Freely Dissolved Concentrations and Sediment-Water Activity Ratios

Oct 24, 2008 - Corresponding author phone: +47-22023159; fax: +47-22230448; e-mail: [email protected]., †. Stockholm University. ... POM-17 was employed in...
0 downloads 12 Views 806KB Size
Environ. Sci. Technol. 2008, 42, 8733–8739

Freely Dissolved Concentrations and Sediment-Water Activity Ratios of PCDD/Fs and PCBs in the Open Baltic Sea G E R A R D C O R N E L I S S E N , * ,†,‡ KARIN WIBERG,§ DAG BROMAN,† HANS PETER H. ARP,‡ YLVA PERSSON,§ KRISTINA SUNDQVIST,§ AND PER JONSSON† Department of Applied Environmental Sciences (ITM), Stockholm University, 10691 Stockholm, Sweden, Norwegian Geotechnical Institute (NGI), Oslo, Norway, and Department of Chemistry, Umeå University, Umeå, Sweden

Received July 3, 2008. Revised manuscript received September 12, 2008. Accepted September 12, 2008.

Aqueous concentrations of polychlorinated dibenzo-p-dioxins and -furans (PCDD/Fs) as well as polychlorinated biphenyls (PCBs) in the open sea have heretofore been measured by filtering and extracting large amounts of water. Measurement of freely dissolved concentrations with this technique is difficult because of corrections for sorption to dissolved organic matter. In this study we use a novel, more economic technique using equilibrium passive samplers consisting of 17-µm thin polyoxymethylene(POM-17),capableofmeasuringfreelydissolved aqueous concentrations (CW) in pristine (i.e., background) locations. POM-17 was employed in an extensive field campaign at five stations in the open Baltic sea to obtain CW at two depths (1 m above the seafloor and 25 m below the surface). Median CW in the overlying water was 2.3 pg toxic equivalents (TEQ)/m3 PCDD/Fs and 15 pg/L sum 7-PCB, with generally less than a factor two variation among sites and depths. Also freely dissolved concentrations of native compounds in the surface sediment porewater (CPW) were determined in laboratory batch experiments. The data were used to derive sediment-water activity ratios, which indicate the diffusive flux direction. It was found that the PCDD/Fs and PCBs were in close equilibrium between the sediment porewater and the overlying water. Comparison of CPW with total sediment concentrations indicated that more than 90% of the compounds were sorbed to sedimentary black carbon.

Introduction Common practice in the measurement of in situ aqueous concentrations of PCBs and polychlorinated dibenzo-pdioxins and -furans (PCDD/Fs) is the filtration of large amounts of water over a glass fiber filter followed by extraction of the total aqueous concentrations CW,total (e.g., refs 1, 2). The freely dissolved aqueous concentration, CW, which is of relevance to bioavailability, is then deduced by accounting * Corresponding author phone: +47-22023159; fax: +47-22230448; e-mail: [email protected]. † Stockholm University. ‡ Norwegian Geotechnical Institute. § Umeå University. 10.1021/es8018379 CCC: $40.75

