Environ. Sci. Technol. 2006, 40, 120-129
PCB Congeners and Dechlorination in Sediments of Sheboygan River, Wisconsin, Determined by Matrix Factorization PHILIP A. BZDUSEK, JIANHANG LU,† AND ERIK R. CHRISTENSEN* Department of Civil Engineering and Mechanics and Center for Great Lakes Studies, University of Wisconsin-Milwaukee, Milwaukee, Wisconsin 53201
Nine sediment cores were collected from the Sheboygan River Inner Harbor, WI, and analyzed for polychlorinated biphenyl (PCB) congeners. Total PCBs ranged from ∼0 to 161 mg/g. Positive matrix factorization (PMF) was applied to the PCB data set to determine source profiles. Two factors were determined to be significant. One factor resembled the original approximated PCB mixture of 50% Aroclor 1248 and 50% Aroclor 1254 and the other factor was a dechlorinated version of the mixture. An anaerobic dechlorination model was applied to the dechlorinated source profiles to quantify possible dechlorination pathways. It was found that dechlorination process H′ provided the best fit for an individual process, and H′ + M provides the best fit for combined processes. PMF source contributions, and plots of PCB concentration versus congener for individual samples, provide evidence of enhanced dechlorination at high concentrations (>40 ppm) and small amounts of dechlorination at low concentrations ( doubly flanked para chlorine > singly flanked para chlorine > singly flanked meta chlorine > unflanked meta or para chlorine on di- or tri- substituted ring > isolated meta or para chlorine (23). Each category from the above chlorine reactivity sequence is assigned a value λ of 0.25, 0.40, 0.55, 0.70, 0.85, 1.0, respectively. This range of numbers provided sufficient bias for preferred reactions, but maintained some randomness. For λ ) 1 the reaction has no preference, and for λ ≈ 0 the reaction has high preference. The following equation was used to determine the preferential reaction order: 1/λ X1/λ 1 + X2 ) 1
coefficient of determination factorsc
Exner
IUPAC ) International Union of Pure and Applied Chemistry. Numbers indicate the position of chlorine atoms on each ring of biphenyl, with a hyphen representing separation of rings. c The coefficients used to find σ were C1 ) 0.001 nmol/g, C2 ) 0.25; df ) 106 × 32 - p(106 + 32); where p ) number of sources. a
b
(3)
where X1 is a random number, and X2 is a mapped random number, i.e., calculated based on a random number. The initial random reaction sequence is determined from X1 with the lowest random number representing the first reaction. The final reaction order is determined from X2 with the highest value of X2 representing the first reaction.
Results and Discussion Dating Analysis. Results of the dating analysis are described in Bzdusek et al. (45). In short, core SR1a, upstream of the Pennsylvania Avenue Bridge, and cores SR5-8, further downstream, show continuous sedimentation (1.2-11.8 cm/ yr) since the late 1950s, whereas net sediment accumulation virtually ceased after 1988 at the intermediate sites of cores
FIGURE 6. PCB source profiles obtained from PMF analysis. SR1-4. The sediment cores were dated by 137Cs and 210Pb methods. Sediment dating was supported by a PCB analysis and U.S. Army Corps of Engineers hydrographic surveys from 1976 to 2002. PCB Congener Profiles. The average PCB congener profile for the 106 samples from the Sheboygan River Inner Harbor is displayed in the left column of Figure 2. The profile is dominated by lower chlorinated congeners including 28/31, 25, 26, 16, and 15/17. This profile does not resemble the original Aroclor 1248 or 1254 profiles, nor a 50% 1248/50% 1254 mixture (Figure 2, right column). Instead, it appears to be a dechlorinated profile originating from a mixture of VOL. 40, NO. 1, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
125
FIGURE 7. Source contributions generated by PMF.
