Determination of personal care products in sewage sludge by pressurized liquid extraction and ultra high performance liquid chromatography-tandem mass spectrometry.

Abstract

This paper describes a method for the determination of a group of personal care products including four UV filters, four preservatives and two antimicrobials in sewage sludge. The method combines pressurized liquid extraction and ultra high performance liquid chromatography-tandem mass spectrometry. Most of the parameters that affect the extraction step such as temperature, pressure, static extraction time, number of cycles, purge time and flush volume were optimized using a fractional experimental design. In the chromatographic step, the compounds were detected by using tandem mass spectrometry with a triple quadrupole analyzer with electrospray ionization in positive and negative modes. The use of small diameter particles (1.8 microm) in the chromatographic column allowed the compounds to be eluted in 9 min. The entire process took a total of 39 min. All recoveries were higher than 72% except for 2,4-dihydroxybenzophenone (a UV filter), whose recovery was 30%. The repeatability and reproducibility between days expressed as RSD (%) (n=3) were less than 8% and 13%, respectively. The LODs and LOQs were lower than 8 microg/kg and 12.5 microg/kg of dry weight (d.w.), respectively. When the method was applied to determine the compounds in sewage sludge from a domestic sewage treatment plant, triclosan (an antimicrobial) and octocrylene (a UV filter) showed the highest levels, 1490 microg/kg (d.w.) and 1842 microg/kg (d.w.), respectively. This paper describes for the first time the determination of parabens and two UV filters (octyldimethyl-p-aminobenzoic acid and benzophenone-3) in sewage sludge.

Full text

Page 1

Journal of Chromatography A, 1216 (2009) 5619–5625
Contents lists available at ScienceDirect
Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma
Determination of personal care products in sewage sludge by pressurized liquid
extraction and ultra high performance liquid chromatography–tandem mass
spectrometry
Antonio Nieto, Francesc Borrull, Rosa Maria Marcé∗, Eva Pocurull
Department of Analytical Chemistry and Organic Chemistry, Universitat Rovira i Virgili, Sescelades Campus, Marcel·lí Domingo s/n, Tarragona 43007, Spain
a r t i c l ei n f o
Article history:
Received 17 February 2009
Received in revised form 20 May 2009
Accepted 25 May 2009
Available online 3 June 2009
Keywords:
Parabens
Antimicrobials
Ultra-violet filters
Sewage sludge
Pressurized liquid extraction
Ultra high performance liquid
chromatography
Tandem mass spectrometry
a b s t r a c t
This paper describes a method for the determination of a group of personal care products including four
UV filters, four preservatives and two antimicrobials in sewage sludge. The method combines pressur-
ized liquid extraction and ultra high performance liquid chromatography–tandem mass spectrometry.
Most of the parameters that affect the extraction step such as temperature, pressure, static extraction
time, number of cycles, purge time and flush volume were optimized using a fractional experimental
design. In the chromatographic step, the compounds were detected by using tandem mass spectrome-
try with a triple quadrupole analyzer with electrospray ionization in positive and negative modes. The
use of small diameter particles (1.8?m) in the chromatographic column allowed the compounds to be
eluted in 9min. The entire process took a total of 39min. All recoveries were higher than 72% except for
2,4-dihydroxybenzophenone (a UV filter), whose recovery was 30%. The repeatability and reproducibility
between days expressed as RSD (%) (n=3) were less than 8% and 13%, respectively. The LODs and LOQs
were lower than 8?g/kg and 12.5?g/kg of dry weight (d.w.), respectively. When the method was applied
to determine the compounds in sewage sludge from a domestic sewage treatment plant, triclosan (an
antimicrobial) and octocrylene (a UV filter) showed the highest levels, 1490?g/kg (d.w.) and 1842?g/kg
(d.w.), respectively. This paper describes for the first time the determination of parabens and two
UV filters (octyldimethyl-p-aminobenzoic acid and benzophenone-3) in sewage sludge.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
The presence of pharmaceuticals and personal care products
(PPCPs) in environmental samples is a topic of increasing interest.
PPCPs are widely used around the world as a means of protecting
and improving human and animal health. There is growing concern
that these compounds pass through sewage treatment plants and
enter the environment [1]. Of particular concern are compounds
that are used in large volumes, persist in the environment, bioac-
cumulate, or have designed bioactivity [2].
Within the PPCP category, personal care products (PCPs) have
been the focus of study less frequently than pharmaceuticals.
Among PCPs, triclosan and triclocarban are antimicrobial com-
pounds used in soap, toothpaste, and other consumer products.
Triclosan is also used as a biocide in sportswear, footwear, carpets,
plastic toys, and kitchenware [3].
Another group of PCPs are UV filters. Sunscreen agents (UV fil-
ters) are chemical compounds that mitigate the deleterious effects
∗Corresponding author. Tel.: +34 977558170; fax: +34 977 558446.
E-mail address: rosamaria.marce@urv.cat (R.M. Marcé).
of sunlight and are used in a variety of cosmetics, specifically in
thosedesignedforsunprotectionsuchassunscreencreams,lotions
and sprays [4,5].
Another group of PCPs is made up of parabens, which are
the most common preservatives used in personal care products.
Parabensarealsousedaspreservativesandbactericidesinpharma-
ceuticalsandfoodproducts.Methylparabenandpropylparabenare
the most widely used and are normally used together due to their
synergistic preservative effects [6–8].
In recent years, the levels and consequences of PCPs in differ-
ent environmental waters have been the subject of several studies;
however, only a few PCPs have been determined in sewage sludge
[9,10] and therefore there is little information available regarding
their presence in these types of samples.
Some UV filters have been determined by gas chromatography–
mass spectrometry (GC–MS) [10,11] and parabens and triclosan by
gas chromatography–tandem mass spectrometry (GC–MS–MS) [3],
but a derivatization step was necessary in all cases. To determine
most PCPs, the best option is liquid chromatography coupled with
mass spectrometry (LC–MS) due to the polarity of PCPs and the
low concentrations at which these compounds have been found
in environmental samples [12–14]. Nowadays, the application of
0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2009.05.061

