Photocatalytic Inactivation of Escherichia coli: Effect of Concentration of TiO2 and Microorganism, Nature, and Intensity of UV Irradiation

Abstract

The efficiency of photocatalytic disinfection, used to inactivate Escherischia coli K12 under different physico-chemical parameters, was examined. The photocatalyst chosen was the semiconductor TiO2 degussa P25 and the irradiation was produced by an HPK 125 lamp. The effect of titania concentration was investigated using two E. coli concentrations. The photocatalyst concentration ranged from 0.1 to 2.5 g/L. The evolution of E. coli inactivation as function of time was discussed depending on the E. coli and TiO2 concentrations. The optimal concentration of the photocatalyst, 0.25 g/L, is lower than that necessary to absorb all photons and to degrade the organic compounds. Some hypotheses are presented to explain this behaviour. The effect of the different domains of UV light (UVA, UVB, and UVC) was also studied and modification of the light irradiation intensity is discussed. No bacteria photolysis was obtained with UVA but the use UVC had, on the contrary, a detrimental effect on bacteria survival. The addition of titania at a low concentration, 0.25 g/L, improved the inactivation of E. coli in the presence of UVA and UVB, but a detrimental effect was observed under UVC. The disinfection efficiency increases as a function of light intensity, whatever the photocatalytic conditions (different TiO2 concentrations and different UV domains). No bacterial growth was observed after disinfection, whether the system contained titania or not. (C) 2007 Published by Elsevier B.V.

Full text

Photocatalytic inactivation of Escherischia coli
Effect of concentration of TiO
2
and microorganism, nature,
and intensity of UV irradiation
A.K. Benabbou
a,b
, Z. Derriche
a
, C. Felix
d
, P. Lejeune
c
, C. Guillard
b,
*
a
Laboratoire physico-chimie des mate
´riaux, catalyse et environnement, de
´partement de Chimie, faculte
´des sciences,
universite
´des sciences et de la technologie d’Oran – B.P 1505 EL M’naouar 31000, Oran, Algeria
b
Institut de recherche sur la catalyse et l’environnement de Lyon (IRCELYON), universite
´Lyon 1-UMR 5256 du CNRS,
2, avenue Albert-Einstein, 69629 Villeurbanne cedex, France
c
Microbiologie, adaptation et pathoge
´nie, UMR 5240 du CNRS-UCB-INSA-BCS, domaine scientifique de la Doua,
10, rue Raphae
¨l-Dubois, ba
ˆtiment Lwoff, 69622 Villeurbanne cedex, France
d
Laboratoire de chimie et biochimie mole
´culaire et supramole
´culaire, universite
´Lyon1-UMR 5246, 43,
boulevard du 11-Novembre-1918, 69622 Villeurbanne cedex, France
Received 26 February 2007; received in revised form 22 May 2007; accepted 23 May 2007
Abstract
The efficiency of photocatalytic disinfection, used to inactivate Escherischia coli K12 under different physico-chemical parameters, was
examined. The photocatalyst chosen was the semiconductor TiO
2
degussa P25 and the irradiation was produced by an HPK 125 lamp. The effect of
titania concentration was investigated using two E. coli concentrations. The photocatalyst concentration ranged from 0.1 to 2.5 g/L. The evolution
of E. coli inactivation as function of time was discussed depending on the E. coli and TiO
2
concentrations. The optimal concentration of the
photocatalyst, 0.25 g/L, is lower than that necessary to absorb all photons and to degrade the organic compounds. Some hypotheses are presented to
explain this behaviour. The effect of the different domains of UV light (UVA, UVB, and UVC) was also studied and modification of the light
irradiation intensity is discussed. No bacteria photolysis was obtained with UVA but the use UVC had, on the contrary, a detrimental effect on
bacteria survival. The addition of titania at a low concentration, 0.25 g/L, improved the inactivation of E. coli in the presence of UVA and UVB, but
a detrimental effect was observed under UVC. The disinfection efficiency increases as a function of light intensity, whatever the photocatalytic
conditions (different TiO
2
concentrations and different UV domains). No bacterial growth was observed after disinfection, whether the system
contained titania or not.
#2007 Published by Elsevier B.V.
Keywords: Photocatalysis; Microorganism; E. coli; Inactivation; Disinfection
1. Introduction
Over the last few years, the provision of clean drinking water
has become a serious problem and an efficient solution must be
found. The present situation involves dealing not only with the
chemical risks but also with a very real biological one. In
addition to their high costs, the conventional water disinfection
technologies, such as chlorination and ozonation, can lead to
the formation of harmful disinfection by-products (DBPs),
among the most dangerous of which are the trihalomethanes
(THMs), well-known for their high carcinogenic potential [1].
One of the most promising new technologies is photocatalysis,
considered as being one of the most important advanced
oxidation technologies (AOT) for water and air purification
[2–4]. It is based on the interaction between light and
semiconductor particles, which produce the highly reactive
oxygen species (ROS), such as OH

