Evidence for Dechlorination of Polychlorinated ... - ACS Publications

May 8, 2012 - ... Polychlorinated Dibenzo-p-Dioxins and -Furans in Wastewater Collection Systems in the New York Metropolitan Area .... es300560q_si_0...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Evidence for Dechlorination of Polychlorinated Biphenyls and Polychlorinated Dibenzo-p-Dioxins and -Furans in Wastewater Collection Systems in the New York Metropolitan Area Lisa A. Rodenburg,*,† Songyan Du,† Hui Lui,†,‡ Jia Guo,† Nicole Oseagulu,† and Donna E. Fennell† †

Department of Environmental Sciences, Rutgers University, 14 College Farm Road, New Brunswick, New Jersey 08901, United States ‡ Department of Biochemistry and Microbiology, Rutgers University, 76 Lipman Drive, New Brunswick, New Jersey 08901, United States S Supporting Information *

ABSTRACT: Polychlorinated biphenyls (PCBs) and polychlorinated dibenzo-pdioxins and dibenzofurans (PCDD/Fs) are persistent organic pollutants targeted by the Stockholm Convention. Both contain aromatic chlorines and are subject to microbial dechlorination. Dechlorination of PCBs in sewers in the Delaware River basin was recently reported. In this work, two data sets on concentrations of PCBs and PCBs+PCDD/Fs in wastewater treatment plant influents and effluents were analyzed to look for evidence that these compounds undergo dechlorination in the sewers of the New York/New Jersey Harbor area. The two data sets come from the Contamination Assessment and Reduction Project (CARP) and were analyzed via Positive Matrix Factorization (PMF). Analysis of the data set containing only PCB concentrations suggests that PCBs are dechlorinated in the sewers of the NY/NJ Harbor via the same pathways observed in the sewers of the Delaware River basin and that advanced dechlorination of PCB mixtures is more likely to occur in combined sewers vs separate sanitary sewers. When the combined data set of PCBs+PCDD/Fs was analyzed, the factor containing PCB dechlorination products also contained high proportions of 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin (HpCDD), a known product of the dechlorination of octachlorodibenzo-p-dioxin (OCDD), and other known dechlorination products of PCDD/Fs. Despite being the most abundant PCDD/F congener in all of the samples in the database, OCDD was a minor component in the dechlorination factor. This provides the first evidence that PCDD/Fs may be dechlorinated in sewers.



INTRODUCTION Polychlorinated biphenyls (PCBs) are toxic, persistent organic pollutants that have been targeted for elimination under the Stockholm Convention.1 Although PCB use and manufacture have been banned in most countries, PCBs are still present at unacceptably high levels in many aquatic systems because they were produced in such large quantities and are extremely persistent. One of the few pathways for their transformation in the environment is dechlorination by anaerobic bacteria, which removes chlorines but does not destroy the PCB backbone.2 This process is usually desirable because it decreases PCB mass, and because it produces lightly chlorinated congeners that typically have lower toxicity and are more amenable to aerobic degradation and volatilization.2 Anaerobic dechlorination of PCBs has been studied extensively in the sediments of lakes, river, and harbors, but in these environments dechlorination is often too slow to significantly impact ambient concentrations of PCBs in the water column. Recently, we published evidence that PCBs are dechlorinated in a variety of environments other than aquatic sediments.3 Our analysis of data on PCB concentrations in effluents from dischargers on the Delaware River indicated that PCBs are © 2012 American Chemical Society

extensively dechlorinated in landfills and sewers as well as in groundwater at contaminated sites. Dechlorination in sewers is particularly fortuitous because of the large flows of wastewater and the large masses of PCBs that enter the sewer. Sewers act as collectors and aggregators of a wide variety of pollutants in urban areas. It has long been known that wastewater treatment plants (WWTPs) remove hydrophobic organic pollutants from wastewater by sequestering them in the sludge. Often 90% or more of the PCBs in the waste stream are removed in this fashion.4−6 Sediment material in sewers also sequesters hydrophobic organic pollutants such as PCBs. For example, Rocher et al.7 calculated that the gross bed solids in sewers contained 98% of the total inventory of polycyclic aromatic hydrocarbons. Sediments are expensive to remove from sewers and therefore are usually removed only when necessary. As a result, sediment can have a very long residence time in the Received: Revised: Accepted: Published: 6612

February 10, 2012 May 1, 2012 May 8, 2012 May 8, 2012 dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

Figure 1. Resolved factors from the PCB-only data set (data set 1).