Page 2

5620
A. Nieto et al. / J. Chromatogr. A 1216 (2009) 5619–5625
advanced LC–tandem MS to environmental analyses has allowed a
broad range of compounds to be determined and has thus permit-
tedthecomprehensiveassessmentofenvironmentalcontaminants.
Different analyzers have been used, but the triple quadrupole ana-
lyzeristhemostsuitabletoobtainlowdetectionlimitsforthetarget
analytes.
Ultra high performance liquid chromatography (UHPLC) also
known as rapid resolution liquid chromatography (RRLC) (trade
name of Agilent Technologies) or ultra performance liquid chro-
matography (UPLC) (trade name of Waters) allows the possibility
of extending the usefulness of this widely used separation tech-
nique.UHPLCusesanalyticalcolumnspackedwith1.8?mparticles,
which offers the advantages of increasing speed, improving sensi-
tivity, selectivity and specificity compared to conventional HPLC
analysis. The higher efficiency of small particles enables shorter
columns to be used, reducing analysis time and solvent consump-
tion.
Up to now, UHPLC has not been extensively applied, but as a
method of determining PCPs, it has been used in a few studies
[15]. The lower degree of band broadening in UHPLC also bene-
fits mass spectrometric detection, concentrating the analyte at the
peak center and thereby increasing response. Thus, methods using
UHPLCinconjunctionwithmassspectrometry(MS)offerimproved
performancefor quantitative
chromatography–tandem mass spectrometry methods [16].
Due to the low concentration of PCPs in environmental sam-
ples, several techniques have been used to extract some of these
compounds from sludge and sediments, including soxhlet extrac-
tion [17], pressurized liquid extraction [10,12,18] and microwave-
assisted extraction [19].
The focus of this paper is the development of a method
for determining a set of personal care products from different
groups in sewage sludge in a single analysis. Four preserva-
tives (methyl paraben, ethyl paraben, propyl paraben and benzyl
paraben), six UV filters (octyldimethyl-p-aminobenzoic acid,
benzophenone-3, 2,4-dihydroxybenzophenone, 2,2?-dihydroxy-4-
methoxybenzophenone, octocrylene and 2-phenylbenzimidazole-
5-sulfonic acid) and two antimicrobials (triclosan and triclocarban)
analyses overexisting liquid
have been determined in sewage sludge using pressurized liquid
extraction followed by ultra high performance liquid chromatogra-
phy–tandem mass spectrometry. To the best of our knowledge this
is the first time that parabens and two UV-filters studied have been
determined in sewage sludge.
2. Methods and materials
2.1. Material and reagents
Four preservatives (methyl paraben, ethyl paraben, propyl
paraben and benzyl paraben), six UV filters (octyldimethyl-
p-aminobenzoicacid,benzophenone-3,
phenone, 2,2?-dihydroxy-4-methoxybenzophenone, octocrylene,
and 2-phenylbenzimidazole-5-sulfonic acid) and two antimicro-
bials (triclosan and triclocarban) were provided by Sigma (St. Louis,
MO, USA). The abbreviations are summarized in Table 1. Stock
solutions of individual standards were prepared by dissolving
each compound in methanol at a concentration of 1000mg/L and
storing them at 5◦C. A mixture of all compounds in methanol
at a concentration of 100mg/L was prepared weekly. Working
solutions were prepared daily.
Ultra pure water was obtained with a Milli-Q water purifica-
tion system (18.2M?cm) (Millipore, Bedford, MA, USA). Acetone,
acetonitrile, dichloromethane and methanol (HPLC-grade) were
acquired from SDS (Peypin, France), nitrogen was from Carburos
Metálicos (Tarragona, Spain) and acetic acid and aluminum oxide
were from Merck (Darmstadt, Germany).
In the PLE optimization, the data analysis was performed using
Statgraphics1Plusversion5.1software(Manugistics,Inc.,Rockville,
MD, USA).
2,4-dihydroxybenzo-
2.2. Sample preparation
The sewage sludge samples were homogenized, frozen,
lyophilizedusingthefreeze-drysystem(Labconco,KansasCity,MO,
USA), sieved through a 125?m screen and stored in a closed flask
at room temperature.
Table 1
MRM conditions used for UHPLC–MS–MS determination of personal care products. Bold face transitions were used to quantify.
CompoundAbbreviationTransitionsFragmentation voltage (V)Collision energy (V) Ionization modeTime window (min)
2-Phenylbenzimidazole-5-sulfonic acidPMDSA
275<194
275<211
200
200
30
25
Positive
0–1.3
Methyl paraben MPB
151<92
151<136
165<136
165<92
80
80
100
100
15
5
15
5
Negative
1.3–2.3
Ethyl paraben EPB Negative
Propyl parabenPPB
179<92
179<136
213<135
213<169
227<92
227<136
243<93
243<123
100
100
130
130
100
100
80
80
15
5
15
15
15
5
15
5
Negative
2.3–4.5
2,4-Dihydroxybenzophenone DHB Negative
Benzyl paraben BPBNegative
2,2?-Dihydroxy-4-methoxybenzophenone DHMB Negative
Benzophenone-3BP-3
229<151
229<105
130
130
15
15
Positive 4.5–5.5
TriclocarbanTCB
313<160
313<126
287<35
289<35
130
130
18
18
5 Negative
5.5–7
15
8
8
TriclosanTCSNegative
OctocryleneOC
384<272
384<228
278<166
278<151
130
130
100
100
5
5
Positive
7–13
Octyldimethyl-p-aminobenzoic acid ODPABA15
15
Positive

