See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223005649

Organic UV filters and their photodegradates. Metabolites, and disinfection

by-products in the aquatic environment

Article in TrAC Trends in Analytical Chemistry · November 2008

DOI: 10.1016/j.trac.2008.08.012

CITATIONS READS

201 1,082

3 authors:

Silvia Díaz-Cruz Marta Llorca

Spanish National Research Council ICRA Catalan Institute for Water Research
144 PUBLICATIONS 7,974 CITATIONS 66 PUBLICATIONS 3,276 CITATIONS

SEE PROFILE SEE PROFILE

Damia Barcelo
ICRA Catalan Institute for Water Research
882 PUBLICATIONS 58,455 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Maternal transfer of halogenated organic compounds in dolphins View project

Uptake and metabolism of Carbamazepine (CBZ) by Clam Ruditapes decussatus and its effects in biochemical responses View project

All content following this page was uploaded by Silvia Díaz-Cruz on 04 June 2018.

The user has requested enhancement of the downloaded file.

Anal Bioanal Chem (2012) 402:2325–2333
DOI 10.1007/s00216-011-5560-8
ORIGINAL PAPER
Analysis of UV filters in tap water and other clean waters
in Spain
M. Silvia Díaz-Cruz & Pablo Gago-Ferrero &
Marta Llorca & Damià Barceló
Received: 12 July 2011 / Revised: 26 October 2011 / Accepted: 3 November 2011 / Published online: 3 December 2011
# Springer-Verlag 2011
Abstract The present paper describes the development
of a method for the simultaneous determination of five
hormonally active UV filters namely benzophenone-3
(BP3), 3-(4-methylbenzylidene) camphor (4MBC), 2-
ethylhexyl 4-(dimethylamino) benzoate (OD-PABA), 2-
ethylhexyl 4-methoxycinnamate (EHMC) and octocry-
lene (OC) by means of solid-phase extraction and gas
chromatography–electron impact ionization–mass spectrom-
etry. Under optimized conditions, this methodology achieved
low method limits of detection (needed for clean waters,
especially drinking water analysis), between 0.02 and 8.42 ng/
L, and quantitative recovery rates higher than 73% in all cases.
Inter- and intraday precision for all compounds were lower
than 7% and 11%, respectively. The optimized methodology
was applied to perform the first survey of UV absorbing
compounds in tap water from the metropolitan area and the
city of Barcelona (Catalonia, Spain). In addition, other types
of clean water matrices (mineral bottled water, well water and
Published in the special issue Mass Spectrometry in Spain with guest
editors José Miguel Vadillo and Damià Barceló.
M. S. Díaz-Cruz (*) : P. Gago-Ferrero : M. Llorca : D. Barceló
Department Environmental Chemistry, Institute of Environmental
Assessment and Water Studies (IDAEA),
Spanish Council of Scientific Research (CSIC),
Jordi Girona 18-26,
08034 Barcelona, Spain
e-mail: sdcqam@cid.csic.es
D. Barceló
Catalan Institute for Water Research (ICRA), H2O building,
Scientific and Technological Park of the University of Girona,
Emili Grahit 101,
17003 Girona, Spain
D. Barceló
King Saud University,
P.O. Box 2455, Riyadh 11451, Saudi Arabia
tap water treated with an ion-exchange resin) were investigat-
ed as well. Results evidenced that all the UV filters
investigated were detected in the water samples analyzed.
The compounds most frequently found were EHMC and OC.
Maximum concentrations reached in tap water were 290
(BP3), 35 (4MBC), 110 (OD-PABA), 260 (EHMC), and
170 ng/L (OC). This study constitutes the first evidence of the
presence of UV filter residues in tap water in Europe.
Keywords Gas chromatography–mass spectrometry .
Hormonally active compounds . Organic UV filters .
Sunscreens . Tap water
Introduction
The rising level of awareness of the harmful effects of sun
radiation has boosted a growth in the use of sun care
products. These substances are used to absorb UVA (315–
400 nm) and UVB (280–315 nm) radiation. Since their
early introduction at the end of the nineteenth century [1],
UV filters have been used in personal and pet care
products, pharmaceuticals, domestic and industrial com-
modities, vehicle maintenance products, and pesticides [2].
UV filters are considered as environmental contaminants of
increasing concern since a number of them are known to elicit
hormonal activity in both aquatic and terrestrial organisms [3–
5]. Estrogenic and androgenic activity of some UV filters
such as 2-ethylhexyl 4-methoxycinnamate (EHMC), 2-
ethylhexyl 4-(dimethylamino) benzoate (OD-PABA), homo-
salate, benzophenone 3 (BP3), 3-(4-methylbenzylidene)
camphor (4MBC), octocrylene (OC), and butyl methoxydi-
benzoylmethane (BM-DBM), has been reported in in vitro
and in vivo assays, in fish and rodent cell lines [6–8]. In
these trials, the compounds were tested as individuals or as
2326
mixtures. However, their effects in humans still remain
unclear. 4MBC and BP3 were identified as the two UV
filters showing the highest estrogenic activity. Consequently,
in Denmark, 4MBC is currently not recommended in
products intended for children use [9].