sewer.8 Our research indicates that dechlorination of PCBs in the sewer, which probably occurs in the sewer sediment, serves as an additional process by which loads of PCBs to ambient waters are reduced. The sewer serves as an anaerobic bioreactor that pretreats the sewage, converting high molecular weight (MW) PCB congeners to lighter, less toxic congeners that may be degraded aerobically or volatilize in the WWTP. A better understanding of the factors that favor dechlorination in sewers would allow utilities to alter the design or management of

sewers to enhance this process, thereby reducing their loads of PCBs to receiving waters. Like PCBs, polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) are persistent organic pollutants targeted by the Stockholm Convention.1 PCBs and PCDD/Fs are structurally similar in that they both contain aromatic chlorines. Several studies indicate that bacterial strains that are capable of dechlorinating PCBs can also dechlorinate PCDD/Fs.9−11 One of the goals of this study was to look for evidence of PCDD/F dechlorination in sewers by analyzing a data set in which both 6613

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

Because of this, the PCB data set was more heterogeneous than the PCDD/F data set. For PMF analysis, two data sets were constructed. These are described in more detail in Supporting Information. The first data set contained PCB measurements in treated effluents and some influents from both the NY and NJ databases. The final PCB data set submitted for PMF analysis contained 64 congeners (or coeluting congener groups) in 149 samples. The second data set contained both PCB and PCDD/F measurements from the New York database only. New Jersey data were not included in this data set because 13 of the 17 PCDD/F congeners were below detection limit in virtually all of the NJ samples. Even in the NY data set, two of the 17 PCDD/F congeners had to be discarded because they were below detection limit in a majority of the samples: 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD) and 1,2,3,7,8,9-hexachlorodibenzofuran (HxCDF). The second data set therefore contained 15 PCDD/F congeners plus 50 PCB congeners in 65 samples. These data sets were examined via PMF 2.0 software.21 Loads of PCBs and PCDD/Fs to the New York/New Jersey Harbor from the WWTPs were calculated by multiplying the geometric mean concentration of each factor by the flow rate of each plant. Flow rates were obtained from the 2007 Interstate Environmental Commission Annual Report.18

PCBs and PCDD/Fs were measured in the same sewage effluent samples collected as part of the Contamination Assessment and Reduction Project (CARP).12 The CARP was initiated in the late 1990s to measure contaminants in suspected sources such as wastewater effluents and stormwater outfalls, as well as in the ambient water column and the sediments of the New York/New Jersey Harbor.12,13 Microbial dechlorination of PCDD/Fs can follow either the peridechlorination pathway, which removes chlorines at the 1, 4, 6, and 9 positions, producing 2,3,7,8-substituted products, or the peri-lateral pathway that removes chlorines at the 2, 3, 7, and 8 positions, resulting in the production of congeners that are not 2,3,7,8-substituted.11,14−16 For this reason, PCDD/F dechlorination is easier to detect when all of the 210 PCDD/F congeners are measured. Unfortunately, CARP, like many other monitoring programs, measured only the 17 2,3,7,8-substituted PCDD/F congeners, so that only the products of the peridechlorination pathway were measured. However, previous work by Barabas et al.17 has demonstrated that such data sets can reveal evidence of dechlorination. Thus, the CARP database provides a unique opportunity to look for evidence of dechlorination of PCDD/Fs in sewers. In this work, data from the CARP on PCB and PCDD/F concentrations in the effluents (and some influents) of WWTPs in the New York City metropolitan area were examined to look for evidence of PCB and PCDD/F dechlorination in sewers.





RESULTS PCB Data Set. For the PCB data set, the correct number of factors was determined to be seven (Figure 1). This conclusion was based on three lines of evidence. First, the 7-factor model yielded a stable model solution with an RSD of 1% for 9 PMF runs with seed values from 1 to 9. (Requesting more factors resulted in RSD values greater than 10%). Second, the seven factors were interpretable. Third, the 7-factor model adequately described the data set. The R2 value for the measured vs modeled concentrations was >0.83 for all congeners and was 0.99987 for ΣPCBs. Aroclor Factors. Based on the congener profiles of the Aroclors from Rushneck et al.,22 factors 2, 4, 5, and 7 strongly resembled Aroclors 1242, 1248, 1254, and 1260, respectively, with R2 values >0.9. These factors represented 7%, 11%, 11%, and 53% of the mass in the data set, respectively. Their relative importance is very different when examined on the basis of loads to the harbor: these four factors represent 22%, 13%, 37%, and 9% of the total load, respectively. The difference between the load and the mass in the data set arises from the importance of heavier Aroclors in the influent and CSO samples, which is related to the fact that the heavier Aroclors are more effectively removed by the WWTPs (see below). Factor 6 somewhat resembled a 70/30 mixture of Aroclors 1254 and 1260 (R2 = 0.77). It represented 12% of the mass in the data set, but only 7% of the load of PCB to the Harbor from these WWTPs. Dechlorination Factors. The remaining two factors appear to be dechlorination signals. Factor 1 is dominated by PCBs 4 (2-2; 46% of ΣPCBs) and 19 (26-2; 9%). (In this notation, numbers before the dash refer to the chlorine positions on ring 1, and the numbers after the dash refer to the chlorine positions on ring 2). Factor 3 is dominated by PCBs 43 + 52 + 73 (2352, 25-25, and 26-35 respectively), 44 + 47 + 65 (23-25, 24-24, 2356, respectively), and 45 + 51 (236-2, 24-26, respectively). These three coeluting congener groups constitute 9%, 13%, and 8% of the ΣPCBs, respectively. These results are very similar to

