Occurrence of Methyl Triclosan, a Transformation Product of the Bactericide Triclosan, in Fish from Various Lakes in Switzerland

Article abstract of DOI:10.1021/es030068p

The bactericide triclosan and methyl triclosan, an environmental transformation product thereof, have been previously detected in lakes and a river in Switzerland. Both compounds are emitted via wastewater treatment plants (WWTPs), with methyl triclosan probably being formed by biological methylation. Passive sampling with semi-permeable membrane devices (SPMDs) showed the presence of methyl triclosan in some lakes, suggesting some potential for bioaccumulation of the compound. In this study, we report the presence of methyl triclosan in fish (white fish, coregonus sp.; roach, rutilus rutilus) from various lakes in Switzerland receiving inputs from WWTPs. Identification of the compound was based on mass spectral (MS) evidence including MS/MS data. The concentrations of methyl triclosan in the fish were up to 35 ng g^-1 on a wet weight basis and up to 365 ng g ^-1 on a lipid basis with concentrations in a relatively narrow range for fish from the same lake (Thunersee, 4-6 ng g^-1 Zuerichsee, 32-62 ng g^-1; Pfaeffikersee, 43-56 ng g^-1; Greifensee, 165-365 ng g^-1, lipid basis). No methyl triclosan (<1 ng g ^-1, lipid basis) was detected in fish (lake trout, salmo trutta) from a remote lake in Sweden (Haebberstjaernen) and in fish (roach) from a small lake in Switzerland with no input from WWTPs (Huettnersee, <2-<5 ng g^-1, lipid basis). The concentrations of methyl triclosan in fish correlated (r² = 0.85) with the ratio of population in the watershed to water throughflow of the lakes (P/Q ratio), which is considered to be a measure for the domestic burden from WWTPs to a lake. Passive sampling with SPMDs confirmed the presence of methyl triclosan in lakes and a river (Zuerichsee and Greifensee; Limmat) but not in a remote mountain lake (Joerisee) and in Huettnersee. The bioconcentration factor (BCF) of methyl triclosan estimated from the fish data and SPMD-derived water concentrations was in the order of 1-2.6 × 10⁵ (lipid basis) and thus in the range of other persistent organic pollutants. SPMDs were found to be reliable for monitoring low concentrations of methyl triclosan in surface water. Methyl triclosan appears to be a suitable marker for WWTP-derived lipophilic contaminants in the aquatic environment and fish.

Full text of DOI:10.1021/es030068p

Environ. Sci. Technol. 2004, 38, 390-395
be reliable for monitoring low concentrations of methyl
Occurrence of Methyl Triclosan, a
Transformation Product of the
Bactericide Triclosan, in Fish from
Various Lakes in Switzerland
M A R I A N N E E . B A L M E R , * , †
T H O M A S P O I G E R , † C H R I S T I A N D R O Z , ‡
K A T H R I N R O M A N I N , ‡
P E R - A N D E R S B E R G Q V I S T , §
M A R K U S D . M U¨ L L E R , † A N D
H A N S - R U D O L F B U S E R †
triclosan in surface water. Methyl triclosan appears to be
a suitable marker for WWTP-derived lipophilic contaminants
in the aquatic environment and fish.
Introduction
Triclosan (5-chloro-2-(2,4-dichlorophenoxy)phenol; struc-
ture, see Figure 1) is an important bactericide used in various
personal care and consumer products such as shampoo, toilet
soap, deodorants, toothpaste, footwear, and plastic ware.
Triclosan is a relatively stable, lipophilic compound. Its
environmental occurrence has been reported, and the
compound has been detected in wastewater, activated sludge,
surface water, and sediments (1-11). However, the envi-
ronmental fate and behavior of triclosan is not fully under-
stood, although photolysis was identified as a major deg-
radation pathway in surface waters (8, 9). The compound is
acutely and chronically toxic to aquatic organisms. For
rainbow trout, a median effective concentration (EC50) of
350 µg L-1 and a no-observed-effect concentration (NOEC)
of 34 µg L-1 is reported (1, 2, 12).
Swiss Federal Research Station, CH-8820 Wa¨denswil,
Switzerland, Cantonal Food Inspectorate, CH-9001 St. Gallen,
Switzerland, and Institute of Environmental Chemistry,
Umea˚ University, S-90187 Umea˚, Sweden
The bactericide triclosan and methyl triclosan, an
There is even less information on methyl triclosan (5-
chloro-2-(2,4-dichlorophenoxy)anisole; Figure 1), an envi-
ronmental transformation product of triclosan. Previously,
the occurrence of methyl triclosan in samples from waste-
water treatment plants (WWTPs) and in lakes and in a river
in Switzerland was reported (9, 11). Both compounds,
triclosan and methyl triclosan, are emitted to surface waters
via WWTPs with methyl triclosan likely being formed by
biological methylation in the plants. Our previous study (9)
revealed that the environmental behavior and fate of the
two compounds is very different: whereas triclosan in its
dissociated form was degraded easily by photolysis, methyl
triclosan was much more persistent. Furthermore, methyl
triclosan was observed in semipermeable membrane devices
(SPMDs) but not triclosan itself. Although the concentrations
of methyl triclosan were very low in surface waters (generally
<1 ng L-1), the concentrations observed in exposed SPMDs
were in the nanogram per membrane range (>15 ng SPMD-1),
suggesting some potential for bioaccumulation of the
compound in fish and other aquatic organisms. In fact, the
presence of methyl triclosan in fish was previously reported
(13). In this study, we report on the occurrence of methyl
triclosan in fish from various lakes in Switzerland. We
investigated to what extent the compound was accumulated
in fish and evaluated whether methyl triclosan is a suitable
chemical marker for the WWTP-derived lipophilic contami-
nants in the aquatic environment and fish.
environmental transformation product thereof, have been
previously detected in lakes and a river in Switzerland.
