Degradation of clofibric acid in acidic aqueous medium by electro-Fenton and photoelectro-Fenton

Article abstract of DOI:10.1016/j.chemosphere.2006.07.039

Acidic aqueous solutions of clofibric acid (2-(4-chlorophenoxy)-2-methylpropionic acid), the bioactive metabolite of various lipid-regulating drugs, have been degraded by indirect electrooxidation methods such as electro-Fenton and photoelectro-Fenton with Fe²⁺ as catalyst using an undivided electrolytic cell with a Pt anode and an O₂-diffusion cathode able to electrogenerate H₂O₂. At pH 3.0 about 80% of mineralization is achieved with the electro-Fenton method due to the efficient production of oxidant hydroxyl radical from Fenton's reaction between Fe²⁺ and H₂O₂, but stable Fe³⁺ complexes are formed. The photoelectro-Fenton method favors the photodecomposition of these species under UVA irradiation, reaching more than 96% of decontamination. The mineralization current efficiency increases with rising metabolite concentration up to saturation and with decreasing current density. The photoelectro-Fenton method is then viable for treating acidic wastewaters containing this pollutant. Comparative degradation by anodic oxidation (without Fe²⁺) yields poor decontamination. Chloride ion is released during all degradation processes. The decay kinetics of clofibric acid always follows a pseudo-first-order reaction, with a similar rate constant in electro-Fenton and photoelectro-Fenton that increases with rising current density, but decreases at greater metabolite concentration. 4-Chlorophenol, 4-chlorocatechol, 4-chlororesorcinol, hydroquinone, p-benzoquinone and 1,2,4-benzenetriol, along with carboxylic acids such as 2-hydroxyisobutyric, tartronic, maleic, fumaric, formic and oxalic, are detected as intermediates. The ultimate product is oxalic acid, which forms very stable Fe³⁺-oxalato complexes under electro-Fenton conditions. These complexes are efficiently photodecarboxylated in photoelectro-Fenton under the action of UVA light.

Full text of DOI:10.1016/j.chemosphere.2006.07.039

Chemosphere 66 (2007) 1660–1669
www.elsevier.com/locate/chemosphere
Degradation of clofibric acid in acidic aqueous medium
by electro-Fenton and photoelectro-Fenton
´
´
´
Ignasi Sires, Conchita Arias, Pere Lluıs Cabot, Francesc Centellas, Jose Antonio Garrido,
*
´
´
Rosa Marıa Rodrıguez, Enric Brillas
Laboratori d’Electroquı´mica dels Materials i del Medi Ambient, Departament de Quı´mica Fı´sica, Facultat de Quı´mica,
`
Universitat de Barcelona, Martı´ i Franques 1-11, 08028 Barcelona, Spain
Received 28 April 2006; received in revised form 17 July 2006; accepted 18 July 2006
Available online 30 August 2006
Abstract
Acidic aqueous solutions of clofibric acid (2-(4-chlorophenoxy)-2-methylpropionic acid), the bioactive metabolite of various lipid-reg-
ulating drugs, have been degraded by indirect electrooxidation methods such as electro-Fenton and photoelectro-Fenton with Fe2+ as
catalyst using an undivided electrolytic cell with a Pt anode and an O2-diffusion cathode able to electrogenerate H2O2. At pH 3.0 about
80% of mineralization is achieved with the electro-Fenton method due to the efficient production of oxidant hydroxyl radical from
Fenton’s reaction between Fe2+ and H2O2, but stable Fe3+ complexes are formed. The photoelectro-Fenton method favors the photo-
decomposition of these species under UVA irradiation, reaching more than 96% of decontamination. The mineralization current effi-
ciency increases with rising metabolite concentration up to saturation and with decreasing current density. The photoelectro-Fenton
method is then viable for treating acidic wastewaters containing this pollutant. Comparative degradation by anodic oxidation (without
Fe2+) yields poor decontamination. Chloride ion is released during all degradation processes. The decay kinetics of clofibric acid always
follows a pseudo-first-order reaction, with a similar rate constant in electro-Fenton and photoelectro-Fenton that increases with rising
current density, but decreases at greater metabolite concentration. 4-Chlorophenol, 4-chlorocatechol, 4-chlororesorcinol, hydroquinone,
p-benzoquinone and 1,2,4-benzenetriol, along with carboxylic acids such as 2-hydroxyisobutyric, tartronic, maleic, fumaric, formic and
oxalic, are detected as intermediates. The ultimate product is oxalic acid, which forms very stable Fe3+-oxalato complexes under electro-
Fenton conditions. These complexes are efficiently photodecarboxylated in photoelectro-Fenton under the action of UVA light.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Drug mineralization; Electro-Fenton; Photoelectro-Fenton; Catalysis; Water treatment
1. Introduction
effluents, surface and ground waters and even in drinking
water at concentrations usually ranging from ng lꢀ1 to
lg lꢀ1. The sources of this pollution involve emission from
production sites, direct disposal of overplus drugs in house-
holds, excretion after drug administration to humans and
animals, treatments throughout the water in fish and other
animal farms and inadequate treatment of manufacturing
waste. Among these compounds, clofibric acid (2-(4-chloro-
phenoxy)-2-methylpropionic acid, 1) has long term persis-
tence in the environment. It is the bioactive metabolite of
clofibrate, etofibrate and etofyllineclofibrate, which are
drugs widely used as blood lipid regulators with therapeutic
doses of about 1–2 g dꢀ1 per person, since they decrease the
There is great interest in the environmental relevance of
pharmaceutical drugs and their metabolites as emerging pol-
lutants in waters (Daughton and Jones-Lepp, 2001; Kum-
¨
merer, 2001; Heberer, 2002; Kolpin et al., 2002; Heberer
and Adam, 2004; Weigel et al., 2004; Tauxe-Wuersch
et al., 2005). Different anti-inflammatories, analgesics,
betablockers, lipid regulators, antimicrobials, antiepileptics
and estrogens have been detected in sewage treatment plant
*
Corresponding author. Tel.: +34 93 4021223; fax: +34 93 4021231.
E-mail address: brillas@ub.edu (E. Brillas).
0045-6535/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2006.07.039

I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
1661
Å
plasmatic concentration of cholesterol and triglycerides
(Buser et al., 1998; Tauxe-Wuersch et al., 2005). Concentra-
tions of 1 up to 10 lg lꢀ1 have been detected in sewage treat-
ment plant influents and effluents and in rivers, lakes, North
Sea, ground and drinking waters (Heberer and Stan, 1997;
Buser et al., 1998; Ternes, 1998; Tixier et al., 2003).
To avoid the potential adverse health effects of drugs
and their metabolites as water pollutants on both humans
and animals, research efforts are underway to develop effi-
cient techniques for achieving their total destruction (Zwie-
ner and Frimmel, 2000; Doll and Frimmel, 2004).
