wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 4

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev ier . com/ loca te /wat res

Electrochemical enhancement of solar photocatalysis:Degradation of endocrine disruptor bisphenol-A on Ti/TiO2

films

Zacharias Frontistis a, Vasileia M. Daskalaki a, Alexandros Katsaounis a, Ioannis Poulios b,Dionissios Mantzavinos a,*aDepartment of Environmental Engineering, Technical University of Crete, Polytechneioupolis, GR-73100 Chania, Greeceb Laboratory of Physical Chemistry, Department of Chemistry, Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece

a r t i c l e i n f o

Article history:

Received 9 December 2010

Received in revised form

8 March 2011

Accepted 15 March 2011

Available online 21 March 2011

Keywords:

BPA

Current

Kinetics

Promotion

Titania

Water

* Corresponding author. Tel.: þ302821037797E-mail address: mantzavi@mred.tuc.gr (D

0043-1354/$ e see front matter ª 2011 Elsevdoi:10.1016/j.watres.2011.03.030

a b s t r a c t

The photoelectrocatalytic oxidation over immobilized Ti/TiO2 films in the presence of

simulated solar light was investigated for the degradation of bisphenol-A (BPA) in water.

The catalyst, consisting of 75:25 anatase:rutile, was prepared by a sol-gel method and

characterized by cyclic voltammetry, X-ray diffraction and scanning electron microscopy.

Experiments were conducted to assess the effect of applied current (0.02e0.32 mA/cm2),

TiO2 loading (1.3e9.2 mg), BPA concentration (120e820 mg/L), initial solution pH (1 and 7.5)

and the aqueous matrix (pure water and treated effluent) on BPA photoelectrocatalytic

degradation which was monitored by high performance liquid chromatography equipped

with a fluorescence detector. The reaction was favored at anodic currents up to 0.04 mA/

cm2 and lower substrate concentrations, but it was hindered by the presence of residual

organic matter and radical scavengers (e.g. bicarbonates) in treated effluents. Moreover,

a pseudo-first order kinetic model could fit the experimental data well with the apparent

reaction constant taking values between 2.9 and 32.4 10�3/min. The degradation of BPA by

pure photocatalysis or electrochemical oxidation alone was also studied leading to partial

substrate removal. In all cases, the contribution of applied potential to photocatalytic

degradation was synergistic with the photocatalytic efficiency increasing between 24% and

97% possibly due to a more efficient separation and utilization of the photogenerated

charge carriers. The effect of photoelectrocatalysis on the ecotoxic and estrogenic prop-

erties of BPA was also evaluated measuring the bioluminescence inhibition of Vibrio fischeri

and performing the yeast estrogen screening assay, respectively.

ª 2011 Elsevier Ltd. All rights reserved.

1. Introduction surface waters, effluents from wastewater treatment plants,

Bisphenol-A (BPA) is widely used as raw material in the

manufacturing of numerous chemical products, such as

polycarbonate plastics and epoxy resins. Due to its wide-

spread usage, it has been detected in treated drinking water,

; fax: þ302821037852.. Mantzavinos).ier Ltd. All rights reserve

landfill leachates, sediments from lakes, rivers and channels

and tissues of aquatic animals (Garoma et al., 2010). BPA is

thought to be associated with endocrine disruption (Kang

et al., 2007), as well as several other adverse effects

including genotoxicity (Park and Choi, 2009), increasing

d.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 4 2997

cancer cells (Jenkins et al., 2009) and sperm count reduction

(Salian et al., 2009).

In recent years, there have been intensive efforts toward

the development of efficient technologies for the removal of

persistent xenobiotics from aqueous matrices. In this

perspective, advanced oxidation processes are likely to play

a key role and several recent studies have dealt with BPA

degradation by ozonation (Garoma et al., 2010), ultrasound

irradiation (Petrier et al., 2010), dark- (Poerschmann et al.,

2010) and photo-Fenton (Zhan et al., 2006) oxidation, electro-

chemical oxidation (Cui et al., 2009), as well as various hybrid

processes (Torres et al., 2008; Huang and Huang, 2009; Torres-

Palma et al., 2010).

Over the past several years, heterogeneous photocatalysis

has received enormous attention for the treatment of various

classes of organic contaminants found in waters and waste-

waters. Of the various semiconductors, TiO2 has almost

exclusively been employed in environmental applications and

the fundamentals of the process can be summarized as

follows (Malato et al., 2009): the electronic structure of TiO2,

consisting of an empty conduction band and a filled valence

band, facilitates the formation of electron/hole pairs when the

semiconductor absorbs photonic energy greater than its band

gap energy of about 3.2 eV, i.e. at wavelengths below about

390 nm. Holes are strong oxidizing agents and electrons are

good reducing agents, therefore both promote redox reac-

tions. Most organic photodegradation reactions utilize the

oxidizing power of holes either directly or indirectly, i.e.

through the formation of hydroxyl radicals and other reactive

oxygen species.

In recent studies, BPA degradation has been investigated

bymeans of UVA (Tsai et al., 2009; Guo et al., 2010) and natural

solar irradiation (Rodriguez et al., 2010) over TiO2 suspensions

with emphasis given on the effect of various operating

conditions on performance. Although solar photocatalysis is

an attractive and sustainable water treatment technology, it

suffers a major drawback, i.e. commercially available, pure

TiO2 can only be activated by UVA light which comprises just

3e5% of the solar spectrum. Enhanced catalytic activity for the

degradation of BPA under visible light was achieved doping

either pure TiO2 (Subagio et al., 2010) or supported on acti-

vated carbon (Yap et al., 2010) with nitrogen, co-doping TiO2

with nitrogen and carbon (Wang and Lim, 2010) and adding

polyethylene glycol (Kuo et al., 2010) in an attempt to narrow

the band gap of titania and/or change its surface properties.