METHODS The CARP program was designed to identify the sources of a variety of pollutants to the sediments of the New York/New Jersey Harbor. The watershed of the harbor is home to about 20 million people and contains New York City, the largest city in the United States. The CARP involved sampling of air, water, sediment, and biota in the Harbor region from 1999 to 2004, including effluents and some influents to WWTPs that discharge into the harbor. 12 The harbor receives more than 2000 million gallons of treated effluent each day from more than 30 WWTPs.18 All of these WWTPs serve sewer systems that contain at least some combined sewers. The CARP data are publicly available by request to the Hudson River Foundation. The data were provided in several Microsoft Access databases, with the New York and New Jersey data in separate databases. The main challenge associated with analyzing the CARP data is the high degree of heterogeneity in the data set. CARP samples were collected by several organizations, including the New Jersey Department of Environmental Protection and the New York State Department of Environmental Conservation, using a variety of different sampling methods. PCDD/Fs and PCBs were not always measured in the same samples. For PCB analysis, some samples were analyzed as whole water samples, so that the reported data represented the dissolved plus particle phases. In other cases, the sampling protocol attempted to separate the dissolved and particulate fractions by pumping water through one (or two) filters and then through one (or two) columns containing XAD-2 resin. For PCDD/F analysis, typically only the filter samples (particle phase) were analyzed. Furthermore, samples were sent to at least four different contract laboratories for analysis via EPA method 161319 for the 17 2,3,7,8-subsituted PCDD/F congeners and method 1668A20 or equivalent for PCBs. Method 1668A allows two possible gas chromatography columns, and both columns were used by the various contract laboratories, which led to different congener coelution patterns. 6614

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

Figure 2. Comparison of the advanced and partial dechlorination factors isolated from the Delaware River Basin Commission (DRBC) and Contamination Assessment and Reduction Project (CARP) databases.

those obtained in our analysis of the PCB congener patterns in effluents of dischargers on the Delaware River,3 where two dechlorination signals were obtained (Figure 2). The first was dominated by PCBs 4 and 19 and was assumed to represent an advanced stage of dechlorination. The second was dominated by PCBs 44 + 47 + 65 and 45 + 51, and was interpreted as a sign of partial dechlorination. In the factors resolved from the CARP data set, PCB 4 dominates the advanced dechlorination signal even more than it did in the Delaware River Basin Commission (DRBC) advanced dechlorination factor. In contrast, PCBs 44 + 47 + 65 and 45 + 51 were less dominant in the CARP partial dechlorination factor than in the DRBC partial dechlorination factor (Figure 2). The relative weakness of the partial dechlorination factor in the CARP data set may be related to the fact that all of the WWTPs in the CARP area serve combined sewers. In our previous work,3 the partial dechlorination signal was most prevalent in effluents from WWTPs that serve separate sewer systems. We speculate that this is because separate sewer systems are less likely to experience methanogenic conditions, since they develop deep beds of cohesive sediment. In contrast, sulfidogenic conditions are more likely to prevail in separate sewers, which generally build up less sediment.23 These two dechlorination signals are present in all of the 32 WWTPs sampled (Figure 3), suggesting that the dechlorination occurred in the sewers and not in the anaerobic digesters employed at about half of these plants. Factors 1 and 3 averaged more than 5% of the sum of PCBs in a majority of the WWTPs sampled. Factor 1 comprised more than 10% of the ΣPCBs in eight of the WWTPs (Bowery Bay, Coney Island, Hunts Point, North River, Poughkeepsie, Rensselaer, and Rockland County municipal WWTPs, and the Fresh Kills Landfill Leachate Treatment Plant). In contrast, only five plants had more than 10% of factor 3, the partial dechlorination factor, in their effluents: Bowery Bay, Edgewater, Fresh Kills, Newtown Creek, and Red Hook. Figure 3 demonstrates that the CARP New York plants, which serve combined sewers, emit a larger fraction of advanced dechlorination products than most of the plants

