Environ. Sci. Technol. 2004, 38, 1681-1685
Atmospheric Emissions of Polybrominated Diphenyl Ethers and Other Persistent Organic Pollutants during a Major Anthropogenic Combustion Event N. J. FARRAR,* K. E. C. SMITH, R. G. M. LEE, G. O. THOMAS, A. J. SWEETMAN, AND K. C. JONES Department of Environmental Science, Lancaster University, Lancaster LA1 4YQ, United Kingdom
Air samples were taken for the analysis of persistent organic pollutants before, during, and after the national U.K. “Bonfire Festival” in November 2000. As expected, ambient levels of polynuclear aromatic hydrocarbons (PAHs) increased sharply in response to the widespread diffusive combustion processes that occurred at the time. Polybrominated diphenyl ethers (PBDEs) also increased at the suburban sampling location, to a greater extent than the PAHs. The rise and fall in PBDE concentrations was rapid, coinciding closely with the PAH “combustion markers”. These data provide evidence for a novel mechanism responsible for dissipation of PBDEs into the environment. It is hypothesized that products treated with the pentaBDE productsnotably household furnishing foams and textilesshave been subject to (unsanctioned) burning on private bonfires; even if the majority of the PBDE burden of such products is debrominated/broken down in the fires, it is shown that only small amounts of the total “stock” of penta product need be emitted to generate the concentrations detected. The mixture of PBDEs in the air during the Bonfire Festival was enriched in higher brominated congeners (e.g., BDE-99, -153, and -154) compared to that in background air. Estimates are made of the masses of compound classes that may have been emitted to the atmosphere during the festival.
Introduction Bonfires are large fires lit in the open air to mark a public celebration, or fires lit to burn garden rubbish. In the United Kingdom (U.K.), bonfires are lit across the country on the Nov 5. “Bonfire night” is a tradition continued for nearly 400 years to mark the capture of Guy Fawkes, who attempted to kill the British monarch, King James I, in 1605. Many thousands of bonfires are litssome at large public displays, others in people’s gardens. They are usually accompanied by fireworks displays. The U.K. atmosphere therefore potentially receives an elevated input of combustion-derived contaminants from numerous diffusive sources over a short period of time. National inventories of certain classes of persistent organic pollutants (POPs) suggest that diffusive, uncontrolled combustion processes may be a significant source of POPs to the * Corresponding author phone: (+44 1524) 592578; fax: (+44 1524) 593985; e-mail:
[email protected]. 10.1021/es035127d CCC: $27.50 Published on Web 02/06/2004
2004 American Chemical Society
atmosphere (1), although such sources can be very difficult to quantify accurately. We were therefore interested to assess the potential impact of this large-scale, diffusive combustion event on the emissions of different POPs. Domestic and public bonfires usually differ in their physical size and makeup. A public display will consist of a rather large bonfire (possibly 5 m or more in height), made predominantly from wood, some of which may presumably be treated. Such fires often burn for several hours and may be left smouldering overnight. Polynuclear aromatic hydrocarbons (PAHs) are incomplete products of combustion and are therefore one compound class of interest. Lee et al. (2) have previously identified Bonfire Night as a source of polychlorinated dibenzo-p-dioxins and -furans (PCDD/Fs). Other compounds may also be emitted. Often domestic bonfires are used as an (unsanctioned) opportunity to burn rubbish, in addition to more common garden waste, for example. Inevitably the combustion conditions, between fires and in a single fire, can vary considerably. This study investigated the episodic release of POPs during the Bonfire Night celebrations of November 2000. The study focused on polybrominated diphenyl ethers (PBDEs), PAHs, and polychlorinated biphenyls (PCBs). It was hypothesized that PBDEs and PCBs may be in rubbish added to bonfires, such as household rubbish and goods, furniture, textiles, etc. Back trajectory analysis (using the HYSPLIT model (3, 4)) was used to investigate the origin of air masses throughout the sampling campaign. A simple assessment of the possible contribution of Bonfire Night to the national emissions estimates was also performed.
