Field Evaluation of a Flow-Through Sampler for Measuring Pesticides

because of the entrainment of blowing snow/ice crystals or large particles. .... Briefly, each GFF and PUF extract was cleaned up with an alumina ...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Field Evaluation of a Flow-Through Sampler for Measuring Pesticides and Brominated Flame Retardants in the Arctic Atmosphere Hang Xiao,†,‡,* Hayley Hung,† Frank Wania,‡ Randy Lao,‡ Edwin Sabljic,§ Ed Sverko,∥ Ying Duan Lei,‡ Phil Fellin,⊥ and Enzo Barresi∥ †

Air Quality Processes Research Section, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4 Department of Physical and Environmental Sciences, University of Toronto Scarborough, 1265 Military Trail, Toronto, Ontario, Canada M1C 1A4 § Department of Biology, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1 ∥ National Laboratory for Environmental Testing, Environment Canada, Burlington, ON, L7R 4A6 ⊥ Airzone One Ltd., 222 Matheson Blvd. E., Mississauga, Ontario, Canada L4Z 1X1 ‡

S Supporting Information *

ABSTRACT: A flow-through sampler (FTS) was codeployed with a super high volume active sampler (SHV) between October 2007 and November 2008 to evaluate its ability to determine the ambient concentrations of pesticides and brominated flame retardants in the Canadian High Arctic atmosphere. Nine pesticides and eight flame retardants, including three polybrominated diphenyl ether (PBDE) replacement chemicals, were frequently detected. Atmospheric concentrations determined by the two systems showed good agreement when compared on monthly and annually integrated time scales. Pesticide concentrations were normally within a factor of 3 of each other. The FTS tended to generate higher PBDE concentrations than the SHV presumably because of the entrainment of blowing snow/ice crystals or large particles. Taking into account uncertainties in analytical bias, sample volume, and breakthrough estimations, the FTS is shown to be a reliable and cost-effective method, which derives seasonally variable concentrations of semivolatile organic trace compounds at extremely remote locations that are comparable to those obtained by conventional high volume air sampling. Moreover, the large sampling volumes captured by the FTS make it suitable for the screening of new and emerging chemicals in the remote atmosphere where concentrations are usually low.

■

alternative.9,10 Although a comparison of four active and passive sampling techniques showed that PAS is capable of providing time-integrated air concentrations,11 environmental variables, such as temperature and wind speed,12,13 could still introduce significant uncertainties into the volumetric air concentrations obtained from PAS. One major limitation of the PAS approach is its relatively low sampling rate, which severely constrains temporal resolution. A previously described flow-through sampler (FTS)14 addresses the need for an improved sampling design, which can significantly shorten the sampling period while maintaining the capability of providing quantitative information in the absence of electrical power. By guiding the wind through a series of polyurethane foam plugs (PUFs) with the help of vanes and an aerodynamic shape, the FTS provides greatly increased sampling rates.14 The sampled air volume is calculated from wind speed, which is measured on top of the sampler using a precalibrated vortex rotor.15 Thus, an FTS can be directly deployed without prior calibration or the need for

INTRODUCTION Persistent organic pollutants (POPs) are relatively stable in the environment, subject to long-range transport (LRT), and potentially bioaccumulative and toxic (http://www.pops.int). Being semivolatile organic compounds (SOCs),1,2 they partition between air, water, soil, and sediment. Atmospheric LRT is generally considered the most rapid route for SOCs to reach remote environments. The Stockholm Convention on POPs advocates air monitoring as a key measure to assess the effectiveness of global control initiatives.3,4 Several established atmospheric monitoring networks, such as the Integrated Atmospheric Deposition Network and the Arctic Monitoring and Assessment Programme, will be instrumental in meeting this need.5,6 To overcome the challenges of very low levels, laboratory and field contamination, and limitations in analytical sensitivity, active high volume (HiVol) sampling techniques are used to capture SOCs in filters and/or sorbents. Conventional HiVol sampling, the most widely used approach for collecting atmospheric SOCs, can reliably quantify concentrations7 at high temporal resolution (ranging from several hours to a week).8 However, it is only suitable for deployment at a limited number of sites because of its high cost and dependence on electricity. In recent years, passive air sampling (PAS) techniques have been developed as a more cost-effective © 2012 American Chemical Society

Received: Revised: Accepted: Published: 7669

April 13, 2012 June 11, 2012 June 17, 2012 June 17, 2012 dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

depuration compounds. At a windy location, an FTS operates similarly to a conventional active sampler, but the driving force is wind instead of an electrical pump. Both indoor and field tests have proven the suitability of the FTS to provide reliable, quantitative atmospheric concentrations for a wide range of SOCs with minimum level of breakthrough.14−16 An FTS field sampling campaign in Tibet not only demonstrated that the FTS performed well under extreme weather conditions but also provided insight into the seasonal variation and atmospheric LRT of pesticides and FRs to the central Tibetan Plateau.17,18 Despite this successful application of FTS, a systematic field evaluation of the FTS’s performance in extremely remote environments was still needed. The aim of this study was to provide such a quantitative evaluation by comparing an FTS’s results with those obtained from a conventional super high volume active sampler (SHV) in the Canadian High Arctic.