The list of permissible compounds by legislation is
regularly updated and 27 such substances (26 organic and 1
inorganic) are currently (from 2007) permitted in the
European Union (EU) [10].
Two pathways were identified for UV filters to enter the
aquatic environment; direct inputs as consequence of
recreational water activities, and indirect inputs mainly
from wastewater as result of the use of cosmetics,
showering and washing, as well as industrial waste.
So far, a number of studies reported levels of organic UV
filters in different aqueous matrices: swimming pool waters
[11–13] bathing waters [14, 15], surface waters [5, 12, 16–
19], wastewaters [14–16, 18–21], and seawater [18]. So far,
the occurrence of EHMC in drinking water has occasionally
been reported [22], but their identification was accidental
(identified along with other compounds) rather than
intentional. In this study, solely one out of the 15 finished
drinking water samples from San Diego (CA, USA) was
found to contain EHMC at 0.45 μg/L [22].
The goal of the present study was to develop and validate a
highly sensitive multi-residual analytical methodology based
on solid-phase extraction (SPE) and purification, and further
analysis by gas chromatography–electron impact–mass spec-
trometry in order to investigate the occurrence of five of the
currently legislated and known to be hormonally active
sunscreen agents, namely BP3, 4MBC, OD-PABA, EHMC,
and OC, at trace level concentrations in different clean waters.
Tap water of Barcelona and surroundings (Spain) as well as
bottled mineral water, ionic-exchange treated tap water and
well water were investigated for the presence of the
aforementioned UV filters.
Experimental section
Chemicals and reagents
High purity analytical standards of BP3 (98% purity), OD-
PABA (98% purity), and OC (97% purity) and the
isotopically labeled benzyl cinnamate (BC) used as internal
standard (IS) (≥99% purity) were purchased from Sigma-
Aldrich (Steinheim, Germany); 4MBC (>99.5% purity) was
obtained from Dr. Ehrenstorfer (Augsburg, Germany),
EHMC, (>99% purity) from Merck (Darmstadt, Germany).
Table 1 shows the structures and some relevant physi-
cochemical properties of the above compounds.
High purity solvents, ethyl acetate (AcEt), dichloro-
methane (CH2Cl2), MeOH, n-hexane, and HPLC-grade
M.S. Díaz-Cruz et al.
water were supplied by Merk (Darmstadt, Germany).
Helium from Air Liquide (Barcelona, Spain) was of 99.99
purity.
SPE was carried out using, Isolute SPE Columns C18
(500 mg), purchased from Biotage (Charlottesville, Virginia,
USA). The 2.7- and 1.0-μm glass fiber filters and 0.45-μm
nylon membrane filters used for filtering the water samples
were purchased from Whatman (Maidstone, England).
Both stock standards and the isotopically labeled
compound were prepared in n-hexane at a concentration
of 200 mg/L. Standards solutions were stored in the dark at
−20 °C. From these solutions, a mixture standard solution
of all UV filters in n-hexane at a concentration of 20 mg/L
of each compound was prepared weekly. Mixed working
solutions were prepared daily by appropriate dilution of the
mixed stock standard solution in n-hexane.
Contamination and photo-degradation considerations
In the current study, several measures were taken in order to
prevent contamination and photo-degradation of samples
and standards. Personnel was informed about avoiding the
use of products containing UV filters (hairspray, make up,
etc.) during sampling and further sample treatments. All
glassware used was previously washed and heated over-
night at 380 °C, and further sequentially rinsed with
different organic solvents and HPLC-grade water, and
immediately used. Furthermore, gloves were worn during
sample preparation; separate solvents and only previously
unopened packages of solvents, chemicals, and other
supplies were used. All samples and stock standard
solutions were always covered with aluminum foil and
stored in the dark in order to prevent the exposure to light
of target compounds. Finally, a triplicate of operational
blanks was processed together with each batch of samples.
Sample collection and sampling sites
In order to assess the occurrence of UV filters, in water
three different mineral bottled waters (acquired in the
supermarket), tap water, well water, and tap water
treated with ion-exchange resins from Barcelona were
collected.
Also, a comprehensive survey of 26 tap waters (six of
them treated with ionic-exchange resins using a domestic
purifying water system) was carried out. Sampling was
carried out in May 2009 during a period of 2 weeks (the
samples were taken every 2 days). Water samples, around
1 L, were taken in amber PTFE bottles pre-rinsed with
CH2Cl2, MeOH, and HPLC-grade water and shipped to the
laboratory under cool conditions. Three sample replicates
were taken in all cases.
UV filters in tap water and other clean waters in Spain 2327