All the aforementioned studies (Tsai et al., 2009; Guo et al.,

2010; Kuo et al., 2010; Rodriguez et al., 2010; Subagio et al.,

2010; Wang and Lim, 2010; Yap et al., 2010) refer to slurry

systems that suffer a serious drawback, i.e. the need to

separate the fine catalyst particles from the treated solution.

This can be overcome immobilizing the catalyst on suitable

supports or carriers; Wang et al. (2009) studied BPA degrada-

tion by UVC irradiation over TiO2 fixed to polyurethane foam

cubes, while Nakashima et al. (2002) immobilized TiO2 on PTFE

mesh sheets to degrade BPA and other endocrine disruptors

under UVA irradiation. Nonetheless, the quantum efficiency

of immobilized catalyst is generally lower than that of the

suspended one due to increased mass transfer limitations of

the contaminant to the catalyst and a decrease of its surface

area (Philippidis et al., 2010).

Away to improve the performance of immobilized TiO2 is to

fix the catalyst on a conductive substrate and apply an external

voltage in a photoelectrochemical cell, thus driving the photo-

generated electrons to the cathode and, consequently, mini-

mizing the rate of electron/hole recombination. The process,

which isalsoreferredtoasphotoelectrocatalysis (PEC),hasbeen

recently demonstrated for the UVA-induced degradation of

aniline (Ku et al., 2010) and the inactivation of Escherichia coli

colonies (Philippidis et al., 2010) onTi/TiO2 electrodes at applied

potential values up to 2 V vs Ag/AgCl, the solar light-induced

degradation of rhodamine B on nitrogen-doped Ti/TiO2 elec-

trodes up to 1.4 V vs saturated calomel electrode (SCE) (Han

et al., 2010) and the solar light-induced degradation of phenol

on Si/TiO2 arrays at 3 V vs SCE (Yu et al., 2009).

ThedegradationofBPAbyPEChasonlymerelybeenreported

in the literature. Li et al. (2005) prepared Au/TiO2/(IneSn oxide)

filmsandassessed the contributionof gold to thedegradationof

16mg/L BPA under UVA irradiation and up to 8 V vs SCE applied

potential; the same research group (Xie and Li, 2006) also tested

Au/TiO2/Tifilmsfor thedegradationof11.2mg/LBPAunderUVA

orvisible irradiationandupto1.5mAappliedcurrent. Inarecent

study, Brugnera et al. (2010) synthesized TiO2 nanotubular

arrays for the degradation of 22.8 mg/L BPA under UVA irradia-

tion and up to 1.5 V vs Ag/AgCl potential.

The scope of this work was to investigate BPA degradation

by solar photoelectrocatalysis on Ti/TiO2 films and compare

its efficiency with photocatalysis (PC) and electrochemical

oxidation (EO). The effect of various operating conditions such

as applied current, BPA concentration, catalyst loading, solu-

tion pH and the water matrix on kinetics has been examined.

2. Materials and methods

2.1. Catalyst preparation

Immobilized TiO2 films on Ti substrates were prepared by

a sol-gel procedure, as follows: 4.2 g of the non-ionic surfac-

tant Triton X-100 (polyoxyethylene-(10) isooctylphenyl ether)

slurry were mixed with 22.8 mL of ethanol, followed by the

addition of 3.9mL of glacial acetic acid and 2.16mL of titanium

isopropoxide under vigorous stirring. Self organization of the

surfactant in this original sol creates organized assemblies

which act as templates defining nanoparticle size. The

surfactant was burned out during calcination. After stirring

for a few minutes, a squared shaped titanium substrate

(2.4� 2.4 cm)was dipped in the above solution andwithdrawn

at a constant speed. The nanocomposite film formed by

dipping was left to dry in air for a fewminutes and then it was

calcined in an oven at 550 �C for 10 min (Bouras and Lianos,

2005). When the film was taken out of the oven it was trans-

parent and optically uniform. The above procedure was

repeated until a certain amount of titania in the range

1.3e9.2 mg (the selection was based on preliminary tests,

previous experience and technical constraints) was deposited

on the Ti substrate; the latter was first sandblasted to ensure

good adhesion of the deposit on its surface followed by

a chemical treatment (using a 1 M oxalic acid solution at the

boiling point for 60 min) in order to obtain a totally clean

surface.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 42998

2.2. Experimental procedure

Photocatalytic experiments were performed using a solar

simulator (Oriel, model 96000) equipped with a 150 W xenon,

ozone free lamp in an open cylindrical pyrex cell at ambient

conditions under continuous stirring. Current in the range

0.02e0.32 mA/cm2 was applied using a galvanostatepotentio-

stat (Amel Instruments, model 2053) with the Ti/TiO2 film

operatingas theanode,whilea zirconiumplatewasusedas the

cathode. The experimental setup is illustrated in Scheme 1.

