ww.sciencedirect.com

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 0

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/watres

Remediation of 17-a-ethinylestradiol aqueoussolution by photocatalysis and electrochemically-assisted photocatalysis using TiO2 and TiO2/WO3

electrodes irradiated by a solar simulator

Haroldo G. Oliveira a, Leticia H. Ferreira b, Rodnei Bertazzoli b,Claudia Longo a,*

a Institute of Chemistry, University of Campinas e UNICAMP, PO Box 6154, 13083-970 Campinas, SP, Brazilb Faculty of Mechanical Engineering, University of Campinas e UNICAMP, PO Box 6122, 13083-970 Campinas, SP,

Brazil

a r t i c l e i n f o

Article history:

Received 3 June 2014

Received in revised form

17 August 2014

Accepted 26 August 2014

Available online xxx

Keywords:

17-a-ethinylestradiol

TiO2

TiO2/WO3

Photocatalysis

Photoelectrochemistry

* Corresponding author.E-mail addresses: cla_longo@hotmail.com

Please cite this article in press as: Oliveiraand electrochemically-assisted photocaResearch (2014), http://dx.doi.org/10.1016

http://dx.doi.org/10.1016/j.watres.2014.08.0420043-1354/© 2014 Elsevier Ltd. All rights rese

a b s t r a c t

TiO2 and TiO2/WO3 electrodes, irradiated by a solar simulator in configurations for het-

erogeneous photocatalysis (HP) and electrochemically-assisted HP (EHP), were used to

remediate aqueous solutions containing 10 mg L�1 (34 mmol L�1) of 17-a-ethinylestradiol

(EE2), active component of most oral contraceptives. The photocatalysts consisted of

4.5 mm thick porous films of TiO2 and TiO2/WO3 (molar ratio W/Ti of 12%) deposited on

transparent electrodes from aqueous suspensions of TiO2 particles and WO3 precursors,

followed by thermal treatment at 450�C. First, an energy diagram was organized with

photoelectrochemical and UVeVis absorption spectroscopy data and revealed that EE2

could be directly oxidized by the photogenerated holes at the semiconductor surfaces,

considering the relative HOMO level for EE2 and the semiconductor valence band edges.

Also, for the irradiated hybrid photocatalyst, electrons in TiO2 should be transferred toWO3

conduction band, while holes move toward TiO2 valence band, improving charge separa-

tion. The remediated EE2 solutions were analyzed by fluorescence, HPLC and total organic

carbon measurements. As expected from the energy diagram, both photocatalysts pro-

moted the EE2 oxidation in HP configuration; after 4 h, the EE2 concentration decayed to

6.2 mg L�1 (35% of EE2 removal) with irradiated TiO2 while TiO2/WO3 electrode resulted in

45% EE2 removal. A higher performance was achieved in EHP systems, when a Pt wire was

introduced as a counter-electrode and the photoelectrodes were biased at þ0.7 V; then, the

EE2 removal corresponded to 48 and 54% for the TiO2 and TiO2/WO3, respectively. The

hybrid TiO2/WO3, when compared to TiO2 electrode, exhibited enhanced sunlight har-

vesting and improved separation of photogenerated charge carriers, resulting in higher

performance for removing this contaminant of emerging concern from aqueous solution.

© 2014 Elsevier Ltd. All rights reserved.

, clalongo@iqm.unicamp.br (C. Longo).

, H.G., et al., Remediation of 17-a-ethinylestradiol aqueous solution by photocatalysistalysis using TiO2 and TiO2/WO3 electrodes irradiated by a solar simulator, Water/j.watres.2014.08.042

rved.

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 02

1. Introduction

The concerns with emerging contaminants in water resources

have been increasing lately. Among these pollutants, which

comprise a broad range of synthetic and naturally occurring

chemicals, the estrogenic hormones and other endocrine-

disrupting compounds have received cautious attention due

to their potential risk to the environment and human health;

also, their occurrence has been detected worldwide in sewage

effluents of densely populated areas (Hamid and Eskicioglu,

2012). In Brazil, for instance, endocrine active chemicals

were found in several regions and in different aquatic eco-

systems including mangroves, rivers and reservoirs

(Montagner and Jardim, 2011). The main sources of estrogens

entering in the aquatic environment include domestic

sewage, which contains the female hormone excretion, hos-

pital wastewater and manufacturing plants effluents. Water

contamination with hormones also results from flushing un-

used drugs down the drain or toilet, because drug take-back

programs are not efficient yet (Nasuhoglu et al., 2012).

The synthetic estrogenic steroid 17a-ethynilestradiol (EE2),

active component of most oral contraceptives and for hor-

mone replacement therapy, consists of a worthy contaminant

of emerging concern. EE2 exhibits high estrogenic potency

and, since it is very resistant to biodegradation in wastewater

treatment plants, it is also considered a persistent pollutant

(Hamid and Eskicioglu, 2012). These important points have

been encouraging studies of EE2 degradation using different

methods, including heterogeneous photocatalysis (HP), Photo-

Fenton and other “Advanced Oxidation Processes”, which are

very efficient to oxidize and eliminate hazardous non-

biodegradable pollutants (Malato et al., 2009).

Solar HP has been considered a promising alternative for

developing innovative and sustainable technologies for

wastewater remediation and can be used as a pre-treatment,

in order to eliminate non-biodegradable compounds, or as a

polishing step to treat recalcitrant organic contaminants

(Chong et al., 2010). Usually, the photocatalyst is used as

suspended colloidal particles in slurry reactors or immobilized

on different substrates. Using slurry reactors, irradiation can

uniformly reach the photocatalyst particles, which guaran-

tees high efficiency even for low pollutant concentrations;

however, another step is necessary for particle separation

from the slurry, which, in addition, should be recuperated for

re-use. Considering this point, the immobilized photocatalyst

is much more practical; moreover, for conductive substrates,

HP can be electrochemically assisted (EHP), which usually

increases efficiency. However, other inconvenient questions

arise for immobilized photocatalyst, associated with the

smaller available surface area per unit volume and also limi-

tations related to contaminant mass transfer from the solu-

tion to the photocatalyst surface (Qu et al., 2013).

TiO2, a wide band gap n-type semiconductor (Ebg of ca. 3.2

ev, photoactive under UV irradiation), is the most studied

photocatalyst for removal of organic pollutants from water.

The catalytic activity results from the photo-induced charge

separation that can occur on the semiconductor surface. First,

absorption of photons with energy� Ebg promotes an electron

to the conduction band (CB) leaving behind a hole (hþ) in the

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 andResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

valence band (VB); then the holes, as well as generated �OH

radicals and other reactive species, produce the photo-

catalytic oxidation of organic compounds,(De la Cruz et al.,

2012).

