Solar Energy 80 (2006) 1376–1381

www.elsevier.com/locate/solener

Solar photocatalysis for detoxification and disinfection

of water: Different types of suspended and fixed
TiO2 catalysts study
D. Gumy, A.G. Rincon, R. Hajdu, C. Pulgarin *

Environmental Science and Technology Institute, Laboratory for Environmental Biotechnology (LBE), Department of Rural
Engineering, Swiss Federal Institute of Technology (EPFL), CH-1015 Lausanne, Switzerland

Received 10 November 2004; received in revised form 7 March 2005; accepted 8 April 2005
Available online 30 September 2005

Abstract

Photocatalysis by titanium dioxide (TiO2), operational in the UV-A domain with a potential use of solar radiation,
could be an alternative to conventional water detoxification and disinfection technologies. However, employing the pho-
tocatalyst as a suspension or slurry makes the scaling-up of the process difficult, as the TiO2 has to be removed from the
decontaminated water to be reused several times. In this work the photocatalytic activity of different types of TiO2 catalyst
(Degussa P-25, Millennium PC-100 and PC-500, Tayca AMT-100 and AMT-600) in suspension or coated on fibrous web
were studied in both decontamination and disinfection experiments at laboratory scale. Gallic acid was chosen as the
model pollutant for detoxification experiments and Escherichia coli as the model microorganism for disinfection experi-
ments. The influence of the surface area and other characteristics of TiO2 are discussed concerning the photocatalytic prop-
erties of TiO2. The role of adsorption is suggested, indicating that the reaction occurs at the TiO2 surface and not in the
solution. Gallic acid degradation kinetics were found to be of the same extent for both TiO2 suspended and fixed, whereas
for the bacterial inactivation efficiency was significantly less important with coated than with suspended TiO2.
Ó 2005 Elsevier Ltd. All rights reserved.

Keywords: Solar water disinfection and detoxification; Photocatalysis; TiO2; Escherichia coli; Gallic acid

1. Introduction suspended in water and irradiated with near UV

Photocatalytic oxidation is a promising technol-
ogy for the disinfection and detoxification of water
and wastewater. When catalytic semiconductor
powders, such as the titanium dioxide (TiO2), are
*
Corresponding author. Tel.: +41 21 693 47 20; fax: +41 21
693 47 22.
E-mail address: cesar.pulgarin@epfl.ch (C. Pulgarin).
(k < 385 nm), hydroxyl radicals (OH) and other
oxidative species like H2O2 and superoxide radicals
(O2 ) are generated. The OHÆ radical is highly toxic
towards microorganisms and very reactive in the
oxidation of organic substances. Photocatalytic
inactivation of bacteria such as Escherichia coli,
Bacillus pumilus and several others, as well as sev-
eral Phage have been investigated (Block and Gosw-
ami, 1997; Matsunaga et al., 1985; Wei et al., 1994;

