Environ. Sci. Technol. 2004, 38, 965-975
Hexachlorocyclohexanes in the North American Atmosphere L I S H E N , † F R A N K W A N I A , * ,† YING D. LEI,† CAMILLA TEIXEIRA,‡ DEREK C. G. MUIR,‡ AND TERRY F. BIDLEMAN§ Department of Chemistry and Department of Physical and Environmental Sciences, University of Toronto at Scarborough, 1265 Military Trail, Toronto, Ontario, Canada M1C 1A4, National Water Research Institute, 867 Lakeshore Road, Burlington, Ontario, Canada L7R 4A6, and Centre for Atmospheric Research Experiments, Meteorological Service of Canada, 6248 Eighth Line, Egbert, Ontario, Canada L0L 1N0
Annually integrated air concentrations of R- and γ-hexachlorocyclohexane (HCH) were determined in 2000/ 2001 at 40 stations across North America using XADbased passive air samplers to understand atmospheric distribution processes on a continental scale. Elevated levels of γ-HCH in the atmosphere of the Canadian Prairies are consistent with the ongoing use of lindane as a seed treatment on canola and confirm the feasibility of detecting the agricultural use of a pesticide using long-term integrated passive air sampling. In contrast to γ-HCH, the atmospheric concentrations of R-HCH show a rather uniform distribution across Canada and the United States, which is expected for a chemical with no current use on the continent. Higher levels in the atmosphere over Atlantic Canada can be explained by R-HCH evaporating from the waters of the Labrador Current, which is supported by the chiral composition of R-HCH and the temperature dependence of its atmospheric concentrations along the east coast of Canada. Similarly, R-HCH is volatilizing from Lake Superior. Atmospheric HCH levels increase with elevation in the Canadian Rocky Mountains. The results suggest that evaporation, in particular from cold water bodies, is an important source of R-HCH to the North American atmosphere. Low levels of HCHs in Central America hint at efficient degradation under tropical conditions. Chiral analysis shows that (+)-R-HCH is often enriched in air over continental areas and at the Pacific Coast, which is opposite to the enantiomeric enrichment in the proximity to the Great Lakes and the Atlantic Ocean. Passive air sampling is a powerful tool to discern the large-scale variability of semivolatile and persistent organic chemicals in the atmosphere.
Introduction As a commercial insecticide, hexachlorocyclohexane (HCH) was used in two formulations: (i) technical HCH, which includes multiple stereoisomers and is dominated by R-HCH (∼80%) and γ-HCH (10-15%), and (ii) lindane, which consists * Corresponding author phone: (416)287-7225; fax: (416)287-7279; e-mail:
[email protected]. † University of Toronto at Scarborough. ‡ National Water Research Institute. § Meteorological Service of Canada. 10.1021/es034998k CCC: $27.50 Published on Web 01/07/2004
2004 American Chemical Society
almost entirely of the γ-isomer. Technical HCH was in widespread use globally until the 1970s. As a result, approximately 6.5 Mt of R-HCH was released to the environment between 1948 and 1997 (1). Technical HCH is no longer in use, and many countries also have banned lindane. The latter is however still widely used in North American agriculture, particularly in Canada (2). R- and γ-HCH are among the major global organochlorine pollutants. Their transport pathways and fate in the environment have been investigated intensively by measuring concentrations in samples from many places in the world, particularly in air and water (e.g., refs 3-8). With a few exceptions (9-11), those efforts focused on a limited region. So far, there are few studies to describe the atmospheric fate of HCHs over a continental scale. The atmosphere plays an important role in the global cycle of organic pollutants (12). Many pollutants are dispersed by atmospheric currents to areas far from their original site of application (13, 14). The atmospheric behavior and distribution of organic pollutants depend on their physical properties and reactivity, the dynamics of the atmosphere, and the location of sources. Substances with relatively low volatility do not stay airborne for very long but attach to atmospheric particles and return quickly to the ground. Substances with relatively high volatility, relatively high Henry’s law constant, and low reactivity (such as the HCHs) can remain in the atmosphere for a long time before settling to the earth’s surface (15). Meanwhile, winds can carry them thousands of kilometers from their starting point so that these substances can establish similar concentrations all over the world. As long as major sources of a substance exist, its atmospheric concentrations will always be highest in the vicinity of these sources, while if emissions decrease or stop, the substance should gradually become more evenly distributed around the world, provided that it is sufficiently persistent (12). In this study, a large-scale network of 40 passive air sampling stations across North America was used to better characterize the atmospheric distributions of R-HCH and γ-HCH on a continental scale. These passive air samplers (PAS) were deployed for an entire year (from summer 2000 to summer 2001) to yield one annually averaged concentration at each station. To our knowledge, this network is the largest and most extensive atmospheric network for semivolatile organic contaminants to ever be in existence. Since R-HCH has been banned for years and γ-HCH is still in use, the divergent observations for these otherwise very similar chemicals should allow us to gain insight into how the global atmospheric system responds to past and present chemical releases. We further determined the relative abundance of the enantiomers of R-HCH in the air samples because chiral signatures have previously been shown to provide information into atmospheric fate and transport. Chiral signatures of R-HCH that had been subject to microbial degradation in soils and water bodies can be preserved upon volatilization and, therefore, allow the discrimination between racemic, primary sources and contaminants recycled from surface media (16, 17).
