Vous êtes sur la page 1sur 14

Journal of Hydrology 472–473 (2012) 314–327

Contents lists available at SciVerse ScienceDirect

Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol

Degradation of thiamethoxam and metoprolol by UV, O3 and UV/O3 hybrid


processes: Kinetics, degradation intermediates and toxicity
D. Šojić a, V. Despotović a, D. Orčić a, E. Szabó b, E. Arany b, S. Armaković a, E. Illés b, K. Gajda-Schrantz b,c,d,
A. Dombi b, T. Alapi b,c, E. Sajben-Nagy e, A. Palágyi e, Cs. Vágvölgyi e, L. Manczinger e, L. Bjelica a,
B. Abramović a,⇑
a
University of Novi Sad, Faculty of Sciences, Department of Chemistry, Biochemistry and Environmental Protection, Trg D. Obradovića 3, 21000 Novi Sad, Serbia
b
Research Group of Environmental Chemistry, Institute of Chemistry, University of Szeged, Rerrich Béla tér 1, 6720 Szeged, Hungary
c
Department of Inorganic and Analytical Chemistry, University of Szeged, Dóm tér 7, 6720 Szeged, Hungary
d
EMPA, Laboratory for High Performance Ceramics, Überlandstrasse 129, 8600 Dübendorf, Switzerland
e
Department of Microbiology, University of Szeged, Közép fasor 52, 6726 Szeged, Hungary

a r t i c l e i n f o s u m m a r y

Article history: A comprehensive study of the degradation of thiamethoxam (THIA) and metoprolol (MET) was conducted
Received 3 April 2012 by using UV-induced photolysis (k = 254 nm), ozonation, and a combination of these methods. In order to
Received in revised form 10 September 2012 investigate how molecular structure of the substrate influences the rate of its degradation, we compared
Accepted 18 September 2012
these three processes for the insecticide THIA and the drug MET (a b1-blocker). Of the three treatments
Available online 8 October 2012
This manuscript was handled by Andras
applied, the UV photolysis and the combination of UV/O3 were found to be most effective in the degra-
Bardossy, Editor-in-Chief, with the dation of THIA, while the UV/O3 process appeared to be the most efficient in terms of MET decay. The deg-
assistance of Eddy Y. Zeng, Associate Editor radation kinetics was monitored by LC–DAD, and spectrophotometry, while the mineralization of the
substrates was studied by TOC analysis. Reaction intermediates were studied in detail and a number
Keywords: of them were identified using LC–MS (ESI+/ESI). Both parent compounds showed slight toxic effects
Thiamethoxam towards algae Pseudokirchneriella subcapitata and bacteria Vibrio fischeri. However, the toxicity of the
Metoprolol solutions containing also the degradation intermediates appeared to be much higher for all the test
Advanced oxidation processes organisms. The inhibition/mortality rates were reduced most efficiently by the UV/O3 procedure. Ames
Kinetics test and Comet assay were used to follow the genotoxicity during the degradation of the studied com-
Ecotoxicity pounds. Genotoxic intermediates were frequently detected in the case of MET in the UV treatment alone
Mutagenicity
or in the presence of ozone. Treatments of THIA samples resulted less frequently in genotoxic interme-
diates. To our best knowledge, this work is the first genotoxicological investigation dealing with the pho-
tolytic degradation process of the studied compounds.
Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction health of aquatic organisms (Banks et al., 2005). In developing


countries, both pesticides and feces contaminate drinking water
There are many research articles dealing with the removal of or- (WR, 1998–1999). Thiamethoxam (THIA) belongs to neonicoti-
ganic micropollutants (Hu et al., 2008; Li et al., 2008; Swaim et al., noids, a relatively new class of insecticides, which have been
2008), including pesticides (Rastogi et al., 2009), as well as phar- extensively used for crop protection against a broad spectrum of
maceuticals (Benotti et al., 2009; Vilhunen et al., 2009; Szabó chewing and sucking pests and for a variety of other purposes such
et al., 2011). Pesticides represent a potential threat to humans as as seed and pet treatment (Maienfisch et al., 2001). Metoprolol tar-
they can contaminate water supplies and directly impact the trate salt (MET) is a selective b1-blocker that is used to treat a vari-
ety of cardiovascular diseases, such as hypertension, coronary
artery disease, and arrhythmias (Ikehata et al., 2006). MET is char-
⇑ Corresponding author. Tel.: +381 214852753; fax: +381 21454065.
acterized by an increasing use in recent years and, as a conse-
E-mail addresses: daniela.sojic@dh.uns.ac.rs (D. Šojić), vesna.despotovic@dh.uns. ac.rs
quence, its occurrence in aqueous effluents is expected to
(V. Despotović), dejan.orcic@dh.uns.ac.rs (D. Orčić), eszabo@chem.u-szeged.hu (E. Szabó),
arany.eszter@chem.u-szeged.hu (E. Arany), sanja.armakovic@dh.uns.ac.rs (S. Armaković), increase as well (Alder et al., 2010; Rivas et al., 2010). Advanced
erzsebet.illes@chem.u-szeged.hu (E. Illés), sranc@chem.u-szeged.hu (K. Gajda-Schrantz), oxidation processes (AOPs) have provided a promising alternative
dombia@chem.u-szeged.hu (A. Dombi), alapi@chem.u-szeged.hu (T. Alapi), sajben@gmail. for the remediation of contaminated water when compared to
com (E. Sajben-Nagy), palagyia@freemail.hu (A. Palágyi), csaba@bio.u-szeged.hu (Cs. other treatment methods (Benitez et al., 2002; Chiron et al.,
Vágvölgyi), manczing@bio.u-szeged.hu (L. Manczinger), bjelica@eunet.rs (L. Bjelica),
biljana.abramovic@dh.uns.ac.rs (B. Abramović).
2000; Venkatadri and Peters, 1993). The most efficient and the

0022-1694/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jhydrol.2012.09.038
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 315

