Determination of triclosan as its pentafluorobenzoyl ester in human plasma and milk using electron capture negative ionization mass spectrometry

Article abstract of DOI:10.1021/ac060666x

A sensitive method for the determination of triclosan in plasma and milk is presented. Following hydrolysis of possible conjugates, triclosan is extracted with n-hexane/acetone, partitioned into alcoholic potassium hydroxide, and converted into its pentafluorobenzoyl ester. After sulfuric acid cleanup, sample extracts are analyzed by gas chromatography/electron capture negative ionization mass spectrometry. The limit of quantification was 0.009 ng/g for a 5-g plasma sample and 0.018 ng/g for a 3-g milk sample. The coefficient of variation for the method was 6%. The method was tested on more than 70 human plasma and milk samples, of which all plasma samples and more than half of the milk samples were above the limit of quantification. The presented method has lowered the limit of quantification for triclosan in human matrixes significantly as compared to previous methods and makes possible the analysis of triclosan in humans under normal exposure conditions.

Full text of DOI:10.1021/ac060666x

Anal. Chem. 2006, 78, 6542-6546
Determination of Triclosan as Its
Pentafluorobenzoyl Ester in Human Plasma and
Milk Using Electron Capture Negative Ionization
Mass Spectrometry
Mats Allmyr,*,†,‡ Michael S. McLachlan,‡ Gunilla Sandborgh-Englund,† and
Margaretha Adolfsson-Erici‡
Department of Applied Environmental Science, Stockholm University, SE-106 91, Stockholm, Sweden, and
Institute of Odontology, Karolinska Institutet, Stockholm, Sweden
A sensitive method for the determination of triclosan in
plasma and milk is presented. Following hydrolysis of
possible conjugates, triclosan is extracted with n-hexane/
acetone, partitioned into alcoholic potassium hydroxide,
and converted into its pentafluorobenzoyl ester. After
Figure 1. Molecular structure of triclosan.
sulfuric acid cleanup, sample extracts are analyzed by gas
chromatography/electron capture negative ionization mass
spectrometry. The limit of quantification was 0.009 ng/g
for a 5-g plasma sample and 0.018 ng/g for a 3-g milk
sample. The coefficient of variation for the method was
6%. The method was tested on more than 70 human
plasma and milk samples, of which all plasma samples
and more than half of the milk samples were above the
limit of quantification. The presented method has lowered
the limit of quantification for triclosan in human matrixes
significantly as compared to previous methods and makes
possible the analysis of triclosan in humans under normal
exposure conditions.
found in human milk and plasma.5,10,11 In humans, triclosan is
metabolized to and excreted as glucuronide and sulfate conju-
gates.8,10 Considering the widespread use of triclosan and the
potential adverse effects from long-term exposure, there is a need
to further study triclosan concentrations in human body fluids.
Few methods for analyzing triclosan in human matrixes are
described in the literature. However, Hoar and Sissons12 presented
a method for analyzing triclosan in plasma, urine, fish tissue and
bile with a reporting limit of 10 ng/mL. More recently, Lin13
analyzed triclosan in human plasma with a reported limit of
detection (LOD) of 10 ng/mL; Bagley and Lin14 reported a limit
of quantification (LOQ) of 10 ng/mL. Ye et al.15 reported an
LC/MS/MS method for the analysis of triclosan in urine with a
LOD of 2 ng/mL.
Triclosan (2,4,4′-trichloro-2′-hydroxydiphenyl ether, CAS 3380-
34-5) (Figure 1) is a lipophilic and phenolic compound (log KOW
) 4.76; pKa ) 7.9).1,2 Due to its antibacterial properties, triclosan
has found widespread use in a variety of consumer products,
including toothpastes, deodorants, soaps, polymers, and fibers.3-5
Following its discovery in the bile of fish living downstream of
sewage treatment plants,5 triclosan has attracted increasing
interest as a potentially toxic environmental contaminant. The
acute toxicity of triclosan to mammals is low, but in vitro studies
on rat and human material have shown that low concentrations
of triclosan may disturb metabolic systems and hormone
homeostasis.6-9 Triclosan is absorbed by humans and has been
To study triclosan in humans under normal exposure condi-
tions, the high reporting limits of previous methods are insuf-
ficient. As a result, we have worked during the last several years
at improving these methods. Recently, our group analyzed
triclosan in milk and plasma at reporting limits of 0.3 ng/g
(6) Hanioka, N.; Omae, E.; Nishimura, T.; Jinno, H.; Onodera, S.; Yoda, R.; Ando,
M. Chemosphere 1996, 33, 265-276.
(7) Schuur, A. G.; Legger, F. F.; van Meeteren, M. E.; Moonen, M. J. H.; van
Leeuwen-Bol, I.; Bergman, A.; Visser, T. J.; Brouwer, A. Chem. Res. Toxicol.
