Catalysis Today 240 (2015) 3038

Contents lists available at ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

Disinfection of urban efuents using solar TiO2 photocatalysis:

A study of signicance of dissolved oxygen, temperature, type
of microorganism and water matrix
Irene Garca-Fernndez, Isabel Fernndez-Calderero, Mara Inmaculada Polo-Lpez,

Pilar Fernndez-Ibnez
Plataforma Solar de Almera-CIEMAT, P.O. Box 22, 04200 Tabernas, Almera, Spain

a r t i c l e

i n f o

Article history:
Received 28 January 2014
Received in revised form 10 March 2014
Accepted 13 March 2014
Available online 13 April 2014
Keywords:
Fusarium solani
Escherichia coli
TiO2 photocatalysis
WW treatment
Temperature
Dissolved oxygen

a b s t r a c t
The enhancement of current technologies used to treat polluted water is one of the most important
challenges in water research. The application of physico-chemical treatments could reduce the load
of chemical and biological pollutants present in WW reducing the pressure over water requirements,
allowing the reclaim of the treated water. Advanced Oxidation Processes (AOPs) and, in particular, photocatalysis using titanium dioxide (TiO2 ) have shown a great potential for chemicals removal as well as for
pathogens reduction in water. Moreover, the use of solar Compound Parabolic Collectors (CPC) reactors
has been also shown to be very effective for water treatment purpose by solar photocatalysis. Nevertheless, the effects of some key parameters in photocatalytic disinfection have not been already investigated
at pilot scale in solar reactors; like dissolved oxygen concentration, water temperature, water matrix
composition and the type of microorganism. The roles of these parameters in photocatalytic processes
are individually known for chemicals degradation, but their relative signicance in water photocatalytic
disinfection has been never studied at pilot scale. The aim of this work was to investigate the inuence of these parameters on the disinfection efciency using a solar 60 L-CPC reactor with suspended
TiO2 (100 mg/L). The following variables were experimentally evaluated: injection of air in the reactor
(160 L/h); different controlled temperatures (15, 25, 35 and 45 C); two very different models of water
pathogen, Escherichia coli (model of fecal water contamination) and Fusarium solani spores (a highly
phytopathogenic fungus); and the chemical composition of the water comparing urban WW efuents
(UWWE) and simulated urban WW efuent (SUWWE). The increase of water temperature (from 15
to 45 C) had a benet on the disinfection rate for both pathogens in all the experimental conditions
evaluated. The air injection led to an important enhancement on the inactivation efciency, which was
stronger for F. solani spores, the most resistant microorganisms to TiO2 photocatalysis. The composition
of the water matrix signicantly affected the efciency of the photocatalytic treatment, showing a better
inactivation rate in SUWWE than for UWWE.
2014 Elsevier B.V. All rights reserved.

1. Introduction
Recently, big efforts have been done to develop alternative
water treatment systems to decontaminate wastewater (WW)
using processes based in reactive oxygen species (ROS). The photocatalytic treatments which use solar light have gained great
attention as they produce ROS to destroy organic contaminants and
kill microorganisms in water with very low energy consumption. In

Corresponding author. Tel.: +34 950 387957; fax: +34 950 363015.

E-mail address: pilar.fernandez@psa.es (P. Fernndez-Ibnez).

http://dx.doi.org/10.1016/j.cattod.2014.03.026
0920-5861/ 2014 Elsevier B.V. All rights reserved.

the last decade, a number of articles have demonstrated the capability of solar photocatalysis to decontaminate and disinfect water
polluted by organic and biological agents [1].
Reuse of treated WW is an alternative resource of water since
freshwater scarcity and lack of access to safe water is a human
sizeable problem today [2]. Wastewater must be treated before
discharge or for restricted reuse; it may contain industrial and agriculture chemical pollutants and also a wide range of pathogens, i.e.
bacteria, viruses and fungi [3].
Agriculture is probably the most affected eld by fungal
pathogens like Fusarium spp. which is especially harmful in intensive agriculture [4]. Fusarium has been reported to be highly

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

resistant to chemical and photocatalytic treatments due to its

spores [5,6]. Fusarium spp. have been associated with human disseminated infections which recently has increased, particularly
in patients with underlying immunosuppressive conditions, as
leukemia, cancer or AIDS patients [7]. On the other hand, WW is
frequently loaded with Escherichia coli, fecal indicator organism
whose presence in water indicates possible contamination with
other enteric pathogens like Salmonella, Shigella or Yersinia, enteric
bacteria that could cause gastrointestinal diseases that generally
presents with diarrhea [8].
Advanced Oxidation Processes (AOPs) are a good alternative
for the traditional disinfection methods which are limited and
have some drawbacks. AOPs are based on physicochemical processes that produce powerful oxidizing species, mainly hydroxyl
radicals ( OH), in situ. Heterogeneous photocatalysis with the semiconductor TiO2 is an AOP that has been used to decontaminate
water containing hazardous pollutants and for disinfection of some
pathogens [9]. TiO2 -photocatalysis in water use the UVA radiation
( < 387 nm) to excite the photocatalyst, that in presence of oxygen and in contact with the water produces OH (summarized in
Eqs. (1) and (2)). Hydroxyl radicals have a high reactivity and therefore a short half-life. Thus the process will be favored when TiO2 is
adsorbed over (or very close to) the cell wall of the microorganism
and this is induced by different electrical charge of microorganism and catalyst [10]. The oxidative action alters cell components
and their functionality, causing a loss of cell integrity, changing the
permeability and diffusion of cellular components to the medium,
ending in cell death.
TiO2 + hv
H2 O + h +
BV

O2(ads) + e
BC

+ h+
BV

(1)

OH + H+
aq

(2)

e
BC

O2(ads)

(3)