Both compounds are emitted via wastewater treatment
plants (WWTPs), with methyl triclosan probably being formed
by biological methylation. Passive sampling with semi-
permeable membrane devices (SPMDs) showed the presence
of methyl triclosan in some lakes, suggesting some
potential for bioaccumulation of the compound. In this
study, we report the presence of methyl triclosan in fish
(white fish, coregonus sp.; roach, rutilus rutilus) from various
lakes in Switzerland receiving inputs from WWTPs.
Identification of the compound was based on mass spectral
(MS) evidence including MS/MS data. The concentrations
of methyl triclosan in the fish were up to 35 ng g-1 on
a wet weight basis and up to 365 ng g-1 on a lipid basis
with concentrations in a relatively narrow range for fish
from the same lake (Thunersee, 4-6 ng g-1; Zu¨richsee,
32-62 ng g-1; Pfa¨ffikersee, 43-56 ng g-1; Greifensee,
165-365 ng g-1, lipid basis). No methyl triclosan (<1 ng
g-1, lipid basis) was detected in fish (lake trout, salmo trutta)
from a remote lake in Sweden (Ha¨bberstja¨rnen) and in
fish (roach) from a small lake in Switzerland with no input
from WWTPs (Hu¨ttnersee, <2-<5 ng g-1, lipid basis).
The concentrations of methyl triclosan in fish correlated
(r2 ) 0.85) with the ratio of population in the watershed to
water throughflow of the lakes (P/Q ratio), which is
considered to be a measure for the domestic burden from
WWTPs to a lake. Passive sampling with SPMDs confirmed
the presence of methyl triclosan in lakes and a river (Zu¨richsee
and Greifensee; Limmat) but not in a remote mountain
lake (Jo¨risee) and in Hu¨ttnersee. The bioconcentration factor
(BCF) of methyl triclosan estimated from the fish data
and SPMD-derived water concentrations was in the order
of 1-2.6 × 105 (lipid basis) and thus in the range of
other persistent organic pollutants. SPMDs were found to
Experimental Section
Chem icals and Reference Com pounds. The sources and
purities of the compounds were as follows: triclosan, purity,
99% (Interkantonale Kontrollstelle fu¨ r Heilmittel, IKS,
Berne, Switzerland). Methyl triclosan was prepared from
triclosan by methylation with diazomethane (see ref 14
13
and cautionary notes therein). A cocktail of C12-PCBs
13
13
including C12-2,4,2′,4′- and C12-2,5,2′,5′-tetrachlorobi-
phenyl (13C12-PCB#47 and 13C12-PCB#52, respectively; 250 pg
µL-1 each) was used as internal standard; in an alternate
13
method, C12-3,4,3′,4′-tetrachlorobiphenyl (13C12-PCB#77)
was used as internal standard (all 13
C12-PCBs from Cambridge
* Corresponding author phone: ++41 1 783 6266; fax: ++41 1
783 6439; e-mail: marianne.balmer@faw.admin.ch.
† Swiss Federal Research Station.
Isotope Laboratories, Cambridge, MA).
Description of the Lakes. Fish from several lakes in
Switzerland were analyzed. Of particular importance for this
study were lake Greifensee, Pfa¨ffikersee, Zu¨ richsee, and
‡ Cantonal Food Inspectorate.
§
Institute of Environmental Chemistry.
9
3 9 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 2, 2004
10.1021/es030068p CCC: $27.50
2004 Am erican Chem ical Society
Published on Web 12/06/2003

For some samples, an alternate procedure using acceler-
ated solvent extraction (ASE, Dionex, Zurich, Switzerland)
was applied: 5.0 g of the homogenized fish filets were mixed
with 10 g of hydromatrix (Chem-Tube Hydromatrix, Varian,
Steinhausen, Switzerland) and extracted with cyclohexane/
dichloromethane (1:1) at room temperature and a pressure
of 1500 psi, during 9 min (3 cycles, with a total solvent volume
of about 40 mL). These lipid extracts were fortified with 8 ng
of 13C12-PCB#77 as internal standard. One-half of the sample
was then subjected to GPC cleanup. Both procedures, column
extraction and ASE, gave comparable results.
FIGURE 1. Structure of the bactericide triclosan and its transforma-
tion product methyl triclosan.
Thunersee. The lakes differ with respect to population in the
watershed, morphology, and hydraulics (Table 1 and refs 15
and 16). Greifensee, Pfa¨ffikersee, and Zu¨ richsee are located
in the Swiss midland region and Thunersee in the prealpine
region. All four lakes receive inputs from WWTPs. More than
99% of the households are connected to WWTPs. In addition,
fish from a remote lake in Sweden (Ha¨bberstja¨rnen) with no
anthropogenic sources nearby and fish from Hu¨ ttnersee, a
small lake near Zu¨ richsee in a more densely populated area
but with no inputs from WWTPs, were analyzed. No fish
were available from the remote mountain lake (Jo¨ risee, 2450
m asl, Switzerland) which was sampled for background
measurements using SPMDs (see below).