Page 3

A. Nieto et al. / J. Chromatogr. A 1216 (2009) 5619–5625
5621
To optimize the extraction procedure, fractions of the sample
(10g) were placed in different beakers, completely covered with
acetone (50mL) and spiked with the analytes at different levels.
The acetone was left to evaporate at room temperature until the
sludges were dry (3h). The mixture was stirred frequently in order
to obtain a homogenous material.
2.3. Pressurized liquid extraction
Extraction was done with an ASE 200 pressurized liquid extrac-
tor (Dionex, Sunnyvale, CA, USA) equipped with 11mL capacity
stainless-steel cells. One cellulose filter followed by 1g of alu-
minum oxide was placed at the bottom of each cell. After loading
the corresponding amount of aluminum oxide and 1g of sam-
ple, the remaining volume in the cell was filled with aluminum
oxide. The aluminum oxide was heated to 120◦C in the oven for
24hbeforeuse.Extractiontime,extractiontemperature,numberof
cycles,pressure,flushvolumeandpurgetimewereoptimizedusing
a fractional factorial experimental design (26−2) after the different
solvents being tested. Experiments to optimize the extraction pro-
cedure were performed by the extraction of 1g of a sewage sludge
sample spiked at 100?g/kg (d.w.).
The extracting solvents were methanol and a mixture of
water (pH 7) and methanol (1:1). The operating conditions were:
extraction temperature of 100◦C; extraction pressure of 140bar;
preheating period of 5min; 2 cycles of 5min with methanol fol-
lowed by 2 cycles of 5min with water (pH 7):methanol (1:1), final
extraction volume ∼25mL; flush volume of 30% of the cell volume
and nitrogen purge of 90s.
The extract was filtered with a microfilter with a pore size of
0.45?m (Teknokroma, Barcelona, Spain), and analyzed by ultra
high performance liquid chromatography–tandem mass spectrom-
etry.
2.4. UHPLC–MS–MS analysis
The chromatographic instrument was an HP1200 liquid
chromatograph–triple quadrupole tandem mass spectrometer
(Agilent Technologies, Waldbronn, Germany) with electrospray
ionization (ESI), an automatic injector, a degasser, a quaternary
pump and a column oven. The chromatographic column was a
Zorbax (5.0cm×0.46cm) with a 1.8?m particle size (Agilent Tech-
nologies), and the volume injected was 50?L. The mobile phase
flow-rate was 0.6mL/min and the column temperature was kept at
50◦C.
We used a binary mobile phase with a gradient elution. Sol-
vent A was Milli-Q water with acetic acid (pH 3) and solvent B was
methanol.Thegradientwasinitially60%B,whichincreasedto100%
in 6min, kept constant for 4min and finally returned to 60% B in
3min. All the compounds eluted within 9min.
Ionization and fragmentation settings were optimized by direct
injection of standard solutions. The parameters optimized for elec-
trospray ionization were: gas temperature (100, 200 and 350◦C),
gasflow(8,10and12L/min),nebulizerpressure(20,30and45psi.)
and capillary voltage (3000, 4000 and 5000V). These ESI condi-
tions were obtained as a compromise using the optimum values
for the majority of the compounds. The optimum ESI conditions
for the positive and negative mode were: capillary voltage 4000V,
nebulizer gas (N2) 45psi, source temperature 350◦C, gas flow (N2)
12L/min. Nitrogen was used as collision gas, and MS–MS was per-
formed in the Multiple Reaction Monitoring (MRM) mode. In order
to maximize sensitivity, six time windows were used and a differ-
ent ionization mode was used in each window. The time windows
and the ionization mode are detailed in Table 1.
For each compound, two characteristic fragments of the depro-
tonated [M−H]−or protonated molecular ion [M+H]+, depending
on the ionization mode, were monitored (Table 1). The ion [M+Na]+
was monitored for OC due to the higher response compared to
[M+H]+. For TCS the ions were two isotopic peaks of trichlorinated
species. The most abundant transition was used for quantification,
while the second most abundant was used as a qualifier. Frag-
mentation voltage and collision energy were optimized for each
compound (Table 1).
3. Results and discussion
3.1. Optimization of UHPLC–MS–MS
Two organic solvents, methanol and acetonitrile, and water at
different pH (3, 6 and 8) were tested to optimize the gradient sep-
aration and the conditions in the MS–MS detector. After injecting
a standard solution of 500?g/L, the best conditions for obtaining a
goodseparationandsymmetricpeakswerefoundwithacidicwater
(pH 3, acetic acid) and methanol.
Different parameters were tested to optimize the conditions for
MS–MS, the first of which were the electrospray ionization condi-
tions. We tested both ionization modes (positive and negative).
To increase the sensitivity of the acquisition and to obtain the
chromatogram combining both ionization modes, we defined six
different time windows, as described in Table 1.
Fragmentation voltage and collision energy were tested in order
to select the transitions in the MRM mode. Table 1 summarizes
the optimum values of these parameters for each compound. Tri-
closan has a fragmentation voltage of 18V because when this value
was increased the response decreased. The higher fragmentation
voltage used with PMDSA resulted in the best response.
The same product ions were observed in all parabens. The first
was m/z 92, which corresponds to the loss of the CO2group and the
methyl, ethyl, propyl or benzyl group depending on the compound.
The other product ion in parabens was m/z 136, which corresponds
to the loss of methyl, ethyl, propyl or benzyl. In some UV filters
we observed the loss of 77, which corresponds to [M−C6H5]+. This
situation appears in BP-3 (229<151) and DHB (213<135). In TCS
both transitions have the product ion m/z 35, which corresponds to
Cl35. The transitions selected are summarized in Table 1.
Linear range was tested by injection of standards at concentra-
tion levels ranging from 0.01 to 500?g/L and the determination
coefficient (r2) values were higher than 0.996 for all compounds.
3.2. Optimization of PLE
Initial conditions were selected from previous studies, where
someUVfiltersorantimicrobialsweredeterminedinsewagesludge
orsediments[10,18,20].Theinitialconditionswere:temperatureof
75◦C, pressure of 100bar, 5min of static extraction time, one cycle,
60s of purge time, 60% of flush volume and 1g of dry sample.
Differentblanksof sewage
PLE/UHPLC–MS–MS under the initial conditions and the chro-
matograms showed some peaks of target analytes at the same
retention time. In each experiment a blank and spiked samples
were analyzed and the signal of the blank was subtracted.
Different pure and binary mixture solvents were tested to opti-
mizetheextractionsolvent.Therecoveriesobtainedwhendifferent
pure solvents and mixtures of solvents were tested are shown
in Table 2. Six different pure solvents were tested: methanol,
dichloromethane, acetone, water, acetonitrile and ethyl acetate.
The results were different for each solvent because of the different
characteristics of the compounds. As expected, the most non-polar
compounds, TCB, TCS, OC and ODPABA had better recoveries when
organic solvents were used as the extraction solvent. MPB, EPB and
PPB had the highest recoveries when water was used. PMDSA, BPB,
DHB, and BP-3 were not extracted with any pure solvent. In gen-
sludge wereanalyzed by