Motivated to contribute to the understanding of photo-

electrochemical processes in HP water treatment, Claudia

Longo and co-workers, at Physical Chemistry Dept, UNI-

CAMP, have been investigating the synthesis and character-

ization of semiconductor oxides and their application as

photoanodes for solar energy conversion, including water

remediation. Studies carried out with porous TiO2 film elec-

trodes and phenol aqueous solution irradiated by a solar

simulator revealed that, after adding a Pt wire as a counter-

electrode, application of a potential bias with a potentio-

stat, or by an external connection to a solar cell, notably

enhanced the efficiency for phenol degradation. Probably, the

e�/hþ recombination can be suppressed in the EHP configu-

ration, enhancing the TiO2 photocatalytic activity for organic

pollutants removal (Oliveira et al., 2010). Later, searching for

oxides with higher sunlight harvesting properties, Fe-doped

TiO2 samples were synthesized by the solegel/hydrother-

mal methodology. The research revealed that a fraction of

iron was inserted as dopant in the TiO2 while another was

segregated as hematite at the Fe:TiO2 grain boundaries. For

such a hybrid photocatalyst, an absorption shift to the visible

range of solar spectrum would not only be expected but also

a lower recombination of photogenerated charge carriers

because, taking into account the relative positions of CB and

VB edges for both semiconductors, photoinduced electrons

should be transferred to the Fe2O3 CB while holes move to-

ward the TiO2 VB. However, besides the improvement of

sunlight absorption, the Fe:TiO2 photocatalytic activity for

phenol degradation was lower than that observed for pristine

TiO2. Then, since Fe2O3 presents a much lower dielectric

constant than TiO2, we proposed that segregated hematite at

the Fe:TiO2 grain boundaries could affect the separation of

photogenerated e-/hþ pairs, as well as electron transport

through the Fe:TiO2 film, decreasing the material perfor-

mance as photocatalyst for phenol oxidation (Santos et al.,

2012). Meanwhile, the photoelectrochemical properties and

the photocatalytic activity for remediation of Rhodamine 6G

(R6G) aqueous solutions were investigated for electrodes

consisting of Degussa P-25 TiO2 and bi-layered WO3eTiO2

porous films. WO3 is an n-type semiconductor photoactive

with the visible range of solar spectrum (Ebg of 2.5e2.7 eV);

also, the relative position of its CB and VB edges is convenient

for a combination with TiO2 (as mentioned for Fe2O3) and,

finally, it presents a high dielectric constant. First, an energy

diagram was assembled in order to evaluate the relative en-

ergy for the CB and VB edges for both semiconductors, as well

as the highest occupied and lowest unoccupied molecular

orbitals (HOMO, LUMO) for R6G. Then, the color removal de-

cays for R6G solutions remediated using both electrodes in

HP and EHP configurations were determined. As expected,

taking into account conclusions from previous studies, for a

given electrode, the EHP configuration was more efficient

than HP and, for a specific configuration, the hybrid

WO3eTiO2 electrode exhibited a higher photocatalytic per-

formance when compared to the TiO2 electrode (Oliveira

et al., 2012).

n of 17-a-ethinylestradiol aqueous solution by photocatalysisTiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 0 3

Considering the promising photocatalyst performance

obtained for WO3eTiO2 photoanodes for R6G dye oxidation,

this contribution reports on the investigation of EE2 removal

from aqueous solutions using a hybrid TiO2/WO3 electrode,

with a W/Ti molar ratio of 12%. WO3 was used as a minor

component because tungsten oxide precursors are much

more expensive than Degussa P25 TiO2. The studies included

the characterization of morphological and structural proper-

ties of the photocatalysts, the organization of an energy dia-

gram forWO3, TiO2 and EE2 and finally, the remediation of EE2

aqueous solutions using the photocatalysts in HP and EHP

configurations.

2. Materials and methods

2.1. Preparation of TiO2, WO3 and TiO2/WO3 particlesand electrodes

Deionized water from a Milli-Q water purification system and

purest grade commercial chemicals were used throughout the

work. Transparent conductive glass-FTO (TCO22-15, 15 U/sq

Solaronix S.A, Switzerland), previously cleaned by sonication

with isopropyl alcohol, was used as substrate for porous films

electrodes.

The TiO2 electrodes were prepared using an aqueous sus-

pension containing Degussa P25 TiO2 particles (according to

the manufacturer, the sample consisted of 80:20 anatase:r-

utile wt% spherical particles with diameter of c.a. 25 nm).

Initially, the oxide (1.0 g), was crushed with acetylacetone

(30 mL); then, an aqueous solution of polyethylene glycol (PEG,

Mw 20 000) was slowly added, resulting in a suspension with

10:30:60 wt% PEG:TiO2:H2O. The TiO2/WO3 electrodes were

prepared using another suspension containing TiO2 and pre-

cursors for WO3. First, tungstic acid powder was previously

dissolved in hydrogen peroxide solution (30%) which resulted

in 1.1 mol L�1 solution. Then, this solution was added to the

TiO2 particles (previously crushedwith acetylacetone) and the

PEG aqueous solution, resulting in a suspensionwith a 12%W/

Ti molar ratio; the same relative amount of PEG and H2O was

used, considering the expected mass of both oxides. Each

suspension was continuously stirred for 2 h and then, small

aliquots were spread onto glass-FTO with a glass rod, using

adhesive tape as spacer. The electrodes were then heated in

an oven from ambient temperature (10�C min�1), maintained

at 350 �C for 30 min and then heated at 450 �C for 30 min. Also,

a small aliquot of each suspension was thermally treated in

order to prepare the powders used for photocatalyst particle

characterizations.

A sample of WO3 particles was also prepared by dissolving

tungstic acid powder (5.0 g) in hydrogen peroxide solution

(100 mL) under magnetic stirring for 2 h. The solution was

dried and the resulting powder was heated at 450 �C for

30 min.

2.2. Characterization of TiO2, WO3 and TiO2/WO3

particles

The crystalline structure of TiO2 and TiO2/WO3 particles was

characterized by X-ray diffraction analysis using Cu Ka

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 and TResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

(l ¼ 1,5406 Ǻ) radiation (XRD Shimadzu 7000). The obtained

diffraction patterns were compared to the data from Joint

Committee on Powder Diffraction Standards Database

(JCPDS).

The specific surface area and total pore volume of the

photocatalyst particles were determined using Brunauer-

Emmett-Teller (BET) methodology by N2 adsorption-

desorption at 77 K (Quantachrome). The samples were

degassed at 120 �C for 15 h before each analysis.