However, 1 is poorly degraded by oxidation methods such
as ozonolysis (Ternes et al., 2002; Andreozzi et al., 2003),
H2O2/UV (Andreozzi et al., 2003), sunlight and UV pho-
tolysis (Packer et al., 2003) and TiO2/UV (Doll and
Frimmel, 2004), as well as after application of different bio-
logical and physico-chemical methods in sewage treatment
plants (Tauxe-Wuersch et al., 2005). More potent oxida-
tion procedures are then needed to be applied to destroy
this compound in wastewaters.
Recently, indirect electrooxidation methods such as elec-
tro-Fenton and photoelectro-Fenton are being developed
for wastewater remediation. In these environmentally clean
electrochemical techniques, hydrogen peroxide is continu-
ously generated in an acidic contaminated solution from
the two-electron reduction of O2 at reticulated vitreous car-
bon (Xie and Li, 2006), mercury pool (Ventura et al.,
2002), carbon-felt (Oturan et al., 1999; Go¨zmen et al.,
2003; Hanna et al., 2005; Irmak et al., 2006) and O2-diffu-
Reaction (3) also enhances the production of OH and
hence, the mineralization of organics.
The electro-Fenton treatment of 100-ml solutions with
1 mM of 1 and 2 mM Fe2+ in 0.01 M HCl has been previ-
ously reported by Oturan et al. (1999), but these authors
only described its decay kinetics and the detection of some
initial aromatic products. To show the possible viability of
the electro-Fenton and photoelectro-Fenton methods to
remove this metabolite in wastewaters, we have carried
out a detailed study on the degradation of acidic aqueous
solutions of 1 in the pH range 2.0–6.0 using 1.0 mM Fe2+
in both procedures. Comparative treatments in the absence
of this catalyst were also made to demonstrate the positive
Å
oxidation action of OH formed from reaction (2). The
effect of current density and clofibric acid concentration
on the degradation process and current efficiency was
explored. Aromatic products were identified by gas chro-
matography–mass spectrometry (GC–MS). The decay of
1 and the evolution of its by-products were followed by
chromatographic techniques. The results obtained in this
study are reported herein.
2. Experimental
2.1. Chemicals
Clofibric acid (1), 4-chlorophenol (2), hydroquinone (3),
4-chlororesorcinol (4), p-benzoquinone (6), 1,2,4-benzenet-
riol (7), 2-hydroxyisobutyric acid (9), tartronic acid (10),
maleic acid (11), fumaric acid (12), formic acid (13) and
oxalic acid (14) were either reagent or analytical grade from
Sigma-Aldrich, Merck, Panreac and Avocado. 4-Chloro-
catechol (5) was synthesized by chlorination of pyrocate-
chol with SO2Cl2 at room temperature, as reported
elsewhere (Boye et al., 2002). Analytical grade sulfuric acid
was purchased from Merck. Anhydrous sodium sulfate
and heptahydrated ferrous sulfate were analytical grade
from Fluka. All solutions were prepared with pure water
obtained from a Millipore Milli-Q system with resistivity
>18 MX cm at 25 ꢁC. Organic solvents and other chemicals
employed were either HPLC or analytical grade from
Panreac.
´
sion (Boye et al., 2002; Brillas et al., 2004a,b; Sires et al.,
2006) cathodes:
O2 + 2Hþ + 2eꢀ ! H2O2
ð1Þ
The oxidizing power of H2O2 is enhanced in the electro-
Fenton method by adding small amounts of Fe2+ as cata-
lyst to the acidic solution. Hydroxyl radical (ÅOH) and Fe3+
are then generated from the classical Fenton’s reaction
between Fe2+ and H2O2 (Sun and Pignatello, 1993):
Fe2þ + H2O2 ! Fe3þ + ÅOH + OHꢀ
ð2Þ
Reaction (2) is propagated from Fe2+ regeneration,
which mainly occurs by reduction of Fe3+ species at the
Å
cathode (Oturan et al., 1999). OH acts as a non-selective,
2.2. Apparatus and analysis procedures
strong oxidant, with ability to react with organics yielding
dehydrogenated or hydroxylated derivatives until their
overall mineralization (conversion into CO2, water and
inorganic ions). In the photoelectro-Fenton process the
solution is irradiated with UVA light to favor: (i) the pho-
todecomposition of complexes of Fe3+ with generated car-
The solution pH was measured with a Crison 2000 pH-
meter. Electrolyses were carried out at a constant current
density (j) of 33, 100 and 150 mA cmꢀ2 with an Amel
2053 potentiostat–galvanostat. All samples withdrawn
from treated solutions were filtered with Whatman
0.45 lm PTFE filters before analysis. The mineralization
of each solution of 1 was monitored by the abatement of
its total organic carbon (TOC), determined on a Shimadzu
VCSN TOC analyzer. Reproducible TOC values were
obtained from analysis of 100-ll aliquots using the stan-
dard non-purgeable organic carbon method. From these
data, the mineralization current efficiency (MCE) for
´
boxylic acids (Zuo and Hoigne, 1992; Brillas et al., 2004a,b;
2+
´
Sires et al., 2006) and (ii) the regeneration of Fe from
additional photoreduction of Fe(OH)2+, which is the pre-
dominant Fe3+ species in acid medium (Sun and Pignatello,
1993):
Fe(OH)2þ + hm ! Fe2þ + ÅOH
ð3Þ

1662
I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
electrolyzed solutions at a given time t was calculated as
follows:
was 35 ꢁC for 2 min, 10 ꢁC minꢀ1 up to 250 ꢁC and hold
time 15 min, and the temperature of the inlet, transfer line
and detector was always 250 ꢁC.
DðTOCÞexp
DðTOCÞtheor
MCE ¼
ꢁ 100
ð4Þ
2.3. Electrolytic system
where D(TOC)exp is the experimental TOC removal and
D(TOC)theor is the theoretically calculated TOC decay
assuming that the applied electrical charge (=current ·
time) is only consumed in the mineralization process of 1.
The decay of 1 and the evolution of aromatic intermedi-
ates were followed by reversed-phase chromatography
using a Waters 600 HPLC liquid chromatograph fitted with
a Spherisorb ODS2 5 lm, 150 · 4.6 mm (i.d.), column at
room temperature, and coupled with a Waters 996 photo-
diode array detector, controlled through a Millennium-32ꢂ
program. For each compound, this detector was selected at
the maximum wavelength of its UV-absorption band.
