Available online at www.sciencedirect.

com

Journal of Chromatography A, 1188 (2008) 132139

Simplified sample preparation method for triclosan and methyltriclosan

determination in biota and foodstuff samples
P. Canosa, I. Rodrguez , E. Rub, M. Ramil, R. Cela
Departamento de Qumica Analtica, Nutricion y Bromatologa, Instituto de Investigacion y Analisis Alimentario,
Universidad de Santiago de Compostela, Avda de las Ciencias s/n., Santiago de Compostela 15782, Spain
Received 3 January 2008; received in revised form 15 February 2008; accepted 19 February 2008
Available online 7 March 2008

Abstract
An improved method for the determination of triclosan (TCS) and methyltriclosan (MTCS) in fish and foodstuff samples is presented. Analytes
were simultaneously extracted and purified using the matrix solid-phase dispersion (MSPD) technique, and then selectively determined by gas
chromatography with tandem mass spectrometry (GCMS/MS). Several combinations of dispersants, clean-up co-sorbents and extraction solvents
were tested in order to obtain lipid-free extracts and quantitative recoveries for TCS and MTCS. Under optimised conditions, 0.5 g samples were
dispersed using 1.5 g of neutral silica in a mortar with a pestle, and transferred to a polypropylene cartridge containing 3 g of silica impregnated
with 10% of sulphuric acid (SiO2 H2 SO4 , 10%, w/w). Analytes were recovered with 10 mL of dichloromethane whereas lipids were oxidized in
the layer of acidic silica. The extract was concentrated to dryness and re-constituted with 1 mL of ethyl acetate. Then, a fraction of 0.5 mL was
mixed with 50 L of N-methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) and injected in the GCMS/MS system. The developed
method provided absolute recoveries between 77 and 120% for different samples spiked at the low ng g1 level, quantification limits in the range of
12 ng g1 and a considerable simplicity in comparison with previously developed sample preparation approaches. Experiments carried out placing
sliced food samples in direct contact with TCS-treated kitchenware surfaces showed the capability of the biocide to migrate into foodstuffs.
2008 Elsevier B.V. All rights reserved.
Keywords: Triclosan; Biota samples; Matrix solid-phase dispersion; Gas chromatography tandem mass spectrometry

1. Introduction
Triclosan [5-chloro-2-(2,4-dichlorophenoxy)phenol], TCS,
is worldwide employed as a broad-spectrum antibacterial agent.
Personal care products (e.g. deodorants, bath gels and especially
tooth paste), plastic toys, kitchenware and textiles may contain
significant amounts (up to 0.3%) of TCS [14]. Nowadays, it is
also being considered as a possible additive in food packaging
polymers and surfaces in contact with foodstuff in the processing
industry [57]. Another potential application consists of bonding
this biocide to siloxane polymers to prepare marine antifouling
coatings [8].
As a result of the above uses, TCS has become an ubiquitous substance in wastewater. Moreover, around half of the TCS
entering wastewater treatment plants (WWTPs) remains sorbed
on sludge particles [9], wherein it can reach concentrations in the

Corresponding author. Tel.: +34 981 563100x14387; fax: +34 981 595012.
E-mail address: qnisaac@usc.es (I. Rodrguez).

0021-9673/$ see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2008.02.064

g g1 range [10,11]. It is also known that, once in the aquatic

environment, TCS undergoes different transformation reactions
through oxidation, photochemical and methylation processes.
The latter, which is probably the most important one for TCS
in WWTPs, leads to the formation of methyltriclosan (MTCS)
[12]. Both, TCS and MTCS are lipophilic species (log Kow 4.78
and 5.4, respectively) of which presence has also been confirmed
in sediments from lakes and rivers receiving treated wastewater
[13,14].
Regarding biota samples, TCS and MTCS have been detected
in edible parts [12,15] and fluids from river and lake fishes
[1618]; moreover, they are bio-accumulated, up to 1000 times,
by algae [19]. TCS has also been identified in breast milk, as a
probable result of topic and oral uses of personal care products
containing this bactericide [18,20,21]. Information regarding the
potential effects of TCS and MTCS on wildlife and human
beings is even scarcer than occurrence data; however, some
recent works pointed that TCS is able to disturb the mechanism and levels of thyroidal hormones in amphibians and rats
[22,23].