GPC was carried out using either a Phenomenex Envi-
roSep-ABC Guard (75 × 22.5 mm) as precolumn and an
EnviroSep-ABC preparative column (300 × 22.5 mm) using
dichloromethane/ n-hexane (35:65) at 3.7 mL/ min or (alter-
nate method) a Biobeads S-X3 (410 × 15 mm) column with
dichloromethane/ cyclohexane (35:65) at 3.0 mL/ min as
mobile phase, respectively (for further details, see ref 19).
Appropriate GPC fractions were collected as established using
the reference compounds and UV detection (retention time
windows, 23-60 and 14.5-27 min, respectively). The GPC
fraction containing the organochlorine compounds were
carefully concentrated and subjected to further cleanup on
a silica minicolumn (Pasteur pipet with 700-800 mg silica
gel 60, Merck, Darmstadt, Germany, dried at 140 °C, and
deactivated with 5% water), using 10 mL of n-hexane/ ethyl
acetate (10:1) as eluent. This cleanup removed interfering
compounds such as fatty acids that would otherwise prevent
proper analysis by gas chromatography-mass spectrometry
(GC-MS). After concentration, the residues were taken up in
50-200 µL of ethyl acetate, and an aliquot of 1-2 µL was
used for GC-MS analysis.
Fish Sam ples. The species analyzed were white fish
(coregonus sp.) and roach (rutilus rutilus) from Zu¨ richsee
and Greifensee, white fish from Pfa¨ffikersee and Thunersee,
roach from Hu¨ ttnersee, and lake trout (salmo trutta) from
the Swedish lake Ha¨bberstja¨rnen. The fish species differed
because the same species were not available from each lake.
The fish were obtained from local fishermen, who were
instructed to handle the fish with great care to prevent
contamination. All fish were weighed and sealed in aluminum
foil and polyethylene bags on the day of catch and kept frozen
at -20 °C until analyzed. All fish were over two years of age
(see Table 2 for details).
Exposures and Extraction of SPMDs. SPMDs were
exposed for ∼3 weeks in Zu¨ richsee, Greifensee, and Hu¨ tt-
nersee and for ∼6 weeks in the remote mountain lake (Jo¨risee)
in Switzerland. In Zu¨ richsee and Greifensee, the SPMDs were
placed at the lake outflows. In Hu¨ ttnersee, the SPMDs were
exposed in the middle of the lake and in Jo¨ risee near the
shore. All SPMDs were exposed at a depth of about 1 m.
Additionally, SPMDs were exposed in the river Limmat, at
a location 15 km below its source at the outflow of Zu¨ richsee,
6 km downstream of the WWTP of the city of Zu¨ rich
(population serviced by the plant, 370 000). SPMDs were
exposed at all locations in summer 2002, in Zu¨ richsee and
the river Limmat additionally in spring 2002. Two membranes
each were deployed in Zu¨ richsee, Greifensee, and Hu¨ ttnersee
in summer and analyzed separately for replicate measure-
ments (for details, see Table 3). The standard SPMDs
consisted of thin layflat semipermeable polyethylene tubings
(2.54 cm ×91.4 cm) containing 1.0 mL of triolein (Exposmeter
AB, Tavelsjo¨ , Sweden). All exposed and two blank SPMDs for
control purposes were dialyzed with cyclopentane/ dichlo-
romethane (95:5) in the same way as previously described
Extraction of Fish Sam ples. For analysis, the edible parts
of the fish were used (filet), scales and fish bones were
removed, skin was not or only partly included. A 25-g sample
was homogenized with 100 g of anhydrous Na2SO4 (dried at
540 °C for 12 h) and mixed with 150 mL of solvent
(cyclohexane/ dichloromethane, 1:1). The homogenized mix-
13
ture was fortified with 80 µL of the C12-PCB cocktail (20 ng
13
of each C12-PCB) and then quantitatively transferred into
a glass column (3 cm × 25 cm). Lipids were extracted with
an additional 200 mL of solvent. The raw extract was
additionally percolated through Na2SO4 to remove remaining
water. The dried extract was then carefully concentrated to
constant weight for determination of the lipid content. The
residue was redissolved in a suitable volume of the GPC
solvent (0.75-6 mL), and an aliquot of 0.75 mL, corresponding
to 170 mg of lipid, was subjected to gel permeation chro-
matography (GPC) for separation of lipid and analytes. The
procedure largely followed those described in the literature
(17, 18).
TABLE 1. Data on Selected Lakes in Switzerland
population
surface
area,
depth
mean water flushing
water
population
per throughflow
of water,
volume,
max (mean), residence rate,e kw througflow, in watershed,
lake
V (km3) A (km2)
h (m)
time,d τ (d)
(d-1
)
Q (m3 s-1
)
P (persons) P/Q (persons m-3 d)
Greifenseea
Zu¨richseea
Thunerseea
0.15
3.83
6.44
0.06
0.01
8.45
68.1
48.4
3.03
0.165
∼0.05
∼0.15
32 (18)
136 (51)
217 (135)
35 (19)
420
440
694
770
120
0.0024
0.0023
0.0014
0.0013
0.0083
4.08
89.2
110
0.9
0.1
107 000
330 000
94 300
12 900
0.303
0.043
0.010
0.174
0.000
0.000
0.000
Pfa¨ffikerseeb
Hu¨ttnerseeb
13 (6)
0
0
0
J o¨riseec
Ha¨bberstja¨rnen (Sweden )c
a
b
c
d
e
Data from ref 15. Reference 16. No hydraulic param eters are available for J o¨risee and Ha¨bberstja¨rnen. Calculated: τ ) V/Q. Inverse of
water residence tim e τ.