The UVeVis diffuse reflectance spectra (UVeVis DRS) were

obtained using a Teflon disc as the reference standard (Cary

SG-Varian apparatus, equipped with an integration sphere

accessory). Considering the variation of semiconductor ab-

sorption (a) with the incident photon energy (hy), the band gap

energy could be estimated from a plot of (a � hn)1/h as a

function of hn. For a direct transition (the momentum is

conserved), h ¼ ½; for indirect transitions (the momentum is

not conserved), h ¼ 2. However, in order to consider scattering

effects (S), the absorption data from DRS were first converted

taking into account the sample and Teflon reflectance values

using the Kubelka-Munk function, (Murphy, 2007) as follows.

a

S¼ ð1� R∞Þ2

2R∞¼ F

�R∞

�(1)

where R∞ ¼ RSample

RTeflon

Then, the band gap energy for the semiconductor particles

was obtained by extrapolation, from plots of [F(R∞) x hn]1/h as a

function of hn.

2.3. Characterization of the TiO2 and TiO2/WO3

electrodes

The film surfaces were examined by scanning electron mi-

croscopy (SEM JEOL JSM 63601V). The electrochemical prop-

erties were investigated using an electrochemical cell

assembled with an optic glass window (ca. 90% transmittance

for irradiation with l > 320 nm). The experiments were per-

formed using a 0.1mol L�1 Na2SO4 aqueous solution (pH¼ 6.9)

as a supporting electrolyte (10 mL), porous electrodes (with

geometrical area of 1 cm2) facing a Pt wire ring as a counter

electrode and Ag/AgCl (in aqueous 3 mol L�1) as a reference

electrode (inside in a Luggin capillary). The measurements

were performed using a potentiostat/galvanostat (Ecochemie

Autolab PGSTAT 302-N), in the dark and under front side

polychromatic irradiation. The irradiation was provided by a

“homemade” solar simulator, assembledwith ametallic vapor

discharge lamp (Metalarc Sylvania HIS-YHX 400 W); consid-

ering the distance from the electrochemical cell (ca. 15 cm),

the polychromatic irradiance was estimated as (130 ± 10) mW

cm�2 and the temperature was maintained at (29 ± 2)�C(Oliveira et al., 2010).

2.4. EE2 removal using TiO2 and TiO2/WO3 electrodes

The photocatalytic activity of irradiated TiO2 and TiO2/WO3

film electrodes for EE2 hormone removal from water was

investigated using an EE2 solution with initial concentration

(C0) of 10 mg L�1 (34 mmol L�1) dissolved in the supporting

electrolyte (aqueous 0.1 mol L�1 Na2SO4). The pH of EE2

n of 17-a-ethinylestradiol aqueous solution by photocatalysisiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 04

solution, measured with a pHmeter, corresponded to 6.9. The

studies were performed using three configurations: (i) het-

erogeneous photocatalysis (HP), using a system consisted by

a photocatalyst film with 1.0 cm2 and 10 mL of EE2 solution;

(ii) electrochemically-assisted HP (EHP), a configuration

where a Pt wire ring was included to the HP system and,

using the potentiostat, þ0.7 V was applied to the photo-

catalyst film; and (iii) photolysis, by irradiation of the EE2

solution under the same conditions but without the photo-

catalysts (a “blank control”). The systems were not stirred

and not bubbled with any gas in these measurements. After

irradiation during 4 h (or different time intervals, for kinetic

investigation), the remediated solution was analyzed by

fluorescence measurements, determination of the total

organic carbon (TOC) and by high performance liquid chro-

matography (HPLC).

2.5. Analysis of the remediated EE2 solutions

The EE2 concentration was determined using fluorescence

spectroscopy (Varien Cary Eclipse Spectrofluorimeter), in

emission scan mode (at 600 nm min�1), with EX slit of 10 and

EM slit of 5. A quartz cell (1 cm) was used, with excitation and

emission wavelengths corresponding to 278 nm and 307 nm

respectively. First, an analytical curve was determined; the

fluorescence peak intensity (F) exhibited a linear relationship

with the hormone concentration (CEE2) according to

F ¼ 6.25 þ 52.0 CEE2. For EE2 concentration ranging from 1.0 to

10.0 mg L�1, the correlation coefficient was R ¼ 0.9985.

The presence of EE2, as well as of possible by-products, was

verified by HPLC (Shimadzu HPLC model 10A VP), using a pre-

column C18 (5 mm, 3.9 mm � 20 mm) and a C18 column

(250� 4.6 mm, 5 mm, Supelco). Amixture consisted of 23/24/53

(V/V) of methanol/acetonitrile/water was used as mobile

phase, with isocratic flow rate of 1.0 mL min�1. The UV de-

tector was set at 210 and 280 nm. Previously, a calibration

curve was elaborated using solutions containing known EE2

concentrations (from 2.0 to 10.0 mg L�1). It was found that the

peak area of the chromatogram linearly increased with EE2

concentration; thus, the HPLC analysis were also used to

quantify the EE2 remaining concentration in treated solutions.

The TOC concentration in remediated EE2 solutions was

estimated with the Analytic Jena TOC Multi N/C 2100 equip-

ment. Previously, the TOC concentration was also determined

for seven solutions with known EE2 concentrations (from 0.5

to 10 mg L�1) and excellent agreement was found comparing

the expected and the experimentally obtained values.

Fig. 1 e UVeVis diffuse reflectance spectra for TiO2, WO3

and TiO2/WO3 particles. In set: variation of the respective

Kubelka-Munk function with photon energy for TiO2 and

WO3.

3. Results

3.1. Characterization of TiO2 and TiO2/WO3 particles

The thermal treatment of the photocatalysts precursor sus-

pensions resulted in crystalline particles, as observed by X-

Ray diffraction patterns available in Figure S1 (supporting

information). Comparison with the JCPDS data revealed that

the white TiO2 particles consisted of anatase and rutile pha-

ses, as expected for this commercial sample. For the yellowish

TiO2/WO3 particles, in addition to the anatase and rutile

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 andResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

peaks, triclinic and hexagonal phases of WO3 can also be

identified (Yang et al., 2005).

The surface areas of TiO2 and TiO2/WO3 particles were

estimated as 50 and 42 m2 g�1, respectively, by N2 adsorption-

desorption isotherms (Supporting information, Fig. S2). Prob-

ably, the WO3 particles grew on the TiO2 surface, obstructing

some pores; thus the surface area of the TiO2/WO3 samplewas

smaller than that exhibited by TiO2. Also, it can be noticed that

the TiO2/WO3 particles exhibited a type-II isotherm, charac-

teristic of mesoporous structure; for TiO2, the observed hys-

teresis in this type-IV isotherm is typical for mesopores with

diameter ranging from 2 to 50 nm (Myers, 1999).