Page 4

5622
A. Nieto et al. / J. Chromatogr. A 1216 (2009) 5619–5625
Table 2
PLE recoveries (n=3) using different solvents. A=water (pH 7), B=methanol, C=acetone, D=dichloromethane, E=acetonitrile and F=ethyl acetate. For other conditions see
the text.
Compounds
ABCDEFA:B (1:3)
A:B (1:1)
B:D (1:1)
PMDSA
MPB
EPB
PPB
DHB
BPB
DHMB
BP-3
TCB
TCS
OC
ODPABA
3––
–
–
–
–
–
–
–
–
–
–
–
1
–
–
–
–
–
–
–
–
–
5
–
–
–
7
3
7
––
–
9
6
–
39
72
60
19
18
15
–
–
12
–
53
100
100
100
14
58
50
–
7
–
16
–
40
51
–
15
–
7
–
–
22
9
24
1
–
–
–
–
–
–
10
8
–
– 12
–
–
33
100
99
86
13
–
–
–
–
–
30
40
65
41
54
98
100
86
20
70
20
–
34
16
RSD (%) ≤8.
eral, higher recoveries were obtained with methanol, with 8 of
the 12 compounds extracted, although some of them had very low
recoveries.
Therefore, we tested three different binary mixtures, as shown
in Table 2, water:methanol (1:1), water:methanol (1:3) and
methanol:dichloromethane (1:1). Higher recoveries were obtained
when methanol was used as extraction solvent for all compounds
except for MPB and EPB whose recoveries were higher when
water:methanol (1:1) was used.
The binary mixture that provided the highest recoveries was
water:methanol (1:1), although the recoveries were less than 58%.
To check if this solvent allowed higher recoveries to be obtained,
differentnumbersofcyclesweretested.AsshowninTable3,nodif-
ferences were found when the number of cycles increased to four.
Although the recoveries were not high for any of the compounds,
water:methanol (1:1) proved the best binary mixture to extract
parabens, and methanol to extract most non-polar compounds. We
therefore decided to test one cycle of methanol followed by one
cycle of water:methanol (1:1).
The extraction of these compounds using methanol and
water:methanol (1:1) was tested with one and two cycles of each
solvent. Table 3 shows that higher recoveries were obtained with
two cycles of each solvent. When three cycles of each solvent were
tested the recoveries were similar to the extraction obtained with
two. PMDSA and DHMB were not extracted under these conditions.
To find the conditions for fast and efficient extraction of the
target compounds from a solid matrix using PLE, a fractional facto-
rial design was chosen to investigate the influence of temperature,
pressure, extraction time, purge time, flush volume and number
of cycles on extraction efficiency. The combination of methanol
Table 3
PLE recoveries (n=3) using different cycles and different solvents. A=water (pH 7),
B=methanol. For other conditions see the text.
Compounds
A:B
(1:1)
1 cycle
A:B
(1:1)
2 cycle
A:B
(1:1)
4 cycle
B
1 cycle+A:B (1:1)
1 cycle
B
2 cycles+A:B (1:1)
2 cycles
PMDSA
MPB
EPB
PPB
DHB
BPB
DHMB
BP-3
TCB
TCS
OC
ODPABA
–––––
14
58
50
20
66
69
28
51
63
34
34
100
52
40
100
15
100
–
7
–
–
3
–
––
29 86
––
7
–
16
–
–
34
16
15
–
–
40
13
2338
35
100
100
100
–
–
35
100
100
86
50
19
RSD (%) ≤9.
followedbywater:methanol(1:1)wasselectedastheoptimumsol-
ventforourexperimentaldesign.Theexperimentaldesignselected
was a 26−2fractional factorial design. In this design the confound-
ing factors were purge time and flush volume. These factors were
considered as confounding factors because, as found in previous
studies [21,22], they do not significantly affect recoveries. Purge
time was confounded with the interaction of temperature, pres-
sure and extraction time and flush volume was confounded with
the interaction of pressure, extraction time and number of cycles.
Intheexperimentaldomainwefixedtwovaluesforeachparam-
eter, one for the low level and another for the high level. These
values were for temperature (50 and 100◦C), pressure (60 and
140bar), extraction time (5 and 15min), cycles (2 and 3), flush vol-
ume (30 and 90%) and purge time (30 and 90s). The Statgraphics
statistical package was used to generate the experimental matrix
andtocalculatethestandardizedmaineffectsofthefactorsconsid-
ered.Paretochartswereusedtoidentifythemostinfluentialfactors.
We considered individual recoveries and the average of all recov-
eries. For example, Fig. 1 shows the Pareto chart for the average of
all recoveries and exhibits information similar to that in the Pareto
chart for each compound analyzed separately. It should be men-
tioned that we cannot analyze the Pareto chart for four compounds
because all the experiments have the same recoveries, 0% (PMDSA
andDHMB)or100%(PBPandTCS).Fortherestofthecompounds,in
most cases, temperature was the most important parameter in the
extraction of these compounds. The recoveries were higher when
the temperature was at the highest level (100◦C). The extraction
time was the second most important parameter in seven of eight
compounds and the recoveries were higher when the extraction
time was at the lowest level (5min). To confirm whether these
were the optimum values, we decided to test another high tem-
perature (125◦C) and another short extraction time (3min) using a
22fullfactorialdesign.Thefactorsandtheirlevelswerefortemper-
ature (100 and 125◦C) and for extraction time (3 and 5min). The
Fig. 1. Standardized Pareto chart obtained from the response of the average of all
recoveries in the fraction factorial design.