However, if oxygen concentration is low in the system, electrons are not caught (Eq. (3)) and then e /h+ pairs can recombine
and inhibit the production of OH. To avoid this, the presence of
oxygen is always required for proper efciency of the photocat
alytic process. Moreover, O2 is also oxidant (ROS) and therefore
acts as a disinfectant causing DNA damage [11].
A number of articles have demonstrated the capability of TiO2
photocatalysis for bacteria, viruses, protozoa, prions or fungi inactivation in water [12,13]. Rincn et al. demonstrated the benets
of using TiO2 for disinfecting distilled and lake water polluted with
E. coli [14]. Seven et al. inactivated E. coli, Pseudomonas aeruginosa,
Staphylococcus aureus, Candida albicans and Saccharomyces cerevisiae using TiO2 with very good results in all cases [15]. Our group
has also studied the photocatalytic susceptibility of a variety of
Fusarium spp. and have demonstrated that TiO2 photocatalysis can
be used with CPC reactors for eliminating pathogens present in
water [16,17].
Water temperature has direct effects on disinfection efcacy,
but it has not investigated up to date in photocatalytic water disinfection. The disinfecting effect of high temperatures is well known
as pasteurization. Thermal increasing beyond the optimal temperature, drastically reduce the viability of microorganisms due to loss
of integrity of proteins, enzymes and genetic material. The optimal
temperature for the most bacteria is between 35 and 40 C; while it
is around 28 C for the most of the fungus [18]. Moreover, the lethal
action of solar mild heat over bacteria has been investigated at temperatures between 40 and 52 C [1921]. Given the importance of
temperature in microbial metabolism and disinfection, up to our
knowledge there is no research done on the effects of temperature of photocatalytic disinfection of water. This contribution will
experimentally undertake this aspect at pilot scale with controlled
temperature.

31

The role of oxygen in metabolic processes is already well known.

Oxygen crosses membranes freely and ROS are generated internally due to aerobic metabolism. This increases the oxidative stress
inside cells [22]. It has been proven that an increase of dissolved
oxygen (DO) in water induced by proper agitation of batch containers of SODIS accelerates the disinfecting effect of solar radiation
[23]. Studies by Reed et al. [24] found a 48 times faster inactivation
rate of fecal bacteria in oxygenated water, compared to deoxygenated water [2527]. Furthermore, these authors demonstrated
that if water was bubbled with nitrogen before to solar exposure, it
resulted in worst inactivation efciency than aired samples [24].
Therefore, oxygen was conrmed as essential to disinfect fecal
coliforms in batch solar disinfection process. In solar photocatalytic disinfection the DO plays also an important role. Rincn
et al. reported an increased photocatalytic inactivation of E. coli
and Bacillus spp. in water with conditions of oxygenation >8 mg/L
(supplying oxygen as bubbled pure O2 ) [28]. This work conrmed
the importance of oxygen even for disinfection in non-controlled
temperature conditions at small scale (4 L). The present work aims
to experimentally evaluate this factor but with controlled temperature at pilot scale (60 L). The disinfection of contaminated
SUWWE and UWWE with Fusarium solani and E. coli using solar
photocatalysis with TiO2 was evaluated at different temperatures
(15, 25, 35 and 45 C) with and without air injection in the solar
reactor.

2. Materials and methods

2.1. Solar CPC pilot plant
The solar CPC pilot plant used consists of two CPC mirror modules titled 37 (Fig. 1a). Each CPC mirror module is made up of 10
borosilicate glass tubes of 1500 mm long, 2.5 mm thick and 50 mm
outer diameter (UVA-transmission: 90%). The CPC mirror is made
of highly reective anodised aluminum with concentration factor
CF = 1 (95% total reectivity). The ratio of irradiated water (45 L) to
total water (60 L) is 75%, with a CPC of 4.5 m2 . Water is recirculated
through the tubes to a tank by a centrifugal pump (150W, Mod.
NH-200 PS PanWorld, USA). Flow was controlled by a Yokogawa
magnetic ow meter (Admag, RXF, Yokogawa Electric Corporation, Japan) at 30 L/min (turbulent ow; Reynolds: 16600) to avoid
catalyst sedimentation. Online sensors for pH, DO and temperature (Crison, Spain) acquired the measurements during the tests
(Fig. 1b). Temperature was controlled at: 15, 25, 35 and 45 C using
a heating electric resistance to increase the water temperature, and
a cooling system (Fig. 1c). DO in the water was increased by two
air pumps that inject air (160 L/h) at two equidistant points of the
solar reactor system (Fig. 1a and b).

2.2. Water sources

Simulated urban WW treatment plant efuent (SUWWE) was
used as a synthetic model of WW efuent with 25 mg/L of dissolved
organic carbon (DOC). This water was chosen because it contains
organic matter (urea, peptone and meet extract) [29]. The ionic
content in this SUWWE was measured (Table 1).
Urban WW treatment plant efuent (UWWE) of Almera, El
Bobar (Spain), was used as real efuent of a secondary treatment. UWWE was freshly collected from the treatment plant in
the morning of each disinfection assay. Every UWWTE stock used
was characterized and the average of the inorganic chemical compounds measured is shown in Table 1. The concentration of ions was
measured with a Dionex DX-600 and Dionex DX-120 (California,
USA) ion chromatographs for anions and cations.

32

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

Fig. 1. Images of the 60 L-CPC reactor at PSA facilities. Front view of the solar CPC reactor (4.5 m2 of collector mirrors) with air injection points indicated (a), side view of CPC
reactor: air injection and DO probe (b), and cooling and heating systems (c).

2.3. Generation and enumeration of F. solani spores

Spores of F. solani strain (CECT 20232) were used. The enumeration and quantication methods used in this work have been
described elsewhere [5,16]. Small pieces (1 cm) of agar containing the culture of F. solani were transferred to a poor nutrient agar
(15 g/L of KCl and 5 g/L of bacteriological agar). After 1520 days
of incubation at 25 C under UV-C radiation (Mercury lamp, 40W)
the spores were recovered by washing Petri dishes with sterile
distilled water yielding an average of 105 spores per mL. Spores
concentration was determined by direct counting with a Neubauer
plate (Brand, Germany) using a phase contrast microscope (Nikon,
Japan) and diluted in the tank of the CPC reactor until desired initial concentration, 103 colony forming units per mL (CFU/mL) for
Table 1
Chemical characterization of SUWWE and UWWE.