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

From an analytical perspective, most of the approaches developed for biota and food samples analysis consume large volumes
of organic solvents to fractionate TCS from lipids and other coextracted species [12,15,19], they are labour-intensive due to the
use of multi-step clean-up procedures [1618] and/or, in some
cases, they do not allow the simultaneous determination of TCS
and MTCS since they are separated in different fractions on
the basis of their pKa values [17,24]. Finally, some methods,
developed with the aim of investigating the migration of TCS
from alimentary packaging films to foodstuff, lack of enough
sensitivity for the analysis of biota samples [6].
The goal of this work was to optimise a simplified sample
preparation strategy, which does not require the acquisition of
dedicated instrumentation, for the extraction of TCS and MTCS
from biota samples, and compatible with their further determination by gas chromatography. Matrix solid-phase dispersion
(MSPD) was selected for the simultaneous extraction and
purification of target species. In this technique, samples are first
blended and dispersed around the particles of a suitable sorbent,
in a mortar with a pestle, and then transferred to a solid-phase
extraction (SPE) cartridge, which might contain also a clean-up
co-sorbent. The efficiency and selectivity of the extraction
can be tuned by appropriate selection of washing and elution
solvents as well as the dispersant and co-sorbent materials. In
the case of biota samples, quantitative recoveries have been
reported for different groups of organic compounds with moderate volumes of organic solvents; moreover, in many cases MSPD
produce a direct analysable extract using chromatographic or
electro-driven separation techniques [25]. In this study, different
combinations of dispersants, clean-up co-sorbents and elution
solvents were tested in order to: (1) minimise the presence
of lipids in the sample extract and (2) achieve quantitative
recoveries for target analytes. After extraction, TCS and MTCS
were determined by gas chromatography combined with
tandem mass spectrometry (GCMS/MS). The performance
of the developed method was evaluated using freeze-dried fish
tissues and foodstuffs with different fat contents.
2. Experimental
2.1. Standards and material
Trace analysis quality solvents: methanol, n-hexane,
dichloromethane (CH2 Cl2 ), ethyl acetate, acetonitrile and chloroform (CHCl3 ), as well as anhydrous sodium sulphate and
concentrated sulphuric acid (98%) were supplied by Merck
(Darmstadt, Germany). TCS and the derivatization reagent
N-methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) were purchased from Aldrich (Milwaukee, WI, USA).
MTCS was acquired from Toronto Research Chemicals
(Toronto, Canada). Individual solutions of TCS and MTCS were
prepared in methanol. Further dilutions and mixtures of them
were made in n-hexane, ethyl acetate and acetonitrile. Derivatization was carried out at room temperature, in 1.5 mL volume
vessels, adding 50 L of MTBSTFA to standard solutions or
sample extracts (0.5 mL volume) in ethyl acetate or acetonitrile.
After 5 min of manual shaking, the mixture (1 L) was injected