mass spectrometric detector using instrumental configurations detailed in ref 18. The full list of chemicals analyzed for in FTS and SHV samples is given in the Supporting Information. Compounds, which had been quantified in both FTS and SHV samples and thus can be used for sampler comparison, include 21 pesticides, 14 PBDEs, and 3 FRs, namely 1,2-bis(2,4,6-tribromophenoxy)ethane (BTBPE), 2-ethyl-1-hexyl 2,3,4,5-tetrabromobenzoate (EHTeBB), and bis(2-ethyl-1-hexyl)tetrabromophthalate (TBPH). These FRs have recently been observed in air samples from Tibet and the Canadian High Arctic.18 To evaluate the deviation in results caused by analytical differences between the 2 laboratories, 10 archived SHV samples and their corresponding leftover extracts after analysis by NLET (called hereafter NLET remnants), including both GFF and PUF extracts, were analyzed in the HAPs laboratory using the same analytical method as that for the FTS samples. The results for archived SHV samples and NLET remnant reanalysis from the HAPs laboratory were compared with those reported by NLET to identify any systematic bias between these two laboratories. Quality Assurance/Quality Control (QA/QC). Two clean PUF discs were spiked with 5−50 ng of standards and treated like real samples to determine analytical method recoveries of FTS samples ranging from 68% to 103%. Differently sized PUF blanks (1 or 3 in.) were taken every month by exposing them to air during sample handling. The average blank amounts found in FTS-PUF plugs with different thickness, in units of picograms, are listed in Table SI2 of the Supporting Information. The amounts trapped in each PUF plug of an FTS sample were reduced by the average amount in field blanks of the corresponding thickness (3 or 1 in.). QA/QC procedures for SHV samples have been described previously.21 Recoveries of PBDE congeners and BTBPE spiked onto SHV-PUF plugs ranged from (80 ± 14)% to (111 ± 14)%. GFF and PUF blanks were taken once every four weeks; their means are also summarized in Table SI2 of the Supporting Information. The amounts detected in particle and gaseous phase extracts of the SHV sample were also adjusted by their respective field blanks. In general, blank levels and recoveries for all chemicals were acceptable.19,20 Data were not adjusted for recoveries. Negative values obtained from blank correction were eliminated before calculation of air concentrations.

■

METHODS Sampling Campaign. Air was sampled at Alert, Nunavut, Canada (82° 30′ N, 60° 20′ W, 200 m asl) using both FTS and SHV techniques. A SHV has been employed to take weekly 7 day integrated samples since 1992;19 here results for samples taken in 2007 and 2008 are presented. Approximately 13 000 m3 of air was sampled every week. Particles and gas phase compounds were trapped separately by a glass fiber filter (GFF) and two polyurethane foam (PUF) plugs, respectively. Only samples taken every other week were analyzed. The FTS was deployed between October 2007 and December 2008 to collect monthly integrated samples using a sampling procedure described previously.17 Briefly, three P10z PUF plugs (Pinta Foamtec, Minneapolis, MN, USA) with a total length of 17.8 cm (7 in., with two 3 in., and one 1 in. PUF plugs) were arranged in series in the sampling train. The air volume collected by the FTS during a 1 month period depends on wind speed and ranged from 2280 m3 to 13 500 m3. The sampling volumes and relevant meteorological parameters for each sampling period are summarized in Table SI1 of the Supporting Information. A total of 31 paired SHV and FTS samples were used for comparison. Instrumental Analysis. Samples taken by the two methods were analyzed in different laboratories using different methods. Each individual PUF plug of an FTS sample, regardless of its thickness, was extracted separately and analyzed at the Hazardous Air Pollutants (HAPs) Lab at Environment Canada (Toronto, Canada). After extraction, 100 ng of mirex were added into each sample before blow-down to 1 mL, whereas all of the sample preparation and instrumental analysis was identical to that in ref 16. GFF and PUF samples from the SHV sampler were extracted and split into two portions by Airzone One Ltd. (Mississauga, Canada) using methods described in detail before.19,20 Half of the extracts were archived in a freezer, whereas the other halves were shipped to Environment Canada’s National Laboratory for Environmental Testing (NLET, Burlington, Canada) for instrumental analysis following procedures described in refs 19 and 21. Briefly, each GFF and PUF extract was cleaned up with an alumina column and fractionated into a hexane and a dichloromethane fraction before quantification using external calibration. Pesticides analysis was performed using a gas chromatograph with an electron capture detector, whereas FRs were quantified with a gas chromatograph equipped with a