0038-092X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.solener.2005.04.026
 D. Gumy et al. / Solar Energy 80 (2006) 1376–1381
Matsunaga and Okochi, 1999; Rincón et al., 2001;
Rincón and Pulgarin, 2003).
The type of TiO2 used plays an important role
during the photocatalytic abatement of bacteria
and organic pollutants since the rate of formation
of surface species like H2O2 and oxidative radicals
is a function of the titania size, BET area, crystalline
phase, uniformity, etc. (Mills and Le Hunte, 1997;
Blake et al., 1999).
The use of TiO2 in suspension is efficient due to
the large surface area of catalyst available for the
reaction. Nevertheless, the catalyst must be removed
following the treatment. Post-treatment removal re-
quires a solid liquid separation stage which adds to
the overall capital and running costs of the plant.
Alternatively, the catalyst may be immobilized onto
a suitable solid support which would eliminate the
need of post-treatment removal, but which would
also create a decrease in the surface area available
for photocatalytic reaction (Robert et al., 1999).
In the present study, we have tested the photoac-
tivities of several types of suspended TiO2, in both
decontamination and disinfection experiments.
Their structures were varied to follow their influence
upon activity. Gallic acid, one of the most frequent
phenolic compounds in wastewater, was chosen as
the model pollutant for detoxification experiments.
E. coli, a classical bacterial indicator of fecal pollu-
tion, was chosen as the model microorganism for
disinfection experiments. One of the catalysts was
also coated on a nonwoven web (both natural and
synthetic fiber) prepared by Ahlstrom (Research &
Services, ZI de lÕAbbaye, 38780 Pont-Evêque,
France) to evaluate the effect of the fixation on
E. coli and gallic acid abatement. The set of the re-
sults and its potential application for water disinfec-
tion are studied within the frame of the
Fig. 1. (a) Photograph of thin film reactors in solar lamp. (b) Photograph of thin film reactor with nonwoven web fixed on an intern PVC
support.
1377
SOLWATER (EU-DGXII, contract: ICA4-CT-
2002-10001) and AQUACAT (EU-DGXII, con-
tract: ICA3-CT2002-10016) INCO projects.
2. Experimental section
2.1. Photoreactors and light sources
All the experiments were carried out under artifi-
cial light using thin film Pyrex glass reactors with an
illuminated volume of 25 ml (Fig. 1(a)). A peristaltic
pump recirculated the water from a glass bottle, act-
ing like a recirculation tank, to the reactors with a
flow rate of 150 ml/min. The total volume of the
system (100 ml) is considered in two parts: the
25 ml irradiated volume and the dead volume (recir-
culation tank + connecting tubing). Internal PVC
supports were put in the reactors to carry the non-
woven web with coated TiO2 (Fig. 1(b)). The reactor
was illuminated from the outside with a solar lamp
(Hanau Suntest AM1). This lamp has a light spec-
tral distribution of about 0.5% of the emitted pho-
tons at wavelengths shorter than 300 nm and
about 7% between 300 and 400 nm. The emission
spectrum between 400 and 800 nm follows the solar
spectrum. Four reactors could be placed in parallel
inside the solar lamp (Fig. 1(a)).
2.2. Catalysts
The experiments with suspended TiO2 were per-
formed with Degussa P-25, two TiO2 catalysts pre-
pared by Millennium Inorganic Chemicals
(Millennium PC-100 and PC-500) and two catalysts
prepared by Tayca Corporation (Tayca AMT-100
and AMT-600) at concentrations of 0.2, 0.6 or
1 gl 1. Their characteristics are given in Table 1.
1378 D. Gumy et al. / Solar Energy 80 (2006) 1376–1381

Table 1
Characteristics of suspended TiO2 samples
Catalyst Supplier Surface area (m2/g) Crystalline form Crystallite size (nm)
PC 100 Millennium 90 Anatase (100%) 15 ± 5
PC 500 Millennium 335 Anatase (100%) 8±3
AMT 600 Tayca 56 Anatase (100%) 30
AMT 100 Tayca 290 Anatase (100%) 6
P-25 Degussa 50 Anatase (80%), rutile (20%) Anatase 24 ± 2, rutile 37 ± 3

P-25 was coated by Ahlstrom at different area load-
ing on the nonwoven web (natural and synthetic fi-
bers, 0.2 mm thick) using an inorganic binder. The
binder was an aqueous dispersion of colloidal
SiO2. The concentrations coated were 5.5, 10.0
and 13 g/m2 and the surface of paper was 60 cm2.
2.3. Chemicals and analysis
Gallic acid and all other organic compounds
were purchased by Fluka (Buchs SG, Switzerland).
TiO2 in suspension was separated by centrifugation
and filtration before analysis of gallic acid. High
performance liquid chromatography (HPLC) was
carried out with a chromatograph Varian 9065
Unit, mounted with a diode array (Varian, Switzer-
land). A spheriosorb silica column (ODS-2) with
acetic acid (10% v/v) acetonitrile gradient evolution
was used.
2.4. Bacterial strain and quantification
The bacterial strain used, E. coli K12 was inocu-
lated in to nutrient broth no. 2 (N. Oxoid 2, Switzer-
land) and grown overnight at 37 °C by constant
agitation under aerobic conditions. Aliquots of the
overnight culture were inoculated into fresh medium
and incubated aerobically at 37 °C during 15 h. At a
stationary growth phase, bacteria cells were col-
bacteria were added to the glass bottle at different
concentrations and mixed by recirculating the water
to constant concentration throughout the system.
All experiments were carried out with Milli-Q water.
Samples were first collected at predetermined times
(t) in the dark. Then the solar lamp was turned on
at an intensity of 500 or 1000 W/m2 and samples
were collected at predetermined times during 2 or
3 h. Each experiment was repeated three times.
3. Results and discussion
3.1. Degradation of gallic acid with different
commercial suspended catalysts
There are various titanium dioxide preparations
commercially available and conflicting reports as
to which preparations are the most efficient as phot-
ocatalysts. Among various characteristics such as
particle size, crystalline form or active site density,
which may intervene on the photocatalytic activity,
the influence of the surface area appeared of prime
importance. Fig. 2 shows the evolution of the con-
20
15
Gallic acid (mg/L)

