PDF Archive

Easily share your PDF documents with your contacts, on the Web and Social Networks.

Share a file Manage my documents Convert Recover PDF Search Help Contact

Environ Toxicol Chem 21, 2002, 2292 2300 .pdf

Original filename: Environ Toxicol Chem 21, 2002, 2292-2300.pdf
Title: entc_21_1129.2292_2300.tp
Author: penta@ultra1

This PDF 1.2 document has been generated by Penta Software Inc. Version 12.0 / Acrobat Distiller Daemon 3.01 for Solaris 2.3 and later (SPARC), and has been sent on pdf-archive.com on 01/11/2015 at 17:59, from IP address 71.17.x.x. The current document download page has been viewed 610 times.
File size: 136 KB (9 pages).
Privacy: public file

Download original PDF file

Document preview

Environmental Toxicology and Chemistry, Vol. 21, No. 11, pp. 2292–2300, 2002
q 2002 SETAC
Printed in the USA
0730-7268/02 $9.00 1 .00

†Department of Chemistry, Box 3065, University of Victoria, Victoria, British Columbia V8W 3V6, Canada
‡Contaminants Science Section, Institute of Ocean Sciences, Fisheries and Oceans Canada, 9860 West Saanich Road,
Sidney, British Columbia V8L 4B2, Canada
( Received 7 November 2001; Accepted 3 May 2002)
Abstract—Semipermeable membrane devices (SPMDs) were placed in the Fraser River near Vancouver, British Columbia, Canada,
between August 6 and September 30, 1996. This location is near a large urban and industrial region (population 2,000,000) and
is expected to be representative of other large, modern cities. After exposure to the ambient water column, SPMD samples were
analyzed for a suite of 36 polybrominated diphenyl ether (PBDE) congeners plus all homologue groups from mono- through hexabrominated. Observed congener patterns differed significantly from that of the commercial penta- and octa-BDE mixtures. A
reconstruction approach was developed based on an aquatic transport model and utilizing published octanol–water partition coefficients, calculated SPMD uptake rates, and predicted water concentrations by using the EcoFate multimedia mass balance aquatic
simulation model for the 13 major PBDE congeners. In combination, composite technical mixtures were created by combining
commercial penta-BDE mixtures (Bromkal 70-5DE and Great Lakes Chemicals DE-71) with commercial octa-BDE mixtures
(Bromkal 79-8DE and Great Lakes Chemicals DE-79) in their relative 2000 North American production volumes. The reconstructed
SPMD patterns more closely approximated the composite technical mixtures and suggest that PBDEs in such an industrial region
arise primarily from penta- and octa-BDE source mixtures.
Keywords—Semipermeable membrane devices
Source apportionment

Flame retardants

Polybrominated diphenyl ethers

Congener patterns

BDE209 dominates sediment patterns in some regions [12].
The major congener present in biota, sediments, and the atmosphere is BDE47 (2,29,4,49-BDE), followed in abundance
by BDE99 and BDE100 (2,29,4,49,5-BDE and 2,29,4,49,6BDE, respectively). These three congeners typically make up
.70 to 80% of SPBDE reported in the literature [6,7].
As with polychlorinated biphenyls (PCBs) and polychlorinated dibenzo-p-dioxins and furans (PCDD/Fs), PBDEs are
not produced as pure compounds. Rather, the technical mixtures have a congener pattern uniquely dependent on production conditions and subsequent environmental weathering.
These congener patterns allow for fingerprinting the source of
such compounds, broadly referred to as source apportionment.
With 209 different congeners possible for mono- through decaBDE, recent advances in multiresidue sample processing and
gas chromatography–high resolution mass spectrometry techniques were necessary to identify and quantitate a sufficient
number of congeners for a reliable PBDE source apportionment study. Sampling programs for such hydrophobic contaminants in a source apportionment study are also a critical consideration. The sampling properties of each compound used
in fingerprinting must be either known or reliably approximated. In other words, the sampling protocols and devices
must be well understood both from a physicochemical perspective as well as with regards to potential sources of contamination. Additionally, rigorous quality assurance–quality
control procedures are required during analysis to prevent unreliable data. If these factors are not properly taken into account, the fingerprint observed will either be unintelligible or
point to the incorrect source.


Recent evidence suggests that polybrominated diphenyl
ethers (PBDEs) have become significant environmental contaminants [1–7]. These compounds are additive flame retardants used in a wide range of materials, including high-impact
plastics, textiles, and foams, at concentrations up to 30% by
weight [8]. Three main types of commercial mixtures are in
widespread use: penta-BDE, octa-BDE, and deca-BDE. PentaBDE contains mainly penta- and hexa-brominated congeners,
octa-BDE is composed of mostly hepta- and octa-brominated
congeners, and deca-BDE is primarily the fully brominated
BDE209 (.90%) with traces of nona-brominated impurities.
These technical mixtures make up 14, 6, and 80%, respectively,
of the estimated 1999 worldwide PBDE production of 68,000
tonnes [9]. In North America, the market demand for PBDEs
in 2000 was estimated at 34,000 tonnes. As with worldwide
demand, deca-BDE was the major product in use in North
America at 24,300 tonnes, albeit at a lower proportion of total
PBDE usage (72%) [10]. Additionally, North America uses a
higher proportion of the penta-BDE (8,290 tonnes or 24%)
and a lower proportion of octa-BDE (1,375 tonnes or 4%) than
worldwide values [10].
However, the degradation pathways of deca-BDE are poorly
understood [6,7,11], and microbial, abiotic, or photochemical
debromination of deca-BDE may lead to production of lower
brominated congeners. Of especial interest are the tetrathrough hexa-BDE congeners that make up the majority of
total PBDEs (SPBDE) found in biota; the fully brominated
* To whom correspondence may be addressed