In a typical experiment, the appropriate amount of BPA

was dissolved in ultrapure water (UPW) (EASYpureRF e

Barnstead/Thermolyne, USA) and the resulting solution (in

the range 120e820 mg/L BPA) was added 5 mM HClO4 as the

supporting electrolyte; 60 mL were then introduced in the

reaction vessel and left for 20 min in the dark to equilibrate

prior to applying solar irradiation and/or current. To assess

the effect of water matrix, experiments were also conducted

with BPA spiked in the treated effluent (TE) of the activated

sludge process of the municipal wastewater treatment plant

of Chania, W. Crete, Greece. The effluent had 8.4 mg/L of

dissolved organic carbon (DOC), while its pH and conductivity

were about 8 and 0.81mS/cm, respectively. The concentration

of chlorides, sulfates, nitrates, nitrites and bicarbonates was

222.1, 60.3, 25.9, 57.1, and 182.1 mg/L, respectively.

2.3. Analytical techniques

Changes in BPA concentration were followed by high perfor-

mance liquid chromatography (HPLC, Waters 2690) equipped

with a Luna 54 (18C2) 100 A column and two detectors con-

nected in series, namely a diode array detector (Waters 996)

and a fluorescence detector (Waters 474). The mobile phase

Scheme 1 e Experimental setup.

consisted of 65:35 acetonitrile:water at a flow rate of 1mL/min

and ambient temperature. BPA was monitored by the fluo-

rescence detector, while the diode array set at 280 nm was

used to identify possible reaction by-products.

DOC was measured by NDIR gas analysis on a Shimadzu

5050A TOC analyzer.

The catalyst was characterized for its phase composition

bymeans of X-ray diffraction (XRD) using a Philips, PW1830/40

powder diffractometer. Scanning electron microscopy (SEM)

was performed using a JEOL JMS-6300-F microscope.

The electrochemical/photoelectrochemical characterization

of theTiO2filmswascarriedout inasingle-compartment, three-

electrodecelldescribed indetail elsewhere (Chatzisymeonetal.,

2010) in conjunction with the solar simulator. The electrolyte

volume in the compartmentwas 40mL. A platinumwire served

as thecounterelectrode,whileHg/Hg2SO4/K2SO4 (MSE) (Ref-621,

Radiometer Analytical) was employed as the reference elec-

trode. Voltammograms were run for at least 3 times since

preliminary experiments showed that steady state conditions

could be reached after the third run.

2.4. Ecotoxicity and estrogenicity assays

The luminescent marine bacteria Vibrio fischeri were used to

assess the acute ecotoxicity of BPA prior to and after PEC

treatment. The inhibition of bioluminescence of V. fischeri

exposed to undiluted BPA solutions for 15 min was measured

using a LUMIStox analyzer (Dr Lange, Germany) and the

results were compared to an aqueous control.

The yeast estrogen screening (YES) bioassay was employed

to assess the effect of PEC on estrogenicity according to the

protocol developed byRoutledge and Sumpter (1996)with some

modifications. In brief, standard solutions of estrogen 17b-

estradiol and sample extracts were produced in ethanol and

10mLofdilutionseriesweredispensed into triplicatewellsof96-

well microtiter plates. After evaporation to dryness at room

temperature, 0.2 mL of growth medium containing chlor-

ophenol red-b-D-galactopyranoside and the yeast cells were

added, followed by incubation at 32 �C for 72 h. The absorbance

of the medium was measured using a microplate reader (LT-

4000MSMicroplate Reader, Labtech). The absorbance at 540 nm

was regarded as estrogenic activity after subtraction of absor-

bance at 640 nm to correct for yeast growth.

3. Results and discussion

3.1. Catalyst characterization

XRD patterns were taken in the range of 2q between 20� and

70� with a scanning rate of 0.05�/s (Fig. 1). As seen, TiO2 films

can be successfully developed on Ti substrate containing

peaks that are attributed to both the anatase and rutile pha-

ses. Phase composition was estimated by the integral inten-

sities of anatase (101) and rutile (110) reflections (Cullity, 1978)

and it was found to be about 75% and 25%, respectively. The

primary crystallite size of TiO2 was estimated at 25 � 3 mm

applying the Scherrer’s equation (Cullity, 1978). On the other

hand, SEM images (Fig. 1) reveal that the Ti surface is highly

porous due to its sandblasting and subsequent chemical

Fig. 1 e XRD patterns (top) and SEM images (bottom) of Ti/TiO2 film and Ti support.

-5

0

5

10

15

20

25

-0.5 0 0.5 1 1.5 2 2.5 3E, V vs MSE

I, µ

A/c

m2

Dark

Solar

Fig. 2 e Linear sweep voltammograms of Ti/TiO2 in the

dark and under solar illumination. Sweep rate [ 50 mV/s;

pHo [ 1.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 4 2999

treatment; this allows the formation of a continuous network

of TiO2 layers with large surface area, where more catalyst

active centers are available for redox reactions to occur.

3.2. Photoelectrochemical characterization

As most applications of photoelectrochemical systems

involve the transfer of electrons across the solid/electrolyte

interface, current density-applied potential recording tech-

niques are commonly used for their characterization. Linear

sweep voltammograms of Ti/TiO2 electrode in 0.1 M HClO4 in

the dark and under solar illumination, recorded between �0.4

andþ3 V vs MSE at a sweep rate of 50 mV/s, are given in Fig. 2.