Materials and Methods Sampling. Three Graseby-Anderson high-volume air samplers were installed in the garden of a home in a residential area, near the Lancaster city center in northwest England (longitude 2.794°, latitude 54.037°) (see Figure 1). No public bonfire displays were held in the vicinity of sampling, although some residential bonfires and associated fireworks displays occurred within the area. Samples were collected for different periods before, during, and after Nov 5, 2000; sampling began on Nov 1 and finished on Nov 13. Samples were taken every day during the campaign and usually over 24 h. Higher resolution (5 h) triplicate sampling was undertaken from Nov 4 to Nov 7, with each sampler yielding a sample of ca. 60-70 m3. The sampling train consisted of a standard PS-1 module containing a combination of two polyurethane foam (PUF) plugs which had been pre-extracted on a Soxhlet system for 18 h using dichloromethane (DCM) and one glass fiber filter (GFF) which had been baked out for 18 h at 450 °C in a muffle furnace. Following pre-extraction and field deployment, PUF plugs were stored in solventrinsed jars while GFFs were placed in solvent-rinsed foil packets. Samples were stored at -20 °C until analysis. Extraction, Cleanup, and quantification. The GFFs and PUFs were Soxhlet extracted separately in DCM for 18 h. The extracts were then reduced to 1 mL and cleaned up, utilizing a combination of an initial silica and neutral alumina column and gel permeation chromatography (GPC), before a final sample volume of ∼500 µL was prepared. At this stage, deuterated PAH internal standards were added; samples were analyzed by gas chromatography-mass spectrometry for PAHs; full details are given elsewhere (5, 6). Following this, internal standards for PCB/PBDE analysis were added (details available in refs 7 and 8) and the samples then blown down to 25 µL in dodecane. Analysis for PCBs and PBDEs was performed by GC-MS, using methods given elsewhere (7, VOL. 38, NO. 6, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1681
FIGURE 1. Back-trajectories for 4 days over the sampling campaign. The Lancaster sample site is indicated by the star, and the colored lines indicate the altitude of the air mass back-trajectory: red, 150 m; blue, 300 m; green, 450 m. 8). The following compounds were detected and quantified in at least some of the samples: BDEs-47, -49, -66, -85, -99, -100, -153, -154, -166, -181, and -190; PAHs naphthalene, 2-methylnaphthalene, 1-methylnaphthalene, biphenyl, 2,6dimethylnaphthalene, acenaphthylene, acenaphthene, 2,3,6trimethylnaphthalene, fluorene, phenanthrene, anthracene, 1-methylphenanthrene, fluoranthene, pyrene, benzo[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[e]pyrene, benzo[a]pyrene, perylene, indeno(123-cd)pyrene, dibenz(ah)anthracene, and benzo(ghi)perylene; PCBs-18, -22, -28, -44, -49, -52, -70, -74, -87, -90, -101, -118, -123, -138, -149, -153, -157, -167, -170, -180, -183, -187, and -188. QA/QC. Six field blanks were set up and handled in a way identical to that of real samples; these accounted for approximately 20% of the total samples taken during the study. Field blank values, for both PUF and GFF, were very good, with individual PAH values equivalent to 0-91 pg/m3 (PUF) and 0-12 pg/m3 (GFF), individual PCB values of 0-2.5 pg/m3 (PUF) and 0-0.6 pg/m3 (GFF), and individual PBDE values of 0-2.0 pg/m3 (PUF); all PBDEs associated with GFFs were below detection. Method detection limits (MDLs) were determined by the summation of the mean field blank value to a value equivalent to 3 times the standard deviation of the blank. Calculated MDLs were as follows: PAHs, 0.8-133 pg/ m3 (PUF) and 0-14 pg/m3 (GFF); PCBs, 0-10.6 pg/m3 (PUF) and 0-1.0 pg/m3 (GFF); PBDEs, 0-2.5 pg/m3 (PUF). MDLs could not be applied to PBDE GFF values. Recoveries were generally very good, averaging 109.5% for the study. The cleanup method was fully validated for the combined analysis of these three compound groups in the samples by 1682
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 6, 2004
undertaking a series of tests to establish recoveries and reproducibility. Eighteen separate high-volume air samples were collected, separately Soxhlet extracted, then bulked, and split into 18 “identical” validation samples, each equivalent to approximately 940 m3 of air. Six PUF samples were taken, three spiked with a “surrogate spike” (13C-labeled PCBs and PCDD/Fs, and deuterated PAHs) and the remaining three injected with a “native spike” (nonlabeled compounds: PAHs, PCBs, OC pesticides, and PBDEs, plus the surrogate spike). All three batches of six PUF samples were treated this way. The same procedure was carried out for GFFs (for particulate-phase compounds). Recoveries were typically ca. 94% (PCBs), 115% (PBDEs), and 128% (PAHs).