most convenient of them involve the use of UV radiation, which The generated HO 
2 reacts then with O3 to produce HO (reac-
promotes the absorbing target molecules to their excited triplet tions (2)–(6)). Accordingly, two main pathways result in the pro-
states. duction of HO during the UV/O3 treatment: photolysis of H2O2
UV-photolysis is widely used for disinfection purposes. It in- (reaction (10)) and HO production via O 3 (reactions (1)–(6)).
volves the interaction of artificial or natural light with the target THIA (Zheng et al., 2006) and MET (Piram et al., 2008; Liu et al.,
compounds, which then induces a series of photochemical reac- 2009) undergo slow direct phototransformation under simulated
tions (Gjessing and Källqvist, 1991; Frimmel, 1994, 1998). The di- sunlight, and UVA, UVB, and simulated sunlight, respectively,
rect photolysis is dependent on the ability of the compounds to whereas both of them undergo slow hydrolysis. MET is readily re-
absorb the emitted light, and on the relevant quantum yields. moved from water by photocatalytic treatment (Yang et al., 2010;
Ozone can also be used for the degradation of organic pollutants Abramović et al., 2011; Romero et al., 2011) and by ozonation
in water. It is a powerful oxidant (electrochemical oxidation poten- (Benner and Ternes, 2009), while the decomposition of THIA has
tial of 2.07 versus 2.8 V for hydroxyl radical (HO)), which is gener- been monitored only in the conditions of direct photolysis
ally produced by an electric discharge in the presence of air or (254 nm) (De Urzedo et al., 2007).
oxygen. A considerable disadvantage of ozone can be its expensive The main objectives of this study were (a) to determine the
production. Similar to other disinfectants for water treatment (e.g. reaction kinetics for THIA (molar absorption maximum at
chlorine or chlorine dioxide), ozone is unstable in water and under- 250 nm) and MET (molar absorption maximum at 225 nm) by
goes reactions with some water matrix components (Von Gunten, using UV-induced photolysis (k = 254 nm), ozonation, and the
2003). However, the unique feature of ozone is its decomposition combination of these methods, (b) to investigate how molecular
into HO radicals which are the strongest oxidants in water (Staeh- structure of the substrate influences the rate of its degradation,
elin and Hoigné, 1985). Two reactions of ozone with dissolved or- bearing in mind their different molar absorption maxima, (c) to
ganic substances can be distinguished in water at low pH: highly identify the major intermediates of both substrates in the three
selective electrophilic attack of molecular ozone, taking place on studied processes, (d) to investigate ecotoxic effects of THIA, MET
the specific part of organic molecules with C@C bonds and/or aro- and their degradation intermediates towards algae Pseudokirch-
matic ring (Von Gunten, 2003), whereas free radicals from ozone neriella subcapitata, zooplankton Daphnia magna and bacteria
decomposition (especially HO) can also react non-selectively with Vibrio fischeri, and (e) to follow the changes of the genotoxic effect
organic compounds (Hoigné and Bader, 1976). Additional free HO during different treatments of the parent molecules.
can be produced in aqueous media from ozone by pH modification
(Doré, 1989) or can be introduced by combining ozone either with
2. Materials and methods
hydrogen peroxide (Duguet et al., 1985) or UV irradiation from e.g.
a low pressure mercury vapor lamp (Staehelin and Hoigné, 1985).
2.1. Chemicals and solutions
When hydroxide ions initiate the reaction, the decomposition of
ozone in water occurs through the following mechanism
All chemicals were of reagent grade and were used without
(Andreozzi et al., 1999):
further purification. The insecticide thiamethoxam {(3-[(2-chloro-
O3 þ HO ! O2 þ HO2 ð1Þ 5-thiazolyl)methyl]tetrahydro-5-methyl-N-nitro-4H-1,3,5-oxadia-
zin-4-imine, CAS No. 153719-23-4, C8H10ClN5O3S, Mr = 291.71,
O3 þ HO2 ! HO2 þ O
3 ð2Þ 99.7% purity} and the drug (±)-metoprolol(+)-tartrate salt {1-[4-
(2-methoxyethyl)phenoxy]-3-(propan-2-ylamino)propan-2-ol tar-
HO2 ¡Hþ þ O
2 ð3Þ trate (2:1), CAS No. 56392-17-7, (C15H25NO3)2 C4H6O6,
Mr = 684.81, P99% purity} were purchased from Sigma–Aldrich;
O 
2 þ O3 ! O2 þ O3 ð4Þ 85% H3PO4 was purchased from Merck; 99.8% OptigradeÒ, HPLC
Gradient Grade acetonitrile (ACN) was a product of Scharlau. The
þ
O 
3 þ H ¡HO3 ð5Þ other chemicals used were: NaH2PO4 (P99%, Spektrum 3D),
Na2HPO4 (P99.0%, Fluka), NaOH (P99.0%, Fluka), NaCl (99.5%,
HO3 ! HO þ O2 ð6Þ Sigma), formic acid (98–100%, Merck), HCl (37%, Scharlau), boric
acid (99.5%, Sigma), Tris (tris(hydroxymethyl)aminomethane,
According to reactions (1) and (2), the initiation of ozone 99.8%, Sigma), and EDTA (99%, Sigma).
decomposition can be artificially accelerated by increasing the Phosphate buffer solution (pH 7.4) was used for toxicity mea-
pH of the solution. surements. The pH of the samples was adjusted with solutions of
UV/O3 treatment is also an AOP and it has been used for practi- NaH2PO4, Na2HPO4, NaOH and HCl.
cal water treatment (Japan Ozone Association, 2004). The advanced Ultrapure water, from a Millipore Milli-Q RG system (Milford,
oxidation effect of UV/O3 treatment is based on the reactions (1)– MA, USA), was used to prepare the solutions in all experiments.
(11) (Beltrán, 2004). In the UV/O3 process, O3 absorbs UV radiation
and produces HO and H2O2 (reactions (7)–(9)), which further
decomposes to HO by direct photolysis (reaction (10)) or dissoci- 2.2. Photochemical reactors
ates to form HO +
2 and H (reaction (11)) (Vilve et al., 2007).
All experiments were carried out using the recirculation photo-
O3 þ hm ! O2 þ O ð7Þ chemical setup shown in Fig. 1. The solution (300 mL) was circu-
lated (375 mL min1) around the low pressure mercury vapor
O þ H2 O ! 2HO ð8Þ lamp covered with a Suprasil sleeve (LightTech, 254/185 nm,
GCL307T5VH/CELL, length 307 mm, 20.5 mm external diameter,
O3 þ H2 O þ hm ! O2 þ H2 O2 ð9Þ 17 W electric power and 4.0 W UV output) in a water-cooled, dou-
ble-walled tubular glass reactor (length 340 mm, inner diameter
H2 O2 þ hm ! 2HO ð10Þ 30 mm).
The solution was continuously stirred with a magnetic bar and
H2 O2 ¡Hþ þ HO2 ð11Þ thermostated at 25 ± 0.5 °C. The UV photon flux determined by
potassium ferrioxalate actinometry (Hatchard and Parker, 1956)
316 D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327

Fig. 1. Experimental apparatus: (1) power supply, (2) Teflon packing ring, (3) low-pressure mercury vapor lamp, (4) envelope (UV photolysis: non-perforated quartz;
ozonation: perforated glass; combination of UV photolysis with ozonation: perforated quartz), (5) reactor, (6) reservoir, (7) magnetic stirrer, (8) pump, (9) thermostat, (10)
oxygen and (11) flow meter.

was I0 = (3.45 ± 0.09)  105 Einstein s1. The photon flux of the 0.8 mL min1) was a mixture of ACN and water (2:8, v/v for THIA
185 nm component of the UV/VUV light (around 6% of the photon and with the following gradient for MET: 0 min 15% ACN which in-
flux of 254 nm light, according to the manufacturers data) was creased to 30% ACN in 5 min, after which 30% ACN was constant for
estimated to be 2.1  106 Einstein s1. Oxygen (Messer, >99.5% 5 min; post time was 3 min), the water being acidified with 0.1%
purity) was bubbled (690 mL min1) through the solution to regu- H3PO4. The use of the gradient mode to follow the degradation
late the dissolved oxygen concentration in the reservoir as well as kinetics of MET was necessary in order to separate the peaks orig-
to generate ozone. For the latter purpose, oxygen was introduced inating from MET and intermediates, and shorten the time of the
between the wall of the lamp and the envelope, which separated LC–DAD analysis. Reproducibility of repeated runs was around
the gas phase and the aqueous solution. The generated ozone 3–10%.
(cgO3 = 3.50  105 mol L1) was then bubbled through the perfora- Absorption spectra were recorded on an Agilent 8453 UV/Vis
tion of the envelope in the solution. Agilent 8453 UV–VIS spectro- spectrophotometer at a fixed slit width (1 nm), using 1 cm quartz
photometer was used to determine the O3 concentration in the gas cells and computer-loaded Agilent UV–Vis software. Kinetics of
phase at 254 nm (e254nm = 2952 mol1 L s1 (Atkinson et al., 1997)), degradation of the oxadiazine ring was monitored at 250 nm (for
using a flow cell. Depending on the material of the envelope, ozon- THIA) and that of aromatic ring transformation at 225 nm (for
ation (perforated glass sleeve), its combination with UV photolysis MET).
(perforated quartz envelope) or UV photolysis only (non-perfo- For the total organic carbon (TOC) analysis, aliquots of 10 mL of
rated quartz sleeve, introducing the oxygen only into the reservoir) the reaction solution were taken at regular time intervals, diluted
of the target compound was investigated (Alapi and Dombi, 2007). to 25 mL and analyzed on an Elementar Liqui TOC II analyzer
In this way, the efficiency of these processes could be compared according to Standard US 120 EPA Method 9060A.
using the same light source. The kinetic measurements were initi-
ated by switching on the light source. Samples were taken at the 2.3.2. Identification of the reaction intermediates
beginning of the experiment and at regular time intervals. All the For the LC–ESI–MS/MS evaluation of intermediates, 20-lL
presented results are the averages of at least three consecutive samples were analyzed on an Agilent Technologies 1200 series
experiments. HPLC with Agilent Technologies 6410A series electrospray ioniza-
tion triple-quadrupole MS/MS, using Agilent Technologies Zorbax
2.3. Analytical procedures XDB-C18 column (50 mm  4.6 mm i.d., particle size 1.8 lm,
40 °C). The mobile phase (flow rate 1.0 mL min1) consisted of
2.3.1. Kinetic studies 0.05% aqueous formic acid and MeOH (gradient, 0 min 30% MeOH,
For the LC–DAD kinetic studies of the THIA and MET degrada- 10–12 min 100% MeOH, post-time 3 min). Analytes were ionized
tion, samples of about 0.50 mL of the reaction solution were taken using an electrospray ion source, a capillary voltage of 4.0 kV and
at the beginning of the experiment and at regular time intervals. A with nitrogen as the drying gas (temperature 350 °C, flow
volume of 20 lL was then injected and analyzed on an Agilent 10 L min1) and nebulizer gas (45 psi). High-purity nitrogen was
Technologies 1100 Series liquid chromatograph, equipped with a used as the collision gas. Full scan mode (m/z range 50–800, scan
UV/Vis DAD set at the absorption maxima (250 nm for THIA and time 100 ms, fragmentor voltage 100 V) in both polarities was used
225 nm MET), and a Zorbax Eclipse XDB-C18 (150 mm  4.6 mm to select precursor ions for starting compound and each degrada-
i.d., particle size 5 lm, 25 °C) column. The mobile phase (flow rate tion intermediate, as well as to examine isotopic peak distribution.
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 317