Table 1 Physicochemical properties of the examined UV filters

Name CAS Nº Structure Log Solubility max

(molecular weight, g/mol) Kow (g/L) (nm)

Benzophenone-3 131-57-7 3,79 0,21 290

(BP3)

228,24

Octocrylene 6197-30-4 7,53 2 x 10-- 4 300

(OC)

361,20

2-ethylhexyl 4- 5466-77-3 5,8 0,15 306

methoxycinnamate
(EHMC)

290,4

3-(4-methylbenzylidene) 36861-47-9 4,95 5,1 x 10-- 3 300

camphor
(4-MBC)
254,37

2-ethylhexyl 4- 21245-02-3 6,15 2,1 x 10-- 3 310

(dimethylamino) benzoate
(OD-PABA)
277,4

Names from http://pharmacos.eudra.org/F3/inci/inciaho.html, in parenthesis, the key system adopted in this paper. Data from [28] and [29]

In this study, the sampling locations were in
differentiated areas of Barcelona city (tap 2, tap 3, and
exchange resin), and the surrounding metropolitan area
(tap 1). These locations were chosen to map at least two
expected major spatial trends in concentrations of UV
filters, since two internal basins, Llobregat and Ter
rivers, supply the drinking water to these areas. Both
rivers start on the southern slopes of the Pyrenees and
end on the Mediterranean Sea. Their respective lengths
and basin areas are: Llobregat 156 km, 4.957 km2, and
700 Hm3/year mean flow; Ter 208 km, 3.010 km2, and
845 Hm3/year mean flow. Both basins are interconnected,
and they constitute the main water resource of Catalonia
(Spain), covering the water supply for urban and industrial
purposes of the heavily populated Barcelona area
(5,000,000 inhabitants). Although both rivers share many
characteristics, they exhibit some differences: while
Llobregat is subjected to industrial, mining and urban
pressures, Ter is more affected by agricultural practices.
Sample preparation
All water samples were filtered through 2.7 and 1.0 μm
glass micro fiber filters, followed by 0.45 μm nylon
membrane filters before extraction. Then, samples were
stored in the dark at 4 °C until analysis (normally within
2 days).
Aliquots of 200 L water samples were loaded onto
Isolute cartridges at 1 mL/min assisted by a 20-fold vacuum
extraction box (J.T. Baker, Phillipsburg, NY, USA) after
being conditioned with 5 mL CH2Cl2/AcEt (1:1), 5 mL
MeOH and 5 mL HPLC-grade water. Cartridges were then
under vacuum dried during 30 min. The target analytes
were eluted from the SPE cartridges with four successive
2.5 mL aliquots of CH2Cl2/AcEt (1:1) at a flow rate of
0.5 mL/min. The extracts were then evaporated under a
gentle stream of nitrogen, at 25 °C, in a TurboVap LV
evaporator (Zymark, Hopkinton, MA, USA), and further
reconstituted in 0.5 mL of n-hexane plus 0.1 mL of a
solution of benzyl cinnamate in n-hexane. Triplicates for
each solution were carried out. The samples were finally
analyzed by GC-MS.
GC-MS analysis
GC-MS analyses were performed on a Trace 2000 Series gas
chromatograph (Thermo Finnigan, Manchester) coupled to a
TRACE MS (Thermo) single quadrupole equipped with an
2328 M.S. Díaz-Cruz et al.