,O
2
HO
2
,
capable of destroying chemical and biological water
contaminants. Many catalysts exist, among the more useful
being ZnO, ZrO
2
,CeO
2
,Fe
2
O
3
, and WO
3
[2,5–7]. However, the
most widely used is TiO
2
, because of its very important
www.elsevier.com/locate/apcatb
Applied Catalysis B: Environmental 76 (2007) 257–263
* Corresponding author. Tel.: +33 4 72 44 62 15; fax: +33 4 72 44 84 38.
E-mail addresses: bakhalile@hotmail.fr (A.K. Benabbou),
Chantal.guillard@ircelyon.univ-lyon1.fr (C. Guillard).
+ Models
APCATB-9998; No of Pages 7
0926-3373/$ – see front matter #2007 Published by Elsevier B.V.
doi:10.1016/j.apcatb.2007.05.026

photoactivity, its lack of toxicity, and its high stability [6,8].
TiO
2
mediated photocatalysis has been widely investigated for
a large number of organic contaminants. Their degradation
pathway and their mineralization have been reported [9–12].
Nowadays, interest is directed to the use of this technique in
water disinfection. Matsunaga et al. [13], reported, for the first
time, that microbial cells could be killed by contact with a
TiO
2
-Pt catalyst under illumination with near UV light for 60 to
120 min. This finding created a new route for sterilization and
led to many photocatalytic disinfection studies using TiO
2
.
Viruses, bacteria, fungi, algae, and cancer cells are included in
the wide spectrum of organisms that were inactivated using this
technique [14]. The photokilling mechanism remains, however,
under discussion. Matsunaga et al. [13,15] proposed that the
direct photochemical oxidation of intracellular coenzyme A to
its dimeric form was the main cause of decreases in respiratory
activity that led to cell death. Other authors [16] have suggested
that the microorganism membrane and the microorganism cell
wall undergo disruption in the presence of irradiated TiO
2
,as
shown by a leakage of intracellular K
+
. More recently, Maness
et al found that TiO
2
photocatalysis promotes the peroxidation
of the E. coli membrane phospholipids and induces major
disorders in the cell membrane [17]. Since 2000, Rincon and
Pulgarin studied thoroughly the effect of various parameters,
such as natures of the support and of the photocatalyst and
bacterial initial concentration on the photocatalytic efficiency
[18–20]. More recently, Guillard et al. [21] compared the
photocatalytic disinfection and the photocatalytic organic
compounds removal. A discussion about the importance of
TiO
2
structure and the bacteria/photocatalyst interaction was
presented. The efficiency of the photocatalytic process against
the avian influenza virus H5N2 was demonstrated.
Microbial contamination does not occur only in water or air
media, the problem also exists for metal and polymeric
surfaces. Many microbes are able to interfere with solid
materials and trigger physiological responses that include
colonization of the material surface and the expression of
specific properties, such as increased resistance to antimicro-
bial agents. This initial contamination can then evolve into a
complex consortium: a highly organized layer of a matrix-
embedded microbial population called a biofilm [22,23]. Many
nosocomial infections are the consequences of surface
contamination of indwelling medical devices, or of hospital
equipment, such as air conditioning and water distribution
networks. The proliferation and resistance to cleaning
procedures of pathogenic germs on the surface of production
equipment used in food-processing industries are also causes of
food-borne infections. In this study, we used an E. coli strain
over-synthesizing curli, a particular type of appendage
rendering the bacteria capable of sticking to surfaces and
forming biofilms [24].
Mechanisms for the primary events occurring at the catalyst
surface have been described [25–27]. The first step in
photocatalysis reactions consists of the generation of the
hole–electron pair through the irradiation of the TiO
2
particles
with photonic energy equal to, or greater than, its band gap
energy (3.2 eV). The electron is then extracted from the valence
band (VB) to the conduction band (CB). This process results in
a positive region in the VB (Hole h
+
) and a free electron (e

)in
the CB. (Eq. (1)):
TiO2þhy!TiO2þeðCBÞþhþðVBÞ:(1)
The hole, at the catalyst surface, reacts with hydroxyl ions
(OH

) and adsorbs water to form free radicals (OH

). (Eqs. (2)
and (3)):
TiO2ðhþÞþOH!TiO2þOH;(2)
TiO2ðhþÞþH2Oads !TiO2þOHþHþ;(3)
The CB electron reduces oxygen to the superoxide ion: O
2
(Eq. (4)). This reaction prevents the e