Published on Web 10/25/2008

 2008 American Chemical Society

for the fraction of contaminants associated with dissolved organic carbon (DOC), CDOC. However, this deduction can be problematic because CDOC values are often not measured directly but are estimated using literature DOC-water distribution coefficients. Due to this potential bias, as well as the cost and difficulty in applying such methods in remote areas, the in situ deployment of equilibrium passive samplers to determine CW presents an attractive alternative (3-6). Because of the relative novelty of equilibrium passive samplers, their usage for field-measurement of in situ CW of PCDD/Fs and PCBs remains uncommon, despite their promising potential. Thus far, they have never been used for PCDD/Fs, and only been used for PCBs at two relatively contaminated harbor locations (4, 6). In the present study, we measured CW of PCBs and PCDD/Fs in remote, background locations in the open Baltic Sea, by in situ deployment of equilibrium passive samplers consisting of ultrathin POM (17 µm; POM-17). In the remainder of this paper CW is used to denote freely dissolved aqueous concentrations. Freely dissolved concentrations in more than one environmental compartment can be used to derive chemical activity gradients and therefore diffusive flux directions between the environmental compartments in which they were deployed (6-9). For the presently studied sedimentwater systems, activity ratios were determined by measuring CW and freely dissolved sediment porewater concentrations, CPW. Knowledge of such activity rations can then be used to identify locations where the sediment acts as a source or a sink of organic pollutants. Sediment-water activity ratios have been studied for PAHs (6, 9) and PCBs (6) at contaminated harbor locations. As far as we know, sediment-water chemical activity ratios have not yet been determined for PCBs at background locations, or for PCDD/Fs at any location. Strong sorption to black carbon (BC) materials such as soot and charcoal can lead to elevated total organic carbon (TOC)-water distribution ratios of hydrophobic organic compounds (2, 7, 10). In the present study, measured CPW values were used along with TOC/BC contents and total sediment PCDD/F and PCB concentrations to calculate sorption coefficients of native PCBs and PCDD/Fs to sediment total organic carbon (TOC) and its BC. Though such sorption coefficients are available for a limited data set of PCDD/Fs and PCBs in contaminated harbor sediments, here a larger set of such data for sediments at pristine locations was established.

Materials and Methods Materials. All solvents used were of glass-distilled purity (Burdick and Jackson, Muskegon, MI). The PCDD/F calibration and internal standards and the calibration standards of PCBs 77, 81, 126, and 169 were obtained from Wellington Laboratories (Ontario, Canada). The remaining PCB calibration standards were from AccuStandard (New Haven, CT), and all PCB internal standards were from Cambridge Isotope Laboratories (Andover, MA). The internal standards included 17 13C-labeled PCDD/Fs (all 2378-substituted ones) and 14 13C-labeled PCBs. The quantification was made by the isotope dilution method and by using the native (12C) calibration standards, which included seventeen PCDD/Fs (all 2378substituted) and fourteen PCBs (same as above). Study Sites. Studies were carried out at seven locations (I-VII) in the open Baltic Sea (Table 1, Figure 1), at depths of 25-122 m. POM studies (aqueous phase) were performed at five stations (I, III, IV, V, VI); sediment studies were performed at five stations (I, II, III, VI, VII). At three stations (I, III, and VI) both POM and sediment studies were done VOL. 42, NO. 23, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8733

TABLE 1. Coordinates and Water Depth at the Seven Open Baltic Sea Stations; TOC, BC, OC Fractions f (%) of Sampled Sediments (Single Measurements); BC/TOC Ratios, Sum 7-PCB Contents, Sum PCDD/F Contents. Also Indicated What Measurement Was Carried out (POM, Sediment, Or Both)

c

I

II

III

IV

V

VI

VII

measurement (POM/sediment)

POM + sed.

sed.

POM + sed.

POM

POM

POM + sed.

sed.

latitude (N) longitude (E) water depth (m) fTOC (%) fBC (%) fOC (%) BC/TOC (%) sum 7-PCBb (ng/kg dw) sum PCDD/Fc (ng TEQ/kg dw)

61°10.25′ 18°06.30′ 83 3.20 0.10 3.10 3.13 1300 10

58°59.90′ 19°08.70′ 115 3.90 0.16 3.74 4.11 4700 23

58°51.84′ 19°34.53′ 108 5.55 0.27 5.28 4.86 8700 32

58°51.72′ 19°34.00′ 122 nma nma nma nma nma nma

58°51.99′ 19°34.34′ 115 nma nma nma nma nma nma

61°54.19′ 17°35.66′ 71 2.35 0.09 2.26 3.83 800 7.6

58°53.60′ 18°58.05′ 115 5.90 0.20 5.70 3.39 14000 43

a nm, Not measured since only POM exposure took place at this site. Sum of 2,3,7,8-substituted PCDD/Fs (16 measured congeners).