These analyses were made by injecting 20-ll aliquots into
the chromatograph and circulating a 50:47:3 (v/v/v) meth-
anol/phosphate buffer (pH = 2.5)/pentanol mixture at
1.0 ml minꢀ1 as mobile phase. Generated carboxylic acids
were followed by ion-exclusion chromatography by inject-
ing 20-ll samples into the above HPLC chromatograph
fitted with a Bio-Rad Aminex HPX 87 H, 300 · 7.8 mm
(i.d.), column at 35 ꢁC. For these measurements, the photo-
diode detector was selected at 210 nm and the mobile phase
was 4 mM H2SO4 at 0.6 ml minꢀ1. Clꢀ concentration in
electrolyzed solutions was determined by ion chromatogra-
phy using a Shimadzu 10Avp HPLC chromatograph fitted
with a Shim-Pack IC-A1S, 100 · 4.6 mm (i.d.), anion col-
umn at 40 ꢁC and coupled with a Shimadzu CDD 10Avp
conductivity detector. These measurements were carried
out with a 2.5 mM phtalic acid and 2.4 mM tris(hydroxy-
methyl)aminomethane solution of pH 4.0 as mobile phase
All electrolyses were conducted in an open, undivided
and thermostated cylindrical cell containing 100 ml of solu-
tion stirred with a magnetic bar. The anode was a 3-cm2 Pt
sheet of 99.99% purity from SEMPSA and the cathode was
a 3-cm2 carbon-PTFE electrode from E-TEK, which was
fed with pure O2 at 12 ml minꢀ1 to generate continu-
ously H2O2 from reaction (1). The electrolytic setup and
the preparation of the O2-diffusion cathode have been
described (Boye et al., 2002). For the experiments with
UVA irradiation, a Philips 6 W fluorescent black light blue
tube was placed at 7 cm above the solution. The tube emit-
ted UVA light in the wavelength region between 300 and
420 nm, with kmax = 360 nm, supplying a photoionization
energy input to the solution of 140 lW cmꢀ2, detected with
a NRC 820 laser power meter working at 514 nm. Both
electro-Fenton and photoelectro-Fenton treatments were
performed after addition of 1.0 mM Fe2+ to the solution.
All trials were carried out at 35.0 ꢁC, which is the maxi-
mum temperature to work with the open electrolytic cell
without significant water evaporation from solution (Boye
et al., 2002).
3. Results and discussion
3.1. Comparative degradation
Comparative electrolyses at 100 mA cmꢀ2 for 6 h were
initially made for solutions containing 179 mg lꢀ1 of 1
(equivalent to 100 mg lꢀ1 of TOC) and 0.05 M Na2SO4
regulated with H2SO4 at pH 3.0 and at 35.0 ꢁC. In these
experiments the solution pH remained practically constant,
reaching final values between 2.8 and 3.0. The change in
at 1.5 ml minꢀ1
.
A 100 ml-solution with 179 mg lꢀ1 of 1 of pH 3.0 was
electrolyzed at 100 mA cmꢀ2 and at 35.0 ꢁC by electro-
Fenton with 1.0 mM Fe2+ for 2 min. The resulting organics
were extracted with 45 ml of CH2Cl2 in three times. The
collected organic solution was dried with anhydrous
Na2SO4, filtered and evaporated to about 2 ml. The
remaining products were separated and identified by
GC–MS using a Hewlett-Packard 5890 Series II gas chro-
matograph fitted with a HP-5 0.25 lm, 30 m · 0.25 mm
(i.d.), column, and a Hewlett-Packard 5989 A mass
spectrophotometer operating in EI mode at 70 eV. The
temperature ramp for this column was 35 ꢁC for 2 min,
10 ꢁC minꢀ1 up to 320 ꢁC and hold time 5 min, and the
temperature of the inlet, transfer line and detector was
250 ꢁC, 250 ꢁC and 290 ꢁC, respectively. To identify the
final carboxylic acids, the above solution was treated under
the same electro-Fenton conditions for 6 h. The resulting
solution was evaporated at low pressure and the remaining
solid was dissolved in 2 ml of ethanol. The esterified acids
were further analyzed by GC–MS using the gas chromato-
graph fitted with a HP-INNOWax 0.25 lm, 30 m ·
0.25 mm (i.d.), column. In this case the temperature ramp
solution TOC with consumed specific charge (Q, in A h lꢀ1
)
for such trials is depicted in Fig. 1a. Note that the electro-
lytic system produces continuouslyÅ hydrogen peroxide
from reaction (1), whereas adsorbed OH is formed at the
Pt anode from water oxidation (Boye et al., 2002; Brillas
et al., 2004a,b; Sire´s et al., 2006):
Å
H2O ! OHads + Hþ + eꢀ
ð5Þ
The use of this system without catalyst corresponds to
the method of anodic oxidation with electrogenerated
H2O2. Fig. 1a shows that this procedure gives a quite slow
TOC removal, attaining 41% of mineralization at 6 h
(Q = 18 A h lꢀ1). This behavior can be accounted for by
the low concentration of ÅOH formed at the Pt surface from
reaction (5), which is the main oxidant of 1 and its by-prod-
ucts. A similar degradation rate can be seen in Fig. 1a when
the solution without catalyst is illuminated with UVA light,
yielding 39% of TOC decay at the end of electrolysis. This
brings to consider that organics are not directly photode-

I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
1663
(Q = 2 A h lꢀ1), whereupon it undergoes a slight drop due
to its oxidation to Cl2 at the Pt anode. These findings indi-
cate that chloro-organics are always degraded with release
of Clꢀ, although they are much more rapidly destroyed in
the two last methods. However, all procedures only lead
to the release of 78–85% of the initial chlorine content of
1 (29.5 mg lꢀ1), suggesting that stable colored polyaromat-
ics formed during degradation contain the remaining
chlorine.
120
100
80
60
40
20
0
a
These results indicate that electro-Fenton only yields
partial decontamination of 1, whereas this pollutant can
be almost completely mineralized by photoelectro-Fenton.
For this last technique, the effect of pH, current density and
metabolite concentration on its oxidizing power was inves-
tigated to clarify its optimum operative conditions.
b
25
20
15
10
5
3.2. Effect of experimental parameters on the
photoelectro-Fenton process
0
0
3
6
9
12
15
18
21
Q / A h l-1
The TOC–Q plots obtained for solutions of 179 mg lꢀ1
of 1 in the pH range 2.0–6.0 degraded by photoelectro-
Fenton at 100 mA cmꢀ2 are depicted in Fig. 2a. The pH
of solutions with initial pH 4.0 and 6.0 underwent a pro-
gressive decrease with time, mainly during the first hour
of electrolysis, due to the generation of acid products and
for this reason, it was continuously regulated within a range
of 0.3 units by adding 1 M NaOH. Fig. 2a shows that the
quickest TOC decay takes place starting from pH 3.0,
whereas for the other solutions, the degradation rate falls
in the order pH 2.0 > pH 4.0 ꢂ pH 6.0. This behavior can
be assocÅiated with the highest generation rate of the main
oxidant OH from Fenton’s reaction (2), since its optimum
pH is 2.8 (Sun and Pignatello, 1993), very close to pH 3.0
where 1 and its by-products are more rapidly destroyed.
