Environ. Sci. Technol. 2003, 37, 5502-5510
Destruction of the World Trade Center and PCBs, PBDEs, PCDD/Fs, PBDD/Fs, and Chlorinated Biphenylenes in Water, Sediment, and Sewage Sludge SIMON LITTEN* New York State Department of Environmental Conservation, 625 Broadway, Fourth Floor, Albany, New York 12233 DENNIS J. MCCHESNEY USEPA, Region 2, Edison, New Jersey M. C. HAMILTON, AND BRIAN FOWLER Axys Analytical Services, Sidney, British Columbia, Canada
Ash-laden runoff samples collected near Ground Zero soon after the September 11, 2001 attack on the World Trade Center (WTC) and subsequent fire demonstrate the release of polychlorinated biphenyls (PCBs), polybrominated dipheyl ethers (PBDEs), polybrominated dibenzo-pdioxins and polybrominated dibenzofurans (PBDD/Fs), polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs), and tetra- and pentachlorinated biphenylenes (PCBPs) from the incident. Relative abundances of PCDD/F congeners in the runoff water and postdisaster lower Manhattan dust samples were different from those seen in pre-disaster NYC combined sewer outfall (CSO) samples. The WTC-related samples showed a greater relative abundance of 2,3,4,7,8-PeCDF than usually seen in CSOs, sludges, and treated wastewaters. This congener may be associated with certain types of incineration. Comparison of sediment and water samples collected in the lower Hudson River before and shortly after September 11, 2001 (9/11) showed no changes in PCB or PCDD/F concentrations or homologue profiles determined down to the parts per quadrillion range. Comparisons of ambient water samples collected post-9/11 with archived samples suggest that the WTC disaster did not significantly impact ambient concentrations of the target chemicals. Ambient concentrations of PBDD/Fs in New York Harbor are similar to those of PCDD/Fs, suggesting that these contaminants deserve increased scrutiny with respect to toxicity, sources, and fate in the environment.
Introduction When the World Trade Center (WTC) was destroyed on September 11, 2001(9/11), the New York State Department of Environmental Conservation (NYSDEC) and the United States Environmental Protection Agency (EPA) immediately marshaled resources to investigate the impacts of the calamity. NYSDEC Division of Water asked if the disaster * Corresponding author phone: (518) 402-8192; fax: (518) 4029029; e-mail:
[email protected]. 5502
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
affected ambient water quality and whether wastewater from the site impacted sludges. Secondarily, we were interested in the release and generation of a few halogenated chemicals: polychlorinated and brominated dibenzo-p-dioxins and furans (PCDD/F and PBDD/Fs), polychlorinated biphenylenes (PCBPs), and polychlorinated diphenyl ethers (PBDEs). Archived samples collected before the disaster were analyzed for some of the exotic substances as information emerged on their abundances in the WTC debris. Since 1998 the NYSDEC, as well as many others, had been conducting an extensive and intensive study of toxic chemicals in New York Harbor as an element of the NY/NJ Harbor National Estuary Contaminant Assessment and Reduction Program (CARP). The monitoring utilized a field technique developed by NYSDEC, the trace organics platform sampler (TOPS), that concentrates organic analytes from large volumes of water. Use of TOPS in conjunction with sensitive laboratory methods (isotopic dilution high-resolution GC/ high-resolution MS), achieved nearly consistent quantitation of toxic hydrophobic organic chemicals such as PCDD/F, PCBs, and chlorinated pesticides from surface water sites throughout the system. As a direct consequence of this work, we had, at the time of the disaster, a great deal of background data against which new impacts could be evaluated. Before 9/11, the NYSDEC water program had collected and analyzed 908 samples of sediment and water for PCBs and 693 samples for PCDD/Fs (using EPA Methods 1168 A and 1613, respectively) from the New York Harbor area. On September 12, 2001 two of us (S.L. and D. McC.) discussed a strategy for measuring the impacts of the disaster on water by sampling combined sewer outfalls (CSOs) near the WTC. We considered that the release and combustion or pyrolization of capacitor and transformer fluids, office furnishings, and computers would result in the environmental distribution of PCBs (1, 2), PCDD/Fs (2, 3), PCBPs (2, 4), PBDEs (5), and PBDD/Fs (6-9). There are well-established analytical methods, and a history of monitoring, and water quality standards for PCBs and PCDD/Fs. The situation with regard to the brominated materials is different. Analytical methods are relatively immature as field and laboratory blank issues have yet to be fully resolved, and the basis for concern, toxicology, is less well developed. The situation with PCBP is even worse than for the brominated substances in that no commercially prepared laboratory standards exist, and, despite apparent similarities to PXDD/Fs, there appear to be no toxicological evaluations.