133

in the gas chromatograph. TCS was converted into the corresponding tert-butyldimethylsilyl derivative, whereas MTCS
remained unaffected [26].
Florisil (60100 mesh), alumina (150 mesh), silica (230
400 mesh) and C18 (70230 mesh) were acquired from Aldrich
and Merck. The normal-phase materials were activated at 130 C
for 48 h and then allowed to cool down in a desiccator. C18
was used directly as received. Silica impregnated with sulphuric acid (SiO2 H2 SO4 ) at three different concentrations: 5,
10 and 20% (w/w), was prepared by mixing the activated sorbent
with the concentrated acid. After homogenisation, acidic silica
was stored in closed amber vessels. Solid-phase extraction cartridges containing 0.5 g of the primarysecondary amine (PSA)
sorbent were acquired from Supelco (Bellefonte, PA, USA).
Empty polypropylene cartridges (15 mL capacity) and 20 m
polyethylene frits were purchased from International Sorbent
Technology (Mid Glamorgan, UK). Syringe filters (Millex GV,
13 mm, 0.22 m) were obtained from Millipore (Billerica, MA,
USA).
2.2. Samples and sample preparation
Elaborated (cheese and boiled ham) and raw (trout and
salmon) food samples were acquired from local supermarkets.
Specimens of rainbow trout (Oncorhynchus mykiss), mackerel (Scomberomorus cavalla) and mussels (Mytilus edulis)
were fished in small rivers and an estuary in the Northwest of
Spain. Optimisation of extraction conditions was carried out
with spiked samples of elaborated foodstuff and freeze-dried
fish muscle (edible parts). Their fat content was determined as
described elsewhere [27].
Spiked samples were prepared by mixing a weighed amount
of each matrix with a standard solution of the analytes in nhexane (about 1 mL of solvent was added per g of sample) in
glass vessels, ca. 20 mL volume. The mixture was thoroughly
homogenised, allowed to stand in a hood overnight and then
vessels closed and stored, under different conditions, for several
days before extraction.
In order to investigate the possible contamination of foodstuffs with TCS, due to the use of kitchenware or surfaces treated
with this bactericide, sliced food samples (cheese and boiled
ham) were placed over a kitchen cutting board with antibacterial
treatment and covered with aluminium foil. After a given time
(between 5 min and 7 h) a fraction of each sample was removed
and processed using optimised conditions.
For extraction, samples (0.5 g) were mixed with 2 g of sodium
sulphate, dispersed in a mortar with a pestle, using 1.52 g
of different materials, and then transferred to a SPE cartridge
containing a layer of co-sorbent at the bottom. Under final working conditions, 1.5 of neutral silica and 3 g of SiO2 H2 SO4
(10%, w/w) were selected as dispersant and co-sorbent, respectively. Analytes were eluted with 10 mL of dichloromethane.
The extract was evaporated to dryness, using a gentle stream of
nitrogen at room temperature, re-dissolved with 1 mL of ethyl
acetate and filtered. An aliquot (0.5 mL) was derivatized using
same conditions as for calibration standards, and injected in the
GCMS/MS system.

134

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

Table 1
MS/MS operation conditions and performance of the GCMS/MS system for MTCS and TCS (as silyl derivative) determination
Compound

Parent ion (m/z)

Product ions (m/z)

Excitation amplitude (V)

Storage level (m/z)

Correlation coefficient (R2 )

LOQs (ng mL1 )

MTCS
TCS

304
347

232, 252, 254

200, 310

1.50
1.50

130
140

0.998
0.999

1
0.5

2.3. Determination
Analytes were determined by GCMS/MS using a Varian
(Walnut Creek, CA, USA) CP 3900 gas chromatograph connected to an ion-trap mass spectrometer (Varian Saturn 2100).
Separations were carried out in a HP-5ms capillary column
(30 m 0.25 mm I.D., df 0.25 m) supplied by Agilent (Wilmington, DE, USA). Helium (99.999%) was used as carrier gas at
a constant flow of 1 mL min1 . The GC oven was programmed
as follows: 50 C (held for 1 min), at 10 C min1 to 270 C
(held for 10 min). The GCMS interface and the ion trap temperatures were set at 270 and 220 C, respectively. Injections
(1 L volume) were made in the splitless mode (splitless time
1 min), with the injector port at 280 C. The mass spectrometer
was operated in the electron impact ionisation mode (70 eV).
MS spectra were recorded in the range from 100 to 550 m/z
units. The most intense ions in the spectra of MTCS, [M + 2]+ ,
and the silylated derivative of TCS, [M + 2-57]+ , were isolated
with a window of 3 m/z units and subjected to collision induced
dissociation.
Recoveries of the proposed sample preparation method were
calculated by external calibration, comparing the responses
obtained for standards and sample extracts. GCMS/MS was
used as quantification technique; moreover, GCMS was also
considered to assess the level of co-extracted interferences under
different experimental conditions. In this case, analytes were
monitored using the sum of signals for the 2 most intense ions:
302 + 304 and 345 + 347 m/z, in the MS spectra of MTCS and
TCS, respectively.