1998, 11, 1075-1081.
(8) Wang, L. Q.; Falany, C. N.; James, M. O. Drug Metab. Dispos. 2004, 32,
1162-1169.
(9) Jacobs, M. N.; Nolan, G. T.; Hood, S. R. Toxicol. Appl. Pharmacol. 2005,
209, 123-133.
(10) Sandborgh Englund, G.; Adolfsson-Erici, M.; Odham, G.; Ekstrand, J. J.
* Corresponding author. E-mail: mats.allmyr@itm.su.se.
† Karolinska Institutet.
‡ Stockholm University.
Toxicol. Environ. Health 2006, in press.
(1) Syracuse Research Corporation, PhysDrop Database Demo, http://www.
syrres.com/esc/physdemo.htm (accessed 20 March 2006).
(2) Merck Index 12:3, 2000.
(11) Hovander, L.; Malmberg, T.; Athanasiadou, M.; Athanassiadis, I.; Rahm, S.;
Bergman, A.; Klasson Wehler, E. Arch. Environ. Contam. Toxicol. 2002,
42, 105-117.
(3) Bhargava, H. N.; Leonard, P. A. Am. J. Infect. Control. 1996, 24, 209-218.
(4) Perencevich, E. N.; Wong M. T.; Harris, A. D. Am. J. Infect. Control 2001,
28, 281-283.
(5) Adolfsson-Erici, M.; Pettersson, M.; Parkkonen, J.; Sturve, J. Chemosphere
2002, 46, 1485-1489.
(12) Hoar, D. R.; Sissons, D. J. Methodol. Dev. Biochem. 1976, 5, 221-226.
(13) Lin, Y. J. Am. J. Dent. 2000, 13, 215-217.
(14) Bagley, D. M.; Lin, Y. J. Am. J. Dent. 2000, 13, 148-152.
(15) Ye, X.; Kuklenyik. Z.; Needham. L. L.; Calafat. A. M. Anal. Chem. 2005,
77, 5407-5413
6542 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006
10.1021/ac060666x CCC: $33.50 © 2006 American Chemical Society
Published on Web 07/22/2006

(personal communication for fresh weight basis) and 0.1 ng/g,
respectively.5,10 In this paper, we give a detailed description of
the analytical methodology, which includes more recent improve-
ments that have further lowered the LOQ.
n-hexane/acetone (9:1, v/v, 2 mL). The combined organic phases
were adjusted to 3 mL before derivatization.
Derivatization. All samples and calibration standards were
treated as follows: Triclosan was converted to its pentafluoroben-
zoyl ester by adding H2O (Milli-Q, 2 mL), KOH (2 M, 50 µL),
pentafluorobenzoyl chloride (10% in toluene, 10 µL) and NaCl (0.3
g; more NaCl was added if emulsion formed upon mixing) to the
extract and shaking the test tube vigorously for 2 min. The extract
was transferred to another test tube, and the aqueous phase was
reextracted with n-hexane (2 mL). H2SO4 (98%, 3 mL) was added
to the extract, and the tube was inverted 60 times. The extract
was transferred to another test tube, and the H2SO4 layer was
reextracted with n-hexane (2 mL). The final extract was reduced
to 2 mL under a gentle flow of nitrogen gas at room temperature
after addition of volumetric standard (CB 106). Approximately 0.5
mL of the extract was transferred to a GC/MS vial and analyzed
as described below.
Triclosan was also converted to its methyl ether and acetate
ester for comparison purposes. The methyl ether was formed by
the reaction of triclosan in n-hexane with diazomethane for 3 h at
4 °C in darkness. Excess reagent and ether were removed under
a gentle flow of nitrogen gas at 30 °C. The acetate ester was
formed by reaction of triclosan in n-hexane with acetic anhydride/
pyridine (1:1) for 30 min at 60 °C. Excess reagent was washed
out from the organic phase with H2O (5 mL).