Experimental Section Sampler Design. A PAS based on the polymeric resin XAD-2 was employed (18). Briefly, the PAS consists of a resin-filled stainless steel mesh cylinder that is suspended in an inverted galvanized steel can with an open bottom. The PAS is deployed at 1.5 m above ground except in locations with a deep snowpack, when deployment height was increased to VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
965
TABLE 1. Sampler Concentrations (ng PAS-1) of r-HCH and γ-HCH, r-/γ-HCH Ratios, and Enantiomer Fractions (EFs) of r-HCH at 40 Stations across North America in 2000/2001 station
concn (ng PAS-1)
location
ratio
EF of r-HCH
no.
name
longitude
latitude
r-HCH
γ-HCH
r/γ
βDEXcst
BGB-172
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Point Petre Burnt Island Alert Bonavista Daniel Harbor East Point Sable Island Kejimkujik St. Leonard Frelighsburg Pukaskwa Thunder Bay Expl. Lake Area McCreary Bratt’s Lake Suffield Kananaskis Bow Lake Rock Isle Lake Donald Station Summerland Saturna Island Cape Beale Eureka Devon Island Kangiqtugaapik Cape Dorset Kuujjuaq Kuujjuarapik Big Creek Toronto Youngstown Solomons Wilmington Turkey Point Muscle Shoals Chetumal Tapachula Belmopan Costa Rica
77°09′ W 82°57′ W 62°17′ W 53°7′ W 57°35′ W 61°58′ W 60°02′ W 65°12′ W 67°50′ W 72°52′ W 86°18′ W 89°26′ W 94°22′ W 93°43′ W 104°42′ W 111°02′ W 115°02′ W 116°19′W 115°47′ W 117°10′ W 119°39′ W 123°08′ W 125°13′ W 85°56′ W 89°40′ W 68°37′ W 76°32′ W 68°25′ W 77°44′ W 80.5° W 79°24′ W 80.5° W 76°30′ W 87° W 84°30′ W 87°40′ W 88° W 92° W 89° W 84° W
43°50′ N 45°48′ N 82°31′ N 48°40′ N 50°14′ N 46°27′ N 43°56′ N 44°26′ N 47°09′ N 45°03′ N 48°35′ N 49°01′ N 49°47′ N 49°39′ N 50°12′ N 50°24′N 51°02′ N 51°40′ N 51°04′ N 51°29′ N 49°34′ N 48°46′ N 48°47′ N 80°00′ N 75°22′ N 70°28′ N 64°14′ N 58°6′ N 55°20′ N 42.7° N 43°39′ N 41° N 38°20′ N 34° N 29°54′ N 34°48′ N 18.5° N 15° N 17° N 10° N
9.1 7.1 6.8 31.6 17.7 21.5 32.1 11.5 11.0 8.3 15.4 11.4 12.3 9.6 15.1 16.9 9.3 12.0 13.4 5.2 14.5 12.3 6.8 9.0 10.2 18.9 17.3 10.5 11.4 5.6 11.6 9.6 17.2 2.9 5.6 10.7 1.1 1.9 0.3 0.9
6.4 3.8 1.0 3.9 1.6 3.5 2.3 3.2 4.2 11.3 4.2 6.7 10.7 25.0 78.1 30.4 4.1 2.6 3.0 0.9 3.2 1.7 0.9 1.5 1.8 2.0 2.9 1.9 2.8 6.7 16.2 9.3 13.7 4.9 5.8 10.4 2.4 12.1 1.1 5.0
1.4 1.9 6.5 8.0 11.2 6.1 14.2 3.6 2.6 0.7 3.7 1.7 1.2 0.4 0.2 0.6 2.2 4.6 4.5 5.5 4.5 7.4 7.5 6.0 5.8 9.5 6.1 5.5 4.1 0.8 0.7 1.0 1.3 0.6 1.0 1.0 0.5 0.2 0.3 0.2
0.489 0.490 0.501 0.471 0.478 0.480 0.470 0.489 0.498 0.500 0.476 0.502 0.509 0.511 0.513 0.515 0.516 0.516 0.517 0.515 0.519 0.516 0.519 0.499 0.496 0.486 0.488 0.494 0.490 0.502 0.500 0.505 0.501 0.507 0.509 0.514 0.512 0.504 -----
0.485