■

RESULTS AND DISCUSSIONS Quantifying Interlaboratory Differences. Because the two types of samples were prepared and analyzed independently at two laboratories using different sampling media and analytical methods, we first needed to ensure that the analytical results were comparable before evaluating the performance of the FTS. Several studies have quantitatively assessed the systematic bias in analytical results that arise when multiple laboratories analyze standards and atmospheric samples for SOCs.20,22−26 Even results for the same samples analyzed by two different laboratories can deviate by a factor of 2−3.20 In Figures SI1−SI3 of the Supporting Information, the analytical results from NLET are plotted against those from HAPs for archived samples (black triangles) and reruns of NLET remnants (red circles) for each sampling period, chemical, and chemical group, respectively. Because a portion of each NLET remnant had already been consumed before the addition of an internal standard at the HAPs laboratory, the reruns 7670

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

the relative importance of snow-bound chemicals and sample volume estimation errors. The average weather conditions and the calculated volume for each FTS sample are listed in Table SI1 of the Supporting Information. 2. Estimation of FTS-Sampled Air Volume. During the first field test of the FTS in Toronto, a relationship was established between the sampled air volume and the wind speed measured by a vortex sensor on top of the sampler.15 This correlation was used to calculate pesticide concentrations in Tibet, which generated results comparable to previous measurements in this region.17 Although sometimes winds at Alert are stronger than in Toronto, the wind speed range recorded is similar to that in Tibet. Hence, it was deemed justified to use the equation from ref 15 to estimate the sampled air volumes. An equation for estimating sample volumes at high wind speeds is still lacking. At low wind, the volume estimated from the recorded wind speed on top of the sampler is generally lower than that obtained from the wind speed measured after passage through the sampling medium,15 that is the sampling volume estimation method used in this study likely underestimated the sampled air volume under relatively calm wind conditions. Uncertainties in the sample volumes directly influence the concentrations derived from the FTS and may contribute significantly to the overall differences observed for the two sampling methods. For comparability, the sampled air volumes were normalized to standard atmospheric conditions (1 atm pressure, 0 °C) with the ideal gas law, using the average temperature and atmospheric pressure during each sampling period. 3. Breakthrough Correction. Large sample volumes may result in the breakthrough loss of more volatile SOCs, such as HCB and the HCHs. Previous studies14,16 have established a method based on frontal chromatographic theory for estimating breakthrough volumes and the theoretical plate number of the sampling train. With these parameters, one can quantify and correct for breakthrough loss. A detailed description and an evaluation of the performance of this correction method are given in the Supporting Information. A summary of the breakthrough corrections for HCB, α-HCH, and γ-HCH is provided in Table SI4 of the Supporting Information; only the final corrected results are discussed further. Whereas the correction yielded estimations of potential ambient concentration ranges for the more volatile SOCs at warmer temperature, it clearly tended to underestimate their breakthrough levels in winter samples: Whereas the theoretical plate prediction suggested an insignificant breakthrough, the frontal profile (ratios of the amount trapped in the back two PUF plugs compare to the first one) indicated otherwise. This is likely due to an overestimation of the breakthrough volume (VB) at lower temperature. A more systematic investigation of the breakthrough behavior of the more volatile compounds under cold environmental conditions is required. Air Concentration at Alert. Tables SI5 and SI6 of the Supporting Information present the final blank-corrected air concentrations from both FTS and SHV for the pesticides and FRs most commonly detected at Alert. Among the chemicals analyzed in samples taken by both methods, 10 out of 21 pesticides and 8 out of 17 FRs were frequently detected: HCB, α-HCH, γ-HCH, α-endosulfan, β-endosulfan, trans-chlordane (TC), cis-chlordane (CC), trans-nonachlor (TN), dieldrin, heptachlor epoxide (HEPX), the dominant BDEs (BDE-47, 99, 100, 153, 209), and the three other FRs. The remaining compounds were either mostly undetectable or their levels in