Na (mg/L)
NH4 + (mg/L)
K+ (mg/L)
Mg2+ (mg/L)
Ca2+ (mg/L)
SO4 2 (mg/L)
Cl (mg/L)
NO3 (mg/L)
PO4 3 (mg/L)
pH
Conductivity (S/cm)
Turbidity (NTU)
DOC (mg/L)
DIC (mg/L)
E. coli (CFU/mL)

SUWWE

UWWE

35.80 1.10
2.70 1.00
3.40 0.60
17.20 0.30
21.63 2.30
9.00 1.40
11.50 2.10
130.40 7.60
12.10 3.00
8.15 0.30
362 12
1.50 0.10
2030
0.54

193.9 4.4
38.6 19.4
27.1 2.0
32.5 2.9
66.1 2.2
69.9 34.7
353.3 27.5
15.1 13.4
47.3 50.1
7.4 0.2
1617.1 226.2
15.1 0.1
1529
5060
1000 2501

DOC = dissolved organic carbon; DIC = dissolved inorganic carbon.

all experiments. Plate counting technique with acidied malt agar

(20 g/L of malt extract, 20 g/L of bacteriological agar and 0.25 g/L of
citric acid) was used to determine the spore concentration throughout the solar experiments. Samples volumes of 50250500 L
were plated by the spread plate method. The detection limit (DL)
of this experimental method was 2 CFU/mL. Fungal colonies were
counted after 48 h of incubation at 28 C in dark.
2.4. Generation and enumeration of E. coli
E. coli K-12 (ATCC 23631) was spiked in SUWWE tests. E. coli
was cultured in Luria broth (SigmaAldrich) and incubated at 37 C
with agitation under aerobic conditions. Bacteria were collected
after 20 h, yielding a concentration of 109 CFU/mL. E. coli suspensions were harvested by centrifugation at 800 g for 10 min. The
bacterial pellet was re-suspended in Phosphate Buffer Saline (PBS)
and diluted in the reactor to have a 106 CFU/mL initial concentration. The samples were enumerated using plated counting through
a serial 10-fold dilutions in PBS and volumes of 20 L were plated in
triplicate on Luria agar Petri dishes (SigmaAldrich). Colonies were
counted after incubation of 24 h at 37 C, with a DL of 4 CFU/mL
[6,16,30]. For UWWE experiments, the naturally occurring E. coli
was evaluated by spreading different sample volumes (25500 L)
on ChromoCult Coliform Agar (Merck) plates (DL = 2 CFU/mL).
2.5. Solar photocatalytic experiments in the CPC reactor
Titanium dioxide (Aeroxide P-25, Evonik, Germany) was used
as received at a concentration of 100 mg/L, according to previous
work [5]. All experiments were done under natural solar radiation
at Plataforma Solar de Almera (Southeast of Spain). Previously to
solar exposure, in SUWWE assays, TiO2 powder and bacteria or

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

fungal suspensions were added to the 60 L-CPC reactor in the dark

and it was kept re-circulating for 15 min (adaptation and homogenization time). For UWWE assays, a similar procedure was followed,
excluding the addition of E. coli suspensions as the naturally occurring E. coli was evaluated instead. The rst sample was taken and
the reactor was uncovered and exposed to solar radiation for 5 h,
under clear sunny days. Samples were taken at regular intervals,
and evaluated according to the previously described enumeration
methods. After bacterial/fungi spiking, a water sample was kept in
the dark at room temperature and plated by that time and again
at the end of the experiment as a control, to check that any factor
merely outside of the solar process is affecting the microorganism
viability. Bacterial and spore re-growth counts were done for all
experiments by leaving the last two experimental samples at room
temperature for 24 h and 48 h overnight after the solar treatments.
Non-regrowth was observed in any case (<DL). All experiments
were done in triplicate in consecutive days, with very similar solar
irradiance. The results of the three repetitions were highly reproducible. The results presented are the average of the replicates with
standard deviation as error bar. Results were analyzed through a
one-way ANOVA (P < 0.05, condence >95%, Origin v7.03, OriginLab Corp., Northampton, USA), reporting a 95% condence level for
the average colony concentration and error. UV radiation was measured with a global UV-A pyranometer (CUV4, Kipp & Zonen) during
all experiments. The inactivation curves are presented versus QUV
(kJ/L), the solar UVA energy per unit of volume received in the
photo-reactor [17].
2.6. Kinetics evaluation
Four different kinetics models have been used to t the
experimental data obtained during the photocatalytic tests: (1) a
log-linear according to the Chick law (Eq. (6)). This model reduces
disinfection to a bimolecular chemical reaction in which microorganisms are treated as molecular species. There is an extensive
application of this equation using other disinfecting agents i.e.
chlorine, ozone, hydrogen peroxide and chloramines. (2) A double
log-linear kinetics (Eq. (7)), with a rst stage of very fast (k1 > k2 )
inactivation and a second phase of attenuated inactivation (k2 ) [28].
(3) A log-linear region followed by a tail (Eq. (8)). The tail shape
represents the bacterial population remaining at the end of the
experiment due to a strong reduction on the photocatalytic activity and/or the presence of a population of cells resistant to the
treatment [31,32]. (4) An initial delay or very smooth decay at the
beginning (shoulder), attributed to lose of cells viability after the
accumulation of oxidative damages during the process, followed by
a log-linear decrease (Eq. (9)) [31,33].
Log

N


Log

N0
N
N0

= k t

(6)


= k1 t; t = [0, t1 ];

Log

N
N0


= k2 t; t = [t1 , t2 ]

dN
N Nres
= k1 N
= ek1 t
N0 Nres
dt
Log

N
N0

= k1 t

NN0 ; 0
N < N0 ; ek1 (NSL)

(7)
(8)


(9)

where N/N0 is the bacteria or spores concentration reductions, ki

is the disinfection kinetic rate and t is the time of treatment, Nres is
the residual population density, and SL = Shoulder length (min1 ).
All experimental data from disinfection tests were tted to the
proposed equations. Those results which led to a minimum R2 coefcient were accepted as the best statistical tting. The results of
these ttings are shown in Table 2, including R2 values. Kinetic
parameters obtained for the different equation are comparable, as