TCS, respectively. Considering an injection volume of 1 L,

these values are similar to the absolute LOQ of 1 pg reported
for MTCS using GC with high-resolution mass spectrometry
(HRMS) detection [16].
3.2. Matrix solid-phase dispersion with lipids retention
A major difficulty for the determination of trace levels of
organic compounds in food and biota samples is the fractionation between target species and lipids. The first of the considered
MSPD strategies attempted the extraction of MTCS and TCS
from dispersed samples whereas lipids remained in the MSPD
cartridge. In order to achieve this aim, several combinations
of C18 and normal-phase materials as dispersants and/or cosorbents with acetonitrile as extraction solvent were tested. The
choice of acetonitrile was made on the basis of several factors, such as its efficiency to extract TCS from solid matrices
[28], compatibility with the further silylation of this specie and

3. Results and discussion

3.1. Performance of GCMS/MS detection
Table 1 summarizes the optimal MS/MS detection conditions
as well as some relevant features in relation to the performance of
the determination technique. Two intense transitions appeared
in the MS/MS spectra of MTCS and TCS. For the first compound, they corresponded to the loss of CH3 Cl (304+ 252+ ,
254+ ) and two chlorine atoms (304+ 232+ ), whereas in the
case of TCS they reflected the loss of one atom of chlorine
(347+ 310+ , 312+ ) and the breakdown of the ether bond
between both aromatic rings in the structure of the silylated
specie (347+ 200+ ), Fig. 1. The linearity in the response of
the GCMS/MS system was evaluated with calibration standards (mixtures of MTCS and TCS as silylated derivative) at 10
different concentration levels in the range of 22000 ng mL1 .
Correlation coefficients of 0.998 and 0.999 and instrumental limits of quantification (LOQs) of 1 and 0.5 ng mL1 , defined for a
signal-to-noise ratio (S/N) of 10, were achieved for MTCS and

Fig. 1. Structures and MS/MS spectra for MTCS (A) and TCS (B) as silylated
derivative.

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

Table 2
Percentage of lipids in the extracts from freeze-dried salmon samples (0.5 g)
using acetonitrile as extraction solvent (15 mL)
Dispersant

Co-sorbent

Percentage of fat (%)

C18
C18
C18
C18
Florisil
Alumina
Silica

Florisil
Alumina
Silica
C18
C18
C18
C18

2.8
2.9
3.1
1.0
0.3
0.4
1.0

The amounts of dispersant and co-sorbent were 2 g.

low affinity for lipids. In all experiments, 0.5 g of freeze-dried

salmon (fat content 59%), 2 g of dispersant and co-sorbent and
15 mL of acetonitrile were employed. The content of lipids in
the extracts was gravimetrically determined, after evaporation
to dryness, and expressed as the ratio between the mass of the
dry residue and the sample intake (0.5 g) multiplied by 100.
Table 2 shows some of the obtained data. The use of normalphase sorbents as dispersants and C18 as clean-up co-sorbent
led to extracts with a lower lipidic content than the reversed
combination. Particularly, the lowest fat residues (about 0.3%,
equivalent to 1.5 mg of lipids in the extract) corresponded to
samples dispersed on Florisil. Although in further experiments
higher masses of Florisil and C18 (up to 4 g) were considered, the
percentage of fat in the extracts remained at the 0.3% level, data
not given. An additional purification of the extracts with SPE
cartridges containing 0.5 g of PSA, a sorbent recommended for
fatty acids retention [29], also failed to reduce the level of lipids
in the final extract. On the other hand, replacement of acetonitrile for less polar solvents, such as ethyl acetate, which has also
been previously tested for TCS extraction from solid matrices
[30], produced extracts containing up to 30% of lipids. Similar levels of residues were noticed using a mixture of n-hexane:
dichloromethane (1:1) as eluent.