FIGURE 1. Sampling locations in the open Baltic Sea. and activity ratios could thus be determined. Sampling was conducted during March-August 2007, and temperatures in the field were 6-10 °C during the exposure (11). Before sampling, a measuring grid of approximately 500 × 500 m was run by means of echosounding at each sampling site to describe the topographical conditions in the area before the exact position for sediment sampling was chosen. Sediment Sampling and Extraction. Surface sediment samples (0-2 cm) were taken with a modified Ponar sampler. Prior to extraction of PCDD/Fs and PCBs, the samples were 8734

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 23, 2008

b

Sum of PCBs 28,52,101,118,138,153, and 180.

homogenized using a commercial blender and weighed (12 g) into precleaned cellulose extraction thimbles, after which internal standards were added. Extraction was carried out in Soxhlet-Dean-Stark by refluxing with 100 mL toluene for at least 15 h. Tetradecane (40-50 µL) was added as a keeper. After reflux a solvent exchange from toluene to n-hexane was carried out. Deployment of Thin POM Samplers. Additive-free polyoxymethylene (POM; Astrup AS, Oslo, Norway) was sliced to 1 cm wide, 17 µm thick strands on a lathe equipped with a high-precision razor blade (9). Concentration-independent compound-specific equilibrium partitioning constants (Kvalues) for 55 µm thin POM, KPOM-55, were established for PCDD/Fs and PCBs in previous studies (SI Tables S1, S2 4, 12). Since the POM-17 is cut from the same material batch as POM-55, we assumed the established KPOM-55-values could be used for POM-17. POM-55 was found to reach equilibrium in the field for all 16-EPA PAHs within 3-4 weeks at 5-10 °C and at current velocities of 1.5 cm/s (4). The same study revealed that KPOM was not significantly different between 8 and 20 °C. Therefore, KPOM-55 measured at 20 °C was used for the data interpretation of POM-17 exposed in water at depths greater than 25 m, where year-round temperatures are around 6-10 °C (11). However, some temperature dependence of KPOM may be expected on the basis of earlier studies on other passive sampler materials (e.g., ref 5). For field exposure, POM-17 was attached to 0.5 m wide rectangular steel frames that were held at 1 m above the sediment (“bottom” water, “B”) and at 25 m below the water surface (“25 m”) by underwater buoys. The whole systems were moored to the bottom by 15 kg weights. After exposure for 154-192 days, the systems were retrieved with the help of 500 m long dragging lines, and 2.0 g POM (9 m) was sampled from all field stations. In all cases, biofouling of the polymer materials was limited. The passive samplers were extracted by horizontal shaking (230 rpm; 96 h) with n-hexane (250 mL) containing internal standards, and purified and analyzed as described below. Extraction efficiencies were tested earlier in recovery tests (13) and revealed that 80-120% of the test compounds were extracted from POM-500 within 48 h (probably much faster for POM-17). Field equilibration times of PCDD/Fs in thin POM were checked by exposure of POM55 for different time periods (179, 270, and 363 days) in the Frierfjord, Norway, under comparable temperature and water current conditions. Sediment BC and TOC Contents. Total organic carbon (TOC) contents of the sampled sediments were determined with catalytic combustion elemental analysis at 1030 °C after microacidification (1 M HCl) to remove inorganic carbonates. Black carbon (BC) contents were determined using the same