(pg/ml)
MLOQ

0.44
AS2000 auto sampler (Thermo) under electron impact (EI)

0.5
1.2

4.5
25
ionization mode. (EI, 70 eV, 200 °C) and selected ion
monitoring (SIM) conditions. The detector voltage was set

Tap water

(pg/ml)
to 400 V.

MLOD

0.14
0.37
0.13
7.4

1.4
Analyte separation was achieved using a DB-5MS
capillary column of 30 m×0.25-mm inner diameter and a

na not analyzed, ILOD instrumental limit of detection, ILOQ instrumental limit of quantification, MLOD method limit of detection, MLOQ method limit of quantification
film thickness of 0.25 μm (J&W Scientific, Folsom, CA,

(pg/ml)
MLOQ

0.60
0.07
0.92
3.25
USA) with the following temperature program: initial

na
temperature 70 °C (holding 2 min) to 150 °C (holding

Water HPLC
2 min) at 20 °C/min to 250 °C at 20 °C/min (holding

(pg/ml)
MLOD

0.011
12 min) and finally to 290 °C at 20 °C/min (holding 5 min).

0.10

0.28
0.98
Helium was used as a carrier gas with a constant flow rate

na
of 1 mL/min and 5 min of solvent delay. The injection was
operated at splitless conditions for 0.8 min. The injector

Tap water

79±20

82±13
111±18
temperature and volume were set to 220 °C and 2 μl,

74±1

83±4
(n=3)
respectively.

Recoveries±S.D (%)
A standard mixture of target compounds was acquired in
MS SCAN mode to select the characteristic fragment ions

Water HPLC
of each one using the NIST/EPA/NIH Mass Spectral

120±1
109±5
82±1
73±1
Method
Library (NIST 98) for confirmation purpose. For further

na
(n=3)
samples, MS acquisition was carried out in SIM mode for Table 2 Performance of the GC-MS developed method for the analysis of the selected UV filters in water samples
improved sensitivity.
The following ions were used to monitor the target UV

Intraday

8.6
8.3
8.9

8.8
10.5
filters (in italics are the ions used for quantification, all the

(n=5)
Precision (%RSD)
rest used for confirmation purpose; ion identification is
indicated for the two ones most intense): m/z 228 [M]+.,
227 [M-H]+, 151 [M-C6H5]+ and 77 [M-C8O3H7]+ (BP3);
Intraday
(n=9)

3.3
4.1
2.9
6.9
254 [M]+., 239 [M-CH3]+, 171 and 128 (4MBC); 277 3
[M]+., 165 [M-C8H16]+. and 148 (OD-PABA); 290 [M]+.,
178 [M-C8H16]+., 161 and 133 (EHMC); 360 [M-H]+, 249
ILOQ (pg)

[M-C8H16]+., 232 and 204 (OC). The IS was monitored at

6.4
6.5
6.7
m/z 238, 192, 131 and 91.
70

30
Quantification was performed by means of internal
standard calibration based on peak areas.
ILOD (pg)

1.9
2.0
2.0
9.1
20

Results and discussion

Extraction and purification optimization

0.9956
0.9972
0.9991
0.9991
0.9971
r2

It is known that extraction efficiency of a SPE procedure

relay mainly on the nature of the adsorbing material, the
Working range (ng/ml)

solvents used and the sample volume loaded. In this study,

we selected silica-based Isolute C18 cartridges previously
50–5,000
5.0–5,000
5.0–5,000
5.0–5,000
10–5,000

conditioned with AcEt/CH2Cl2 (1:1), MeOH and water, and

Instrumental

eluted with four 2.5-ml aliquots of the AcEt/CH2Cl2 (1:1)

mixture solution. According to literature data [12], this
combination yielded good recovery rates in the analysis of
the individual compounds.
In order to determine the optimum sample volume to be
OD-PABA
4(E)-MBC

extracted, aliquots of 50, 100, 200, and 400 mL of a 10 ng/

EHMC

mL UV filters mixture standard solution in HPLC-quality

BP3

OC

water were loaded onto the SPE cartridges. Breakthrough

UV filters in tap water and other clean waters in Spain 2329

Table 3 Mean concentrations (n=3) of the UV filters examined in the water samples analyzed

Sample 4-MBC (ng/L) OD-PABA (ng/L) EHMC (ng/L) OC (ng/L)

Bottled mineral water A n.d. n.d. n.d. n.d.

