Research Polychlorinated Naphthalenes in Swedish Background Air A N N A - L E N A E G E B A¨ C K , * U L L A W I D E Q V I S T , U L F J A¨ R N B E R G , A N D LILLEMOR ASPLUND Institute of Applied Environmental Research (ITM), Stockholm University, SE-106 91 Stockholm, Sweden
Polychlorinated naphthalenes (PCNs) with four to eight chlorines were studied in air collected at two background stations in Sweden, one southerly and one northerly. Air was sampled with a high-volume sampler, and gas-phase adsorbents and filters were analyzed separately. The sum of TeCNs to HxCNs in the gaseous phase ranged between 1 and 10 pg/m3 with significantly higher concentrations at the southern location Hoburgen. HpCNs and OCN were below the detection limit. The highest concentrations were found in two samples from Hoburgen with the air masses coming from SW and W and during warm weather (+11 °C). The lowest concentration was found in a sample from the northern location Ammarna¨ s at cold weather (-22 °C) when the air came from the east. A correlation was found between logP and 1/T indicating that temperature has a larger effect than location on the concentration in the gas phase. The TeCNs constituted 5075% in the gaseous phase. In most filter samples TeCNs and PeCNs were below the limit of quantitation. PUF samples with air trajectories from W to NE had relatively higher concentrations of late eluting TeCNs and PeCNs, while in samples with winds from SE to S the early eluting congeners dominated. Samples with early eluting congeners were mainly collected at lower temperature. TriCNs constituted the dominant homologue group both in the gaseous and particulate phase of air samples as well as in bulk deposition from a rural monitoring station south of Stockholm. The octanol-air partition coefficient described the gas/particle interaction well for samples collected at temperatures down to -8 °C.
Introduction Polychlorinated naphthalenes (PCNs) comprise 75 compounds. Technical PCN mixtures known as e.g. Halowax and Nibren wax have physical and chemical properties similar to those of polychlorinated biphenyls (PCBs). They have been widely used in a variety of applications such as capacitor fluids, electrical insulators, and engine oil additives and for impregnation of wood, paper, and textiles, as flameretardants and insecticides (1). PCN exposure has been associated with chloracne and liver damage in humans (1). Toxicological studies have shown that several of the PCN congeners have a dioxin-like activity. This activity strongly depends on the molecular structure, and the most potent * Corresponding author phone: +46 8 674 71 63; fax: +46 8 674 76 37; e-mail:
[email protected]. 10.1021/es030165i CCC: $27.50 Published on Web 08/25/2004
2004 American Chemical Society
congeners are 1,2,3,4,6,7-HxCN, 1,2,3,5,6,7-HxCN, and 1,2,3,4,5,6,7-HpCN (2, 3). PCNs are released to the environment from waste incineration processes (4-6), slag residue from copper ore smelters (7), and from some chloro-alkali processes (8) in addition to the emission from products. They are also present as impurities in PCB mixtures (9, 10). Due to their common occurrence, PCNs are widespread contaminants and have been detected in different environmental media, both far from as well as near industrial areas (11-19). PCNs have also been detected in urban and rural air (20-23) and in air from the east Arctic Ocean (24). PCNs have been found in biological samples and sediment in the Swedish environment (8, 11, 15). They have also been detected in human milk (25) and in plasma from people with varying intake of fish (26). Long-range atmospheric transport (LRAT) has been recognized as an important pathway for several persistent organic pollutants, e.g. PCBs, PCDDs/PCDFs, and pesticides. These pollutants can be air transported from source regions at low latitudes over very long distances reaching areas even as remote as the Arctic (27). PCNs in air were studied to examine differences in concentrations, congener composition, and gas/particle distribution between one location in southern Sweden influenced by air from continental, industrialized areas, and a more unpolluted northern location. Both locations were considered to have minor or no influences from local sources. Samples collected during stable weather situations were selected to reveal any significant compositions of PCNs, which could be related to the origin of the air and thus indicate specific source areas. The PCN congener composition in bulk deposition collected at Aspvreten, a rural monitoring station, was compared with air samples.