ensure that the PAS is never covered by snow. Contaminants are taken up in the resin from the atmosphere by diffusion, whereby previous experiments established independence of the sampling rate over a wide range of wind speeds (18). Measurements of the sorption coefficients for the XAD-2 resin (19) as well as year-long calibration experiments in Arctic and southern Canada (18) confirmed that HCHs do not reach equilibrium between the atmospheric gas phase and the resin. This makes it feasible to interpret the amount quantified in the PAS in terms of volumetric air concentrations, using a sampling rate that is largely independent of chemical, wind speed, and temperature (18). Sampling Sites. Thirty-one of the network’s stations are in Canada, five are in the United States, two are in Mexico, one is in Belize, and one is in Costa Rica (Table 1 and Figure 1). The stations roughly follow two transects. A north-south transect in the eastern part of the continent covers 72 degrees of latitude (10-82° N) from the Arctic to Central America. An east-west transect in southern Canada ranges from eastern Newfoundland to Vancouver Island, covering 72 degrees of longitude (53-125° W). Four of these stations constitute an altitudinal transect in the Southern Canadian Rocky Mountains, covering a range of elevation of more than 1500 m within a radius of approximately 100 km. The selection of these sites was based on a number of criteria: They should be representative of the atmospheric contamination status of a fairly large region and thus should not be close to 966
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004
0.500 0.479 0.478 0.470 0.484
0.498 0.520
0.513 0.510 0.509 0.518 0.502 0.498 0.489 0.491 0.496 0.490
significant point sources of organic contaminants. They should further be close to a climate station recording temperature, wind speed, and humidity on a regular basis. Finally, they should be safe (i.e., nobody unauthorized should be able to tamper with the sampler during the time of field deployment). Where possible, the samplers were located at existing air monitoring sites, for example, from the Canadian Acid Precipitation Monitoring Network (CAPMON) (stations 8, 13, and 22) and the Integrated Atmospheric Deposition Network (IADN) (stations 1 and 2). Sampling Procedure. XAD-2 underwent a rigorous cleaning procedure, and the PAS were filled and assembled in a dedicated clean room at the National Water Research Institute in Burlington, ON (18). Two PAS, detailed installation instructions, disposable cameras, and laboratory gloves were sent in May-July 2000 by courier or mail to local contact people, who installed duplicates at their sampling site. In some cases, we visited the sampling sites ourselves and installed the PAS. After 1 yr, the contacts received retrieval instructions and another set of gloves and were asked to return by courier or mail the XAD-filled mesh cylinder to NWRI in Burlington, ON, using the same airtight Teflon shipping tubes in which they had received the samplers the previous year. There, they were stored frozen until analysis. Eight of the stations received additional XAD mesh containers, which remained in the Teflon shipping containers throughout the 1-yr deployment period, during which they were taped
VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
FIGURE 1. Annual mean sampler concentrations of r-HCH (A) and γ-HCH (B) in the North American atmosphere collected from summer 2000 to summer 2001.