lected by centrifugation at 500g for 10 min at 4 °C

and the bacterial pellet was washed three times with 10
a tryptone solution. Cell suspensions were diluted
with Milli-Q water in the glass bottle describe above
to the required cell density corresponding to 105–106 5
darkness
adsorption
colony forming units per millilitre (CFU/ml). Sam-
ples were spotted onto plates and spread using stan- 0
-100 -80 -60 -40 -20 0 20 40 60 80 100 120
dard techniques. Plates were incubated at 37 °C for
Time (min)
24 h prior to enumeration.
Fig. 2. Adsorption and photodegradation (1000 W/m2) of gallic
acid (20 mg l 1) with different TiO2 in suspension (0.2 g/l): PC100
2.5. Photocatalysis experiments
(h), PC500 (j), AMT100 (m), AMT600 (n), P25 () and
without TiO2 addition (s) as a function of time. Each point
At the beginning of the experiments while reac- represents an average value of three experiments. The bars show
tors were kept in the dark, all the chemicals or the the standard deviation.

centration of gallic acid as a function of time during
experiments with the five different catalysts tested.
Adsorption (dark) experiments were carried out
for all the catalysts. It appears that the adsorption
of gallic acid is strongly correlated with the specific
area of the catalyst (Fig. 2, darkness adsorption).
TiO2 with larger surface area have a higher adsorp-
tion capacity.
As we have measured only gallic acid still present
in solution (HPLC measurements), it is difficult to
dissociate the influence of the TiO2 characteristics
on both adsorption and photodegradation pro-
cesses. Nevertheless, as already observed by other
researchers (Guillard et al., 2003a), the initial rate
of disappearance, for the four catalysts constituted
only of anatase crystalline phase (Tayca and Millen-
nium catalysts), seems to be correlated with the
amount of adsorbed gallic acid (and therefore with
the catalysts surface area) (Fig. 3). This result sug-
gests that the photoreaction occurs, in our case,
mainly in the adsorbed phase at the surface of
TiO2 and not in the solution as it has been some-
times claimed for other compounds (Parra et al.,
2002). In spite of the fact that P-25 has the smallest
surface area and a smallest quantity of adsorbed
gallic acid (Figs. 2 and 3), a higher initial rate of dis-
appearance is observed for P-25 compared with PC-
100 and AMT-600. Generally, anatase is considered
to be the photoactive form, while rutile is consid-
ered to have low photocatalytic activity. However,
for reasons that are not yet understood, mixtures
of anatase and rutile (e.g., P-25, which consists of
about 80% of anatase and 20% of rutile) have a bet-
ter photo-activity than either phase by itself (Rincón
and Pulgarin, 2003).
1.4
1.2
P25
r0 (µmol L-1 min-1)
1.0
0.8
0.6
0.4
AMT600
0.2 PC100
0.0
0 1 2 3
qads (µmol)
Fig. 3. Initial rate of gallic acid disappearance with different TiO2
in suspension (0.2 g/l) as a function of the TiO2 amount
adsorbed.
D. Gumy et al. / Solar Energy 80 (2006) 1376–1381 1379
3.2. Influence of the amount of fixed and
suspended TiO2 on gallic acid photocatalytic
degradation
Supported photocatalyst is rather advantageous
for technical systems especially under continuous
flow, since usual recovery steps such as filtration
and decantation can be avoided. In order to investi-
gate the effect of the catalyst fixation on the gallic
acid degradation rate, experiments were performed
with different amount of P-25 TiO2 suspended or
coated on the fibrous web (Fig. 4). The different
concentration tested, the respective amount of cata-
lyst in the system and the corresponding pseudo-
first order constant are summarized in Table 2.
Fig. 4(a) and (b) show that in both cases the
adsorption is linearly correlated with the amount
of TiO2. As shown in the inserts of Fig. 4 and in Ta-
ble 2, the kinetics of disappearance with suspended
or fixed TiO2 are almost similar and linearly corre-
lated with the amount of TiO2. According to results
found by other researchers (Guillard et al., 2003b),
it seems that, for the same TiO2 amount, the catalyst
fixation does not affect the rate of photodegradation
but only the adsorption capacity of TiO2.
3.3. Inactivation of E. coli with different TiO2 in
suspension
Bacterial inactivation by artificial light (500 W/
m2) occurs in absence or presence of catalyst as
shown in Fig. 5. The inactivation by light (trace
(m)) is enhanced by the presence of all the different
TiO2. The total time of bacterial abatement is short-
er in the presence than in absence of TiO2. This re-
sult is in accordance with previous findings (Rincón
et al., 2001; Rincón and Pulgarin, 2003; Mills and
Le Hunte, 1997; Blake et al., 1999; Robert et al.,
1999). On the opposite to the degradation of gallic
AMT100 acid, it seems that the different characteristics of
the TiO2, such as the surface area and the particle
PC500
size, do not influence their bacterial inactivation effi-
ciencies. This result suggests that the mechanisms of
interaction at the surface of TiO2 between the oxida-
tive species generated and bacteria are different that
those observed with organic compounds.
4 5 6 7 8 3.4. Inactivation of E. coli with fixed TiO2
Fig. 6 shows the inactivation of E. coli as a func-
tion of time with catalyst in suspension or coated on
fibrous web and without TiO2 addition at an
1380 D. Gumy et al. / Solar Energy 80 (2006) 1376–1381