TABLE 3. Average Values and Standard Deviations of S for 100 Reaction Sequences Considering 7 Single Dechlorination Processes and 5 Multiple Dechlorination Processes (Two-Factor Solution) dechlorinationa process
random reaction sequence
preferential reactionb sequence
H M H′ Q P N LP H′ + M H′ + Q H′ + P H′ + N H′ + LP
9698 ( 47 39292 ( 473 8540 ( 191 24223 ( 1348 14459 ( 156 46696 ( 75 46728 ( 479 4700 ( 1611 8922 ( 972 8052 ( 519 6486 ( 465 7476 ( 590
9709 ( 52 39219 ( 417 8557 ( 171 23737 ( 1351 14460 ( 163 46712 ( 78 46696 ( 421 3988 ( 1513 8831 ( 1086 8018 ( 469 6472 ( 484 7474 ( 512
a Various dechlorination processes based from Bedard (43). b Reaction sequences are altered according to eq 3 and Williams (23).
Aroclors 1248 and 1254. In both Aroclors 1248 and 1254, PCB congeners 25 and 26 have very low concentrations, however both are important contributors to the average sample PCB concentrations (Figure 2), indicating dechlorination. The percentage of Aroclors 1248 and 1254 in the original 1248/1254 mixture was determined by comparing the PCB profiles of samples with low concentrations to various combinations of the two Aroclors. By trial and error the mixture was estimated as 50 ( 5% GE Aroclor 1248 and 50 ( 5% GE Aroclor 1254 (44). The congener profile of the 50/50 mixture (Figure 2) clearly resembles the profiles of the lower concentrations (j 3 ppm) post-1973 samples in Figures 3 and 4. For example, samples SR1a-1, SR1a-3, and SR4-3 are all similar to the 50/50 mixture with an increase in some lower chlorinated congeners, such as PCBs 25 and 26, indicating some dechlorination even at concentrations 39 ppm and represent dechlorinated profiles. Samples SR4-7 (39 ppm) and SR4-15 (42 ppm) have relatively higher concentration (or abundance) of highly chlorinated
congeners when compared to those of SR4-9 (85 ppm), SR411 (59 ppm), SR4-13 (70 ppm), and SR4-14 (114 ppm), an indication of enhanced dechlorination at higher concentrations, which is consistent with the literature. For example, laboratory studies by Sokol et al. (46) demonstrate a clear threshold concentration between 35 and 45 ppm below which no dechlorination is observed. Above the threshold concentration dechlorination was a function of sediment PCB concentration. Perhaps some of the minor dechlorination at lower concentrations (SR4-5, 5 ppm) is a result of the proximity of dechlorinating organisms at higher concentration samples (SR4-7-14). Plots of PCB congener versus concentration for various samples from core SR7 are shown in Figure 5. Sample SR7-1 (2.0 ppm) has an unusual profile with a high concentration of PCB congener 18. Samples SR7-4 (1.9 ppm), SR7-5 (0.3 ppm), and SR7-6 (1.6 ppm) resemble the original mixture (Figure 2). SR7-7 (6.1 ppm) and SR7-8 (5.3 ppm) have higher concentrations of PCBs and PCB congener 18. This PCB maximum coincides with remediation dredging in the Upper Sheboygan River in 1989 and 1990. It is apparent from the PCB maximum that PCB congener 18 is preferentially released during dredging and has accumulated downstream. This is consistent with a study by Bergen et al. (17) who found VOL. 40, NO. 1, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
127
evidence of low molecular weight PCBs being transported the furthest during remediation of the New Bedford Harbor. Also, note the difference between the profiles of the samples impacted by dredging (SR7-7 and SR7-8, Figure 5) and dechlorinated samples such as SR4-9-SR4-14 (Figure 4). In dechlorinated samples PCB congener 28/31 is the primary PCB congener, while PCB 18, the primary contributor to dredged samples, is less important. Diagnostic Tools. Results of COD and Exner function are shown in Table 2. A significant improvement in both COD and Exner function are observed from one-factor to twofactors. From two-factors to three-factors the only significant improvement is PCB congener 18, the COD of which increases from 0.06 to 0.83. The improvement in the Exner function from 0.28 to 0.25 is small. Thus, based on the diagnostic tools a two-factor solution is sufficient to adequately represent most congeners. PMF Model PCB Source Profiles. Results of PMF modeling indicate two significant factors, which