B n.d. n.d. n.d. n.d.
C n.d. n.d. n.d. n.d.
Tap water n.d. <LOQ 870±46 290±59
Well water n.d. n.d. 770±83 n.d.
Ionic-exchange resins treated water n.d. <LOQ n.d. n.d.

Results are corrected for recoveries and pre-concentration factor. In both cases, the LOQ considered refers to that calculated for tap water, which
was 1.2 pg/mL
n.d. not detected, <LOQ below the method limit of quantification

curves were made, where the volume loaded was repre-
sented against the corresponding recovery rates calculated.
Results showed that the highest recovery was obtained after
the extraction of 200 mL of water sample; this volume was,
therefore, selected for the analysis of real samples.
Method validation
The optimized method was evaluated in terms of linearity;
accuracy, precision, and sensitivity. The quality parameters
of the whole method are listed in Table 2.

Well Water

EHMC

BC

Tap Water
OD- PABA
EHMC
BC OC

BC Standard Mixture

4- MBC OD- PABA

EHMC OC

Fig. 1 GC-MS chromatograms corresponding to well water and tap water samples containing UV filters and to a UV filters mixture standard
solution (10 ng/mL)
2330 M.S. Díaz-Cruz et al.

Analyses of blank samples were performed to test for
interferences or contamination from solvents, chemicals,
glassware, and the target analytes as aforementioned. The
absence of analytical signals corresponding to the target
analytes rejects the possibility of procedural contamination
that could alter the results.
For internal-standard-based quantification, calibration
curves were constructed using a least-square linear
regression analysis. In order to find the linear response
range, injection of standard mixtures of the UV filters
in n-hexane at concentrations from 5 to 5,000 ng/mL were
carried out. The GC-MS method exhibited good linearity
for each individual compound. All the squared correlation
coefficients (r2) were higher than 0.995.
Triplicate samples of 200 mL of a 0.150 ng/mL UV
filters mixture standard solution, in final volume of tap
water, were prepared in order to determine the extrac-
tion efficiency of the method. This concentration level
was selected taking into account the environmental
levels reported in the literature for similar water
matrices. In parallel, a triplicate of the same tap water
sample was analyzed. The mean UV concentration
values obtained were subtracted from the mean UV
filters concentrations found in the spiked water samples
for recoveries determination.
The recoveries of the analytes from the spiked samples
varied depending on the nature of sample water and
compound, and ranged from 73% to 120%. The mean
recovery rates are presented in Table 2. In HPLC-grade
water, the results evidenced high recovery for 4MBC
(which has the lowest logKow), followed by OD-PABA,
while OC (which has the highest logKow) showed the
lowest recovery. The recoveries obtained in tap water
presented similar levels (74–82%) with exception of OC
which showed the highest recovery rate (111%).
The intraday precision was evaluated in terms of relative
standard deviation (RSD) by injecting a UV filter standard
mixture solution at 10 ng/mL on five different days. Values

Table 4 UV filter concentra-

tions in tap water analyzed by Concentration±S.D. (ng/L)
GC-MS analysis
Day BP3 4MBC OD-PABA EHMC OC