generally yield concentrations that are strongly correlated with, but slightly lower than, the concentrations in their corresponding archived samples. This was expected. For this reason, only the regressions between archived samples and NLET-reported results were used for estimating interlaboratory differences. For both GFFs and PUFs, the analytical results from HAPs were remarkably similar to those reported by NLET for all commonly detected chemicals (Figure SI1 of the Supporting Information). This confirmed that differences in the recovery rates between the two laboratories were insignificant despite a number of additional sample cleanup and fractionation procedures in the NLET method. NLET results were plotted against HAPs measurements of both archived samples and reruns of NLET remnants for each chemical and chemical group (OCPs, PBDEs, and new FRs) in Figures SI2 and SI3 of the Supporting Information, respectively. Linear regression results with intercepts of zero (Table SI3 of the Supporting Information) show that, except for EHTeBB and BTBPE, the differences between NLET and HAPs results of archived samples were well within a factor of 2, with slopes ranging from 0.535 to 1.77. In other words, the data for OCPs and PBDEs by the two laboratories were very comparable with strong linear correlations: OCP results from the two laboratories can almost be treated as identical; the PBDE results are also within a factor of 1.8 with slightly higher results given by NLET. Results from the two laboratories vary for the FRs: Whereas TBPH results are similar; NLET reported concentrations for BTBPE and EHTeBB that are more than 3 and 6 times higher than HAPS data, respectively. In general, as interlaboratory differences vary between chemicals, they need to be considered individually when comparing the results obtained by the two types of samplers. Factors Influencing Sampler Comparison. Besides analytical differences between laboratories, other factors could impact the outcome of the comparison by influencing what the FTS actually sampled or how the FTS-derived air concentrations were calculated. 1. Weather Conditions. A distinctive design difference between the FTS and most active air sampling devices is its horizontal, nonobstructive configuration, which allows the sampling medium to be directly exposed to the atmosphere. This configuration permits the FTS to achieve significantly higher sampling rates than conventional PASs, but it also makes the FTS more sensitive to weather conditions, such as snow, rain, or fog. These weather conditions could result in the entrainment of snow/water droplets/aerosols with associated SOCs into the FTS sampling medium. The severity of impact would vary with a chemical’s partitioning characteristics. FTSderived concentrations may therefore be biased high, especially for chemicals with a tendency to partition out of the gas phase and during time periods with precipitation or high concentrations of particles. The impact of snow storms on FTS sampling is slightly different from other weather phenomena. During snowy periods (December 2007, February to March 2008), ice was observed to accumulate in the front of the FTS. The snow/ice crystals trapped in the sampling train would not only introduce more snow/ice-sorbed chemicals into these specific samples but also cause clogging of the airflow through the sampling medium and thus an overestimation of sampled air volumes. Snowrelated artifacts can therefore result in either an overestimation or an underestimation of the air concentrations depending on 7671

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

define a wind-speed-weighted average concentration CWA for the SHV measurements,

most samples were similar to those in blanks. The following discussions will focus on the dominant compounds. Other than the three commonly detected FRs, among the emerging FRs, which were screened for in the SHV samples, only pentabromobenzene, pentabromotoluene, allyl 2,4,6tribromophenyl ether, and antidechlorane plus were occasionally detected. However, the blank-corrected concentrations were mostly very low, that is levels were not significantly different from the blanks, similar to previous screening results.18 Concentrations for selected SOCs obtained using the FTS and SHV are compared in Figure 1; results for all commonly

C WA =

∑ (VC i SHVi)/ ∑ Vi

(1)

where Vi is the volume of air sampled by the FTS during week i when concentration CSHVi was measured with the SHV (Table SI6 of the Supporting Information). Correspondingly, the airvolume based percentage for each FTS sampling period covered by SHV measurements (pi) can be defined as: pi =

∑ (Vi )/V0·100%

(2)

where V0 is the total volume sampled during an FTS sample period (Table SI1 of the Supporting Information). A Bland-Altman plot is a common method for analyzing the agreement between two methods.11 In such a plot (Figure SI6 of the Supporting Information), the air concentration difference between the two techniques for the same sampling period ΔC = CWA − CFTS is plotted against their mean CAVG = (CFTS + CWA)/2. The average ΔC and the 95% confidence interval (1.96 times the standard deviation of ΔC) are also calculated, which estimates the average agreement between FTS and SHV and the variability between all paired samples, respectively. Figure SI6 of the Supporting Information clearly shows that for all chemicals except HCB, for which the FTS consistently obtained lower values than the SHV, the monthly air concentrations were not significantly different from each other. As the CAVG for pesticides increases, ΔC shows no obvious decreasing or increasing trend. However, for PBDEs, significantly decreasing trends were observed, which are shown as linear regressions (blue lines) in the figures; such a trend is particularly evident for BDE-209 (particle-phase data). In other words, the FTS samples significantly more PBDEs than the SHV when their air concentrations are high. Compared with FTS results, one especially high SHV value was found for EHTeBB and TBPH. This is likely due to the interlaboratory bias. In Figure 2 and Figure SI7 of the Supporting Information, CFTS is plotted against CWA for selected chemicals. The black dotted diagonal lines represent perfect agreement. Except for HCB (Figure SI7 of the Supporting Information), the data points for all chemicals are more or less evenly distributed on either side of that line. Significant linear correlations were found between the OCP air concentrations, but not for most of FRs. The statistical results and slopes are given in Table SI7 of the Supporting Information. The slopes of the regressions for pesticides vary from 0.56 to 1.06, except for the minimum value of 0.32 for HCB. In other words, CFTS and CWA of pesticides are generally within a factor of 2 of each other. The only FR that showed significant linear correlations between the two methods is TBPH. Further, paired two-sample t tests (Table SI8 of the Supporting Information) indicate that only the concentrations for HCB, CC, and TN given by these two techniques differed significantly. Despite this, the concentration variation and therefore also the absolute differences for CC and TN were very small. The difference in HCB concentrations is mainly due to the underestimation of breakthrough. For each FTS sampling period, CFTS and CWA are plotted against each other in Figure 3. Pesticides, PBDEs, and other FRs are marked with different symbols in each panel. HCB results were excluded from this analysis. These comparisons allow for a better understanding of how the FTS sample volume estimation and SHV coverage percentage pI influence