Page 5

A. Nieto et al. / J. Chromatogr. A 1216 (2009) 5619–5625
5623
Table 4
PLE recoveries obtained with the 22experimental design. For other conditions see
the text.
Experiment1234
Temperature (◦C)
Extraction time (min)
PMDSA
MPB
EPB
PPB
DHB
BPB
DHMB
BP-3
TCB
TCS
OC
ODAPBA
100125100125
3
0
3
0
5
0
5
0
42
73
103
13
104
100
96
102
30
104
72
80
102
30
106
100
100
104
27
102
0540
35
70
102
52
74
60
50
100
61
73
79
77
103
104
108
61
65
104
70
73
other conditions were fixed depending on the values obtained in
the previous Pareto chart from the fractional factorial design. These
conditionsweretwocycles,pressureof140bar,flushvolumeof30%
of the volume of the cell and a purge time of 90s.
The recoveries obtained in the full factorial design are summa-
rized in Table 4, which shows that the highest recoveries for most
compounds were obtained in experiment 3 (100◦C and 5min). In
some cases, the recoveries in experiments 1 and 2 (extraction time
of 3min) were lower than when the extraction time was 5min. At
5min, when the temperature increased, the recoveries of MPB and
EPBincreasedto100%,however,therecoveriesofBP-3,TCB,OCand
ODPABA decreased. As shown in Table 4, PMDSA and DHMB were
not extracted and were therefore excluded from this study.
In the end, the optimum conditions for extracting personal care
products with PLE were set at 100◦C and 100bar with two cycles of
methanolfollowedbytwocyclesofwater:methanol(1:1,v/v),5min
of static extraction time, 90s of purge time and 30% flush volume.
Allrecoveriesexceeded72%exceptinthecaseofDHB,whoserecov-
ery was 30%. A pre-concentration step before the LC analysis was
not necessary because a low method detection limit was obtained,
thereby saving time, energy and also minimizing matrix effect in
the UHPLC–MS–MS. Applying the entire method (extraction and
chromatographic process) took only 39min.
4. Method validation
As mentioned above, some target analytes were found in the
sample when a PLE extract of different blanks of sewage sludge
was analyzed by UHPLC–MS–MS. For this reason the calibration
curves were obtained by injection of standard solutions instead of
by analyzing a sample using the entire method. Table 5 shows the
validation data obtained.
The identification and confirmation criteria for the analysis of
our target analytes were based on a series of Commission Decision
[23]. This decision provides rules for the analytical methods to be
used in analyzing the presence of residues in products of animal
origin. To confirm the presence of the compounds, the retention
time of the compounds and relationship between the two transi-
tions were compared. For instance, for methyl paraben the second
transition was 36% of the first transition in the standard injection
and in the samples this value was between 30% and 38%. A differ-
ence of less than ±20% was considered acceptable according to EU
directive.
The matrix effects were checked while the method was being
developed. We measured the recoveries of the analytes by spiking
thematrixextractwiththem.Tocalculatetheserecoveries,wesub-
tracted the signal of the blank and then we compared these signals
with the signal of a standard solution. A decrease was assumed to
be caused by matrix components in the extract. At the levels stud-
ied(100?g/kg(d.w.)and25?g/kg(d.w.))thematrixeffectwasless
than 15% and was therefore not considered.
The recoveries were also determined at 25?g/kg (d.w.), with
a difference of less than 5% respects to the recoveries obtained at
100?g/kg (d.w.).
Ourrecoverieswereconsistentwiththosefoundinotherstudies
in which UV filters and antimicrobials were determined separately.
For example, studies that used PLE or shaking extraction to extract
UV filters in sewage sludge and sediments reported similar recov-
eries to those obtained by our method for OC, ODPABA and BP-3
[9,10,24]. To our knowledge, there are no other studies in the liter-
ature concerning the determination of parabens in sewage sludge
using PLE as an extraction technique. However, some studies have
reportedrecoveriesinindoordustofover70%forallparabenswhen
PLE and matrix solid-phase dispersion were used as extraction
techniques [3,20]. Antimicrobials (TCS and TCB) are often studied
together with other groups of compounds and the recoveries are
alwayshigherthan80%indifferentkindsofsamplessuchassewage
sludge [18], sediments [12] or indoor dust [3,20]. When TCS was
analyzed using an MAE with GC–MS–MS [19], recoveries of more
than81%wereobtainedinsewagesludgeandsediments.Therefore,
our study is relevant in that good recoveries were obtained using
the same extraction process for groups of compounds which have
different characteristics.
Theprecisionofthemethodwasevaluatedbyextractingconsec-
utive extraction of 3 spiked samples at 25?g/kg (d.w.) (within day
repeatability) and three spiked sludge samples at 25?g/kg (d.w.)
Table 5
Validation data (n=3).
Recoveries (%)a
Linear rangeb(?g/L)LODsc(?g/kg) LOQsd(?g/kg) RSD (%)e
RSD (%)f
MPB
EPB
PPB
DHB
BPB
BP-3
TCB
TCS
OC
ODPABA
72
80
102
30
106
79
77
103
104
108
250–0.25
250–0.1
250–0.1
250–0.1
250–0.25
250–0.1
250–0.1
250–0.5
500–0.25
500–0.1
3
1.75
1.75
2
3
1.5
1.25
8
3.5
1.75
6.3
2.5
2.5
2.5
6.3
2.5
2.5
12.5
6.3
2.5
4
3
2
6
2
1
4
8
3
3
5
5
7
10
5
1
8
13
8
4
aSample spiked at 25?g/kg (d.w.).
bInstrumental linear range.
cLimit of detection of the method.
dLimit of quantification of the method.
eRepeatability (n=3).
fReproducibility between days (n=3).