are plotted in Figure 6. A three-factor PMF solution provided no additional information separating loading 2 of 2 into two similar factors. The PMF generated source profiles in Figure 6 display two distinct profiles. Loading 1 of 2 (Figure 6) is a slightly dechlorinated version of the original 50:50 Aroclor 1248/ Aroclor 1254 mixture (Figure 2). Loading 2 of 2 is a dechlorinated version of the original mixture. Notice the similarity between the loading 1 of 2 profile and the lower concentration sample profiles such as SR1a-1, SR4-3, SR712, etc. (Figures 3-5). Loading 2 of 2 is very similar to high concentration dechlorinated samples, such as SR4-9-SR414 (Figure 4). The pattern of low concentration samples representing the original Aroclor mixture and high concentration samples representing the dechlorinated profiles is consistent throughout the PMF source contributions. PMF Model PCB Source Contributions. PMF source contributions are shown graphically in Figure 7. Score 1 of 2 represents the original mixture and is present in most samples, especially the lower concentration samples. For example, cores SR7 and SR8 are represented almost entirely by score 1 of 2 and the maximum concentration in these cores is 6 ppm. Other lower concentration contributions are noted from samples SR2-3-SR2-13, SR6-1-SR6-13, SR4-15, etc. On the other hand, score 2 of 2 follows the total PCB profile with peaks at the highest concentrations. For example, SR1-4-SR1-8 have concentrations greater than 50 ppm as do SR4-8-SR4-14. This finding is consistent with a laboratory study by Sokol et al. (46) and a field study of Lake Hartwell sediments (24). Anaerobic Dechlorination Model Application Results. Results of the anaerobic dechlorination model runs are tabulated in Tables 3 and 4 and displayed in Figure 8. From Table 3 it is apparent that dechlorination processes H′ and H are the best single dechlorination processes and H′ + M is the best dechlorination process combination. Dechlorination process N (Figure 8) is important for only higher chlorinated congeners, while dechlorination process M alone is unable to improve the higher chlorinated congeners fit or produce the products 25 and 26. Process H′ works well for both higher and lower chlorinated congeners and is complemented well with process M. If preferential reaction sequences are considered process H′ + M has an even better fit (S ) 4700 f S ) 3988). This improvement can be seen in Figure 8 (bottom plots) for congener 26. For all other single and combination dechlorination processes the preferential reaction sequence has very little impact (Table 3). The reactions that are primarily responsible for the improved fit (using preferential reaction sequences) involve PCB congener 70 (25-34). When H′ is the only dechlorination process only one reaction with PCB 70 is possible PCB 70 (25-34) f PCB 26 (25-3). Thus, the reactions PCB 70 f PCB 128
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 1, 2006
26 and PCB 67 (245-3) f PCB 26 (25-3) are able to sufficiently supply PCB 26 (25-3) as seen in Figure 8 (top right corner). However, when combined with process M, PCB congener PCB 70 (25-34) can dechlorinate to PCB congeners PCB 33 (34-2) and PCB 31 (25-4). Given a preference for the flanked para reaction PCB 70 f PCB 26 the average number of chlorines removed for this reaction increases from 29.6 to 38.8 (Table 4), while for the flanked and unflanked meta reactions PCB 70 (25-34) f PCB 33 (2-34) and PCB 70 (2534) f PCB 31 (25-4) they decrease, 16.7 to 13.6 and 21.1 to 13.9, respectively. The five most important dechlorination reactions include PCB 66 (24-34) f PCB 25 (24-3), PCB 18 (25-2) f PCB 4 (2-2), PCB 44 (23-25) f PCB 16 (23-2), PCB 118 (245-34) f PCB 67 (245-3) and PCB 67 (245-3) f PCB 26 (25-3) (Table 4). The processes that dechlorinate the reactions are mixed between H′ and M and the reactive chlorines are either unflanked meta or flanked para. Some potentially toxic monoortho congeners (PCB 118 (245-34) f PCB 67 (245-3) f PCB 26 (25-3), are dechlorinated to the lower chlorinated PCB congeners 25 and 26 (47, 48). Reduction of dioxin-like toxicity is demonstrated through the reaction PCB 110 (23634) f PCB 59 (236-3).