Results Back-Trajectories and Meteorology. Air mass back-trajectories were calculated using the HYSPLIT model (NOAA Air Research Laboratories; 3, 4) for each sample. Air sampled prior to Bonfire Night originated from the west, off the Irish Sea. This was also the case during the early part of Nov 5. However, through the day a low-pressure air mass started to develop (Figure 1), and this situation continued for the remainder of the campaign. Air was therefore originating from southern areas of the U.K. and central Europe. Temperatures remained fairly stable, with an average of 8 °C. The average temperature difference between night and day was 2.4 °C, with a minimum of 5 °C and a maximum of 11 °C being seen on Nov 3. It rained during the bonfire weekend period, which presumably acted to reduce the air concentrations, by scavenging of gas- and particle-phase compounds.
FIGURE 2. Atmospheric concentration profiles of PBDEs and PAHs over the sampling campaign. Sections of the graph indicate the origin of the air masses, as calculated using the HYSPLIT model.
FIGURE 3. Concentration profiles of PAHs seen before and during the period of peak combustion. Details of the Bonfire Period and the Timing of the Emissions Pulse. In 2000, Nov 5 was a Sunday. This is important to note, because it had the effect of prolonging bonfire celebrations over the weekend period. Most bonfires were actually organized for Saturday (Nov 4) evening; some took place on Sunday (Nov 5) itself. The peak emissionsson the night of Nov 4swere at a time when the air mass trajectories show air masses coming predominantly from the west. PAHs. The PAH trends are shown in Figure 2. At the beginning and end of the campaign, ∑PAH air concentrations were ∼20-30 ng m-3. This is typical of the location and season and in line with data collected at an established airmonitoring station at Hazelrigg, ca. 4 km away (9, 10). However, concentrations increased during the bonfire event to reach a maximum of ∼180 ng m-3, approximately a 4-fold increase over the background concentrations. The compounds naphthalene, 2-methylnaphthalene, biphenyl, and phenanthrene dominated the PAH profile during the period of peak emission, accounting for over 50% of the ∑PAH. The same compounds (with the exception of biphenyl) also
dominated the samples taken during “background” periods (Figure 3). As expected, Figure 2 clearly shows that the peak air concentrations occurred late on the night of the main bonfire celebrations. Back-trajectories showed the air at that time originated from the west, a region of low atmospheric POP emission (10). Consequently, the peak on the night of Nov 4 largely reflects “local emissions” from the Lancaster conurbation and surrounding areas. Levels also remained high during the day of Nov 5, as relatively contaminated air arrived at the site after traveling over the most populated southern and central parts of England (see Figure 1). Subsequently, concentrations declined to background levels over a few hours, as cleaner advected air arrived. The annual average “clearance time” of the U.K. atmosphere is 2-3 days, due to advection, but can be as short as 10-12 h (11). PBDEs. The trends in PBDEs closely mirrored those of the PAHs (see Figure 2). ∑PBDE concentrations observed before and after Bonfire Night were also in agreement with data recorded previously at nearby Hazelrigg (12, 13). At these times, concentrations were a few pg m-3 (Figure 4), in line VOL. 38, NO. 6, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1683
FIGURE 4. PBDE concentration profiles prior to and during the Bonfire Night event. Clean air masses, of Atlantic origin, are thought to be responsible for the limited number of PBDE congeners seen during the prebonfire profile.