9
967
to the posts holding the sampling shelters. These samples served as field blanks, as they had undergone the same shipping and handling as the exposed samplers. Extraction and Quantification. The XAD-2 from the sampling container was transferred to an elution column, solvent-extracted, and fractionated as described in detail in ref 18. This reference also provides details of the quantification by gas chromatography-electron capture detection, and the quality assurance steps involving procedure blanks, resin blanks, and field blanks. Resin blanks and field blanks were higher than the procedural blanks. The passive air sampling concentrations (in units of ng PAS-1) are presented as blankcorrected averages of duplicates using the averages of 19 resin blanks and 8 field blanks. The average levels in the blanks were 18% of the lowest level in the sample for R-HCH and 7% for γ-HCH. The mean-normalized differences between duplicates were 0.21 for R-HCH and 0.15 for γ-HCH. Time-averaged volumetric air concentrations (CA; in ng m-3) were estimated by dividing the sampler concentration (in ng PAS-1) by the product of the deployment period (365 d) and the PAS sampling rate (0.52 m3 d-1 PAS-1) (18). Chiral Analysis for r-HCH. Determination of the enantiomeric composition of R-HCH was performed with a Hewlett-Packard 5890 GC-5989B MS engine in the negative chemical ionization mode. Separation was carried out using a Rt-βDEXcst (30 m, 0.32 mm i.d., 0.25 µm film thickness; Restek, Bellefonte, PA) and a BGB-172 column (30 m, 0.32 mm i.d., 0.25 µm; BGB Analytik, AG, Switzerland). Selected ions 255 and 257 were monitored. The (+)-R-HCH elutes first on the Rt-βDEXcst and (-)-R-HCH elutes first on the BGB-172. The elution order was confirmed with enriched (+)-R-HCH standard from AXACT Lab (Commack, NY). The GC temperature programs were 90 °C for 1 min, 20 °C/min to 160 °C, 1 °C to 225 °C, held for 15 min (Rt-βDEXcst); and 90 °C for 1 min; 20 °C/min to 170 °C; 2 °C to 195 °C, 20 °C/min to 225 °C, held for 40 min (BGB-172). Injector temperature was 220 °C, detector temperature was 220 °C, ion source temperature was 150 °C, and quadrupole temperature was 100 °C. Helium was used as carrier with methane as reagent gas. Chiral signatures are expressed using enantiomer fractions (EFs) (20), calculated as the ratio of the amount of (+)-RHCH to the total R-HCH. The EFs of racemic standard for R-HCH were 0.502 ( 0.002 for the Rt-βDEXcst column and 0.504 ( 0.002 for the BGB-172 column, calculated as the averages for ions 255 and 257 from replicate injection of standards. The isotope ratio for ions 255/257 obtained from standards was 1.55 with a standard deviation of (0.03. As a quality control, the measured ion ratio of R-HCH in the sample was accepted only if it was within the 95% confidence interval of the average ion ratio from the calibration standard.
Results Air Concentrations of r- and γ-HCH. R- and γ-HCH were detected in the samplers from all 40 stations. The concentrations of R- and γ-HCH, the R-/γ-ratio, and the EFs for R-HCH are listed in Table 1. The spatial distribution across the continent of these parameters is shown in maps in Figures 1-3. Sampler concentrations are in the range of 0.3-32 ng PAS-1 for R-HCH and of 1.0-78 ng PAS-1 for γ-HCH. This corresponds to volumetric air concentration between 1.5 and 170 pg m-3 for R-HCH and between 5 and 400 pg m-3 for γ-HCH. The spatial distribution of R- and γ-HCH across the North American atmosphere is very different. Even though the concentrations of both isomers range over approximately 2 orders of magnitude, Figure 1 reveals a much more uniform distribution of R-HCH than for γ-HCH. This is reflected in a much larger kurtosis (22.4 for γ-HCH vs 2.2 for R-HCH) and skewness (4.4 for γ-HCH vs 1.1 for R-HCH) of the γ-HCH data set, which means that its distribution is much more 968
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004
peaked and asymmetrical. R-HCH concentrations are mostly in the range of 5-15 ng PAS-1 (corresponding to 25-80 pg m-3), with slightly higher concentrations in the maritime provinces of eastern Canada and lower levels in Central America (Figure 2). γ-HCH levels are generally lower than 10 ng PAS-1 (corresponding to approximately 50 pg m-3), with much higher levels in the Prairie Provinces of Canada. The overall continental average concentrations are 11.4 ng PAS-1 (corresponding to 60 pg m-3) for R-HCH and 7.8 ng PAS-1 (corresponding to 41 pg m-3) for γ-HCH. Relative Abundance of r- and γ-HCH. The relative abundance of the two isomers is often expressed by the R/γ ratio. The R-/γ-HCH ratio in the technical HCH mixture is around 4-7. R-/γ-HCH ratios in air vary appreciably in different regions of North America, ranging from 0.2 to 14 (Table 1). The continental average ratio is 3.6. A continental map of this ratio (Figure 2) reveals relatively high values along the Canadian coasts with the Pacific, Arctic, and Atlantic Ocean and low ratios in central Canada, the southeastern United States, and Central America. Relatively high ratios at some distance from the ocean are recorded around the Great Lakes (especially at the station to the northeast of Lake Superior) and in the mountains of western Canada. Enantiomeric Composition of r-HCH. The results of the chiral analysis of R-HCH, expressed using EFs, are presented in Table 1 and Figure 3. EFs for the two air samples from Belize and Costa Rica are not reported because the concentrations of R-HCH were too low for an accurate determination of its chiral composition. In some air samples, interferences prevented the determination of the EFs of R-HCH with the BGB-172 column. The EFs of R-HCH in air range from 0.471 to 0.519 and show a notable variation in