Page 6

5624
A. Nieto et al. / J. Chromatogr. A 1216 (2009) 5619–5625
Fig. 2. MRM chromatogram obtained by PLE/UHPLC–MS–MS of 1g of sewage sludge collected in July 2007. For experimental conditions see text.
in three days (reproducibility between days). The relative standard
deviation varied between 1% and 8% and 1% and 13%, respectively.
The limit of quantification (LOQ), calculated as the concentra-
tion of the lowest point of the calibration curve, ranged from 2.5 to
12.5?g/kg (d.w.). The LOQ for triclosan is almost one order of mag-
nitude higher than other compounds, as in other studies reported
[18]. The lower sensitivity of determination for triclosan is due to
the poor yield of the product ion (m/z 35) relative to the transi-
tionmonitoredforothercompounds.Thelimitsofdetection(LODs)
were defined for a ratio signal to noise of 3 for all compounds and
were lower than 8?g/kg (d.w.). The LODs and LOQs obtained in our
studywerecomparabletoandslightlybetterthanthoseobtainedin
otherstudiesusingdifferentdetectionsystemsandextractiontech-
niques. Antimicrobial compounds (TCS and TCB) were determined
in sewage sludge using PLE/LC–MS–MS with a triple quadrupole
analyzer with detection limits of 1.5 and 0.2?g/kg (d.w.) for TCS
and TCB, respectively [18].
Toourknowledge,onlytheOC(UVfilter)hasbeendeterminedin
sewage sludge, but using a shaking extraction process and GC–MS
[9]. In that study, the LOD was 6?g/kg (d.w.)—a slightly higher
result than our LOD for OC (3.5?g/kg (d.w.)). But there are stud-
ies in which UV filters were determined in other kind of samples.
For instance, when the UV filters ODPABA, OC and BP-3 were deter-
mined in lake sediments using PLE/GC–MS, the detection limits
were between 2 and 6?g/kg (d.w.) [10]; however in the literature
no data for these compounds using LC–MS–MS were found.
5. Method application
Themethoddevelopedwasappliedtothreesewagesludgesam-
ples collected in January, July and September of 2007 from a STP in
the city of Tarragona. Table 6 shows the results of this study. EPB
and DHB were excluded from the table because they did not appear
in any samples.