TABLE 1. Summary of Data on the Gas/Particle-Phase Distribution before and during the Bonfire Event
compound
typical background level (pg/m3)
elevated peak emission level (pg/m3)
typical percent in the gas phase under normal conditions seen during this study
typical percent in the gas phase during the bonfire event
phenanthrene biphenyl benzo[a]pyrene PCB-28 PCB-52 PCB-180 PBDE-47 PBDE-99
10000 3500 300 15 4 2 2 2
30000 23000 1000 45 14 2 7.5 14
98-99 99-100 2-10 2 3 18 n/a n/a
98 100 3 4 4 40 95% during the time of peak concentrations. Comments on Mechanisms of Release. Combustion conditions vary substantially between bonfires, depending on their composition, moisture content, size, and so on. It is therefore difficult to generalize about their temperatures. Conditions in many bonfires may be akin to those in natural forest fires, where wood and leaves are combusted. Research
into natural fires carried out by CSIRO in Australia has provided continuous vertical temperature profile data for a range of vegetation types and combustion conditions. These data suggest that typical temperatures can range from 400 to 1000 °C, measured 0.5 m above the ground. However, temperatures can be variable, depending on fuel type and age, as well as environmental factors such as in-forest and open wind speed, rainfall, solar radiation, temperature, and relative humidity (21). Temperatures within open fires, where “backyard trash” is burned, may typically be in the 500-700 °C range (22). These conditions are important when seeking to understand the release of compounds from the bonfires. It is well-known that PAHs are formed during incomplete combustion and released to the atmosphere. Emissions can be in the gas phase, caused by the elevated temperatures in fires, and in the particle phase, associated with soot/ash. The likely sources of PBDEs in the bonfires are discarded textiles and furnishings, burned in the smaller scale bonfires at people’s homes. Sakai et al. (23) suggested that temperatures of 900 °C are required to ensure full degradation of PBDEs. However, at the lower temperatures likely in many bonfires, it appears that a fraction of the PBDEs may be emittedswithout being destroyed (debrominated) by the fire. Presumably some of this may be in the gas phase, by volatilization under the elevated temperatures of the fire. In addition, particulates of textile fibers, polymers, soft furnishings, charred foam, etc. may also be emitted. The latter may well be more important, given the increase in particle-bound PBDEs detected in the atmosphere during the peak emission. Estimated Emissions. The ambient data were used to derive a simple estimate of the possible national emissions of these compounds during the bonfire period. This estimate is inevitably subject to many uncertainties. The following assumptions were made: (i) that the measured concentrations are the result of a uniform emission within the region, vertically mixed throughout an atmospheric column of 1 km; (ii) given the population and size of the Lancaster area, that the mass of contaminant emitted per capita could be estimated; (iii) that the estimate in (ii) could be scaled up for the total U.K. population (60 million people), assuming a uniform emission per capita. The results of these calculations suggest that ∼1.3 kg of PBDE, 2700 kg of PAHs, and 5 kg of PCBs could have been emitted over the festival period. It should be noted that this represents a very “small” mass of PBDE, as the following illustrative calculation shows. An estimated 1500 tons of penta-BDE has been used within the U.K. (24). An estimated 30% of this has been used in polyurethane in furniture and a further 20% in carpets, beddings, textiles. and packaging materials. Therefore, ca. 750 tonnes is estimated to reside in these products within the U.K.; 1.3 kg is only 0.0002% of this stock. These categories of use may all be subject to some unsanctioned burning on domestic bonfires. The mass of chemical released to air from this stock during Bonfire Night will depend on the mass of material put on bonfires and the combustion destruction efficiency. As a simple example, if 0.1% of this stock (e.g., 750 kg) is burned on Bonfire Night, with a combustion efficiency of 99%, then 7.5 kg of PBDE would be emitted to the U.K. atmosphere. Assuming that 30% of this is BDE-99 and that the volume of the U.K. atmosphere up to 1000 m is ∼2.3 × 1014 m3 (11), this mass would generate an air concentration of ∼10 pg m-3, if evenly dispersed. This is in line with the air concentrations recorded during the event (see Figure 4). Obviously, the assumptions of the percentage of the treated “stock” that may be burned
and the destruction efficiency of the combustion process may both be very different in reality. However, these assumptions serve to illustrate the potential for diffusive combustion processes to impact ambient levels of PBDEs, PAHs, and PCBs.
Acknowledgments We are grateful to the U.K. Department of the Environment, Food and Rural Affairs (DEFRA) Air Quality Division for financial support. We gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and/or READY website (http://www.arl.noaa.gov/ready.html) used in this study.