Then, the product ion scan MS2 mode (fragmentor voltage 100 V, tail. The head is composed of intact DNA, while the tail consists
scan time 200 ms, collision energy 0–30 V in increments of 10 V) of damaged (fragmented) pieces of DNA. The literature describes
was used to elucidate the structure of each degradation the comet assay in an alkaline or in a neutral version. In our study,
intermediate. the alkaline version was used to estimate the DNA damage of
Escherichia coli cells exposed to various treatments. The
2.4. Toxicity tests experiments were performed according to Singh et al. (1988), with
some modifications. Briefly, samples of the treated and untreated
The toxicity towards P. subcapitata (formerly known as Selena- E. coli cells were embedded into low melting point agarose gel.
strum capricornutum) was determined using Algaltoxkit FTM After gelification, the blocks were incubated at 8 °C, in the dark,
(MicroBioTests Inc.) microbiotests according to the ISO Standard in a lysing solution (2.5 mol L1 NaCl, 0.1 mol L1 EDTA,
8692 and OECD Guideline 201. In each case, the cell density was 0.01 mol L1 Tris, 1% Triton X-100, pH 10) for 16 h. Blocks were
calculated from the optical density of the algae in different washed three times (5 min each) in cold mixture of 10 mmol L1
samples, measured in 10 cm path-length cuvettes at 670 nm, using Tris and 1 mmol L1 EDTA buffer (1 mL) and placed onto the lateral
a Hewlett–Packard (HP8452A) diode array spectrophotometer. The surface of the teeth of an electrophoresis comb. Agarose solution
initial cell density was in each case min. 10,000 mL1. The growth (0.76%) was prepared in diluted (10) alkaline buffer and heated
rate (l) was calculated by subtracting the natural logarithm of the for complete homogenization. The comb with the attached sample
number of the cells after 72 h incubation (lnx72) from the natural blocks were fitted in a proper electrophoresis tray and the cooled
logarithm of the number of the cells before the incubation (lnx0) agarose solution was poured in it to cast a gel. After gelification,
and dividing it with the time of incubation (tI) (12). the comb was gently removed, leaving the sample blocks behind
and embedded in the agarose gel slab. The electrophoresis was
lnx72  lnx0
l¼ ð12Þ carried out in an alkaline buffer (12 g NaOH and 2 mL 0.5 mol L1
tI
EDTA per liter deionized water, pH 8.0) at 7 V cm1 for 20 h. After
During the incubation at 21 °C, the samples containing algae electrophoresis, the gel was neutralized by soaking in a solution
were illuminated from the bottom with 6000 lx. The inhibition containing 30 mmol L1 NaCl and 50 mmol L1 Tris–HCl buffer
values were calculated by comparing the growth rate of the algae (pH 6.0). Finally, the gel was stained in ethidium bromide in
in the samples and in the blank solution (l0) (13). Tris–Borate–EDTA buffer pH 8.0 (10 g Tris base, 5.5 g boric acid,
l0  l and 4 mL of 0.5 mol L–1 EDTA per L) and visualized in a UV-
I¼  100 ð13Þ transilluminator.
l0
The toxicity towards D. magna was determined using
Daphtoxkit F™ magna (MicroBioTests Inc.) tests according to the 3. Results and discussion
ISO Standard 6341. The ephippia were used after 72–90 h of
hatching time at 21 °C, applying 6000 lx bottom illumination and 3.1. Degradation of THIA and MET by UV, O3 and UV/O3
after 2 h of pre-nourishment. The zooplankton containing samples
were incubated under the same conditions for 48 h, and 8–10 The results of the comparative study of the degradation of THIA
neonates were added to each sample. The number of the neonates and MET in distilled water under UV, O3 and UV/O3 irradiation for 2
was counted using a magnifier table-gooseneck 3/8 (VWR). and 10 min, respectively, shown in Fig. 2a and b, reflect the effect
Mortality values were given as the percentage of the dead and/or of the structure of the substrate on the rate of its transformation.
immobilized neonates compared to their initial number. The kinetics of the degradation of THIA and MET was followed by
The toxicity towards V. fischeri luminescent bacteria was LC–DAD, as the corresponding instrument was constantly avail-
determined by measuring the natural light emission of these able, whereas the identification of the reaction intermediates was
microorganisms using a HACH-LANGE LUMIStox 3000 lumino- carried out using the more expensive LC–MS/MS instrument. As
meter. The inhibition of the light emission in the presence of the can be seen significant transformation of THIA was observed under
sample was determined against a non-toxic control. NaCl was UV-irradiation, which is understandable bearing in mind its
added to each sample to obtain a 2% solution. The samples were absorption maximum (250 nm) and the emission maximum of
diluted to the half with phosphate buffer (pH 7.4) to reach a pH the UV lamp (k = 254 nm). Namely, the concentration of THIA mea-
value of 7 ± 0.2. Samples containing bacteria were incubated at sured by LC–DAD decreased rapidly in the first 2 min of photolysis,
15 °C for 15 min before the toxicity measurements. Samples for and the transformation efficiency achieved was about 99%. As far
the toxicity tests were taken at the same degradation degree of as MET and similar compounds are concerned, the UV spectra
the parent compounds for the three treatments. show that most of the b-blockers absorb only in the UVC range,
thus they cannot be photodegraded in environmental waters by di-
2.5. Mutagenicity tests rect photolysis (Piram et al., 2008). Moreover, as can be seen in
Fig. 2, the efficiency of MET photolytic transformation achieved un-
2.5.1. Ames test der UVC radiation in the first 10 min was in this case only 7%.
Ames tests were performed basically as described by Maron and As already mentioned, using the combination of ozone with UV
Ames (1983). The tester strains employed were Salmonella radiation, several individual reactions take place: direct ozonation,
typhimurium TA98, TA1535 and TA1537. Strains TA98 and direct photolysis, and oxidation by HO. These reactions may also
TA1537 detect mutagens causing frameshift mutations, while cause synergistic effects. The aim of our adoption of this AOP
TA1535 detects base substitution mutations. was to generate HO, thereby enhancing the oxidation rate of
MET with respect to the single ozonation process. In this case,
2.5.2. Comet assay HO is generated during the (photo)decomposition of ozone. The
Comet assay or Single Cell Gel Electrophoresis (SCGE) assay was combination of ozone with UV radiation resulted in an almost
developed in the late 1980s, and was described first by Singh et al. complete transformation of MET in 10 min, as 84% of the substrate
in 1988, as a fast and effective way of measuring DNA damage in was transformed. In the case of THIA, the 2-min ozonation com-
the individual cells. The name comes from the resulting image, bined with UV-photolysis gave again a high transformation effi-
obtained after the process as a ‘‘comet’’, with a distinct head and ciency of about 98%.
318 D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327

A comparison of the mineralization capacities of UV, O3 and


UV/O3 processes shows that mineralization was slower than
transformation of the initial substrates or of the substances with
oxadiazine/aromatic moiety. Namely, in the case of THIA, TOC
measurements showed that about 40% of organic compounds still
remained in the system after 240 min under UV/O3, 55% in the case
of O3, and in the case of UV even 85%. In the case of MET, after
240 min, the TOC content of the solutions was 15% under UV/O3,
75% in the presence of O3, while under UV, practically, no
mineralization occurred.

