so far for MCCPs. The systematically lower concentrations of SCCPs and MCCPs obtained in our study with the APCI-qTOF-HRMS based method compared to the GC/ECNI-HRMS based method could be due to the presence of chemical interferences in the sample extracts for which the GC/ECNI-HRMS method is more vulnerable than the APCI-qTOF-HRMS method;39,40 whereas the APCI-qTOF-HRMS method is not particularly sensitive to chlorinated aromatic compounds.20 In future studies, the parallel analysis of synthetic mixtures of CPs and spiked samples with both analytical methods might provide additional information about the reason for the systematically lower concentrations of SCCPs and MCCPs obtained with the new analytical method. To assess the influence of the goodness of the pattern deconvolution on the calculated CP concentrations, a sensitivity analysis was performed and is described in detail in the Supporting Information. A reconstruction of the SCCP pattern with a R2 = 0.50 would result in a discrepancy of the quantified concentration by a factor of 4. Considering that in our study the two independent analytical methods also resulted in a difference in concentrations of SCCPs of approximatively a factor of 4, we suggest R2 = 0.50 as a cutoff for a meaningful quantification of SCCPs with the pattern deconvolution method. For MCCPs, R2 = 0.5 would result in a quantification error of maximum a factor of 2. For LCCPs, the difference in quantified concentrations was below a factor of 1.5, independent of the standard combination used.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional documentation about the analytical method, the deconvolution procedure, and the results. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +41 44 632 5951. Present Address §

U.B.: Helmholtz Centre for Environmental Research - UFZ, Department of Analytical Chemistry, Permoserstraße 15, DE04318 Leipzig, Germany. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Margot Reth and Zdenek Zencak for their contribution to the initial APCI-qTOF-HRMS method development. Special thanks go to Annika Jahnke and Markus Zennegg for providing and preparing the field sample extracts. Anders Borgen is acknowledged for providing test samples for the method development. The Swiss National Science Foundation is acknowledged for the travel grant provided to Christian Bogdal (Grant IZK0Z2_154108/1).

■

REFERENCES

(1) Fiedler, H. In Chlorinated Paraffins; DeBoer, J., Ed.; Springer: London, 2010; pp 1−40. (2) Muir, D.; Stern, G.; Tomy, G. In New Types of Persistent Halogenated Compounds; Paasivirta, J., Ed.; Springer: London, 2000; pp 203−236. (3) Chen, M.-Y.; Luo, X.-J.; Zhang, X.-L.; He, M.-J.; Chen, S.-J.; Mi, B.-X. Environ. Sci. Technol. 2011, 45, 9936−9943. (4) Persistent Organic Pollutants Review Committee. Draft Risk Profile for Short-Chained Chlorinated Paraffins; Secretariat of the Stockholm Convention: Geneva, 2007; http://www.pops.int/ documents/meetings/poprc/drprofile/drp/DraftRiskProfile_SCCP. pdf. (5) ECHA. 2008; http://echa.europa.eu/de/candidate-list-table (accessed Jan. 20, 2015). (6) Zencak, Z.; Borgen, A.; Reth, M.; Oehme, M. J. Chromatogr., A 2005, 1067, 295−301. (7) Rusina, T. P.; Korytar, P.; de Boer, J. Int. J. Environ. Anal. Chem. 2011, 91, 319−332. (8) Reth, M.; Zencak, Z.; Oehme, M. J. Chromatogr., A 2005, 1081, 225−231. (9) Tomy, G. T.; Stern, G. A. Anal. Chem. 1999, 71, 4860−4865. (10) Tomy, G. T.; Stern, G. A.; Muir, D. C. G.; Fisk, A. T.; Cymbalisty, C. D.; Westmore, J. B. Anal. Chem. 1997, 69, 2762−2771. (11) Coelhan, M.; Saraci, M.; Parlar, H. Chemosphere 2000, 40, 685− 689. (12) Korytar, P.; Covaci, A.; Leonards, P. E. G.; de Boer, J.; Brinkman, U. A. T. J. Chromatogr., A 2005, 1100, 200. (13) Leonards, P.; Koekkoek, J. Analysis of medium chain chlorinated paraffins: Comparison of GCxGC-ECD vs. GC-ECNIMS. Proceedings of the 34th International Symposium of Halogenated Persistent Organic Pollutants Dioxin 2014, Madrid, Spain, August 31− September 5, 2014.