Reconstructing source PBDE congener patterns

Semipermeable membrane devices (SPMDs) are well suited
to sampling hydrophobic environmental contaminants present
not only in the water column [1,12–21], but also in sediments
[16,19,20,22], air [23–27], and soils [28]. Much of the theoretical understanding of SPMDs has been elucidated in pioneering work by Huckins, Gale, and coworkers at the U.S.
Geological Survey (USGS) [29–32]; hence, we hereafter refer
to their understanding of SPMD sampling as the USGS model.
As a result of their well-documented uptake characteristics,
SPMDs can be used to estimate time-integrated averages of
contaminant concentrations [16,29,31,33]. Because SPMD uptake of chemicals takes place only through transient membrane
pores, absorption is restricted to truly dissolved compounds
and not particulate- or colloid-associated compounds
[29,30,33]. These theoretical insights allow the use of SPMDs
to reconstruct water column congener patterns for compounds
such as PBDEs, which in combination with aquatic transport
models, would permit an attempt to apportion PBDE sources
in a region.
In this study, we deployed seven SPMDs in the Fraser River
near Vancouver, British Columbia, Canada (population
2,000,000) during the summer of 1996. The Fraser River is
the most important watercourse from a fisheries perspective
on the west coast of North America. It drains 230,400 km2 of
British Columbia and supports large populations of anadromous salmonid species [34]. Previous work by our group has
shown high levels of PCBs and PCDD/Fs in fish, sediments,
and SPMDs in the Fraser River due to the large concentration
of urban and industrial activities along its length [22,34,35].
The Fraser River is also the major freshwater input to the Strait
of Georgia, a poorly flushed inland sea between Vancouver
Island and the British Columbia mainland. We have previously
shown that high levels of not only PBDEs, but also PCBs,
PCDD/Fs, and organochlorine pesticides, are present in aquatic
biota from this region [1], of which contaminant inputs from
the Fraser River are a major contributor.
To account for the PBDE congener patterns observed in the
environment, three schools of thought exist. The PBDE patterns could originate from usage of the more mobile pentaand octa-BDE mixtures (as compared to deca-BDE), in some
proportion to their relative annual production and usage values.
Conversely, because deca-BDE dominates the annual PBDE
production quantities (;80%), environmental debromination,
transport, and bioaccumulation of deca-BDE degradation products may account for observed congener patterns. An alternate
theory postulates that the large quantities of commercial tetraBDE products used in the 1970s and early 1980s still reside
in sediments and biota. A combination of all processes is most
likely; however, one process may be clearly dominant and it
is of interest to investigate this possibility.
To observe PBDE congener patterns in the water column,
we deployed SPMDs in an urban and industrial Canadian river.
This article examines the use of an aquatic transport model,
EcoFate, and the USGS model of SPMD sampling in an attempt to reconstruct source PBDE congener patterns.

SPMD construction and deployment
Semipermeable membrane devices were prepared by using
the protocols discussed in detail previously [22]. The SPMDs
were placed in perforated 20-L plastic food buckets, three
SPMDs to one bucket, before immersion in the water column.
The food buckets were anchored with heavy chain and the

Environ. Toxicol. Chem. 21, 2002


containers were attached to log booms or pilings with rope.
Seven SPMDs were deployed in the Fraser River near Vancouver from August 6 to September 30, 1996, at a low-tide
depth of 2 to 3 m for a total exposure time of 55 days (Fig.
1). Care was taken to avoid attachment to creosoted timbers,
which are abundant in the lower Fraser River. One SPMD was
located at MacMillan Island, near Fort Langley (Langley; S1),
which is 28.5 km upstream from the city of New Westminster
and the industrial activities in the lower Fraser River. However,
this site is still subject to tidal influences and possible upstream
transport of contaminants.
The remaining six SPMDs were located in the lower Fraser
River west of the city of New Westminster. Three SPMDs were
located on the highly industrialized north arm of the Fraser
River. One of these SPMDs (North Arm; S2) was situated
approximately 1 km west of where the Fraser River bifurcates
into the north and south arms. Other north arm SPMDs were
deployed at the railway bridge to Mitchell Island (Mitchell;
S3), approximately half-way between separation of the north
and south arms and where the north arm discharges into the
Strait of Georgia, and near the office of the North Fraser Harbour Commission (Harbour; S4). Three further SPMDs were
deployed on the less industrialized south arm of the Fraser
River between New Westminster and where the south arm
discharges into the Strait of Georgia. One site was approximately 1 km downstream of the bifurcation into the north and
south arms in Annacis Channel (East Bridge, BC, Canada)
near a muddy beach (Annacis 1; S5). The other two south arm
SPMD sites were at Purfleet Point at the southwestern corner
of Annacis Island, 3 km below a major sewage treatment outfall (Annacis 2; S6), and further downstream on the south bank
of the Fraser River near Chatterton Chemicals (Chatterton; S7).
Concentrations of PBDEs in these SPMDs are available elsewhere [1] and only the congener patterns are presented in Table
1. In addition to the SPMDs deployed in the water column,
another set of SPMDs was exposed to ambient air during the
deployment at the sampling sites (;30 min) to serve as field
blanks and to reveal possible atmospheric contamination.

SPMD collection and processing
Procedures on the extraction and processing of SPMD samples are given in detail in our previous work [22]. Briefly, the
extracts were reduced to 1 ml followed by further cleanup on
alumina and carbon fiber columns. The eluants from the carbon
fiber column were concentrated to less than 10 ml and spiked
with a 13C-labeled method performance standard (1,000 pg
[13C]-3,39,4,49-tetrachlorodiphenyl ether) before analysis to
monitor recovery of the 13C-labeled PBDE internal standards
and the analytes during the sample workup procedure. The
final combined volume, including sample extract plus recovery
standard, was approximately 30 ml. Samples were processed
in batches of 6 to 10 that consisted of a procedural blank and
a field blank. The recoveries of the 13C-labeled PCDE surrogates ranged from 40 to 120%, within the allowable limits.
All congener concentrations were corrected for percent recovery of the internal standards.

Gas chromatography–high resolution mass spectrometry
Clean PBDE extracts were analyzed by gas chromatography–high resolution mass spectrometry with a VG-Autospec
high-resolution mass spectrometer (Micromass, Manchester,
UK) equipped with a Hewlett-Packard Model 5890 Series II


Environ. Toxicol. Chem. 21, 2002

S. Rayne and M.G. Ikonomou

Fig. 1. Semipermeable membrane device sampling locations and compartments used in the EcoFate aquatic model. Study sites are located
in Canada.

gas chromatograph (Avondale, PA, USA) and a CTC A200S
autosampler (CTC Analytics, Zurich, Switzerland). Four commercial PBDE mixtures were also analyzed and the masses of
individual congeners and their contribution to the sum of the
13 congeners under consideration here are presented in Table
2. These commercial mixtures are Great Lakes Chemical (West
Lafayette, IN, USA) commercial penta-BDE (DE-71) and octaBDE (DE-79) products, and Bromkal’s commercial penta-BDE
(70-5DE) and octa-BDE (79-8DE) mixtures (Cambridge Isotope Laboratories, Andover, MA, USA). Details on gas chromatography–high resolution mass spectrometry conditions and
the quality assurance–quality control protocols for multiple
congener PBDE analysis are given elsewhere [1,2].