3.2. LC–MS/MS identification of degradation intermediates

3.2.1. Thiamethoxam
A total of seven peaks (labeled TP1–TP7) of THIA degradation
intermediates were detected (Table 1 and Fig. 3). Collision-induced
dissociation patterns were less informative than those of MET, and
hence only two intermediates (TP1 and TP5) were completely
identified. For the other compounds, only partial identification of
certain structural features was possible. For all peaks except TP2,
TP4 and TP7, signals could be obtained only in the positive mode,
which could be expected due to the presence of numerous nitrogen
atoms with one electron pair. Retention times of all degradation
intermediates (0.48–0.92 min) were shorter than that of THIA
(1.19 min), indicating their increased polarity due to the molecule
cleavage and/or introduction of polar group.
Peak TP1, eluting at 0.48 min, corresponds to a compound with
a monoisotopic molecular mass (Mmi) of 115 units. Based on the
molecular mass and isotopic peaks profile, it contains an odd num-
Fig. 2. The evolution profiles of the transformation of (a) THIA and (b) MET
ber of nitrogen atoms and no chlorine atoms. In the positive ioni-
(c0 = 0.05 mmol L1) in the presence of UV, O3 and UV/O3, monitored by LC–DAD, zation (PI) MS2 spectrum, consecutive losses of CH2@O (Dm/
SPH and TOC. Treatment time: 2 min for THIA and 10 min for MET. z = 30) and CH2@NH (Dm/z = 29) were observed. The spectral char-
acteristics are in agreement with those previously reported for 5-
methyl-1,3-thiazol-2(3H)-one (De Urzedo et al., 2007) (Fig. 4).
The O3-based processes can be considered as the result of two Peak TP5, eluting at 0.63 min, represents a compound with
different contributions: the direct ozonation reaction and the rad- Mmi = 246, which is by 45 units lower than the molecular mass of
ical pathway (Chang et al., 2009). In the case of THIA removal, the THIA and indicates the replacement of the nitro group with
O3 method was not effective, as only 3.5% degraded during the first hydrogen. The molecular mass also indicates an even number of
2 min of treatment. The slower degradation of THIA under O3 com- nitrogen atoms, while the intensity of the A + 2 isotopic peak
pared to UV/O3 was expected (Xu et al., 2010; Sillanpää et al., confirms that chlorine atom is preserved in the molecule. In the
2011). On the other hand, the accelerated degradation rate under PI MS2 spectrum, several fragment ions were detected: m/z 217
UV/O3 can be explained by UV decomposition of O3, resulting in [M+H30]+, m/z 174 [M+H73]+, m/z 161 [M+HCH2O56]+, and
the formation of H2O2 and HO (reactions (9) and (10)). Concerning m/z 132 [M+H7342]+. The losses of formaldehyde (Dm/z = 30),
MET and based on the limited information in the literature, it CH3N@CHOCH3 (Dm/z = 73), NH@C@NCH3 (Dm/z = 56) and
seems that all b-blockers are fairly reactive with ozone (Ternes, NH@C@NH (Dm/z = 42) all point out to the 3-methyl-1,3,5-
1998; Benner and Ternes, 2009). These pharmaceuticals contain a oxadiazinan-4-imine moiety, i.e. they confirm the removal of the
secondary amine group and a weakly/moderately activated aro- nitro substituent from THIA. Thus, the peak TP5 represents
matic ring that are probable the targets of molecular ozone and 3-[(2-chloro-1,3-thiazol-5-yl)methyl]-5-methyl-1,3,5-oxadiazinan-
HO attacks. During the 10-min ozonation, 27% of MET was 4-imine, a compound already identified as THIA degradation inter-
transformed. mediate (De Urzedo et al., 2007) (Fig. 4).
In order to identify the type of intermediates formed during Peak TP4 (tR = 0.54 min) corresponds to a compound with
degradation, UV spectra were recorded in the range of 200– Mmi = 218. The intensity of A + 2 isotopic peak confirms the loss
400 nm for both THIA and MET. It was found that the degradation of a chlorine atom, while the even Mmi indicates an even number
rate of both substrates measured by spectrophotometry (SPH) was of nitrogen atoms. In contrast to the majority of THIA degradation
lower compared to their transformation rates obtained by LC–DAD intermediates, this compound gave a signal in negative ionization
under UV, O3 and UV/O3 processes (Fig. 2a and b). This is under- (NI), indicating the presence of acidic functional groups. However,
standable if we bear in mind that SPH monitors the kinetics of deg- due to the very simple, non-informative NI MS2 spectrum, featur-
radation of the oxadiazine/aromatic moiety for THIA and MET, ing only two fragment ions at m/z 97 [MH120] and m/z 79
respectively (starting compounds and their degradation intermedi- [MH120H2O], it was not possible to obtain any structural
ates that retain the respective moiety), whereas LC–DAD measures information.
only the changes of the THIA or MET concentration. As can be seen A compound with a high Mmi of 397 units, designated TP6,
from Fig. 2, the UV/O3 system was the most efficient in this case eluted at 0.59 min. On the basis of the molecular mass and isotopic
too. With such technique, 90% of THIA and its intermediates con- peaks, it can be concluded that the compound contains one
taining the oxadiazine ring were transformed after 2 min of treat- chlorine atom and an odd number of nitrogen atoms. The molecu-
ment, while in the case of MET, 50% of the aromatic intermediates lar weight, 106 units higher than that of THIA, indicates the
were transformed in 10 min. introduction of additional groups in molecule. The absence of
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 319

Table 1
MS/MS fragmentation data for THIA degradation intermediates.

Peak label tR (min)a Mmi (g mol1) Mode Precursor ion m/z Vcol (V) Product ions m/z (rel. abundance)
THIA 1.19 291 PI 292 0 292(96), 246(11), 211(100), 210(32), 132(6)
10 211(100), 210(17), 181(24), 180(10), 132(22)
20 211(42), 181(100), 180(15), 152(15), 132(56), 122(23), 69(14)
30 181(68), 152(65), 151(22), 132(100), 125(28), 122(53), 108(32), 71(17), 69(41)
TP1 0.48 115 PI 116 0 116(100), 86(8)
10 116(32), 86(26), 57(100)
20 57(100)
TP2 0.52 195b NI 194 0 196(100)
10 196(95), 161(100)
20 161(100)
TP3 0.53 155b PI 156 0 156(100), 126(22)
10 156(14), 126(100), 98(22), 69(15)
20 126(20), 98(12), 69(100), 55(18)
30 69(100), 55(76)
TP4 0.54 218 NI 217 0 217(100), 97(17)
10 97(100)
20 97(100)
30 97(100), 79(46)
TP5 0.63 246 PI 247 0 247(100), 161(6)
10 247(75), 217(16), 174(19), 161(100), 132(28)
20 174(12), 161(46), 132(100), 57(5)
30 161(12), 133(10), 132(100), 71(9), 57(8)
TP6 0.59 397 PI 398 0 398(100)
10 398(100), 325(31), 208(15), 132(12)
20 398(15), 368(14), 325(47), 208(100), 178(52), 132(99)
30 178(35), 132(100)
TP7 0.92 253 PI 254 0 254(100), 167(21), 138(6)
10 254(7), 181(9), 167(100), 138(78), 117(11)
20 138(100), 87(13)
30 138(100), 111(9)
NI 252 0 252(100), 165(79)
10 252(8), 165(100)
20 165(100)
a
The dead time of the column is 0.42 min.
b
Only one ion was visible in MS1 spectrum, thus Mmi assignation is uncertain.

NO2 loss (Dm/z = 46) early in the fragmentation pathway suggests explanation of the great difference in the mineralization capacities
the possible absence of the nitro group. Consecutive losses of of the UV and UV/O3 processes, although the rate of the degrada-
CH2@O (Dm/z = 30) and CH3N@CH2 (Dm/z = 43) point out to the tion of THIA is practically the same. In the O3-based process
presence of oxadiazinane ring. No further features of molecular (Fig. 3b), only some of the intermediates (TP1–TP4) that were
structure could be determined. identified using the UV and UV/O3 processes formed, and this
Finally, the peak eluting at 0.92 min (TP7) represents a com- was possible at significantly longer treatment times. In this
pound with Mmi = 253, with no chlorine atoms and having an treatment, the intermediate TP4 formed in a significantly higher
odd number of nitrogen atoms. Both PI and NI MS2 spectra contain concentration than in the application of the other two processes,
only a few fragment peaks with significant abundance. The only which indicates that the efficiency of the O3 process in the
reaction observed in NI was the unidentified neutral loss of degradation of this intermediate was significantly lower.
86 mass units. Similar loss, of 87 mass units, as well as an ion of
m/z 87 formed through a complementary fragmentation pathway, 3.2.2. Metoprolol
was observed in the PI chromatograms. As with TP6, no further By the use of the LC–MS/MS technique, a total of 10 peaks (la-
identification was possible. beled MP1–MP10) of MET degradation intermediates were de-
For the peaks TP2 and TP3, it was not possible to unambigu- tected (Table 2 and Fig. 5). Since the collision-induced
ously determine molecular masses nor to obtain relevant dissociation patterns of MET degradation intermediates are well-
structural information from the mass spectra. defined (Borkar et al., 2012), it was possible to identify the majority
As can be concluded from Fig. 3, the intermediates formed from of peaks using PI and NI MS2 spectra. The retention times of inter-
THIA in the UV treatment (Fig. 3a) and in its combination with mediates MP1–MP9 (0.47–1.28 min) were shorter than that of MET
ozonation (Fig. 3c) are the same and of similar concentrations in (1.48 min), which was due to the molecule cleavage and formation
the beginning of the degradation (first 2 min). This could be of polar moieties (alcohols and aldehydes). However, intermediate
expected since the dominant process is direct photolysis, and the MP10 eluted after MET (2.18 min), which indicates its hydropho-
kinetics of THIA degradation is the same as described in Section 3.1. bicity, and which is in line with the assumed structure (additional
However, the efficiency of the degradation of the intermediates i-Pr group).
formed using UV radiation only was significantly lower, which Peak MP1 represents a compound with Mmi = 133, indicating
was especially pronounced in the case of the intermediate TP1, the presence of one nitrogen atom. In the PI MS2 spectrum, a series
whose concentration increased with time in the first 10 min, of fragments were observed, corresponding to the losses of alco-
whereas in the case of the UV/O3 process its concentration began holic OH as water (Dm/z = 18) and of N-bound isopropyl as pro-
to decrease already after 2 min. This is at the same time an pene (Dm/z = 42): 116 [M+HH2O]+, 92 [M+HC3H6]+, 74
320 D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327

Fig. 3. Degradation kinetics of THIA and the appearance/disappearance of the intermediates (TP1–TP7) in the presence of (a) UV, (b) O3 and (c) UV/O3 detected by LC–MS/MS
(ESI+).