Instrumentation. For gas chromatography/electron capture
(GC/ECD) determinations, a Varian Star (Varian, Walnut Creek,
CA) equipped with a spilt/splitless injector and a DB5-MS capillary
column (15 m; i.d 0.25 mm; 0.10-µm film thickness; J&W, USA)
coupled to a 63Ni-ECD detector was used. The temperature of the
split/splitless injector was 275 °C, and the detector temperature
was 320 °C. The column temperature was held at 90 °C for 1
min, ramped 20 °C min-1 to 300 °C, and held for 10 min. Helium
was used as the carrier gas (1 mL/min), and argon-methane (20
mL/min), as the makeup gas.
For gas chromatography/mass spectrometry/electron capture
negative ionization (GC/ECNI/MS), a HP5890A gas chromato-
graph connected to a Finnigan SSQ 7000 quadropole mass spec-
trometer was used. The extract (1 µL) was injected in splitless
mode on a DB5-MS capillary column (15 m; i.d 0.25 mm; 0.10-µm
film thickness; J&W, USA). The temperature of the split/split-
less injector was 280 °C, and the transfer line temperature was
300 °C. The column temperature was held at 90 °C for 2 min,
ramped 20 °C min-1 to 315 °C, and held for 10 min. The ion source
temperature was 180 °C. The electron energy was 70 eV. Helium
was used as the carrier gas (1.4 mL/min), and ammonia, as the
reagent gas (7000 mTorr). The ions monitored for quantification
and identification purposes were m/z 287, 289, and 482 for the
triclosan pentafluorobenzoyl ester and its fragments; m/z 299, 301,
and 494 for the 13C-labeled triclosan pentafluorobenzoyl ester and
its fragments; and m/z 326 and 328 for CB 106. The surrogate
standard was used to correct the measured concentrations.
EXPERIMENTAL SECTION
Chemicals. Solvents and chemicals used in the extraction and
cleanup of the samples: n-hexane (LiChrosolv) (Merck, Darms-
tadt, Germany) was of liquid chromatography grade; acetone
(Suprasolv) (Merck) was of gas chromatography grade; hydro-
chloric acid (HCl, 37%, w/w), sodium chloride (NaCl) (Scharlau,
Barcelona, Spain), sulfuric acid (H2SO4, 98%) (Fisher Scientific,
Loughborough, UK), and potassium hydroxide (KOH) (Eka
Chemicals, Bohus, Sweden) were of pro analysis quality; ethanol
(99.5%) (Kemetyl, Haninge, Sweden); pentafluorobenzoyl chloride
(PFBCl, g 98.5%) (Fluka, UK). Water was purified employing a
Milli-Q Reagent Water System (Millipore Corporation, Billerica,
MA). Universal indicator (Merck) was used for pH measurement.
The gases used were helium (99.997%), nitrogen (99.998%),
argon-methane (90:10, 99.9997%/99.995%) (Air Liquid, Kungsa¨n-
gen, Sweden), and ammonia (99.998%) (AGA Gas AB, Sundbyberg,
13
Sweden). C12-labeled triclosan was purchased from Cambridge
Isotope Laboratories (Andover, MA). Triclosan (Irgasan DP300)
(calibration standard) was a gift from Ciba-Geigy. CB 106
(2,3,3′,4,5-pentachlorobiphenyl) was synthesized by Peter Haglund
(now at Umeå University).
Samples. The analytical method described in the following
was applied to 72 plasma and 71 milk samples from 36 Swedish
women.16 The study was approved by the Regional Ethical Review
Board, Huddinge University Hospital, Sweden (dnr: 395/03).
Informed consent by the mothers was compulsory for their
participation in the study.
Hydrolysis, Extraction, and Cleanup. Test tubes were
machine-washed and heated to 420 °C for 4 h before use. During
analysis, the caps of the test tubes were lined with a double layer
of aluminum. The test tubes were centrifuged for 5 min at 2000
rpm (900 g) after each mixing of aqueous and organic phases to
accelerate phase separation.
The sample (3 g milk, 5 g plasma, or 3 mL H2O as the blank)
was transferred to a 15-mL screw capped test tube and fortified
with 13C-labeled triclosan. To cleave any metabolic conjugates of
triclosan, the sample was hydrolyzed in H2SO4 (9 M when in
sample) on a heating unit (0.5 h, 60 °C). The sample was cooled
to room temperature and extracted twice with n-hexane/acetone
(9:1, v/v, 6 + 4 mL) by inverting the tube 60 times by hand.