33

all are obtained under same experimental (reactor, illumination,

protocols, etc.) conditions.
3. Results
3.1. Control of water temperature and dissolved oxygen in the
CPC reactor
Previous to carry out the solar experiments, the capability
heating-cooling system to maintain the water temperature at 15,
25, 35, and 45 C constant was evaluated. The experiments were
conducted with and without air injection, measuring at the same
time DO. These tests were run in the CPC reactor for 5 h under natural solar radiation in distilled water at 30 L/min of ow rate and
results are shown in Fig. 2. The water temperatures measured in the
CPC reactor for each case were: 15.0 1.4, 25.0 1.9, 35.0 1.0 and
45.0 1.5 C. In all cases, the DO values measured were higher for
experiments with air sparging than for those without injection of
air, except for 15 C, where no signicant differences were observed
(Fig. 2).
3.2. Effect of temperature and dissolved oxygen on the
inactivation of E. coli and F. solani spores
Control tests in the dark with air injection were done previously to solar experiments in SUWWE to determine the inuence
of each temperature and air injection over the viability of E. coli and
F. solani spores in the 60 L-CPC reactor with a ow rate of 30 L/min
and 100 mg/L of TiO2 . The concentration of both pathogens, i.e.
106 CFU/mL for E. coli and 103 CFU/mL for F. solani spores, remained
constant at all temperatures evaluated for 5 h, except for 45 C,
where less than 1-log reduction in the concentration of E. coli was
observed (data not shown). Therefore, no signicant detrimental
effect of the operational conditions was observed over bacteria and
spores.
3.2.1. Case 1: SUWWE
Fig. 3a shows the photocatalytic inactivation of E. coli at 15, 25,
35 and 45 C without air injection. An enhancement of the E. coli
inactivation efciency was observed as temperature rises from 15
to 45 C. In the case of 15 C, a 4-log reduction was observed with
a QUV = 29.2 kJ/L, and the DO varied from 9.0 to 7.8 mg/L. At 25 C, a
5-log decrease was found, although the DL was not achieved with
40 kJ/L of QUV , and DO ranged from 6.0 to 5.0 mg/L. A 6-log reduction was observed at 35 and 45 C with a QUV of 23.3 and 14 kJ/L,
respectively. In these experiments, DO varied from 6.3 to 5.1 mg/L
at 35 C, and from 5.1 to 3.7 mg/L at 45 C. Due to the slight uctuations of DO measurements from the beginning to the end of the
experiment, we report the average DO for each experimental condition. The DO values decreased as temperature increased, i.e. 8.1,
5.5, 5.8 and 4.4 mg/L at 15, 25, 35 and 45 C, respectively.
The solar photocatalytic inactivation of E. coli at 15, 25, 35 and
45 C injecting air is shown in Fig. 3b. Again, the inactivation kinetics
enhanced when the temperature increased from 15 to 35 C, while
very similar results at 45 and 35 C were observed. At 15 C, a 4.5-log
reduction was obtained (QUV = 37.6 kJ/L, DO = 8.9 mg/L). In the case
of 25, 35 and 45 C, DL was achieved with 25.9, 11.4 and 12.8 kJ/L
of QUV with DO concentration of 6.8, 6.0 and 5.2 mg/L, respectively.
Fig. 4a shows the inactivation of F. solani spores at different
temperatures without air injection. The concentration of fungi was
reduced only 2-log at 15 and 25 C. At 35 C a 3-log decrease (reaching the DL) was observed with QUV = 44 kJ/L. Best inactivation result
was obtained at 45 C, where the DL was reached with 18 kJ/L of
QUV . DO average values at 15, 25, 35 and 45 C in these experiments were 7.1, 6.4, 5.2, and 5.5 mg/L, respectively. Fig. 4b shows F.

34

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

Table 2
E. coli and F. solani spores inactivation rates (k) during solar photocatalytic disinfection of SUWWE and UWWE.
T ( C)

k1 (min1 )

R2 1

k2 (min1 )

R2 2

SL (min)

Log (Nres )

Model #

E. coli, without air injection, SUWWE (Fig. 3a)

0.033 0.002
15
25
0.051 0.007
35
0.069 0.004
0.107 0.030
45

0.998
0.965
0.947
1.000

0.010 0.001
0.015 0.004

0.970
0.957

1.83
0.30

3
3
2
2

E. coli, with air injection, SUWWE (Fig. 3b)

0.026 0.004
15
0.053 0.011
25
35
0.060 0.020
0.117 0.012
45

0.974
0.920
0.867
0.990

0.006 0.004

0.011 0.003

0.690

0.966

1.20

3
2
1
2

F. solani, without air injection, SUWWE (Fig. 4a)

0.007 0.001
15
25
0.009 0.001
35
0.013 0.002
45
0.020 0.005

0.9679
0.9778
0.9830
0.9686

0.003 0.002

0.866

60

1
4
2
1

F. solani, with air injection, SUWWE (Fig. 4b)

0.013 0.001
15
0.023 0.001
25
0.041 0.001
35
0.059 0.001
45

0.995
0.998
0.999
0.999

1
1
1
1

E. coli, without air injection, UWWE (Fig. 5a)

0.007 0.0005
15
0.013 0.001
25
35
0.021 0.003
0.013 0.001
45

0.9588
0.977
0.955
0.957

90

4
1
1
1

E. coli, with air injection, UWWE (Fig. 5b)

0.008 0.001
15
0.017 0.002
25
0.025 0.004
35
45
0.024 0.003

0.976
0.977
0.966
0.973

1
1
1
1

F. solani, without air injection, UWWE (Fig. 6a)

0.001 0.001
15
25
0.005 0.001
0.005 0.001
35
0.022 0.003
45

0.966
0.997
0.953
0.980

0.0031 0.0003

0.9678

30

1
4
1
2

F. solani, with air injection, UWWE (Fig. 6b)

15
0.001 0.001
25
0.004 0.001
35
0.005 0.001
45
0.024 0.002

0.817
0.955
0.967
0.983

1
1
1
1

k = inactivation rate (linear regression of Log (concentration) versus time); R2 = regression coefcient; Model 1 = Log-lineal (k1 ); Model 2 = double log-lineal (k1 , k2 ); Model
3 = Log-lineal (k1 ) + tail (Log(Nres )); Model 4 = shoulder (SL) + log-lineal (k1 ).