3-(isopropylamino)propane-1,2-diol, already identified as a MET


degradation intermediate (Fig. 6) (Abramović et al., 2011).
For the peak MP2 (tR = 1.05 min), it was not possible to obtain
MS2 spectra either in the PI or NI mode. Based on its low and even
monoisotopic molecular mass of 158 units, it could only be con-
cluded, that it represents an intermediate of chain cleavage and
it does not contain nitrogen atom. Also, based on the fact it gives
rise to the [MH] ion, the presence of an acidic group (phenolic
hydroxyl or carboxylic group) could be expected.
The remaining peaks (MP3–MP9, Fig. 6) share the common frag-
mentation pattern (Borkar et al., 2012), featuring loss of propene
(Dm/z = 42) and consecutive losses of isopropylamine and water
(Dm/z = 77) from the protonated molecule, as well as the series
of low molecular mass fragments: 56 (C3H6N), 72 (C3H6NO), 74
(C3H8NO), 98 (C6H12N), 105 (C8H9), 116 (C6H14NO), and 133
(C9H9O). The fragment of 116 m/z value points out to a preserved
2-hydroxy-3-(isopropylamino)propoxy chain. The mass difference
between the [M+H77]+ fragment and [C6H5OC3H4]+ (m/z 133)
Fig. 4. Structural formula of THIA and its degradation intermediates.
thus provides an insight into the changes in the 2-methoxyethyl
side chain (hereafter: chain B).
The peak MP3 (tR = 0.69 min) corresponds to a compound with
[M+HH2OC3H6]+, 72 [CH2@NCH(CH3)2+H]+, 56 [M+H2H2 Mmi = 237, which is by 30 units less than that of MET, and the ob-
OC3H6]+. Based on the fragmentation pattern and literature served loss of CO (161 ? 133) in the MS2 spectrum indicates that
data, it was concluded that the peak MP1 represents the chain B underwent oxidative cleavage, giving rise to a formyl
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 321

Table 2
MS/MS fragmentation data for MET degradation intermediates.

Peak tR Mmi Mode Precursor ion Vcol Product ions m/z (rel. abundance)
label (min)a (g mol1) m/z (V)
MET 1.48 267 PI 268 0 268(100)
10 268(100), 116 (11)
20 268(52), 191(35), 159(74), 133(73), 121(67), 116(100), 98(59), 74(99), 72(85), 56(77)
MP1 0.47 133 PI 134 0 134(100)
10 134(100), 116(18), 92(28), 74(39), 72(24), 56(17)
20 134(8), 92(11), 74(38), 72(19), 57(7), 56(100)
30 74(8), 56(100)
MP3 0.69 237 PI 238 0 238(100)
10 238(100), 196(21), 161(15), 56(12)
20 161(31), 149(44), 133(26), 105(69), 74(32), 72(29), 56(100)
30 105(72), 56(100)
MP4 0.59 239 PI 240 0 240(100)
10 240(100), 198(11), 163(18), 133(20), 116(10), 98(11), 74(9), 72(16), 56(8)
20 163(35), 133(88), 133(93), 107(39), 105(36), 98(23), 74(67), 72(84), 56(100)
30 133(21), 107(22), 105(100), 79(30), 77(39), 74(31), 72(26), 56(79)
MP5 0.80 251 PI 252 0 252(100)
10 252(100), 210(9), 175(12)
20 252(18), 175(33), 163(64), 149(22), 137(59), 133(56), 116(25), 105(16), 98(19), 74(48),
72(40), 56(100)
30 105(17), 91(18), 74(19), 72(15), 56(100)
MP6 0.63 253 PI 254 0 254(100)
10 254(100), 212(10), 177(11)
20 212(14), 177(59), 151(22), 133(26), 116(14), 105(74), 74(24), 72(37), 56(100)
30 121(10), 105(22), 79(22), 56(100)
MP7 0.72 253 PI 254 0 254(100)
10 254(100), 177(20), 72(22)
20 254(72), 177(93), 159(50), 133(86), 98(100), 72(64), 56(89)
MP8 1.28 281 PI 282 0 282(100)
10 282(100), 205(8)
20 282(74), 205(40), 159(68), 145(100), 133(34), 121(37), 116(30), 98(45), 74(60), 72(59),
56(72)
MP9 0.66 283 PI 284 0 284(100)
10 284(100), 116(22), 74(5)
20 284(78), 207(6), 175(8), 163(15), 145(8), 133(43), 116(20), 105(5), 74(100), 72(41), 56(31)
MP10 2.18 325 PI 326 0 326(100)
10 326(100), 284(13), 132(5), 132(6)
20 326(21), 284(19), 191(24), 159(26), 132(100), 114(21), 88(25)
NI 324 0 324(100), 172(8), 151(6)
10 324(43), 172(89), 151(100)
20 172(13), 151(100)
30 151(100), 119(27), 106(26)
a
The dead time of the column is 0.42 min.

group. Thus, the peak MP3 represents 4-[2-hydroxy-3-(isopropyla- or 1-{4-[2-hydroxy-3-(isopropylamino)propoxy]phenyl}ethanone.


mino)propoxy]benzaldehyde. Similarly, the peak MP4 (tR = However, no definite identification is possible at the moment.
0.59 min, Mmi = 239), has a molecular mass which is by 28 units The two peaks of the compounds with Mmi = 253, designated
lower than that of MET and features the loss of HCHO MP6 and MP7, were detected at 0.63 min and 0.72 min. The
(163 ? 133) in the PI MS2 spectrum. This indicates that chain B molecular mass, 14 units lower than that of MET, suggests the
has been converted to hydroxymethyl, i.e. the peak MP4 represents loss of a single CH2 unit. Based on the absence of methanol loss
1-[4-(hydroxymethyl)phenoxy]-3-(isopropylamino)propan-2-ol. (Dm/z = 32) in the PI MS2 spectrum, it can be concluded that both
Both MP3 (Abramović et al., 2011) and MP4 (Benner and Ternes, compounds lack terminal methyl group from the side chain B. For
2009) have already been identified as MET degradation interme- both peaks, consecutive losses of H2O (Dm/z = 18) and C2H2
diates. Spectral characteristics of MP4 are in agreement with (Dm/z = 26) were observed, although the reaction sequences dif-
reference data (Benner and Ternes, 2009). fered, with MP7 dehydrating first (177 ? 159 ? 133) and MP6
The compound MP5, eluting at 0.80 min, has a molecular mass of losing C2H2 first through rearrangement (177 ? 151 ? 133). The
251 units, i.e. 16 units lower than that of MET, suggesting either the isomer with the abundant fragment 159 (MP7) was previously
loss of an oxygen atom (unlikely) or the loss of a CH2 unit, followed identified as 1-[4-(2-hydroxyethyl)phenoxy]-3-(isopropylami-
by a double bond formation. The observed fragmentation pathways no)propan-2-ol (Benner and Ternes, 2009), leading to the
– the loss of ketene (175 ? 133) and HCHO (163 ? 133) – may conclusion that MP6 may be 1-[4-(1-hydroxyethyl)phenoxy]-
suggest that the side chain B has been transformed to either 2- 3-(isopropylamino)propan-2-ol.
oxoethyl- or acetyl-group, i.e. that the compound MP5 is either The compound marked MP8, eluting at 1.28 min, has a molecu-
{4-[2-hydroxy-3-(isopropylamino)propoxy]phenyl}acetaldehyde lar mass which is by 14 units higher than that of MET, implying the
322 D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327

Fig. 5. Degradation kinetics of MET and the appearance/disappearance of the intermediates (MP1–MP10) in the presence of (a) UV, (b) O3 and (c) UV/O3 detected by LC–MS/
MS (ESI+).