Figure 1. Comparison of air concentrations of BDE-47, TBPH, and αendosulfan in air sampled at Alert, Nunavut, Canada, measured by a flow-through sampler (red bar) and a Super High Volume sampler (diamonds).

detected chemicals are given in Figure SI4 of the Supporting Information. For most compounds, both the absolute levels and the seasonal pattern reported by the two sampling methods generally agree well. In the following sections, we quantitatively compare the concentrations derived from the two methods with a focus on temporal resolution, different chemicals/ chemical groups, and isomer ratios. Comparison of Time-Integrated Air Concentrations on a Monthly Scale. We compared FTS and SHV techniques at the monthly temporal resolution of the FTS sampling. A comparison on a weekly scale is given in Figure SI5 of the Supporting Information. As the sampling rate of the FTS is wind-speed dependent, the air concentrations are biased toward windy periods. To compare the weekly SHV samples taken during a month with the one FTS sample for that month, we should not simply take the average of the SHV concentrations but weigh the importance of the SHV samples by the volume of air sampled by the FTS during the weeks of SHV sampling (Table SI6 of the Supporting Information). Accordingly, we 7672

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

suggest that FTS may give slightly higher results for particle-/ snow-bound substances due to a configuration that does not prevent inclusion of snow/ice or particles. Most of the slopes for pesticides ranged between 0.35 and 2.0 (Table SI9 of the Supporting Information), except for the minimum and maximum values of 0.205 and 2.26 for December 2007 and August 2008 samples, respectively. High concentration variability during summer may be the main reason for the discrepancy for August 2008 samples because its pi was only 15%. Contrarily, the December 2007 FTS sample had a pi of 90%, thus, the disagreement is more likely due to the overestimation of the sampled air volume resulting from snow/ice build-up on the FTS during sampling. Yet, the snow/ ice build-up will bring more particle-bound or surface-absorbed chemicals into the FTS. Thus, by no means coincidental, the December 2007 FTS sample has generally higher PBDE concentrations than that of the SHV. Under normal circumstances, the difference between the pesticide concentrations derived from the two methods is within a factor of 4 but for the December 2007 sample, the discrepancies are slightly higher (factor of 4.8). For PBDEs, although the FTS tends to report higher values due to its configuration, the absolute results are not only well correlated with but also well within an order of magnitude of the SHV measurements. Considering that the FTS and the SHV samples have different sampling durations and were analyzed in two different laboratories using different methods, the agreement in their results is quite good. Comparison of Average Air Concentrations for the Whole Sampling Campaign. Annually averaged air concentrations obtained by FTS and SHV are compared in box-andwhisker plots in Figure SI8 of the Supporting Information. Whereas the techniques gave similar maxima, averages, and ranges for pesticides, the FTS generally gave slightly higher maxima and averages for PBDEs. The opposite was true for the other FRs, where the FTS yielded slightly lower results. The latter was mostly due to the analytical discrepancies between the two laboratories because the results were quite comparable after correcting for interlaboratory bias. Simple linear regressions forced through zero were performed for the average concentrations from different techniques for these three chemical groups in Figure 4, which clearly showed strong linear relationships. On an annually averaged basis, the pesticide concentrations from the FTS were almost identical to those from the SHV (slope of 1.02, R2 = 0.991, n = 8, p < 0.0001). The deviations between the two sampling methods were slightly larger for PBDEs and FRs, with corresponding slopes of 2.52 (R2 = 0.998, n = 5, p < 0.0001) and 0.497 (R2 = 0.882, n = 3, p = 0.0401), respectively. Considering the interlaboratory analytical bias, PBDE results obtained with the FTS were about 5 times higher than those obtained with SHV, whereas FTS measurements for other FRs varied from being similar to about 3 times higher. Similar to what was observed in ref 11, the differences in concentrations obtained by different sampling techniques became smaller with longer averaging periods. Because the data cannot be assumed to distribute normally, nonparametric hypothesis tests for paired sample were conducted for all chemicals. At the 95% confidence level, the distributions of results from these two samplers are not significantly different from each other, with Sign and Wilcoxon tests giving p values of 0.8036 and 0.7820, respectively. Comparison of Isomer Ratios. Isomer ratios can be used to trace sources of SOCs.27−29 We further evaluated the FTS

Figure 2. Comparison between flow-through sampler-derived air concentrations with wind-speed-weighted concentrations obtained by Super HiVol sampling for different chemicals. Red lines represent significant regressions between the results of the two techniques.