The influence of current density on the oxidation
ability of this method was examined by electrolyzing solu-
tions with 179 mg lꢀ1 of 1 of pH 3.0 at 33, 100 and
150 mA cmꢀ2. As can be seen in Fig. 2b, a progressive
increase in Q from 7 to 27 A h lꢀ1 for achieving total
decontamination takes place when j increases. However,
the time needed for overall mineralization drops from 7 h
at 33 mA cmꢀ2 to 5.5 h at 150 mA cmꢀ2. The faster miner-
alization rate with time when j raises can be ascribed to a
Fig. 1. (a) Total organic carbon and (b) concentration of accumulated
chloride ion vs. specific charge for the degradation of 100-ml solutions
with 179 mg lꢀ1 clofibric acid (1) and 0.05 M Na2SO4 of pH 3.0 at
100 mA cmꢀ2 and at 35.0 ꢁC using an undivided cell with a 3-cm2 Pt anode
and a 3-cm2 carbon-PTFE O2-diffusion cathode. Method: (d) anodic
oxidation with electrogenerated H2O2; (j) anodic oxidation with electro-
generated H2O2 under a 6-W UVA irradiation with kmax = 360 nm; (m)
electro-Fenton with 1.0 mM Fe2+ in the solution; (r) photoelectro-
Fenton with 1.0 mM Fe2+ and UVA light.
composed by UVA light. A different behavior can be
observed in Fig. 1a when 1.0 mM Fe2+ is added as catalyst.
For the electro-Fenton process, TOC is rapidly reduced by
79% at 6 h due to the fast reaction of organics with the great
Å
amounts of OH produced from Fenton’s reaction (2). In
contrast, the photoelectro-Fenton process leads to quicker
TOC decay with almost overall mineralization (>96%
TOC removal) at the end of electrolysis. This trend can be
related to: (i) the rapid photolysis of some stable complexes
of Fe3+ with generated carboxylic acids under electro-Fen-
´
ton conditions (Zuo and Hoigne, 1992; Brillas et al., 20Å04a;
´
Sires et al., 2006) and/or (ii) the enhanced generation of OH
from additional photoreduction of Fe(OH)2+ from reaction
(3). Note that the starting pale yellow solution changed to
pale orange color at the end of both electro-Fenton and
photoelectro-Fenton degradations. This is indicative of
the formation of soluble colored polyaromatics in small
extent, which can be not be destroyed by oxidant ÅOH pro-
duced by reactions (2), (3) and (5).
Å
greater production of OH at the Pt anode from reaction
(5) and in the medium from reaction (2) due to the electro-
generation of more H2O2 by the O2-diffusion cathode from
reaction (1) (Brillas et al., 2004a). The increase in Q for
total decontamination under these conditions is indicative
Å
Mineralization of 1 is accompanied by its overall dechlo-
rination. Ion chromatograms for the above treated solu-
tions only displayed a defined peak related to Clꢀ ion. No
other chlorine–oxygen ions such as ClOꢀ3 and ClO4ꢀ were
detected by this technique. Fig. 1b shows a gradual accumu-
lation of Clꢀ up to 23 mg lꢀ1 for 4 h (Q = 12 A h lꢀ1) by
anodic oxidation with electrogenerated H2O2, whereas for
electro-Fenton and photoelectro-Fenton, a Clꢀ concentra-
tion of about 25 mg lꢀ1 is already attained at 40 min
of a slower relative generation of oxidant OH due to the
acceleration of non-oxidizing reactions of this radical, for
example, its oxidation to O2 at the Pt anode and its recom-
bination into H2O2.
The great oxidizing power of the photoelectro-Fenton
method was confirmed by degrading up to 0.56 g lꢀ1 (close
to saturation) of 1 at pH 3.0 and at 100 mA cmꢀ2. The
TOC–Q plots thus obtained are shown in Fig. 2c. As can
be seen, more than 96% of mineralization is achieved after

1664
I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
120
100
80
60
40
20
0
C10H11ClO3 + 17H2O ! 10CO2 + Clꢀ + 45Hþ + 44eꢀ ð6Þ
a
b
c
Taking into account reaction (6) to calculate the theoret-
ical TOC removal, the mineralization current efficiency of
electrolyzed solutions was determined from Eq. (4). The
MCE values thus obtained for the different treatments
reported in Fig. 1a are depicted in Fig. 3a. The efficiency
for both anodic oxidation procedures is very small, reach-
ing a maximum value of 3.3–3.8% at 2 h (Q = 6 A h lꢀ1), as
expected from their low oxidation ability. In contrast, this
parameter attains a value of 25% and 23% at the early
stages (20 min) of electro-Fenton and photoelectro-Fenton
processes, respectively. When electrolysis is prolonged, a
dramatic drop in MCE can be observed in Fig. 3a for such
treatments, indicating the geneÅration of products that are
more difficultly oxidized with OH than the initial pollu-
tant. The efficiency for the photoelectro-Fenton method
is clearly higher from 1 h of electrolysis (Q = 3 A h lꢀ1),
because it is able to destroy most products, including com-
plexes of Fe3+ with generated carboxylic acids that are sta-
ble under electro-Fenton conditions.
For the photoelectro-Fenton treatment at the optimum
pH 3.0, the MCE value always decays with rising cur-
rent density, as can be seen in Fig. 3b. For example, after
1 h of electrolysis of 179 mg lꢀ1 of 1, decreasing efficiencies
of 46% (Q = 1 A h lꢀ1), 20% (Q = 3 A h lꢀ1) and 14%
(Q = 4.5 A h lꢀ1) are found at increasing j values of 33,
100 and 150 mA cmꢀ2, respectively. This tendency corrob-
orates the enhancement of parallel non-oxidizing reactions
of ÅOH (e.g., its anodic oxidation to O2 and its recombina-
100
80
60
40
20
0
300
250
200
150
100
50
0
0
3
6
9
12 15 18 21 24 27 30 33
Q / A h l-1
Fig. 2. Effect of experimental parameters on TOC abatement vs. specific
charge for the treatment of 100 ml of different clofibric acid (1) solutions
with 1.0 mM Fe2+ at 35.0 ꢁC by photoelectro-Fenton. In plot (a),
concentration of 1: 179 mg lꢀ1; initial solution pH: (d) 2.0; (r) 3.0; (j)
4.0; (m) 6.0; current density: 100 mA cmꢀ2. In plot (b), concentration of 1:
179 mg lꢀ1; solution pH: 3.0; current density: (d) 33; (r) 100; (j)
150 mA cmꢀ2. In plot (c), initial concentration of 1: (d) 557 (close to
saturation); (j) 358; (r) 179; (m) 89 mg lꢀ1; solution pH: 3.0; current
30
25
a
20
15
10
5
density: 100 mA cmꢀ2
.
consumption of 30 A h lꢀ1 (10 h), 24 A h lꢀ1 (8 h),
18 A h lꢀ1 (6 h) and 15 A h lꢀ1 (5 h) for 557, 358, 179 and
89 mg lꢀ1 of 1, respectively. The drop in Q with decreasing
clofibric acid concentration could be simply associated with
the presence of lower amount of organics. However, results
of Fig. 2c evidence the removal of more TOC at a given
time with increasing initial pollutant content. For example,
after 2 h of electrolysis, TOC is reduced by 33, 81, 143 and
218 mg lꢀ1 starting from 89, 179, 358 and 557 mg lꢀ1 of 1,
respectively. Since the same production of OH is expected
from reactions (2), (3) and (5) in these trials, it seems plau-
sible to consider that its competitive non-oxidizing reac-
tions become slower and more OH concentration can
then react with pollutants.