/h
+
recombination, in the
absence of other electron acceptors.
O2þe!O2:(4)
The further reduction of O
2
produces H
2
O
2
, as described
in (Eq. (5))[28]:
O2 þeþ2Hþ!H2O2:(5)
The super oxide ion and its protonated form subsequently
dismute to yield hydrogen peroxide or a peroxide anion [29]
(Eqs. (6)–(8)):
O2 þHþ!HO2;(6)
O2 þ4HO2!2OHþ3O2þH2O2;(7)
2HO
2!O2þH2O2;(8)
It has also been shown that the addition of hydrogen
peroxide can increase the photodegradation rate under certain
conditions [30], probably by the formation of OH

radicals via
the Harber–Weiss reaction (Eq. (9)) or through the reduction of
H
2
O
2
by the CB e

(Eq. (10))[29–31]:
O2 þH2O2!OHþOHþO2;(9)
H2O2þe!OHþOH:(10)
On the other hand, recombination of OH

radicals can lead
again to the production of hydrogen peroxide (Eq. (11)):
OHþOH!H2O2:(11)
Among all the ROS that are generated, according to the
reactions described above, the OH

radical would be the most
important oxidant species that is responsible for bacteria
inactivation [29].
2. Materials and methods
2.1. Materials
All solutions were prepared in deionised water. Glassware
used in these experiments was washed with distilled water and
then autoclaved at 185 8C for 6 h. TiO
2
Degussa P 25 was used
as the photocatalyst. Its density is equal to 3.8 g/cm
3
, the
A.K. Benabbou et al. / Applied Catalysis B: Environmental 76 (2007) 257–263258
+ Models
APCATB-9998; No of Pages 7

average diameter size is 20 to 30 nm and its surface area is
50 m
2
/g.
2.2. Experimental procedures
The disinfection experiments were carried out in a 90 ml
pyrex reactor, with 20 ml of solution. The experimental
apparatus is shown in Fig. 1. To ensure complete mixing,
the TiO
2
/E. coli slurry was magnetically stirred. Three sets of
disinfection experiments were carried out, varying the
parameters of (1) titania weight, (2) light intensity, and (3)
UV light domains. The disinfection experiments were carried at
room temperature, using air as oxidant. The TiO
2
concentration
ranged from 0.1 to 2.5 g/L. The UV light source was an HPK
125 W lamp, which emits in the 200–400 nm range. Different
optical filters were used to modify the lamp emission spectrum.
A 0.52 filter was used to cut off any radiation below the
wavelength of 340 nm (UVA). The pyrex filter was used to
obtain radiations with a wavelength greater than 290 nm (UVB)
and the quartz filter was used to provide irradiations in the range
of 200–400 nm (UVC). The light intensity was modified using
grids, with different sizes of mesh, which were put on the lamp.
Light intensity was measured with a radiometer VLX-3W,
equipped with 365, 312, and 254 nm captors. To measure the
intensity, captors were put at the same distance from the light
source, as was the solution in the reactor for irradiations. With
UVA and UVB radiations, the reactor containing the bacterial
suspension to be disinfected was maintained at a distance of
3 cm from the water circulation cell. In the case of UVC
radiations, the light source was placed above the reactor to
avoid a radiation cut off at 290 nm by the reactor itself, which
was made of pyrex glass. In these conditions, the distance
between the suspension and the light source was 9 cm. Values
of lamp intensities in the different UV domains, and with
different types of grids, are listed in Tables 1 and 2, respectively.
The disinfection durability was tested at the point of total
microorganism inactivation. The inactivated cells were kept in
the dark, with stirring, and samples were taken over a period of
48 h. E. coli samples, with initial concentrations of (low) 10
5
to
10
6
and (high) 10
8
to 10
9
cfu/mL, were chosen as the
microorganism model. The disinfection experiments were
repeated three times to check the reproducibility of the results.
2.3. Analysis and culture of bacteria
The bacterial strain used in this work is PHL1273, a
derivative of the E. coli K-12 strain MG1655. It was constructed
by G. Jubelin (unpublished) by transformation of the strain
PHL818 [32] using the plasmid pPHL127. PHL 818 carries the
chromosomal mutation omp R234 which results in the
overproduction of curli, a particular type of fimbriae allowing
the bacteria to adhere to abiotic surfaces [33]. pPHL127 is a
derivative of the pPROBE-gfp[LVA] (stratagene) containing the
csgBA promoter in front of the gfp[LVA] reporter gene. E. coli
was inoculated into Luria Bertani medium [34], mixed with
water 1/1, and grown overnight at 30 8C, with constant
agitation, under aerobic conditions. Bacterial growth was
monitored by measuring the optical density at 600 nm. At a
stationary growth phase, bacteria were harvested by centrifu-
gation in a 1.5 mL Eppendorf tube at 10
3
rounds per min for
3 min, and washed twice with 1.5 mL of water. An E. coli stock
suspension was prepared by resuspending the final pellets in
1.5 mL of water. The initial populations of E. coli ranged,
approximately, from 10
5
to 10
6
and from 10
8
to 10
9
cfu/mL
(colony forming unit) and they were obtained by diluting the
stock suspension. The cell concentrations were determined by
the spread plate method, with nutrient agar grown at 37 8C for
18 h. 0.1 mL of suspension was withdrawn at each sampling
and was diluted to 1/10, 1/100, and 1/1000. Some of the
samples were spread without diluting. Finally, 100 mL of the
diluted and undiluted suspensions were spread to count the
number of E. coli. Three replicate plates were used at each
dilution. Measurements of bacteria dimensions were carried out
using fluorescence microscopy. The microorganisms were
spread on a glass plate that had been previously coated with the
nutrient agar.
Fig. 1. Experimental apparatus.
Table 1
Values of lamp intensities in the three UV domains
UVA (l>340 nm)
Height (cm) 1 2 3
Intensity (mw/cm
2
) 7 5.2 3.85
UVB (l>290 nm)
Height (cm) 1 2 3
Intensity (mw/cm
2
) 13.3 10.9 8.5
UVC (l>200 nm)
Height (cm) 9
Intensity (mw/cm
2
) 13.2
Table 2
Values of lamp intensities with the grids
UVA
Grids 1 2 3
Intensity (mw/cm
2
) 1.95 1 0.5
UVB
Grids 1 2 3
Intensity (mw/cm
2
) 4.15 2 1
UVC
Grids 1 2 3
Intensity (mw/cm
2
) 6.9 3.6 1.8
A.K. Benabbou et al. / Applied Catalysis B: Environmental 76 (2007) 257–263 259
+ Models
APCATB-9998; No of Pages 7