Tap 1 1 295±68 33±6 115±11 197±7 167±7

2 55±3 14±4 75±2 256±14 145±55
3 41±6 7.4±0.6 70±1 38±0.4 37±1
4 n.d. n.d. <LOQ 3.2±0.4 39±3
5 n.d. n.d. <LOQ 4.5±0.6 39±1
6 n.d. n.d. <LOQ 4.9±0.7 39±1
7 n.d. n.d. <LOQ 4.1±1.9 35±3
Tap 2 1 98±19 19±3 92±3 55±2 50±2
2 48±10 11±1 71±3 42±3 39±3
3 66±31 12±7 68±2 47±11 41±10
4 n.d. n.d. 2.1±0.7 3.4±0.8 33±0
5 n.d. n.d. <LOQ 2.9±0.9 33±2
6 n.d. n.d. 1.9±0.2 4.8±1.7 37±1
7 n.d. n.d. <LOQ 4.2±0.8 39±5
Tap 3 1 130±14 35±23 114±39 74±3 58±10
2 57±6 14±2 67±27 36±15 42±5
3 70±32 n.d. 6.3±2.0 10±2 38±1
4 <LOQ n.d. <LOQ 4.2±1.3 35±5
5 n.d. n.d. <LOQ 3.2±0.6 37±1
6 35±3 n.d. 3.1±0.9 6.9±0.6 42±7
Ionic-exchange resin-treated water 1 54±13 15±3 79±5 49±4 42±1
2 22±20 10±1 58±21 35±12 39±3
Results are corrected for recov- 3 n.d. n.d. <LOQ 2.1±0.2 34±2
eries and pre-concentration 4 n.d. n.d. <LOQ 1.5±0.4 33±4
factor 5 n.d. n.d. <LOQ 1.7±0.4 34±3
n.d. not detected, LOQ method 6 n.d. n.d. <LOQ 10±1 57±37
limit of quantification
UV filters in tap water and other clean waters in Spain 2331

were below 11% (see Table 2). Interday precision was
assessed by measuring the same standard mixture solution
ninefold during the same day. In this case, the RSD values
were within the range 3%–6%
Instrumental detection limits were calculated from the
injection of a UV filter mixture standard solution at the
lowest concentration used to build the calibration curves.
The values achieved were between 1.9 and 20 pg (see
Table 2).
MLODs and method limits of quantification (MLOQs)
were calculated from the analysis of spiked HPLC-grade and
tap water samples (extracted by SPE as described in the
“Sample preparation” section) as three and ten times the
signal-to-noise ratio, respectively. As listed in Table 2,
MLODs were in the range 0.010–0.98 ng/L in HPLC-grade
water, and between 0.13 and 7.4 ng/L in tap water.
To date, there is no available data in the literature on
MLODs of these compounds in well and bottled mineral
water. However, similar values were obtained for 4MBC and
EHMC in lake [16] and sea water [15], respectively.
Regarding drinking water, Loraine and Pettigrove [22],
reported a MLOD of 0.28 μg/L for EHMC by GC-MS
analysis, which is much higher (50-fold) than that achieved
in this study (5.55 ng/L).
Practical application
In order to test the applicability of the method, the occurrence
of UV filters in different clean water matrices: one well water
sample, three different tap waters, one tap water treated with
exchange resins, and three different trade mark bottled mineral
water samples were determined. Table 3 summarizes the
OC was also detected in tap water at a lower concentration
(290 ng/L). The world-wide approved OC is mainly used in
water-resistant sunscreen products, cosmetics and hair days,
but also imparts outstanding light stability to plastics and other
paintings, being especially suitable for polystyrene and
polyesters. As EHMC, 10% is the maxima concentration
allowed for OC in cosmetics.
The chromatograms corresponding to two contami-
nated samples are represented in Fig. 1. Both, EHMC
and OC according to several in vitro and in vivo studies
elicit anti-estrogenic and very potent anti-androgenic
activity [3, 4, 23–26].
OD-PABA was also found to be present in tap water
and ion-exchange-treated tap water, but below the
MLOQ. 4MBC was not observed above the detection
limits in any water sample analyzed.
This finding encouraged us to perform a more
comprehensive study on the occurrence of these UV
filters in tap water and tap water treated with an ionic-
exchange resin.
Occurrence of UV filters in tap water
A total of 26 tap water samples from Barcelona city and
surroundings were analyzed in this case study. Table 4
and Fig. 2 summarize the results obtained in this study. As
it can be seen, during the first 3 days, waters contained the
0.9
tap 1
0.8
tap 2

mean concentrations found. Recoveries and pre-

0.7
concentration factor were taken into account for concentra- tap 3

tion calculations. Results showed that EHMC (which is one

Acumulated Conc. (ug/L)