Acknowledgments This work was supported by the U. S. National Science Foundation grant BES-0107402.
Literature Cited (1) David, M.; Brondyk, L. M.; Sonzogni, W. C. Distribution of PCB congeners in Sheboygan River (Wisconsin) sediments. J. Great Lakes Res. 1994, 20, 510-522. (2) Imamoglu, I.; Li, K.; Christensen, E. R.; McMullin, J. K. Sources and dechlorination of polychlorinated biphenyl congeners in the sediments of Fox River, Wisconsin. Environ. Sci. Technol. 2004, 38, 2574-2583. (3) Imamoglu, I.; Li, K.; Christensen, E. R. Modeling polychlorinated biphenyl congener patterns and dechlorination in dated sediments from the Ashtabula River, Ohio, USA. Environ. Toxicol. Chem. 2002, 21, 2283-2291. (4) Brown, J. F. J.; Bedard, D. L.; Brennan, M. J.; Carnahan, J. C.; Feng, H.; Wagner, R. E. Polychlorinated biphenyl dechlorination in aquatic sediments. Science 1987, 236, 709-712. (5) Quensen, J. F., III; Tiedje, J. M.; Boyd, S. A. Reductive dechlorination of polychlorinated biphenyls by anaerobic microorganisms from sediments. Science 1988, 242, 752-754. (6) Bedard, D. L.; May, R. J. Characterization of the polychlorinated biphenyls in the sediments of Woods Pond: Evidence for microbial dechlorination of Aroclor 1260 in situ. Environ. Sci. Technol. 1996, 30, 237-245. (7) Sokol, R. C.; Kwon, O.-S.; Bethoney, C. M.; Rhee, G.-Y. Reductive dechlorination of polychlorinated biphenyls in St. Lawrence River sediments and variations in Dechlorination Characteristics. Environ. Sci. Technol. 1994, 28, 2054-2064. (8) Pakdeesusuk, U.; Lee, C. M.; Coates, J. T.; Freedman, D. L. Assessment of natural attenuation via in situ reductive dechlorination of polychlorinated biphenyls (PCBs) in sediments of the Twelve Mile Creek Arm of Lake Hartwell, SC. Environ. Sci. Technol. 2005, 39, 945-952. (9) Tiedje, J. M.; Quensen, J. F., III; Chee-Sanford, J.; Schimel, J. P.; Boyd, S. A. Microbial reductive dechlorination of PCBs. Bioremediation 1993, 4, 231-240. (10) Quensen, J. F., III; Mousa, M. A.; Boyd, S. A.; Sanderson, J. T.; Froese, K. L.; Giesy, J. P. Reduction of aryl hydrocarbon receptormediated activity of polychlorinated biphenyl mixtures due to anaerobic microbial dechlorination. Environ. Toxicol. Chem. 1998, 17, 806-813. (11) Alder, A. C.; Haggblom, M. M.; Oppenheimer, S. R.; Young, L. Y. Reductive dechlorination of polychlorinated biphenyls in anaerobic sediments. Environ. Sci. Technol. 1993, 27, 530-538. (12) Bedard, D. L.; Bunnell, S. C.; Smullen, L. A. Stimulation of microbial para-dechlorination of polychlorinated biphenyls that have persisted in Houstanic River sediment for decades. Environ. Sci. Technol. 1996, 30, 687-694. (13) Sanders, G.; Hamilton-Taylor, J.; Jones, K. C. PCB and PAH dynamics in a small rural lake. Environ. Sci. Technol. 1996, 30, 2958-2966.