Reconstruction procedures
Concentrations of PBDE congeners were determined on a pg/
g of triolein basis. These values were then normalized by dividing
by the level of total PBDEs (SPBDEs) to obtain a percent in
total SPBDEs value for each of 36 congeners analyzed for. A
listing of the 36 congeners is provided elsewhere [2]. Only the
following 13 congeners where $30% of the sample values were
above the method detection limit were used in the reconstruction
model: 4,49-BDE15; 29,3,4-BDE33/2,4,49-BDE28 (co-eluting
congeners); 2,4,49,6-BDE75; 2,29,4,59-BDE49; 2,29,4,49-BDE47;
2,39,4,49-BDE66; 2,29,4,49,6-BDE100; 2,39,4,49,6-BDE119;
2,29,4,49,5-BDE99; 2,29,4,49,6,69-BDE155; 2,29,4,49,5,69-BDE154;
and 2,29,4,49,5,59-BDE153.

Table 1. Congener patterns for the semipermeable membrane devices (SPMDs) deployed in the Fraser River, Canada. Values are percent in the
sum of the 13 congeners presented here. Values are masses of analyte per gram triolein in the SPMD [1]a



North arm



Annacis 1

Annacis 2









BDE 5 bromodiphenyl ether.

Environ. Toxicol. Chem. 21, 2002

Reconstructing source PBDE congener patterns
Table 2. Congener patterns for the four technical mixtures examined.
Masses reported are in gram analyte per 100 g of commercial mixture
(equivalent to percent by wt). Values in parentheses are percent in
the sum of the 13 congeners presented herea
Great Lakes Chemical
DE-71 (%)

DE-79 (%)

70-5DE (%) 79-8DE (%)

ND (0.0)
ND (0.0)
ND (0.0)
ND (0.0)
ND (0.0)
0.29 (0.3)
ND (0.0)
0.18 (0.2)
ND (0.0)
ND (0.0)
ND (0.0)
ND (0.0)
ND (0.0)
0.23 (0.3)
ND (0.0)
0.44 (0.5)
30.57 (33.6) ND (0.0)
28.01 (30.8) 0.09 (0.5)
ND (0.0)
0.20 (0.2)
ND (0.0)
0.40 (0.4)
ND (0.0)
7.60 (8.4)
ND (0.0)
8.02 (8.8)
1.30 (11.8)
ND (0.0)
0.05 (0.3)
ND (0.0)
46.45 (51.1) 2.40 (21.8)
43.71 (48.1) 0.13 (0.8)
ND (0.0)
ND (0.0)
ND (0.0)
0.19 (0.2)
4.08 (37.1)
2.10 (2.3)
2.37 (14.0)
3.97 (4.4)
3.23 (29.3)
6.03 (6.6) 14.35 (84.5) 3.50 (3.8)
90.95 (100) 16.99 (100) 90.94 (100) 11.01 (100)

DE 5 diphenyl ether; BDE 5 bromodiphenyl ether; ND 5 not

To reconstruct the SPMD patterns, the following calculations were performed. As shown in Figure 2, congener-specific
environmental partitioning of PBDEs from their source to
within the SPMDs can be divided into two models. The first
model is EcoFate, a time-dependent multimedia mass balance
simulation model of the environmental distribution and foodchain bioaccumulation of organic contaminants in aquatic systems [36], available at http://www.rem.sfu.ca/ecofate/cofate.
ehtml. EcoFate has an available multimedia model for the
234,000-km2 Fraser River basin from its headwaters near
Prince George (BC, Canada) to its discharge into the Strait of
Georgia at Vancouver. In EcoFate, the Fraser River from Prince
George to Vancouver is divided into 18 smaller interconnected
nonlayered compartments varying in length from 15 to 60 km.
Details on the hydrologic parameters for the Fraser River in
each compartment (i.e., flow rates, river width and depth, temperature, concentrations of suspended solids, and pH) and the
food-chain bioaccumulation model are available elsewhere


[36] and within the software. For the present investigation, a
steady-state input of each PBDE congener was assumed to
take place at the start of compartment 17, or near Mission (BC,
Canada) approximately 60 km upstream of where the Fraser
River discharges into the Strait of Georgia.
To approximate a relevant mass input of each PBDE congener at Mission we constructed a commercial PBDE technical
mixture comprised of the 2000 total market demand values of
penta- and octa-BDE comercial mixtures in North America.
Great Lakes Chemical’s commercial penta-BDE mixture (DE71) and Bromkal’s commercial octa-BDE mixture (79-8DE)
were used as representative commercial mixtures. Market demand values of penta- and octa-BDE technical mixtures in
North America in 2000 were reported to be 8,290 and 1,375
tonnes, respectively [10]. Deca-BDE is the largest commercial
PBDE mixture in use in North America at 24,300 tonnes in
2000 [10]. However, with .90% BDE209, its log octanol–
water partition coefficient (KOW) of approximately 10 and vapor pressure of 1.3 3 10212 Pa [37] likely will hinder aquatic
transport. Deca-BDE may be giving rising to the majority of
lesser brominated PBDEs presently observed in the environment through debromination processes (e.g., photolytic or metabolic). If debromination is the major source of current PBDE
patterns in the water column and SPMDs, then we would not
likely be able to successfully reconstruct a PBDE profile consistent with penta- and octa-BDE technical mixtures. Debromination of deca-BDE commercial mixtures may give rise to a
congener pattern consistent with present use patterns of pentaand octa-BDEs, but this is unlikely. Having constructed the
DE-71/79-8DE mixture, each congener was assigned an input
proportional to Vancouver’s population (;2,000,000) relative
to that of North America (;300,000,000), relative to its mass
contribution in the DE-71/79-8DE mixture, proportional to the
rate of penta-/octa-BDE demand in North America in 2000,
and assuming that approximately 1% of consumed PBDEs
leach into the environment each year. We omitted the population of Mexico because PBDE usage is unclear in that region.
For example, BDE47 (30.15% of SPBDE in the DE-71/798DE mixture) was assigned a mass input of 53 g/d (30.15%

Fig. 2. Aquatic transport and semipermeable membrane device uptake model used in the study. PBDE 5 polybrominated diphenyl ether; USGS
5 U.S. Geological Survey; P8L 5 vapor pressure; SW 5 water solubility; KH 5 Henry’s law constant; RS 5 SPMD uptake rate.