Experimental Section Sampling. The first post-9/11 rain event (4.8 cm fell in Central Park) occurred on September 14, 2001. EPA’s field team found no evidence of CSO activity but they did find street-cleaning washdown runoff near Ground Zero. The catastrophic collapse of the WTC generated thousands of tons of finely pulverized concrete, glass, gypsum, and combustion ash that blanketed the immediate vicinity. As part of the cleanup work, NYC Department of Sanitation workers used Hudson River water (supplied by a New York Fire Department fireboat) to flush the dust and ash off the streets. EPA’s field team sampled this runoff at the foot of Rector Street (at the southern side of the WTC complex) on September 14, 2001 by holding precleaned amber bottles under street drains. At that time there was in intermittent flow of between 1000 and 2000 L/min of a slurry of dust and water running into the Hudson River. Sample collection sites are shown on Figure 1. 10.1021/es034480g CCC: $25.00
2003 American Chemical Society Published on Web 11/15/2003
FIGURE 1. Sampling sites and their geographic relation to the World Trade Center. Concentrations of solids varied widely as sanitation workers hosed down the street. Most of the washdown was captured in a storm drain. The suspended solids concentration in the dirtiest sample was 9700 mg/L. A second storm occurred on September 20 (2.1 cm at Central Park). Runoff water was again collected by holding bottles below street drains. Washdown had ceased and the amount of water running off of Rector St. was a trickle. Total amounts of runoff water that entered the Hudson are unknown but are suspected to have been very low due to the absence of commercial activity and the evacuation of many residents reducing the amount of wastewater being generated. Rain events would therefore have been less likely to generate CSOs. EPA’s surveillance team did not observe any CSO activity in lower Manhattan. To further evaluate the impact of the disaster on Harbor water quality, Hudson River/East River surface water monitoring was conducted by EPA using 4-Lgrab samples on September 20 from three locations off the WTC site (Hudson North, Hudson West, and Hudson South), a location near the George Washington Bridge, and a site in the East River off South St. These samples were taken with pre-cleaned Kemmerer bottles. EPA collected dust and ash from streets and sidewalks at 11 sites around the WTC between September 11 and October 8. Dust samples were obtained by scooping the material with a clean scoop into pre-cleaned certified 8-oz amber glass jars in accordance with SW846 protocol for solids sampling. These data have been reported in draft form (10). New York City Department of Environmental Protection (NYCDEP) collected sludges (not dewatered) from the wastewater treatment plant receiving sewage from southern Manhattan (Newtown Creek WPCF) on September 19 and 25, 2001. These samples were compared with composites of dewatered sludges collected during each day of February 2001. On September 16 NYSDEC collected sediment samples from Piers 6 and 25 prior to their dredging to accommodate barges for debris removal. These samples, from approximately 13 cm into the sediment, were taken to determine placement of dredge spoils and are thought to be more representative of long-term deposition than the immediate impact of the disaster. Ponar-collected sediment (2 L) was poured into a tray where 10 evenly spaced subsamples were composited. An archived sediment sample taken in 1998 from a heavily industrialized channel in the western Harbor, the Arthur Kill (off the northern end of Pralls Island), was also analyzed for the expanded range of analytes.