(14) Muir, D. C. G.; Omelchenko, A.; Grift, N. P.; Savoie, D. A.; Lockhart, W. L.; Wilkinson, P.; Brunskill, G. J. Spatial trends and historical deposition of polychlorinated biphenyls in Canadian midlatitude and Artic lake sediments. Environ. Sci. Technol. 1996, 30, 3609-3617. (15) Gevao, B.; Hamilton-Taylor, J.; Murdoch, C.; Jones, K. C.; Kelly, M.; Tabner, B. J. Depositional time trends and remobilization of PCBs in lake sediments. Environ. Sci. Technol. 1997, 31, 32743280. (16) Chiarenzelli, J. R.; Alexander, C.; Isley, A.; Scrudato, R.; Pagano, J.; Ramirez, W. Polychlorinated biphenyls in nonaccumulating, century-old sediments: sources, signatures, and mechanism of introduction. Environ. Sci. Technol. 2001, 35, 2903-2908. (17) Bergen, B. J.; Rahn, K. A.; Nelson, W. G. Remediation at a marine Superfund site: surficial sediment PCB congener concentration, composition, and redistribution. Environ. Sci. Technol. 1998, 32, 3496-3501. (18) Willman, E. J.; Manchester-Neesvig, J. B.; Armstrong, D. E. Influence of ortho-substitution on patterns of PCB accumulation in sediment, plankton, and fish in a freshwater estuary. Environ. Sci. Technol. 1997, 31, 3712-3718. (19) Sanders, G.; Jones, K. C.; Hamilton-Taylor, J.; Dorr, H. Historical inputs of polychlorinated biphenyls and other organochlorines to a dated lacustrine sediment core in rural England. Environ. Sci. Technol. 1992, 26, 1815-1821. (20) Elder, J. F.; James, R. V.; Steuer, J. J. Mobility of 2,2′,5,5′tetrachlorobiphenyl in model systems containing bottom sediments and water from the lower Fox River, Wisconsin. J. Great Lakes Res. 1996, 22, 697-706. (21) Chiarenzelli, J. R.; Scrudato, R. J.; Wunderlich, M. L. Volatile loss of PCB Aroclors from subaqueous sand. Environ. Sci. Technol. 1997, 31, 597-602. (22) Achman, D. R.; Hornbuckle, K. C.; Eisenreich, S. J. Volatilization of polychlorinated biphenyls from Green Bay, Lake Michigan. Environ. Sci. Technol. 1993, 27, 75-87. (23) Williams, W. A. Microbial reductive dechlorination of trichlorobiphenyls in anaerobic sediment slurries. Environ. Sci. Technol. 1994, 20, 630-635. (24) Bzdusek, P. A.; Christensen, E. R.; Lee, C. M.; Pakdeesusuk, U. PCB congeners and dechlorination in sediments of Lake Hartwell, South Carolina determined from cores collected in 1987 and 1998. Environ. Sci. Technol. 2006, 40, 109-119. (25) Paatero, P. Least squares formulation of robust nonnegative factor analysis. Chemom. Intell. Lab. Syst. 1997, 37, 23-35. (26) Paatero, P.; Tapper, U. Positive matrix factorization: a nonnegative factor model with optimal utilization of error estimates of data values. Environmetrics 1994, 5, 111-126. (27) Chueinta, W.; Hopke, P. K.; Paatero, P. Investigation of sources of atmospheric aerosol at urban and suburban residential areas in Thailand by positive matrix factorization. Atmos. Environ. 2000, 34, 3319-3329. (28) Song, X. H.; Polissar, A. V.; Hopke, P. K. Sources of fine particle composition in the northeastern US. Atmos. Environ. 2001, 35, 5277-5286. (29) Kim, E.; Hopke, P. K.; Edgerton, E. J. Source identification of Atlanta aerosol by positive matrix factorization. Air Waste Manage. Assoc. 2003, 53, 731-739. (30) Kim, E.; Hopke, P. K.; Larson, T. V.; Covert, D. S. Analysis of ambient particle size distributions using unmix and positive matrix factorization. Environ. Sci. Technol. 2004, 38, 202-209. (31) U. S. EPA. Superfund Record of Decision: Sheboygan Harbor & River, Sheboygan, WI; EPA/ROD/R05-00/030; U.S. Environmental Protection Agency: Washington, DC, 2000. http:// www.epa.gov/oerrpage/superfund/sites/rods/fulltext/ r0500030.pdf. (32) Wisconsin Department of Natural Resources. Sheboygan River food chain and sediment contaminant assessment, Final Project Report; WI DNR: Milwaukee, WI, 2000. http://www.epa.gov/ glnpo/sediment/FoodChain/finalrpt.pdf.