20 20
3
18 18
2 2.5
16 16 2

k (1/h)

k (1/h)
Gallic acid (mg/L)

Gallic acid (mg/L)

14 14 1.5
1 y = 0.0278x
2 y = 0.0281x
12 R = 0.9951 12 1
2
R = 0.9673
10 10 0.5
0
0 50 100 0
8 8 0 50 100
Ti O2 (mg)
Ti O2 (mg)
6 6

4 darkness 4 darkness
adsorption adsorption
2 2

0 0
-60 -30 0 30 60 90 120 150 180 -60 -30 0 30 60 90 120 150 180
Time (min) Time (min)
(a) Fixed (b) Suspended

Fig. 4. (a) Adsorption and photodegradation (500 W/m ) of gallic acid (20 mg l ) with TiO2 P25 coated on 60 cm2 of web at different
2 1

area loading: 13 (g/m2) (n), 10 (g/m2) (d) 5.5 (g/m2) (h) as a function of time. Each point represents an average value of three
experiments. The bars show the standard deviation. (b) Adsorption and photodegradation (500 W/m2) of gallic acid (20 mg l 1) with TiO2
P25 in suspension at different concentration: 0.2 (g/l) (m), 0.6 (g/l) (s) and 1.0 (j) (g/l) as a function of time. Each point represents an
average value of three experiments. The bars show the standard deviation. The inserts show plots of pseudo-first order rate constants of
gallic acid disappearance with P25 coated on fibrous web (a) and P25 in suspension (b) as a function of the amount of TiO2 in the total
system (100 ml).

Table 2
Concentration, amount and corresponding pseudo-first order rate constants for gallic acid disappearance of suspended and fixed TiO2 P-
25
Suspended P-25 TiO2 (100 ml solution) Fixed P-25 TiO2 (60 cm2 paper in 100 ml solution)
TiO2 amount (mg) TiO2 conc. (g/l) k (1/h) TiO2 amount (mg) TiO2 area loading (g/m2) k (1/h)
20 0.2 0.80 33 5.5 0.97
60 0.6 1.59 60 10.0 1.64
100 1.0 2.83 78 13.0 2.17

1.0E+06
1.E+07

1.0E+05
1.E+06
Survival (CFU/mL)
Survival (CFU/mL)

1.E+05 1.0E+04

1.E+04
1.0E+03

1.E+03
1.0E+02
1.E+02
1.0E+01
1.E+01

1.0E+00
1.E+00
0 20 40 60 80 100 120
0 30 60 90 120 150 180
Time (min) Time (min)

Fig. 5. Photoinactivation (500 W/m2) of E. coli with different Fig. 6. Photoinactivation (1000 W/m2) of E. coli with TiO2 P25
TiO2 in suspension (1 g/l): PC100 (j), PC500 (h), AMT100 (d), coated on fibrous web (13 g/m2) (s), with TiO2 in suspension
AMT600 (s), and without (m) TiO2 addition as a function of (0.2 g/l) (h), without TiO2 addition (j) and with TiO2 P25
time. Each point represents an average value of three experi- coated on fibrous web (13 g/m2) in dark condition () as a
ments. The bars show the standard deviation. function of time. Each point represents an average value of three
experiments. The bars show the standard deviation.