Finally, the behavior of the compound eluting at 2.18 min (des-


ignated MP10), with Mmi = 325, differed from the other identified
MET degradation intermediates (Fig. 6). Unlike the other com-
pounds, MP10 gave a signal in NI, indicating the presence of acidic
functional groups (phenolic and/or carboxylic). The only abundant
fragment ions in the NI MS2 spectrum were at m/z 151 and 172, the
former probably being 4-(2-methoxyethyl)phenolate anion, and
the latter representing the complementary ion, formed by the loss
of 4-(2-methoxyethyl)phenol from [MH]. The presence of the ion
m/z 119, formed by the loss of methanol molecule from the ion m/z
151, supports the assumed structure. While a greater number of
abundant ions were observed in the PI MS2 spectrum, the only
structural feature that could be deduced was the N-bound isopro-
pyl group (326 ? 284). Based on the molecular mass, it is possible
that the compound MP10 represents 1-[hydroxyl-4-(2-methoxy-
Fig. 6. Structural formula of MET and its degradation intermediates. ethyl)phenoxy]-3-(diisopropylamino)propan-2-ol. To the best of
our knowledge, no MET degradation intermediate of a molecular
mass 325 has been identified so far.
introduction of one oxygen atom and abstraction of two hydrogen
As can be seen from Fig. 5, the intermediates formed from MET
atoms, i.e. the introduction of an oxo group. The observed consec-
in all three treatments are the same, but only the kinetics of their
utive losses of 46 mass units (formic acid) and 26 mass units (eth-
formation/disappearance is different. This indicates that in a par-
yne) indicate that the terminal methyl group of the chain B was
ticular treatment, in contrast to THIA, there is no significant differ-
oxidized, i.e. the peak MP8 is 2-{4-[2-hydroxy-3-(isopropylami-
ence in the kinetics of the disappearance of MET and formation/
no)propoxy]phenyl}ethyl formate. The lack of methyl (Dm/z = 15)
disappearance of its intermediates.
and the loss of methanol (Dm/z = 32), observed in the MET spectra,
support this assumption.
The peak MP9, eluting at 0.66 min, corresponds to a compound 3.3. Toxicity measurements
with Mmi = 283, i.e. to a monohydroxylated derivative of MET. The
observed loss of methanol (207 ? 175) indicates that the terminal 3.3.1. Toxicity towards algae during the degradation
methoxy group of the chain B is preserved, while the subsequent The toxicity of the parent compounds towards the algae culture
loss of ketene (175 ? 133) indicates that the additional oxygen is was expressed by the normalized inhibition values of the solutions
bound to the chain B. The compound with a hydroxyl in the after 72 h of incubation at 21 °C. THIA itself inhibited slightly the
b-position (relative to the benzene ring) would represent an growth of P. subcapitata, while the effect of MET was not significant
hemiacetal, but since it is very unstable, it is more likely that the (under 20%). However, at the end of the treatment (when the par-
peak MP9 represents an a-hydroxy derivative, i.e. 1-[4-(1-hydro- ent molecules were completely transformed) the toxicity of the
xy-2-methoxyethyl)phenoxy]-3-(isopropylamino)propan-2-ol. solutions was in both cases (and especially in case of MET), higher
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 323

Fig. 7. Normalized inhibition of Pseudokirchneriella subcapitata during the degradation of (a) THIA and (b) MET using UV (D), O3 (s), and UV/O3 (h).

Fig. 8. Mortality rate of Daphnia magna during the degradation of (a) THIA and (b) MET applying UV (D), O3 (s) and UV/O3 (h).
324 D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327

Fig. 9. The inhibition values for Vibrio fischeri during the degradation of (a) THIA and (b) MET applying UV (D), O3 (s) and UV/O3 (h).

than the normalized inhibition values of the initial solutions to the decomposition of the by-products formed from THIA or
(Fig. 7). The likely reason for this is the formation of degradation MET (Figs. 3c and 5c, respectively). However, in the case of MET,
intermediates that are more toxic than the parent compound. the toxicity of the solution was significant even at the end of the
The formation of some aliphatic acids during the UV, O3 and the treatment – most probably because of the presence of some ali-
UV/O3 treatments cannot be excluded too, as it was shown before phatic intermediates formed from the target compound (Alibone
(Alapi et al., 2008, in press). In accordance with these assumptions, and Fair, 1981; Smith et al., 1987; Zhao et al., 1998; Berglind
the pH of the initial solution decreased from pH 7 to 6–5.5 dur- et al., 2010). During the UV/O3 process, the toxicity maxima and
ing the treatments. Although Newsted (2004) reported that these the maximum detected concentration of the major intermediates
values are still within the optimum pH range for algae, according (TP1, TP5 and TP3 of THIA and MP1 of MET) correlated well
to the work of Smith et al. (1987), this change in the pH of the solu- (Figs. 3c, 5c and 7).
tion can significantly affect the growth rate of algae, which sup- During the ozonation, the toxicity was continuously high in par-
ports the results of the present study. allel with the high, only slightly decreasing, concentration of TP1,
During the UV photolysis of THIA the inhibition showed a sharp TP3 and TP4 in the case of THIA, and MP1 in the case of MET
maximum. In all other cases for both compounds it was very broad (Figs. 3b, 5b and 7).
or not present at all (THIA, O3). The highest effect observed during
the UV treatment of THIA indicates the formation of by-products
that are more toxic to P. subcapitata than the parent compounds. 3.3.2. Toxicity towards zooplankton during the degradation
The decreasing toxicity of THIA solutions treated by UV photolysis The toxicity of the parent compounds towards D. magna was ex-
could correlate with the decreasing concentration of TP5 observed pressed by the mortality rate of the zooplankton after 48 h of incu-
by LC–MS/MS (Figs. 3a and 7a). However, the intermediate TP1 bation at 21 °C. From the results it could be concluded that D.
could contribute to the high toxicity of the solution at the end of magna was more sensitive to the degradation by-products and less
the treatment. Among the UV photoproducts of MET, the maximal sensitive to the parent compounds THIA and MET than P. subcapi-
concentration of MP2–MP10 was around the detected toxicity tata (Figs. 7 and 8). The mortality rate increased with the treatment
maximum, and the slow decrease of the algal growth inhibition time in almost all cases, suggesting the presence of toxic by-prod-
could be related to the intermediate MP1, present in the solution ucts (TP1, TP3–TP5 and MP1, Figs. 3 and 5) or a weakly acidic pH
even at the end of the photolysis (Figs. 5a and 7b). (caused by the formation of aliphatic acids), which could have also
It has been shown previously that among the investigated treat- a significant effect on these microcrustaceans (Alibone and Fair,
ments the combined one (UV/O3) was the most efficient in the 1981; Zhao et al., 1998). The only exception was the UV/O3 treat-
mineralization of the target molecules and their degradation inter- ment of MET. After 120 min, the mortality rate decreased signifi-
mediates (Alapi et al., in press). A fast increase of the inhibition was cantly, indicating an efficient mineralization of the toxic
observed in the first 5 and 25 min of the treatment (for THIA and compounds in the solution, which is in good agreement with the
MET, respectively), and the decrease at the end could be ascribed LC–MS/MS data (Figs. 5c and 8b).
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 325

Table 3
Ames test results for THIA degradation (number of revertants/107 cells).

Treatment time (min)


0 0.5 1 2 10 30 60 120 180
UV TA1537 8 14 7 19 8
TA1535 12 61 141 243 152
TA98 2 20 14 5 8
O3 TA1537 8 5 11 6 8
TA1535 14 34 38 33 16
TA98 13 84 26 151 463
UV/O3 TA1537 5 9 14 11 10
TA1535 14 16 49 76 72
TA98 6 14 24 12 9
Fig. 10. Comet assay of THIA and MET samples after UV or UV/O3 treatment. X –
untreated samples (controls).