The photocatalyst abilities for solar harvesting were first

evaluated using UVeVis diffuse reflectance; the WO3 sample

was also considered for comparison. The absorption edges

were estimated by extrapolation from the curves of optical

absorption as function of wavelength (Fig. 1) as 390 nm for

TiO2, 490 nm for WO3 and 449 nm for TiO2/WO3 (with molar

ratio W/Ti ¼ 12%). Thus, the TiO2/WO3 absorption edge is an

intermediate value of those exhibited by the pristine oxides,

as has been mentioned in the literature (for instance, 475 nm

for TiO2/WO3 samples with W:Ti of 24:76 (Lin et al., 2008).

Considering that TiO2 and WO3 exhibit indirect transitions

resulted in the Kubelka-Munk functions (F(R∞) x hn)1/2 repre-

sented as inset in Fig. 1. From these curves, the Ebg values for

TiO2 and WO3 were estimated by extrapolating the linear

portion and corresponded respectively to 3.1 and 2.5 eV, in

agreement with the absorption edges from UVeVis DRS.

These Ebg values are coherent with those reported in the

literature. WO3 usually is considered as a semiconductor with

indirect transition, with band gap ranging from 2.5 to 2.7 eV

(Higashimoto et al., 2008). Indirect transition is also usually

reported for Degussa P-25 TiO2, however, for other TiO2 sam-

ples, the DRS data can be described taking into account direct

or indirect transitions, resulting in Ebg values ranging from 3.0

to 3.5 eV, depending on the predominant crystalline phase

(Reyes-Coronado et al., 2008).

n of 17-a-ethinylestradiol aqueous solution by photocatalysisTiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 0 5

3.2. Morphological and electrochemical properties ofTiO2 and TiO2/WO3 electrodes

The photocatalyst films deposited onto the conductive surface

of transparent electrodes, with a geometric area of 1 cm2,

exhibited uniform surfaces. The density of immobilized par-

ticles corresponded to 1.2 mg cm�2 for TiO2 and 1.5 mg cm�2

for TiO2/WO3; considering the molar mass of these oxides,

both samples exhibited ca. 16 mmol of photocatalyst per cm2.

Fig. 2 shows some representative SEM images of the films

cross-section and surface. Both the films are constituted by

agglomerated particles with diameter ranging from 50 to

200 nm, exhibiting similar morphology and thickness

(4.5 ± 0.5 mm).

The electrochemical properties of these porous electrodes

were then investigated using 0.1 mol L�1 Na2SO4 aqueous so-

lutions as supporting electrolyte. In the dark, the open circuit

potential (OCP) exhibited by the TiO2 electrode was 0.18 V;

under polychromatic irradiation, the OCP was �0.25 V, thus

Fig. 2 e Scanning electron micrographs of the surface and cross

deposited onto glass-FTO substrate.

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 and TResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

the resulting photopotential corresponds to �0.43 V. As ex-

pected for n-type semiconductor electrodes, the negative

photopotential resulted from electron injections into the TiO2

conduction band (Oliveira et al., 2010). A negative photo-

potential was also observed for the TiO2/WO3 electrode; the

OCP in the dark and under irradiation corresponded to 0.30 V

and �0.10 V, respectively. From cyclic or linear sweep vol-

tammetry measurements it was observed that both the elec-

trodes exhibited positive values for photocurrents, as also

expected for n-type semiconductor electrodes at potentials

higher than the flat band potential (Efb). In this condition,

holes moved to the electrode surface and oxidize species in

solution, producing an anodic phototocurrent. Thus, such

materials should exhibit promising application as photo-

anodes for oxidation of organic compounds (Bard and

Faulkner, 2001; Oliveira et al., 2010, 2012).

The currentevoltage profiles exhibited by the TiO2 and

TiO2/WO3 electrodes in the dark and under polychromatic

irradiation are represented in Fig. 3 (20 mVs�1). In the dark

-sectional of TiO2 (a, b, c) and TiO2/WO3 (d, e, f) films

n of 17-a-ethinylestradiol aqueous solution by photocatalysisiO2/WO3 electrodes irradiated by a solar simulator, Water

Fig. 3 e Cyclic voltammograms (20 mV s¡1) for (a) TiO2

electrode in the dark, in 0.1 mol L¡1 Na2SO4 aqueous

solution and in supporting electrolyte containing

10 mg L¡1 EE2; in set: chemical structure of EE2 molecule;

(b) for TiO2 and TiO2/WO3 electrodes in EE2 solution under

irradiation.

Fig. 4 e Energy diagram for the interface semiconductorjEE2 hormone in aqueous solution, considering the

hormone highest occupied and the lowest unoccupied

molecular orbitals, as well as the valence and conduction

band edges for WO3 and TiO2.

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 06

(Fig. 3a), in supporting electrolyte, both electrodes exhibited

almost identical voltammograms, with a small capacitive

current limited by the hydrogen and oxygen evolution re-

actions (onset at ca. 0.0 V for HER and 1.2 V for OER, in relation

to Ag/AgCl reference electrode). Comparison of the voltam-

mograms obtained for the supporting electrolyte and for the

solution containing 10 mg L�1 of EE2 hormone revealed an

oxidation current peak with onset at 0.8 V that can be attrib-

uted to EE2 oxidation. This current gradually increased with

applied potential (for instance, corresponded to 6 mA cm�2 at

1.1 V), and superimposed with that for OER. Probably, the EE2

oxidation occurred with a considerable overpotential on these

semiconducting electrodes, considering that this reaction was

observed at 0.55 V using carbon nanotube-modified electrodes

in aqueous buffer solution, pH ¼ 7 (Vega et al., 2007). Since

almost identical curves were obtained in the supporting

electrolyte (data not shown), Fig. 3b shows only the voltam-

mograms registered under irradiation for both electrodes in

EE2 solution. The photocurrent remained almost constant

from potentials ranging from 0.1 to 0.8 V (ca. 18 and

22 mA cm�2 for TiO2 and TiO2/WO3 electrodes, respectively)

and gradually increased with applied potential; at 1.1 V, the

photocurrent values exhibited by irradiated TiO2 and TiO2/

WO3 electrodes were 22 and 25 mA cm�2. Besides the EE2

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 andResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

oxidation peak cannot be identified in the voltammograms

obtained under irradiation, EE2 molecules can be oxidized by

photogenerated holes or hydroxyl radicals produced at elec-

trode surface, resulting in the slighter higher photocurrent

values. Comparison of both voltammograms revealed that the

TiO2/WO3 electrode exhibited higher photocurrent than the

TiO2 electrode.