3. Results and discussion
First, the bacterial suspension was irradiated under UVA,
with a light intensity of 3.85 mW/cm
2
and in the absence of
TiO
2
, for 5 h. No E. coli inactivation was observed.
3.1. Effect of TiO
2
concentration
Experiments to investigate the TiO
2
concentration effect on
photocatalytic disinfection were performed using two different
concentrations of E. coli, a concentration of 10
7
–10
8
cfu/mL
and another about one hundred times lower, at around 10
5
–
10
6
cfu/mL. The same experiments were carried out three times
to check their reproductibility.
3.1.1. Effect of TiO
2
concentration using an E. coli
concentration of 10
7
–10
8
cfu/mL
Fig. 2a shows that, at an E. coli concentration which ranged
from 10
7
to 10
8
cfu/mL, there was total inactivation of E. coli
after 3 h of reaction using 0.25 g/L of TiO
2
, whereas the
concentration remained near 10
4
cfu/mL after the same
exposure to irradiation with a TiO
2
concentration ten times
higher, that is, 2.5 g/L. Depending on the TiO
2
concentration
used, the E. coli inactivation process consisted of three or four
steps:
during the first 5 to 10 min of irradiation, inactivation was
very slow. This is an induction period, where the active
species generated begins to attack the membrane but not
sufficiently to cause serious damage to the bacterial outer
membrane leading to the loss of its permeability. The
bacterial membrane is gradually oxidised, though the
microorganism is resisting this attack with the usual self-
defence and auto-repair mechanisms. The latter involves the
generation of repair enzymes by the bacteria [29]. At the end
of this first period, the repeated ROS attacks on the E. coli
membrane may result in a perforation which accelerates the
bacterial inactivation process. This induction period lasted
for 10 min using 2.5, and 1.5 g/L of TiO
2
and for 5 min using
0.5, 0.25, and 0.1 g/L of titania;
during the period between 5/10 and 60 min, microorganism
inactivation was accelerated. The anti-stress enzymes are no
longer able to protect the bacterial membrane against
oxidation and, hence, membrane perforation can occur.
During this period the inactivation of bacteria could also be
due to the formation of by-products, such as acids or
aldehydes. Moreover, it can be suggested that the bacterial
inactivation is not only due to the active species generated
during the photocatalytic process but also to other
phenomena, such as the photo–Fenton reaction, which could
contribute to the acceleration of bacterial inactivation. In fact,
when the membrane is perforated, ferrous, and ferric ions
could be released and can react to form OH