3.4. Mutagenicity
Table 4
Ames test results for MET degradation (number of revertants/107 cells).
3.4.1. Ames tests
Treatment time (min) Some genotoxic degradation intermediates appeared at signifi-
0 30 60 120 240 cant levels in the course of particular treatments. Treatments of
TA1537 5 679 25 80 7 THIA samples resulted in genotoxic intermediates less frequently
UV TA1535 8 304 281 199 95 compared to MET (Tables 3 and 4). The degradation intermediates
TA98 11 568 450 366 34 of THIA produced revertants, first of all with the TA1535 tester
TA1537 8 5 11 6 8 strain: this mutant detects base substitution mutations. In the case
O3 TA1535 11 28 83 71 47 of MET, the highest level of genotoxic compounds was detected
TA98 7 8 11 17 58
when the UV treatment alone or in the presence of ozone was car-
TA1537 4 49 31 112 49 ried out (Table 4). In this case, the degradation intermediates that
UV/O3 TA1535 7 243 39 149 86
appeared at 30 and 120 min proved to be most genotoxic.
TA98 5 63 8 159 51
To our best knowledge, the present work is the first in which
genotoxicological investigations were performed regarding the
photolytic degradation process of these two compounds, so that
our results cannot be compared with those of other studies.
3.3.3. Toxicity towards bacteria during the degradation
The toxicity of the target compounds towards V. fischeri was ex-
pressed by the inhibition values of the solutions (diluted to one 3.4.2. Comet assay
half with phosphate buffer) after 15 min of incubation at 15 °C. The modified alkaline version of the comet assay used in this
From the results it could be deduced that bacteria were the least study proved to be a reliable approach for studying DNA damage
sensitive to the presence of THIA, MET or their degradation by- caused by the treatment substrate. The UV/O3 treatment generated
products (Fig. 9). a substantial amount of DNA-breaking compounds, first of all at
The curves shown in Fig. 9a have a maximum, indicating the the beginning of the treatments (Fig. 10). Petersen et al. (2000)
formation and decomposition of molecules that exhibit a higher studied the role of different reactive oxygen species formed in
toxic effect than THIA. From these results it could be concluded the UVA irradiation, i.e. the UVA-induced DNA breaks. The UVA
that these compounds were present during the whole UV and O3 irradiation increased the intracellular levels of H2O2, detected by
treatment and that they decomposed most efficiently only in the a fluorescent probe, which caused oxidative damage of DNA, single
combined treatment. The mentioned findings are in agreement strand-breaks, and alkali-labile sites, measured by alkaline single
with the previously presented results (Fig. 3). cell gel electrophoresis (comet assay). The exogenous H2O2 was
In the case of MET, similar tendencies could be observed during also able to induce DNA damage. So, it is very likely that the
the UV/O3 treatment (Fig. 9b). Interestingly, the highest inhibition DNA breaks detected in UV-treated samples in our experiments
value was also detected during the combined treatment, although by comet assay could be in part caused by the presence of H2O2,
the concentration of the most significant degradation by-product in addition to the transformation intermediates. Because the molar
of MET (MP1) was the highest during ozonation (Fig. 5). This expe- absorption coefficient of H2O2 at 254 nm is relatively high
rience underlines again the importance of the possible formation of (18.6 mol1 L cm1) (Nicole et al., 1990) and the quantum yield
aliphatic intermediates, which could also affect the biolumines- of its UV photolysis is 0.98 ± 0.05 (Hunt and Taube, 1952), the
cence of V. fischeri (Berglind et al., 2010). The UV photolysis and harmful effect of H2O2 could be overcome by irradiating the solu-
ozonation did not affect significantly the toxicity of the initial solu- tions with 254 nm light after the end of the UV/O3 treatment,
tion of MET, maybe because they were not efficient enough in the which would lead to the degradation of H2O2 formed during the
mineralization of the target molecules, as it can also be seen from UV/O3 process (Arany et al., 2012).
Fig. 5a and b. However, in the case of THIA, the UV treatment and
the combined method resulted in a higher toxicity compared to 4. Conclusions
ozonation. The inhibition values decreased in parallel with the
decomposition of TP5, indicating probably a higher sensitivity of The results of this study clearly indicate that the concentration
bacteria to the degradation intermediate TP5, formed in a measur- of THIA measured by LC–DAD decreased rapidly in the first 2 min
able concentration during the UV and UV/O3 treatments (Figs. 3 of photolysis, and the transformation efficiency achieved was
and 9a). The same sensitivity towards TP5 could not be detected about 99%. However, with this method only 7% of MET was trans-
either with zooplanktons or algae (Figs. 7a and 8a). formed after the first 10 min, which is understandable bearing in
326 D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327

mind the absorption maximum for THIA (250 nm) and MET Atkinson, R., Baulch, D.L., Cox, R.A., Hampson, R.F., Kerr, J.A., Rossi, M.J., Troe, J., 1997.
Evaluated kinetic, photochemical and heterogeneous data for atmospheric
(225 nm) and the emission maximum of the UV lamp
chemistry. Supplement V. IUPAC subcommittee on gas kinetic data evaluation
(k = 254 nm). The combination of ozone and UV radiation resulted for atmospheric chemistry. J. Phys. Chem. Ref. Data 26, 521–784.
in the 84% efficiency of degradation of MET after 10 min of treat- Banks, K.E., Hunter, D.H., Wachal, D., 2005. Diazinon in surface waters before and
ment, and about 98% of THIA after 2 min. However, in the first after a federally-mandated ban. J. Sci. Total Environ. 350, 86–93.
Beltrán, F.J., 2004. Ozone Reaction Kinetics for Water and Wastewater Systems. CRC
2 min of ozonation, only 3.5% of THIA was transformed. In the case Press, Boca Raton, FL, USA, pp. 193–226.
of MET, whose molecule contains a secondary amine group and a Benitez, F.J., Acero, J.L., Real, F.J., 2002. Degradation of carbofuran by using ozone,
weakly/moderately activated aromatic ring that are probable tar- UV radiation and advanced oxidation process. J. Hazard. Mater. 89, 51–65.
Benner, J., Ternes, T.A., 2009. Ozonation of metoprolol: elucidation of oxidation
gets of molecular ozone and HO, 27% of the substrate was trans- pathways and major oxidation products. Environ. Sci. Technol. 43, 5472–5480.
formed in 10 min of ozonation. The rates of degradation of both Benotti, M.J., Stanford, B.D., Wert, E.C., Snyder, S.A., 2009. Evaluation of a
substrates in all three processes measured by SPH were compared photocatalytic reactor membrane pilot system for the removal of
pharmaceuticals and endocrine disrupting compounds from water. Water Res.
to LC–DAD results, and the mineralization was slower than the 43, 1513–1522.
transformation of the initial substrates or substances having oxadi- Berglind, R., Leffler, P., Sjöström, M., 2010. Interactions between pH, potassium,
azine/aromatic moiety. The LC–ESI–MS/MS data showed that the calcium, bromide and phenol and their effects on the bioluminescence of Vibrio
fischeri. J. Toxicol. Environ. Health, Part A 73, 1102–1112.
degradation of both substrates resulted in the formation of several Borkar, R.M., Raju, B., Srinivas, R., Patel, P., Shetty, S.K., 2012. Identification and
intermediates (7 for THIA and 10 for MET). characterization of stressed degradation products of metoprolol using LC/Q-
As far as toxicity tests are concerned, it can be concluded that TOF-ESI-MS/MS and MSn experiments. Biomed. Chromatogr. 26, 720–736.
Chang, C.C., Chiu, C.Y., Chang, C.Y., Chang, C.F., Chen, Y.H., Ji, D.R., Yu, Y.H., Chiang,
algae and bacteria showed sensitivity towards the parent com-
P.C., 2009. Combined photolysis and catalytic ozonation of dimethyl phthalate
pounds THIA and MET, although their toxic effect was not higher in a high-gravity rotating packed bed. J. Hazard. Mater. 161, 287–293.
than 30%. However, the degradation intermediates (primarily Chiron, S., Fernandez-Alba, A., Rodriguez, A., Garcia-Calvo, E., 2000. Pesticide
TP1, TP3, TP4 and TP5 in case of THIA, and MP1 in case of MET) chemical oxidation: state-of-the-art. Water Res. 34, 366–377.
De Urzedo, A.P.F.M., Diniz, M.E.R., Nascentes, C.C., Catharino, R.R., Eberlin, M.N.,
formed during all treatments, were more toxic. Therefore, the Augusti, R., 2007. Photolytic degradation of the insecticide thiamethoxam in
decomposition of the target molecules, aimed at their complete aqueous medium monitored by direct infusion electrospray ionization mass
mineralization, should be performed with great care. Although spectrometry. J. Mass Spectrom. 42, 1319–1325.
Doré, M., 1989. Chimie des Oxydants et Traitement des Eux. Tec. & Doc, Paris.
the combined UV/O3 treatment was the most effective technique Duguet, J.P., Brodard, E., Dussert, B., Mallevialle, J., 1985. Improvement in the
as far as mineralization is concerned, its advantage from the toxic- effectiveness of ozonation of drinking water through the use of hydrogen
ity point of view is rather small. peroxide. Ozone: Sci. Eng. 7, 241–248.
Frimmel, F.H., 1994. Photochemical aspects related to humic substances. Environ.
The investigations of mutagenicity and DNA breaking revealed Int. 20, 373–385.
that some very harmful degradation intermediates of both THIA Frimmel, F.H., 1998. Impact of light on the properties of aquatic natural organic
and MET are produced in the photolytic processes. In this respect, matter. Environ. Int. 24, 559–571.
Gjessing, E.T., Källqvist, T., 1991. Algicidal and chemical effect of UV-radiation of
new ecotoxicological problems can be encountered, especially in water containing humic substances. Water Res. 25, 491–494.
the degradation of MET. Hatchard, C.G., Parker, C.A., 1956. A new sensitive chemical actinometer. II.
Potassium ferrioxalate as a standard chemical actinometer. Proc. Royal Soc. A.
235, 518–536.
Acknowledgements Hoigné, J., Bader, H., 1976. The role of hydroxyl radical reactions in ozonation
processes in aqueous solutions. Water Res. 10, 377–386.
Hu, Q., Zhang, C., Wang, Z., Chen, Y., Mao, K., Zhang, X., Xiong, Y., Zhu, M., 2008.
This document has been produced with the financial assistance Photodegradation of methyl tert-butyl ether (MTBE) by UV/H2O2 and UV/TiO2. J.
of the European Union (Project HU-SRB/0901/121/116 OCE- Hazard. Mater. 154, 795–803.
EFPTRWR Optimization of Cost Effective and Environmentally Hunt, J.P., Taube, H., 1952. The photochemical decomposition of hydrogen peroxide.
Quantum yields, tracer and fractionation effects. J. Am. Chem. Soc. 74, 5999–
Friendly Procedures for Treatment of Regional Water Resources). 6002.
The contents of this document are the sole responsibility of the Ikehata, K., Naghashkar, N.J., El-Din, M.G., 2006. Degradation of aqueous
University of Novi Sad Faculty of Sciences and can under no cir- pharmaceuticals by ozonation and advanced oxidation processes: a review.
Ozone: Sci. Eng. 28, 353–414.
cumstances be regarded as reflecting the position of the European Japan Ozone Association, 2004. Applications of ozone as advanced oxidation
Union and or the Managing Authority. processes. In: Somiya, I. (Ed.), Ozone Handbook. Sanyusyobo, Yokohama, pp.
317–325.
Li, K., Hokanson, D.R., Crittenden, J.C., Trussell, R.R., Minakata, D., 2008. Evaluating
References UV/H2O2 processes for methyl tert-butyl ether and tertiary butyl alcohol
removal: effect of pretreatment options and light sources. Water Res. 42, 5045–
5053.
Abramović, B., Kler, S., Šojić, D., Laušević, M., Radović, T., Vione, D., 2011.
Liu, Q.-T., Cumming, R.I., Sharpe, A.D., 2009. Photo-induced environmental
Photocatalytic degradation of metoprolol tartrate in suspensions of two TiO2-
depletion processes of b-blockers in river waters. Photochem. Photobiol. Sci.
based photocatalysts with different surface area. Identification of intermediates
8, 768–777.
and proposal of degradation pathways. J. Hazard. Mater. 198, 123–132.
Maienfisch, P., Angst, M., Brandl, F., Fischer, W., Hofer, D.M., Kayser, H., Kobel, W.,
Alapi, T., Dombi, A., 2007. Comparative study of the UV and UV/VUV-induced
Rindlisbacher, A., Senn, R., Steinemann, A., Widmer, H., 2001. Chemistry and
photolysis of phenol in aqueous solution. J. Photochem. Photobiol. A: Chem.
biology of thiamethoxam: a second generation neonicotinoid. Pest Manage. Sci.
188, 409–418.
57, 906–913.
Alapi, T., Gajda-Schrantz, K., Ilisz, I., Mogyorósi, K., Sipos, P., Dombi, A., 2008.
Maron, D.M., Ames, B.N., 1983. Revised methods for the Salmonella mutagenicity
Comparison of UV- and UV/VUV-induced photolytic and heterogeneous
test. Mutat. Res. 113, 173–215.
photocatalytic degradation of phenol, with particular emphasis on the
Newsted, J.L., 2004. Effect of light, temperature, and pH on the accumulation of
intermediates. J. Adv. Oxid. Technol. 11, 519–528.
phenol by Selenastrum capricornutum, a green alga. Ecotoxicol. Environ. Saf. 59,
Alapi, T., Berecz, L., Arany, E., Dombi, A., in press. Comparison of the UV-induced
237–243.
photolysis, ozonation and their combination at the same energy input using a
Nicole, I., De Laat, J., Dore, M., Duguet, J., Bonnel, C., 1990. Utilisation du
self-devised experimental apparatus. Ozone: Sci. Eng.
rayonnement ultraviolet dans le traitement des eaux: measure du flux
Alder, A.C., Schaffner, C., Majewsky, M., Klasmeier, J., Fenner, K., 2010. Fate of b-
photonique par actinometrie chimique au peroxide d’hydrogene. Water Res.
blocker human pharmaceuticals in surface water: comparison of measured and
24, 157–168.
simulated concentrations in the Glatt Valley Watershed, Switzerland. Water
Piram, A., Salvador, A., Verne, C., Herbreteau, B., Faure, R., 2008. Photolysis of b-
Res. 44, 936–948.
blockers in environmental waters. Chemosphere 73, 1265–1271.
Alibone, M.R., Fair, P., 1981. The effects of low pH on the respiration of Daphnia
Petersen, A.B., Gniadecki, R., Vicanova, J., Thorn, T., Wulf, H.C., 2000. Hydrogen
magna Straus. Hydrobiologia 85, 185–188.
peroxide is responsible for UVA-induced DNA damage measured by alkaline
Andreozzi, R., Caprio, V., Insola, A., Marotta, R., 1999. Advanced oxidation processes
comet assay in HaCaT keratinocytes. J. Photochem. Photobiol. B 59, 123–131.
(AOP) for water purification and recovery. Catal. Today 53, 51–59.
Rastogi, A., Al-Abed, S.R., Dionysiou, D.D., 2009. Effect of inorganic, synthetic and
Arany, E., Oppenländer, T., Gajda-Schrantz, K., Dombi, A., 2012. Influence of H2O2
naturally occurring chelating agents on Fe(II) mediated advanced oxidation of
formed in situ on the photodegradation of ibuprofen and ketoprofen. Curr. Phys.
chlorophenols. Water Res. 43, 684–694.
Chem. 2, 286–293.
D. Šojić et al. / Journal of Hydrology 472–473 (2012) 314–327 327