135

The efficiency of the extraction under the most favourable

conditions: Florisil (2 g) as dispersant, C18 (2 g) as co-sorbent
and acetonitrile (15 mL) as elution solvent, was evaluated with
samples, freeze-dried salmon, spiked at the 300 ng g1 level.
Extracts were concentrated to a final volume of 5 mL and an
aliquot (0.5 mL) mixed with 50 L of MTBSTFA. The yield of
the extraction was determined by GCMS/MS, using standards
derivatized under same conditions as sample extracts; moreover,
some injections were made in the GCMS mode to investigate
the complexity of the extracts. Although the above described
methodology provided quantitative recoveries (97.0 11% and
99.1 13% for MTCS and TCS, respectively), intense peaks
appeared in the GCMS chromatograms. By comparison with
the US National Institute of Standards and Technology (NIST)
mass spectra database, most of them were identified as tertbutyldimethylsilyl derivatives of saturated and non-saturated
fatty acids, Fig. 2. Since, at medium term, those co-extracted
species might impair the performance of GC column, particularly if extracts are concentrated to a low final volume (ca. 1 mL),
an alternative clean-up methodology was investigated.
3.3. Matrix solid-phase dispersion with lipids removal
In this approach, a layer of acidified silica was used as
clean-up co-sorbent. Initial experiments were performed using
SiO2 H2 SO4 (10%, w/w) and the same spiked, fatty matrix
as in the above section. Samples (0.5 g) were first mixed with
2 g of sodium sulphate, dispersed on 1.5 g of neutral silica
and then transferred to a MSPD cartridge containing different
amounts of the above clean-up co-sorbent. The elution solvent
was dichloromethane (15 mL). Considering 1 g of co-sorbent,
extracts still contained 12% of lipids; whereas, percentages
of only 0.050.06% (equivalent to 0.3 mg of dry residue) were
attained for 3 g. This lipidic residue is similar to that reported
for the MSPD extraction of polychlorinated biphenyls (PCBs)
and polybrominated diphenyl ethers (PBDEs) from fat samples

Fig. 2. GCMS chromatogram for the acetonitrile extract from a spiked salmon sample (300 ng g1 ). Florisil (2 g) and C18 (2 g) were used as dispersant and cosorbent, respectively. The sample intake and the final extract volume were 0.5 g and 5 mL. Peaks labelled as C8 to C18 correspond to silylated fatty acids with the
indicated number of carbons.

136

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

Table 4
Recoveries for spiked trout samples (300 ng g1 ) as function of storage conditions, n = 3 replicates

Fig. 3. Influence of the concentration of SiO2 H2 SO4 on the responses for

MTCS and TCS in the extracts from spiked freeze-dried salmon. Data for
triplicate extractions using dichloromethane (15 mL) as extraction solvent.

using n-hexane as elution solvent and SiO2 H2 SO4 (44%, w/w)

as co-sorbent [31]. Chromatograms for spiked samples proved
that MTCS and TCS survived to the strong oxidative conditions
employed in the MSPD process; therefore, this clean-up strategy
was adopted for the rest of this study and submitted to detailed
optimisation.
3.3.1. Concentration of sulphuric acid
Fig. 3 depicts the responses (peak areas) for MTCS and TCS
in the extracts from spiked (300 ng g1 ) salmon samples as
function of the concentration of sulphuric acid in the layer of
co-sorbent. As observed, the average responses for both species
remained constant (Fig. 3), which suggests that (1) they are stable under conditions employed in the extraction process and,
moreover, (2) they are not retained in the layer of carbon resulting
from the oxidation of fat and other co-extracted sample components. Regarding the percentage of lipids in the extracts no
additional reduction was noticed after increasing the content of
sulphuric acid from 10 to 20% (Fig. 3); therefore, the first was
selected as the optimal value for further experiments.
3.3.2. Extraction solvent and volume
In addition to CH2 Cl2 , n-hexane and CHCl3 were evaluated
for the extraction of MTCS and TCS from three different spiked
samples: freeze-dried salmon and trout as well as fresh cheese.
For each sample and solvent, three 5 mL volume fractions were
collected from the MSPD cartridge. They were evaporated to
dryness, re-constituted with 1 mL of ethyl acetate and mixed

Sample code

Storage conditions
(time, temperature)

1
2
3
4
5

Non-storeda

7 days, 20 C
7 days, 4 C
7 days, 20 Cb
7 days, 4 Cb

% Recovery SD
MTCS

TCS

107.8 7.3
94.8 8.6
106.8 5.5
n.d.
n.d.