In order to evaluate the possible application of TiO2 and

TiO2/WO3 electrodes for EE2 photocatalytic oxidation, an en-

ergy diagram for the hormonejsemiconductor interface was

assembled, considering the positions of conduction and

valence bands for the semiconductors and the LUMO and

HOMO energies for the hormone.

For semiconductors, the CB edge can be associated with

the flat band potential (Efb), which can be estimated from

photocurrent measurements at different applied potentials,

as briefly described in supporting information. Figure S3

shows the variation of the square of the current with

applied potential for the irradiated TiO2, WO3 and TiO2/WO3

electrodes. Assuming that the Efb is associated to the photo-

current onset (Radecka et al., 2008), the Efb for the TiO2 and

WO3 electrodes can be estimated respectively as �0.3 and

0.3 V (in relation to Ag/AgCl reference electrode). The Efbvalues can be converted considering the correlation for energy

scales, [E (eV) ¼ �4.5 eV ee ESHE (V)], where e is the elementary

charge and ESHE is the potential with respect to the standard

hydrogen electrode (Bard and Faulkner, 2001). Then, the CB

edges for these TiO2 and WO3 electrodes correspond respec-

tively to �4.4 and �5.0 eV. Now, since the energy gap from the

CB and VB edges is associated with the optical band gap

estimated from DRS (3.1 and 2.5 eV), the VB edges for TiO2 and

WO3 electrodes correspond respectively to �7.5 and �7.7 eV,

as represented in Fig. 4. These values are in reasonable

agreement with those reported in the literature; in aqueous

electrolyte with pH ¼ 1, for TiO2, the CB and VB edges ranges

respectively from�4.2 to�4.4 eV and from�7.3 to�7.6 eV; for

WO3, the bottom of CB varied from�4.6 to�4.8 eV and, for VB,

from �7.2 to �7.4 eV (Bard and Faulkner, 2001; Gratzel, 2001).

For organic compounds, the HOMO level usually can be

identified with its oxidation potential, whereas the HOMO-

n of 17-a-ethinylestradiol aqueous solution by photocatalysisTiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 0 7

LUMO energy gap is associated with the absorption in UVeVis

spectrum (Evans, 2008). The redox potential for EE2 oxidation

in the supporting electrolyte was then determined using Pt as

a working electrode. From the cyclic voltammogram pre-

sented as Figure S4 (supporting information), the EE2 redox

potential was identified as 0.75 V (in relation to Ag/AgCl);

converting to the energy scale, this value corresponds to

�5.5 eV. The EE2 UVeVis absorption spectrum exhibited ab-

sorption peaks and bands around 205, 225 and 280 nm

(Figure S5, supporting information). Considering that

[E(eV) ¼ 1241/l(nm)], the corresponding energy values are 6.1,

5.5 and 4.4 eV; thus, the LUMO energy levelmust be situated in

the range varying from �1.1 to 0.6 eV. These values are in

reasonable agreement with those estimated by density func-

tional theory for EE2, which corresponded to �6.05 and �0.39

for HOMO and LUMO respectively (Rokhina and Suri, 2012).

Considering first the relative positions of the semi-

conductor CB and VB edges, compared to those for the TiO2

electrode, the photoinduced charge separation should be

improved for the mixed TiO2/WO3 electrode, since electrons

injected into the TiO2 CB must move to the WO3 CB, while the

holes should be displaced in the opposite direction (fromWO3

VB to TiO2 VB) (Higashimoto et al., 2008). Finally, taking into

account the relative HOMO level for EE2 and the semi-

conductor VB, it can be concluded that this hormone can be

oxidized by the photogenerated holes at the semiconductor

surfaces.

Fig. 5 e Remediation of EE2 aqueous solution with

C0 ¼ 34 mmol L¡1 (10 mg L¡1) under polychromatic

irradiation at (29 ± 2) �C by photolysis and by

heterogeneous photocatalysis (HP) or electro-assisted HP

(EHP) using TiO2 and TiO2/WO3 electrodes: (a) EE2

concentration decay (determined by fluorescence) and (b)

logarithm of relative EE2 concentration.

3.3. Photocatalytic activity of TiO2 and TiO2/WO3

electrodes for EE2 removal

The activity of TiO2 and TiO2/WO3 electrodes for photo-

catalytic EE2 removal from water was then investigated using

systems for HP and EHP configurations, with 1.0 cm2 elec-

trodes for remediation of 10 mL of EE2 aqueous solution

(C0 ¼ 10 mg L�1, 34 mmol L�1). For comparison, an EE2 solution

irradiated without any photocatalyst was also investigated

(“control” to evaluate the EE2 removal by photolysis). For EHP

configuration, the front side irradiated semiconductor elec-

trode was biased at þ0.7 V in order to suppress the e�/hþ

recombination and electrochemically assist the photo-

catalytic EE2 oxidation. This potential, insufficient to achieve

the EE2 electrochemical oxidation (as observed from Fig. 3a),

was already used for remediation of phenol and Rhodamine

6G aqueous solutions with EHP system (Oliveira et al., 2010,

2012) because this value corresponds to the typical OCP

exhibited by an irradiated dye-sensitized TiO2 solar cell (Longo

and De Paoli, 2003). Thus, in upcoming studies, the potentio-

stat could be replaced by an external connection to a solar cell.

First, the EE2 adsorptions at the photocatalyst surface were

determined in the dark. For the TiO2 electrode, after 15 min,

the EE2 concentration decreased from 10 to 9.7 mg L�1; thus,

the quantity of adsorbed EE2 corresponded to 5 nmol cm�2. A

higher amount was determined for the TiO2/WO3,

8 nmol cm�2. These results seem coherent; besides the TiO2/

WO3 particles exhibited lower surface areas than for TiO2 (42

and 50m2 g�1, respectively), the TiO2/WO3 electrode exhibited

a slightly higher amount of immobilized particles than the

TiO2 electrode.

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 and TResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

After, periodic measurements of fluorescence and TOC

concentration were performed for the remediated EE2 solu-

tions. Fig. 5a revealed that, for the “control”, the EE2 concen-

tration slowly decayed with irradiation time from 34.0 to

31.1 mmol L�1, resulting in 9% EE2 removal by photolysis after

4 h of irradiation. From the TOC concentration decay (Fig. S6,

supporting information), the corresponding mineralization

was 4%. For the solutions remediated using the HP and EHP

configurations, for the initial 60 min, a linear decay of EE2

concentration can be observed (Fig. 5a). During this period, for

the TiO2 electrode irradiated in HP and EHP systems, EE2

decreased from the initial C0 ¼ 34 mmol L�1 to 26.1 and to

24.7 mmol L�1, respectively (EE2 removal of 23 and 27%). A

higher performance was found for the TiO2/WO3 electrode;

after 60 min under irradiation using HP and EHP configura-

tions the EE2 remaining concentrations were 24.1 and

22.3 mmol L�1 (respectively 29 and 33% of EE2 removal). During

this period, the TOC concentrations were almost unaltered

(Figure S6, supporting information). For the overall 4 h treat-

ment period, the TOC concentration decrease was always less

pronounced than the corresponding EE2 fluorescence decay.