Triclosan was partitioned into an aqueous KOH solution (0.5 M,
50% ethanol, 2 mL), which was mixed with the extract on a vortex
mixer, followed by inverting the test tube for 5 min on a rotary
mixer. The organic phase containing neutral compounds was
discarded after centrifugation. The aqueous KOH solution with
the deprotonated triclosan was acidified with HCl (2 M, 1 mL),
and triclosan was extracted twice with n-hexane/acetone (9:1, v/v,
4 + 2 mL) by mixing on a vortex mixer. The extract was cleaned
up by adding H2SO4 (13.7 M, 4 mL) and inverting the test tube
for 5 min on a rotary mixer. The extract was transferred to a
second test tube, and the H2SO4 layer was reextracted with
RESULTS AND DISCUSSION
In the following, triclosan concentrations, if not from spiking,
refer to the sum of triclosan in either unchanged or conjugated
form present in plasma or milk.
Hydrolysis and Extraction. In humans, triclosan is metabo-
(16) Allmyr, M.; Adolfsson-Erici, M.; McLachlan, M. S.; Sandborgh Englund, G.
Sci. Total Environ., submitted.
lized to and excreted as glucuronide and sulfate conjugates.8,10
Analytical Chemistry, Vol. 78, No. 18, September 15, 2006 6543

Figure 2. Molecular structure of the triclosan pentafluorobenzoyl
derivative.
Figure 4. Full scan ECNI mass spectra of triclosan and 13C-labeled
triclosan from a milk sample, m/z 270-500. [M]- ) the molecular
ions for the derivatives. The molecular ion structure of the derivative
is depicted. The zigzag sign in the molecular structure denotes where
the major fragmentation occurs.
Figure 3. Relative response for the different triclosan derivatives
from the GC/ECNI/MS analysis. Square ) pentafluorobenzoyl tri-
closan (relative response factor, k ) 2.3), diamond ) acetyl triclosan
(k ) 0.5), circle ) methyl triclosan (k ) 0.08). Ctric/Cvol and Atric/Avol
) concentration and area ratio between triclosan and the volumetric
standard.
response of the pentafluorobenzoyl and the acetyl derivative
relative to the volumetric standard CB 106 dropped by ∼50 and
13%, respectively, as compared to the first analysis. The response
for the methylated triclosan relative to the volumetric standard
remained unchanged. The results suggest that the slow degrada-
tion of the pentafluorobenzoyl derivative could reduce the method
sensitivity and cause errors in quantification if the derivatized
extracts were stored for longer periods prior to instrumental
analysis. However, if 13C-labeled triclosan is used as the surrogate
standard, the degradation of the native triclosan and the surrogate
standard should occur at the same rate, and the determination of
triclosan should remain unaffected.
Pentafluorobenzoylation was selected as the derivatizing method,
primarily because of the high sensitivity of the derivative on the
instruments tested. In addition, the method is quick and easy,
and it allows an effective cleanup with H2SO4. The use of
hazardous chemicals such as diazomethane is avoided as well.
Method Specificity. The ECNI of the triclosan pentafluo-
robenzoyl ester resulted in the molecular ion m/z 482 and 484
and a major dissociation to the fragment ions m/z 287 and 289.
The corresponding ions for the 13C-labeled triclosan derivative
were m/z 494 and 496 and m/z 299 and 301 for the fragments
(Figure 4). Figure 4 depicts the full scan mass spectrum (m/z
270-500) for a milk sample. Both the full scan spectra of the
triclosan and the 13C-labeled triclosan derivatives show the
characteristic molecular ion cluster patterns for trichlorinated
compounds. Full scan spectra for standards and samples were
used for the identification of triclosan. However, for the quantifica-
tion of triclosan, the fragment ions m/z 287 and 289 for the
triclosan derivative and m/z 299 and 301 for the 13C-labeled
triclosan derivative were used. The absolute and relative abun-
dance of these ions was used for identification and quantification
of triclosan. Figure 5 depicts the GC/ECNI/MS-SIM chromato-
grams for triclosan and 13C-triclosan from a triclosan calibration
standard and from the milk sample with the lowest concentration
above the LOQ (0.018 ng/g) in this study.