80.0
92.1
97.9
101.1
108.7

7.4
12.9
7.6
3.3
2.6

n.d. under detection limits.

a Extracted 12 h after addition.
b Spiked only with TCS.

with the derivatization reagent (0.5 mL of extract plus 50 L

of MTBSTFA) previously to GCMS/MS determination. Each
series of experiments was carried out in triplicate. The less polar
solvent, n-hexane, was useless as extraction solvent since neither
MTCS nor TCS were detected in the extracts from any of the
samples. Their distribution pattern in the consecutive fractions
of CH2 Cl2 and CHCl3 is shown in Table 3. For both solvents,
MTCS was only detected in the first fraction. In the case of
TCS, around 90% of the compound was recovered with 5 mL of
CH2 Cl2 , whereas it appeared distributed in the three fractions of
CHCl3 , except for the sample of freeze-dried trout. On the basis
of these results, CH2 Cl2 was selected as the extraction solvent
and its volume limited to 10 mL.
3.3.3. Analytes stability
A crucial aspect during the development of a new sample
preparation procedure, particularly when hard extraction and/or
clean-up conditions are employed, is to guarantee the stability of target species. It is also relevant to assess the absence of
degradation or inter-conversion reactions during sample storage. Table 4 summarizes the recoveries of the proposed method
for freeze-dried trout samples spiked with MTCS and TCS at
300 ng g1 . Although values obtained for the three first samples
(codes 13) did not show statistically significant differences as
function of storage conditions, slightly higher recoveries were
achieved for MTCS than for TCS. In order to investigate the
potential methylation of a small percentage of TCS, either during
sample extraction and/or storage, two new fractions of the same

Table 3
Distribution of TCS and MTCS in the consecutive fractions, 5 mL each, obtained from samples spiked at the 300 ng g1 level, n = 3 replicates
Solvent

Sample

Normalised response with SD values within parenthesis

TCS
Fraction 1

CH2 Cl2

Salmon
Trout
Cheese

95 (10)
88 (3)
97 (12)

CHCl3

Salmon
Trout
Cheese

42 (5)
n.d.
40 (1)

n.d. not detected.

MTCS
Fraction 2

Fraction 3

4 (2)
9 (1)
2 (0.2)
48 (3)
53 (7)
54 (4)

Fraction 1

Fraction 2

Fraction 3

1 (0.3)
3 (0.6)
1 (0.1)

100 (2)
100 (9)
100 (11)

n.d.
n.d.
n.d.

n.d.
n.d.
n.d.

10 (1)
47 (14)
6 (0.1)

100 (3)
100 (11)
100 (5)

n.d.
n.d.
n.d.

n.d.
n.d.
n.d.

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

137

Fig. 4. Overlay of GCMS/MS chromatograms for a boiled-ham sample. A, procedural blank; B, non-spiked sample; C, same sample spiked with 10 ng g1 of
MTCS and TCS.

sample were spiked only with TCS (300 ng g1 ) and stored again
for 7 days at different temperatures. MTCS was not detected in
the extracts from any of both samples (codes 4 and 5, Table 4).
Therefore, it can be concluded that MTCS and TCS are stable
during the sample preparation process and that spiked samples,
at least in case of freeze-dried fish tissue, can be stored for at
least 1 week at 4 C.
3.4. Performance of the method
Extraction yields of the developed method were evaluated
using samples of freeze-dried salmon and trout and two fresh
foodstuffs, sliced cheese and boiled ham, with different fat
contents. Each sample was divided in three fractions, two of
them were spiked with target species at two different levels:
10 and 50 ng g1 , and the third considered as a blank. Spiked
and non-spiked fractions of each sample were aged for 1 week
at 4 C before extraction. As regards non-spiked samples, only
TCS was detected around the LOQs of the method in one
of the two foodstuff samples, Fig. 4. Recoveries for spiked
samples are summarized in Table 5. For MTCS they ranged

between 85 and 120% whereas values from 77 to 112% were

attained for TCS. Their associated standard deviations remained
below 12% in the four investigated matrices. Considering that
these data were not corrected with the use of internal surrogates, they are favourable in comparison with the 50% recovery
value reported by Allmyr et al. [24] for samples spiked with
13 C labelled TCS, and similar to those achieved by SanchesSilva et al. for foodstuffs spiked with TCS at the g g1 level.
The inter-day variability of the method was studied using a
sample of freeze-dried salmon spiked at the 25 ng g1 level
and extracted in duplicate during 4 consecutive days. The relative standard deviations of MTCS and TCS concentrations
were 4.2 and 7.2%, respectively. Considering a sample intake
of 0.5 g and using 1 mL of ethyl acetate to reconstitute sample extracts, the LOQs of the developed method were 1 and
2 ng g1 for TCS and MTCS. These figures are in the same
order of magnitude than the LOQ reported by Balmer et al.
[12] for MTCS (15 ng g1 , depending on the matrix fat content) using a larger mass of sample (525 g), and the values
obtained for TCS and MTCS in algae samples (510 ng g1 )
[19].

