Article pubs.acs.org/est
Deposition of Brominated Flame Retardants to the Devon Ice Cap, Nunavut, Canada Torsten Meyer,† Derek C. G. Muir,*,‡ Camilla Teixeira,‡ Xiaowa Wang,‡ Teresa Young,‡ and Frank Wania† †
Department of Physical and Environmental Sciences, University of Toronto Scarborough, 1265 Military Trail, Toronto, Ontario, M1C 1A4, Canada ‡ Aquatic Ecosystem Protection Research Division, Environment Canada, Burlington, Ontario L7R 4A6, Canada S Supporting Information *
ABSTRACT: Brominated flame retardants (BFRs) can be transported to Arctic regions via atmospheric long-range transport, however, relatively little is known about their deposition to terrestrial environments. Snow cores from the Devon Ice Cap in Nunavut, Canada served to determine the recent depositional trends of BFRs. Snow pits were dug in 2005, 2006, and 2008. Dating using annual snow accumulation data, ion chemistry, and density measurements established that the pits covered the period from approximately 1993 to spring 2008. Samples were extracted under clean room conditions, and analyzed using GC-negative ion MS for 26 tri- to decabromodiphenyl ethers (BDEs), as well as other BFRs, nonbrominated flame retardants, and industrial chemicals. Decabromodiphenyl ether (BDE-209) was the major congener present in all samples followed by nona-BDEs (BDE-207, BDE206, and BDE-208), both accounting for 89% and 7% of total BDE, respectively. BDE-209 concentrations were in most cases significantly correlated (P < 0.05) to tri- to nona-BDE homologues, and the strength of the correlations increased with increasing degree of bromination. Prior to or after deposition BDE-209 may be subject to debromination to lighter congeners. Deposition fluxes of BDE-209 show no clear temporal trend and range between 90 and 2000 pg·cm−2·year−1. Back trajectory origin in densely populated areas of northeastern North America is significantly correlated (P < 0.005) with the BDE-209 deposition flux. Several other high production volume and/or alternative BFRs such as hexabromocyclododecane (HBCD), 1,2-bis(2,4,6dibromophenoxy)ethane (BTBPE), pentabromo ethyl benzene (PBEBz), and pentabromobenzene (PBBz), as well as the industrial chemical 1,3,5-tribromobenzene (135-TBBz) were found consistently in the snow pits.
■
INTRODUCTION Ice and snow cores from midlatitude and Arctic ice caps have been widely used to study recent and historical atmospheric concentrations of anthropogenic organic contaminants. Both gaseous and particle-bound organic compounds are subject to efficient snow scavenging.1 Substantial amounts of particles are also deposited to Arctic glaciers via dry deposition.2,3 Changes in concentrations with depth in the accumulated snow are interpreted as reflecting changes in the chemistry of the atmosphere over the interval of deposition.4 Such temporal trends have been documented for persistent organic pollutants,5−7 polycyclic aromatic hydrocarbons (PAHs),8,9 current use pesticides,10 and brominated flame retardants, including BDEs.11 The Devon Ice Cap receives pollutants via atmospheric long-range transport from North American and Eurasian sources.3 Particularly between January and April, the atmosphere of Devon Island experiences Arctic haze episodes, during which atmospheric conditions can resemble those in the polluted midlatitudes.2,12 Organic substances of low volatility (octanol-air partition coefficients log KOA > ∼13), such as © 2011 American Chemical Society
higher brominated BDEs (hexa- to deca-homologues), are primarily associated with particles in the atmosphere of the high Arctic. However, semivolatile and lower brominated BDEs (trito penta-homologues) (log KOA < ∼13) may be partially present as vapors.13,14 The interpretation of chemical concentration patterns in ice and snow cores requires the consideration of postdepositional processes. Changes in physical snow properties and temperature can drive the revolatilization of semivolatile substances from snow back to the atmosphere 15−17 aided by wind ventilation within the upper snow layers.18 Semivolatile chemicals are subject not only to remobilization, as downhill winds ablating recently deposited snow from the upper part of the Devon Ice Cap to areas of lower altitude,19 may also relocate some less volatile substances. Finally, during limited Received: Revised: Accepted: Published: 826
August 19, 2011 December 5, 2011 December 7, 2011 December 7, 2011 dx.doi.org/10.1021/es202900u | Environ. Sci. Technol. 2012, 46, 826−833
Environmental Science & Technology
Article
Figure 1. Sampling site on the Devon Ice Cap and densely populated source areas (dark blue color).
the time period 1995−2005. Also, Svalbard and the Canadian High Arctic are influenced by source regions in North America and Eurasia to a different relative extent.