different regions of North America. EFs in Canada west of Ontario are all higher than 0.5, indicating a relative higher abundance of (+)R-HCH, with particularly high values in British Columbia and the Canadian Rocky Mountains. EFs are also above 0.5 in the samples from the southeastern United States and the Mexican Yucatan Peninsula. EFs below 0.5, indicative of relative depletion of (+)-R-HCH, occur in eastern and Arctic Canada, with particularly low EFs in the Maritime Provinces and east of Lake Superior. Racemic signatures are found at the two northernmost stations (Alert and Eureka on Ellesmere Island in the Canadian High Arctic), in Toronto and surroundings, in Maryland, and in Tapachula in southwest Mexico. Regional Characteristics. The above observations clearly indicate some regional characteristics in terms of the distribution and relative abundance of HCHs. Western Canadian (British Columbia and the mountains) air is characterized by (+)-enantiomer-enriched R-HCH at levels corresponding to the continental average. Low levels of γ-HCH lead to relatively high R/γ ratios. The Canadian Prairies also have continental average concentration of R-HCH, which is again enriched in (+)-enantiomer. However, very high levels of γ-HCH cause very low R/γ ratios. In the Great Lakes region, both isomers have levels around the continental average, leading to R/γ ratios close to 1. The enantiomeric composition of R-HCH in Great Lakes air is highly inhomogeneous, with EFs above, below, and at 0.5. Atlantic Canada experiences high concentrations of R-HCH, which is strongly depleted in the (+)-enantiomer. At the same time low γ-HCH levels result in very high R/γ ratios in the Maritimes. Interestingly, a very similar pattern as observed in Atlantic Canada is also found at Pukaskwa to the northeast of Lake Superior. R-HCH concentrations in Arctic Canada are close to the continental average, and EFs are below 0.5 except in the very far north. High R/γ ratios in the Arctic result from very low levels of γ-HCH. Similar to the Great Lakes region, R- and γ-HCH levels in the southeast United States are close to average, and ratios are around 1. EFs there indicate some enrichment
FIGURE 2. r-/γ-HCH ratios in the North American atmosphere from summer 2000 to summer 2001. in the (+)-enantiomer of R-HCH. Finally, in Central America very low levels of R-HCH are observed. Since γ-HCH levels are low to average, R/γ ratios are also quite low.
Discussion In this study, a newly developed passive air sampling system was used to measure atmospheric levels of R-HCH and γ-HCH. To evaluate the performance of this technique for quantitatively determining volumetric air concentrations, PAS-derived air concentrations of R-HCH and γ-HCH were compared with data reported by other studies that had used active high-volume sampling. For comparison, the amounts of R-HCH and γ-HCH measured in the replicate samplers deployed for 1 yr were converted to annual mean volumetric air concentrations (in units of pg m-3). Table 2 lists the air concentrations from this study with those from other recent measurements for R-HCH and γ-HCH. The agreement between the PAS-derived air concentrations and those derived from high-volume sampling is good (i.e., comparable to the agreement between studies employing high-volume sampling). The exceptions are the stations at Burnt Island and Point Petre, where the PAS-derived air concentrations are higher than those obtained by high-volume sampling, which is possibly due to a breakthrough of HCHs on the polyurethane foam plugs used for high-volume sampling at these sites (18). The average deviation of the PAS data from the high-volume data for the remaining sites is 30%. This gives us confidence that the patterns observed with the PAS are a true reflection of the large-scale spatial variability in air concentrations.
It is an interesting challenge to explain the regional differences in the distribution, relative abundance and enantiomeric composition of the HCHs across the North American atmosphere. The chemical parameters controlling atmospheric fate are quite similar for the two isomers. Both have similar volatility (24), as reflected in similar vapor pressure and octanol-air partitioning coefficients (Table 3). γ-HCH has a slightly lower air-water partitioning coefficient than R-HCH (Table 3), suggesting that it is scavenged somewhat more efficiently. Both react slowly with atmospheric oxidants, and atmospheric lifetimes based on the reaction with OH radicals are in the range of a few months (R-HCH, 120 d; γ-HCH, 96 d) (25). Calculations of long-range transport potential by TaPL3 2.10 (15), ELPOS 1.1.0 (26), and Chemrange 2.1 (27, 28) indicate a very high potential for widespread dispersal of both isomers in both the atmosphere and oceans (Table 3). The atmospheric transport distance for R-HCH is somewhat higher than that of γ-HCH because of a lower atmospheric reaction rate and less efficient rain scavenging. The differences in partitioning and persistence between the two isomers is however not sufficient to explain the widely divergent distribution patterns on a continental scale. Instead, the patterns must have their root in the location and timing of release into the global environment, which is very different for R- and γ-HCH. Emission clearly helps to explain the atmospheric distribution of the γ-isomer. Lindane is still widely used in North America, particularly in Canada. It is used mostly for treatment of seeds, especially canola seed (2). The Canadian Prairie Provinces, in particular Saskatchewan, are major VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
969
FIGURE 3. Enantiomeric composition of r-HCH in the North America atmosphere from summer 2000 to summer 2001.