solani inactivation at same temperatures with air injection. Similar tendency as previously observed was obtained for different
temperatures (1545 C). DL was reached for all cases, with an accumulated solar UV-A energy of 19.4, 8.8, 5.2 and 3.0 kJ/L at 15, 25,
35 and 45 C, with DO values of 7.8, 6.7, 5.7, and 5.1 mg/L, respectively. Best spores inactivation efciency was observed again at the
highest temperature (45 C).
3.2.2. Case 2: UWWE
Fig. 5a shows photocatalytic inactivation of natural occurring
E. coli without air injection at different temperatures. Complete
removal of E. coli, from 103 104 CFU/mL to DL was achieved in all
cases except for 15 C, where a 1.5-log reduction was observed. The
QUV values required to reach the DL at 25, 35 and 45 C were 43.8,
18.2 and 42.5 kJ/L, respectively. Best inactivation results were found
at 35 C. DO concentration decreased as temperature increase, from
9.2 mg/L at 15 C to 5.1 mg/L at 45 C. The same experiments with
injected air are shown in Fig. 5b. DL was achieved in all cases except
for 15 C, where a 2-log reduction of E. coli was observed with
23.8 kJ/L of QUV . 31.5 kJ/L was required for complete inactivation
at 25 C. Best inactivation results were found at 35 C with a QUV
of 17.5 kJ/L, while 31.1 kJ/L were necessary to reach DL at 45 C. DO

concentration decrease as temperature increase from 9.4 mg/L at

15 C to 6.4 mg/L at 45 C.
Fig. 6a shows the photocatalytic inactivation of F. solani spores
in UWWE at different temperatures without air injection. At 15 C,
the concentration of spores remained almost constant for 5 h. Very
similar kinetics, which resulted in 1-log spore reduction after 5 h,
were observed at 25 and 35 C; although 35 C lead to a slightly
faster inactivation. Complete Fusarium inactivation was attained
only at 45 C, a 3-log decrease until the DL, with 54.7 kJ/L of QUV .
Same experimental results in UWWE with air sparging are shown
in Fig. 6b. The concentration of spores remained constant at 15 C.
1-log reduction was observed at 25 C (47.1 kJ/L), a 2-log decay was
observed with 49.1 kJ/L at 35 C, while complete inactivation was
observed at 45 C, with 19.7 kJ/L. The DO values of these experiments had similar tendency than in previous experiments with
E. coli.
Very similar results in E. coli inactivation were observed with
and without air injection in both types of WW (Figs. 3 and 5),
although there is an enhancement with air at 45 C (Table 2),
highlighting the importance of temperature in the photocatalytic
disinfection in spite that this temperature was not responsible for
thermal inactivation (Section 3.2).

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

35

Fig. 2. Registration of temperature in the CPC reactor in experiments conducted at water temperature of 15 C (a), 25 C (b), 35 C (c), and 45 C (d). Solids lines (- - -) and
(- - -) represent DO concentration values measured in injected and no injected air experiments in each temperature.

For the case of F. solani, the air supply and thermal increase produced an enhancement in the photocatalytic efciency (Table 2,
Figs. 4 and 6). Although the air sparging in the reactor does not
produce signicant increases in the measured DO in water, it positively affects the photocatalytic efciency due to local increases
of DO in some parts of the reactor. The open design of this reactor
and the air spargers did not permit increase the DO measurement,
but to enhance the photocatalysis efcacy observed only when the
microorganism is enough resistant.
4. Discussion
Our previous studies on photocatalytic disinfection of water
with CPC systems did show very promising results [5,16,17], nevertheless the temperature in the whole reactor system was not
controlled. The ambient temperature at this place ranges from 15
to 40 C (during the experiments) and the water temperature can
vary from 5 to 55 C in the CPC reactor, depending of the season. The
inactivation results at xed temperatures (15, 25, 35, and 45 C),
demonstrated that inactivation efciency of E. coli and F. solani
spores is enhanced when the water temperature increased.
It is well known that photo-excitation of TiO2 is not affected by
the temperature in the range of 2080 C [34]. Nevertheless, the
ionic products of water (OH , H3 O+ ) increase with temperature
from 20 to 60 C [35], which induces an increase in the concentration of hydroxyl anions. In aqueous solution, the photo-generated
holes in the TiO2 valence band are trapped by hydroxyl anions to
form OH (Eq. (2)). Thus, the rise of OH as temperature increases
makes the hole trapping and the hydroxyl radicals generation to
be more efcient. Actually, Janus et al. used 2-hydroxytepephthalic
acid to demonstrate the correlation between hydroxyl radicals formation and temperature increase, from 20 to 60 C, which resulted

in an increase on the photocatalytic activity of TiO2 [36]. This

explains the better photocatalytic inactivation efciency at higher
temperatures that we observed in this work.
Thermal death of E. coli and F. solani is discarded as a reason for
their inactivation during the solar experiments. Control tests in the
dark (Section 3.2) demonstrated there are no damages resulting in
viability loses for both pathogens during 5 h.
E. coli growths in the range of 1047 C, with optimum temperature of 3744 C [18]. When temperature is far from optimal,
signicant changes in the metabolism of cells occur. At temperatures below 10 C, the uidity of cell membrane is drastically
reduced and the metabolism slows down. At the optimum temperature, the metabolic activity is maximal; the genetic material is
unpacked and completely active, and therefore more vulnerable to
oxidative stress due to increased surface of DNA [37]. The reduced
SOS response in E. coli when temperature increases from 10 to 20 C,
being constant until 40 C has been proven [38]. In our results, when
temperatures raised from 15 to 35 C, we observed an increase in
efciency of the disinfection process, which can be explained by the
increased activity of the metabolism of bacteria cells; while no signicant different results were observed between 35 and 45 C, as at
these temperatures there are no metabolic differences in the bacteria. Moreover, at higher temperatures, the defence mechanisms of
bacteria decrease their capability.
Similar effect of temperature was observed for F. solani.
Spores are biological structures of resistance against stress factors
present in different environments. Some parameters as temperature (ranged from 25 to 30 C), acid pH, UV-light, water activity and
organic and inorganic chemical compounds, favor the spore germination [39]. Temperatures above 20 C promote the germination
process, which involves the modication of the cell wall for water
inux inside the spore core. During germination, the structure of

36

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

a) 0

a)
0

F. solani (Log N/N0)

-2
-3
-4
-5
-6

E. coli (Log N/N0)

b)

-1

-2

DL

DL

10

15

20

25

30

35

-3

40

10

15

QUV (kJ/L)

25

30

35

40

45

30

35

40

45

b)

-1
-2
-3
-4
-5
-6

20

QUV (kJ/L)

F. solani (Log N/N0)

E. coli (Log N/N0)

-1

DL

-1

-2
DL

10

15

20

25

30

35

40

QUV (kJ/L)
Fig. 3. Photocatalytic inactivation of E. coli under natural sunlight in SUWWE
at different temperatures: 15 C ( ), 25 C ( ), 35 C ( ), and 45 C ( ) without
(a) and with (b) air injection. Kinetic model ttings are also shown (solid line).
DL = 4 CFU/mL.