On October 5 NYSDEC and EPA resampled a Harbor segment by boat, including the area adjacent to Ground Zero, using NYSDEC shop-made equipment (a trace organics platform sampler, TOPS) which preconcentrates hydrophobic substances from very large volumes of water to establish comparability with pre-disaster ambient water quality data from the same area. TOPS samples were analyzed using laboratory procedures identical to those that had been employed during six pre-disaster NY/NJ Harbor CARP monitoring events. We designed TOPS to be totally enclosed to reduce atmospheric contamination. Water is pumped from an intake through a glass-fiber cartridge filter (Parker Hannifin, nominal porosity of 1 µm) to capture suspended particles. A tow fish suspended off the vessel’s side held the TOPS intake into the current and away from the wake and other influences generated by the sampling vessel. A magnetic impeller pump pulled water from the intake on the tow fish and passed it through a 100-µm mesh Nytex plankton net to remove large zooplankton. A sensor measured the filter back-pressure and triggered TOPS to shut down at 15 psi. Some of the filtered water was wasted and some was pumped through two Teflon columns arranged in series (each containing 65 g of the synthetic resin XAD-2). Flow meters recorded the total volumes wasted between the filter and XAD and the amount leaving the XAD. New filters were cleaned by baking at 450 °C for 4 h. Baked filters were immediately wrapped in fired aluminum foil and stored at -18 °C prior to use. The XAD resin was size sorted and cleaned by a series of solvent washings (methanol extraction for 48 h, dichloromethane rinsing for 96 h, and drying). Field filtration was typically accomplished at about 3 L/min. Filtrate was pumped through the XAD at 600 mL/min (3 bed volumes/min). All XAD discussed here was purchased pre-cleaned and packed in the Teflon columns by Axys Analytical Services of Sidney, BC. The amount of water filtered is the sum of the outputs of both flow meters, and the amount that was exposed to XAD is the output of the second flow meter. As a precaution against flow meter failure, rates of pumping were measured at least three times per sample for each flow meter by timing the interval required to fill a 20-L plastic carboy. At the end of the sampling cruise, the carboys were weighed on a calibrated electronic scale and reweighed after emptying. After sampling, the XAD columns were immediately sealed and double bagged. They were refrigerated prior to being shipped out for analysis. The glass fiber cartridge was removed VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5503
TABLE 1. Numbers of Samples Analyzed for PCBPs, PBDD/Fs, and PBDEs type
sampling location
date
ambient ambient ambient ambient ambient ambient ambient ambient ambient ambient ambient ambient ambient ambient ambient blank sediment sediment sediment sludge sludge ambient ambient ambient ambient ambient WTC WTC
Hackensack R., mid-tidal Hackensack R., mouth Hudson R. at Poughkeepsie Hudson R. at Poughkeepsie Hudson R. between Tappen Zee and Harlem R. Hudson R. south of Harlem R. Hudson R. south of Harlem R. Hudson R. south of Harlem R. Lower East R., Brooklyn Bridge To Hell Gate Newark Bay Passaic R., mouth, bottom Passaic River, mid-tidal Passaic River, mid-tidal Upper Bay Upper Bay EPA water blank Arthur Kill at end of Pralls Island Pier 25 Pier 6 Newtown Creek WPCF Digested Sludge Newtown Creek WPCF Digested Sludge East River, South St. George Washington Bridge Hudson River South Hudson River West Hudson River North Rector St. runoff Rector St. runoff
9/2/99 7/7/99 3/1/99 4/17/99 7/10/99 10/5/01 3/16/99 8/12/99 7/27/99 8/11/99 7/21/99 10/18/00 8/25/99 3/18/99 8/11/99 9/20/01 9/98 9/18/01 9/18/01 9/19/01 9/25/01 9/20/01 9/20/01 9/20/01 9/20/01 9/20/01 9/14/01 9/20/01