However, since the EE2 fluorescence is related to its estrogenic

activity, due to the interaction of the molecular functional

groups with the estrogenic receptor, even partial EE2

n of 17-a-ethinylestradiol aqueous solution by photocatalysisiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 08

degradation could be worthwhile for reducing the EE2 effects

in the environment (Ohko et al., 2002; Zhang et al., 2006).

Considering the initial 120 min, the EE2 concentration

decay can be adjusted by pseudo-first order kinetics, ln (C/

C0) ¼ �kap� t, as represented in Fig. 5b. The apparent rate

constants (kap) for EE2 removal using TiO2 in HP and EHP

systems correspond to 0.0040 and 0.0053 min�1 respectively.

Higher values were found for the TiO2/WO3 electrode, 0.0052

and 0.0059 min�1 in HP and EHP configurations. The photo-

catalytic degradation of organic pollutants often shows a

pseudo-first order kinetics, depending only on the pollutant

concentration, because the hydroxyl radical concentration at

the photocatalyst surface is high and almost constant (Malato

et al., 2009). However, this is valid only if mass transfer pro-

cesses are faster than the reaction on the photocatalyst sur-

face; for diluted pollutant solutions, the diffusion from the

bulk liquid phase to the photocatalyst surface should affect

the reaction rate (Chong et al., 2010). Probably, this is the

reason linearity is not observed in Fig. 5b after longer reme-

diation periods (for instance, the EE2 concentration is ca.

17 mmol L�1 after irradiation for 150 min). Furthermore, the

photocatalytic oxidation of organic compounds frequently

can be described by a LangmuireHinshelwood mechanism,

which accounts for pollutant adsorption at the catalyst sur-

face. This hypothesis could be verified by determining the

profile of pollutant concentration decay using different values

of its initial concentration (Oliveira et al., 2010). Unfortunately,

since the EE2 solubility limit in water is quite low, ca.

13 mg L�1 (Shareef et al., 2006), it was not possible to perform

these studies using EE2 solutions with higher concentrations.

The resulting EE2 removal percentages after 4 h of reme-

diation using different systems are represented in Fig. 6. The

TOC removal values, as well as the EE2 removal estimated

from fluorescence, are the average of experiments performed

in triplicate; for EE2 removal estimated from HPLC, the mea-

surements were carried out just once. Moreover, HPLC anal-

ysis exhibited a single peak for the remediated EE2 solutions,

which has the same retention time as a fresh EE2 solution.

Fig. 6 e EE2 removal determined by fluorescence, TOC and

HPLC techniques after remediation for 4 h under

polychromatic irradiation at (29 ± 2) �C by photolysis,

heterogeneous photocatalysis (HP) or electro-assisted HP

(EHP) using TiO2 and TiO2/WO3 electrodes.

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 andResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

Comparison of the results presented in Fig. 6 revealed that, for

the same conditions, TOC elimination is much smaller than

EE2 removal estimated by fluorescence and HPLC measure-

ments. The poor TOC decreases suggest that the EE2 complete

oxidation reactions have not occurred in high extensions for

any configuration. Also, some of the intermediate products

should present the same retention time than the EE2 mole-

cule, considering that the EE2 removal calculated by HPLC is

always lower than the corresponding value estimated by

fluorescence. The fluorescence and HPLC analysis performed

for the EE2 solutions irradiated during 4 h without any pho-

tocatalyst identified ca. 90% of the initial hormone concen-

tration, and only 4% of TOC removal. These results revealed

that the EE2 hormone in water cannot be degraded by

photolysis.

Comparing the results obtained using the HP configuration,

the TiO2/WO3 electrode presented better performance than

the TiO2 electrode. This result is coherent with the energy

diagram for these semiconductors, as already discussed

considering Fig. 4. This behavior was also observed comparing

both electrodes in the EHP configuration; moreover, it can be

observed that the TiO2/WO3 electrode biased at þ0.7 V pre-

sented higher current density than the TiO2 electrode under

similar conditions (Figure S7, supporting information). Finally,

comparison of the results obtained for TiO2 electrode revealed

that, in relation to the HP configuration, the EHP system

exhibited higher efficiency for EE2 removal. This was also

observed for the TiO2/WO3 electrode. Polarization of the n-

type semiconductor electrode at positive potentials inhibits

the electron-charge recombination and then improves the

efficiency for EE2 photocatalytic oxidation by ca. 20%, as pre-

viously observed for phenol and for the R6G dye (Oliveira et al.,

2010, 2012).

The literature presents several reports discussing EE2

degradation by different methodologies. Degradation by

photolysis, for instance, was carried out using radiation with

wavelengths of 365 and 254 nm (UVA and UVC, respectively)

for treatment of EE2 solutions in acetonitrile (300 mmol L�1);

after 1 h, EE2 removal was lower than 20% (Kralchevska et al.,

2012). The EE2 decay in acetonitrile solutions was also inves-

tigated using irradiated TiO2 as photocatalyst; after 3 h, the

EE2 removal corresponded to 25% for UVA and 60% for UVB

radiation. The effect of EE2 concentration on the kinetics for

its removal was also discussed, taking into account the

diffusion of EE2 molecules from solution to the photocatalyst

surface (Puma et al., 2010). On the other hand, for EE2 aqueous

solutions, the studies usually are carried out using low EE2

concentrations. Remediating an EE2 solution that initially

contained 0.1mg L�1 with UVA radiation for 30min resulted in

only 20% of EE2 removal. However, the efficiency increased to

90% when immobilized TiO2 was used as photocatalyst in a

quartz flow reactor (Tanizaki et al., 2002). TiO2 immobilized on

a titanium alloy irradiated by UVA was used as photocatalyst

for treatment of aqueous EE2 solutions (C0 ¼ 10 mg L�1); the EE2

decay presented pseudo-first order kinetics with kap of

0.086 min�1, resulting in 100% of EE2 removal after 50 min

(Coleman et al., 2004). A Ti/SnO2 (2� 3 cm2) electrodewas used

for electrochemical treatment of 100 mL of aqueous EE2 so-

lution (C0 ¼ 10 mg L�1) under magnetic stirring; after 120 min

under galvanostatic control (10 mA cm�2), the EE2

n of 17-a-ethinylestradiol aqueous solution by photocatalysisTiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 0 9

concentration decayed to 1.1 mg L�1 (Feng et al., 2010). Thus,

taking into account the diversity in experimental conditions, a

comparison of the results we obtained with those reported in

literature is not straightforward.