Literature Cited (1) U.S. Environmental Protection Agency. Exposure Reassessment Document Volume 1: Sources of dioxin-like compounds in the US; External Review Draft, EPA/600/6-88/005Ca; U.S. Environmental Protection Agency: Washington, DC, September 2000; www.epa.gov/ncea/dei.html. (2) Lee, R. G. M.; Green, N. J. L.; Lohmann, R.; Jones, K. C. Environ. Sci. Technol. 1999, 33, 2864-2871. (3) Draxler, R. R.; Rolph, G. D. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model. Accessed via NOAA ARL READY Website. http://www.arl.noaa.gov/ready/hysplit4. html, NOAA Air Resources Laboratory, Silver Spring, MD, 2003. (4) Rolph, G. D. Real-time Environmental Applications and Display System (READY) Website. http://www.arl.noaa.gov/ready/ hysplit4.html, NOAA Air Resources Laboratory, Silver Spring, MD, 2003. (5) Smith, K. E. C.; Green, M.; Thomas, G. O.; Jones, K. C. Environ. Sci. Technol. 2001, 35, 2141-2150. (6) Smith, K. E. C.; Thomas, G. O.; Jones, K. C. Environ. Sci. Technol. 2001, 35, 2156-2165. (7) Thomas, G. O.; Sweetman, A. J.; Parker, C. A.; Kreibich, H.; Jones, K. C. Chemosphere 1998, 36, 2447-2459. (8) Gouin, T.; Thomas, G.; Cousins, I.; Barber, J.; Mackay, D.; Jones, K. C. Environ. Sci. Technol. 2002, 36, 1426-1434. (9) Lee, R. G. M.; Hung, H.; Mackay, D.; Jones, K. C. Environ. Sci. Technol. 1998, 32, 2172-2179. (10) Lee, R. G. M.; Jones, K. C. Environ. Sci. Technol. 1999, 33, 705712. (11) Sweetman, A. J.; Cousins, I. T.; Seth, R.; Jones, K. C.; Mackay, D. Environ. Toxicol. Chem. 2002, 21, 930-940. (12) Lee, R. G. M.; Thomas, G. O.; Jones, K. C. Organohalogen Compd. 2002, 58, 101-104. (13) Lee, R. G. M.; Thomas, G. O.; Jones, K. C. Environ. Sci. Technol. 2004, 38, 699-706. (14) Jaward, F. M.; Farrar, N. J.; Harner, T.; Sweetman, A. J.; Jones, K. C. Environ. Sci. Technol., 2004, 38, 34-41. (15) Sjo¨din, A.; Jakobsson, E.; Kierkegaard, A.; Marsh, G.; Sellstro¨m, U. J. Chromatogr., A 1998, 822, 83-89. (16) Alaee, M.; Arias, M.; Sjo¨din, A.; Bergman, A. Environ. Int. 2003, 29, 683-689. (17) Watanabe, I.; Sakai, S. Environ. Int. 2003, 29, 665-682. (18) Breivik, K.; Sweetman, A.; Pacyna, J. M.; Jones, K. C. Sci. Total Environ. 2002, 290, 181-198. (19) Breivik, K.; Sweetman, A.; Pacyna, J. M.; Jones, K. C. Sci. Total Environ. 2002, 290, 199-224. (20) Tanner, R. L.; Parkhurst, W. J.; Valente, M. L.; Humes, K. L.; Jones, K.; Gilbert, J. Atmos. Environ. 2001, 35, 6539-6547. (21) www.bbm.csiro.au/vesta/f_tower. (22) Gullett, B. K.; Lemieux, P. M.; Lutes, C. C.; Winterrowd, C. K.; Winters, D. L. Chemosphere 2001, 43, 721-725. (23) Wevers, M.; De Fre, R.; Desmedt, M. Chemosphere 2004, 54, 1351-1356. (24) Alcock, R. E.; Sweetman, A. J.; Prevedouros, K.; Jones, K. C. Environ. Int. 2003, 29, 691-698.
Received for review October 10, 2003. Revised manuscript received December 23, 2003. Accepted December 30, 2003. ES035127D
VOL. 38, NO. 6, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1685