Table 5
Recoveries for spiked samples stored at 4 C for 7 days before extraction, n = 4 replicates
Sample

Fat content (%)

Addition level (ng g1 )

% recovery SD
MTCS

TCS

Freeze-dried Salmon

59

10
50

119.9 3.5
109.5 7.6

112.3 9.0
98.2 6.6

Freeze-dried trout

31

10
50

110.8 11.7
95.2 10.7

106.7 8.9
79.4 3.6

Cheese

40

10
50

103.4 11.7
110.3 7.0

95.1 9.2
112.6 8.9

Boiled ham

10
50

84.9 8.6
92.2 7.8

77.0 11.3
84.8 6.5

138

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

The excellent stability TCS and MTCS under strong oxidative

conditions (similar to that presented by well-known persistent
pollutants, e.g. PCBs) raises the concern in relation to their
persistence and bio-accumulative character. On the other hand,
experiments performed with TCS-containing kitchenware and
foodstuff samples confirmed the capability of this bactericide
to migrate from treated surfaces to food. This behaviour should
be taken into account in order to consider TCS as a suitable,
or unsuitable, anti-microbial additive in food packaging films
and any other materials in direct or indirect contact with foodstuff, either in domestic environments or in the food processing
industry.
Fig. 5. Time-course of TCS concentrations (n = 2 extractions) in sliced foodstuff
samples in contact with a cutting board treated with this bactericide.

3.5. Method application

The developed method was applied to analyse six freezedried samples of rainbow trout, mackerel and mussels obtained
in two small rivers (rainbow trout) and an estuary from an industrialized area in the Northwest of Spain. None of them contained
detectable amount of target species.
On the other hand, the presence of trace levels of TCS in
some sliced foodstuff samples (see previous section) suggested
that they might be contaminated due to direct contact with TCScontaining kitchenware utensils (e.g. knives, cutting boards, etc.)
or indirectly, due to the use of sponges or food service wipes to
clean surfaces in contact with foodstuff. To assess this possibility, sliced food samples were placed in contact with a cutting
board, incorporating TCS as anti-bactericide additive, acquired
from a local market. Fig. 5 shows the levels of TCS incorporated by two different matrices as function of the contact time.
As observed, concentrations in the range of 4050 ng g1 were
detected after a contact time as short as 5 min. Moreover, the
levels of TCS in both samples increased with time, reaching a
value over 600 ng g1 for the sample with the highest fat content
(43%) after 7 h.
4. Conclusions
The sample preparation approach optimised in this work
constitutes the first application of the MSPD technique to the
simultaneous extraction and purification of TCS and its methylated by-product (MTCS) from freeze-dried fish tissues and fresh
food samples. Although the use of sulphuric acid, as liquid
reagent and impregnated in silica, had been previously proposed
to the purification of TCS and MTCS extracts from biota samples, this is the first time that this clean-up strategy is combined
on-line with the extraction step. Major advantages of the proposed method are (1) simplicity, (2) low consumption of organic
solvents, (3) extraction of TCS and MTCS in the same fraction
and (4) a higher sample throughput in comparison with previously published alternatives. Moreover, it provides suitable
recoveries, precision and quantification limits for the determination of both compounds in real life samples.