serving separate sanitary sewers, such as most of the DRBC plants. This supports the conclusion in our previous work3 that advanced dechlorination is more prevalent in combined sewers. However, the fact that the CARP NJ plants (which also serve combined sewers) emit fewer dechlorination products than most of the other plants indicates that there are other, as yet unidentified, factors that influence the extent of dechlorination. These factors could include the age of the sewer system; the slope of the sewer, which governs the rate of flow of the sewage and therefore the rate of accumulation of sewer sediment; and the quality of the sewage itself (amount of carbon and other nutrients). In the CARP data set, the advanced dechlorination factor comprises 6.0% of the loads of PCBs to the harbor from these WWTPs, and the partial dechlorination factor comprises 6.4% of the loads. Our previous study found that 19% of the loads of PCBs to the Delaware River consisted of dechlorinated PCBs.3 Thus dechlorination is less important overall in the CARP data set. This arises partly because some of the plants with the largest flows have relatively low amounts of dechlorination products in their effluents. The three plants with the largest flow in the CARP data set (Passaic Valley Sewerage Commission, Ward’s Island, and Newtown Creek) comprise about 35% of the total flow, but less than 13% of the PCBs in their effluents consist of dechlorination products. Again, this suggests that there are other, as yet unidentified, factors that influence the extent of dechlorination. It is well documented that dechlorination of PCBs occurs in the sediments of the Upper Hudson River, and this dechlorination signal can be observed in the water column of the lower Hudson River/ CARP study area.24 However, this cannot account for the dechlorination observed in the treated effluents. This issue is discussed in more detail in the Supporting Information. Toxic Equivalency Quotients. PCBs are reasonably anticipated to be carcinogens and also display a range of noncancer health effects.25 The reduction in toxicity due to these many types of health effects is difficult to estimate. However, PCBs also display dioxin-like toxicity which can be quantified. The PCB data set contained five of the dioxin-like 6615

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

congeners (PCBs 77, 105, 118, 156 + 157, and 167), dechlorination to the advanced dechlorination signal represented a reduction in TEQ of more than 98% for all of the Aroclor factors, and reduction to the partial dechlorination signal resulted in reduction in TEQ of between 32% and 86% depending on the starting Aroclor. These calculations highlight the important impact of dechlorination in reducing not just mass but dioxin-like toxicity of PCB mixtures. Influent Samples. One sample of influent was collected at each of ten of the New York WWTPs. These samples were labeled as “CSO” samples in the database, since all of the New York WWTPs have combined sewer systems. As with the Delaware River data set, the CARP data reveal that the dechlorination signal is present in higher concentrations in the influent than in the effluent at all ten plants, suggesting the most or all of the dechlorination occurred in the collection system, i.e., the sewers. For each of these ten plants, the removal (R) of PCBs from the wastewater stream was calculated as ⎞ ⎛C R = ⎜ eff − 1⎟ ·100% ⎠ ⎝ C in

(1)

where Ceff is the average concentration of the factor in the effluent, and Cin is the concentration of the factor in the influent sample. For the sum of PCBs, the average (±standard deviation) removal was 95 ± 4%. The average (±standard deviation) removal for factors 1 through 7 was 63 ± 26%, 89 ± 7%, 89 ± 6%, 95 ± 4%, 89 ± 8%, 98 ± 2%, and 97 ± 4%. The removal of factor 1 was thus more variable that the removal of the other factors, and application of the F-test demonstrates that the % removal of factor 1 is significantly (p < 0.05) less than the removal of the other factors. It is possible that some dechlorination of PCBs occurred in the anaerobic sludge digesters employed at all 10 of these plants, which would increase the dechlorination products in the effluent and lead to lower calculated removal. However, if the lower removal of dechlorination products is instead due to the reduced hydrophobicity of the products, this has implications for the removal of PCB toxicity by the wastewater collection and treatment system. Even though dechlorination results in a significant decrease in TEQ, this may be partially offset by the greater mass of dechlorination products released to receiving waters due to the WWTP’s lesser ability to remove these dechlorinated congeners from the waste stream. PCB+PCDD/F Data Set. The correct number of factors derived from the combined data set containing 65 PCB and PCDD/F congeners in 65 samples was determined to be five (here designated factors A through E to avoid confusion with the PCB-only model). The five-factor model was chosen because it resulted in a reasonably low RSD of the G matrix of 8.7%, it produced meaningful and interpretable factors, and it satisfactorily described the data matrix. The R2 value for the measured versus modeled concentrations of the sum of all 65 analytes was 0.998, and was >0.8 for 62 of the 65 analytes. The R2 values for the three remaining analytes (PCB 4, PCBs 61 + 70 + 74 + 76, and PCB 194) increased to >0.8 when one or two outliers were excluded. The R2 value for OCDD was 0.88 and was better than 0.92 for the rest of the PCDD/Fs. Identification of Factors. Four of the five factors generated were dominated by PCBs, which comprised more than 99% of the mass in those factors (Figure 4). Only factor E was dominated by PCDD/Fs, which comprised 54% of its mass. As

Figure 3. Comparison of the percent of the ΣPCBs in the effluent that consisted of the advanced (white) and partial (black) dechlorination factors in each WWTP from the CARP and DRBC databases, sorted by state. Among the DRBC plants, only six serve combined sewer systems: the three plants of the Philadelphia Water Department (PWD), DELCORA, Wilmington, and Camden County Municipal Utilities Authority (CCMUA). All of the CARP plants serve combined sewers.

