Chemical Engineering Journal xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Bacterial inactivation and degradation of organic molecules by titanium

dioxide supported on porous stainless steel photocatalytic membranes
C. Adán, J. Marugán ⇑, S. Mesones, C. Casado, R. van Grieken
Department of Chemical and Environmental Technology (ESCET), Universidad Rey Juan Carlos, C/ Tulipán s/n, 28933 Móstoles (Madrid), Spain

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Photocatalysis/microfiltration hybrid
system with TiO2–PSS membranes.
 Average pore size of membrane plays
a significant role on photocatalytic
efficiency.
 Optimal TiO2 loading is a compromise
of light absorption, pressure and
surface area.
 Photocatalytic membranes show a
significant inactivation of the
retained bacteria.
 The feasibility of the hybrid process
for biofouling control has been
confirmed.

a r t i c l e i n f o a b s t r a c t

Article history: A novel integrated photocatalysis/microfiltration hybrid system based on the use of TiO2 immobilized by
Available online xxxx filtration on porous stainless steel supports has been demonstrated to provide a good efficiency in the
removal of bacteria for water disinfection applications. A synergistic effect derived of the physical
Keywords: removal of microorganisms coupled to its photocatalytic inactivation was proved to be achieved. The
Microfiltration membrane pore size and the TiO2 loading were found to play a significant role on the efficiency of the
Photocatalysis process and the value of the transmembrane pressure. An optimal value of 8.5 g
m 2 TiO2 supported over
Stainless-steel
the membranes provides good photocatalytic activity keeping the transmembrane pressures in the range
TiO2
Thin film
of 1.5–2 bar. Moreover the photocatalytic activity of the membrane was proved to have as an additional
Escherichia coli advantage the possibility of the simultaneous oxidation of the chemical pollutants present in the feed
water. Continuous operation of the photocatalytic membrane reactor in dead-end configuration confirms
the accumulation of bacteria on the membrane leading to an increase in the transmembrane pressure,
whereas the photocatalytic functionality demonstrated a significant inactivation of the retained cells,
confirming the feasibility of the hybrid process for biofouling control.
Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction advantages: small footprint, small size (compact modules), lower

energy consumption, admission of fluctuations in feed quality
Membranes play an important role in the separation technolo- and its excellent separation efficiencies. This technology provides
gies and represent promising alternatives to conventional water many advantages over conventional treatments, as demonstrated
and wastewater treatment processes [1–3], with well-known by the rapidly increase in their industrial use in potable water pro-
duction in the past decade [4,5]. In general, filtration uses the dif-
ference of pressure between the feed and permeate side as the
⇑ Corresponding author. Tel.: +34 91 664 7466; fax: +34 91 488 7068. driving force to transport the solvent through the membrane. In
E-mail address: javier.marugan@urjc.es (J. Marugán). this way, the particles and dissolved components are retained

http://dx.doi.org/10.1016/j.cej.2016.04.091
1385-8947/Ó 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
2 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx

based on the membrane properties such as pore size, shape, and Despite the extensive work on photocatalytic membranes for
surface charge [6]. The most general and known classification of removal of organic contaminants [10,22] and anti-bacterial and
membranes is based on the pore size. This classification distin- anti-fouling effects [21,23,24], to the best of our knowledge there
guishes microfiltration (50–500 nm), ultrafiltration (2–50 nm), are no reports on the application of photocatalytic membranes
nanofiltration (62 nm), and reverse osmosis (0.3–0.6 nm) [6]. based on TiO2 supported on porous stainless steel (PSS) mem-
Depending on the application and the required pore size the base branes for the disinfection of water and simultaneous removal or
materials of the membranes are different [1]. Polymeric mem- organic pollutants. PSS membranes present significant advantages
branes (i.e., polysulfone (PS), polyethersulfone (PES), polypropy- due to their mechanical stability, chemical resistance and possibil-
lene (PP), polyethylene (PE), polytetrafluoroethylene (PTFE), ity of fine tuning of their properties during the preparation step.
polyvinyl chloride (PVC), cellulose acetate (CA) and Nafion mem- Ames et al. [25] investigated the characterization of stainless steel
branes are commonly used for ultrafiltration, nanofiltration and membranes for aqueous waste separation processes. Li et al. [26]
reverse osmosis. For microfiltration applications organic, metallic reported the preparation of TiO2 films incorporated onto 316L
or ceramic membranes are usually employed [7–9], being the cera- PSS pieces of 1 lm pore size by dip coating to study the synthesis
mic materials supports more frequently used than polymers or conditions of the films, whereas Potter et al. [17] studied tubular
metals due to their higher resistance to oxidation [8–10]. TiO2–PSS membranes with a nominal pore size of 0.2 lm for the
The microfiltration process is commonly used to remove big filtration of anionic dyes in tap and deionized water solutions.
particles and also works as a disinfection barrier to assist in the However, there is a lack of investigations on the photocatalytic
removal of pathogenic microorganisms [1,6,11] that may cause applications of metallic membranes. The novelty of this work is
problems in further treatment steps, moreover, microfiltration is focused in the further investigation of the PSS membranes and
also a common pretreatment step of other membrane processes. its behavior in a hybrid photocatalytic/microfiltration TiO2–PSS
Nevertheless, this technology has the main problem of the mem- membrane system. The preparation and photocatalytic perfor-
brane fouling (deposition of retained particles at the membrane mance of TiO2–PSS microfiltration membranes for water disinfec-
surface or inside the pores) which causes several adverse effects tion applications, including not only the removal and inactivation
on membrane systems as gradual reduction in the permeate flux, of microorganisms, but also the possibility of simultaneous oxida-
increase in transmembrane pressure (TMP), biodegradation of tion of chemicals is described in this work. The influence of param-
membrane materials and poor permeability [2,12], which can lead eters such as TiO2 loading and transmembrane pressure on the
as well to operational problems as rising energy demand, increase process efficiency is also addressed, together with the analysis of
of the cost in chemical cleaning agents and finally system failure the process both in batch and continuous operation mode using
[1,13]. To control biofouling the most common method is periodic deionized water (DW) and a simulated wastewater treatment
cleaning of the membranes by acid or basic chemical solutions and plant effluent (SWTPE).
backflushing with water that raises the cost of the process [12].
The combination of filtration processes with Advanced Oxida-
2. Materials and methods
tion Technologies (AOTs) has received a growing interest in the last
years [14]. Among the different AOTs, titanium dioxide photocatal-
2.1. Microfiltration system and photocatalytic membrane reactor
ysis constitutes one of the most promising and studied processes
for the generation of reactive oxidant species upon illumination
Microfiltration and photocatalytic experiments were carried out
of a semiconductor material. The high efficiency of TiO2 to generate
in the experimental setup schematized in Fig. 1. The system consist
hydroxyl radicals when it is irradiated with UV light, together with
of a 4 L feed tank, a positive displacement pump couple to a pulsa-
its environmentally friendly properties and low cost, make this
tion damper and two pressure and flow gauges located before and
material the most popular and studied photocatalyst. Therefore,
after the photocatalytic membrane reactor. The liquid flow rate
the incorporation of TiO2 to the membranes enables the coupling
provided by the pump can be regulated by a variable-frequency
of both processes by the simultaneous action of pollutant rejection
drive and a piston displacement system. By default, experiments
and photocatalytic degradation of contaminants.
were carried out using a flow rate of 1.1 L
min 1. The photocat-
Titania has been previously reported to have antibacterial and
alytic membrane reactor consists of an annular reactor of 15 cm
antiadhesion properties for microorganisms due to the photogen-
long, with a 3 cm in diameter glass inner tube and 6 cm in diame-
erated reactive oxidant species, as hydroxyl radicals and superox-
ter poly-methylmethacrylate outer tube, being the membrane
ide that can completely degrade most of organic pollutants,
(5 cm in outer diameter) located concentrically between both walls
bacteria, fungus and even viruses from water [15,16]. The possibil-
(Fig. 1b and c). Illumination of the TiO2 layer in the inner surface of
ity of reducing membrane fouling by the incorporation of a photo-
the membrane is provided by a black light lamp (Philips TL 6 W
catalyst makes extensible the useful lifetime of the membranes.
with an emission maximum at 365 nm) located in the axis of the
These hybrid photocatalysis/membrane processes are conducted
annular reactor.
in photocatalytic membrane reactors (PMR) where the catalyst
Both concentrate and permeate outlet streams of the reactor,
can be immobilized in a membrane (fixed bed reactors) [14,17]
can be driven to an external storage tank (in continuous operation
or suspended in a feed tank (slurry reactors) and the membrane
mode) or recycle to the feed tank (in batch operation mode). A
might act as a barrier not only for the degraded by-products but
valve in the outlet of the concentrate stream allows operation of
also for the catalysts particles [18–20]. In the case of fixed bed
the system under cross flow or dead end filtration modes.
reactors the surface of the membrane constitutes the catalyst sup-
port and must be illuminated with UV light. The TiO2 nanoparticles
can be either deposited over the membrane surface or trapped in 2.2. Preparation of photocatalytic membranes
the polymeric membrane [7,10,21], being the first option more
commonly used in microfiltration membranes. The photocatalytic The 316 L porous stainless steel (PSS) membranes were sup-
process occurs in the TiO2 layer providing the features of biofouling plied by Shijiazhuang Beot Inorganic Membrane Separation Equip-
control and disinfection of the permeate stream. This PMR config- ment Co. Ltd with nominal average pore sizes of 0.20 and 0.50 lm
uration, however, reduces the photocatalytic active surface areas and will be reference as PSS-0.2 and PSS-0.5, respectively. They
as a consequence of catalysts immobilization which usually results were fabricated by synterization of stainless steel microparticles
in a loss of activity [2]. conforming a final tubular shape geometry (15 cm length, 5 cm