Acknowledgements
Financial support from the Spanish Government, Xunta de
Galicia and FEDER founds (projects DGICT CTQ2006-03334
and PGIDIT06PXIB237039PR) is acknowledged. PC thanks the
Spanish Ministry of Education for a FPU grant.
References
[1] D. Sabaliunas, S.F. Webb, A. Hauk, M. Jacob, W.S. Eckhoff, Water Res.
37 (2003) 3145.
[2] E.M. Fiss, K.L. Rule, P.J. Vikesland, Environ. Sci. Technol. 41 (2007) 2387.
[3] M. Orhan, D. Kut, C. Gunesoglu, Indian J. Fibre Textile Res. 32 (2007)
114.
[4] Z. Moldovan, Chemosphere 64 (2006) 1808.
[5] T. Moretro, T. Sonerud, E. Mangelrod, S. Langsrud, J. Food Protect. 69
(2006) 627.
[6] A. Sanches-Silva, R. Sendon-Garca, J. Lopez-Hernandez, P. PaseiroLosada, J. Sep. Sci. 28 (2005) 65.
[7] D. Chung, S.E. Papadakis, K.L. Yam, Int. J. Food Sci. Technol. 38 (2003)
165.
[8] S.B. Choi, J. Jepperson, L. Jarabek, J. Thomas, B. Chisholm, P. Boudjouk,
Macromol. Symp. 249250 (2007) 660.
[9] J. Heidler, R.U. Halden, Chemosphere 66 (2007) 362.
[10] S. Morales, P. Canosa, I. Rodrguez, E. Rub, R. Cela, J. Chromatogr. A
1082 (2005) 128.
[11] K. Bester, Water Res. 37 (2003) 3891.
[12] M.E. Balmer, T. Poiger, C. Droz, K. Romanin, P.A. Bergqvist, M.K. Muller,
H.R. Buser, Environ. Sci. Technol. 38 (2004) 390.
[13] H. Singer, S. Muller, C. Tixier, L. Pillonel, Environ. Sci. Technol. 36 (2002)
4998.
[14] A. Kronimus, J. Schwarzbauer, L. Dsikowitzky, S. Heim, R. Littke, Water
Res. 38 (2004) 3473.
[15] H.R. Buser, M.E. Balmer, P. Schmid, M. Kohler, Environ. Sci. Technol. 40
(2006) 1427.
[16] K. Valters, H. Li, M. Alaee, I. DSa, G. Marsh, A. Bergman, R.J. Letcher,
Environ. Sci. Technol. 39 (2005) 5612.
[17] M. Alaee, I. DSa, E. Bennet, R. Letcher, Organohalogen Compd. 62 (2003)
136.
[18] M. Adolfsson-Erici, M. Pettersson, J. Parkkonen, J. Sturve, Chemosphere
46 (2002) 1485.
[19] M.A. Coogan, R.E. Edziyie, T.W. La Point, B.J. Venables, Chemosphere
67 (2007) 1911.
[20] M. Allmyr, M. Adolfsson-Erici, M.S. McLachlan, G. Sandborgh-Englund,
Sci. Total Environ. 372 (2006) 87.
[21] A.D. Dayan, Food Chem. Toxicol. 45 (2007) 125.
[22] N. Veldhoen, R.C. Skirrow, H. Osachoff, H. Wigmore, D.J. Clapson, M.P.
Gunderson, G. Van Aggelen, C.C. Helbing, Aquat. Toxicol. 80 (2006) 217.
[23] K.M. Crofton, K.B. Paul, M.J. DeVito, J.M. Hedge, Environ. Toxicol.
Pharmacol. 24 (2007) 194.

P. Canosa et al. / J. Chromatogr. A 1188 (2008) 132139

[24] M. Allmyr, M.S. McLachlan, G. Sandborgh-Englund, M. Adolfsson-Erici,
Anal. Chem. 78 (2006) 6542.
[25] E.M. Kristenson, L. Ramos, U.A.Th. Brinkman, Trends Anal. Chem. 25
(2006) 96.
[26] P. Canosa, S. Morales, I. Rodrguez, E. Rub, R. Cela, M. Gomez, Anal.
Bioanal. Chem. 383 (2005) 1119.
[27] P. Manirakiza, A. Covaci, P. Schepens, J. Food Compd. Anal. 14 (2001)
93.

139

[28] P. Canosa, I. Rodrguez, E. Rub, R. Cela, Anal. Chem. 79 (2007)

1675.
[29] O. Shimelis, Y. Yang, K. Stenerson, T. Kaneko, M. Ye, J. Chromatogr. A
1165 (2007) 18.
[30] P. Canosa, D. Perez-Palacios, A. Garrido-Lopez, M.T. Tena, I. Rodrguez,
E. Rub, R. Cela, J. Chromatogr. A 1161 (2007) 105.
[31] A. Martnez, M. Ramil, R. Montes, D. Hernanz, E. Rub, I. Rodrguez, R.
Cela, J. Chromatogr. A 1072 (2005) 83.