Finally, it would be interesting to discuss possible appli-

cations for EHP systems. Currently, the cheapest method of

waste disposal from pharmaceuticals, including prescription

drugs, is landfilling, which presents several inconveniences

(Assamoi and Lawryshyn, 2012). However, for some pharma-

ceutical plant industries, the indicated disposal option still is

incineration, an expensive technology with as many negative

environmental impacts as landfilling, due to the emission of

toxic compounds. Moreover, the cost for such technologies

also includes waste transport from industry to the final

disposal facilities (Bring et al., 2011). Thus, HP-based tech-

nologies could be used for remediation of industrial effluents

“in-situ”. This perspective was considered for the wastewater

of the oral contraceptive production plant of WYETH, St-

Laurent, Canada. A Pyrex photoreactor with 2 g L�1 TiO2 par-

ticles under UVC irradiation was used for treatment of 0.2 L of

aqueous effluents containing levonorgestrel and EE2 (3.5e5

and 0.8e1.6 mg L�1, respectively), individually or in complex

matrices. The hormone removal efficiency was excellent for

solutions containing each hormone individually and 48% in

the wastewater matrix, revealing the applicability of UVC

photocatalysis for reduction of hormone content in waste-

waters (Nasuhoglu et al., 2012). However, taking into account

that a slurry reactor was used, steps for particles separation

still have to be considered in order to evaluate the feasibility of

this technology.

In this study, the highest performance for hormone

removal (54%) was observed for the hybrid electrode in the

EHP configuration; briefly, when an irradiated TiO2/WO3

electrode (1 cm2) biased at 0.7 V was used for remediation of

EE2 aqueous solutions (10 mL) during 4 h, the EE2 concentra-

tion decayed, in average, from 10 to 3.6 mg L�1. In spite of the

relatively poor TOC removal (only 22% of the initial value), the

fluorescence decay was considerable, suggesting that the es-

trogenic activity also decreased for the remediated hormone

solutions. These can be considered promising results; since

the viability was demonstrated, the EHP configuration should

be enhanced for its practical application; a solar cell could be

connected to provide the power to electrochemically-assist

hormone oxidation on the photocatalyst electrode and a

continuous flow system could help overcome the mass

transfer limitations. Lastly, developing of a water treatment

technology based on a solar EHP system would provide a

sustainable alternative for in-situ remediation of industrial

effluents containing contaminants of emerging concern.

4. Conclusions

TiO2/WO3 porous film electrodes, with a molar ratio W/Ti of

only 12%, exhibited higher performance for EE2 removal from

aqueous solution by heterogeneous photocatalysis using a

solar simulator than similar photoanodes of pure TiO2. This

result is coherent with the energy diagram organized for the

semiconductors and EE2. Considering the band gap energies,

adding WO3 enhanced sunlight harvesting; moreover,

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 and TResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

photoinduced charge separation should be improved due to

the relative positions of TiO2 andWO3 conduction and valence

band edges. The electrochemically-assisted photocatalysis

was achieved by introducing a Pt wire in the HP system; then,

when the photoanode was biased at þ 0.7 V, the efficiency for

EE2 removal was increased for both photoelectrodes because

the applied potential suppresses charge recombination. After

4 h under polychromatic irradiation, the hybrid TiO2/WO3

reduced the EE2 concentration from 10 to 3.6 mg L�1. In

addition the total organic carbon removal was only 22%, the

fluorescence decay suggests a decrease in estrogenic activity.

Since electrochemically-assisted photocatalysis could also be

obtained by connection with a solar cell, this study indicates

that EE2 estrogenic activity could be achieved under solar

irradiation, which would be a promising alternative for

remediation of pharmaceutical industrial effluents with con-

taminants of emerging concern.

Acknowledgments

The EE2 sample provided by Organon-Schering Plus and the

assistance of Claudia Martelli, Daniel Razzo, Erika D. Silva,

Priscila A. da Silva and Roy E. Bruns are gratefully acknowl-

edged. This work was supported by CAPES, CNPq, FAPESP and

National Institute of Science, Technology and Innovation on

Advanced Complex Materials (INOMAT).

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2014.08.042.

r e f e r e n c e s

Assamoi, B., Lawryshyn, Y., 2012. The environmental comparisonof landfilling vs. incineration of MSW accounting for wastediversion. Waste Manag. 32 (5), 1019e1030.

Bard, A.J., Faulkner, L.R., 2001. Eletrochemical Methods,Fundamentals and Applications, second ed. JohnWiley & SonsInc, New Jersey USA.

Brink, K., Poulsen, T.G., Skov, H., 2011. Energy and greenhouse gasbalance for solid waste incineration plant: a case study. WasteManag. Res. 29 (10), 13e19.

Chong, M.N., Jin, B., Chow, C.W.K., Saint, C., 2010. Recentdevelopments in photocatalytic water treatment technology: areview. Water Res. 44 (10), 2997e3030 27.

Coleman, H.M., Routledge, E.J., Sumpter, J.P., Eggins, B.R.,Byrne, J.A., 2004. Rapid loss of estrogenicity of steroidestrogens by UVA photolysis and photocatalysis over animmobilized titanium dioxide catalyst. Water Res. 38 (14e15),3233e3240.

De la Cruz, N., Gim�enez, J., Esplugas, S., Grandjean, D.,Alencastro, L.F., Pulgarı́n, C., 2012. Degradation of 32 emergentcontaminants by UV and neutral photo-fenton in domesticwastewater effluent previously treated by activated sludge.Water Res. 46 (6), 1947e1957.

Evans, D.H., 2008. One-electron and two-electron transfers inelectrochemistry and homogeneous solution reactions. Chem.Rev. 108 (7), 2113e2144.

n of 17-a-ethinylestradiol aqueous solution by photocatalysisiO2/WO3 electrodes irradiated by a solar simulator, Water

wat e r r e s e a r c h x x x ( 2 0 1 4 ) 1e1 010

Feng, Y., Wang, C., Liu, J., Zhang, Z., 2010. Electrochemicaldegradation of 17-alpha-ethinylestradiol (EE2) and estrogenicactivity changes. J. Environ. Monit. 12 (2), 404e408.

Gratzel, M., 2001. Photoelectrochemical cells. Nature 414,338e344.