Page 7

A. Nieto et al. / J. Chromatogr. A 1216 (2009) 5619–5625
5625
Table 6
Results in ?g/kg (d.w.) of sewage sludge samples analyzed. Relative standard devi-
ation (%RSD) is in brackets (n=3).
January 2007July 2007 September 2007
MPB
PPB
BPB
BP-3
TCB
TCS
OC
ODPABA
202 (5)
7 (9)
5 (4)
20 (3)
7 (2)
1490 (8)
700 (4)
132 (5)
80 (4)
10 (5)
46 (5)
6 (5)
––
20 (4)
6 (3)
1300 (7)
1800 (9)
153 (5)
10 (4)
5 (2)
1328 (5)
1842 (9)
170 (4)
The levels of some compounds were similar in all samples
throughout the year. For instance, PPB showed consistent levels
of between 6 and 10?g/kg (d.w.). Lee et al. [25] found the high-
est levels of MPB and PPB in influent and effluent water, but in the
literature we could not find studies reporting the concentrations
of parabens in sewage sludge samples. In our case the parabens
MPB and PPB were present in all samples at levels between 6 and
202?g/kg (d.w.).
TCB and TCS also showed similar concentrations in all sam-
ples: between 1300 and 1490?g/kg (d.w.) and between 5 and
7?g/kg (d.w.), respectively. Chu and Metcalfe [18] found simi-
lar concentrations in municipal biosolids for triclosan and even
higher concentrations of TCB. When TCS was studied in differ-
ent kinds of sludges (primary, biological or disinfected) using
a MAE/GC–MS–MS, concentrations between 418 and 5400?g/kg
(d.w.) were found in a STP in Spain [19].
One of the UV filters (OC) showed twice the concentration
in the sample collected in summer. In winter, OC was detected
at 700?g/kg (d.w.) while in summer the levels increased to
1800?g/kg (d.w.). Although we have only a few samples, this result
may be explained by the fact that OC is used in sun blocks to
protect the skin and its use in the tourist area of Tarragona
increases in summer. The concentration of 1800?g/kg (d.w.) was
low in comparison with that found in the study of Plagellat et al. [9]
wherehigherconcentrations(between320and18740?g/kg(d.w.))
were detected in different sewage sludges from Switzerland. The
different design of the sewage treatment plant may be one of the
reasons of this difference. Two of the UV filters studied (OC and
BP-3) were also found in lake sediments in Germany at levels
between 60 and 90?g/kg (d.w.) [10]. UV filters have even been
found in fish samples at levels of a few ?g/kg of OC and BP-3
[24,26]. The high concentration in fish results is obviously due to
the presence of these compounds in water and to the bioaccumu-
lation potential of the lipophilic UV-filters. This occurs because the
elimination power of STPs is not 100%, as is confirmed by Balmer et
al. [24] who found an elimination power of between 68% and 99%
for BP-3 and OC.
As an example, Fig. 2 shows the MRM chromatograms of a sam-
ple collected in July 2007.
6. Conclusions
The suitability of the PLE technique with UHPLC–MS–MS with
triple quadrupole analyzer for the determination of preservatives,
UV filters and antimicrobials all together in a single analysis of
sewage sludge has been demonstrated for the first time. The
method can be qualified as a rapid method, with 30min of extrac-
tion time and 9min of separation analysis. The limit of detection