9
VOL. 38, NO. 2, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 1

TABLE 2. Methyl Triclosan Concentrations (ng g-1, Lipid Basis) in Fish from the Swiss Lakes and the Remote Swedish Lake
lake
fish species
date of catch
weight (g)
fat (%)
methyl triclosan (ng g-1 lipid)
Zu¨richsee
white fish
white fish
white fish
white fish
white fish
roach
roach
roach
roach
roach
Nov 14, 2001
Nov 14, 2001
Nov 14, 2001
J an 18, 2002
J an 18, 2002
J an 18, 2002
J an 18, 2002
J un 5, 2002
86
133
237
270
275
272
298
136
145
338
7.5
6.8
5.5
2.6
0.8
3.1
5.2
1.1
1.6
1.3
55
38
47
48
32
46
40
62
56
Aug 15, 2002
Aug 15, 2002
46
average
white fish
white fish
roach
roach
roach
47 ( 9.0b
165
211
307
365
200
300
258 ( 77b
5.5
Greifensee
J an 30, 2002
J an 30, 2002
J an 30, 2002
J an 30, 2002
Aug 9, 2002
Aug 9, 2002
470
453
373
366
338
192
0.9
1.1
1.6
0.4
0.3
0.8
roach
average
white fish
white fish
white fish
average
white fish
white fish
average
roach
Thunersee
J an 22,2002
J an 22, 2002
J an 22, 2002
177
170
170
7.6
8.6
1.1
4.2
6.4
5.4 ( 1.1b
43
Pfa¨ffikersee
Sep 1, 2002
Sep 1, 2002
651
561
2.8
4.8
56
50 ( 9.2b
<5
Hu¨ttnersee
Sep 8, 2002
Sep 8, 2002
Aug 2000
268
412
0.2
0.2
1.4
roach
lake trout
<2
<1
Ha¨bberstja¨rnen
n.d.a
a
b
n.d.) not determ ined. Average ( standard deviation.
TABLE 3. Methyl triclosan Concentrations in the Exposed SPMDs
lake/river
date
duration (d) no. of SPMDs
average water temperature (°C) methyl triclosan (ng SPMD-1
)
J o¨risee
Hu¨ttnersee
Zu¨richsee
J ul 30-Sep 9, 2002
J ul 8-J ul 26, 2002
Apr 30-May 21, 2002
Aug 14-Sep 4, 2002
J ul 17-Aug 7, 2002
Apr 30-May 21, 2002
Aug 15-Sep 4, 2002
41
18
21
21
21
21
20
2
2
1
2
2
1
1
10
21
12
21
22
12
21
<1, <1
<1, ∼1
25
28, 32
56, 85
87
Greifensee
Lim m at
137
(9, 19). The GPC cleanup of concentrated extracts was carried
out according to the alternate method reported above for
the fish samples.
The amounts of methyl triclosan were determined from
13
peak area ratios relative to C12-PCB#47 (13C12-PCB#77 in
the alternate method) and in reference to suitable standard
solutions. Quality control fish samples (roach, Zu¨ richsee, 1.3%
lipids) fortified with 100 ng of methyl triclosan per 5 g of fish
showed acceptable recoveries (76-108%); the recoveries of
methyl triclosan from fortified SPMD extracts were 88-103%.
The limit of detection for methyl triclosan in fish was <1 ng
g-1 when calculated on a lipid basis for fish with >1% lipid,
up to 5 ng g-1 for fish with less lipids, and <1 ng SPMD-1 in
SPMDs. Method blanks and fish fortified with 200 ng of
triclosan per 5 g of fish showed no (<1 ng g-1, lipid basis)
methyl triclosan, and thus no contamination or conversion
of triclosan to methyl triclosan during sample preparation
occurred. The data reported are not corrected for recoveries.
GC-MS Analysis. GC-MS analysis was performed on a VG
Tribrid magnetic sector mass spectrometer (MassLab Group,
Manchester, U.K.) under electron-impact ionization (EI, 70
eV, 200 °C) and full-scan (m/ z 35-435; mass resolution M/ ∆M
) 500), selected-ion-monitoring (SIM), or selected-reaction-
monitoring (SRM, MS/ MS) conditions, using a 25-m DB-5
fused silica column (0.32 mm i.d.) and split/ splitless injection
(250 °C). The column was temperature programmed as
follows: 70 °C, 2-min isothermal, 20 °C/ min to 140 °C, then
at 5 °C/ min to 280 °C, followed by a 10-min isothermal hold
at this temperature to elute late-eluting steroids (cholesterol)
and other components. Data acquisition was started at 140
°C, and retention times were measured from this point. Under
these conditions, methyl triclosan was eluted at 12.7 min. In
the SIM mode, the molecular ions (M+•) at m/ z 302 for methyl
Results and Discussion
Occurrence of Methyl Triclosan in Fish. A chlorinated
component subsequently identified as methyl triclosan was
observed in all fish, except the control fish from the re-
mote lake in Sweden and the fish from the small lake in
Switzerland (Hu¨ ttnersee) with no inputs from WWTPs, as
illustrated in Figure 2. The chromatograms also show the
13
triclosan and C12-tetra-PCBs were monitored; Cl-isotope
satellite ions (m/ z 304) and fragment ions (m/ z 252 and 254)
were also monitored for confirmatory purposes. In the SRM
mode, the formation of specific daughter ions generated from
selected parent ions (i.e., the molecular ion M+•) was
monitored using argon as collision gas (for details, see ref
20). The ion transitions 302+ f 252+, 304+ f 252+, and 304+
f 254+ (loss of CH3Cl) were used to monitor methyl triclosan,
and the ion transition 302+ f 267+ (loss of Cl) was used to
13
presence of the earlier-eluted C12-tetra-PCB internal stan-
dards (13C12-PCB#47 and 13C12-PCB#52), which responded at
the same m/ z value as methyl triclosan. The identification
of the new component as methyl triclosan was based on the
fact that the compound coeluted on the GC column with the
13
detect C12-tetra-PCBs.