Hamid, H., Eskicioglu, C., 2012. Fate of estrogenic hormones inwastewater and sludge treatment: a review of properties andanalytical detection techniques in sludge matrix. Water Res.46 (18), 5813e5833.

Higashimoto, S., Ushiroda, Y., Azuma, M., 2008. Electrochemicallyassisted photocatalysis of hybrid WO3/TiO2 films: effect of theWO3 structures on charge separation behavior. Top. Catal. 47(3e4), 148e154.

Kralchevska, R., Milanova, M., Bistan, M., Pintar, A.,Todorovsky, D., 2012. The photocatalytic degradation of 17a-ethynylestradiol by pure and carbon nanotubes modified TiO2

under UVC illumination. Central Eur. J. Chem. 10 (4),1137e1148.

Lin, C.F., Wu, C.H., Onn, Z.N., 2008. Degradation of 4-chlorophenolin TiO2, WO3, SnO2, TiO2/WO3 and TiO2/SnO2 systems. J.Hazard. Mater. 154 (1e3), 1033e1039.

Longo, C., De Paoli, M.-A., 2003. Dye-sensitized solar cells: asuccessful combination of materials. J. Braz. Chem. Soc. 14 (6),889e901.

Malato, S., Fern�andez-Ib�a~nez, P., Maldonado, M.I., Blanco, J.,Gernjak, W., 2009. Decontamination and disinfection of waterby solar photocatalysis: recent overview and trends. Catal.Today 147 (1), 1e59.

Montagner, C.C., Jardim, W.F., 2011. Spatial and seasonalvariations of pharmaceuticals and endocrine disruptors in theAtibaia River, S~ao Paulo State (Brazil). J. Braz. Chem. Soc. 22 (8),1452e2146.

Murphy, A.B., 2007. Band-gap determination from diffusereflectance measurements of semiconductor films, andapplication to photoelectrochemical water-splitting. Sol.Energy Mater. Sol. Cells 91 (14), 1326e1337.

Myers, D., 1999. Surface, Interfaces and Colloids: Principles andApplication, second ed. Jonh Wiley & Sons Inc, TorontoCanada.

Nasuhoglu, D., Berk, D., Yargeau, V., 2012. Photocatalytic removalof 17a-ethinylestradiol (EE2) and levonorgestrel (LNG) fromcontraceptive pill manufacturing plant wastewater under UVCradiation. Chem. Eng. J. 185e186, 52e60.

Ohko, Y., Iuchi, K., Niwa, C., Tatsuma, T., Nakashima, T.,Iguchi, T., Kubota, Y., Fujishima, A., 2002. 17 b-estradioldegradation by TiO2 photocatalysis as means of reducingestrogenic activity. Environ. Sci. Technol. 36 (19), 4175e4181.

Oliveira, H.G., Fitzmorris, B.C., Longo, C.,Z., Zhang, J.Z., 2012.Photoelectrochemical and photocatalytic properties of TiO2,WO3 and WO3-TiO2 porous films in the photodegradation of

Please cite this article in press as: Oliveira, H.G., et al., Remediatioand electrochemically-assisted photocatalysis using TiO2 andResearch (2014), http://dx.doi.org/10.1016/j.watres.2014.08.042

rhodamine 6G in aqueous solution. Sci. Adv. Mater. 4 (5e6),673e680.

Oliveira, H.G., Nery, D.C., Longo, C., 2010. Effect of appliedpotential on photocatalytic phenol degradation usingnanocrystalline TiO2 electrodes. Appl. Catal. B: Environ. 93(3e4), 205e211.

Puma, G.L., Puddu, V., Tsang, H.K., Gora, A., Toepfer, B., 2010.Photocatalytic oxidation of multicomponent mixtures ofestrogens (estrone (E1), 17b-estradiol (E2), 17a-ethynylestradiol (EE2) and estriol (E3)) under UVA and UVCradiation: photon absorption, quantum yields and rateconstants independent of photon absorption. Appl. Catal. B:Environ. 99 (3e4), 388e397.

Qu, X., Alvarez, P.J.J., Li, Q., 2013. Applications of nanotechnologyin water and wastewater treatment. Water Res. 47 (12),3931e3946.

Radecka, M., Rekas, M., Trenczek-Zajac, A., Zakrzewska, K., 2008.Importance of the band gap energy and flat band potential forapplication of modified TiO2 photoanodes in water photolysis.J. Power Sources 181 (1), 46e55.

Reyes-Coronado, D., Rodrı́guez-Gattorno, G., Espinosa-Pesqueira, M.E., Cab, C., Coss, R., Oskam, G., 2008. Phase-pureTiO2 nanoparticles: anatase, brookite and rutile.Nanotechnology 19 (14), 145605e145615.

Rokhina, E.V., Suri, R.P.S., 2012. Application of density functionaltheory (DFT) to study the properties and degradation ofnatural estrogen hormones with chemical oxidizers. Sci. TotalEnviron. 417, 280e290.

Santos, R.S., Faria, G.A., Giles, C., Leite, C.A.P., Barbosa, H.S.,Arruda, M.A.Z., Longo, C., 2012. Iron insertion and hematitesegregation on Fe-doped TiO2 nanoparticles obtained fromsol�gel and hydrothermal methods. Appl. Material Interfaces4 (10), 5555e5561.

Shareef, A., Angove, M.J., Wells, J.D., Bruce, Johnson, B.B., 2006.Sorption of bisphenol A, 17a-ethynylestradiol and estrone tomineral surfaces. J. Colloid Interface Sci. 297 (1), 62e69.

Tanizaki, T., Kadokami, K., Shinohara, R., 2002. Catalyticphotodegradation of endocrine disrupting chemicals usingtitanium dioxide photosemiconductor thin films. Bull.Environ. Contam. Toxicol. 68, 732e739.

Vega, D., Agui, L., Gonzalez-Cortes, A., Yanes-Seden~o, P.,Pingarr�on, J.M., 2007. Electrochemical detection of phenolicestrogenic compounds at carbon nanotube modifiedelectrodes. Talanta 71 (3), 1031e1038.

Yang, H., Shi, R., Zhang, K., Hu, K., Tang, A., Li, X., 2005. Synthesisof WO3/TiO2 nanocomposites via sol-gel method. J. AlloysCompd. 398 (1e2), 200e202.

Zhang, X., Chen, P., Wu, F., Deng, N., Liu, J., Fang, T., 2006.Degradation of 17 a-ethinylestradiol in aqueous solution byozonation. J. Hazard. Mater. B 133 (1e3), 291e298.

n of 17-a-ethinylestradiol aqueous solution by photocatalysisTiO2/WO3 electrodes irradiated by a solar simulator, Water