TABLE 2. Comparison of Annual Mean Air Concentrations for r-HCH and γ-HCH (pg m-3) Derived from Passive Air Sampling Data with Other Recent Measurements concentration (pg m-3) region Arctic Lake Superior Lake Huron Lake Ontario Alabama Belize
location
date
r-HCH
γ-HCH
ref
Alert Alert Pukaskwa Brule River Eagle Harbor Burnt Island Burnt Island Point Petre Point Petre Muscle Shoals Muscal Shoals Belmopan Belmopan
summer 2000-summer 2001 1998 summer 2000-summer 2001 1998 1998 summer 2000-summer 2001 1998 summer 2000-summer 2001 1998 summer 2000-summer 2001 October 1996-May 1997 summer 2000-summer 2001 summer 1996
35.8 44.60 81.1 78.19 83.11 37.4 23.06 47.9 21.25 56.4 92 ( 68 1.6
5.5 7.02 23.0 12.50 15.55 20.8 9.43 35.1 11.91 57.0 50 ( 26 6.0 44 ( 13
this study (14) this study (21) (21) this study (21) this study (21) this study (22) this study (23)
TABLE 3. Physical-Chemical Properties at 25 °C (24) and Estimated Long-Range Transport Potentiala for r- and γ-HCH CTD in km (TaPL3) R-HCH γ-HCH
pL (Pa)
log KOW
log KAW
log KOA
airb
0.245 0.0757
3.94 3.83
-3.53 -3.91
7.46 7.74
17946 9732
CTD in km (ELPOS)
SR in % (Chemrange)
waterc
airb
waterc
air
waterd
1549 2126
22307 12572
409 758
46.9 41.0
47.9 42.1
a CTD, characteristic travel distance; SR, spatial range. b Wind speed, 14.4 km/h. c Emission to water; water speed, 0.013 × wind speed. d Emission to air.
canola-growing areas (29). Lindane was used in the United States in the mid-1990s, and some of our sampling sites (stations 34-36) were close to the areas of use (e.g., Georgia) 970
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004
(30). We have no information on lindane use in the United States during the time period of sampler deployment. In agreement with this usage pattern, the three sampling stations
in the Canadian Prairies had significantly higher γ-HCH concentrations than other sites, and Bratt’s Lake, SK, had the highest level of γ-HCH among all samples. The results indicate that the Canadian Prairies are major sources of γ-HCH to the North American atmosphere (31). Although low levels of R-HCH and γ-HCH are found in air in Central America, the concentrations of γ-HCH are relatively high as compared to the very low concentrations of R-HCH, in particular in Tapachula, Mexico. Very low R/γ ratios might indicate some use of lindane in Central America. The lower levels of R-HCH and γ-HCH in Central America may be either due to lower amounts of lindane being used or more efficient removal from the tropical atmosphere (32). At first glance, emissions aid little in explaining the observed distribution of R-HCH in the North American atmosphere. Technical HCH usage, and thus the emission of R-HCH, was mostly concentrated in Asia. China, India, and the Soviet Union were among the major users of technical HCH (33). Yet the highest levels of R-HCH are observed in the Maritime Provinces of Canada, the part of North America furthest removed from the Asian source regions. The western part of Canada, which is much closer to Asia and presumably could have been directly impacted by Asian sources (34, 35), has no elevated R-HCH concentrations. The reason γ-HCH shows highest levels in or in close proximity to source regions, whereas R-HCH does not, must be related to the fact that γ-HCH usage is recent and ongoing, whereas technical HCH usage has been phased out more than a decade ago. This suggests that the atmospheric concentrations of R-HCH are no longer determined by primary sources but are now mostly a result of re-evaporation from terrestrial and especially aquatic surfaces. We will show that this is supported by the spatial distribution of the R-HCH concentrations and by the enantiomeric signatures. If re-volatilization is controlling air concentrations, these should be highest where the potential for evaporation is highest. The rate of volatilization is a function of the concentration of the chemical in the surface medium (water/ soil/plants) and the environmental parameters influencing evaporation, most notably temperature and wind speed. We would expect higher rates of volatilization and thus higher air concentrations when surface concentrations, temperatures, and wind speeds are high. Knowledge of the variability in the R-HCH concentrations in surface media across North America is incomplete. However, it is known that concentrations in the terrestrial environment (plants and soils) generally tend to be very low (36-39). The work of Bidleman, MacDonald, Jantunen and co-workers (40-43) has further established that the oceans and large lakes can constitute major reservoirs for R-HCH. r- and γ-HCH in