Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Source Apportionment of Atmospheric Polychlorinated Biphenyls in New Jersey 1997- 2011 Pornsawai Praipipat, Qing Yu Meng, Robert J. Miskewitz, and Lisa A. Rodenburg Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b04572 • Publication Date (Web): 04 Jan 2017 Downloaded from http://pubs.acs.org on January 4, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 22
Environmental Science & Technology
1 1
Source Apportionment of Atmospheric Polychlorinated Biphenyls in New
2
Jersey 1997- 2011
3
Pornsawai Praipipat1,2, Qingyu Meng3, Robert J. Miskewitz1, and Lisa A. Rodenburg1* 1
4
Department of Environmental Sciences, Rutgers University, 14 College Farm
5 6
Road, New Brunswick, NJ 08901, USA 2
Current affiliation: Department of Environmental Science, Faculty of Science, Khon Kaen
7 8
University, Khon Kaen 40002, Thailand 3
School of Public Health, Rutgers University, Piscataway, New Jersey 08854, USA
9
* Corresponding author, formerly Lisa A. Totten. Email:
[email protected], Phone:
10
848-932-5774 Fax: 732-932-8644
11 12
Abstract
13
Concentrations of polychlorinated biphenyls (PCBs) in the Delaware River currently
14
exceed the Water Quality Criteria of 16 pg/L for the sum of PCBs due in part to atmospheric
15
deposition. The purpose of this work was to use a source apportionment tool called Positive
16
Matrix Factorization (PMF) to identify the sources of PCBs to the atmosphere in this area and
17
determine whether their concentrations are declining over time. The data set was compiled by
18
the Delaware Atmospheric Deposition Network (DADN) from samples taken in Camden, NJ
19
from 1999 to 2011 and New Brunswick, NJ from 1997 to 2011. The PMF analysis revealed four
20
resolved factors at each site. The factors that dominate the PCB burden in the atmosphere at
21
both Camden and New Brunswick resemble Aroclor 1242. These factors declined in
22
concentration during some portions of the monitoring period, but this decline slowed or stopped
ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 22
2 23
during 2003-2011. None of the factors displayed consistent declines in concentration throughout
24
the monitoring periods, and some factors actually increased in concentration during some
25
periods. This suggests natural attenuation alone will not control atmospheric PCB
26
concentrations, and additional efforts are needed to control PCB atmospheric emissions as well
27
as the numerous other sources of PCBs to the estuary.
28 29
Introduction
30
Polychlorinated biphenyls (PCBs) are toxic, persistent, and bioaccumulative chemicals
31
that are targeted for phase out under the Stockholm Convention on Persistent Organic
32
Pollutants.1 PCB levels in most US environmental compartments have declined dramatically
33
since PCBs were banned in the 1970s.2-5 However, PCB concentrations in the Delaware River
34
currently exceed the Water Quality Criteria of 16 pg/L for the sum of PCBs due in part to
35
atmospheric deposition.6 The sources of atmospheric PCBs are not clear. They may include old
36
PCB-containing equipment, joint sealants, caulks, waste incineration, storage and disposal
37
facilities, superfund sites and accidental releases.7-17 Because the sources are not known, it is not
38
clear what measures would be effective in reducing atmospheric PCB concentrations. Several
39
recent studies have examined the long-term temporal trends in atmospheric PCB
40
concentrations.4, 18-26 However, most of these studies measured a short list of PCB congeners24, 25,
41
27
42
sources of atmospheric PCBs,28-31 it is important to investigate temporal trends in urban
43
atmospheric PCB concentrations. Only one other study 13 has examined temporal trends for a
44
comprehensive list of PCB congeners and conducted source apportionment of the congeners to
45
examine temporal trends in atmospheric source categories. This previous study used data from
and/or examined trends in rural or remote areas. Since urban areas appear to be the primary
ACS Paragon Plus Environment
Page 3 of 22
Environmental Science & Technology
3 46
the Integrated Atmospheric Deposition Network (IADN) from Chicago, IL, USA. 13 The present
47
work examines temporal trends for a long list of PCB congeners and their source categories for
48
urban (Camden) and suburban (New Brunswick) sites in New Jersey using data from the
49
Delaware Atmospheric Deposition Network (DADN). The DADN gathered data on atmospheric
50
PCB concentrations in support of the Total Maximum Daily Loads (TMDLs) for PCBs
51
developed by the Delaware River Basin Commission (DRBC) for the entire tidal portion of the
52
Delaware River and Delaware Bay. The DADN included as many as eleven sites over various
53
years. In the present work, data from the New Brunswick and Camden sites were chosen for
54
examination via Positive Matrix Factorization (PMF) because they have the longest data history.