Environ. Toxicol. Chem. 21, 2002

S. Rayne and M.G. Ikonomou

Table 3. Values of the reconstruction factors used in the aquatic modela




105 (1.00)
105 (2.21)
106 (4.90)
106 (4.90)
106 (4.90)
106 (4.90)
106 (6.76)
106 (10.84)
106 (10.84)
106 (23.99)
106 (23.99)
106 (23.99)

RS in L·d21 EcoFate factor
(see Eqn. 1)



See Figure 2 for explanation of abbreviations.

3 (9,665,000 kg/year) 3 (2,000,000/300,000,000) 3 (1/365
d) 3 1%). The accuracy of this approximation in assigning
absolute mass inputs for each congener in the EcoFate model
is not critical, as long as the relative inputs of each congener
are consistent with the proposed penta-/octa-BDE mixture described above. The objective of this study was to examine
relative PBDE congener patterns for source apportionment
purposes, not to determine absolute environmental compartment concentrations of these compounds in the Fraser River.
The purpose of using the EcoFate model in this manner
was to assess potential losses of each PBDE congener from
the water column into other environmental compartments during aquatic transport (e.g., volatilization, sedimentation, or bioaccumulation). Losses into these compartments would remove
PBDEs from the potential pool available to be sampled by the
SPMDs. Hence, a loss factor was calculated for each congener
according to its relative predicted water column concentration
at the end of compartment 18 (where the Fraser River discharges into the Strait of Georgia), compared to its starting
concentration at the beginning of compartment 17 (near Mission). This factor is represented as the numerator in Equation
1 below. These loss factors are presented in Table 3, and were
quite small, suggesting that the PBDE congener pattern in the
water column downstream of the input, and that of the source,
were quite similar. Hence, preferential losses of lower brominated congeners to the atmosphere through volatilization
and preferential losses of higher brominated congeners to sedimentation were either minimal or offset each other.
The EcoFate model predicts that the aqueous congener pattern of PBDEs at the SPMD sampling locations is quite similar
to our constructed technical mixture source; however, it is also
necessary to account for preferential partitioning of PBDE
congeners onto particulate organic carbon (POC) and dissolved
organic carbon (DOC). For this, we used the understanding of
SPMD uptake processes largely developed by scientists at the
USGS. The SPMDs, as shown in Figure 2, can only sample
truly dissolved compounds [14,15,17–19,29,30,38] and those
compounds associated with larger POC or DOC moieties will
be much less available for diffusion into the SPMD
[14,18,19,29]. The physicochemical relationship best describing the extent to which hydrophobic compounds are bound to
organic carbon moieties in the water column is the organic
carbon–water partitition coefficient (KOC). Values of KOC are
not available for PBDE congeners, but KOC can be approximated by KOW [14,16,39] and to a first approximation, there
is generally little difference between KOC and KOW for similarly

halogenated aromatic contaminants such as PCBs, PCDD/Fs,
chlorophenols, and chlorobenzenes [14,16,39]. Because we
only were interested in examining relative PBDE profiles in
the freely dissolved phase, rather than converting KOW values
to KOC values in our model by using the equations available
in the literature [14], we chose to use KOW values to approximate to relative partitioning of PBDE congeners between the
organic carbon phase and the freely dissolved phase of the
water column. The use of this equilibrium approximation has
been previously validated to obtain estimates of dissolved
PCBs, chlorobenzenes, and chlorophenols with a log KOW
range from 3.72 to 7.27 [14–16,29], encompassing our range
of PBDE log KOW of 5.47 to 6.85. Although partitioning into
DOC may occur to a lesser extent than into POC [18,40], these
errors are likely minimal compared to those inherent in estimating partitioning coefficients and such calculations in general [18].
Additionally, values of KOW for each congener were obtained via a linear regression equation developed for predicting
PBDE KOW values presented in the literature [41]. The relatively poor fit for this regression equation suggests that any
errors in not using KOC for the partitioning between POC/DOC
and the freely dissolved phase were small compared with errors
in calculating KOW. The relative abundance of each congener
in the freely dissolved phase will be related to the inverse of
its KOW, such that higher brominated congeners will be more
restrained on the POC/DOC phase than their lower brominated
counterparts [14,16,29]. Values for KOW and relative KOW are
presented in Table 3. Although BDE209 has been recently
reported to be taken up by SPMDs in the water column [12],
it was not considered here because its large log KOW (ø10)
suggests it will be primarly bound to POC and DOC and thus
be largely unavailable for SPMD uptake. Furthermore,
BDE209 is relatively difficult to quantify, as evidenced by a
recent worldwide interlaboratory study on PBDE analysis [42],
and further work is likely needed on BDE209 SPMD uptake
rates and analytical methods before reliable data are available.
Finally, uptake rates for each PBDE congener will be different and must be taken into account when reconstructing
SPMD patterns for source apportionment purposes. Generally,
SPMD uptake rates are a function of the hydrophobicity of
the compound (KOW) and water temperature. Work on the
SPMD uptakes rates of polyaromatic hydrocarbons (PAHs) has
provided a multiple linear regression equation for calculating
the uptake rates of similar aromatic compounds provided that
the KOW for the analyte of interest and water temperature are
known [17]. Use of this equation, along with the KOW values
provided in the literature [41] and an average water temperature of 208C, resulted in the calculated values for SPMD
uptake rates (RS) in Table 3. Although this regression equation
was developed for PAHs [17], it has also been successfully
used for the SPMD uptake rates of chlorophenols [16]. As
shown in Table 3, values for RS decrease with increasing KOW
for the PBDE congeners of interest. Similar results have been
reported by our group, and others, for PCDD/Fs and PCBs in
the range of log KOW from approximately 5.5 to .8 [14,18,20].
Steric hindrance may help explain lower uptake rates for compounds with a log KOW . 5.5, because these molecules have
˚ ) near or exceeding pore sizes in the SPMD
diameters (;9 A
membrane [14,18,20]. To approximate water concentrations of
dissolved hydrophobic contaminants, levels observed in
SPMDs are generally divided by the RS for the analyte of
interest [14,16,18,29], consistent with Equation 1 below. No

Environ. Toxicol. Chem. 21, 2002

Reconstructing source PBDE congener patterns


Fig. 3. Observed and reconstructed polybrominated diphenyl ether congener patterns in semipermeable membrane devices (SPMDs) and the
following penta-/octa-bromodiphenyl (BDE) commercial mixes: (A), DE-71/DE-79 Great Lakes Chemical; (B), 70-5DE/79-8DE Bromkal; (C),
DE-71/79-8DE Great Lakes Chemical/Bromkal; and (D), 70-5DE/DE-79 Bromkal/Great Lakes Chemical. Error bars are 95% confidence limits
on the mean.