PCB congeners (PCBs 105, 118, 156 + 157, and 167). Based on the toxic equivalency factors (TEFs) recently adopted by the U.S. EPA,26 Factor 3 (the partial dechlorination signal) has the lowest toxic equivalency quotient (TEQ) of the seven factors. Dechlorination to factor 3 decreases the TEQ for the Aroclor factors by 13−92%. Dechlorination to factor 1 represents a 17−82% reduction in TEQ for most factors (although it actually represents an increase in TEQ for factor 7, the Aroclor 1260 factor). In contrast, the reduction in mass upon dechlorination is just 9−30% for the advanced dechlorination signal. Thus the reduction in TEQ is greater than the reduction in mass. In comparison, the reduction in TEQ for the factors isolated from the DRBC data set3 is even more dramatic. In that data set, which included six dioxin-like 6616

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

Figure 4. Resolved factors for the data set containing both PCB and PCDD/F congeners. PCBs are plotted as fraction of Σ50PCBs, and PCDD/Fs are plotted as fraction of Σ15PCDD/Fs. The percent of the total mass of each factor that consists of PCDD/F congeners is shown in the upper right of each panel.

Passaic River, which is within the New Jersey portion of the CARP study area. Because OCDD is the most abundant PCDD/F congener in the sewers, it is a likely substrate for dechlorination. The presence of a documented dechlorination product of OCDD, and the relative absence of the dechlorination substrate, OCDD, in the same factor that contains dechlorinated PCBs is a strong indication that dechlorination of OCDD occurs in sewers. Similarly, highly chlorinated PCDF congeners such as octachlorodibenzo-p-furan (OCDF) and the HpCDFs are unusually scarce in factor A. They comprise about 0.5% of Σ15PCDD/Fs in factor A, compared to >1.3% for the other four factors. Dechlorination of weathered OCDF with 1,2,3,4,7,8,9HpCDF postulated as the product has been observed in Kymijoki sediment both in situ in Finland as well as in microcosms developed from this sediment and spiked with OCDF.29 Lower chlorinated congeners including 1,2,3,6,7,8HxCDF and 1,2,3,6,7,8-HxCDD are more abundant in factor A. Barabas et al.17 likewise noted these two congeners as products of in situ dechlorination of PCDD/Fs in the Passaic River. Our results raise the possibility that the dechlorination observed by Barabas et al.17 in the Passaic River sediment occurred not in

in most environmental samples, octachlorodibenzo-p-dioxin (OCDD) was the most abundant PCDD/F congener in all of the samples, comprising 56−88% of the sum of the 15 PCDD/ F congeners in the data set (Σ15PCDD/Fs). Because of its abundance, factor analysis studies of PCDD/Fs frequently observe that OCDD dominates in all of the resolved factors.27,28 Its presence is therefore not useful in identifying the factor, and for this reason, OCDD is sometimes excluded from the data set.17,27 OCDD does dominate the PCDD/F congener pattern of four of the five PCB+PCDD/F factors, but it is unusually scarce in factor A, where it comprises just 3% of Σ15PCDD/Fs. Factor A is dominated by PCB 4 (23% of the total PCB+PCDD/F mass), indicating that it represents advanced dechlorination. Figure 4 demonstrates that the PCDD/F signal of factor A is dominated by 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin (HpCDD), which comprises 79% of Σ15PCDD/Fs in this factor. The ratio of 1,2,3,4,6,7,8-HpCDD to OCDD in factors B through E is less than 0.2, but is 23 in factor A, the dechlorination factor. Barkovskii and Adriaens14 observed 1,2,3,4,6,7,8- HpCDD as a product of the peri-dechlorination of OCDD in microcosms constructed using sediment from the 6617