55
PMF is an advanced source apportionment tool developed by Paatero and Tapper32 that
56
has been used to identify PCBs sources in water, sediment, and air.10-12, 33-36 The present work
57
builds on previous studies in which PMF was used to apportion PCB sources to the water
58
column35, point source dischargers37, and sediment36 of the Delaware River Basin. The present
59
work also builds on our previous effort to apportion PCB sources in the air of Chicago.13
60
Therefore, the goals of this work were to apportion the major sources of PCBs to the air in New
61
Jersey and the Delaware River Basin and to determine whether atmospheric PCB concentrations
62
are declining over time in order to predict when the concentrations of PCBs in the Delaware
63
River might drop below the water quality standard, thus achieving the TMDL. Another goal was
64
to compare the PCB sources in Camden (Philadelphia) to another major US city, Chicago.11, 13
65
Methodology
66
Study Sites
67 68
Camden, NJ is located across the river from Philadelphia, PA which is currently the fifth largest city in the US, and was 4th largest in 1970 during the years of peak PCB usage.38 In a
ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 22
4 69
previous work, PCB sources in the Camden DADN samples from 1999 to 2002 were
70
investigated using PMF.11 The present study updates the previous one and includes data gathered
71
from 1999 to 2011, allowing an evaluation of temporal trends. The DADN network did not
72
operate in Camden from August 2002 to March 2004 due to lack of funding. New Brunswick is a
73
suburban city in New Jersey with a population of about 55,000 in the 2010 census.39 New
74
Brunswick and nearby Piscataway host the main campus of Rutgers, the State University of New
75
Jersey, which is host to about 40,000 students.
76
PCB measurements
77
The methods used to measure PCBs in the DADN samples have been described
78
previously and are summarized in the Supporting Information section 1.40, 41 In brief, 24-hour
79
air samples were collected using high-volume air samplers with quartz fiber filters (QFF) to
80
capture the particle phase and polyurethane foam (PUF) to the collect PCBs in the gas phase.
81
PUFs and QFFs were extracted in a Soxhlet apparatus, cleaned up on florisil, and analyzed for 60
82
PCB peaks representing 93 congeners via electron capture detection on an Agilent 6890 gas
83
chromatograph. Methods remained the same throughout the data collection period.
84
PMF analysis
85
For this study, Positive Matrix Factorization (PMF) version 2.0 software (Yp-Tekniika
86
Ky Co., Helsinki, Finland) was used to examine the gas phase PCB concentrations only (particle
87
phase concentrations were ignored). In contrast, our previous PMF analysis of PCBs in Camden
88
utilized the gas plus particle phase concentrations, however, that analysis isolated a factor that
89
represented particle phase PCBs.11 The gas phase typically contains about ten times higher PCB
90
concentrations than the particle phase, so excluding the particle phase excludes only a small
91
fraction of the total atmospheric PCB mass.28 Data from the two DADN sites were analyzed in
ACS Paragon Plus Environment
Page 5 of 22
Environmental Science & Technology
5 92
separate PMF runs. (A pooled data set from both sites was analyzed, but the results produced
93
only three factors and are not shown here.) The relative standard deviation of the recoveries of
94
the applicable surrogate was used as the uncertainty for the detected concentrations of individual
95
congeners. Non-detects were substituted with a random number between zero and the detection
96
limit, and assigned three times the uncertainty of the detected concentrations.