adjustments were made for biofouling in the present study.
This approximation has proven reasonable in previous studies
of similar compounds [16], although it may help explain some
of the remaining variation in our models.
The mathematical transformations in the reconstruction
process were as follows. The following four penta-/octa-BDE
mixes were constructed by using the four individual pentaand octa-BDE commercial mixtures discussed above (DE-71/
DE-79, 70-5DE/79-8DE, DE-71/79-8DE, and 70-5DE/DE-79).
For each of the 13 congeners, concentrations in individual
SPMDs were converted to a molar basis by dividing by the
respective molecular masses. Once converted to molar concentrations (picomoles PBDE congener per gram SPMD triolein), each congener’s molar concentration was multiplied by
the reconstruction factors as follows
1 OO PBDEs) 2
(C @ O PBDEs)
(C @ O PBDEs) 2
1R 21K 2

(C W@
(C W@

CSource 5 CSPMD


Source EcoFate




Source EcoFate,min






where CW and SPBDE are the concentration of analyte and

sum of concentrations of the 13 PBDE congeners in the water
column at the source and SPMD sampling location, respectively, as calculated by EcoFate; RS is the SPMD uptake rate;
RS,min is the minimum value of RS over the range of congeners
examined; KOW is the octanol–water partition coefficient; and
KOW,min is the minimum value of KOW over the range of congeners examined. The resulting values for CSource were then
converted to a percent of total PBDEs value (sum of 13 reported congeners) by multiplying by each analyte’s molecular
weight and normalizing concentrations to a percent basis.
These calculations resulted in the reconstructed SPMD patterns
shown in Figure 3.

Reconstructed PBDE congener patterns from SPMDs deployed in Fraser River near Vancouver suggest that commercial
penta- and octa-BDE mixtures are the source of PBDEs in
aquatic systems from this urban and industrial area. To the
best of our knowledge, all other studies to date have compared
congener patterns from the various environmental compartments (e.g., sediments, biota, atmosphere, and biosolids)
[37,44–49] directly to the commercial penta-mixtures. The
deca-BDE mixture is not expected to be as mobile after release
because its log KOW of approximately 10 and vapor pressure


Environ. Toxicol. Chem. 21, 2002

of 1.3 3 10212 Pa prevent significant transport [7], although
transport on suspended sediments may be important. Similarly,
hepta- through nona-BDE congeners are typically not found
in appreciable amounts in environmental compartments, as
would be expected from their high log KOW values and low
vapor pressures. However, the small proportion of penta- and
hexa-congeners (e.g., BDEs 99, 100, 153, 154, and 155) found
in the commercial octa-mixtures (e.g., Bromkal 79-8DE) must
be considered in evaluating observed patterns. When using
four individual penta- and octa-BDE mixtures (penta-BDE:
DE-71 and 70-5DE; octa-BDE: DE-79 and 79-8DE), the following four commercial penta-/octa-BDE mixes were constructed in the same proportions as the 2000 North American
market demand levels for these products (8,290 and 1,375
tonnes, respectively): DE-71/DE-79, 70-5DE/79-8DE, DE-71/
79-8DE, and 70-5DE/DE-79. The reconstructed PBDE source
patterns were compared to these technical mixture composites.
Congener patterns of the individual technical mixtures are
shown in Table 2. The two penta-BDE mixtures (DE-71 and
70-5DE) have approximately the same congener patterns, although 70–5DE is slightly enriched in tetra- and penta-BDE
congeners and depleted in hexa-BDEs. Levels are presented
as gram analyte per 100 g of commercial mixture. As is shown
in Table 2, approximately 91% of the two penta-mixtures are
composed of the 13 PBDE congeners we observe in the environment. In contrast to the two penta-BDE mixtures, the two
octa-BDE mixtures (DE-79 and 79-8DE) have very different
congener patterns. The DE-71 is depleted in penta-BDEs compared to 79-8DE, while having larger quantities of hexa-BDEs.
As noted by the sum of the 13 congeners, only 17% and 11%
of DE-79 and 79-8DE, respectively, are made up of monothrough hexa-BDEs. The remainder of BDE congeners for
these two mixtures are hepta- through nona-BDEs. In this regard, DE-79 is more pure than 79-8DE, with less monothrough hexa-BDE impurities than 79-8DE.
Additionally, BDE153 is much more prevalent than
BDE154 in DE-79, whereas 79-8DE has similar quantities of
these two congeners. This distinct difference in congener patterns between the two octa-BDE mixtures suggests different
synthetic methods in their production, such that one method
(for DE-79) favors BDE154 over BDE153, whereas the production method for 79-8DE is less selective between BDEs
154 and 153. This may be the result of kinetic versus thermodynamic control during industrial production (i.e., BDE153
is less sterically hindered than BDE154, so BDE153 is favored
under thermodynamic control). In SPMDs, we tend to observe
approximately equal contributions from BDEs 154 and 153,
which suggests that 79-8DE is the source of hexa-BDEs to the
Fraser River (see below). Further evidence for this claim lies
in the congener patterns of the two penta-BDE mixtures. Both
DE-71 and 70-5DE have greater contributions from BDE153
than BDE154, so that no quantity of penta-BDE mixture can
balance out an excess of BDE153 over BDE154, and can only
increase the relative proportion of BDE153.
The reconstructed PBDE patterns in Figure 3 show that
some form of commercial penta-/octa-BDE mix is the likely
source of PBDEs to the Fraser River. As with many large urban
and industrial regions, PBDEs in Vancouver are expected to
be released directly to local waterways (e.g., the Fraser River)
via sewage and other industrial discharges, storm runoff, and
potentially to a lesser extent by atmospheric deposition. Upon
release to aqueous environments, several physicochemical parameters are expected to govern the resulting congener patterns