the spore becomes vulnerable to environmental factors, including

oxidative damage of photocatalysis [29]. For this reason, the photocatalytic disinfection of spores was faster at higher temperatures.
In summary, the enhanced inactivation photocatalytic efciency
observed at higher temperatures can be explained by three reasons
acting simultaneously: (i) an increase in generated OH by TiO2 ;
(ii) and increased activity of the chemical reaction kinetics associated to the metabolic activity; and (iii) the non-lethal inuence of
temperature on E. coli and spores of F. solani metabolisms.
If we compare E. coli (Figs. 3 and 5) and F. solani (Figs. 4 and 6)
results, including the kinetic constants (Table 2) in same water type
at same temperature, it is clear that E. coli is much more sensitive than F. solani spores to photocatalytic treatment, as published
in other articles under different reactor and water conditions [6].
This difference is attributed to the resistant structure and composition of Fusarium spores. Contrary to E. coli cells (vegetative forms),
fungal spore walls are rigid structures composed of polymeric sugars, proteins and glycoproteins. Additionally, spores wall contain
also an outer xylan layer [6]. This multifunctional structure confers
a high resistance against different stress factors. Previous experimental studies on solar photocatalysis with TiO2 also demonstrated
that UV dose required for removal of different types of microorganism depends on cell structure [13,40]. The measured DO was very
close to saturation value at the given temperature, i.e. from 8 mg/L
to 5 mg/L, for 15 to 45 C, respectively. The inactivation results
with injected air showed an enhancement versus the absence of

-3

10

15

20

25

QUV (kJ/L)
Fig. 4. Photocatalytic inactivation of F. solani spores under natural sunlight in
SUWWE at different temperatures: 15 C ( ), 25 C ( ), 35 C ( ), and 45 C ( )
without (a) and with (b) air injection. Kinetic model ttings are also shown (solid
line). DL = 2 CFU/mL.

injected air for both microorganisms (Table 2), although the only
difference in DO measurements was 1 mg/L for all tested temperatures. The role of increased values of DO in the photocatalytic
process is not only to avoid electron/hole pairs recombination in
the photocatalyst but also to increase the levels oxygen that diffuses inside cells increasing the ROS and internal oxidative stress
[41]. Air bubbling and mechanical agitation have been investigated
by other authors for solar water disinfection without photocatalyst,
i.e. SODIS treatment. Kehoe et al. [23] attributed an enhancement
of disinfection efcacy of E. coli after agitation of SODIS bottles to
an increase of dissolved oxygen in water [42]. Reed [43] reported
positive effects of DO concentration on inactivation of E. coli and
E. faecalis. They compared SODIS results in air-equilibrated (oxygenated) water with anaerobic (deoxygenated) water; helium was
injected in SODIS bottles for these experiments. They reported
that inactivation process was blocked under anaerobic conditions
(0.0187 min1 ) while in air-equilibrated samples the reaction was
very fast (0.071 min1 ) [43]. Other authors have demonstrated that
the effect of increased concentrations of oxygen in a carrier gas
from 0% to 0.75% provoked an important reduction in MS2 viability [44]. High levels of ROS are known to be stressing for cells
causing irreversible damage to cellular components like reduction
of DNA stability and changes on proteins and lipids structure and
activity [45].

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

a)

a)
0

F. solani (LogN/N0)

E. coli (LogN/N0)

-1

-2

-3 DL
0

10

15

20

25

30

35

40

45

-1

-2
DL

-3

50

10

QUV (kJ/L)

20

30

40

50

40

50

QUV (kJ/L)

b)

b)
0

F. solani (LogN/N0)

E. coli (LogN/N0)

37

-1

-1

-2

-2

-3

DL

DL

-4

10

15

20

25

30

35

40

45

50

QUV (kJ/L)

-3

10

20

30

QUV (kJ/L)

Fig. 5. Photocatalytic inactivation of E. coli under natural sunlight in UWWE at different temperatures: 15 C ( ), 25 C ( ), 35 C ( ), and 45 C ( ) without (a) and
with (b) air injection. Dark controls are represented with the corresponding empty
symbol. DL = 2 CFU/mL.

Fig. 6. Photocatalytic inactivation of F. solani spores under natural sunlight in

UWWE at different temperatures: 15 C ( ), 25 C ( ), 35 C ( ), and 45 C ( ) without (a) and with (b) air injection. Kinetic model ttings are also shown (solid line).
DL = 2 CFU/mL.

The previous work [5] on F. solani inactivation under similar

operational conditions without air neither thermal control showed
worst results. A much higher QUV (29.9 kJ/L in distilled water and
41.5 kJ/L in well water) was required to achieve the DL, while in
the present contribution with sparged air in SUWWE, the DL was
achieved with 3 kJ/L at 45 C (the fastest result) and less than 20 kJ/L
at 15 C (the worst case). This means a drastic reduction of treatment time from 4 h to 30 min for 60 L of polluted water.
Inactivation rate constants (Table 2) showed that the most
favorable conditions for E. coli (0.117 0.012 min1 ) and F. solani
(0.059 0.001 min1 ) removal in SUWWE were 45 C with air
injection. For UWWE, the best inactivation rate constants for
naturally occurring E. coli was 0.025 0.004 min1 at 35 C, and
0.024 0.002 min1 for F. solani spores at 45 C with air injection.
Results in Table 2, show a well-dened trend to increase inactivation efciency when temperature increases and air sparging is
applied. Also, as explained before, inactivation rate for E. coli was
higher than for F. solani in all cases evaluated. All inactivation rates
were lower in UWWE than in SUWWE. Real efuents content a relatively high amount of chemical compounds which may decrease
the photocatalytic activity due to the action ionic species like sulfates, nitrates, chlorides and phosphates, which partially scavenge
ROS and poison the catalyst competing for the active sites of TiO2
surface [5]. Also a high content on inorganic carbon in the water