9
3 9 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 2, 2004

position (21). The SRM chromatograms for the monoisotopic
transition 302+ f 252+ and its two isotopic transitions 304+
f 252+ and 304+ f 254+ confirmed the presence of methyl
triclosan in these fish. The relative abundances of the peaks
were consistent with those expected for a trichloro compound
losing a neutral fragment with one Cl (abundance ratio
observed, 1.00:0.68:0.34; theoretical, 1.00:0.70:0.37).
Methyl triclosan was only detected in fish from lakes with
inputs from WWTPs. The concentrations in fish (wet weight
basis) ranged from <0.2 ng g-1 for the fish from the remote
lake in Sweden to 35 ng g-1 for fish from Greifensee. When
the concentrations were calculated on a lipid basis, a clear
pattern was observed: the lowest concentrations in fish from
Swiss lakes with inputs from WWTPs were observed in the
fish from Thunersee (4-6 ng g-1). Higher concentrations
were observed in fish from Pfa¨ffikersee (43-56 ng g-1) and
Zu¨ richsee (32-62 ng g-1), and the highest concentrations
were observed in fish from Greifensee (165-365 ng g-1, see
Table 2). The lipid-based concentrations in fish from the
same lakes were in relatively narrow ranges, although there
was a tendency of somewhat higher concentrations in roach
than in white fish and in smaller than larger fish. Species and
weight are known to be factors that influence bioconcen-
tration (22), but the observed variations were not considered
to be statistically relevant. In fish from Hu¨ ttnersee, a lake
with no inputs from WWTPs, methyl triclosan was not
detected (<5 ng g-1, lipid basis). Our results are consistent
with an earlier Japanese study, where increasing concentra-
tions of methyl triclosan in fish (1-38 ng g-1, wet weight
basis) were observed downstream an approximately 40-km
stretch of the Tama River (13).
Methyl triclosan in SPMDs and SPMD-Derived Con-
centrations in Water. Because of the generally low concen-
trations of methyl triclosan in surface water, direct mea-
surement is difficult. To get more information about relative
concentrations in the lakes, SPMD sampling was used. SPMDs
measure time-weighted averages of concentrations of dis-
solved lipophilic pollutants in water. Furthermore, they can
serve as a model for the bioconcentration of compounds in
fish. In Table 3 we report methyl triclosan concentrations in
the exposed SPMDs. Methyl triclosan was not detected in
SPMDs from the remote mountain lake Jo¨ risee (<1 ng
SPMD-1) and barely in those from Hu¨ ttnersee (<1 and ∼1
ng SPMD-1). However, methyl triclosan was detected in the
SPMDs exposed in Zu¨ richsee (25-32 ng SPMD-1), Greifensee
(56 and 85 ng SPMD-1), and the river Limmat (87 and 137
ng SPMD-1). The concentrations in the SPMDs increased in
the order Jo¨ risee ∼ Hu¨ ttnersee < Zu¨ richsee < Greifensee <
Limmat. The highest concentrations were observed in the
SPMDs exposed in the river Limmat, about 15 km below the
lake outflow, 6 km below the WWTP of Zu¨ rich. The
concentrations in these SPMDs were a factor 3-4 higher
than above the WWTP (at the outflow of Zu¨ richsee),
confirming that this WWTP is a major source of methyl
triclosan. For lakes, where fish and SPMDs were analyzed,
the methyl triclosan concentrations in the SPMDs (ng
SPMD-1) increased in the same order and were in the same
range as those in fish (ng g-1, lipid basis), even though the
SPMDs likely had not reached equilibrium during the 3 weeks
of exposure.
FIGURE 2. EI SIM chromatograms (m/z 302) showing elution of
methyl triclosan (retention time, 12.7 min) and the two 13C -tetra-
12
PCBs (internal standards; retention time, 9.4 and 9.7 min) in fish
from (a) Ha1bberstja1rnen, the remote lake in Sweden (lake trout),
(b) Thunersee (white fish), (c) Zu1richsee (white fish), and (d)
Greifensee (white fish). Concentrations in ng g-1 lipid basis; peak
intensities are normalized to the largest peak in each chromatogram.
Note the increasing peak area ratios methyl triclosan/internal
standard (13C12-PCB#47) in chromatograms a-d.
FIGURE 3. Partial EI mass spectrum (m/z 200-400) of the component
observed in fish from Greifensee, identified as methyl triclosan.