the agreement between the two methods. Strong linear correlations were found between CFTS and CWA for most FTS sampling periods. Simple linear regressions forced through the origin were used to fit the data for pesticides, PBDEs, and all chemicals combined (Table SI9 of the Supporting Information). (There are no linear regressions solely for new FRs because not all three FRs were detectable at all times.) Statistically significant regression lines are included in Figure 3. When considered separately, pesticides and PBDEs generally show significant linear correlations between the two methods; with the linearity of the regressions for pesticides generally better than those for the PBDEs. For samples taken in July 2008, pesticide concentrations from the two techniques are similar, but CWA for PBDEs are up to an order of magnitude higher than CFTS. After accounting for interlaboratory bias, these differences are smaller, only up to a factor of 5. As pi for the July FTS sample was only 17%, such discrepancies may be due to the short-term variability of snow/particles and/or snow/particle-bound compounds. Without considering the interlaboratory differences, the PBDE concentrations derived from FTS are approximately 45, 18, and 126 times higher than SHV results for samples taken in December 2007, February and March 2008, respectively, whereas the pesticide results for the same period compare very well. Since pi for samples during these months was high (90%, 84%, and 41% respectively), large differences between these two methods are unlikely the result of short-term variability. FTS entrainment of particle- or snowbound chemicals during Arctic haze events may play a more significant role in this discrepancy. These observations again 7673

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

Figure 3. Comparison between FTS-derived air concentrations with wind-speed-weighted concentrations obtained by Super HiVol sampling for different monthly sampling periods, where pesticides and flame retardants are marked with different symbols.

samplers. Only blank-corrected concentrations above 0.01 pg/ m3 were used to calculate isomer ratios. Figure 5 and Figure SI9 of the Supporting Information compare the FTS and SHV concentration ratios of HCH and chlordane isomers, as well as BDE-47/99, BDE47/209, and EHTeBB/TBPH. The two sampling techniques achieved similar absolute concentration values and concentration variation patterns in time for the two HCH isomers (Figure SI4 of the Supporting Information) and, therefore, also show similar α/γ-HCH ratios (Figure 5). The breakthrough corrected α/γ-HCH ratios obtained with the FTS varied from 4.6 to 24, a range only slightly narrower than the range of 1.0 to 28 obtained from SHV samples. A narrower range for the FTS is consistent with its longer sampling duration. Both FTS and SHV clearly indicate a summer peak in the α/γ-HCH ratio, which is consistent with the welldocumented volatilization of α-HCH from the Arctic Ocean during the warmer summer months.30−32 The ice cover during winter prevents volatilization resulting in a low α-HCH concentration and a low α/γ-HCH ratio. FTS and SHV also gave similar results for the ratios CC/TC, BDE-47/99, BDE47/209, and EHTeBB/TBPH (Figure SI9 of the Supporting Information). Unlike other pairs of chemicals that have one dominant isomer, the ratio of EHTeBB/TBPH observed in the SHV samples varied from 0.13 to 12 with an average of 3.8; the

Figure 4. Correlations between the annually averaged air concentrations derived from Super HiVol and Flow Through sampling for pesticides, PBDEs, and new flame retardants.

sampling efficiency for closely related chemicals, independent of the error involved in estimating the sampled air volume, through comparing the isomer ratios obtained from the two 7674

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

chemicals in the remote atmosphere at extremely low concentrations. Even in the absence of network power, the FTS constitutes a viable and economical method for reliably collecting SOCs from large volumes of air.

■

ASSOCIATED CONTENT

* Supporting Information S

Detailed information about original concentration data, breakthrough analysis and other supplementary results. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Thanks to the Northern Contaminant Program (Aboriginal Affairs and Northern Development Canada) and the Government of Canada’s Chemicals Management Plan for funding and to the Canadian Forces Station Alert for supporting data collection.

■

REFERENCES

(1) Brown, T. N.; Wania, F. Screening chemicals for the potential to be persistent organic pollutants: A case study of Arctic contaminants. Environ. Sci. Technol. 2008, 42, 5202−5209. (2) Muir, D. C. G.; Howard, P. H. Are there other persistent organic pollutants? A challenge for environmental chemists. Environ. Sci. Technol. 2006, 40, 7157−7166. (3) Global Monitoring Plan for Effectiveness Evaluation (UNEP/POPS/ COP.4/31) 2009, 13. (4) UNEP, Guidance for a Global Monitoring Programme for Persistent Organic Pollutants, 2004. (5) Buehler, S. S.; Hites, R. A. The Great Lakes’ integrated atmospheric deposition network. Environ. Sci. Technol. 2002, 36, 354A−359A. (6) Hung, H.; Kallenborn, R.; Breivik, K.; Su, Y.; Brorström-Lundén, E.; Olafsdottir, K.; Thorlacius, J. M.; Leppänen, S.; Bossi, R.; Skov, H.; Manø, S.; Patton, G. W.; Stern, G.; Sverko, E.; Fellin, P. Atmospheric monitoring of organic pollutants in the Arctic under the Arctic Monitoring and Assessment Programme (AMAP): 1993−2006. Sci. Total Environ. 2010, 408, 2854−2873. (7) Dobson, R.; Scheyer, A.; Rizet, A. L.; Mirabel, P.; Millet, M. Comparison of the efficiencies of different types of adsorbents at trapping currently used pesticides in the gaseous phase using the technique of high-volume sampling. Anal. Bioanal. Chem. 2006, 386, 1781−1789. (8) Hoff, R. M.; Muir, D. C. G.; Grift, N. P. Annual cycle of polychlorinated biphenyls and organohalogen pesticides in air in Southern Ontario. 1. Air concentration data. Environ. Sci. Technol. 1992, 26, 266−275. (9) Wania, F.; Shen, L.; Lei, Y. D.; Teixeira, C.; Muir, D. C. G. Development and calibration of a resin-based passive sampling system for monitoring persistent organic pollutants in the atmosphere. Environ. Sci. Technol. 2003, 37, 1352−1359. (10) Pozo, K.; Harner, T.; Wania, F.; Muir, D. C. G.; Jones, K. C.; Barrie, L. A. Toward a global network for persistent organic pollutants in air: Results from the GAPS study. Environ. Sci. Technol. 2006, 40, 4867−4873. (11) Hayward, S. J.; Gouin, T.; Wania, F. Comparison of four active and passive sampling techniques for pesticides in air. Environ. Sci. Technol. 2010, 44, 3410−3416.