Experimental Section Sampling. Air samples were collected at two meteorological observation stations, one at the southern point of Gotland, Hoburgen (H) (56°55′ N, 18°09′ E) in the central Baltic Proper and the other in Ammarna¨s (A) (65°49′ N, 16°43′ E) situated in the north of Sweden near the Swedish mountains and close to the polar circle. Sampling was performed simultaneously at both stations, once a week, between September 1990 and March 1991, giving a total of 25 samples from each station. Air trajectories at 850 hPa (72 h) were obtained from the Swedish Meteorological and Hydrological Institute (SMHI) for both sampling sites, twice during each sampling occasion (Figure 1). Seven samples from each location were selected for analysis. Additional air and bulk deposition samples, collected for the Swedish Monitoring Program at a background station, Aspvreten, situated about 100 km south of Stockholm (58°48′ N, 17°24′ E), were also analyzed. Sampling data are given in Tables 1 and 2. Sample H7F was lost during cleanup and thus not analyzed. Air sampling was carried out using a high-volume sampler equipped with glass fiber filter (Millipore, AP) 290 mm diameter, followed by two polyurethane foam plugs (PUF, 78 × 75 mm, F ) 22-25 kg/m3, D.P. Sunde, Norway). PUFs were precleaned before use by Soxhlet extraction in toluene for 24 h followed by acetone for 48 h and filters were precleaned by heating to 450 °C for 48 h. Samples were collected for about 48 h, giving 1400-2200 m3 of air. The first PUF was analyzed for PCN gas-phase concentrations, and the second one was checked for breakthrough. The filters VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4913
FIGURE 1. Air trajectories for seven sampling occasions at Hoburgen (H) and Ammarna1 s (A).
TABLE 1. Collection Data for Seven Sampling Occasions at Hoburgen (H) and Ammarna1 s (A) sample
date
H1 A1 H2 A2 H3 A3 H4 A4 H5 A5 H6 A6 H7 A7
900921-23 901005-07 910125-27 910201-03 910208-10 910301-03 910308-10
wind air volume direction (m3) 1754 1459 2004 1456 1994 1485 2160 1500 2160 1440 2050 1434 1896 1414
temp °C mean (min, max)
SW 10.7 (9.0, 13.6) E to NE 2.1 (0.6, 4.4) W 11.8 (10.0, 13.5) N to W 3.2 (0.6, 6.8) NW to W 2.6 (0.6, 4.0) NW to W -5.0 (-13.0, 0.6) circle (E) -2.2 (-5, 1.8) circle (W) -13.3 (-25.8, -5.4) SE to S -3.0 (-5.2, -1.6) E -22 (-30.8, -13.4) E to S 0.8 (0.0, 1.2) E to SE -7.8 (-3.0, -10.8) S 1.7 (1.2, 2.5) S -6.7 (-9, -3.8)
were analyzed separately. Deposition samples were collected with an open, PTFE coated sampler (bulk sampler), with a collection area of 1 square m. The sampler had the shape of a gently sloping funnel with 10 cm edges (28). Rainwater was passed through two PUFs on which the organic compounds were trapped. After each collection period, the surface of the sampler was wiped with clean glass fiber filters, which were included in the sample. Samples were stored in a -18 °C freezer, exclusively used for these samples until cleanup and analysis. Chemicals and Reference Compounds. Chemicals. nHexane and acetone were glass-distilled and checked for 4914
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004
purity. Dichloromethane was of HPLC quality (Lab-Scan Ltd.), and sulfuric acid was of 98% purity (BDH). Reference Compounds Used for Quantitation and Method Evaluation. TriCN congeners (13, 15, 16, 17, 20, 21, 22, 24, 26), TeCN congeners (42, 46), and PeCN-52 were kind gifts from Prof. U. A. Th. Brinkman, Free University Amsterdam, The Netherlands. PCN congeners 66, 67, 71, 73, and 75 were generous gifts from Dr. Eva Jacobsson and Prof. A° ke Bergman, Department of Environmental Chemistry, Stockholm University, Sweden (29). The technical products Halowax 1014 and 1099 were produced by Koppers Chemical (U.S.A.). Internal and Injection Standards. 