The low dark current density between�0.4 and 1.8 V is typical

of the n-type semiconductive behavior of the synthesized TiO2

specimens, while the observed increase at potentials greater

than þ1.8 V vs MSE can be attributed to water decomposition

and oxygen evolution. At potentials more negative than

�0.4 V, both a cathodic peak and an ill-defined anodic humpor

peak occur (not shown), with the former corresponding to the

reduction of surface Ti(IV) species and the latter to the re-

oxidation of surface Ti(III) species.

On the contrary, upon illumination a significant increase in

the anodic current density above �0.4 V vs MSE occurred, as

a result of the photogenerated holes. According to the

literature (Morrison, 1980; Finklea, 1988; Memming, 2001),

illumination of a semiconductoreelectrolyte interface with

light energy greater than its band gap energy generates elec-

tronehole pairs at the electrode surface. The simultaneous

application of a bias positive to the flat-band potential

produces a bending of the conduction and valence bands

0

0.2

0.4

0.6

0.8

1

C/C

o

PECPCEO

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 43000

which, in turn, causes a more effective separation of the

photogenerated carriers within the space charge layer; this

increases the photocurrent that begins to flow, thus

promoting the oxidative degradation process. The potential

gradient forces the electrons toward the counter electrode,

thus leaving the photogenerated holes to react with H2O and/

or OH� to yield hydroxyl radicals or to attack directly the

organics present in the solution, i.e.:

Anode (working electrode):

TiO2 þ hn/TiO2 � e�cb þ TiO2 � hþ

vb (1)

0 30 60 90 120 150 180time, min

Fig. 3 e Variation of normalized BPA concentration during

PEC, PC and EO. Conditions: [BPA]o [ 300 mg/L in UPW;

I [ 0.08 mA/cm2; pHo [ 1; TiO2 loading [ 2.6 mg.

TiO2 � hþvb þH2Os/TiO2 �OHs$þHþ (2)

TiO2 � hþvb þOH�

s /TiO2 �OHs$ (3)

TiO2 � e�cb þ TiO2 � hþ

vb/recombination (4)

Cathode (counter electrode):

2H2O þ 2e� / H2 þ 2OH� (5)

where the subscripts cb and vb denote the conduction and

valence bands, respectively and hþ and e� denote the photo-

generated holes and electrons, respectively.

3.3. Efficiency of PEC in relation to PC and EO

Fig. 3 shows normalized BPA concentrationetime profiles

during PEC, PC and EO treatments at 300 mg/L initial concen-

tration and pHo ¼ 1. PC oxidation occurs slowly yielding only

30% conversion after 180 min of reaction; nonetheless,

conversion becomes quantitative when 0.08 mA/cm2 of

current is applied. The enhanced PEC performance may only

partly be attributed to electrochemical degradation reactions

since EO alone results in about 10% conversion. The applica-

tion of a positive current producing a potential greater than

the flat-band potential of the Ti/TiO2 electrode decreases the

charge recombination process (eq. (4)), which is essential for

promoting substrate degradation.

3.4. Effect of applied current

Fig. 4 shows the effect of varying current in the range

0.02e0.32 mA/cm2 on BPA degradation by PEC and EO treat-

ments at 300 mg/L initial concentration and pHo ¼ 1. There

appears to be a threshold current at about 0.02 mA/cm2 below

which the promotion of PC is not evident, i.e. 37% and 31% BPA

conversion was recorded after 180 min at 0.02 and 0 mA/cm2,

respectively. Increasing current from 0.02 to 0.04 mA/cm2 had

a pronounced effect on PC leading to complete BPA removal

after 120e180 min, while applying higher currents did not

practically improve degradation (e.g. in terms of final

conversion).

Unlike PEC, increasing current from 0.02 to 0.16 mA/cm2

enhanced EO performance, i.e. the final BPA conversion was

5%, 18% and 24% at 0.02, 0.04 and 0.16 mA/cm2, respectively

(Fig. 4b). Notably, operation at 0.32 mA/cm2 led to decreased

performance and this may be attributed to the fact that the

electrode partially disintegrated loosing 25% of its mass after

180 min of reaction. This phenomenon did not occur at lower

currents as evidenced weighing the electrodes before and

after the reaction. Conversely, PEC at 0.32 mA/cm2 was not

affected by the electrode’s partial disintegration and this may

be due to the fact that (i) the remaining electrode is still

active enough to induce fast BPA degradation (see Section

3.5), and/or (ii) the leached TiO2 (0.67 mg) contributes signif-

icantly to the PC reaction rate, thus compensating any

activity loss due to dissolution. This was verified in prelimi-

nary PC runs with 0.65 mg of Degussa, P-25 TiO2 either

immobilized on Ti or suspended in the reaction mixture; the

immobilized catalyst gave far lower conversions than the

respective slurry system.

Previous studies (Li et al., 2005; Brugnera et al., 2010) have

shown that BPA degradation by PEC can be described by

pseudo-first order kinetics, i.e.:

�dCdt

¼ kPECC5lnCo

C¼ kPECt (6)

where kPEC is an apparent kinetic constant. If the results of

Fig. 4a are plotted in the form of eq. (6), straight lines passing

through the origin (not shown) fit the experimental data very

well. From the slopes of the resulting lines, kPEC values can

be computed and they are summarized in Table 1 alongside

the coefficients of linear regression; as seen, kPEC at 0.02 mA/

cm2 is an order of magnitude lower than at 0.04e0.16 mA/

cm2.