the Atmosphere of the Eastern Arctic and Atlantic Canada. Field investigations (42, 43), global modeling exercises (44), and budget calculations (43) have identified the Arctic Ocean in particular as the last major refuge for R-HCH on a global scale. R-HCH was dispersed globally and transferred to the Arctic Ocean during the period of technical HCH usage. Whereas it was rapidly lost from other world oceans after cessation of emissions as a result of degradation and evaporation, it was preserved in the Arctic Ocean because low temperatures and an ice cover greatly reduced these loss processes. Although net evaporation of R-HCH from the Arctic Ocean began in the mid-1990s (41), this process is slow and outflow with water through the Canadian Arctic archipelago has been identified as the major loss process of R-HCH from the Arctic Ocean (43). This water continues to flow further south along the east coast of Canada as the Labrador Current. Evaporation of R-HCH from Arctic Ocean water flowing south can explain the observations on the distribution of R-HCH in Arctic and Atlantic Canada.
FIGURE 4. Natural logarithm of the partial pressure of r-HCH and γ-HCH in air measured along the east coast of Canada against the reciprocal of annual mean water temperature. Along the northeastern coast of the North American continent, levels of R-HCH increase southwards from 6.8 ng PAS-1 on northern Ellesmere Island at 82° N to 32.1 ng PAS-1 around eastern Newfoundland at 43° N (Figure 1A). This increase is consistent with the steady increase in temperature along that transect, which should enhance the rate of volatilization of R-HCH. If the origin of the R-HCH in the atmosphere of eastern Canada is indeed temperature-driven air-water exchange, the measured partial pressures (p) should have a relationship with temperatures (T). Given a constant water concentration (Cw), the relationship between p and T can be expressed by ln p ) m/T + b, where the slope (m) is related to the phase transition energy and is believed to be an indicator of the extent to which atmospheric concentrations are controlled by temperature-dependent air-surface exchange versus long-range transport (45). To test this hypothesis, the annual mean air temperature for eight stations (stations 4-7 and 25-28) along the transport path of the Labrador Current and surface seawater temperature in proximity to each station were generated using ArcInfo 8 (ESRI). The sampler concentrations were converted to air concentrations (in pg m-3) as described above and then converted to partial pressures (in Pa) using the ideal gas law. The partial pressures of R-HCH were then regressed against the reciprocal of the annual mean water temperatures, which is considered to be close to that of the air-water interface (9). Even though Alert (station 3) and Eureka (station 24) are at the coast of the Arctic Ocean, the air concentrations from these two stations were not included in the regression because ice cover prevents volatilization almost year-round. The regressions between ln p and 1/T (Figure 4), which for comparison was also performed for γ-HCH, yielded slopes of -7119 (r 2 ) 0.54) for R-HCH and -2755 for γ-HCH (r 2 ) 0.14). It has been suggested that the slope of ln p versus 1/T can be used to interpret the relative importance of volatilization from local surfaces versus long-range transport (45). Atmospheric levels controlled by long-range transport show a shallow slope or less temperature dependence, whereas steeper slopes result from volatilization from local soils or water bodies. The regression results suggest that the levels of R-HCH are controlled almost exclusively by air-water exchange, whereas γ-HCH levels appear to be controlled to a much larger extent by transport than by vaporization from seawater in the region. If the water concentration is assumed to be constant in the waters of the Labrador Current, it is possible to estimate an apparent enthalpy of air-water exchange (∆AWH) from VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
971
TABLE 4. Water Concentrations (µg m-3) of r-HCH at Stations Close to the Atlantic Ocean Estimated Assuming Equilibrium with the Measured Air Concentration stations
T (K)
H (Pa m3 mol-1)
pexp (Pa)
Cw (µg m-3)
Cape Dorset Kuujjuaq Bonavista Daniel Harbor East Point Sable Island
274.2 274.7 275.9 276.5 279.4 283.1
0.100 0.105 0.118 0.124 0.162 0.227
7.1E-10 4.4E-10 1.4E-09 7.6E-10 9.4E-10 1.4E-09
2.1 1.2 1.8 3.4 1.7 1.8