Hence, both unchanged and conjugated triclosan may be present
simultaneously in the body.10 To determine the total amount of
triclosan in milk and plasma, hydrolysis was conducted to cleave
triclosan from any conjugates that might be present. The hot
sulfuric acid hydrolysis method for triclosan conjugates used in
this work was previously described by Hoar and Sissons.12 Initially
a vortex mixer was used for the first extraction step for both milk
and plasma. However, in some cases, problematic emulsions were
formed, especially in plasma. This was avoided by instead inverting
the test tube 60 times by hand.
Derivatives. Derivatization was used for enhancing the
chromatographic and spectrometric properties of triclosan. We
compared the pentafluorobenzoylated triclosan (Figure 2) with
the acetylated and methylated triclosan with respect to its
sensitivity on GC/ECD and GC/ECNI/MS. Applying GC/ECD,
the relative response factor (relative to the volumetric standard)
of the pentafluorobenzoylate was 4 and 7 times higher than the
acetylate and the methylate, respectively. On GC/ECNI/MS, the
relative response factor of the pentafluorobenzoylate was 5 and
28 times higher than the acetylate and methylate, respectively
(Figure 3). This suggests that pentafluorobenzoylation is a suitable
method for enhancing the chromatographic and, in particular, the
spectrometric properties of triclosan.
The use of concentrated sulfuric acid to destroy matrix
components of biological samples is a simple and effective method
often employed in environmental analysis. The different deriva-
tives were tested regarding their stability toward H2SO4 treat-
ment (>98%, 3 mL, inverting the tube 60 times by hand). The
results showed that acetylated triclosan was degraded completely
in H2SO4, but both the methylate and pentafluorobenzoylate
showed satisfactory stability (results not shown).
Stability of the derivatives over time was also compared by
storing a set of derivatized extracts in a refrigerator and repeatedly
analyzing them on GC/ECD. Over the course of 1 month, the
Blanks. Blanks were analyzed together with each set of 16
samples. All blanks contained triclosan at levels well above the
LOD (see definition in next section). Initially, the levels were
6544 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

Figure 6. Triclosan concentrations (ng/g of fresh weight) in human
blood plasma and milk. Cross ) plasma, circle ) milk, solid lines )
LOQ in plasma and milk, respectively, in the present study; dashed
lines ) reporting limits for previous studies: (A) ) Hoar and Sissons
(1976); Lin (2000); Bagley and Lin (2000), (B) ) Adolfsson-Erici et
al. (2002), (C) ) Sandborgh et al. (in press), (D) ) present study.
Figure 5. GC/ECNI/MS-SIM chromatograms showing triclosan
(upper chromatograms, fragment ions m/z 287 and 289) and 13C-
triclosan (lower chromatograms: fragment ions m/z 299 and 301) from
(A) a triclosan calibration standard and (B) the milk sample with the
lowest triclosan concentration above the LOQ in this study.
repeatability (recovery 42% (CV 22%)); however, the repeatability
was still very good (see below).
Method Repeatability. The method repeatability test was
assessed by repeatedly analyzing triclosan in three different milk
samples: one with a high, one with an intermediate, and one with
a low triclosan concentration. The replicates were not analyzed
together; rather, they were analyzed on different occasions with
freezing and thawing of the sample between analyses. The higher
concentration was among the highest concentrations in milk in
this study, and the lower concentration was near the LOQ for milk.
The coefficient of variation was 6% for the high concentration
(mean ) 0.84 ng/g, n ) 7), 1% for the intermediate concentration
(mean ) 0.31 ng/g, n ) 3), and 5% for the low concentration
(mean ) 0.020 ng/g, n ) 3). The instrumental precision was
evaluated by analyzing two samples three times distributed over
a sequence of 184 injections on the GC/ECNI/MS. In both cases,
the instrumental coefficient of variation was 1%.
unacceptably high. To identify the source, the Teflon-coated test
tube caps used in the procedure were rinsed with solvent, which
was then analyzed. Elevated levels of triclosan were found, even
though the caps had been machine-washed with alkaline detergent
after previous use. The use of aluminum foil to line the caps
significantly reduced the method blank. The blanks were also
consistent, the coefficient of variation of (Atriclosan/A C-triclosan) in
the blanks amounting to 5% (n ) 5) and 14% (n ) 5) in milk and
plasma analysis, respectively.
Limit of Quantification (LOQ) and Limit of Detection
(LOD). The limit of detection was defined as a signal height of
three times the standard deviation of the chromatographic noise.