Likewise, data from the respective PC experiments were

also fitted to pseudo-first order kinetics (not shown) from

which the respective kPC constants were computed; these,

alongside kPEC values, were employed to quantify the extent of

photoelectrochemical enhancement (E ), as follows:

E ¼ kPEC � kPC

kPEC(7)

As also seen in Table 1, the extent of enhancement is only

about 24% at 0.02 mA/cm2 but it increases to 90.5 � 1.2% at

currents between 0.04 and 0.16 mA/cm2.

Table 1 e Apparent kinetic constants (eq. (6)) and extentof enhancement (eq. (7)) associated with PEC degradationof BPA at various conditions. Numbers in brackets showcoefficients (r2) of linear regression (eq. (6)). ND: notdetermined because PC run was not performed.

Co,mg/L

I,mA/cm2

pHo Matrix Catalystloading, mg

kPEC, 10�3/

minE,%

300 0.02 1 UPW 2.6 2.9 (0.930) 24.1

300 0.04 1 UPW 2.6 20.8 (0.999) 89.4

300 0.08 1 UPW 2.6 20.6 (0.974) 89.3

300 0.16 1 UPW 2.6 26.4 (0.983) 91.7

300 0.16 1 UPW 9.2 23.6 (0.996) 72

300 0.16 1 UPW 1.3 20.5 (0.998) 90.2

120 0.16 1 UPW 2.6 32.4 (0.987) 97.2

500 0.16 1 UPW 2.6 19.5 (0.988) 89.2

820 0.16 1 UPW 2.6 11.2 (0.999) ND

300 0.16 7.5 UPW 2.6 22.8 (0.998) 53.1

300 0.16 8 TE 2.6 10.1 (0.977) 85.1

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120 150 180time, min

C/C

o

PC-2.6mgPC-9.2mgPEC-2.6mgPEC-9.2mgPC-1.3mgPEC-1.3mg

Fig. 5 e Variation of normalized BPA concentration during

PEC and PC as a function of catalyst loading. Conditions:

[BPA]o [ 300 mg/L in UPW; pHo [ 1; I [ 0.16 mA/cm2.

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120 150 180time, min

C/C

o 0 mA/cm20.02 mA/cm20.04 mA/cm20.16 mA/cm20.32 mA/cm2

a

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120 150 180time, min

C/C

o

0.02 mA/cm2

0.04 mA/cm2

0.16 mA/cm2

0.32 mA/cm2

b

Fig. 4 e Variation of normalized BPA concentration during

(a) PEC and (b) EO as a function of applied current.

Conditions: [BPA]o [ 300 mg/L in UPW; pHo [ 1; TiO2

loading [ 2.6 mg.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 4 3001

3.5. Effect of catalyst loading

Fig. 5 shows the effect of increasing catalyst loading from 1.3

to 9.2 mg on BPA degradation by PEC and PC treatments at

300 mg/L initial concentration and pHo ¼ 1. An increase in

catalyst loading had practically little effect on PEC degradation

with kPEC being 23.5� 3 10�3/min at any loading. Nevertheless,

the extent of enhancement decreased from 91 � 0.8% at

1.3e2.6 mg to 72% at 9.2 mg; this is due to the fact that the

respective kPC value, which remained unchanged between 1.3

and 2.6 mg, increased by as much as three times between 2.6

and 9.2 mg catalyst.

In an immobilized PC system, the reactant diffuses from

the bulk solution through a boundary layer to reach the elec-

trode double layer, where it gets adsorbed onto the active sites

of the catalyst and reacts. The optimum film thickness

depends on the light penetration depth and the width of the

space charge layer. At reduced catalyst loadings, fewer surface

sites are available for reactions including the generation of

fewer holes and electrons, as well as secondary oxidizing

species. In the case of PEC, the efficient separation of charge

carriers and, consequently, their availability to react for longer

periods seems to be critical, thus compensating the reduced

rate of generation at lower loadings.

3.6. Effect of BPA concentration

The effect of initial BPA concentration on PEC degradationwas

also investigated at 0.16 mA/cm2, 2.6 mg catalyst loading and

pHo ¼ 1 and the results are shown in Fig. 6 and Table 1. As

seen, increasing BPA concentration resulted in decreased

kinetics; for example, kPEC was 32.4, 26.4, 19.5 and 11.2 10�3/

min at 120, 300, 500 and 820 mg/L, respectively. At a constant

catalyst loading, performance will be dictated by the catalyst

sites to substrate molecules ratio. The fact that degradation

decreases at higher initial BPA concentrations may be asso-

ciated with the availability of fewer active sites, thus trig-

gering a competitive adsorption between BPA and reaction by-

products (whose determination was not feasible with the

analytical protocols employed in this study) onto the catalyst

surface which thereby decreases the concentration of

oxidizing species attacking BPA (Jain and Shrivastava, 2008).

0

150

300

450

600

750

0 30 60 90 120 150 180time, min

BP

A c

on

ce

ntra

tio

n, µ

g/L

120µg/L

300µg/L

500µg/L

820µg/L

Fig. 6 e Variation of BPA concentration in UPW during PEC

as a function of initial concentration. Conditions:

I [ 0.16 mA/cm2; pHo [ 1; TiO2 loading [ 2.6 mg.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 43002

3.7. Effect of pH and the water matrix

In further studies, BPA degradation in UPW was studied at

pHo ¼ 7.5 and the results are shown in Fig. 7. Comparing the

performance at acidic and near-neutral conditions (Figs. 4 and

7 and Table 1), it is evident that PEC degradation was nearly

equally fast at either conditions with kPEC being 26.4 and

22.8 10�3/min at pHo 1 and 7.5, respectively. (It should be

pointedouthere that althoughpHwas left uncontrolled during

the reactions it did not change more than 0.1e0.2 units.)