S. Rayne and M.G. Ikonomou

observed in the water column by SPMDs. These relationships
are schematically represented by an aquatic transport model
shown in Figure 2. After entering the water column, PBDEs
may fractionate, or be lost, into three major environmental
compartments: bottom sediments, biota, and the atmosphere.
Assuming that degradation is negligible [7] over the relatively
short time span (on the order of several days) PBDEs spend
in the Fraser River water column from Mission to the Strait
of Georgia, each partitioning equilibria can be governed by a
single physicochemical property.
The EcoFate model [36] was used to model these losses as
each PBDE congener travels down the Fraser River from a
location approximately 60 km upstream of Vancouver to where
the river discharges into the Strait of Georgia. This software
has a model of the Fraser River (and its major tributary, the
Thompson River) developed to examine the aquatic transport
of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Because of the relative
similarity in environmental partitioning between chlorinated
dioxins and PBDEs, the EcoFate model of the Fraser River
was used with its default parameters. Inputs of PBDE congeners were relative to their contribution in the four composite
penta-/octa-BDE mixes discussed above. As noted previously,
a steady-state input of each PBDE congener was assumed to
take place at the start of compartment 17 (near Mission), and
the loss of each congener was determined by comparing this
concentration with the concentration where the Fraser River
discharges into the Strait of Georgia (end of compartment 18).
Losses of PBDE congeners in the EcoFate model can be
approximated as follows. Partitioning between the atmosphere
and the water column is largely a function of the Henry’s law
constant, KH. Inputs of KH into the EcoFate model were taken
from the linear regression equation predicting this property for
the PBDE congeners of interest [41]. Both bottom sediments
and biota, because of their high organic contents, have partitioning functions governed by the value of KOW , or KOC .
Compounds with relatively large KOW values (e.g., log KOW .
1 or log KOC . 1) will prefer to reside in sediments and biota
rather than in the water column. For biota, bioconcentration
factors, which are functions of KOW or KOC, have been found
to be superior in describing the partitioning of organic pollutants between the water column and aquatic organisms [50].
The KOW values for PBDE congeners found in the environment
are difficult to determine because of the difficulty in obtaining
analytical-grade standards and the very large partition coefficients, which are generally .106. A linear regression equation
between the log KOW values and number of bromine substituents on PBDEs has been developed recently [41]. This equation was used to calculate the log KOW values for the 13 PBDE
congeners found at significant levels in the Fraser River
SPMDs (Table 3). Because the measured log KOW of BDE100
did not fit well with the literature regression line prediction
(log KOW 5 6.51), the measured log KOW of 6.30 [41] was used
for the calculations described previously.
Also of importance is ensuring that model inputs of PBDEs
to the Fraser River do not approach reported water solubility
limits for these compounds. Near the solubility limit, compounds will precipitate out of the water column onto suspended
and bottom sediments. This would lead to changes in the water
column congener pattern that were not accounted for by the
EcoFate model. As reported elsewhere, flows in the Fraser
River vary on average from a maximum of 10,000 m3/s during
the freshet in June to a minimum of 700 m3/s between October
and February [51]. Use of literature water solubility (SW) val-

Reconstructing source PBDE congener patterns

ues for the major congeners in Table 3 [41] illustrates that
more than five times the worldwide annual PBDE production
volume would need to be released into the Fraser River each
year for congener-specific saturation conditions to occur. Although this approximation assumes complete dissolution and
mixing at the source, which is unrealistic, the results clearly
show that much higher quantities of PBDEs would have to be
assumed to enter the Fraser River than were presently used
for solubility issues to arise.
Additionally, the relatively short hydraulic retention time
expected in the Fraser River between suspected PBDE discharge
regions and the SPMDs, combined with the low reported vapor
pressures [52] for these congeners, suggests that equilibration
with and significant losses to the atmosphere are negligible on
short time scales, as shown by the EcoFate results. Biota are
expected to accumulate PBDEs in a similar manner to organic
sediments with some relationship between the bioconcentration
factor and log KOW [50]. The complex and unknown transformations of PBDEs taking place in aquatic organisms during the
combined effects of uptake, metabolism, and excretion make
this compartment particularly difficult to model. However, the
large quantities of suspended sediment in the Fraser River in
the area of interest (0.09–0.21 g/L [36]) suggest that losses of
these compounds into biota are negligible compared to partitioning onto sediments. As noted above, partitioning onto suspended sediments and losses to bottom sediments are accounted
for by the model in Figure 2.
Any model used to reconstruct PBDE source patterns would
have to account for the observed contribution of BDE47, because this is consistently the most prevalent congener reported
in the literature. Our model is able to reconstruct the contribution of BDE47 that is consistent with a commercial penta-/
octa-BDE source, regardless of which combination of technical
mixtures is used (Fig. 3). Similar results were observed for
BDE153. The contribution of the second major congener,
BDE99, also was significantly closer to the commercial mix
after reconstruction. Reconstruction resulted in all congeners
except BDEs 100 and 155 having a contribution closer to that
of the respective commercial mix than any of the observed
SPMDs. As noted above, the observed log KOW for BDE 100
was quite distant from the reported regression equation, and
no value for BDE 155 was used in calculating the regression
line [41]. Thus, given that the KOW value is important in our
reconstruction approach, small errors in the reported log KOW
values will be magnified greatly during the inverse logarithm
transformation. Overall, the congeners of note above are BDEs
47 and 99, which make up approximately 85% of total PBDEs
in the SPMD samples. Their contributions are significantly
improved with respect to the commercial mixes when using
our approach.
As a general rule, the observed SPMD pattern tends to be
overrepresented by the lower brominated congeners (di- through
tetra-BDEs) while being depleted with respect to congeners with
five or more bromine substituents. This distinction from commercial PBDE mixtures has also been observed elsewhere
[1,2,44,49]. Similar results have been reported for PCBs, PCDD/
Fs, and polychlorinated dibenzothiophenes, where the abundances of SPMD homologue groups generally decrease with
increasing halogenation. This is typically in direct contrast to
the congener patterns observed in sediments [13,16,19,22],
where homologue contributions increase with increasing halogenation, as would be expected by preferential sorption of congeners with higher KOW values. As noted above, compounds