intensity of 1000 W/m2. The required time to inac- (trace h), even with a lower amount of suspended
tive bacteria using immobilized TiO2 (trace s) was TiO2 (78 mg of immobilized TIO2 vs. 20 mg of sus-
higher than that observed with suspended TiO2 pended TiO2, Table 2). In the dark, during 2 h for
 D. Gumy et al. / Solar Energy 80 (2006) 1376–1381
both TiO2 fixed or suspended, inactivation or bacte-
rial adsorption on fibrous web do not occur (trace
).
As there is no adsorption of bacteria on the fixed
TiO2 and that reactions occur certainly mainly at
(or close to) the surface of TiO2 and not in the solu-
tion, the average distance between bacteria and cat-
alyst appears to be an important limiting factor for
the photocatalytic inactivation of bacteria. The
average half-distance between bacteria and immobi-
lized TiO2 increases and the probability of attack by
OH decreases as compared to suspended TiO2.
4. Conclusions
Water disinfection and detoxification using pho-
tocatalytic process have been studied here. These
two processes were found to be influenced in a dif-
ferent way by the physical characteristics of the cat-
alyst and by its coating on support. The results
suggest that the mechanisms of interaction between
TiO2 and the targeted contaminants (organics or
microorganisms) are not the same in the two pro-
cesses. It is important to take these differences into
account for the design of a process for the photocat-
alytic treatment of water.
Acknowledgements
The authors wish to thank the OFES for its
financial assistance (Projects Nos. 02.0030 and
02.0316 within the SOLWATER and AQUACAT
Contract Nos. ICA4-CT.2002-10001 and ICA3-
CT-2002-10016 of the European Community) and
Ahlstrom Research & services for supplying photo-
catalytic nonwoven webs. Our special thanks to
Environnement-EPFL-Colombie cooperation pro-
gram supported by VPRI-cooperation in EPFL.
1381
References
Blake, D., Maness, P., Huang, Z., Wolfrum, E., Huang, J., 1999.
Application of the photocatalytic chemistry of titanium
dioxide to disinfection and the killing of cancer cells. Sep.
Purif. Methods 28, 1–50.
Block, S.S., Goswami, D.Y., 1997. Chemically enhanced sunlight
for killing bacteria. J. Sol. Energy Eng. 119, 85–91.
Guillard, C., Lachheb, H., Houas, A., Ksibi, M., Elaloui, E.,
Herrmann, J.-M., 2003a. Influence of chemical structure of
dyes, of pH and of inorganic salts on their photocatalytic
degradation by TiO2 comparison of the efficiency of powder
and supported TiO2. J. Photochem. Photobio., A 158, 27–36.
Guillard, C., Disdier, J., Monnet, C., Dussaud, J., Malato, S.,
Blanco, J., Maldonado, M., Herrmann, J., 2003b. Solar
efficiency of a new deposited titania photocatalyst: chloro-
phenol, pesticide and dye removal applications. Appl. Catal.
B: Environ. 46, 319–332.
Matsunaga, T., Okochi, M., 1999. TiO2-mediated photochemical
disinfection of Escherichia coli using optical fibers. Environ.
Sci. Technol. 29, 501–505.
Matsunaga, T., Tomoda, R., Nakajima, T., Wake, H., 1985.
Photo-electrochemical sterilization system that uses photo-
semiconductor powders. FEMS Microbiol. Lett. 29, 211–214.
Mills, A., Le Hunte, S., 1997. An overview of semiconductor
photocatalysis. J. Photochem. Photobiol., A 108, 1–35.
Parra, S., Olivero, J., Pulgarin, C., 2002. Relationships between
physicochemical properties and photoreactivity of four bior-
ecalcitrant phenylurea herbicides in aqueous TiO2 suspension.
Appl. Catal. B: Environ. 36, 75–85.
Rincón, A.G., Pulgarin, C., 2003. Photocatalytical inactivation of
E. coli: effect of (continuous-intermittent) light intensity and
of (suspended-fixed) TiO2 concentration. Appl. Catal. B:
Environ. 44, 263–284.
Rincón, A.G., Adler, N., Peringer, P., Pulgarin, C., 2001.
Interaction between E. coli inactivation and DBPs precur-
sors—in the photocatalytic process of drinking water with
TiO2. J. Photochem. Photobiol., A 139, 233–241.
Robert, D., Piscopo, A., Heintz, O., Weber, J.V., 1999. Photo-
catalytic detoxification with TiO2 supported on glass-fibre by
using artificial and natural light. Catal. Today 54, 291–296.
Wei, C., Lin, W.Y., Zainal, Z., Williams, N.E., Zhu, K., Kruzic,
A.P., Smith, R.L., Rajeshwar, K., 1994. Bactericidal activity
of TiO2 photocatalyst in aqueous-media—toward a solar-
assisted water disinfection system. Environ. Sci. Technol. 28,
934–938.