The LOQ was defined by the triclosan content in the blanks. As
noted above, the variability in the blanks was low. Consequently,
the LOQ was defined as 4 times the blank level. This corresponded
to 0.06 ng per sample, or 0.018 ng/g of milk and 0.009 ng/g of
plasma for ∼3 g of milk and 5 g of plasma.
13
In addition to the repeated determination of triclosan in milk,
10 g of milk with only trace amounts of triclosan was fortified
with triclosan to a nominal concentration of 7.4 ng/g. The
measured triclosan concentration in that sample was 7.1 ng/g
(mean, n ) 3). These results show that the repeatability of the
method was good and that the reliability was satisfying.
Method Recovery. The relative recovery of native triclosan
compared with the 13C-labeled triclosan was assessed by spiking
five aliquots of a milk sample to nominal triclosan concentrations
of 0.6 ng/g and an additional five to a concentration of 4.6 ng/g
of milk. 13C-labeled triclosan was then added, and the samples
were worked up and analyzed as described above. The relative
recovery for both levels of triclosan was 95%.
Plasma (n ) 72) and milk (n ) 71) samples were fortified with
13C-labeled triclosan and analyzed as described. The mean absolute
recovery of the 13C-labeled triclosan was 46% (CV ) 23%) in plasma
(85% in blanks) and 49% (CV ) 18%) in milk (91% in blanks). The
main loss of triclosan is believed to occur in the derivatization
step, in which varying amounts of precipitate formed in the sample
extract, possibly hindering the derivatization of the triclosan. This
precipitate was not seen in the blanks, which were treated in the
same way as the samples, and for which the recovery was
considerably higher. It would probably enhance the recovery if
an additional cleanup procedure were introduced before the
derivatization step. Variability in recovery of the surrogate
standard was also observed in seven replicates analyzed for
Method Applications. The present work was aimed at
developing a method for the determination of triclosan, possible
metabolic conjugates in plasma and milk, or both; however, the
method or parts from it may also be applicable to other matrixes.
Provided that the extract is clean enough, the derivatization of
triclosan to its pentafluorobenzoylate should be suitable for any
analysis of triclosan when using GC/ECNI/MS or GC/ECD;
however, the use of a GC/ECD for determination excludes the
possibility of using 13C-labeled triclosan as the surrogate standard.
More than 70 plasma and milk samples were analyzed for
triclosan with this method. Figure 6 depicts the results. Triclosan
was present in the plasma samples at concentrations ranging over
almost 4 orders of magnitude. All of the concentrations were above
the LOQ. In milk, triclosan was found in detectable amounts in
all samples, but in 42% of them, the concentration was below the
LOQ. Overall, the concentrations in milk were lower than in
plasma, but also the sample size was smaller for milk (3 g) than
it was for plasma (5 g). If a method would be required for
quantifying triclosan in milk from the vast majority of women, it
Analytical Chemistry, Vol. 78, No. 18, September 15, 2006 6545

would be necessary to increase the sample size or reduce the
method blank.
The development work resulting in the method presented in
this paper has lowered the LOQ for triclosan in human body fluids
by 3 orders of magnitude. This opens the possibility to study the
level of triclosan exposure in the general population and to better
explore the sources of that exposure and the behavior of triclosan
in the body.
The reporting limits from published methods are also shown
in Figure 6. If the methods reported by other groups12-14 (line A)
had been used to analyze these samples, all of the milk samples
and 85% of the plasma samples would have been below the LOQ
or LOD. The earlier methods developed in our group5,10 would
have performed better but would still not have allowed for the
quantification of triclosan in most of the milk samples and ∼45%
of the plasma samples.
The increased sensitivity of the present method was primarily
due to the use of pentafluorobenzoylation as a means to enhance
the spectrometric properties of triclosan. Furthermore, the pen-
tafluorobenzoylation of triclosan allows an effective final cleanup
of the extract with concentrated sulfuric acid, which reduces the
interferences in the GC/MS analysis. A significant prerequisite
for the method improvement was the reduction of triclosan in the
blanks, which, in fact, defined the LOQ.
ACKNOWLEDGMENT
The project was supported by the Swedish Research Council
for Environment, Agricultural Sciences and Spatial Planning (Nos.
2002-0706 and 2004-1187), the Swedish Patent Revenue Fund for
Research in Preventive Dentistry, and grants from Karolinska
Institutet.
Received for review April 10, 2006. Accepted June 6,
2006.
AC060666X
6546 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006