(33) U. S. Army Corps of Engineers. Annual report/contract dredging report; Detroit District, Project Operations Section: Detroit, MI, 2003. Available online at http://www.lre.usace.army.mil/ OandM/text/sheboy.pdf. (34) Van Camp, R. P. Sedimentation Patterns of Lake Erie and the Ashtabula and Black Rivers, Ohio, USA. M. S. Thesis, 80 pp., University of Wisconsin-Milwaukee, Milwaukee, WI, 1999. (35) Li, A.; Ab Razak, I.; Christensen, E. R. Toxic organic contaminants in the sediments of the Milwaukee Harbor Estuary, Phase III, Kinnickinnic River sediments. Final Report to U. S. Army Corps of Engineers; University of Wisconsin-Milwaukee: Milwaukee, WI, 1995. (36) U. S. EPA. Test Methods for Evaluating Solid Waste, Physical/ Chemical Methods; SW846, Method 8082; U.S. Environmental Protection Agency: Washington, DC, 1996. Available online at http://www.epa.gov/epaoswer/hazwaste/test/pdfs/8082.pdf. (37) Lawson, C. L.; Hanson, R. J. Solving Least Squares Problems; Prentice-Hall: Englewood Cliffs, NJ, 1974. (38) Magar, V. S.; Johnson, G. W.; Brenner, R. C.; Quensen, J. F., III; Foote, E. A.; Durell, G.; Ickes, J. A.; Peven-McCarthy, C. Longterm recovery of PCB-contaminated sediments at the Lake Hartwell Superfund site: PCB dechlorination. 1. End-member characterization. Environ. Sci. Technol. 2005, 39, 3538-3547. (39) Rachdawong, P.; Christensen, E. R. Determination of PCB sources by a principal component method with nonnegative constraints. Environ. Sci. Technol. 1997, 31, 2686-2691. (40) Imamoglu, I. PCB Sources and Degradation in River Sediments Determined by Receptor Modeling. Ph.D. Dissertation, Department of Civil Engineering and Mechanics, University of Wisconsin-Milwaukee, Milwaukee, WI, 2001. (41) Malinowski, E. R. Factor Analysis in Chemistry, 2nd ed.; WileyInterscience: New York, 1991. (42) Bedard, D. L.; Quensen, J. F., III. Microbial reductive dechlorination of polychlorinated biphenyls. In Microbial Transformation and Degradation of Toxic Organic Chemicals; Cerniglia, C. E., Ed.; Wiley-Liss Inc.: New York, 1995; pp 127-216. (43) Bedard, D. L. Polychlorinated biphenyls in aquatic sediments: Environmental fate and outlook for biological treatment. In Dechlorination: Microbial Processes and Environmental Applications; Bossert, I. D., Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2003; pp 443-465. (44) Frame, G. M.; Cochran, J. W.; Bowardt, S. S. Complete PCB congener distributions for 17 Aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congener-specific analysis. J. High Resolut. Chromatogr. 1996, 19, 657-668. (45) Bzdusek, P. A.; Lu, J.; Christensen, E. R. Evidence of fine-grained sediment transport and deposition in Sheboygan River, Wisconsin, based on sediment core chemical tracer profiles. Water Resour. Res. 2005, 41, W04016. (46) Sokol, R. C.; Bethoney, C. M.; Rhee, G.-Y. Effect of Aroclor 1248 concentration on the rate and extent of polychlorinated biphenyl dechlorination. Environ. Toxicol. Chem. 1998, 17, 1922-1926. (47) Safe, S. Toxicology, structure-function relationship, and human and environmental health impacts of polychlorinated biphenyls: progress and problems. Environ. Health. Perspect. 1993, 100, 259-268. (48) Safe, S.; Safe, L.; Mullin, M. J. Polychlorinated biphenyls: congener specific analysis of a commercial mixture and a human milk extract. Agric. Food Chem. 1985, 33, 24-29.
Received for review January 14, 2005. Revised manuscript received October 18, 2005. Accepted October 21, 2005. ES050083P
VOL. 40, NO. 1, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
129