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

In contrast to the PCB mixtures, we cannot estimate the change in TEQ upon dechlorination for the PCDD/Fs except to note that if dechlorination primarily converts OCDD (with a TEF of 0.0003) to 1,2,3,4,6,7,8- HpCDD (with a TEF of 0.01), then dechlorination increases the TEQ of the mixture. However, since dechlorination could produce non-2,3,7,8substituted PCDD/Fs which theoretically have no “dioxin-like” toxicity, dechlorination may actually reduce the overall toxicity of the PCDD/F mixture.11 Influent samples. PCB+PCDD/F concentrations in influent samples were available for four WWTPs: Coney Island, Newtown Creek, Owl’s Head, and Port Richmond. At these four plants, percent removal of the highest molecular weight factors (B, C, D, and E) was always greater than 85%. Removal of factor A, the advanced dechlorination factor, was more variable, ranging from 30% at Coney Island to 80% at Port Richmond. As noted above, the lesser removal of the dechlorination factor is probably related to the fact that it contains lower MW congeners that are generally more soluble, although there could be some contribution from dechlorination of PCBs and PCDD/Fs in the anaerobic digesters employed at all of these plants. As noted above, less efficient removal of dechlorination products during wastewater treatment leaves greater concentrations of these products in the effluents, which can counteract any reduction in toxicity achieved via dechlorination. Implications. The high degree of heterogeneity in the CARP database of PCB measurements is regrettable and limits the utility of the CARP data. The use of multiple contract laboratories is not the main driver of heterogeneity within the CARP PCB data. Our earlier analysis used data collected by facilities that discharge to the Delaware River and reported to the Delaware River Basin Commission (DRBC). The DRBC data set was likewise generated from multiple contract laboratories chosen by the dischargers. DRBC was able to generate a more homogeneous data set by constructing a thorough and thoughtful set of guidelines that each laboratory followed. These guidelines stipulated the GC column to be used in the analysis as well as the maximum acceptable detection limits, among other things. CARP suffered from a lack of cooperation among multiple state and federal agencies. Despite these limitations, analysis of the CARP data provided useful insights into the dechlorination of persistent organic pollutants in sewer systems. Taken together, analysis of these two data sets via PMF suggests that dechlorination of PCBs occurs in the sewers of the NY/NJ Harbor area via the same pathways observed in most of the sewer systems in the Delaware River basin. The results suggests that dechlorination of PCBs has only a small effect on the mass load of PCBs, since the reduction in mass of the PCB congeners due to dechlorination is small and partially offset by decreased removal efficiency during the waste treatment process. However, the reduction in TEQ for dioxin-like PCB congener is substantial, perhaps as large as 20% compared to similar emissions of unweathered PCB formulations. The wastewater collection systems in the NY/NJ Harbor area contain primarily combined sewers. Dechlorination of PCBs therefore appears to occur in virtually all combined sewer systems and most separate sewer systems. The analysis of the combined PCB+PCDD/F data set provides tantalizing evidence that PCDD/Fs are dechlorinated in sewers, and suggests that 1,2,3,4,6,7,8-HpCDD or the ratio of 1,2,3,4,6,7,8-HpCDD to OCDD may be used as a tracer of the dechlorination of PCDD/Fs. This observation is likely to be

the river but in the combined sewers that frequently overflow into the tidal Passaic. These combined sewers flow to the Passaic Valley Sewerage Commission (denoted as “PVSC” in Figure 3). 1,2,3,4,7,8- and 1,2,3,6,7,8-HxCDFs are also abundant in factor A; Adriaens and Grbic-Galic30 observed these congeners as products of the dechlorination of 1,2,3,4,6,7,8-HpCDF in microcosms using sediment from the Hudson River. Also, Liu and Fennell observed that Dehalococcoides ethenogenes strain 195 can dechlorinate 1,2,3,4,7,8-HxCDF via a peri-lateral pathway to produce non2,3,7,8-substituted congeners.11 Thus it is possible that 1,2,3,4,6,7,8-HpCDF is a dechlorination intermediate in the sewer system. In general, factor A contains a wider variety of lower molecular weight PCDD/F congeners than any of the other factors, suggesting that sequential dechlorination steps may produce a wide variety of PCDD/F congeners, and that OCDD is not the only substrate for dechlorination in the sewers. As noted above, only the peri-dechlorination products were measured in the CARP database. Evidence that dechlorination occurs by this pathway in sewers suggests that the peri-lateral dechlorination pathway may also occur in sewers. Also, dechlorination products with fewer than four chlorines may also be present, but were not measured in these samples. The PCB portions of the combined PCB+PCDD/F factors were compared with the Aroclors. Only factor D clearly resembled a single Aroclor (Aroclor 1260, R2 = 0.96). As a result, factor D also resembled factor 7 of the PCB-only model (R2 = 0.97). The other four factors did not resemble a single Aroclor or even a linear combination of Aroclors. Factor A resembled factor 1 of the PCB-only model (R2 = 0.78), since both were dominated by PCB 4 and represent advanced dechlorination. The other three PCB+PCDD/F factors did not strongly resemble any of the PCB-only factors. We speculate that this change in the composition of the factors suggests that when more than one class of contaminant is investigated, the PMF model is more likely to identify secondary sources, i.e., processes that transport the contaminants from their locations of release to the receptor point, rather than primary sources such as Aroclors. TEQs. For factors A, B, and C, the total TEQ was dominated by TEQ associated with the PCDD/Fs, even though they constituted less than 0.5% of the mass in those factors. The PCB associated TEQ in factors A, B, and C represented just 2.0%, 2.9%, and 6.3% of the overall TEQ respectively. The TEQ of factor D, which represents Aroclor 1260, was split equally between the PCB-associated TEQ and PCDD/Fassociated TEQ even though PCDD/Fs make up just 0.015% of the mass in this factor. Not surprisingly, the TEQ of Factor E, which consists of 54% PCDD/Fs by mass, is dominated by the PCDD/F-associated TEQ. The PCB-associated TEQ of this factor was just 0.02% of the total. Factor D, the Aroclor 1260 factor, had the lowest overall TEQ. Factor A, the advanced dechlorination factor, had the next lowest TEQ which was nine times higher than the factor D TEQ. The factor B TEQ was 32 times higher than factor D, the factor C TEQ was 44 times higher, and the TEQ of factor E, the mass of which is dominated by PCDD/Fs, was about 2600 times higher than the TEQ of factor D. 2,3,7,8-TCDD has the highest TEF among the PCDD/Fs, but since it was not included in this data set, the major source of TEQ among the PCDD/Fs was 1,2,3,7,8-PeCDD, usually followed by 1,2,3,4,6,7,8- HpCDD, and then by 1,2,3,6,7,8-HxCDD. 6618