Interestingly and unlike PEC, the individual processes were

favored at pHo ¼ 7.5 with the final conversion being 85% and

63% for PC and EO, respectively, while the corresponding

values at acidic conditions were only 31% and 24%. In terms of

kinetics, kPC increased by asmuch as about 5 times going from

acidic to near-neutral conditions and this explains the

substantial drop of process enhancement from 91.7% to 53.1%.

At thepHvalues under consideration, BPA is inmolecular form

in the solution since its pKa value is 9.6e10.2 (Bautista-Toledo

et al., 2005). In this respect, the pH effect cannot be explained

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120 150 180time, min

C/C

o

PC/TEEO/TEPEC/TEPC/UPWEO/UPWPEC/UPW

Fig. 7 e Variation of normalized BPA concentration during

PEC, PC and EO in UPW at pHo [ 7.5 and TE at pHo [ 8.

Conditions: [BPA]o [ 300 mg/L; I [ 0.16 mA/cm2; TiO2

loading [ 2.6 mg.

on the basis of electrostatic interactions between the ampho-

teric catalyst surface (which would be positively charged at

pHo¼ 1 andnegatively chargedat pHo¼ 7.5) and theuncharged

BPA molecules. The reduced performance at pHo ¼ 1 could

partially be due to the fact that the formation of hydroxyl

radicals is suppressed at lower pH values (Hapeshi et al., 2010).

In a final set of experiments, process performance was

evaluated in a secondary treated effluent at inherent condi-

tions (i.e. pHo¼ 8 and 0.81mS/cmconductivity) and the results

are also shown in Fig. 7. The presence of 8.4 mg/L of organic

carbon (this is 35 times the carbon contained in 300 mg/L BPA)

impeded degradation and this was more pronounced for PEC

and PC. As seen in Table 1, kPEC in TE decreased by about 50% in

relation to UPW, while the respective reduction for kPC was

7-fold. Nevertheless, the PEC degradation of BPA became

quantitative after 180minof reaction. Photogenerated reactive

species are partly consumed to attack effluent organic matter

(this is consistent with the fact that reaction rates decrease

with increasing carbon concentration) and this explains the

reduced performance in TE samples compared to UPW

samples. Furthermore, the presence of inorganics like bicar-

bonates that act as radical scavengersmay, to some degree, be

responsible for decreased conversions in TE.

3.8. Removal of ecotoxicity and estrogenicity

The effect of PEC on the ecotoxic and estrogenic properties of

BPA is shown in Fig. 8. BPA is only slightly toxic to marine

bacteria with 38% inhibition of bioluminescence being recor-

ded at 300 mg/L; this value decreased to 26% after 120 min of

reaction, which corresponds to quantitative BPA removal

(Fig. 7). In parallel, BPA exhibits 5 mg/L of equivalent estro-

genicity and this is reduced by 25% after 60 min; these results

imply that PEC degradation by-products are less ecotoxic and

estrogenic than BPA.

It should be noted here that identification of likely reaction

by-products was not feasible with the analytical tools and

protocols employed in thiswork. Nonetheless, an attemptwas

made to perform a carbon balance in the liquid phase

following BPA and DOC profiles during PEC degradation in

0

10

20

30

40

50

0 30 60 90 120time, min

In

hib

itio

n, %

0

1

2

3

4

5V. fischeriEstrogenicity E

qu

iv

ale

nt e

stro

ge

nic

ity

, µ

g/L

Fig. 8 e Variation of inhibition to V. fischeri (left axis) and

estrogenicity (right axis) during PEC in UPW at pHo [ 7.5.

Conditions: [BPA]o [ 300 mg/L; I [ 0.16 mA/cm2; TiO2

loading [ 2.6 mg.

0

800

1600

2400

3200

4000

0 30 60 90 120time, min

Ca

rb

on

c

on

ce

ntra

tio

n, µ

g/L

DOC

BPA

Fig. 9 e Organic carbon balance during PEC of 5 mg/L BPA in

UPW. Conditions: I [ 0.16 mA/cm2; TiO2 loading [ 2.6 mg;

pHo [ 7.5.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 4 3003

UPW; the experiment was carried out at an elevated BPA

concentration of 5mg/L (this corresponds to 4mg/L of DOC) to

allow for reliable DOC analysis. As clearly seen in Fig. 9, most

of the remaining organic carbon in the liquid phase is due to

unreacted BPA; for instance, the carbon contained in BPA

accounts for 82% of DOC after 120 min of reaction, where 53%

of BPA has been removed. This implies that most of the by-

products do not accumulate in the liquid but they are rapidly

mineralized to carbon dioxide and water.

4. Conclusions

The degradation of bisphenol-A, an emerging aqueous phase

pollutant, has been investigated by means of solar irradiation

over fixed Ti/TiO2 catalysts enhanced by the simultaneous

application of an electric field. The major conclusions drawn

from this study are summarized as follows:

(1) The proposed process is advantageous since (i) it utilizes

renewable energy, thus promoting sustainability, and (ii)

catalyst immobilization enables its easy recovery at the

end of the treatment.