radical as
presented in equations ((12a) and (12b)).
Fe2þþH2O2!Fe3þþOH
|fflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflffl}
FeðOHÞ2þ
þOH(12a)
FeðOHÞ2þ!
hnþFe2þþOH(12b)
The membrane perforation may open a way to the leakage of
inner bacterial components. Inactivation rates were greater than
99.9% for all titania concentrations. However, it is important to
note that about 10
3
to 10
4
bacteria were not inactivated by using
a too important concentration of TiO
2
.
between 60 and 150 min there is no further improvement in
inactivation efficiency for TiO
2
concentrations of 2.5, 1.5,
and 0.1 g/L while total inactivation occurs using 0.25 and
0.5 g/L of TiO
2
. This second induction period is more
difficult to explain. Nevertheless, several hypotheses can be
suggested; at high photocatalyst concentration, a screen
effect can take place. Actually the bacteria are wrapped by
TiO
2
in opposition to what happened when organic
molecules, such as pesticide, are degraded. In this last case,
it is the organic molecules which are adsorbed on TiO
2
and
this phenomenon is not observed; due to the high initial
concentration of microorganisms, it can be suggested that an
important amount of organic molecules are formed. These
molecules could be in competition with bacteria as regards
the ROS. This hypothesis is in agreement with the irradiation
curve observed when using a lower initial concentration of
Fig. 2. (a) Effect of TiO
2
concentration on the inactivation of 10
7
–10
8
cfu/mL E
.coli suspensions. Insert: irradiation time required to decrease the bacteria
concentration from 10
8
to 10
5
cfu/mL, as a function of titania concentration. (b)
Effect of TiO
2
concentration on the inactivation of 10
5
–10
6
cfu/mL E. coli
suspensions. Insert: time required to decrease the bacteria concentration from
10
6
to 10
2
cfu/mL, as a function of titania concentration.
A.K. Benabbou et al. / Applied Catalysis B: Environmental 76 (2007) 257–263260
+ Models
APCATB-9998; No of Pages 7

microorganisms (Fig. 2); at high titania concentration,
terminal reactions may also be implicated in the rate
decrease of E. coli inactivation (Eq (a) and (b)). As described
in Eq (a), the radical OH

will readily self-recombine to form
H
2
O
2
. This last will produce HO
2
(Eq (b)) through its
interaction with another OH

radical. The hydroperoxyl
radical obtained would be less reactive and would not
contribute to the inactivation process [35].
OHþOH!H2O2(a)
H2O2þOH!H2OþHO2(b)
for the TiO
2
concentrations equal to 0.5 and 0.25 g/L, the
inactivation process continues. The bacteria concentration
decreases to a nondetectable level. However, the necessary
time required to reach a total inactivation of the bacteria
increases as the amount of TiO
2
increases.
This phenomenon could be due to the scattering effect of
titania. The greater the amount of photocatalyst present, the
more it forms several layers on the bacteria membrane. As a
consequence, there would be increasingly limited light
diffusion to the E. coli. Thus, if the hole and the electrons
which are produced by the irradiation are not used in the
process they would quickly recombinate with titania, given the
large quantity of photocatalyst present. The optimum TiO
2
concentration was found to be equal to 0.25 g/L. At a lower
concentration, 0.1 g/L, the efficiency decreased (insert of
Fig. 2a). This amount of TiO
2
is not sufficient to absorb all the
photons.
Contrary results have been obtained by different authors in
earlier studies [36–38] which have indicated higher efficiencies
at higher TiO
2
concentrations. Min et al. [28] found that E. coli
inactivation at the concentration of 1 g/L is approximately
twice as effective as that at 0.1 g/L. However, a detrimental
effect was noticed when the TiO
2
concentration was increased
from 1 to 2 g/L. The work of Rincon and Pulgarin [29] revealed
an analogous result. These variations could be attributed not
only to the difference of E. coli strain tested by the authors, but
also to the various kind of lamps giving the UV lights and the