Environ. Toxicol. Chem. 21, 2002


with higher KOW values present in the water column are more
likely associated with POC and DOC, and thus not available
for SPMD uptake [19,29]. The reconstruction process in Figure
2, which uses physicochemical and SPMD uptake properties,
accounts for some of this variation. However, inclusion of a
commercial octa-BDE mixture, which contains small amounts
of penta- and hexa-BDEs, is necessary in an attempt to match
observed environmental patterns of PBDEs to a commercial
Disagreement over the source of PBDEs found in biota and
sediments is apparent in the literature and at scientific conferences. The three schools of thought are as follows: PBDE
patterns found in environmental compartments result from present inputs of the commercial penta- and octa-BDE mixtures,
of which the tetra- through hexa-brominated congeners are
found in the highest levels; environmental debromination of
the commercial deca-BDE mixture, which is produced in significantly greater quantities than the commercial penta- and
octa-BDE mixtures, has produced the congener patterns seen
in the environment; and historically used dispersive commercial tetra-BDE products still reside in sediments and biota, and
continued cycling of these products from sediments to biota
produces the dominance of tetra-BDE congeners observed today. The results presented above suggest that the first school
is likely most realistic, namely that PBDE congener patterns
observed in sediments and biota near urban and industrial
regions arise from a weighted mix of commercial penta- and
octa-BDE mixtures in present use.
Acknowledgement—M.G. Ikonomou acknowledges the Department
of Fisheries and Oceans Environmental Science Strategic Research
Fund, Toxic Substances Research Initiative, and the National Contaminants Program for financial support and the Regional Dioxin Laboratory staff for sample analyses and technical assistance.
1. Ikonomou MG, Rayne S, Fischer M, Fernandez MP, Cretney W.
2002. Occurrence and congener profiles of polybrominated diphenyl ethers (PBDEs) in environmental samples from coastal
British Columbia, Canada. Chemosphere 46:649–663.
2. Ikonomou MG, Rayne S, Addison RF. 2002. Exponential increases of the brominated flame retardants, polybrominated diphenyl
ethers, in the Canadian Arctic from 1981–2000. Environ Sci Technol 36:1886–1892.
3. Strandberg B, Dodder NG, Basu I, Hites RA. 2001. Concentrations and spatial variations of polybrominated diphenyl ethers
and other organohalogen compounds in Great Lakes air. Environ
Sci Technol 35:1078–1083.
4. Schaefer A. 2001. Lake Michigan heavily contaminated with
PBDEs. Environ Sci Technol 35:139A–140A.
5. Manchester-Neesvig JB, Valters K, Sonzogni WC. 2001. Comparison of polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) in Lake Michigan salmonids. Environ Sci Technol 35:1072–1077.
6. Renner R. 2000. What fate for brominated fire retardants. Environ
Sci Technol 34:222A–226A.
7. de Boer J, de Boer K, Boon JP. 1999. Polybrominated biphenyls
and diphenylethers: In Paasivirta J, ed, The Handbook of Environmental Chemistry: New Types of Persistent Halogenated
Compounds. Springer-Verlag, New York, NY, USA, pp 62–95.
8. van Esch GJ. 1994. Polybrominated Diphenyl Ethers. Environmental Health Criteria 162. World Health Organization, Geneva,
9. Arias PA. 2001. Brominated flame retardants—An overview. Proceedings, Second International Workshop on Brominated Flame
Retardants, Stockholm, Sweden, May 14–16, pp 3–4.
10. BSEF. 2000. An Introduction to Brominated Flame Retardants.
Bromine Science and Education Forum, Brussels, Belgium.
11. Hooper K, McDonald TA. 2000. The PBDEs: An emerging environmental challenge and another reason for breast-milk monitoring programs. Environ Health Perspect 108:387–392.


Environ. Toxicol. Chem. 21, 2002

12. Booij K, Zegers BN, Boon JP. 2002. Levels of some polybrominated diphenyl ether (PBDE) flame retardants along the Dutch
coast as derived from their accumulation in SPMDs and blue
mussels (Mytilus edulis). Chemosphere 46:683–688.
13. Gale RW, Huckins JN, Petty JD, Peterman PH, Williams LL,
Morse D, Schwartz TR, Tillitt DE. 1997. Comparison of the uptake of dioxin like compounds by caged channel catfish and semipermeable membrane devices in the Saginaw River, Michigan.
Environ Sci Technol 31:178–187.
14. Meadows JC, Echols KR, Huckins JN, Borsuk FA, Carline RF,
Tillitt DE. 1998. Estimation of uptake rate constants for PCB
congeners accumulated by semipermeable membrane devices and
brown trout (Salmo trutta). Environ Sci Technol 32:1847–1852.
15. Hofelt CS, Shea D. 1997. Accumulation of organochlorine pesticides and PCBs by semipermeable membrane devices and Mytilus edulis in New Bedford Harbor. Environ Sci Technol 31:
16. Granmo A, Ekelund R, Berggren M, Brorstrom-Lunden E,
Bergqvist PA. 2000. Temporal trend of organochlorine marine
pollution indicated by concentrations in mussels, semipermeable
membrane devices, and sediment. Environ Sci Technol 34:3323–
17. Huckins JN, Petty JD, Orazio CE, Lebo JA, Clark RC, Gibson
VL, Gala WR, Echols KR. 1999. Determination of uptake kinetics
(sampling rates) by lipid-containing semipermeable membrane
devices (SPMDs) for polycyclic aromatic hydrocarbons (PAHs)
in water. Environ Sci Technol 33:3918–3923.
18. Luellen DR, Shea D. 2002. Calibration and field verification of
semipermeable membrane devices for measuring polycyclic aromatic hydrocarbons in water. Environ Sci Technol 36:1791–
19. Echols KR, Gale RW, Schwartz TR, Huckins JN, Williams LL,
Meadows JC, Morse D, Petty JD, Orazio CE, Tillitt DE. 2000.
Comparing polychlorinated biphenyl concentrations and patterns
in the Saginaw River using sediment, caged fish, and semipermeable membrane devices. Environ Sci Technol 34:4095–4102.
20. Rantalainen AL, Cretney WJ, Ikonomou MG. 2000. Uptake rates
of semipermeable membrane devices (SPMDs) for PCDDs,
PCDFs and PCBs in water and sediment. Chemosphere 40:147–
21. Vrana B, Schu¨u¨rmann G. 2002. Calibrating the uptake kinetics
of semipermeable membrane devices in water: Impact of hydrodynamics. Environ Sci Technol 36:290–296.
22. Rantalainen A-L, Ikonomou MG, Rogers IH. 1998. Lipid-containing semipermeable membrane devices (SPMDs) as concentrators of toxic chemicals in the lower Fraser River, Vancouver,
British Columbia. Chemosphere 37:1119–1138.
23. Ockenden WA, Corrigan BP, Howsam M, Jones KC. 2001. Further developments in the use of semipermeable membrane devices
as passive air samplers: Application to PCBs. Environ Sci Technol
24. Lohmann R, Corrigan BP, Howsam M, Jones KC, Ockenden WA.
2001. Further developments in the use of semipermeable membrane devices (SPMDs) as passive air samplers for persistent
organic pollutants: Field application in a spatial survey of PCDD/
Fs and PAHs. Environ Sci Technol 35:2576–2582.
25. Ockenden WA, Sweetman AJ, Prest HF, Steinnes E, Jones KC.
1998. Toward an understanding of the global atmospheric distribution of persistent organic pollutants: The use of semipermeable
membrane devices as time-integrated passive samplers. Environ
Sci Technol 32:2795–2803.
26. Ockenden WA, Prest HF, Thomas GO, Sweetman A, Jones KC.
1998. Passive air sampling of PCBs: Field calculation of atmospheric sampling rates by triolein-containing semipermeable
membrane devices. Environ Sci Technol 32:1538–1543.
27. Thomas G, Sweetman AJ, Ockenden WA, Mackay D, Jones KC.
1998. Air–pasture transfer of PCBs. Environ Sci Technol 32:936–
28. Strandberg B, Wagman N, Bergqvist PA, Haglund P, Rappe C.
1997. Semipermeable membrane devices as passive samplers to
determine organochlorine pollutants in compost. Environ Sci
Technol 31:2960–2965.
29. Huckins JN, Manuweera GK, Petty JD, Mackay D, Lebo JA.
1993. Lipid-containing semipermeable membrane devices for
monitoring organic contaminants in water. Environ Sci Technol
30. Gale RW. 1998. Three-compartment model for contaminant accumulation by semipermeable membrane devices. Environ Sci
Technol 32:2292–2300.
31. Petty JD, Orazio CE, Huckins JN, Gale RW, Lebo JA, Meadows