(2) The concurrent use of two or more processes to improve

performance is conceptually attractive if the overall effect

is synergistic rather than additive. This seems to be the

case since the application of relatively low currents (i.e. up

to 0.04e0.08 mA/cm2) promotes dramatically the rate of

the respective photocatalytic process with the level of

enhancement reaching values up to 97%.

(3) Promotion of photocatalysis is thought to be associated

with the efficient separation of photogenerated holes from

electrons, thus maximizing their utilization as primary

oxidants or source to generate secondary oxidants like

hydroxyl (and other) radicals.

(4) Performance can be affected by parameters like applied

current, catalyst loading, solution pH and the composition

of the water matrix; the presence of dissolved species may

partly impede the removal of BPA from secondary treated

effluents.

Acknowledgments

The authors wish to thank Dr D. Fatta-Kassinos (University of

Cyprus) and Dr E. Routledge (Brunel University, UK) for

supplying yeasts for the YES test.

r e f e r e n c e s

Bautista-Toledo, I., Ferro-Garcia, M.A., Moreno-Castilla, C., VegasFernandez, F.J., 2005. Bisphenol A removal from water byactivated carbon. Effects of carbon characteristics andsolution chemistry. Environmental Science & Technology 39(16), 6246e6250.

Bouras, P., Lianos, P., 2005. Photodegradation of dyes in aqueoussolutions catalyzed by highly efficient nanocrystalline titaniafilms. Journal of Applied Electrochemistry 35 (7e8), 831e836.

Brugnera, M.F., Rajeshwar, K., Cardoso, J.C., Zanoni, M.V.B., 2010.Bisphenol A removal from wastewater using self-organizedTIO2 nanotubular array electrodes. Chemosphere 78 (5),569e575.

Chatzisymeon, E., Fierro, S., Karafyllis, I., Mantzavinos, D.,Kalogerakis, N., Katsaounis, A., 2010. Anodic oxidation ofphenol on Ti/IrO2 electrode: experimental studies. CatalysisToday 151 (1e2), 185e189.

Cui, Y.-H., Li, X.-Y., Chen, G., 2009. Electrochemical degradation ofbisphenol A on different anodes. Water Research 43 (7),1968e1976.

Cullity, B.D., 1978. Elements of X-Ray Diffraction. Addison-Wesley, Reading, MA.

Finklea, H.O., 1988. Semiconductor Electrodes. Elsevier, NewYork, pp. 71.

Garoma, T., Matsumoto, S.A., Wu, Y., Klinger, R., 2010. Removal ofbisphenolAand its reaction-intermediates fromaqueoussolutionby ozonation. Ozone Science & Engineering 32 (5), 338e343.

Guo, C., Ge, M., Liu, L., Gao, G., Feng, Y., Yang, Y., 2010. Directedsynthesis of mesoporous TiO2 microspheres: catalysts andtheir photocatalysis for bisphenol A degradation.Environmental Science & Technology 44 (1), 419e425.

Han, L., Xin, Y., Liu, H., Ma, X., Tang, G., 2010.Photoelectrocatalytic properties of nitrogen doped Ti/TiO2

photoelectrode prepared by plasma based ion implantationunder visible light. Journal of Hazardous Materials 175 (1e3),524e531.

Hapeshi, E., Achilleos, A., Vasquez, M.I., Michael, C.,Xekoukoulotakis, N.P., Mantzavinos, D., Kassinos, D., 2010.Drugs degrading photocatalytically: kinetics and mechanismsof ofloxacin and atenolol removal on titania suspensions.Water Research 44 (6), 1737e1746.

Huang, Y.-F., Huang, Y.-H., 2009. Identification of producedpowerful radicals involved in the mineralization of bisphenolA using a novel UV-Na2S2O8/H2O2-Fe(II, III) two-stageoxidation process. Journal of Hazardous Materials 162 (2e3),1211e1216.

Jain, R., Shrivastava, M., 2008. Photocatalytic removal ofhazardous dye cyanosine from industrial waste usingtitanium dioxide. Journal of Hazardous Materials 152 (1),216e220.

Jenkins, S., Raghuraman, N., Eltoum, I., Carpenter, M., Russo, J.,Lamartiniere, C.A., 2009. Oral exposure to bisphenol Aincreases dimethylbenzanthracene-induced mammarycancer in rats. Environmental Health Perspectives 117 (6),910e915.

Kang, J.H., Aasi, D., Katayama, Y., 2007. Bisphenol A in the aquaticenvironment and its endocrine-disruptive effects on aquaticorganisms. Critical Reviews in Toxicology 37 (7), 607e625.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 2 9 9 6e3 0 0 43004

Kuo, C.-Y., Wu, C.-H., Lin, H.-Y., 2010. Photocatalytic degradationof bisphenol A in a visible light/TiO2 system. Desalination 256(1e3), 37e42.

Ku, Y., Chiu, P.-C., Chou, Y.-C., 2010. Decomposition of aniline inaqueous solution by UV/TiO2 process with applying biaspotential. Journal of Hazardous Materials 183 (1e3), 16e21.

Li, X.Z., He, C., Graham, N., Xiong, Y., 2005. Photoelectrocatalyticdegradation of bisphenol A in aqueous solution usinga AueTiO2/ITO film. Journal of Applied Electrochemistry 35(7e8), 741e750.