set-up geometry used in the different works. In our case the
bacteria is more adherent and this would increase the quantity
of microorganisms present at the titania surface.
3.1.2. Effect of TiO
2
concentration using an E. coli
concentration of 10
5
–10
6
cfu/mL
Three TiO
2
concentrations were used, namely 1.5, 0.25, and
0.1 g/L. Curves related to the inactivation of the microorgan-
isms are presented in Fig. 2b. Time for total microorganism
inactivation was shorter at low concentrated suspension (10
5
–
10
6
cfu/mL). An induction period for the TiO
2
concentrations
of 1.5 and 0.1 g/L was detected resulting, respectively, from the
scattering effect and the small quantity of photocatalyst present,
leading to a decrease in ROS production. These observations
are correlated with the self-defence and auto-repair mechan-
isms. Total E. coli inactivation was achieved after 90 min of
irradiation using 0.25 and 0.1 g/L, whereas the time period was
180 min with a concentration of TiO
2
equal to 1.5 g/L.
Whatever the E. coli concentration, the optimal concentra-
tion of TiO
2
used to inactivate E. coli was 0.25 g/L. This value
does not correspond to a total absorption of the lamp-emitted
photons and it does not correspond to the optimum TiO
2
concentration found by Herrmann for organic compounds [2].It
seems that bacteria needs less titania for their inactivation than
organic compounds. It is important to note this significant
difference in the behaviour of titania towards chemicals or
microorganisms. Due to their very small size, organic
compounds are able to diffuse between the titania particles,
and to be adsorbed onto them, whereas the opposite occurs with
microorganisms. In fact, bacteria are more than 30 times bigger
than TiO
2
particles.
3.1.3. Discussion of the optimal quantity of TiO
2
Although 0.25 g/L of TiO
2
degussa P-25 is not enough to
absorb all the photons, according to previous studies [2],
nevertheless, in our experimental conditions, it is the optimal
concentration necessary to inactivate E .coli. Three hypotheses
could explain this phenomenon:
1) bacteria diffuse a part of the UV light, rendering it
unavailable for TiO
2
activation;
2) due to the difference in size between bacteria and TiO
2
particles, it is the semiconductor that wraps around the
bacteria and not the opposite. It is possible that a
concentration of 0.25 g/L is sufficient to totally cover the
E. coli bacteria. Above this TiO
2
concentration, a multilayer
of TiO
2
could be formed. So, we tried to estimate the amount
of TiO
2
required to totally wrap around the bacteria. This
value would be theoretical and approximate because TiO
2
particles are often agglomerated. The bacteria were assumed
to have a cylindrical shape. Its dimensions, determined by
fluorescence microscopy, were about 1.7 mm long and
0.9 mm wide. Thus, the total microorganism surface area
was 6.10
12
m
2
.TiO
2
particles, on the other hand, are
spherical with a diameter of 20 to 30 nm. We considered that
aTiO
2
particle would have a surface area of 4 10
16
m
2
(or 9 10
16
m
2
). Based on the assumption that the
bacterial surface is totally covered by a monolayer of
TiO
2
particles, the ratio between the surface of the
microorganism and that of the particle indicates the number
of titania particles that are around each E. coli. So, the
theoretical quantity of TiO
2
covering one bacteria can be
estimated to 1.44 10
10
gTiO
2
/bacteria (6.34 10
11
g
TiO
2
/bacteria). Experimentally, 2.5 10
12
gTiO
2
/bacteria
were present, considering a 20 mL bacterial suspension at a
concentration of 1O
8
cfu/mL containing 0.25 g/L of TiO
2
,
this indicates that the optimal amount of TiO
2
used was not
sufficient to cover all the bacteria. So, this does not provide
an explanation for the behaviour observed in our experi-
ments;
3) the release of Fe
2+
during the photocatalytic process, lead to
the formation of the complex Fe (OH)
2+
. This last absorbs
A.K. Benabbou et al. / Applied Catalysis B: Environmental 76 (2007) 257–263 261
+ Models
APCATB-9998; No of Pages 7