samples under study decrease the efciency. Negative effect of

carbonates and bicarbonates on photocatalytic activity through
OH scavenging and photo-absorption is well known [46]. Dissolved Organic Matter naturally present in UWWE may be also
detrimental as it competes with microbial cells for generated OH
radicals. Water turbidity affects also negatively to the photocatalytic process shading catalyst particles and microorganisms by
light scattering phenomena. The average UWWE turbidity was
15.2 0.1 NTU, while in SUWWE was 1.5 0.1 NTU (Table 1).
As expected for photocatalytic disinfection with TiO2 , rst-order
kinetics can be considered the most common behavior for all the
results (Table 2). Only few cases are represented by biphasic model
(model 2, double log-linear) and less frequently by linear behavior
continued with a residual concentration (model 3, log-linear + tail).
As all these equations are based on rst-order kinetics (simple or
modied) the kinetic constants can be directly compared to assess
the best disinfection results. The tail shape can be explained for
the presence of remaining population of bacteria which was not
attacked by the oxidative damage of the process due to a number
of reasons like: (i) the catalyst particle aggregates around bacteria
cell, shades it and prevents the formation the hydroxyl radical in
the cell wall surface; (ii) the oxidative attack of hydroxyl radicals
is more likely to occur at high initial concentration of bacteria (106
and 103 CFU/mL) than when viable bacteria population is very low

38

I. Garca-Fernndez et al. / Catalysis Today 240 (2015) 3038

(10010 CFU/mL); therefore, there is a clear deceleration of the process at the end of the reaction [31]. This tail is more evident when
there is a strong competence between DOM and bacteria for the
hydroxyl radicals, as suggested previously by other authors [29,31].
Only very few experiments (Table 2) can be described with initial
shoulder shape followed by linear. This is attributed to the multitarget action of photocatalytic processes, which requires a number
of oxidative attacks prior to any sign of bactericidal effect (in our
case: in terms of viability loses). When the oxidative conditions
are very good (air injection, and higher temperature) this phase is
very fast, the shoulder does not exist (model 1, 2, 3); when this
phase takes longer, the shoulder appears (model 4). For the case of
a more resistant microorganism and without air sparging, shoulder
appears.
Our results show in general better inactivation efcacies than
data reported in the literature for E. coli using TiO2 slurry in
pilot plant reactors. Agull-Barcel et al. reported a k-value of
0.0135 0.0013 min1 for naturally occurring E. coli in 10 L-CPC
reactor with TiO2 /solar (100 mg/L) in real WW efuents [40]. Van
Grieken et al. showed also for E. coli inactivation by TiO2 (100 mg/L)
slurry in WW efuent 2 kinetic decay constants of 5.6 104 min1
and 16.6 104 min1 ; and in synthetic WW a decay rate of
5.7 104 min1 . These authors used a bench annular reactor with
a lamp (365 nm) to treat 1 L of WW [47]. Cho et al. used a 0.6 m2 -CPC
reactor with TiO2 (100 mg/L) to inactivate E. coli, reporting k values
of 4.88 103 , 1.25 102 , 3.28 103 (min1 ) for three different
days [48]. Rincn and Pulgarn reported a double linear model, with
k1 = 0.483 min1 , k2 = 0.300 min1 for E. coli at 2025 C in a photoreactor of 4 L with UV lamps and 1 g/L of TiO2 [28]. Unlikely, there
are few studies on the photocatalytic Fusarium spores. In our previous work we reported a kinetic rate constant calculated against
(instead of time) QUV of 0.40 0.07 (L/kJ) using 100 mg/L of TiO2
in distilled water, and 0.32 0.02 (L/kJ) [17]. In this work, if we
calculate the kinetic constants of Table 2 against QUV , we obtain
0.792 0.075 L/kJ for SUWWE, and 0.139 0.012 L/kJ for UWWE
with injected air and 45 C. This are very promising results if we
consider the complexity of the WW.
5. Conclusions
Solar photocatalysis with TiO2 suspensions in a solar reactor for
water disinfection can be enhanced in real and simulated WW efuent contaminated with E. coli and F. solani spores by the increasing
of temperature. This has implications in the interpretation of results
from these reactors; when the low temperatures of the water range
from 5 to 15 C and the kinetics are very slow; but sometimes
this is attributed wrongly to other factors. Therefore, the changes
in temperature of photocatalytic reactors for water disinfection
might be considered to evaluate the efciency of the process. The
energy costs of little increase of water temperature are very low
and provide signicant treatment time reductions.
The air sparging causes also an improvement in the bacterial
and spore photocatalytic inactivation, which is specially marked
in F. solani spores. This is in line with the important role of DO in
photocatalytic water disinfection, which should be also considered
for solar reactors design for water disinfection.
Acknowledgments
The authors wish to thank the Spanish Ministry of Economy and
Competitiveness for nancial support under the AQUASUN project
(reference: CTM2011-29143-C03-03). We also acknowledge the
contribution of Elisa Ramos and Agustn Carrin in the experimental work. IGF would like to thank the University of Almera and
CIEMAT-PSA for her PhD research grant.

References

[1] S. Malato, P. Fernndez-Ibnez,

M.I. Maldonado, J. Blanco, W. Gernjak, Catal.
Today 147 (2009) 159.
A.M. Mayes,
[2] M.A. Shannon, P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J. Marins,
Nature 452 (2008) 301310.
[3] S.K. Sharma, R. Sanghi (Eds.), WW Reuse and Management, Springer, New York,
2013, p. 534, http://dx.doi.org/10.1007/978-94-007-4942-9 6.

M. de Cara, J. Tello, Water Res. 43 (7) (2009)

[4] C. Sichel, P. Fernndez-Ibanez,
18411850.

[5] M.I. Polo-Lpez, P. Fernndez-Ibnez,

I. Garca-Fernndez, I. Oller, I. salgadoTrnsito, C. Sichel, J. Chem. Technol. Biotechnol. 85 (2010) 10381048.

[6] I. Garca-Fernndez, M.I. Polo-Lpez, I. Oller, P. Fernndez-Ibnez,

Appl. Catal.
B: Environ. 121122 (2012) 2029.
[7] A.B.A. Teixeira, M.L. Moretti, P. Trabasso, A. von Nowakonski, F.H. Aoki, A.C.
Vigorito, M. Miyaji, K. Nishimura, H. Taguchi, A.Z. Schreiber, Micopathologia
160 (2005) 291296.
[8] K. Sung-Youn, S. Hun-Gu, H. Sang, R. Sangryeol, K. Dong-Hyun, Int. J. Food
Microbiol. 153 (2012) 171175.