0.6 anionic
of the most widely used sunscreen agent in personal care and exchange
industrial products because both protects in the UVA and resin
0.5
UVB regions, and augments the absorbing capacity of other
organic UV filters) was present in well and tap water at 770 0.4
and 870 ng/L, respectively. These high concentration values
for EHMC are in line with that reported by Loraine and 0.3
Pettigrove [22]. In this study, raw and treated drinking water
from four water filtration plants in San Diego (USA) were 0.2
investigated for pharmaceuticals and personal care products.
The concentration of EHMC reported, 450 ng/L, was ca. the 0.1
half of that found here. EHMC is widely used in sunscreen
products and lip balms. The maxima concentration, accord- 0.0
ing to the Spanish legislation, is 10% (w/w), the highest 1
2
percentage allowed for sunscreen agents in cosmetics. The 3
4
5
well and tap water sampling sites selected in the present 6
7 day
study were located in the north area of Barcelona city, where
the water supplied is mainly from Llobregat River, highly Fig. 2 Total UV filter concentrations detected in taps 1, 2, and 3 and
affected by both urban pressure and industrial activities. in water treated with an ionic-exchange resin
2332
highest load of UV filters, with the presence of all of them
at concentrations between 10 and 300 ng/L in tap waters,
and from 2 to 80 ng/L in tap water treated with an ionic-
exchange resin. The other samples were characterized by
the presence of EHMC and OC, at similar mean concen-
trations, around 4 and 40 ng/L, respectively.
Figure 2 shows the accumulated concentration per day
for every sample matrix. A clear pattern of UV filters
contamination was observed. The first 3 days in the four
sampling points the load of UV filters was high with
values in the range 0.8–0.25 μg/L, tending to drop down
and maintain to lower levels, around 0.05 μg/L, associated
to EHMC and OC presence.
Comparison of the results pointed out that tap 1 showed
high concentration levels. For instance, BP3 concentration
was more than threefold that found in the other sampling
points. Therefore, it appears that water from tap 1 was from
a different origin than the rest. This finding is supported by
the fact that to the metropolitan area where tap 1 point is
located, the highest proportion of water to be supplied is
taken from Llobregat River with solely a small contribution
from Ter River. On the contrary, in the two other tap waters
(tap 2 and tap 3) the water supply comes mainly from Ter
River. Therefore, Llobregat River appears to contribute
with a higher load of UV filters than Ter River. This may be
supported by the higher urban and industrial pressure
exerted on Llobregat River, as explained before.
The resin-treated tap water had the same origin as had
tap 1, i.e., mainly from Llobregat River. Thus, the use of
ionic-exchange resins for removing sunscreen agents from
the tap water appears to be effective for all the compounds,
especially the most polar BP3, whose concentration was
reduced to the half.
All these results suggest that conventional water treat-
ment processes appears to be inefficient at removing the
residues of UV filters. Levels of these compounds in
drinking water are unregulated and are not routinely
monitored. Little is known about potential human health
effects associated with chronic exposure to trace levels of
organic wastewater related contaminants through pathways
such as drinking water. Further studies on their removal
efficiency, especially for EHMC during wastewater treat-
ments should be carried out because of the potential risk
posed by this estrogenic compound. Some studies on
mixture effect doses below the no-observed adverse effect
level have been carried out, mainly focused on endocrine
disrupting compounds [27]. Experimental evidence showed
that mixture effects resulted from pollutants that each did
not induce observable effects or produced very small effects
if they were simultaneously present.
The outcomes of our work may serve as a basis to
promote further studies on the persistence/removal of UV
filters, and to develop new water treatment strategies to
M.S. Díaz-Cruz et al.