0
50
b
40
30
20
10
0
Å
Å
0
3
6
9
12 15 18 21 24 27 30 33
Q / A h l-1
Fig. 3. Dependence of mineralization current efficiency calculated from
Eq. (4) on specific charge for the degradation of 100 ml of clofibric acid (1)
solutions of pH 3.0 at 35.0 ꢁC. Plot (a) corresponds to the different
treatments shown in Fig. 1a at 100 mA cmꢀ2. Plot (b) corresponds to the
photoelectro-Fenton treatment of: (d) 557; (j) 358; (s, r, n) 179; (m)
89 mg lꢀ1 of 1 with 1.0 mM Fe2+ at (s) 33; (d, j, Ç, m) 100; (n)
3.3. Mineralization current efficiency
The mineralization of 1 yields carbon dioxide and Clꢀ as
final products. The overall reaction can be written as
follows:
150 mA cmꢀ2
.

I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
1665
tion into H2O2) when j raises, yielding a smaller proportion
200
160
120
80
4
3
2
1
0
a
of this oxidant with ability to destroy pollutants. Fig. 3b
also illustrates that the efficiency for the photoelectro-
Fenton degradation at pH 3.0 and at 100 mA cmꢀ2
undergoes a progressive increase with rising metabolite
concentration, indicating the removal of larger amounts
Å
of organics with OH because its non-oxidizing reactions
0
30
60
90 120 150
40
become slower. Thus, the MCE values at 1 h
(Q = 3 A h lꢀ1) are 8.2%, 20%, 32% and 45% for 89, 179,
358 and 557 mg lꢀ1 of 1, respectively. Under these condi-
tions, the highest efficiency of 50% is obtained at the begin-
ning (20 min) of the degradation of the more concentrated
solution.
The above results allow concluding that the photoelec-
tro-Fenton method is viable for treating acidic wastewaters
containing clofibric acid up to close saturation at optimum
pH 3.0. This technique becomes more efficient when the
content of this pollutant increases and current density
decreases.
time / min
0
600
500
400
300
200
100
0
0
60
120
180
240
300
4
b
3
2
1
0
0
2
4
6
8
10 12
time / min
0
4
8
12
16
20
3.4. Clofibric acid decay and kinetic analysis
time / min
Fig. 4. Clofibric acid (1) decay with electrolysis time at pH 3.0 and at
35.0 ꢁC. Plot (a) presents the degradation of 100 ml of a solution with
179 mg lꢀ1 of 1 at 100 mA cmꢀ2 by (d) anodic oxidation with electro-
generated H2O2; (j) anodic oxidation with electrogenerated H2O2 and
UVA light. Plot (b) shows the treatments by: (Æ) electro-Fenton of the
same solution with 1.0 mM Fe2+ at 100 mA cmꢀ2; photoelectro-Fenton
of: (d) 557; (j) 358; (s, Ç, n) 179; (m) 89 mg lꢀ1 of 1 with 1.0 mM Fe2+
at (s) 33; (d, j, Ç, m) 100; (n) 150 mA cmꢀ2. The corresponding kinetic
analysis assuming a pseudo-first-order reaction for 1 is given in the inset
panels.
Å
The kinetics of the reaction between 1 and OH gener-
ated in the different methods tested was comparatively
studied by electrolyzing 179 mg lꢀ1 of this compound at
pH 3.0, at 100 mA cmꢀ2 and at 35.0 ꢁC. Its concentration
was determined by reversed-phase chromatography, where
it exhibits a well-defined absorption peak with a retention
time (tr) of 7.9 min. As can be seen in Fig. 4a, the concen-
tration of 1 undergoes a similar fall by anodic oxidation
without and with UVA irradiation, disappearing from
the medium in 240 min in both cases. This confirms that
1 is not directly photolyzed by UVA light. Good straight
lines were obtained when the above concentration decays
were fitted to a pseudo-first-order kinetic equation, as
depicted in the inset of Fig. 4a. From this analysis, an aver-
age pseudo-first-order rate constant (k) of (4.7 0.1) ·
10ꢀ4 sꢀ1 (square regression coefficient, R2 = 0.991) is found
for both anodic oxidation treatments. This behavior sug-
100 mA cmꢀ2 is achieved at longer time when its initial
concentration rises. Thus, it disappears after 3, 7, 12 and
18 min for 89, 179, 358 and 557 mg lꢀ1, respectively. Their
kinetics analysis (see the inset of Fig. 4b) gives decreasing
k-values of 3.88 · 10ꢀ2 sꢀ1 (R2 = 0.992), 1.26 · 10ꢀ2 sꢀ1
(R2 = 0.997), 5.6 · 10ꢀ3 sꢀ1 (R2 = 0.996) and 4.3 ·
10ꢀ3 sꢀ1 (R2 = 0.995). The decay in k with raising the con-
tent of 1 indicates the gradual acceleration of competitive
Å
gests the production of a constant concentration of OH
from reaction (5) at the Pt anode during electrolysis.
Fig. 4b shows a much quicker decay of 1 under compa-
rable electro-Fenton and photoelectro-Fenton treatments
of 179 mg lꢀ1 of 1 at 100 mA cmꢀ2, as expected if the pro-
duction of oxidant ÅOH from Fenton’s reaction (2) is much
greater than that of reaction (5) at the Pt anode. In both
cases 1 is destroyed at a similar rate, being completely
removed in approximately 7 min. Kinetic analysis of these
data also agrees with a pseudo-first-order reaction of the
metabolite (see inset of Fig. 4b), leading to an average k-
value of (1.35 0.10) · 10ꢀ2 sꢀ1 (R2 =Å0.996). This behav-
ior indicates a very low generation of OH by reaction (3)
under the action of UVA light.
Å
reactions between OH and by-products, thus enhancing
TOC removal and MEC values as experimentally found
(see Figs. 2c and 3b). Fig. 4b evidences a more rapid decay
of 179 mg lꢀ1 of 1 with rising j and their kinetic analysis
shows greater k-values of 6.5 · 10ꢀ3 sꢀ1 (R2 = 0.9996),
1.26 · 10ꢀ2 sꢀ1 (R2 = 0.997) and 1.81 · 10ꢀ2 sꢀ1 (R2 =
0.990) at higher current densities of 33, 100 and
150 mA cmꢀ2, respectively. Note that k does not vary
proportionallyÅwith j, confirming the reaction of a smaller
proportion of OH with pollutants when j rises, since it is
more quickly wasted by parallel non-oxidizing reactions.