13C12-Labeled 3,3′,4,4′TeCB (CB-77), 3,3′4,4′,5-PeCB (CB-126), and 3,3′,4,4′,5,5′HxCB (CB-169) (Cambridge Isotope Laboratories Woburn, MA), 2,2′,5,6′-TeCB (CB-53), 2,2′,4,5′,6-PeCB (CB-103), 2,3,3′,4,5,5′-HxCB (CB-159), and 2,3,3′,4,4′,5,5′-HpCB (CB189) were of 99% purity solution, purchased from Larodan Fine Chemicals. Extraction and Cleanup. Air PUF plugs were Soxhlet extracted for 24 h with 450 mL of n-hexane and filters with 450 mL of acetone. Prior to extraction, known amounts of about 1 ng of each 13C12-labeled chlorobiphenyls were added as internal standards to the adsorbents and filters. Other internal standards were added for determining the recovery of other groups of compounds eluting in the different fractions described below. Deposition PUF plugs and filters were Soxhlet extracted together in 450 mL of acetone. A volume of 800 mL of water was added to the extract, followed by liquid-liquid extraction twice with hexane, 120 and 60 mL, respectively. Extract volumes were reduced by rotary evaporation to 5 mL and treated with the same volume of sulfuric acid. The acid phase was re-extracted once with n-hexane. All laboratory glasswares were heated to 450 °C prior to use. Extensive fractionation was performed to enable the analysis of several compound classes. This study presents the results for PCNs. The cleanup used was a modification of a method described by Jansson et al. (30). Gel Permeation Chromatography (GPC). To isolate the chlorinated paraffins (CP) from the target compounds, extracts were fractionated on a high-performance size exclusion GPC system, using two serially coupled gel permeation columns (PL-GEL, 5 µm, 50 Å, 300 × 7.5 mm, Polymer Laboratories, U.K.). The HPLC system was provided with an UV detector set at 260 nm. The extracts were reduced to 50 µL in a gentle stream of nitrogen before injection. The columns were eluted with n-hexane:dichloromethane (1:1), 0.7 mL/min, as a mobile phase. Two fractions were manually collected with PCNs, PCBs, and pesticides eluting in the second (24-45 min). Silica Gel. PCNs and PCBs were separated from the pesticides on an open silica gel column (10 mm i.d. 4.5 g, Kiselgel 60, particle size 0.040-0.063 mm, Merck) according to a method described earlier (31). The second GPC fraction was reduced by nitrogen evaporation to a volume of 0.5 mL. PCNs and PCBs were eluted with 30 mL of n-hexane followed by 40 mL of hexane/diethyl ether (3:1) containing the pesticides. PYE. TeCNs- to OCN and non-ortho CBs were isolated from the bulk of PCBs by fractionation of the first silica gel fraction on two HPLC columns in series, packed with 2-(1pyrenyl)ethyldimethyl silylated silica (Cosmosil 5-PYE, 150 × 4.6 mm, particle size 5 mm, Nacalai Tesque, Japan). The mobile phase was n-hexane, saturated with water. Prior to injection the extracts were reduced to 20 µL with nitrogen in a test tube with a narrow end. Four fractions were collected, the flow rate was 0.5 mL/min for the first two fractions, and the two last were back-flushed at 1.2 mL/min. Fraction 4 contained the TeCNs to OCN, and TriCNs eluted in fractions
TABLE 2. Collection Data for Air, Deposition, and Sediment Samples sample i.d. SA SF SD1 SD2
air, PUF, pooled air, filter, pooled bulk deposition, pooled bulk deposition, mainly snow
1, 2, and 4. The chromatographic conditions are summarized in Supporting Information Table S1. Prior to GC/MS analysis CB-103 was added to fractions 2 and 4, and CB-189 to fraction 1 as injection standards. The volume was reduced to 20 µL of which 1 µL was injected. Air and deposition samples from Aspvreten (Stockholm) were fractionated on silica gel and PYE-columns. HRGC/HRMS. The analyses were performed on a Finnigan MAT 95, high-resolution magnetic sector mass spectrometer (Thermo Quest) equipped with a HP 5890A (Hewlett-Packard) gas chromatograph. Gas chromatographic separation was performed on a 60 m DB-5 (5% phenyl, methyl silicone) fused silica capillary column (0.25 mm i.d., 0.25 µm film