the slope of the regression (Figure 4). The rate of R-HCH volatilization is small relative to the size of the water reservoir, which suggests that it is not unreasonable that the change in Cw resulting from volatilization is negligible relative to the change in p. The ∆AWH for R-HCH calculated from the field data is 59 kJ mol-1 and therefore virtually identical to the ∆AWH of 57 kJ mol-1 determined for R-HCH in artificial seawater over the temperature range of 0.5-23 °C (46). This means that the variability in the air concentrations of R-HCH along the coast of eastern Canada can be explained by equilibrium partitioning between the waters of the Labrador Current and the atmosphere. Strictly speaking, water concentrations in the Labrador Current can only be assumed constant, if the Arctic Ocean water is not diluted during its transport down to the Eastern Seaboard. Arctic Ocean water exiting through the archipelago had R-HCH concentrations of approximately 3 µg m-3 (47, 48). In addition to the outflow from Hudson Bay and the archipelago, the Labrador Current has a component that derives from the East Greenland Current and which mixes with water from the Arctic Ocean outflow near the top of Baffin Bay and within Davis Strait. On the basis of field measurements near Svalbard and the Greenland Sea (11, 49), this water likely had R-HCH concentrations somewhat below 1 µg m-3 and thus would dilute the Arctic Ocean water. Indeed, recent measurements suggest an R-HCH concentration of 1.5 ( 0.5 µg m-3 for Davis Strait (L. Jantunen, personal communication), consistent with a mixing of water from the Arctic Ocean and the East Greenland Current. We estimated equilibrium water concentrations of R-HCH from the air concentrations measured at stations 4-7, 27, and 28 using Henry’s law (46). Stations 25 and 26 are not included because the estimated water temperature close to these two stations is lower than 0.5 °C. Equilibrium partial pressures above seawater are assumed to be equal to measured partial pressures for the sampling sites. Table 4 summarizes water temperatures (T), Henry’s law constants (H), measured partial pressures (p), and estimated equilibrium water concentrations (Cw) of R-HCH at each sampling site. The water concentrations calculated based on air-water equilibrium are very close, only ranging from 1.2 to 3.4 µg m-3 and averaging 2 µg m-3 at the six locations. These estimated levels of R-HCH in surface water are thus what we would expect from a mixing of Arctic Ocean and East Greenland Current water and what has been measured in Davis Strait. This further confirms that air and seawater in the region are close to equilibrium with respect to R-HCH, which is also in agreement with fugacity ratios of R-HCH in the range of 0.81.26 (50). The chiral signatures provide further evidence that volatilization from the ocean is controlling R-HCH concentrations along the east coast of Canada. The appearance of nonracemic R-HCH in air is often explained by evaporation of partially degraded R-HCH from water. Jantunen and Bidleman (41, 49) observed selective depletion of (+)-R-HCH in Arctic Ocean water, resulting in an average EF in the eastern Arctic Ocean of 0.465 (50). Degradation half-lives for (+)972
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004
R-HCH and (-)-R-HCH in the eastern Arctic Ocean were reported to be 5.9 and 23.1 yr, indicating that (+)-R-HCH degrades faster than (-)-R-HCH (50). In our study, the air in eastern Canada was also found to be depleted in (+)-RHCH, with EFs in the Maritimes ranging from 0.470 to 0.480 (Figure 3). This is thus consistent with evaporation of R-HCH from seawater. That the EFs are higher than those in water is presumably a result of dilution of the nonracemic R-HCH originating from the seawater with R-HCH from elsewhere. This is also consistent with the observation that the enantiomeric composition of R-HCH along the East Coast becomes more and more racemic with increasing latitudes (Figure 3). This can be explained by considering the effect of temperature and ice cover. The relative importance of evaporation from the ocean water should increase with decreasing ice cover and increasing temperature. The enantiomeric signatures of R-HCH in air over frozen portions of the high Arctic (Alert and Eureka) show racemic composition (Figure 3), even though the surface water is not racemic (41, 49). At these latitudes, a permanent ice cover inhibits the evaporation of R-HCH from water, and R-HCH must instead derive from long-range atmospheric transport, which is often racemic, as photolysis and hydroxyl radical reaction are not enantioselective. Mixing of air masses that are enriched in the opposite enantiomers (e.g., air with EF >0.5 from the Bering Sea and air with EF