Figure 5. Comparison of selected isomer ratios in Arctic air derived from flow-through sampler and super HiVol.

corresponding FTS values are 0.35 to 18 with an average of 2.4. As the two major components in Firemaster 550 (FM 550), the mass ratio of EHTeBB to TBPH in FM 550 is approximately 4;33 and EHTeBB is presumably the more volatile of the two compounds. 34 TBPH is also added to DP-45, a liquid flame retarded plasticizer.35,36 Previous research has suggested that the TBPH levels at Alert are partly due to sources other than FM 550.18 The isomer ratios from both the FTS and SHV samples are consistent with this hypothesis. This first field evaluation of the performance of the FTS under extreme arctic conditions provided a quantitative understanding of differences in its sampling efficiency for different chemical groups. The sampled air volume calculation method established in a previous field test in midlatitudes was judged reliable, unless snow build-up reduced the air flow through the sampling medium. Total air concentrations derived from the FTS were comparable to those obtained from a conventional active air sampler when compared on monthly and annually averaged time scales. For pesticides that are mostly in the gas phase, the concentrations derived from FTS and SHV were normally within a factor of 3. The FTS tended to sample more coarse particles and/or snow/ice, especially during the Arctic Haze period, thus generating higher concentrations for particle-bound substances. The correction method established in the previous study was useful in estimating breakthrough levels for HCHs, but tended to underestimate the loss under extremely cold conditions, especially for HCB. Generally, the FTS performed well under the extreme conditions of the Canadian High Arctic, being exposed to high winds and very low temperatures. Large sampling volumes make the FTS suitable for the screening of new and emerging 7675

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676

Environmental Science & Technology

Article

(31) Jantunen, L. M.; Helm, P. A.; Kylin, H.; Bidleman, T. F. Hexachlorocyclohexanes (HCHs) in the Canadian Archipelago. 2. Airwater gas exchange of alpha- and gamma-HCH. Environ. Sci. Technol. 2008, 42, 465−470. (32) Bidleman, T. F.; Jantunen, L. M.; Falconer, R. L.; Barrie, L. A.; Fellin, P. Decline of hexachlorocyclohexane in the Arctic atmosphere and reversal of air-sea gas exchange. Geophys. Res. Lett. 1995, 22, 219− 222. (33) Stapleton, H. M.; Allen, J. G.; Kelly, S. M.; Konstantinov, A.; Klosterhaus, S.; Watkins, D.; Maclean, M. D.; Webster, T. F. Alternate and new brominated flame retardants detected in U.S. house dust. Environ. Sci. Technol. 2008, 42, 6910−6916. (34) Ruan, T.; Wang, Y.; Wang, C.; Wang, P.; Fu, J.; Yin, Y.; Qu, G.; Wang, T.; Jiang, G. Identification and evaluation of a novel heterocyclic brominated flame retardant tris(2,3-dibromopropyl) isocyanurate in environmental matrices near a manufacturing plant in Southern China. Environ. Sci. Technol. 2009, 43, 3080−3086. (35) Lam, J. C. W.; Lau, R. K. F.; Murphy, M. B.; Lam, P. K. S. Temporal trends of hexabromocyclododecanes (HBCDs) and polybrominated diphenyl ethers (PBDEs) and detection of two novel flame retardants in marine mammals from Hong Kong, South China. Environ. Sci. Technol. 2009, 43, 6944−6949. (36) Andersson, P. L.; Ö berg, K.; Ö rn, U. Chemical characterization of brominated flame retardants and identification of structurally representative compounds. Environ. Toxicol. Chem. 2006, 25, 1275− 1282.