UV light and gives OH

which favours the photocatalytic
process.
3.2. Effect of the UVA light intensity
The influence of the light intensity was investigated using
the optimal TiO
2
concentration, which was equal to 0.25 g/L.
Increased efficiencies were obtained at higher intensities. Fig. 3
shows a decrease in the E .coli inactivation rate by decreasing
the light intensity from 3.85 to 0.48 mw/cm
2
. An induction
period was observed in the first 10 min for the lower intensities,
which suggests that the self-defence and autorepair mechan-
isms for protecting the bacteria were more efficient at a low
intensity, but still insufficient as regards the accumulation of
irradiation. By decreasing the intensity from 3.85 to 0.48 mw/
cm
2
, the time necessary to totally inactivate E .coli increased
from 90 to 180 min. The insert in Fig. 3 shows an inactivation
time that increased when the light intensity decreased, these
values having been calculated for a reduction in E .coli from 10
6
to 10
3
cfu/mL. A similar result was observed for the intensity
influence on the E .coli inactivation using a TiO
2
concentration
equal to 1.5 g/L.
3.3. Effect of the type of UV light used (UVA, UVB, and
UVC), in the presence or absence of TiO
2
The effect of the different UV lights was studied under
similar intensities. Using the same lamp and the appropriate
grids, similar intensities were obtained. I
UVA
,I
UVB,
and I
UVC
were equal to 3.8, 4, and 3.6 mW/cm
2
respectively. As shown in
Fig. 4, the inactivation of E. coli is more efficient under UVC
irradiation. The concentration of bacteria decreases to a non-
detectable level within only 20 min of photocatalytic treatment.
At an equivalent intensity, the UVB and UVA were less
effective than the UVC. Total inactivation was achieved after
60 min with UVB and 90 min using UVA irradiation. UVB
were more bactericidal than UVA, which could result from the
fact that, in addition to generate ROS at the TiO
2
surface, UVB
are implicated in DNA lesions that cause inhibition of DNA
replication and bacterial mutation [39]. However, it is
interesting to note that the initial inactivation of bacteria
observed in the presence of UVA or UVB was very similar. It is
only after 30 min that UVB became more efficient. This
behaviour can be explained with respect to the self-defence and
autorepair processes. A comparison between photocatalytic and
purely photolytic microorganism inactivation after 10 min of
treatment is given in the insert of Fig. 4. From an initial
concentration of 10
6
cfu/mL, bacterial survival decreased to
about 10
3
and to 4 10
2
cfu/mL, using the UVA/TiO
2
and
UVB/TiO
2
systems respectively, while nonsignificant inactiva-
tion was observed when using only UVA or UVB. On the other
hand, the presence of titania had a detrimental effect on the
efficiency of UVC rays. The UVC photolysis was seen to be
much more efficient than UVC/TiO
2
photocatalysis. A final
bacterial concentration of 37 cfu/mL remained in the presence
of the photocatalyst, while no microorganisms were detected
using pure UVC photolysis. During the phototreatment, the
TiO
2
particules would have protected the bacteria from UVC
radiation through a screening effect. This suggests that
photocatalysis could be a useful process for inactivating
bacteria using solar irradiation.
3.4. Durability of treatment
Disinfection durability experiments were performed to
determine the efficiency of the photodegradation of the
microorganisms. Two recovery processes are known to exist
for bacteria: dark repair and photoreactivation [18,40]. There is
said to be reactivation when bacteria that have lost their
cultivability during irradiation can activate resistance to
stressful conditions and recover their cultivability when they
are introduced into more favourable conditions [41,42].A
bacterial growth test was not performed for the UVA photolysis
experiment since cell inactivation did not occur. No bacterial
growth was detected for all the systems studied. UVC
accelerated the bacterial inactivation, and did not allow any
bacterial recovery in the dark. In the presence of TiO
2
, which
Fig. 3. Intensity effect on E. coli inactivation in the UVA domain at a TiO
2
concentration of 0.25 g/lL. Insert: time needed to decrease the bacteria con-
centration from 10
6
to 10
3
cfu/mL, as a function of irradiation intensity.
Fig. 4. Photocatalytic E. coli inactivation in the presence of UVA, UVB, and
UVC rays, at the same light intensity. The TiO
2
concentration was 0.25 g/L.
Insert: the concentration of E. coli in solution after 10 min of bacterial
irradiation in the presence or absence of TiO
2
using different UV lights.
A.K. Benabbou et al. / Applied Catalysis B: Environmental 76 (2007) 257–263262
+ Models
APCATB-9998; No of Pages 7