■

CONCLUSIONS The presented APCI-qTOF-HRMS analysis and deconvolution procedure offers some benefits compared to established analytical methods for CPs in environmental samples, particularly the possibility to quantify single CP congener groups. The APCI-qTOF-HRMS method could further be used as a screening tool for large sets of samples, because the analytical part is very quick. Whole signal intensity would H

DOI: 10.1021/ac504444d Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(14) Nilsson, M. L.; Bengtsson, S.; Kylin, H. Environ. Pollut. 2012, 163, 142−148. (15) Geiß, S.; Einax, J. W.; Scott, S. P. Clean: Soil, Air, Water 2010, 38, 57−76. (16) Zencak, Z.; Reth, M.; Oehme, M. Anal. Chem. 2004, 76, 1957− 1962. (17) Coelhan, M. Anal. Chem. 1999, 71, 4498−4505. (18) Koh, I. O.; Rotard, W.; Thiemann, W. H. P. Chemosphere 2002, 47, 219−227. (19) Steinberg, S. M.; Emerson, D. W. Environ. Monit. Assess. 2012, 184, 2119−2131. (20) Zencak, Z.; Oehme, M. Rapid Commun. Mass Spectrom. 2004, 18, 2235−2240. (21) Jahnke, A.; Mayer, P.; McLachlan, M. S.; Wickström, H.; Gilbert, D.; MacLeod, M. Environ. Sci.: Processes Impacts 2014, 16, 464−472. (22) Jahnke, A.; MacLeod, M.; Wickström, H.; Mayer, P. Environ. Sci. Technol. 2014, 48, 11352−11359. (23) Diefenbacher, P.; Bogdal, C.; Gerecke, A. C.; Schmid, P.; Scheringer, M.; Hungerbühler, K. Organohalogen Compd. 2014, 76, 246−249. (24) Zennegg, M.; Munoz, M.; Schmid, P.; Gerecke, A. C. Environ. Int. 2013, 60, 202−208. (25) Tannenbaum, H. P.; Roberts, J. D.; Dougherty, R. C. Anal. Chem. 1975, 47, 49−54. (26) Zencak, Z.; Reth, M.; Oehme, M. Anal. Chem. 2003, 75, 2487− 2492. (27) Lawson, C. L.; Hanson, R. J. Solving Least Squares Problems; Prentice-Hall Inc.: Englewood Cliffs, New Jersey, USA, 1974. (28) Tomy, G. T.; Fisk, A. T.; Westmore, J. B.; Muir, D. C. G. Rev. Environ. Contam. Toxicol. 1998, 158, 53−128. (29) Heath, E.; Brown, W. A.; Jensen, S. R.; Bratty, M. P. J. Ind. Microbiol. Biotechnol. 2006, 33, 197−207. (30) Zencak, Z.; Oehme, M. TrAC, Trends Anal. Chem. 2006, 25, 310−317. (31) Olofsson, U.; Bignert, A.; Haglund, P. Water Res. 2012, 46, 4841−4851. (32) Maulshagen, A.; Hamm, S.; Petersen, M.; Elsholz, O.; Fengler, S.; Seel, P. Organohalogen Compd. 2003, 62, 371−374. (33) Barber, J. L.; Sweetman, A. J.; Thomas, G. O.; Braekevelt, E.; Stern, G. A.; Jones, K. C. Environ. Sci. Technol. 2005, 39, 4407−4415. (34) Pellizzato, F.; Ricci, M.; Held, A.; Emons, H.; Böhmer, W.; Geiss, S.; Iozza, S.; Mais, S.; Petersen, M.; Lepom, P. TrAC, Trends Anal. Chem. 2009, 28, 1029−1035. (35) Tomy, G. T.; Westmore, J. B.; Stern, G. A.; Muir, D. C. G.; Fisk, A. T. Anal. Chem. 1999, 71, 446−451. (36) van der Veen, I.; Crum, S.; De Boer, J. Interlaboratory Study on the Analysis of Chlorinated Paraffins in Environmental Matrices Phase II; Report W-14/18, version 1; Institute for Environmental Studies: Amsterdam, 2014;www.normandata.eu/sites/default/files/files/ Events/2014/ QUASIMEME%20Interlaboratory%20Study%20CPs%20Report%20 phase%202.pdf. (37) Stevenson, G.; Yates, A.; Gillet, R.; Keywood, M.; Galbally, I.; Borgen, A. Organohalogen Compd. 2011, 73, 1367−1369. (38) van der Veen, I.; Cofino, W. P.; Crum, S.; De Boer, J. Interlaboratory Study on the Analysis of Chlorinated Paraffins in Environmental Matrices Phase 1; Report W-12/11, version 1; Institute for Environmental Studies: Amsterdam, 2012; http://norman.ineris.fr/ public_docs/qa_qc/report_chlorinated_paraffins_ils_quasimeme_ 2011.pdf. (39) Arbogast, B.; Budde, W. L.; Deinzer, M.; Dougherty, R. C.; Eichelberger, J.; Foltz, R. D.; Grimm, C. C.; Hites, R. A.; Sakashita, C.; Stemmler, E. Org. Mass Spectrom. 1990, 25, 191−196. (40) Stemmler, E. A.; Hites, R. A.; Arbogast, B.; Budde, W. L.; Deinzer, M. L.; Dougherty, R. C.; Eichelberger, J. W.; Foltz, R. L.; Grimm, C.; Grimsrud, E. P.; Sakashita, C.; Sears, L. J. Anal. Chem. 1988, 60, 781−787.

I

DOI: 10.1021/ac504444d Anal. Chem. XXXX, XXX, XXX−XXX