from its housing with acetone-wiped metal tongs and laid on pre-cleaned aluminum foil. Particles may slough off the filter and collect in the bottom of the filter housing. To retain these, a pressure filter apparatus was set up whereby high purity nitrogen gas was used to push the water remaining in the housing through a GF/F glass fiber filter that had been pre-fired for 4 h at 450° C and kept at -18° C prior to use. As the water passed through the pressure filter, it was caught in the emptied filter housing, swirled, and re-filtered twice, or until all visible particles had been captured. The GF/F flat filter was combined with the cartridge filter in an aluminum foil wrapping, double bagged, and frozen. Particle trapping efficiency was assessed by measuring particulate organic carbon (POC) at the inlet and outlet to the filter. XAD capture efficiency was assessed two ways: by injecting C13-labeled analogues into the XAD and by separately analyzing two XAD columns arranged in series. At the conclusion of sampling, the TOPS was cleaned by recirculating hot detergent water for 10 minutes, pumping out the soapy water, and rinsing with continuously supplied hot tap water. Field blanks were taken at this point by pumping 4 L of reagent grade water through a filter and XAD. On every sampling cruise, external polyethylene tubing and the silicone tubing in the peristaltic pump were replaced. Tubing in the main body of the TOPS was either stainless steel or Teflon and was not replaced. The average volume of water filtered in lower Hudson River sites for PCBs and PCDD/ Fs was 808 L. Table 1 lists the samples analyzed for PCBPs, PBDD/Fs, and PBDEs. Analytical. Axys Analytical Services performed all the analyses on grab samples, sediments, XAD, and glass fiber cartridges. TOPS glass fiber filters were analyzed separately from TOPS XAD columns. Sediments and glass fiber cartridges were air-dried and extracted with toluene. XAD columns were extracted with 80:20 toluene/acetone. Whole water samples were filtered at Axys and then extracted separately, but the extracts were combined for analysis. Sediment and dried sludge samples were Soxhlet extracted at Axys. Dust samples 5504
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
PCBPs
1
1 1 1 1 1 1 1 1 1 1 1 1 1
PBDD/F 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
PBDE
1
1 1 1 1 1 1 1 1 1 1 1 1 1 1
were analyzed by USEPA contractors using EPA SW-846 method 8080. PCBs and PBDEs were analyzed by US EPA method 1668A. Samples were spiked with labeled surrogate and cleanup standards, cleaned up on gel permeation and Florisil (2% deactivated) columns, and treated for sulfur with activated copper. Further cleanup was performed on alumina (1% deactivated) and acid/base silica columns; the extracts were reduced in volume and spiked with an aliquot of the labeled recovery (internal) standard. Analysis for PCBs was conducted by high-resolution gas chromatography/high-resolution mass spectrometry (HRGC/HRMS) on a VG magnetic sector highresolution MS equipped with an HP 5890 gas chromatograph, a CTC autosampler, and a VAX data system running VG software. An SPB-Octyl (30 m, 0.25-mm, i.d., 0.25-µm film thickness) chromatography column was coupled directly to the MS source. The MS was operated at 10 000 (static) mass resolution in the electron impact (EI) mode using multiple ion detection, acquiring at least two ions for each target and surrogate compound. Target concentrations were determined by isotope dilution using VG OPUSQUAN software. Samplespecific detection limits (SDLs) were determined from the analysis by converting the minimum detectable signal to a concentration following the same procedures used to convert target peak responses to concentrations. In the case of PCBs, 159 chromatographic domains encompassing all 209 congeners and 18 labeled surrogates were spiked into samples in the laboratory. There were 41 PBDE chromatographic domains encompassing 44 congeners, and four PBDE surrogates were added in the laboratory. PBDEs were analyzed on an HP 6890 HRGC using a DB-5HT (30 m, 0.25-mm i.d., 0.1-µm film thickness) and an Autospec Ultima high-resolution MS running Micromasss OPUSQUAN software. PCDD/Fs, PBDD/Fs, and PCBPs were analyzed by EPA Method 1613B. This is an isotopic dilution HRGC/HRMS method carried out on a VG Ultima high-resolution mass spectrometer equipped with a HP 5890 gas chromatograph with a CTC autosampler and a VAX 4000 data system. A split/ splitless injection sequence was used. Chromatographic
TABLE 2. Comparison of PCB Concentrations (ng/L) from WTC Runoff with Ambient and CSO Samples (Sample Detection Levels (SDLs) and Method Blanks (MB) Are Also Shown for the Rector St. Samples) Rector St.
total mono di tri tetra penta hexa hepta octa nona deca a
Hudson R. ambient
CSO
9/14
9/14a
9/20
avg SDL
avg MB
avg. of 8
std. dev.
avg. of 16
std. dev.
23,800 394 1,100 1,520 1,680 3,110 6,700 7,310 1,720 277 9.61
8,410 195 528 615 619 1,230 2,540 2,050 554 70.3 4.46
96.4 2.13 6.78 8.86 9.04 15.7 27.8 20.3 5.06 0.71 0.06
0.0405 0.372 0.492 3.08 21.4 9.86 4.44 2.54 0.505 0.0886
0.047 0.104 0.0911 0.0577 0.0401 0.0403 0.0202 0.0063