led to the same result, no recovery of the microorganisms was
observed. This result indicates that the oxidative species
developed at the titania surface caused severe damage to the
cells. According to the results obtained, it would be possible to
use solar energy that contains UVA to activate the TiO
2
and so
disinfect water.
4. Conclusions
Photocatalytic water disinfection seems to be a promising
alternative for bacterial elimination. However, some parameters
have to be taken into consideration:
1) an optimal TiO
2
concentration value was determined and the
system efficiency would decrease if this parameter was not
applied;
2) bacterial inactivation was related to the radiation intensity,
with higher microorganism inactivation obtained at higher
intensities;
3) bacterial inactivation efficiency depends on the type of UV
light used. At equivalent intensities, the UVC/TiO
2
system
was more lethal for bacteria than the UVA/TiO
2
and UVB/
TiO
2
systems. Moreover, UVC photolysis was more efficient
than photocatalysis. This result indicates that the photo-
catalysis process would be of interest when solar energy can
be used. However, bacterial inactivation is not the only
parameter to be considered. Studies should be developed to
determine whether these photolytic or photocatalytic pro-
cesses generate other contaminants. It is also necessary to
analyse the organic and inorganic intermediates formed
during these two processes to investigate whether TiO
2
can
improve UVC photolysis. The dark event studies, following
bacterial inactivation, demonstrated that no bacterial culti-
vability was recovered by the microorganisms after 48 h.
Hence, the phototreatment could guarantee the durability of
the disinfection. This process may also open up a new route in
the sterilisation of surfaces contaminated by biofilms.
Acknowledgements
We are grateful to the Algerian Ministry of Education and
Research for the Ph.D scholarship and the University of Lyon
for its financial support (BQR: Bonus Qualite
´Recherche)
References
[1] C.E. Lawrance, P.R. Taylor, B.J. Trock, A.A. Reilly, Nat. Cancer Inst. 72
(1984) 563.
[2] J.M. Herrmann, Catal. Today 53 (1999) 115.
[3] D. F. Ollis C. R. Acad Sci. Paris, Se
´rie IIc, Chimie/Chemistry 405 (2000) 3.
[4] P. Pichat, Catal. Today 313 (1994) 19.
[5] O. Seven, B. Dindar, S. Aydemir, D. Metin, M.A. Ozinel, S. Icli, J.
Photochem. Photobiol., A 165 (2004) 103.
[6] A. Malowski, W. Wardas, Curr. Top. Biophys. 25 (1) (2001) 19.
[7] S. Chatterjee, S. Sarkar, S.N. Battacharyya, J. Photochem. Photobiol., A
81 (1994) 199.
[8] K. Djebbar, T. Sehili, Pestic. Sci. 54 (1998) 269.
[9] M.R. Hoffmann, S.T. Martin, W.Y. Choi, D.W. Bahnemann, Chem. Rev.
95 (1995) 69.
[10] C. Guillard, J. Photochem. Photobiol., A 135 (2000) 65.
[11] C. Guillard, S. Horikoshi, N. Watanabe, H. Hidaka, P. Pichat, J. Photo-
chem. Photobiol., A 149 (2002) 155.
[12] M. Carrier, N. Perol, J.M. Herrmann, C. Bordes, S. Horikoshi, O. Paisse,
B. Baudat, C. Guillard, Appl. Catal. B Environmental 65 (2006) 11.
[13] T. Matsunaga, R. Tomada, T. Nakajima, H. Wake, FEMS Microbiol. Lett.
29 (1985) 211.
[14] D.M. Blake, P.-C. Maness, Z. Huang, E.J. Wolfrum, J. Huang, Sep. Purif.
Methods 28 (1) (1999) 1.
[15] T. Matsunaga, R. Tomoda, Y. Nakajima, N. Nakamura, T. Komine, Appl.
Environ. Microbiol. 54 (1988) 1330.
[16] T. Saito, T. Iwase, T. Morioka, J. Photochem. Photobiol., B 14 (1992) 369.
[17] P.-C. Maness, S. Smolinski, D.M. Blake, Z. Huang, E.J. Wolfrum, W.A.
Jacobi, Appl. Environ. Microbiol. 65 (1999) 4094.
[18] A.-G. Rincon, C. Pulgarin, Appl. Catal. B Environmental 49 (2004) 99.
[19] A.G. Rincon, C. Pulgarin, Appl. Catal. B Environmental 51 (2004) 283.
[20] D. Gumy, C. Morais, P. Bowen, C. Pulgarin, S. Giraldo, R. Hadju, J. Kiwi,
Appl. Catal. B Environmental 63 (2006) 76.
[21] C. Guillard, T-H. Bui, C. Felix, V. Moules, B. Lina, P. Lejeune. Comptes
Rendus de l’Acade
´mie des Sciences, Se
´rie IIC Chemistry, a
`paraı
ˆtre.
[22] J.W. Costerton, Z. Lewandowski, D. DeBeer, D. Caldwell, D. Korber, G.
James, J. Bacteriol. 176 (1994) 2137.
[23] P. Lejeune, Trends Microbiol. 11 (2003) 179.
[24] G. Jubelin, A. Vianney, C. Beloin, J.M. Ghigo, J.C. Lazzaroni, P. Lejeune,
C. Dorel, J. Bacteriol. 187 (2005) 2038.
[25] D.F. Ollis, Environ. Sci. Technol. 19 (1985) 480.
[26] N. Serpone, Sol. Energy Mater. Sol. Cells 38 (1995) 369.
[27] P. Pichat, M.A. Fox, in: M.A. Fox, M. Chanon (Eds.), Photoinduced
Electron Transfer, Elsevier, Amsterdam, 1988, , p. 241, Part D.
[28] C. Min, C. Hyenmi, C. Wonyong, Y. Jeyong, Water Res. 38 (2004) 1069.
[29] A. Rincon, C. Pulgarin, Appl. Catal. B Environmental 44 (2003) 263.
[30] P. Pichat, C. Guillard, A. Laurence, A.-C. Renard, O. Plaidy, Sol. Energy
Mater. Sol. Cells 38 (1995) 391.
[31] E.J. Wolfrum, D.F. Ollis, in: G.R. Helz, R.G. Zepp, D.G. Crosby (Eds.),
Aquatic and Surface Photochemistry, Lewis, Boca Raton, 1994, p. 451.
[32] C. Prigent-Combaret, E. Brombacher, O. Vidal, A. Ambert, P. Lejeune, P.
Landini, C. Dorel, J. Bacteriol. 183 (2001) 7213.
[33] O. Vidal, R. Longin, C. Prigent-Combaret, C. Dorel, M. Hooreman, P.
Lejeune, J. Bacteriol. 180 (1998) 2442.
[34] J.H. Miller, Spring Harbour Laboratory, Cold Spring Harbor N.Y. (1972)
352.
[35] O. Legrini, E. Oliveros, M.A. Braun, Chem. Rev. 93 (1993) 671.
[36] C. Wei, W. Lin, Z. Zainal, N.E. Zhu, K. Kruzic, R.L. Smith, K. Rajeshwar,
Environ. Sci. Technol. 28 (1994) 934.
[37] D. Bahnemann, D. Bockelmann, R. Golish, Sol. Energy Mater. 24 (1997)
564.
[38] M. Bekbolet, Water Sci. Technol. 35 (1997) 95.
[39] A.B. Britt, Annu. Rev. Physiol. Plant. Mol. Biol. 47 (1996) 75.
[40] H. Liltved, B. Landfald, Water Res. 30 (1996) 1109.
[41] W. Harm, Cambridge University Press, Cambridge, 1980.
[42] G.J. Leyer, E.A. Johnson, Appl. Environ. Microbiol. 59 (1993) 1842.
A.K. Benabbou et al. / Applied Catalysis B: Environmental 76 (2007) 257–263 263
+ Models
APCATB-9998; No of Pages 7