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx 3

Fig. 1. (a) Schematic representation of the experimental setup, (b) detailed lateral view of the photocatalytic membrane reactor and (c) cross-sectional view of the
photocatalytic membrane reactor.

diameter, 0.25 cm thickness), corresponding to an apparent geo- during approximately 12 h to remove the particles trapped inside
metrical internal surface of 235.6 cm2. For the immobilization of the pores. The cleaning water has to be changed several times dur-
TiO2, the metallic membranes were firstly sonicated in ethanol ing this process until reaching particle free water with negligible
for 30 min to clean the surface from impurities and sonicated again turbidity. The preparation and cleaning process has been repeat-
twice in distilled water. The membranes were then placed into the edly validated and proved to be reproducible.
photocatalytic reactor described above. Evonik AeroxideÒ TiO2-P25
was used as photocatalytic material to prepare the suspensions.
The selected amount of TiO2-P25 powder was added to the tank 2.3. Membrane characterization
and vigorous stirred during 5 min. Afterwards, the suspension
was pumped to the reactor in a dead-end operation mode, forcing Surface and cross section morphologies of the PSS membranes
the flow to cross the membrane. The filtration ends when the rejec- were analyzed by an Environmental Scanning Electron Microscope
tion of the TiO2 photocatalyst particles reaches values above 99.9% Philips XL 30 ESEM provided with an EDX detector. Membranes
and the turbidity of the water in the tank is falls below 2.0 NTU, were previously mounted using an epoxy resin, and then sectioned
corresponding to a negligible TiO2 concentration. The experiments with a diamond saw. The specimens were ground to a 1200 grit fin-
were done with different TiO2 aqueous suspensions with a particle ish and polished with a 1 lm diamond paste.
concentration that vary from 0 to 0.325 g
L 1 to obtain catalyst The porosity studies were completed by mercury intrusion
loadings in the range of 0–0.8 g per membrane, corresponding to porosimetry (MIP) using CE Instruments Pascal 140/240 and apply-
a surface loading of TiO2 of 0–34.0 g
m 2. The deposited TiO2 layer ing the Washburn equation. It is assumed a cylindrical non-
was proved to be stable under all tested experimental conditions intersecting pore model, with the values recommended by the
and remains firmly attached to the membrane wall surface after IUPAC of 1418 and 484 mN m 1, for the contact angle and surface
the experiments. If required, membranes can be cleaned by brush- tension of mercury, respectively. This provides data corresponding
ing the surface TiO2 layer and applying an ultrasound treatment to pores from about 300 mm down to 7.5 nm diameter.