Note interference from a coeluting penta-PCB (signals at m/z 324
and 254 for molecular and fragment ions).
reference standard of methyl triclosan and on mass spec-
trometric evidence (full-scan and MS/ MS data). In Figure 3,
we show an EI mass spectrum of this component in one of
the fish samples (white fish, Greifensee). The mass spectrum
clearly shows signals at m/ z 302 (Cl3) and 252, corresponding
to the molecular ion (M+•) and the (M - CH3Cl)+• “dioxin-
like” fragment ion of methyl triclosan, respectively. There
are, however, additional ions from coeluting interfering
compounds such as from a penta-PCB, most likely from
The concentration in a SPMD (CSPMD, ng SPMD-1) is a
function of the mean concentration of the dissolved com-
pound in water (Cwater, ng L-1), the sampling rate (RS, L day-1
SPMD-1), and the time of exposure (t, day) according to eq
1 (adapted from refs 23 and 24)
+•
PCB#99 (M+• at m/ z 324 and (M - Cl2) at m/ z 254).
For further confirmation, some fish samples were re-
analyzed using MS/ MS. MS/ MS previously indicated several
ion transitions of high signal abundance for methyl triclosan.
Of particular importance is the ion transition M+• f (M -
CH3Cl)+• (302+ f 252+, and its isotopic transitions). The loss
of CH3Cl is characteristic for methyl ethers of ortho-phenoxy
phenols that have a further Cl substituent in the ortho′-
CSPMD ) CwaterRSt
(1)
whereby the sampling rate RS itself is a function of parameters
such as temperature, phase transfer, etc. Reported sampling
9
VOL. 38, NO. 2, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 3

rates for various lipophilic pollutants are in the range of about
1-8 L day-1 SPMD-1 (25, 26). For this study, we assumed
relatively low sampling rates of 3-4 L day-1 because in most
lakes some biofouling was observed. Assuming sampling
volumes of ∼70 L for the 18-21 day exposures of the SPMDs
in Hu¨ ttnersee, Zu¨ richsee, and Greifensee and ∼100 L in
Jo¨ risee (41 days exposure and less biofouling but lower
temperature), the methyl triclosan concentrations in water
were estimated at <0.02 ng L-1 in Jo¨ risee and Hu¨ ttnersee, at
0.4-0.5 ng L-1 in Zu¨ richsee, and at 0.8-1.2 ng L-1 in
Greifensee. The concentrations estimated for Greifensee are
in good agreement with those actually measured and reported
in an earlier study (0.8 ng L-1, ref 9). SPMD sampling seems
to be a reliable method for the measurement of the compound
in surface water.
Bioconcentration of Methyl Triclosan in Fish. The lipid-
based concentrations of methyl triclosan observed in fish
are considerably higher than the concentrations in lake water,
suggesting significant bioaccumulation of the compound.
For aquatic organisms, the potential uptake mechanisms of
lipophilic contaminants are (i) direct uptake from water
through exposed surfaces, mainly gills (passive uptake,
bioconcentration), and (ii) uptake through the consumption
of food (biomagnification, ref 22). According to various
authors (27-29), bioconcentration in fish is of much higher
relevance than biomagnification for compounds with log
FIGURE 4. Methyl triclosan concentrations in fish (ng g-1, lipid
basis) from various Swiss lakes and the remote Swedish lake plotted
vs population in the watershed per throughflow of water (P/Q,
persons m-3 d). The correlation coefficient r2 was calculated from
the averaged concentrations of all fish from a lake, but values for
each fish sample are depicted: Ha1 ) Ha1bberstja1rnen, Hu1 )
Hu1ttnersee, Th ) Thunersee, Pf ) Pfa1ffikersee, Gr ) Greifensee,
Zu1 ) Zu1richsee. Note the partly expanded scale.
K
OW e 5, because only a minor fraction of such compounds
τ is the mean water residence time (day), and kw ) 1/ τ is the
flushing rate (day-1).
is sorbed to organic matter (<10% as estimated in ref 9).
Therefore, it is most likely that bioconcentration rather than
biomagnification is the main process responsible for the
presence of methyl triclosan in fish. The lipid-based bio-
concentration factor BCFL (L kg-1) can thus be defined as
Assuming the proportionality factor f is equal for the lakes
studied, the methyl triclosan concentrations are expectedly
increasing in the order Jo¨ risee ∼ Hu¨ ttnersee < Thunersee <
Zu¨ richsee < Pfa¨ffikersee < Greifensee (based on the P/ Q
ratios listed in Table 1). A linear relationship of the con-
centrations in lake water vs P/ Q was previously observed for
caffeine (31) and polycyclic musk compounds (32), which
were used as markers for domestic contamination of surface
water.
BCFL) Cfish/ Cwater
(2)
where Cfish is the steady-state concentration in fish (lipid
basis, ng kg-1) and Cwater is the time-averaged dissolved
concentration in water (ng L-1). Steady-state conditions were
assumed for methyl triclosan in our study, because all fish
were of similar age (g2 years) and concentrations of methyl
triclosan in a lake can be considered as relatively time
independent. Therefore, eq 2 was used for a rough estimation
of BCFL values.