method on small combusted samples (10 mg) that were ballground to optimize oxygen access. Samples were combusted at 375 ( 1 °C for 18 h under abundant oxygen access, following exactly the procedures outlined by Gustafsson et al. (14). Total Sediment Concentrations. Nondried sediment (5 g) mixed with anhydrous Na2SO4 (2-5 g dry weight) was extracted with toluene (400 mL) and internal standard in a Soxhlet-Dean-Stark for 17 h. The extracts were cleaned and analyzed as described below. Freely Dissolved Porewater Concentrations. Sediment (nondried, 20 g dry weight) was shaken horizontally (180 rpm; 70 days; 10 ( 1 °C; 250 mL all-glass flasks) in the laboratory with distilled water (250 mL), POM-17 (2 g), NaN3 (1 g/L) and NaCl (concentrations similar to Baltic Sea water, 6 g/L). No PCDD/Fs or PCBs were added as we studied native in situ compounds. POM/sediment ratios were comparable to those in earlier studies (7, 9, 12, 13, 16) but higher than in solid-phase microextraction experiments (3). It was assumed that 70 days was long enough for POM equilibration, as (i) shorter equilibration times (10 days) were needed for 6-ring PAHs using a thicker POM-55 sampler (9), (ii) only 14 days was required for equilibrium with PAH, PCBs, and PCDD/Fs using 51 µm thick LDPE (7), a sampler that sorbs approximately 5-10 times stronger than POM (9), and (iii) 23-30 days of shaking was found to be sufficient even for thicker POM (500 µm)-sediment mixtures (13). Also on a theoretical basis, it can be argued that 70 days is sufficient to equilibrate slowly and very slowly desorbing fractions (desorption rate constants 10-3 and 10-4-10-5 h-1, respectively (15)) with water and POM, since 70 days is sufficient to desorb all of the slowly desorbing fraction, and 2-20% of the very slow one. After equilibration, the POM strips were sampled, cleaned, and extracted as described above. Extract cleanup and analysis were done as described below. Freely dissolved sediment pore water concentrations (CPW) were deduced from the PCDD/F and PCB contents in the POM samplers and measured KPOM values. PCDD/F and PCB depletion of the sediment phase by the passive samplers was quantified with the amount sorbed to the POM and the total sediment contents. This depletion was 1.0 ( 0.6% for PCDD/Fs and 3 ( 1% for PCBs, only exceeding the oftenused 5% criterion (7, 16) for PCB-28 in two sediments and PCB-52 in one. Moreover, depletion was corrected for in the calculation of CPW. Even a hypothetical significant depletion of the AOC will probably not influence the measured CPW, since a novel equilibrium will be established between POM and AOC (16). We did not correct for any salting out due to the 0.6% salt concentration (0.1 M), as this influence would be minor (around 0.03 log unit). Cleanup and Analysis. For the sediment extracts, four open liquid chromatographic columns and one step for reducing the sulfur content were used for the cleanup. First, a prewashed (n-hexane) multilayer column with KOH-silica (3 g), neutral silica (3 g), 40% H2SO4-silica (6 g), and Na2SO4 (3 g) was employed (Silica 60, Merck, Darmstadt, Germany), and the analytes were eluted with 60 mL n-hexane. Thereafter, sulfur was removed by adding activated copper pretreated by hydrochloric acid and washed with methanol. The second column was as the first column, but included only half a mounts of the adsorbents. In the third column, the extracts were passed over a mixture of activated carbon (AX21): Celite (carbon from Anderson Development Co., MI and Celite 545 from Fluka, Switzerland), and the poly-ortho and the monoortho PCBs were collected in the first fraction by forward elution with 40 mL n-hexane:dichloromethane (1/1 v/v) and PCDD/Fs and non-ortho PCBs were collected in the second fraction by reversed elution using toluene (40 mL). Finally, a miniaturized version of the first and second column was employed, and recovery standards (four 13C-labeled PCDD/ Fs and two 13C-labeled PCBs) were added. Clean-up and

fractionation of the POM samples required only the third and fourth open liquid columns (and no sulfur removing step). More details about the open column procedures are found in Danielsson et al. (17). Instrumental analyses were performed using an HRGC/HRMS system (an Agilent 6890N GC with Waters Autospec Ultima MS (Milford, MA) equipped with a 60 m × 0.25 mm Rtx-5MS or a SLB-5 ms column (Restek, Bellefonte, PA; Supelco, Bellefonte, PA) with a 0.25 µm thick film. For better chromatographic resolution of some of the PCDD/F congeners, all samples were also run on an SP-2331 column (Supelco, same dimensions). The limit of detection (LOD) was calculated as three times the average area of noise in GC-MS chromatograms. The LOD varied from 0.3 to 3 pg/sample for the various PCDD/F and PCB congeners. Results for 1234789-HpCDF are not reported due to insufficient signal-to-noise ratios. QA/QC. Average recoveries of the internal standards were 92 ( 10% for PCDD/Fs and 92 ( 12% for PCBs. Laboratory blanks generally contained