[9] J.A. Byrne, P. Fernndez-Ibnez,

P.S.M. Dunlop, D.M.A. Alrousan, J.W.J. Hamilton,
Int. J. Photoenergy (2011) 12p.
[10] D. Gumy, C. Morais, P. Bowen, C. Pulgarin, S. Giraldo, Rw. Hajdu, J. Kiwi, Appl.
Catal. B: Environ. 63 (2006) 7684.
[11] D. Rubio, E. Nebot, J.F. Casanueva, C. Pulgarn, Water Res. 47 (2013) 63676379.
[12] I. Paspaltis, K. Kotta, R. Lagoudaki, N. Grigoriadis, I. Poulios, T. Sklaviadis, J. Gen.
Virol. 87 (2006) 31253130.

[13] C. Sichel, M. de Cara, J. Tello, J. Blanco, P. Fernndez-Ibnez,

Appl. Catal. B:
Environ. 74 (2007) 152160.
[14] A. Rincn, C. Pulgarn, Appl. Catal. B: Environ. 63 (2006) 222231.
[15] O. Seven, B. Dindar, S. Aydemir, D. Metin, M.A. Ozinel, S. Icli, J. Photochem.
Photobiol. A: Chem. 165 (2004) 103107.

[16] C. Sichel, J. Tello, M. de Cara, P. Fernndez-Ibnez,

Catal. Today 129 (2007)
152160.

[17] P. Fernndez-Ibnez,
C. Sichel, M.I. Polo-Lpez, M. de Cara-Garca, J.C. Tello,
Catal. Today 144 (12) (2009) 6268.
[18] G.J. Tortora, B.R. Funke, C.L. Case, Microbiology, An Introduction, 9th ed., Pearson International, San Francisco, 2007.
[19] L. Humpheson, M.R. Adams, W.A. Anderson, M.B. Cole, Appl. Environ. Microbiol.
64 (1998) 459464.
[20] A. Benito, G. Ventoura, M. Casadei, T. Robinson, B. Mackey, Appl. Environ. Microbiol. 65 (1999) 15641569.
[21] M. Berney, H.U. Weilenmann, A. Simonetti, T. Egli, J. Appl. Microbiol. 101 (2006)
828836.
[22] J.A. Imlay, S. Linn, Science 240 (2007) 13021309.
[23] S.C. Kehoe, T.M. Joyce, P. Ibrahim, J.B. Gillespie, R.A. Shahar, K.G. McGuigan,
Water Res. 35 (2001) 10611065.
[24] R.H. Reed, S.K. Mani, V. Meyer, Lett. Appl. Microbiol. 30 (6) (2000) 432436.
[25] A.J. Acher, B.J. Juven, Appl. Environ. Microbiol. 33 (5) (1977) 10191022.
[26] A.J. Acher, G. Ayoub, Wat. Sci. Res. AQUA 46 (13) (1997) 218223.
[27] R.H. Reed, Adv. Appl. Microbiol. 54 (2004) 333365.
[28] A.G. Rincn, C. Pulgarn, Catal. Today 101 (2005) 331344.
[29] M.I. Polo-Lpez, I. Garca-Fernndez, T. Velegraki, A. Katsoni, I. Oller, D. Mantza
Appl. Catal. B: Environ. 111112 (2012) 545554.
vinos, P. Fernndez-Ibnez,

[30] C. Sichel, J. Blanco, S. Malato, P. Fernndez-Ibnez,

J. Photochem. Photobiol. A:
Chem. 189 (2007) 239246.
[31] J. Marugn, R. van Grieken, C. Sordo, C. Cruz, Appl. Catal. B: Environ. 82 (12
(16)) (2008) 2736.

[32] E. Ortega-Gmez, P. Fernndez-Ibnez,

M.M. Ballesteros Martn, M.I. PoloLpez, B. Esteban-Garca, J.A. Snchez Prez, Water Res. 46 (2012) 61546162.
[33] K.D. Omatoyo, S. Elias, A.T. Maya, D.G. Yogi, Appl. Catal. B: Environ. 98 (2010)
2738.
[34] J.M. Herrmann, Topics Catal. 34 (14) (2005) 4965.
[35] Physicochemical
compendium
(Poradnikzykochemiczny),
WidawnictwoNakowo-Techniczne, Warszawa, 1974.
[36] M. Janus, E. Kusiak-Nejman, A. WaldemarMorawski, Reac. Kinet. Mech. Cat 106
(2012) 289295.
[37] H. Truble, E. Ssckmann, J. Am. Ceram. Soc. 94 (13) (1972) 44994510.
[38] A. Favre, V. Chams, A. Caldeira de Arujo, Biochimie 68 (1986) 857864.
[39] S. Marn, V. Sanchs, A. Teixido, R. Saenz, A.J. Ramos, I. Vinas, N. Magan, Can. J.
Microbiol. 42 (1996) 10451050.

[40] M. Agull-Barcel, M.I. Polo-Lpez, F. Lucena, J. Jofre, P. Fernndez-Ibnez,

Appl.
Catal. B: Environ. 136137 (2013) 341350.
[41] T.M. Delvin, Textbook of Biochemistry With Clinical Correlations, 4th ed., John
Wiley, New York, 1997.
[42] S.C. Kehoe, T.M. Joyce, P. Ibrahim, J.B. Gillespie, R.A. Shahar, K.G. McGuigan,
Water Res. 35 (4) (2001) 10611065.
[43] R.H. Reed, Lett. Appl. Microbiol. 24 (1997) 276280.
[44] N.H. Alshraideh, M.Y. Alkawareek, W.G. Gorman, W.G. Graham, B.F. Gilmore, J.
Appl. Microbiol. 115 (2013) 14201426.
[45] J.A. Imlay, Annu. Rev. Biochem. 77 (2008) 755776.
[46] S. Canonica, T. Kohn, M. Mac, F.J. Real, J. Wirz, U. von Guten, Environ. Sci. Technol.
39 (2005) 91829188.
[47] R. van Grieken, J. Marugn, C. Sordo, C. Pablos, Catal. Today 144 (2009) 4855.
[48] I.-H. Cho, I.-Y. Moon, M.-H. Chung, H.-K. Lee, K.-D. Zoh, Water Science and
Technology: Wat. Supply 2 (1) (2002) 181190.