Assuming bioconcentration as the main uptake process
of a compound by fish, the expected lipid-based steady-
state concentration in fish (Cfish) is then
Cfish ) BCFLCwater ) BCFLf P/ Q
(4)
The lipid-based concentrations in fish exceeded the
SPMD-derived concentrations in the water of the respective
lakes by a factor of 1-2.6 × 105 (log BCFL) 5.0-5.4, averaged
data for fish and water from Greifensee and Zu¨ richsee). These
bioconcentration factors are consistent with the lipophilic
properties of methyl triclosan (log Kow ) 5.0, ref 30) and fit
into the linear correlation typically observed with persistent
organic compounds between log BCFL and log Kow (for an
overview, see refs 22 and 27). Assuming an average fat content
in fish of 2%, the bioconcentration factor for methyl triclosan
on a wet weight basis (BCFW) was estimated to be 2000-5200
L kg-1 (log BCFW ) 3.3-3.7).
and under the assumption that f and BCFL are constant, the
concentrations in fish are proportional to P/ Q. The actual
concentrations of methyl triclosan (lipid basis) in fish from
the Swiss lakes increased in the order Hu¨ ttnersee <Thunersee
< Zu¨ richsee ∼ Pfa¨ffikersee < Greifensee, and the concentra-
tions thus increased with P/ Q. When these concentrations
were plotted versus P/ Q, a reasonable correlation was found
(r2 ) 0.85, calculated from the averaged concentrations of all
fish from a lake), consistent with the assumption that methyl
triclosan inputs are from a source proportional to population
such as from WWTPs (Figure 4). However, the plot also shows
lower concentrations in the fish from Pfa¨ffikersee than
expected from the direct linear relationship. The reason for
the comparatively lower concentrations is not known, but
these fish were the largest analyzed and were of the white
fish species, both factors possibly contributing to lower
concentrations (see above).
In a recent study (33), white fish Coregonus sp. from eight
lakes in Switzerland were analyzed for polybromodiphenyl
ether (PBDE) flame retardants (36-165 ng g-1, lipid basis).
A correlation between these concentrations and the ratio
surface area/ volume of the lake (A/ V; r2 ) 0.70) was
interpreted as an indication for atmospheric input of these
compounds to the lakes. When we plotted these PBDE
concentrations versus P/ Q, the correlation decreased (r2 )
0.45). In contrast, the correlation of methyl triclosan con-
centrations in our fish with P/ Q was clearly better than a
Methyl Triclosan as Chem ical Marker for WWTP-Derived
Lipophilic Contam inants in Fish. Provided that there is no
other elimination from the lake than flushing, the concen-
tration (Cwater) of a persistent chemical emitted with waste-
water to a lake is expected to be proportional to the population
in the watershed (P, persons) and inversely proportional to
the throughflow of water (Q, m3 day-1
)
Cwater ) f P/ Q
(3)
whereby f is a proportionality factor defined by the actual
amount of parent compound (triclosan) used in the water-
shed (g person-1 day-1) and its proportion converted to
methyl triclosan and transferred to the lake. For a lake, Q is
defined as Q ) V/ τ ) Vkw where V is the lake volume (m3),
9
3 9 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 2, 2004

correlation with A/ V (r2 ) 0.67), suggesting different input
pathways for methyl triclosan and PBDE into these lakes.
As stated above, eqs 3 and 4 are valid for a chemical with
no other elimination from the lake than flushing. Presently,
little is known about the environmental fate of methyl
triclosan. However, the compound is expected to be relatively
stable toward photolysis under environmental conditions
(9). Furthermore, sorption to particulate organic carbon
(POC) and subsequent sedimentation, estimated as described
in ref 9, did not significantly affect the observed correlation
of concentrations and P/ Q (according to eq 4).
The correlation of the methyl triclosan concentrations in
fish with P/ Q was reasonable despite the fact that fish of
different species, age, and size were included. This dem-
onstrates the usefulness of the compound as a chemical
marker for lipophilic WWTP-derived contaminants in the
lakes studied.
(13) Miyazaki, T.; Yamagishi, T.; Matsumoto, M. Bull. Environ.
Contam. Toxicol. 1984, 32, 227-232.
(14) Vogel, A. Vogel’s Textbook of Practical Organic Chemistry, 4th
ed.; Longman: London, U.K., 1978.
(15) Liechti, P. Zustand der Seen in der Schweiz; Bundesamt fu¨ r
Umwelt, Wald und Landschaft (BUWAL): Bern, Switzerland,
1994.
(16) Kupper, U.; Koch, M.; Meier, W.; Niederhauser, P. Ober-
fla¨chengewa¨sser und Abwasserreinigungsanlagen; Office for
Waste, Water, Energy, and Air of the Canton of Zurich (AWEL):
Zurich, Switzerland, 1998.
(17) Zook, D. R.; Buser, H. R.; Bergqvist, P. A.; Rappe, C.; Olsson, M.
Ambio. 1992, 21, 557-560.
(18) Buser, H. R.; Zook, D. R.; Rappe, C. Anal. Chem. 1992, 64, 1176-
1183.
(19) Bergqvist, P. A.; Strandberg, B.; Ekelund, R.; Rappe, C.; Granmo,
A. Environ. Sci. Technol. 1998, 32, 3887-3892.
(20) Buser, H. R.; Mu¨ ller, M. D. Environ. Sci. Technol. 1997, 31, 1960-
1967.
(21) Tulp, M. Th. M.; Olie, K.; Hutzinger, O. Biomed. Mass Spectrom.
1977, 4, 310.
(22) Geyer, H. J.; Rimkus, G. G.; Scheunert, I.; Kaune, A.; Kettrup, A.;
Zeeman, M.; Muir, D. C. G.; Hansen, L. G.; and Mackay, D. In
Bioaccumulation-New Aspects and Developments; Beek, B., Ed.;
Springer: Berlin, Germany, 2000; pp 1-166.
(23) Booij, K.; Sleiderink, H. M.; Smedes, F. Environ. Toxicol. Chem.