reduce their occurrence in waters, especially those intended
for drinking water purpose.
Conclusions
The determination of five hormonally active UV filters in
different water matrices was successfully performed by SPE
followed by GC-MS. The optimized method proved to be
sensitive enough for application in environmental studies
(trace levels) of UV filters. The developed method allowed
investigating the occurrence of these compounds in a total
of 34 clean water samples. The results revealed that the UV
filters EHMC and OC were present in measurable quanti-
ties in tap water from Barcelona. BP3, 4MBC and OD-
PABA were detected on certain days as well. On the other
hand, these data suggest that ion-exchange treatment
decreases the concentration of the target UV filters, mainly
BP3 in tap water samples. Measuring the concentration of
these chemicals in tap water will help to guide future
studies on their removal in water treatment processes and
on their ecotoxicity. Therefore, in the design of future
monitoring programs, in relation to the assessment of water
quality, the monitoring of UV filters should be included.
The results here reported constitute the first data on the
occurrence of residues of UV filters in drinking water in
Europe.
Acknowledgments This work has been financially supported by the
Spanish Ministry of Science and Innovation through the projects
CEMAGUA (CGL2007-64551/HID) and SCARCE (Consolider-
Ingenio 2010 CSD2009-00065), and by the Spanish Ministry of
Environment and Rural and Marine Media (MMAMRM-010/PC08/3-
04). Pablo Gago-Ferrero acknowledges Junta para la Ampliación de
Estudios (JAE) his fellowship.
References
1. Ubrach F (2001) J Photochem Photobiol 64:99–104
2. Eriksson E, Andersen HRR, Ledin A (2008) Water Air Soil Pollut
8:473–479
3. Kunz PY, Fent K (2006) Toxicol Appl Pharmacol 217:86–99
4. Kunz PY, Fent K (2006) Aquat Toxicol 79:305–324
5. Fent K, Zenker A, Rapp M (2010) Environ Pollut 158:1817–1824
6. Ciszmas L, McDonald TJ, Phillips TD, Guillespie AM,
Lingenfelter RA, Kubena LF, Phillips TD, Donnelly KC
(2004) Environ Sci Technol 38:5127–5123
7. Coronado M, De Haro H, Deng X, Rempel MA, Lavado R,
Schlenk D (2008) Aquat Toxicol 90:182–187
8. Schlecht C, Klammer H, Jarry H, Wuttke W (2004) Toxicology
205:123–130
9. Danish EPA. http://www.mst.dk/Nyheder/Pelressemeddelelser/
2001/11270000.htm
10. European Commission. Council Directive 76/768/EEC of 27 July
1976 and its successive amendments. http://europa.eu.int/eurlex/
en/lif/reg/en_ergister_133016.html
11. Lambropoulou DA, Giokas DL, Sakkas VA, Albanis TA,
Karayannis MI (2002) J Chromatogr A 967:243–253
UV filters in tap water and other clean waters in Spain
12. Sakkas VA, Giokas DL, Lambropoulou DA, Albanis TA (2003) J
Chromatogr A 1016:211–222
13. Parisis NA, Giokas DL, Vlessidis AG, Evmiridis NP (2005) J
Chromatogr A 1097:17–24
14. Balmer ME, Buser HR, Müller MD, Poiger T (2005) Environ Sci
Technol 39:953–962
15. Giokas DL, Sakkas VA, Albanis TA, Lambropoulou DA (2005) J
Chromatogr A 1077:19–27
16. Poiger T, Buser HR, Balmer ME, Bergqvist PE, Müller MD
(2004) Chemosphere 55:951–963
17. Jeon HK, Chung Y, Ryu JC (2006) J Chromatogr A 1131:192–
202
18. Kawaguchi M, Ito R, Endo N, Sakui N, Okanouchi N, Saito K,
Sato N, Shiozaki T, Nakazawa H (2006) Anal Chim Acta
557:272–277
19. Rodil R, Quintana JB, Mahía PL, Lorenzo SM, Rodríguez DP
(2008) Anal Chem 80:1307–1315
View publication stats
2333
20. Kupper T, Plagellat C, Bärndli RC, De Alencastro LF, Grandjean
D, Tarradellas J (2006) Water Res 40:2603–2612
21. Onesyos KM, Yu JT, Bouuwer EJ (2009) Biodegradation
20:441–466
22. Loraine GA, Pettigrove ME (2006) Environ Sci Technol
40:687–695
23. Schreurs R, Lanser P, Seinen W (2002) Arch Toxicol 76:257–
261
24. Heneweer M, Muusse M, van der Berg M, Sanderson JT (2005)
Toxicol Appl Pharmacol 208:170–177
25. Kunz Y, Galicia HF, Fent K (2006) Toxicol Sci 90:349–361
26. Weisbrod CJ, Kunz PY, Zenker AK, Fent K (2007) Toxicol Appl
Pharmacol 225:255–266
27. Kortenkapm A (2008) Int J Androl 31:233–240
28. SciFinder, American Chemical Society (2008)
29. De Orsi D, Giannini G, Gagliardi L, Porrà R, Berri S, Bolasco A,
Carpini I, Tonelli D (2006) Chromatographia 64:509–515