summer melt as commonly occurs in the Arctic, relatively water-soluble chemicals may be transported downward along with percolating meltwater, whereas more hydrophobic substances are left behind.20,21 In snow with low particle content, even fairly hydrophobic substances such as phenanthrene may be mobilized by meltwater.22 Despite those limitations deposition histories at relatively high temporal resolution can still be reconstructed at well-studied sites such as the Devon Ice Cap.23,24 In 2004, technical penta- and octa-BDE mixtures were banned in the EU,25 and banned or voluntarily phased out in the USA.26 Deca-BDE is the major “additive” brominated flame retardant currently in use, and there is concern that it can be debrominated to congeners that are more bioaccumulative and toxic.27−30 BDEs have been found in various environmental compartments in the Arctic including air, sediment, and biota.31 Recent emission reductions have so far not resulted in consistent temporal trend of tri- to hepta-BDEs throughout the Arctic. Concentrations of BDE-209 in the atmosphere of Alert, High Arctic Canada, were found to increase over the time period 2002−2005.32 Recently, other high production volume and/or alternative halogenated flame retardants such as HBCD, tetrabromobisphenol A (TBBPA), dechlorane plus (DP), BTBPE and decabromo diphenyl ethane (DBDPE) have received increasing attention.33 The only published study on flame retardants in an Arctic ice core reported deposition from 1953 to 2005 to the Holtedahlfonna ice cap in Svalbard, Norway.11 The most abundant flame retardants were BDE-209 and HBCD, whose presence in Svalbard was mainly attributed to transport from Eurasia during the Arctic haze season. NonaBDEs, octa-BDEs as well as other high production volume BDE congeners such as BDE-47 and BDE-99 were present but not quantified due to high detection limits. Here we present depositional trends of a wide range of brominated flame retardants, as well as several nonbrominated industrial chemicals, covering the period from approximately 1993 to spring 2008. Our study complements the study by Hermanson et al.,11 because the latter represents a longer-term period of deposition with no temporal resolution for the recent past, whereas here we present data with temporal resolution for
■
METHODS Sample Collection. Samples were collected in May of each year 2005, 2006, and 2008 from the Devon Ice Cap, Nunavut (75° 20.4 N, 82° 40.2 W) (Figure 1). Snow pits were dug with depths of 5 m in 2005, 6.8 m in 2006, and 7 m in 2008, each located several kilometers upwind from the nearest temporary research site of the CryoSat-2 line34 at about 1800 m above sea level. Sampling sites were two to five kilometers apart from each other. In 2005 and 2006 samples were taken horizontally, in 2005 every 25 cm from top to the bottom, and in 2006 every 25 cm (from surface to 3 m depth) and every 20 cm (from 3 m depth to bottom), using a stainless steel corer of 8.1 cm diameter. The samples were stored in 4 L polypropylene (PP) bottles and later pooled to obtain larger sample volumes. In 2008 duplicate samples representing about 16 L of snow were taken vertically at 50 cm intervals along the face of the pit using 4 L PP bottles, thereby covering the entire vertical stretch. The sample comprising the depth range from 3 to 3.5 m of the 2008 snow pit was lost in the field and not available for analysis. Additional smaller samples were taken at 10 cm intervals for density and ion analysis. All samples were shipped frozen by airfreight to the Canada Centre for Inland Waters (CCIW), Burlington, Canada and kept at −20 °C until analysis. The sampling procedure of the 2006 campaign is also described in detail in Young et al.35 Sample Preparation and Analyses. The full method for quantitative analysis and QA/QC is described elsewhere 11,36 and further details are provided in the Supporting Information, SI. Briefly, individual snow/ice samples were melted in a clean room at CCIW (positively pressured, HEPA and carbon filtered air) and the meltwater (6−11 L per sample) was extracted using XAD-2 resin columns. Extracts from the 2005 and 2006 snow pits were screened for 15 individual BDEs, as well as HBCD, BTBPE, PBEB, and DBDPE using GC-negative ion mass spectrometry (GC-NIMS). HBCD was quantified based on a γ-HBCD standard as it was most likely thermally 827
dx.doi.org/10.1021/es202900u | Environ. Sci. Technol. 2012, 46, 826−833
Environmental Science & Technology
Article
Figure 2. Concentrations of BDEs in snow (water equiv.) from the 2008 pit. The snow pit covers deposition from approximately late 1993 to April 2008.
isomerized to the α isomer in the GC injection port. The extracts from the 2008 snow pit were analyzed for 26 individual BDEs as well as 22 non-BDE flame retardants and industrial chemicals. The full list of analytes, as well as method detection limits (MDL), are given in SI Table S1. The subsamples were analyzed for the ions SO42−, Cl−, Na+, and Ca2+ in CCIW’s National Laboratory for Environmental Testing, using standard methods.37 Snow Pit Dating. Years of deposition were estimated by combining large amounts of physicochemical data including snow densities, snow grain structure and ice layers observed and recorded while working in the pit (SI Figures S1, S2), ion concentration records, annual deposition data from various studies, as well as a detailed data set from a snow pit close to our 2008 pit and investigated by a different group using different methods.24 Historical accumulation data of water equivalents were taken from Boon et al. (ref 34 and refs therein); modeled high resolution mass balance estimates for the Devon Ice Cap from Gardner and Sharp.38 A detailed description of the dating procedure, associated calculations, and uncertainties can be found in the SI. Airsheds. In order to trace the origin of the air masses arriving at the sampling site on the Devon Ice Cap, five day back trajectories were calculated every 6 h at 50, 100, and 200 m height for the time period from 1994 to 2008 and compiled to create back trajectory probability density maps (“airsheds”). Airshed images were generated with the Geographic Information System Manifold System 8.0 using a method described in Westgate et al.39 The fractions of back trajectories that had their origin in densely populated areas of North
America and Eurasia were calculated by superimposing the airsheds with population density maps40 using Manifold (Figures 1, 6, and S6 of the SI). Areas were considered densely populated when the number of inhabitants within one geographic grid cell (1° latitude, 1° longitude) exceeded 100, 000 (Figure 1).
■
RESULTS AND DISCUSSION Concentrations of Brominated Diphenyl Ethers. The concentrations of the individual BDE congeners are presented in Figures 2 (2008 samples) and 3 (2005 and 2006 samples). The concentrations of the BDEs comprising all three years of sampling were 680−100,000 pg·L−1 for deca-BDE, 56−1400 pg·L−1 for BDE-208, 81−2200 pg·L−1 for BDE-207, 99−1500 pg·L−1 for BDE-206,