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
4 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx

Fig. 2. Scanning Electron Microscopy images of the PSS membranes: (a) cross section of PSS-0.5, (b) surface of PSS-0.5, (c) cross section of PSS-0.2, and (d) surface of PSS-0.2.

(a) PSS-0.2 (b) PSS-0.2

PSS-0.5 PSS-0.5
0.25 0.05
Cumulative Intrusion Volumen (mL g )
-1

0.20 0.04
Log. Differential Intrusion

0.15 0.03

0.10 0.02

0.05 0.01

0.00 0.00
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16

Pore size diameter (µm) Pore size diameter (µm)

Fig. 3. (a) Pore size diameter and, (b) cumulative pore volume of PSS membranes determined by mercury intrusion porosimetry.

2.4. Photocatalytic degradation of methanol to stabilize the emission. In the meantime, 4 L of pure water was
recirculated to the membrane to get a stationary flow. Then,
After incorporation of TiO2 to the membrane, the photocatalytic methanol was added to the reservoir tank to get an initial concen-
activity of the titania membrane was evaluated on the oxidation of tration of 100 mM. The pump was then turned on and the experi-
methanol to formaldehyde in a batch operation mode. Before start- ment starts taking samples from the tank at different irradiation
ing the experiment, the UVA lamp was switched on during 15 min times. The quantification of formaldehyde on the samples collected

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx 5

6
100 1200 16
Rejection 15 PSS-0.5
PSS-0.2 14 5
PSS-0.2
PSS-0.5 1000
80 13

Transmembrane pressure (bar)

Reaction rate, r x 10 (M s )
-1
12
4
800 11

8
TiO2 rejection (%)

Turbidity (NTU)
10
60 3
9
600 8
7 2
40
6
400 5
Tank turbidity
1 PSS-0.5
4
20 3 PSS-0.2
Permeate turbidity 200
2 0
0.005 0.010 0.015 0.020 0.005 0.010 0.015 0.020
-1 -2 -1 -2
0 0 Permeate flux (L·min ·cm ) Permeate flux (L·min ·cm )
0 5 10 15 20 25 30
Fig. 6. Dependence on the permeate flow rate of the reaction rate (a) and the
Time (min)
transmembrane pressure (b) for membranes loaded with 0.2 g of TiO2.
Fig. 4. Rejection percentage of 0.25 g
L 1 of TiO2 particles during membrane
incorporation in the 0.5 lm and 0.2 lm membrane.
water (DW) and simulated wastewater [28] diluted to a total
organic carbon (TOC) value of 15 mg
L 1 to simulate a secondary
wastewater treatment plant effluent (SWTPE). This water matrix
(a) (b) avoids the interference of the osmotic effects that usually appears
50
Pressure detection limit
PSS-0.5
-1
in disinfection experiments with pure water. The photoactivity
PSS-0.2 0.2 g·L
6 runs were preformed following the same procedure than for
Suspension
methanol oxidation experiments. Samples were taken at time
40
5 intervals to determine the concentration of E. coli both in the per-
Transmembrane pressure (bar)

Reaction rate, r x 10 (M/s)

-1
0.15 g·L meate stream and in the reservoir tank. Quantification of viable
bacteria in samples was carried out following a standard serial
8

4 30
dilution procedure and plating in agar dishes with Luria–Bertani
growth medium for counting after 24 h incubation at 37 °C. More
-1
0.1 g·L
3
20 details about the analytical procedure can be found elsewhere [16].
0.05 g·L
-1 Photocatalytic microfiltration of bacteria was carried out using
2
two operational configurations: (i) experiments in batch mode,
10 performed in a similar way to methanol oxidation experiments
1 recirculating 4 L of the bacterial suspension; (ii) experiments in
PSS-0.5
PSS-0.2 -1
continuous operation, conducted using 40 L of the bacterial sus-
0.013 g·L
0 0 pension in a one-pass continuous flow. All experiments were done
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
in dead end filtration mode at a constant flow rate of permeate.
-2 -2
TiO2 loading (g·m ) TiO2 loading (g·m )