The concentration–time plots obtained for the photo-
electro-Fenton treatment of different metabolite contents
and current densities at pH 3.0 are also presented in
Fig. 4b. As can be seen, the complete removal of 1 at
3.5. Identification and evolution of intermediates
A solution of 179 mg lꢀ1 of 1 of pH 3.0 was treated by
electro-Fenton at 100 mA cmꢀ2 and at 35.0 ꢁC for 2 min

1666
I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
and the remaining organics were extracted and analyzed by
GC–MS. The MS spectrum displayed peaks related to sta-
ble aromatics such as 4-chlorophenol (2) (m/z = 128 (100,
M+), 130 (33, (M + 2)+)) at tr = 17.0 min, hydroquinone
(3) (m/z = 108 (100, M+)) at tr = 21.5 min, 4-chlorocate-
chol (5) (m/z = 144 (100, M+), 146 (33, (M+2)+)) at
tr = 18.2 min and p-benzoquinone (6) (m/z = 110 (53,
M+)) at tr = 4.1 min. In addition, an intense peak ascribed
to a chloro-derivative, with m/z = 214 (12, (M+2)+), 212
(36, M+), 184 (22), 169 (100) and 144 (49) as main frag-
mentation, was detected at tr = 14.2 min. Although this
product was not identified by pure standards, it can be
reasonably assigned to a dehydrated species of 2-(4-chloro-
2-hydroxyphenoxy)-2-methylpropionic acid (8), a hydro-
xylated product of 1 that can be transformed into 5
(molecular peak = 144). The silylated derivatives of 2, 3,
5 and 8 were also detected by GC–MS after derivatization
with N,O-bis-(trimetylsilyl)acetamide.
Reversed-phase chromatograms of treated solutions
exhibited peaks related to the products 2 at tr = 5.0 min,
5 at tr = 3.1 min and 6 at tr = 2.0 min, along with other
additional peaks associated with 4-chlororesorcinol (4) at
tr = 2.8 min and 1,2,4-benzenetriol (7) at tr = 1.8 min. All
these aromatics were unequivocally identified by compar-
ing their tr-values and UV–vis spectra, measured on the
photodiode detector, with those of pure products. Note
that only 2, 5 and 6 have been previously reported as prod-
ucts of 1 during its electro-Fenton degradation in 0.01 M
HCl (Oturan et al., 1999).
5
4
3
2
1
0
a
0
60 120 180 240 300 360 420
8
7
6
5
4
3
2
1
0
b
0
2
4
6
8
10
12
14
time / min
Fig. 5. Evolution of the concentration of aromatic intermediates detected
during the degradation of 100 ml of 179 mg lꢀ1 clofibric acid (1) solutions
of pH 3.0 at 100 mA cmꢀ2 and at 35.0 ꢁC. In plot (a), anodic oxidation
with electrogenerated H2O2. In plot (b), electro-Fenton (hollow symbols)
and photoelectro-Fenton (solid symbols), both with 1.0 mM Fe2+
Compound: (d, s) 4-chlorophenol (2); (m) 4-chlororesorcinol (4);
(j, h) 4-chlorocatechol (5); (Ç, Æ) p-benzoquinone (6); (.) 1,2,4-benze-
netriol (7).
.
The evolution of aromatic intermediates during the
treatment of 179 mg lꢀ1 of 1 at pH 3.0 and at 100 mA cmꢀ2
by anodic oxidation with electrogenerated H2O2 is shown
in Fig. 5a. Under these conditions, all products are poorly
accumulated and persist during long time, as expected from
the slow removal of 1 in 240 min (see Fig. 4a). Compounds
5, 6 and 7 are detected up to 300, 360 and 240 min, respec-
tively, after reaching 4.2, 2.1 and 2.8 mg lꢀ1 as maximum
at 30–40 min, whereas 2 and 4 attain ca. 2.5 mg lꢀ1 at
30 min and disappear after 180 min. In contrast, the same
products are much more quickly formed and destroyed
under comparable electro-Fenton and photoelectro-Fenton
´
Brillas et al., 2004a; Sires et al., 2006), whereas 9 is expected
to be released in the first degradation stages of 1. This was
confirmed from the GC–MS analysis of organics produced
after 2 min of the electro-Fenton treatment of 179 mg lꢀ1
of 1 at pH 3.0 and at 100 mA cmꢀ2, since the MS spectrum
after derivatization exhibited a peak of the disylilated
derivative of 9 (m/z = 248 (10, M+)) at tr = 10.2 min. The
photoelectro-Fenton treatment of 50 mg lꢀ1 of 9 at pH
3.0, at 100 mA cmꢀ2 and at 35.0 ꢁC corroborated its oxida-
tion to acid 14. This acid can also be generated from the
Å
´
degradations due to the greater generation of OH from
independent degradation of 10–12 (Sires et al., 2006). The
reaction (2). Fig. 5b shows that in both cases the prod-
uct 2 is accumulated up to 7.3 mg lꢀ1 at 1 min and per-
sists to 10–12 min, whereas 5 and 6 are formed in smaller
extent and destroyed in 7 and 10 min, respectively. The
fact that all products show a similar evolution in both
electro-Fenton and photoelectro-Fenton processes con-
firms that they are not photolyzed under UVA illumi-
nation.
production of 13 and 14 as ultimate carboxylic acids was
confirmed by electrolyzing 179 mg lꢀ1 of 1 at pH 3.0 and
at 100 mA cmꢀ2 under electro-Fenton conditions for 6 h.
The GC–MS analysis after esterification of the remaining
acids with ethanol revealed the presence of an intense peak
corresponding to diethyl oxalate (m/z = 146 (2, M+)) at
tr = 7.9 min, and a very weak peak related to ethyl formate
(m/z = 74 (10, M+)) at tr = 10.5 min.
Ion-exclusion chromatograms of electrolyzed solutions
showed well-defined peaks ascribed to carboxylic acids
such as 2-hydroxyisobutyric (9) at tr = 12.6 min, tartronic
(10) at tr = 7.7 min, maleic (11) at tr = 8.1 min, fumaric
(12) at tr = 16.1 min, formic (13) at tr = 14.0 min and oxa-
lic (14) at tr = 6.6 min. Acids 10–13 come from the oxida-
tion of the aryl moiety of aromatics (Boye et al., 2002;
As can be seen in Fig. 6a for the anodic oxidation treat-
ment with electrogenerated H2O2 of 179 mg lꢀ1 of 1 at pH
3.0 and at 100 mA cmꢀ2, large amounts of acids 9–14 are
slowly accumulated without apparent degradation, except
for acid 10 that reaches a maximum content of 57 mg lꢀ1
at 180 min. After 360 min of electrolysis, 23.8, 25.1,
3.5, 1.7, 11.0 and 14.1 mg lꢀ1 of 9, 10, 11, 12, 13 and 14,

I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
1667
60
50
40
30
20
10
that all detected carboxylic acids give 27 mg lꢀ1 of soluble
TOC, a value much lower than 59 mg lꢀ1 determined for
the final degraded solution (see Fig. 1a). That means that
this solution contains high contents of undetected prod-
ucts, probably hardly oxidizable aromatics.