(12) Gouin, T.; Wania, F.; Ruepert, C.; Castillo, L. E. Field testing passive air samplers for current use pesticides in a tropical environment. Environ. Sci. Technol. 2008, 42, 6625−6630. (13) May, A. A.; Ashman, P.; Huang, J.; Dhaniyala, S.; Holsen, T. M. Evaluation of the polyurethane foam (PUF) disk passive air sampler: Computational modeling and experimental measurements. Atmos. Environ. 2011, 45, 4354−4359. (14) Xiao, H.; Hung, H.; Harner, T.; Lei, Y. D.; Johnston, G. W.; Wania, F. A flow-through sampler for semivolatile organic compounds in air. Environ. Sci. Technol. 2007, 41, 250−256. (15) Xiao, H.; Hung, H.; Harner, T.; Lei, Y. D.; Wania, F. Field testing a flow-through sampler for semivolatile organic compounds in air. Environ. Sci. Technol. 2008, 42, 2970−2975. (16) Xiao, H.; Hung, H.; Lei, Y. D.; Wania, F. Validation of a flowthrough sampler for pesticides and polybrominated diphenyl ethers in air. Atmos. Environ. 2009, 43, 2401−2409. (17) Xiao, H.; Kang, S.; Zhang, Q.; Han, W.; Loewen, M.; Wong, F.; Hung, H.; Lei, Y. D.; Wania, F. Transport of semivolatile organic compounds to the Tibetan Plateau: Monthly resolved air concentrations at Nam Co. J. Geophys. Res. D Atmos. 2010, 115 D16310; DOI 10.1029/2010JD013972. (18) Xiao, H.; Shen, L.; Su, Y.; Barresi, E.; DeJong, M.; Hung, H.; Lei, Y. D.; Wania, F.; Reiner, E. J.; Sverko, E.; Kang, S. C. Atmospheric concentrations of halogenated flame retardants at two remote locations: The Canadian High Arctic and the Tibetan Plateau. Environ. Pollut. 2012, 161, 154−161. (19) Fellin, P.; Barrie, L. A.; Dougherty, D.; Toom, D.; Muir, D.; Grift, N.; Lockhart, L.; Billeck, B. Air monitoring in the Arctic: Results for selected persistent organic pollutants for 1992. Environ. Toxicol. Chem. 1996, 15, 253−261. (20) Su, Y.; Hung, H.; Stern, G.; Sverko, E.; Lao, R.; Rosenberg, B.; Fellin, P.; Li, H.; Xiao, H. Assessing bias in a change of analytical laboratory under a long-term air monitoring program for organic pollutants in the Arctic: a quality assurance/quality control analysis. J. Environ. Monit. 2011, 13, 3111−3118, DOI: 10.1039/C1EM10513B. (21) Su, Y.; Hung, H.; Sverko, E.; Fellin, P.; Li, H. Multi-year measurements of polybrominated diphenyl ethers (PBDEs) in the Arctic atmosphere. Atmos. Environ. 2007, 41, 8725−8735. (22) Su, Y.; Hung, H. Inter-laboratory comparison study on measuring semi-volatile organic chemicals in standards and air samples. Environ. Pollut. 2010, 158, 3365−3371. (23) Bidleman, T. F. Interlaboratory analysis of high molecular weight organochlorines in ambient air. Atmos. Environ. 1981, 15, 619− 624. (24) Bidleman, T. F.; Cussion, S.; Jantunen, L. M. Interlaboratory study of toxaphene analysis in ambient air. Atmos. Environ. 2004, 38, 3713−3722. (25) Harner, T.; Kucklick, J. Interlaboratory study for the polychlorinated naphthalenes (PCNs): Phase 1 results. Chemosphere 2003, 51, 555−562. (26) Wu, R.; Backus, S.; Basu, I.; Blanchard, P.; Brice, K.; DryfhoutClark, H.; Fowlie, P.; Hulting, M.; Hites, R. Findings from quality assurance activities in the integrated atmospheric deposition network. J. Environ. Monit. 2009, 11, 277−296. (27) Hoff, R. M.; Muir, D. C. G.; Grift, N. P. Annual cycle of polychlorinated biphenyls and organohalogen pesticides in air in Southern Ontario. 2. Atmospheric transport and sources. Environ. Sci. Technol. 1992, 26, 276−283. (28) Bidleman, T. F.; Jantunen, L. M. M.; Helm, P. A.; BrorströmLundén, E.; Juntto, S. Chlordane enantiomers and temporal trends of chlordane isomers in arctic air. Environ. Sci. Technol. 2002, 36, 539− 544. (29) Shen, L.; Wania, F.; Lei, Y. D.; Teixeira, C.; Muir, D. C. G.; Bidleman, T. F. Atmospheric distribution and long-range transport behavior of organochlorine pesticides in North America. Environ. Sci. Technol. 2005, 39, 409−420. (30) Shen, L.; Wania, F.; Lei, Y. D.; Teixeira, C.; Muir, D. C. G.; Bidleman, T. F. Hexachlorocyclohexanes in the North American atmosphere. Environ. Sci. Technol. 2004, 38, 965−975. 7676

dx.doi.org/10.1021/es301481w | Environ. Sci. Technol. 2012, 46, 7669−7676