Fig. 5. Effect of TiO2 loading on: (a) transmembrane pressure as a function of TiO2
loading (errors bars correspond to the standard deviation of TMP data along
3. Results and discussion
methanol oxidation experiments); (b) reaction rates (hyperbolic fitting in dashed
lines). 3.1. Characterization of the photocatalytic membranes

SEM micrographs of the PSS membranes are showed in Fig. 2. It

at time intervals from the tank was carried out through a colori- can be clearly observed the porosity and roughness derived of the
metric method [27] in a spectrophotometer Biochrom, model Libra synterization fabrication procedure. The different average pore
S22. All experiments were carried out in deionized water (MilliQÒ, sizes depend on average size of the particles used as row material
18.2 MX
cm) at natural pH with air being bubbled in the reservoir in the synterization process, being slightly smaller in the PSS-0.2 in
tank to keep saturation concentration of oxygen in water. comparison with the PSS-0.5 membrane (Fig. 1b and d). As a con-
sequence the pores seem smaller in the PSS-0.2 (Fig. 1c) than in the
2.5. Photocatalytic inactivation of Escherichia coli PSS-0.5 membrane (Fig. 1a). In all cases, EDX analysis provides an
average composition in agreement with the 316L stainless steel
Lyophilized Escherichia coli K12 strains were selected as raw material of 67.8% Fe, 18.9% Cr, 11.8% Ni, 1.1% Mo and 0.4% Mn.
microorganism model for the microfiltration and photocatalytic Fig. 3 displays total pore volume and pore size distributions
bacterial inactivation experiments. E. coli was provided by the obtained from mercury intrusion measurements analysis of the
Colección Española de Cultivos Tipo (CECT 4624, corresponding PSS membranes. Both membranes present a wide range pore size
to ATCC 23631). Firstly, a bacterial suspensions with a concentra- distribution of macropores with a maximum at about 6.0 and
tion about 109 CFU
mL 1 was prepared by inoculation of a liquid 8.0 lm for the PSS-0.2 and the PSS-0.5 membrane, respectively,
nutrient medium (Miller’s LB Broth, Scharlab) and incubation on and similar porosity around 27%. This large pore size distribution
a rotary shaker under stirring during 24 h at 37 °C. The initial bac- is not consistent with the manufacturer’s specifications in that
terial concentration for the experiments of 106 CFU
mL 1 was the range is significantly larger than that indicated by Shijiazhuang
adjusted by dilution. Experiments were carried out using distilled Beot Inorganic Membrane Separation Equipment Co. Ltd. However,

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
6 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx

0 0
10 10 SWTPE (membrane)
DW (membrane)

-1 -1
10 10

-2 -2
10 SWTPE (no membrane) 10
Viable E. coli CFU, C/Co

Viable E. coli CFU, C/Co

SWTPE (membrane)
DW (No membrane)
-3 DW (membrane) -3
10 Theoretical curve 10

-4 -4
10 10

-5 -5
10 10

a) b)
-6 -6
10 Bacterial detection limit 10 Bacterial detection limit

0 20 40 60 80 100 0 20 40 60 80 100
Time (min) Time (min)

Fig. 7. Microfiltration experiments with a PSS-0.5 membrane and E. coli in an initial concentration of 106 CFU
mL 1 for; (a) tank and, (b) permeate streams in DW and SWTPE.
Dashed line represents the theoretical evolution of total E. coli retention calculated for a volumetric flow of 1.1 L
min 1 and, (c) SEM image of an E. coli bacterium.