a
A very different behavior is found for carboxylic acids in
the electro-Fenton and photoelectro-Fenton processes of
the above solution of 1. These products are rapidly
degraded by both treatments at 100 mA cmꢀ2, so that only
the ultimate acid 14 is largely accumulated (see Fig. 6b),
although less than 0.1 mg lꢀ1 of acid 13 is also detected
at the end of electro-Fenton. Fig. 6b illustrates that 9
and 13 persist in large extent to ca. 60 min when smaller
0
80
70
60
50
40
30
20
10
0
b
Å
amount of OH is produced by photoelectro-Fenton at
33 mA cmꢀ2. In both methods complexes of acid 14 with
Fe3+ generated from reaction (2) are expected to be formed
(Zuo and Hoigne, 1992).ÅThese Fe3+-oxalato complexes are
´
difficulty oxidized with OH in electro-Fenton, remaining
ca. 60 mg lꢀ1 of 14, corresponding to 16 mg lꢀ1 of TOC,
at 360 min (see Fig. 6b). Since the resulting solution con-
tains 21 mg lꢀ1 of TOC (see Fig. 1a), one can conclude that
the stable colored chlorinated polyaromatics formed yield
about 5 mg lꢀ1 of TOC. In contrast, Fig. 6b shows the
complete mineralization of acid 14 in photoelectro-Fenton,
because Fe3+-oxalato complexes are efficiently photode-
carboxylated under the action of UVA light (Zuo aꢀn1d
0
60 120 180 240 300 360 420
time / min
Fig. 6. Time-course of the concentration of carboxylic acids detected
under the same conditions as in Fig. 5. In plot (a), anodic oxidation with
electrogenerated H2O2. In plot (b), electro-Fenton (hollow symbols) and
photoelectro-Fenton (solid symbols), both with 1.0 mM Fe2+. Compound:
(Ç) 2-hydroxyisobutyric acid (9); ( ) tartronic acid (10); (m) maleic acid
(11); (j) fumaric acid (12); (.) formic acid (13); (d, s) oxalic acid (14).
Concentrations of 9 and 13 in photoelectro-Fenton were determined at
33 mA cmꢀ2
.
´
Hoigne, 1992). The remaining solution TOC (<4 mg l
,
Fig. 1a) can then be ascribed to the stable colored
chlorinated polyaromatics generated during degradation
of 1.
respectively, corresponding to 11.0, 7.5, 1.5, 0.7, 2.9 and
3.8 mg lꢀ1 of TOC, are found. This balance indicates
OH
O
OH
OH
COOH
CHOH
COOH
10
OH
3
O
OH
6
OH
- Cl–
+
OH
OH
OH
OH
COOH
OH
OH
OH
- Cl–
OH
COOH
11
OH
OH
Cl
4
OH
7
Cl
2
OH
- Cl–
+
HOOC
OH
CH3
CH3
C
O
COOH
COOH
H3C
C
O
H3C
COOH
12
OH
OH
OH
Cl
5
+
OH
HCOOH
13
Cl
Cl
8
1
OH
CH3
COOH
OH
OH
C
COOH
H3C
CO2
COOH
14
OH
9
hν
-Fe2+
Fe3+
Fe3+-oxalato
complexes
Fig. 7. Proposed reaction pathway for clofibric acid (1) degradation in acidic aqueous medium by electro-Fenton and photoelectro-Fenton processes with
Fe2+ as catalyst.

1668
I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
3.6. Proposed degradation pathway
ciently photolyzed to CO2 in photoelectro-Fenton under
the action of UVA light.
Fig. 7 presents a plausible pathway for the degradation
of 1 by the electro-Fenton and photoelectro-Fenton pro-
cesses with Fe2+ as catalyst. The sequence involves all
intermediates detected in this work, including those that
are only identified in anodic oxidation with electrogener-
ated H2O2 since they are quickly destroyed, without accu-
Acknowledgements
The authors thank the financial support from MEC
(Ministerio de Educacion y Ciencia, Spain) under project
CTQ2004-01954/BQU and the grant given to I. Sires by
DURSI (Departament d’Universitats, Recerca i Societat
´
´
Å
mulation, in the above methods. Electrogenerated OH is
stated as the main oxidant and the possible parallel oxida-
tion of complexes of Fe3+ with other products containing
OH groups, different from acid 14, is not indicated for sake
of simplicity.
´
de la Informacio, Generalitat de Catalunya) to do this
work.
References
The process is initiated either by the breaking of the
C(1)–O bond of clofibric acid to form the phenol 2 with
loss of acid 9, or the direct hydroxylation on its C(2)-posi-
Andreozzi, R., Caprio, V., Marotta, R., Radovnikovic, A., 2003.
Ozonation and H2O2/UV treatment of clofibric acid in water: a kinetic
investigation. J. Hazard. Mater. B 103, 233–246.
Å
tion to give 8. Further parallel attack of OH on the C(4)-,
Boye, B., Dieng, M.M., Brillas, E., 2002. Degradation of herbicide 4-
chlorophenoxyacetic acid by advanced electrochemical oxidation
methods. Environ. Sci. Technol. 36, 3030–3035.
C(3)- and C(2)-positions of 2 yields the benzenediols 3,
with release of Clꢀ ion, 4 and 5, respectively. Product 5
is also formed from the oxidation of 8 with loss of acid
9. The subsequent hydroxylation with dechlorination of 4
and 5 leads to the benzenetriol 7. This product is also
formed from ÅOH attack on 3, which is oxidized in parallel
to 6. Further degradation of 6 and 7 leads to a mixture of
acids 10, 11, 12 and 13. The latter acid is directly mineral-
ized to CO2, whereas the three former ones are indepen-
dently transformed into acid 14, which is also generated
from the oxidation of 9. The ultimate Åcarboxylic acid 14
is very slowly converted into CO2 by OH since it forms
very stable Fe3+-oxalato complexes under electro-Fenton
conditions. These species can be photodecarboxylated with
loss of Fe2+ under the action of UVA light (Zuo and
Brillas, E., Banos, M.A., Camps, S., Arias, C., Cabot, P.L., Garrido, J.A.,
˜
Rodr´ıguez, R.M., 2004a. Catalytic effect of Fe2+, Cu2+ and UVA light
on the electrochemical degradation of nitrobenzene using an oxygen-
diffusion cathode. New J. Chem. 28, 314–322.
´
´
Brillas, E., Boye, B., Sires, I., Garrido, J.A., Rodrıguez, R.M., Arias, C.,
Cabot, P.L., Comninellis, C., 2004b. Electrochemical destruction of
chlorophenoxy herbicides by anodic oxidation and electro-Fenton
using a boron-doped diamond electrode. Electrochim. Acta 49, 4487–
4496.