Malato, S., Fernandez-Ibanez, P., Maldonado, M.I., Blanco, J.,Gernjak, W., 2009. Decontamination and disinfection of waterby solar photocatalysis: recent overview and trends. CatalysisToday 147 (1), 1e59.

Memming, R., 2001. Semiconductor Electrochemistry. Wiley-VCH,Weinheim, New York, pp. 60.

Morrison, S.R., 1980. Electrochemistry at Semiconductors andOxidized Metal Electrodes. Plenum Press, New York, pp. 126.

Nakashima, T., Ohko, Y., Tryk, D.A., Fujishima, A., 2002.Decomposition of endocrine-disrupting chemicals in water byuse of TiO2 photocatalysts immobilized onpolytetrafluoroethylene mesh sheets. Journal ofPhotochemistry & Photobiology A: Chemistry 151 (1e3),207e212.

Park, S.Y., Choi, J., 2009. Genotoxic effects of nonylphenol andbisphenol A exposure in aquatic biomonitoring species:freshwater crustacean, daphnia magna, and aquatic midge,chironomus riparius. Bulletin of EnvironmentalContamination & Toxicology 83 (4), 463e468.

Petrier, C., Torres-Palma, R., Combet, E., Sarantakos, G., Baup, S.,Pulgarin, C., 2010. Enhanced sonochemical degradation ofbisphenol-A by bicarbonate ions. Ultrasonics Sonochemistry17 (1), 111e115.

Philippidis, N., Nikolakaki, E., Sotiropoulos, S., Poulios, I., 2010.Photoelectrocatalytic inactivation of E. coli XL-1 blue coloniesin water. Journal of Chemical Technology & Biotechnology 85(8), 1054e1060.

Poerschmann, J., Trommler, U., Gorecki, T., 2010. Aromaticintermediate formation during oxidative degradation ofbisphenol A by homogeneous sub-stoichiometric Fentonreaction. Chemosphere 79 (10), 975e986.

Rodriguez, E.M., Fernandez, G., Klamerth, N., Maldonado, M.I.,Alvarez, P.M., Malato, S., 2010. Efficiency of different solaradvanced oxidation processes on the oxidation of bisphenol Ain water. Applied Catalysis B: Environmental 95 (3e4),228e237.

Routledge, E.J., Sumpter, J.P., 1996. Estrogenic activity ofsurfactants and some of their degradation products assessedusing a recombinant yeast screen. Environmental Toxicology& Chemistry 15 (3), 241e248.

Salian, S., Doshi, T., Vanage, G., 2009. Perinatal exposure of rats tobisphenol A affects the fertility of male offspring. Life Sciences85 (21e22), 742e752.

Subagio, D.P., Srinivasan, M., Lim, M., Lim, T.-T., 2010.Photocatalytic degradation of bisphenol-A by nitrogen-dopedTiO2 hollow sphere in a vis-LED photoreactor. AppliedCatalysis B: Environmental 95 (3e4), 414e422.

Torres, R.A., Sarantakos, G., Combet, E., Petrier, C., Pulgarin, C.,2008. Sequential helio-photo-Fenton and sonication processesfor the treatment of bisphenol A. Journal of Photochemistry &Photobiology A: Chemistry 199 (2e3), 197e203.

Torres-Palma, R.A., Nieto, J.I., Combet, E., Petrier, C., Pulgarin, C.,2010. An innovative ultrasound, Fe2þ and TiO2 photoassistedprocess for bisphenol A mineralization. Water Research 44 (7),2245e2252.

Tsai, W.-T., Lee, M.-K., Su, T.-Y., Chang, Y.-M., 2009.Photodegradation of bisphenol-A in a batch TiO2 suspensionreactor. Journal of Hazardous Materials 168 (1), 269e275.

Wang, R., Ren, D., Xia, S., Zhang, Y., Zhao, J., 2009. Photocatalyticdegradation of bisphenol A (BPA) using immobilized TiO2 andUV illumination in a horizontal circulating bed photocatalyticreactor (HCBPR). Journal of Hazardous Materials 169 (1e3),926e932.

Wang, X., Lim, T.-T., 2010. Solvothermal synthesis of CeNcodoped TiO2 and photocatalytic evaluation for bisphenol Adegradation using a visible-light irradiated LED photoreactor.Applied Catalysis B: Environmental 100 (1e2), 355e364.

Xie, Y.-B., Li, X.-Z., 2006. Degradation of bisphenol A in aqueoussolution by H2O2-assisted photoelectrocatalytic oxidation.Journal of Hazardous Materials 138 (1), 526e533.

Yap, P.-S., Lim, T.-T., Lim, M., Srinivasan, M., 2010. Synthesis andcharacterization of nitrogen-doped TiO2/AC composite for theadsorption-photocatalytic degradation of aqueous bisphenol-A using solar light. Catalysis Today 151 (1e2), 8e13.

Yu, H., Chen, S., Quan, X., Zhao, H., Zhang, Y., 2009. Siliconnanowire/TiO2 heterojunction arrays for effectivephotoelectrocatalysis under simulated solar light irradiation.Applied Catalysis B: Environmental 90 (1e2), 242e248.

Zhan, M., Yang, X., Xian, Q., Kong, L., 2006. Photosensitizeddegradation of bisphenol A involving reactive oxygen speciesin the presence of humic substances. Chemosphere 63 (3),378e386.