S. Rayne and M.G. Ikonomou














JC, Echols KR, Cranor WL. 2000. Considerations involved with
the use of semipermeable membrane devices for monitoring environmental contaminants. J Chromatogr A 879:83–95.
Huckins JN, Petty JD, Lebo JA, Almeida FV, Booij K, Alvarez
DA, Clark RC, Mogensen BB. 2002. Development of the permeability/performance reference compound approach for in situ
calibration of semipermeable membrane devices. Environ Sci
Technol 36:85–91.
Gustavson KE, Harkin JM. 2000. Comparison of sampling techniques and evaluation of semipermeable membrane devices
(SPMDs) for monitoring polynuclear aromatic hydrocarbons
(PAHs) in groundwater. Environ Sci Technol 34:4445–4451.
MacDonald DD, Ikonomou MG, Rantalaine AL, Rogers IH, Sutherland D, Van Oostdam J. 1997. Contaminants in white sturgeon
(Acipenser transmontanus) from the upper Fraser River, British
Columbia, Canada. Environ Toxicol Chem 16:479–490.
Macdonald RW, Ikonomou MG, Paton DW. 1998. Historical inputs of PCDDs, PCDFs, and PCBs to a British Columbia interior
lake: The effect of environmental controls on pulp mill emissions.
Environ Sci Technol 32:331–337.
Gobas FAPC, Pasternak JP, Lien K, Duncan RK. 1998. Development and field validation of a multimedia exposure assessment
model for waste load allocation in aquatic ecosystems: Application to 2,3,7,8-tetrachlorodibenzo-p-dioxin and 2,3,7,8-tetrachlorodibenzofuran in the Fraser River watershed. Environ Sci
Technol 32:2442–2449.
Strandberg B, Dodder NG, Basu I, Hites RA. 2001. Concentrations and spatial variations of polybrominated diphenyl ethers
and other organohalogen compounds in Great Lakes air. Environ
Sci Technol 35:1078–1083.
Booij K, Sleiderink HM, Smedes F. 1998. Calibrating the uptake
kinetics of semipermeable membrane devices using exposure
standards. Environ Toxicol Chem 17:1236–1245.
Di Toro DM. 1985. A particle interaction-model of reversible
organic-chemical sorption. Chemosphere 14:1503–1538.
Burkhard LP. 2000. Estimating dissolved organic carbon partition
coefficients for nonionic organic chemicals. Environ Sci Technol
Tomy G, Tittlemier S, Braekevelt E, Halldorson T, Stern G. 2001.
The physico-chemical properties of some brominated flame retardants. Proceedings, Second International Workshop on Brominated Flame Retardants, Stockholm, Sweden, May 14–16, pp
de Boer J. 2000. Report on the BSEF PBDE interlaboratory study.
Organohalogen Compound 45:118–121.
Mckim J, Schmieder P, Veith G. 1985. Absorption dynamics of
organic-chemical transport across trout gills as related to octanol
water partition-coefficient. Toxicol Appl Pharm 77:1–10.
Hale RC, La Guardia MJ, Harvey EP, Gaylor MO, Mainor TM,
Duff WH. 2001. Persistent pollutants in land-applied sludges.
Nature 412:140–141.
Manchester-Neesvig JB, Valters K, Sonzogni WC. 2001. Comparison of polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) in Lake Michigan salmonids. Environ Sci Technol 35:1072–1077
Allchin CR, Law RJ, Morris S. 1999. Polybrominated diphenylethers in sediments and biota downstream of potential sources in
the UK. Environ Pollut 105:197–207.
Meironyte D, Noren K, Bergman A. 1999. Analysis of polybrominated diphenyl ethers in Swedish human milk. A time-related trend study, 1972–1997. J Toxicol Environ Health A 58:
Jansson B, Andersson R, Asplund L, Litzen K, Nylund K, Sellstrom U, Uvemo U-B, Wahlberg C, Wideqvist U, Odsjo T, Olsson
M. 1993. Chlorinated and brominated persistent organic compounds in biological samples from the environment. Environ Toxicol Chem 12:1163–1174.
Christensen JH, Platz J. 2001. Screening of polybrominated diphenyl ethers in blue mussels, marine and freshwater sediments
in Denmark. J Environ Monit 2001:543–547.
Meylan WM, Howard PH, Boethling RS, Aronson D, Printup H,
Gouchie S. 1999. Improved method for estimating bioconcentration/bioaccumulation factor from octanol/water partition coefficient. Environ Toxicol Chem 18:664–672.
Yin K, Harrison PJ, Beamish RJ. 1997. Effects of a flucuation in
Fraser River discharge on primary production in the central Strait
of Georgia, British Columbia, Canada. Can J Fish Aquat Sci 54:
Wong A, Lei YD, Alaee M, Wania F. 2001. Vapor pressures of
the polybrominated diphenyl ethers. J Chem Eng Data 46:239–

Related documents

environ toxicol chem 21 2002 2292 2300
environ sci technol 36 2002 1886 1892
chemosphere 46 2002 649 663
environ sci technol 37 2003 2847 2854
science news ringed seals
environ sci technol 38 2004 4293 4299

Related keywords