Buser, H.R., Muller, M.D., Theobald, N., 1998. Occurrence of the
pharmaceutical drug clofibric acid and the herbicide mecoprop in
various Swiss lakes and in the North Sea. Environ. Sci. Technol. 32,
188–192.
Daughton, C.G., Jones-Lepp, T.L., (Eds.), 2001. Pharmaceuticals and
Personal Care Products in the Environment. Scientific and Regulatory
Issues. ACS Symposium Series, Washington.
Doll, T., Frimmel, F.H., 2004. Kinetic study of photocatalytic degrada-
tion of carbamazepine, clofibric acid, iomeprol and iopromide assisted
by different TiO2 materials—determination of intermediates and
reaction pathways. Water Res. 38, 955–964.
´
Hoigne, 1992).
4. Conclusions
Go¨zmen, B., Oturan, M.A., Oturan, N., Erbatur, O., 2003. Indirect
electrochemical treatment of bisphenol A in water via electrochemi-
cally generated Fenton’s reagent. Environ. Sci. Technol. 37, 3716–
3723.
Hanna, K., Chiron, S., Oturan, M.A., 2005. Coupling enhanced water
solubilization with cyclodextrin to indirect electrochemical treatment
for pentachlorophenol contaminated soil remediation. Water Res. 39,
2763–2773.
Heberer, T., 2002. Tracking persistent pharmaceutical residues from
municipal sewage to drinking water. J. Hydrol. 266, 175–189.
Heberer, T., Adam, M., 2004. Transport and attenuation of pharmaceu-
tical residues during artificial groundwater replenishment. Environ.
Chem. 1, 22–25.
It has been demonstrated that the photoelectro-Fenton
method with Fe2+ and UVA light as catalysts is able to
mineralize more than 96% of 1 in aqueous medium of pH
3.0. Its efficiency rises with increasing metabolite content
and with decreasing j. This procedure is then viable for
treating acidic wastewaters containing this pollutant. In
contrast, the electro-Fenton method only yields about
80% of decontamination due to the formation of products
Å
hardly oxidizable with OH, which is mainly formed from
reaction (2). Comparative treatment by anodic oxidation
Å
with electrogenerated H2O2 confirms that OH is produced
Heberer, T., Stan, H.J., 1997. Determination of clofibric acid and N-
(phenylsulfonyl)-sarcosine in sewage, river, and drinking water. Int. J.
Environ. Anal. Chem. 67, 113–124.
Irmak, S., Yavuz, H.I., Erbatur, O., 2006. Degradation of 4-chloro-2-
methylphenol in aqueous solution by electro-Fenton and photoelectro-
Fenton processes. Appl. Catal. B: Environ. 63, 243–248.
Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D.,
Barber, L.B., 2002. Pharmaceuticals, hormones, and other organic
wastewater contaminants in U.S. Streams, 1999–2000: A national
reconnaissance. Environ. Sci. Technol. 36, 1202–1211.
in much smaller extent at the Pt anode from water oxida-
tion. Clꢀ ion is released during mineralization. The decay
of 1 always follows a pseudo-first-order kinetics with simi-
lar rate constant for electro-Fenton and photoelectro-
Fenton. Aromatic products and generated carboxylic acids
have been identified by GC–MS. From their quantification
by HPLC chromatography, the different oxidation ability
of both methods can be explained from the behavior of
acid 14. This acid forms very stable Fe3+-oxalato com-
plexes under electro-Fenton conditions, which can be effi-
Kummerer, K. (Ed.), 2001. Pharmaceuticals in the Environment. Sources,
¨
Fate and Risks. Springer, Berlin.

I. Sire´s et al. / Chemosphere 66 (2007) 1660–1669
1669
Oturan, M.A., Aaron, J.J., Oturan, N., Pinson, J., 1999. Degradation of
Removal of pharmaceuticals during drinking water treatment. Envi-
ron. Sci. Technol. 36, 3855–3863.
Tixier, C., Singer, H.P., Oellers, S., Muller, S.R., 2003. Occurrence and fate
¨
chlorophenoxyacid herbicides in aqueous media, using
electrochemical method. Pestic. Sci. 55, 558–562.
a novel
Packer, J.L., Werner, J.J., Latch, D.E., McNeill, K., Arnold, W.A., 2003.
Photochemical fate of pharmaceuticals in the environment: Naproxen,
diclofenac, clofibric acid, and ibuprofen. Aquat. Sci. 65, 342–351.
of carbamazepine, clofibric acid, diclofenac, ibuprofen, ketoprofen, and
naproxen in surface waters. Environ. Sci. Technol. 37, 1061–1068.
Ventura, A., Jacquet, G., Bermond, A., Camel, V., 2002. Electrochemical
generation of the Fenton’s reagent: application to atrazine degrada-
tion. Water Res. 36, 3517–3522.
´
´
Sires, I., Garrido, J.A., Rodrıguez, R.M., Cabot, P.L., Centellas, F.,
Arias, C., Brillas, E., 2006. Electrochemical degradation of paracet-
amol from water by catalytic action of Fe2+, Cu2+, and UVA light on
electrogenerated hydrogen peroxide. J. Electrochem. Soc. 153, D1–D9.
Sun, Y., Pignatello, J.J., 1993. Photochemical reactions involved in the
total mineralization of 2,4-D by iron(3+)/hydrogen peroxide/UV.
Environ. Sci. Technol. 27, 304–310.
Weigel, S., Berger, U., Jensen, E., Kallenborn, R., Thoresen, H.,
Huhnerfuss, H., 2004. Determination of selected pharmaceuticals
¨
and caffeine in sewage and seawater from Tromsø/Norway with
emphasis on ibuprofen and its metabolites. Chemosphere 56, 583–592.
Xie, Y.B., Li, X.Z., 2006. Interactive oxidation of photoelectrocatalysis
and electro-Fenton for azo dye degradation using TiO2–Ti mesh and
reticulated vitreous carbon electrodes. Mater. Chem. Phys. 95, 39–50.
Tauxe-Wuersch, A., De Alencastro, L.F., Grandjean, D., Tarradellas, J.,
2005. Occurrence of several acidic drugs in sewage treatment plants in
Switzerland and risk assessment. Water Res. 39, 1761–1772.
Ternes, T.A., 1998. Occurrence of drugs in German sewage treatment
plants and rivers. Water Res. 32, 3245–3260.
´
Zuo, Y., Hoigne, J., 1992. Formation of hydrogen peroxide and depletion
of oxalic acid in atmospheric water by photolysis of iron(III)-oxalato
complexes. Environ. Sci. Technol. 26, 1014–1022.
Ternes, T.A., Meisenheimer, M., McDowell, D., Sacher, F., Brauch, H.J.,
Haist-Gulde, B., Preuss, G., Wilme, U., Zulei-Seibert, N., 2002.
Zwiener, C., Frimmel, F.H., 2000. Oxidative treatment of pharmaceuticals
in water. Water Res. 34, 1881–1885.