and limit of quantification were lower than 8?g/kg (d.w.) and
12.5?g/kg (d.w.), respectively. In the analysis of three different
samplesfromaSTP,eightdifferenttargetanalytesweredetermined.
The UV filter, OC, showed the highest concentration in the sum-
mer sample (1842?g/kg (d.w.)) and triclosan showed the highest
concentration in the winter sample (1490?g/kg (d.w.)). For the
first time, this paper demonstrated the presence of two UV filters
(ODPABA and BP-3) and parabens in sewage sludge samples.
Acknowledgements
The authors wish to thank the personnel of the water treatment
plant in Tarragona for their cooperation in all aspects of this study.
This study was financially supported by the Dirección General
de Investigación of the Ministry of Science and Technology, project
CTM2008-06847-C02-01/TECNO. A.N. gratefully acknowledges the
financial support of the Ministry of Science and Technology BES-
2006-1332.
References
[1] D.L. Giokas, A. Salvador, A. Chisvert, Trends Anal. Chem. 26 (2007) 360.
[2] A. Kot-Wasik, J. Debska, J. Namiesnik, Trends Anal. Chem. 26 (2007) 557.
[3] P. Canosa, I. Rodríguez, E. Rubí, R. Cela, Anal. Chem. 79 (2007) 1675.
[4] A. Chisvert, M.C. Pascual-Martí, A. Salvador, J. Chromatogr. A 921 (2001)
207.
[5] D.Orsi,G.Giannini,L.Gagliardi,R.Porrà,S.Berri,A.Bolasco,I.Carpani,D.Tonelli,
Chromatographia 64 (2006) 509.
[6] A.M. Peck, Anal. Bioanal. Chem. 386 (2006) 907.
[7] P.Canosa,I.Rodríguez,E.Rubí,M.H.Bollaín,R.Cela,J.Chromatogr.A1124(2006)
3.
[8] M. Borremans, J.V. Loco, P. Roos, L. Goeyens, Chromatographia 59 (2004) 47.
[9] C. Plagellat, T. Kupper, R. Furrer, L.F. Alencastro, D. Grandjean, J. Tarradellas,
Chemosphere 62 (2006) 915.
[10] R. Rodil, M. Moeder, Anal. Chim. Acta 612 (2008) 152.
[11] P. Cuderman, E. Heath, Anal. Bioanal. Chem. 387 (2007) 1343.
[12] A. Agüera, A.R. Fernández-Alba, L. Piedra, M. Mézcua, M.J. Gómez, Anal. Chim.
Acta 480 (2003) 193.
[13] R.U. Halden, D.H. Paull, Environ. Sci. Technol. 38 (2004) 4849.
[14] R.U. Halden, D.H. Paull, Environ. Sci. Technol. 38 (2005) 1420.
[15] B.Kasprzyk-Hordern,R.M.Dinsdale,A.J.Guwy,Anal.Bioanal.Chem.391(2008)
1293.
[16] T. Yoshida, R.E. Majors, J. Sep. Sci. 29 (2006) 2421.
[17] S. Morales, J.L. Luque-García, M.J. Ramos, A. Fernández-Alba, M.D. Luque de
Castro, Anal. Chim. Acta 552 (2005) 50.
[18] S. Chu, C.D. Metcalfe, J. Chromatogr. A 1164 (2007) 212.
[19] S. Morales, P. Canosa, I. Rodríguez, E. Rubí, R. Cela, J. Chromatogr. A 1082 (2005)
128.
[20] P. Canosa, D. Pérez-Palacios, A. Garrido-López, M.T. Tena, I. Rodríguez, E. Rubí,
R. Cela, J. Chromatogr. A 1161 (2007) 105.
[21] A. Nieto, F. Borrull, E. Pocurull, R.M. Marcé, J. Sep. Sci. 30 (2007) 979.
[22] A. Nieto, F. Borrull, R.M. Marcé, E. Pocurull, J. Chromatogr. A 1174 (2007)
125.
[23] OfficialJournaloftheEuropeanCommunitiesL221,8–36,CommissionDecision
(2002/657/EC) of 12 August 2002 concerning the performance of analytical
methods and the interpretation of results, Brussels, Belgium.
[24] M.E. Balmer, H. Buser, M.D. Müller, T. Poiger, Environ. Sci. Technol. 39 (2005)
953.
[25] H.B. Lee, T.E. Peart, M.L. Svoboda, J. Chromatogr. A 1094 (2005) 122.
[26] M. Meinerling, M. Daniels, Anal. Bioanal. Chem. 386 (2006) 1465.