1998, 17, 1236-1245.
(24) Petty, J. D.; Orazio, C. E.; Huckins, J. N.; Gale, R. W.; Lebo, J. A.;
Meadows, J. C.; Echols, K. R.; Cranor, W. L. J. Chromatogr., A
2000, 879, 83-95.
(25) Huckins, J. N.; Petty, J. D.; Lebo, J. A.; Orazio, C. E.; Prest, H. F.;
Tillitt, D. E.; Ellis, G. S.; Johnson, B. T.; Manuweera, G. K. In
Techniques in Aquatic Toxicology; Ostrander, G. K., Ed.; Lewis
Publishers: Boca Raton, FL, 1996; pp 625-655.
(26) Huckins, J. N.; Petty, J. D.; Orazio, C. E.; Lebo, J. A.; Clark, R. C.;
Gibson, V. L.; Gala, W. R.; Echols, K. R. Environ. Sci. Technol.
1999, 33, 3918-3923.
(27) Connell, D. W. In Reviews of Environmental Contamination and
Toxicology; Ware, G. W., Ed.; Springer: New York, 1988; Vol.
102, pp 117-160.
Acknowledgments
We thank R. Ruf, Au, E. Zollinger, Riedikon, U. Ochsenbein,
Canton of Berne, H. Vogt, Fishery Authority Pfa¨ffikon, J. Suter,
Turbenthal, E. Hauser and G. Scott, Wa¨denswil, for the various
fish. We also thank the personnel of the Water Police
(Seepolizei) of the City of Zurich, Electrical works of the
Canton of Zurich (EKZ), Dietikon, and the Office for Waste,
Water, Energy, and Air, Canton of Zurich (AWEL), Zurich, for
assisting in the deployment of the SPMDs and I. J. Buerge
for critical review of the manuscript. We acknowledge the
continued support of the Swiss Federal Agency for Environ-
ment, Forests and Landscape (BUWAL), Berne.
Literature Cited
(1) Adolfsson-Erici, M.; Pettersson, M.; Parkkonen, J.; Sturve, J.
Organohalogen Compd. 2000, 45, 83-86.
(2) Offhaus, K.; Klingl, H.; Scherb, K.; Wachs, B. In Mu¨ nchner Beitra¨ge
zur Abwasser-, Fischerei- und Flussbiologie 30, “Schadstoffe im
Oberfla¨chenwasser und Abwasser”; Bayerische Landesanstalt fu¨ r
Wasserforschung: Mu¨ nchen; R. Oldenburg Verlag: Mu¨ nchen,
1978.
(28) Connell, D. W. In Bioaccumulation of Xenobiotic Compounds;
Connell, D. W., Ed.; CRC Press: Boca Raton, FL, 1990; pp 59-
73.
(29) ECETOC. The Role of Bioaccumulation in Environmental Risk
Assessment: The Aquatic Environment and Related Food Webs;
European Centre for Ecotoxicology and Toxicology of Chemicals
(ECETOC): Brussels, Belgium, 1996.
(30) Valo, R.; Salkinoja-Salonen, M. J. Gen. Appl. Microbiol. 1986,
32, 505-517.
(31) Buerge, I. J.; Poiger T.; Mu¨ ller, M. D., Buser, H. R. Environ. Sci.
Technol. 2003, 37, 691-700.
(32) Buerge, I. J.; Buser, H. R.; Mu¨ ller, M. D.; Poiger, T. Environ. Sci.
Technol. 2003, 37, 5636-5644.
(3) Lopez-Avila, V.; Hites, R. A. Environ. Sci. Technol. 1980, 14, 1382-
1390.
(4) Paxeus, N. Water. Res. 1996, 30, 1115-1122.
(5) Van Stee, L. L. P.; Leonards, P. E. G.; Vreuls, R. J. J.; Brinkman,
U. A. T. Analyst 1999, 124, 1547-1552.
(6) Hites, R. A.; Lopez-Avila, V. Anal. Chem. 1979, 51, 1452A-1456A.
(7) Okumura, T.; Nishikawa, Y. Anal. Chim. Acta 1996, 325, 175-
184.
(8) Tixier, C.; Singer, H. P.; Canonica, S.; Mu¨ ller, S. R. Environ. Sci.
Technol. 2002, 36, 3482-3489.
(9) Lindstro¨ m, A.; Buerge, I. J.; Poiger, T.; Bergqvist, P. A.; Mu¨ ller,
M. D.; Buser, H. R. Environ. Sci. Technol. 2002, 36, 2322-2329.
(10) Singer, H.; Mu¨ ller, S.; Tixier, C.; Pillonel, L. Environ. Sci. Technol.
2002, 36, 4998-5004.
(33) Zenegg, M.; Kohler, M.; Gerecke, A. C.; Schmid, P. Chemosphere
2003, 51, 545-553.
(11) McAvoy, D. C; Schatowitz, B.; Jacob, M.; Hauk, A.; Eckhoff, W.
S. Environ. Toxicol. Chem. 2002, 21, 1323-1329.
(12) Orvos, D. R.; Versteeg, D. J.; Inauen, J.; Dapdevielle, M.;
Rothenstein, A.; Cunningham, V. Environ. Toxicol. Chem. 2002,
21, 1338-1349.
Received for review May 23, 2003. Revised manuscript re-
ceived October 1, 2003. Accepted October 2, 2003.
ES030068P
9
VOL. 38, NO. 2, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 5