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

dechlorinates diverse chlorinated aromatic pollutants. Environ. Sci. Technol. 2004, 38 (7), 2075−2081. (11) Liu, F.; Fennell, D. E. Dechlorination and detoxification of 1,2,3,4,7,8-hexachlorodibenzofuran by a mixed culture containing Dehalococcoides ethenogenes strain 195. Environ. Sci. Technol. 2008, 42 (2), 602−607. (12) Contamination Assessment and Reduction Project (CARP). Data Archive: Water, sediment and biota data collected from 1999−2003. CD-ROM; Hudson River Foundation: New York, 2007. (13) Contamination Assessment and Reduction Project (CARP). A Model for the Evaluation and Management of Contaminants of Concern in Water, Sediment, and Biota in the NY/NJ Harbor Estuary: Contaminant Fate & Transport & Bioaccumulation Sub-models; Hudson River Foundation: New York, 2007. (14) Barkovskii, A. L.; Adriaens, P. Microbial dechlorination of historically present and freshly spiked chlorinated dioxins and diversity of dioxin-dechlorinating populations. Appl. Environ. Microbiol. 1996, 62 (12), 4556−4562. (15) Fennell, D. E.; Nijenhuis, I.; Wilson, S. F.; Zinder, S. H.; Häggblom, M. M. Dehalococcoides ethenogenes strain 195 reductively dechlorinates diverse chlorinated aromatic pollutants. Environ. Sci. Technol. 2004, 38, 2075−2081. (16) Bunge, M.; Adrian, L.; Kraus, A.; Opel, M.; Lorenz, W. G.; Andreesen, J. R.; Gorisch, H.; Lechner, U. Reductive dehalogenation of chlorinated dioxins by an anaerobic bacterium. Nature 2003, 421 (6921), 357−360. (17) Barabas, N.; Goovaerts, P.; Adriaens, P. Modified polytopic vector analysis to identify and quantify a dioxin dechlorination signature in sediments. 2. Application to the Passaic River. Environ. Sci. Technol 2004, 38, 1821−1827. (18) Interstate Environmental Commission. 2007 Annual Report; New York, 2008. (19) USEPA. Method 1613: Tetra- through Octa-Chlorinated Dioxins and Furans by Isotope Dilution HRGC/HRMS; Washington, DC, 1994. (20) USEPA. Method 1668, Revision A: Chlorinated Biphenyl Congeners in Water, Soil, Sediment, and Tissue by HRGC/HRMS; Washington, DC, 1999. (21) Paatero, P. User’s Guide for Positive Matrix Factorization Programs PMF2 and PMF3. Part 1: Tutorial; 2003. (22) Rushneck, D. R.; Beliveau, A.; Fowler, B.; Hamilton, C.; Hoover, D.; Kaye, K.; Berg, M.; Smith, T.; Telliard, W. A.; Roman, H.; Ruder, E.; Ryan, L. Concentrations of dioxin-like PCB congeners in unweathered Aroclors by HRGC/HRMS using EPA Method 1668A. Chemosphere 2004, 54, 79−87. (23) Hvitved-Jacobsen, T. Sewer Processes: Microbial and Chemical Process Engineering of Sewer Networks; CRC Press: New York, 2003. (24) Rodenburg, L. A.; Du, S.; Xiao, B.; Fennell, D. E. Source apportionment of polychlorinated biphenyls in the New York/New Jersey Harbor. Chemosphere 2011, 83, 792−798. (25) Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for Polychlorinated Biphenyls (PCBs); U.S. Department of Health and Human Services, Public Health Service: Atlanta, GA, 2000. (26) U.S. Environmental Protection Agency. Recommended Toxicity Equivalence Factors (TEFs) for Human Health Risk Assessments of 2,3,7,8-Tetrachlorodibenzo-p-dioxin and Dioxin-Like Compounds; EPA/ 100/R-10/005; Washington, DC, 2010. (27) Barabas, N.; Adriaens, P.; Goovaerts, P. Modified polytopic vector analysis to identify and quantify a dioxin dechlorination signature in sediments. 1. Theory. Environ. Sci. Technol. 2004, 38, 1813−1820. (28) Bright, D. A.; Cretney, W. J.; MacDonald, R. W.; Ikonomou, M. G.; Grundy, S. L. Differentiation of polychlorinated dibenzo-p-dioxin and dibenzofuran sources in coastal British Columbia, Canada. Environ. Toxicol. Chem. 1999, 18 (6), 1097−1108. (29) Lui, H. Microbial reductive dechlorination of weathered polychlorinated dibenzo-p-dioxins and dibenzofurans in contaminated sediments. Rutgers, the State University of New Jersey, New Brunswick, NJ, 2010.