thickness, J&W Scientific), using a splitless injection (split closed for 2 min) at 260 °C. The oven was temperature programmed as follows: 90 °C (2 min) - 30 °C /min - 200 °C (2 min) - 2 °C/min - 285 °C. The mass spectrometer was operated in the electron impact (EI) mode with an electron energy of 70 eV. The ion source and the HRGC/HRMS interface were held at 260 °C and 280 °C, respectively. High-resolution data were acquired with (R ) 8000), using perfluorokerosene (PFK) for continuous calibration of the instrument. Selective Ion Monitoring (SIM) technique with four time windows was used, monitoring the exact masses of the two most intense ions in the molecular ion cluster. The ions monitored for target compounds and internal standards are given in Supporting Information Table S2. Quantitation. The identification of the compounds was based on HRMS data and the ratio between the two chloroisotopes. The retention times of the eluting peaks were compared to single CN reference compounds used for quantitation, congeners in the Halowax 1014 mixture and in a fly ash extract. The labeling of the compounds was done according to Wiedmann and Ballschmiter (32), see Supporting Information, Table S3. A complete congener specific quantitation was not possible due to the lack of several reference substances and the fact that some congeners remain unresolved. Quantitation was based on the sum of the peak heights of the two most intense ions of the molecular ion isotope cluster. Losses during extraction and cleanup were corrected for by using the recovery of the internal standards, the 13C12-labeled chlorobiphenyls for each individual sample. Reference compounds used for quantitation are presented in Supporting Information Table S2. In the air and deposition samples from Aspvreten the concentrations of TriCNs eluting in fractions 1 and 2 and TeCNs to OCN in fraction 4 from the PYE column were quantitated using the injection standards CB-189 (fraction 1) and CB-103 (fractions 2 and 4), respectively, for volume correction. The internal standards, added originally, could not be used for correction of recovery, since parts of the samples had been consumed for PCB analysis after silica gel cleanup and the exact fractions of them used for PCN analysis was not known. Instead, the average recoveries of the internal standards CB-53, CB-159, and 13C12labeled chlorobiphenyls added to the Hoburgen and Ammarna¨s samples and eluting in fractions 1, 2, and 4, respectively, from the PYE-column were used for that purpose.
date
sampling site
Aug-Oct 1995 Aug-Oct 1995 Aug-Oct 1995 Feb 1995
Stockholm, Aspvreten Stockholm, Aspvreten Stockholm, Aspvreten Stockholm, Aspvreten
Results and Discussion Recovery and Repeatability. The average recovery of the internal standards added before extraction, the 13C12-labeled CB-77, CB-126, and CB-169, were 83%, 91%, and 98%, respectively, with a relative standard deviation of 20%. The recoveries of CB-53 and CB-159 were 92% and 86%, respectively. Four subsamples of an extract of PUF plugs and filters, pooled from several air samples, were analyzed at different occasions to check the repeatability of the method. The relative standard deviations were 7-18% for the TeCNs, 1018% for the PeCNs, and 4-12% for the HxCNs. Detection Limits and Breakthrough. Two PUF plugs and two filters were cleaned, wrapped in aluminum foil, placed in capped cans, stored at the laboratory, and analyzed as blanks. The concentrations in blank samples were calculated by using an average of collected air volumes, 1730 m3. The instrumental detection limit (LOD), defined as the signalto-noise ratio ) 3, corresponded to 1 fg/m3. The amount of individual TeCNs and PeCNs in PUF blank plugs corresponded to 5-30 fg/m3 and 0-9 fg/m3, respectively. The concentrations of PCNs in the backup PUF plug were in the same range as in the blank sample, and the TeCNs constituted