97
To prepare the data for PMF analysis, first congeners and samples in which a large
98
number of measurements were below detection limit (i.e. more than 130 non-detects at New
99
Brunswick and more than 145 non-detects at Camden) were removed. Detection limits for
100
individual PCB congeners (or co-eluting groups) ranged from 0.013 to 1468 pg m-3 at Camden
101
and from 0.002 to 612 pg m-3 at New Brunswick for PUFs. This yielded 54 PCB congeners in
102
234 samples at Camden and 48 PCB congeners in 307 samples at New Brunswick. After
103
preliminary PMF runs, several samples were removed because they were outliers in the
104
preliminary PMF solutions due to more than 20 congeners below detection. Thus the final data
105
matrices consisted of 49 congeners in 226 samples at Camden, and 46 PCB congeners in 307
106
samples at New Brunswick. These congeners represented the tetra through nona homologues.
107
The number of factors was determined to be four each for Camden and New Brunswick (see
108
Supporting Information section 2 for justification). For clarity, these are hereafter referred to as
109
factors 1C, 2C, 3C, and 4C (from Camden) and factors 1NB, 2NB, 3NB, and 4NB (from New
110
Brunswick).
111
Results and Discussion
112
At Camden, average (± standard deviation) gas-phase 93PCB concentrations (Figure S-
113
1) fell from 4600 ± 3300 pg/m3 in 1999 to 1600 ± 1000 pg/m3 in 2011. New Brunswick, average
114
gas-phase 93PCB concentrations (Figure S-1) similarly declined from 330 ± 220 pg/m3 in 1997
ACS Paragon Plus Environment
Environmental Science & Technology
Page 6 of 22
6 115
to 110 ± 60 pg/m3 in 2011. Camden is therefore similar to other urban areas of the US, such as
116
Chicago 23, in having relatively high atmospheric PCB concentrations. In comparison to sites
117
from the Integrated Atmospheric Deposition Network which surrounds the North American
118
Great Lakes, concentrations at New Brunswick are more similar to sites that have some urban
119
influence, such as Sleeping Bear and Sturgeon Point, but are higher than truly remote sites such
120
as Brule River or Eagle Harbor.23
121
Before doing the PMF analysis, it is important to investigate the raw data for trends.
122
Gas-phase PCB concentrations almost always display a temperature dependence that roughly
123
follows a Clausius-Clapeyron relationship, because more energy is available to drive their
124
volatilization at higher temperatures.4, 13, 23, 26, 42, 43 In particular, the measured gas-phase
125
concentrations of each congener were investigated as a function of both temperature and time via
126
the widely-used approach4, 13, 23, 26, 42, 43 of fitting the natural log of concentration (ln Cgas) versus
127
time (t in days) and inverse temperature (1/T in Kelvin) via a least-squares regression:
128
1 ln C gas a0 a1 a2t T
(Eqn. 1)
129
where a0, a1 and a2 represent the coefficients for the intercept, temperature, and time parameters,
130
respectively. As is typical of gas-phase PCB concentrations, all PCB congeners displayed a
131
significant correlation (p < 0.05) with temperature, with higher PCB concentrations at higher
132
temperatures (Supporting Information Figure S-2 and S-3). As observed in other studies, lower
133
molecular weight congeners displayed weaker temperature dependence43, in keeping with their
134
lower enthalpies of vaporization.44 Figure S-4 demonstrates that the temperature dependence of
135
most PCB congener concentrations was greater and a stronger function of vapor pressure at the
136
urban site, Camden, than at suburban New Brunswick. A stronger temperature dependence of
137
gas-phase PCB concentrations in urban versus rural areas has been observed in other studies.43, 45,
ACS Paragon Plus Environment
Page 7 of 22
Environmental Science & Technology
7 138
46
Simcik et al. 45 have suggested that this occurs because urban monitoring sites are nearer to
139
PCB sources. In contrast, the effect of temperature on PCB concentrations is muted at rural
140
locations because the PCBs must travel farther to reach them.
141
Most congeners displayed significant negative values of a2 over the full monitoring
142
period, implying that their concentrations are decreasing over time. At Camden, PCBs 40,
143
85+136, 153+132, 137+176+130, 163+138, 158, 187+182, 183, 185, 202+171+156, 199,
144
170+190, 201, 203+196, and 194 displayed no significant correlation with time (i.e. neither
145
increasing nor decreasing). PCBs 195+208 and 206 were increasing over time (i.e. a2 was
146
positive and significant p