Rivas, F.J., Gimeno, O., Borralho, T., Carbajo, M., 2010. UV-C radiation based methods Venkatadri, R., Peters, R.W., 1993. Chemical oxidation technologies: ultraviolet
for aqueous metoprolol elimination. J. Hazard. Mater. 179, 357–362. light/hydrogen peroxide, Fenton’s reagent, and titanium dioxide assisted
Romero, V., De la Cruz, N., Dantas, R.F., Marco, P., Giménez, J., Esplugas, S., 2011. photocatalysis. Hazard. Waste Hazard. Mater. 10, 107–149.
Photocatalytic treatment of metoprolol and propranolol. Catal. Today 161, 115– Vilhunen, S., Bosund, M., Kääriäinen, M.-L., Cameron, D., Sillanpää, M., 2009. Atomic
120. layer deposited TiO2 films in photodegradation of aqueous salicylic acid. Sep.
Sillanpää, M.E.T., Kurniawan, T.A., Lo, W.H., 2011. Degradation of chelating agents in Purif. Technol. 66, 130–134.
aqueous solution using advanced oxidation process (AOP). Chemosphere 83, Vilve, M., Hirvonen, A., Sillanpää, M., 2007. Ozone-based advanced oxidation
1443–1460. processes in nuclear laundry water treatment. Environ. Technol. 28, 961–968.
Singh, N.P., McCoy, M.T., Tice, R.R., Schneider, E.L., 1988. A simple technique for Von Gunten, U., 2003. Ozonation of drinking water – Part I. Oxidation kinetics and
quantitation of low levels of DNA damage in individual cells. Exp. Cell Res. 175, product formation. Water Res. 37, 1443–1467.
184–191. Word Resources (WR) 1998–1999, Part I: Environmental Change and Human
Smith, P.D., Brockway, D.L., Stancil, F.E., 1987. Effects of hardness, alkalinity and pH Health. <http://www.wri.org/publication/content/8640> (04.09.12).
on the toxicity of pentachlorophenol to Selenastrum capricornutum (Printz). Xu, B., Chen, Z., Qi, F., Ma, J., Wu, F., 2010. Comparison of N-nitrosodiethylamine
Environ. Toxicol. Chem. 6, 891–900. degradation in water by UV irradiation and UV/O3: efficiency, product and
Staehelin, J., Hoigné, J., 1985. Decomposition of ozone in water in the presence of mechanism. J. Hazard. Mater. 179, 976–982.
organic solutes acting as promoters and inhibitors of radical chain reactions. Yang, H., An, T., Li, G., Song, W., Cooper, W.J., Luo, H., Guo, X., 2010. Photocatalytic
Environ. Sci. Technol. 19, 1206–1213. degradation kinetics and mechanism of environmental pharmaceuticals in
Szabó, R., Megyerei, Cs., Mazellier, P., Dombi, A., Gajda-Schrantz, K., 2011. Photolytic aqueous suspension of TiO2: a case of b-blockers. J. Hazard. Mater. 179, 834–
degradation of ibuprofen and ketoprofen. Chemosphere 84, 1658–1663. 839.
Swaim, P., Royce, A., Smith, T., Maloney, T., Ehlen, D., Carter, B., 2008. Effectiveness Zhao, Y.H., Ji, G.D., Cronin, M.T.D., Dearden, J.C., 1998. QSAR study of the toxicity of
of UV advanced oxidation for destruction of micro-pollutants. Ozone: Sci. Eng. benzoic acids to Vibrio fischeri, Daphnia magna and carp. Sci. Total Environ. 216,
30, 34–42. 205–215.
Ternes, A.T., 1998. Occurrence of drugs in German sewage treatment plants and Zheng, L.-Q., Liu, G.-G., Sun, D.-Z., 2006. Study on the hydrolysis and photolysis of
rivers. Water Res. 32, 3245–3260. thiamethoxam. J. Harbin Inst. Technol. 38, 1005–1008.

Vous aimerez peut-être aussi