particularly useful in data sets that measured only the 17 2,3,7,8-substituted PCDD/F congeners. This analysis bolsters the general view that environmentally important microbial processes are occurring in sewers. Other anaerobic bacterial processes, such as additional reductive dehalogenation pathways and mercury methylation,31 may also occur in sewers.



ASSOCIATED CONTENT

S Supporting Information *

Details on the construction of data sets and the possible influence of dechlorination in the sediments of the Upper Hudson River. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: 732-932-9800 x 6218; fax: 732-932-8644; e-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Simon Litten for his help in understanding the CARP data. Simon served as the mastermind of the NY CARP data collection effort and recently retired from the NYSDEC. We wish him a blissful retirement.



REFERENCES

(1) United Nations Environment. Program Final Act of the Conference of Plenipotentiaries on The Stockholm Convention On Persistent Organic Pollutants; United Nations Environment Program: Geneva, Switzerland, 2001; p 44. (2) Bedard, D. L. Polychlorinated biphenyls in aquatic sediments: Environmental fate and outlook for biological treatment. In Dehalogenation: Microbial Processes and Environmental Applications; Haggblom, M. M., Bossert, I. D., Eds.; Kluwer Academic Publishers: Boston, MA, 2003. (3) Rodenburg, L. A.; Du, S.; Fennell, D. E.; Cavallo, G. J. Evidence for widespread dechlorination of polychlorinated biphenyls in groundwater, landfills, and wastewater collection systems. Environ. Sci. Technol. 2010, 44, 7534−7540. (4) Blanchard, M.; Teil, M. J.; Ollivon, D.; Legenti, L.; Chevreuil, M. Polycyclic aromatic hydrocarbons and polychlorobiphenyls in wastewaters and sewage sludges from the Paris area (France). Environ. Res. 2004, 95 (2), 184−197. (5) Katsoyiannis, A.; Samara, C. Persistent organic pollutants (POPS) in the sewage treatment plant of Thessaloniki, northern Greece: Occurrence and removal. Water Res. 2004, 38 (11), 2685− 2698. (6) Pham, T. T.; Proulx, S. PCBs and PAHs in the Montreal Urban Community (Quebec, Canada) wastewater treatment plant and in the effluent plume in the St Lawrence River. Water Res. 1997, 31 (8), 1887−1896. (7) Rocher, V.; Azimi, S.; Moilleron, R.; Chebbo, G. Hydrocarbons and heavy metals in the different sewer deposits in the Te Marais’ catchment (Paris, France): Stocks, distributions and origins. Sci. Total Environ. 2004, 323 (1−3), 107−122. (8) Ashley, R. M.; Hvitved-Jacobsen, T. Management of Sewer Sediments. In Wet-Weather Flow in the Urban Watershed; Field, R., Sullivan, D., Eds.; Lewis: New York, 2002. (9) Adrian, L.; Dudková, V.; Demnerová, K.; Bedard, D. L. Dehalococcoides sp. strain CBDB1 extensively dechlorinates the commercial polychlorinated biphenyl mixture Aroclor 1260. Appl. Environ. Microbiol. 2009, 75, 4516−4524. (10) Fennell, D. E.; Nijenhuis, I.; Wilson, S. F.; Zinder, S. H.; Haggblom, M. M. Dehalococcoides ethenogenes strain 195 reductively 6619

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620

Environmental Science & Technology

Article

(30) Adriaens, P.; Grbic-Galic, D. Reductive dechlorination of PCDD/F by anaerobic cultures and sediments. Chemosphere 1994, 29, 2253−2259. (31) Zhao, X.; Rockne, K. J.; Drummond, J. L.; Hurley, R. K.; Shade, C. W.; Hudson, R. J. M. Characterization of methyl mercury in dental wastewater and correlation with sulfate-reducing bacterial DNA. Environ. Sci. Technol. 2008, 42 (8), 2780−2786.

6620

dx.doi.org/10.1021/es300560q | Environ. Sci. Technol. 2012, 46, 6612−6620