other works has found similar results when analyzing stain-less photocatalytic degradation of methanol was negligible. Fig. 5b also
steel membranes [25]. shows reference experiments with TiO2 suspensions (without
Incorporation of titanium dioxide to the membrane by dead end membrane) for comparison purposes. The reaction rate observed
filtration of TiO2 suspensions under batch operation showed in with slurry TiO2 increase linearly with the catalyst loading, not
both cases a negligible turbidity in the permeate from the very showing limitation by radiation absorption in the studied range
beginning. Consequently, the TiO2 photocatalyst can be completely of titania concentration. As expected, the suspensions reached
removed from the liquid and deposited in the inner wall of the much higher activities in comparison with the TiO2–PSS systems.
membrane by a simple one step filtration method. The average However, the immobilized systems have the advantage of having
agglomerate size of P25-TiO2 in water at natural pH is 0.7 lm all the catalyst irradiated in the reactor volume and the photocat-
[29], what means that, despite the presence of larger pores, the alyst can be used more efficiently and for subsequent times [31],
membrane acts as a multilayer lattice retaining all the TiO2 parti- avoiding the separation step of the TiO2 particles and facilitating
cles that are unable to follow the tortuous channels of the porous the continuous operation of the process.
structure. Fig. 4 displays the evolution versus time of the rejection The reaction rates of the TiO2–PSS membranes grow as the
of 0.25 g
L 1 of TiO2 particles for both PSS membranes followed by addition of titanium dioxide increases improving methanol oxida-
turbidity measurements. The initial turbidity of the TiO2 suspen- tion. However, up to a value of around 0.2–0.3 g of TiO2 (8.5–
sion in the tank was of 1150 NTU (corresponding to the initial 12.7 g m 2), photocatalytic activity reaches a plateau, indicating a
TiO2 concentration of 0.25 g
L 1), and the turbidity of permeate gradual saturation of photonic absorption for a given incident light
was in the range of 0.22–0.45 NTU throughout the experiment. flux, in agreement with other authors [10,14,19] that attributed
For the flow rate of 1.1 L
min 1 used in the experiment, the total this phenomenon to the sum of some factors as the increase of
filtration of the TiO2 particles in the suspension was achieved in internal mass transfer resistance due to inter-particle agglomera-
less than 15 min. tion and the light shielding effect as TiO2 loading increases [10].
These data can be successfully described by a hyperbolic saturation
3.2. Photocatalytic activity of PSS–TiO2 membranes for methanol curve similar to the Langmuir–Hinshelwood kinetic rate law
oxidation (dashed lines). Although similar tendencies are shown by both
membranes, it is noticeable that slightly higher reaction rates are
As shown in the previous section, the incorporation of a TiO2 achieved with the PSS-0.5 membrane. The differences in reaction
layer to the membrane can be easily accomplish by filtration of rates can be attributed to differences observed in the TMP achieved
suspensions of TiO2 nanoparticles with the desired loading. As in both PSS membranes. Peng et al. [33] justified the decrease in
expected, increasing amounts of titanium dioxide incorporated to the photocatalytic reaction rate for increasing TMP considering
the membrane lead to increasing transmembrane pressure (TMP) that higher TMP reduces the cross-flow velocity and the mass
values (Fig. 5a). An apparent linear dependence was observed for transfer rate in the liquid–solid system was decreased accordingly
both membranes, being the slope of the PSS-0.2 membrane slightly [33]. However, in the present work a constant flow rate has been
higher than that of the PSS-0.5. used in all the experiments. Consequently, the decrease in the
The efficiency of both membranes in the photocatalytic oxida- activity produced by higher TMP values could be reasonably attrib-
tion of methanol was examined in dead-end filtration mode with uted to the decrease in the available surface area caused by the
a constant flow rate of 1.1 L
min 1 (Fig. 5b). The photooxidation densification of the TiO2 layer due to agglomeration of the particles
rate (ro) was evaluated from the slope of the curves of the and reduction of the interparticle porosity.
formaldehyde concentration vs. irradiation time following a well- To study this phenomenon, experiments of methanol photoox-
established procedure [30–32]. The blank experiment (reaction idation with different fluxes and pressures were carried out for the
rate of the PSS membrane without TiO2) reveals that non- previously estimated optimal titania loading. Fig. 6a shows the

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx 7

10
0 (a) 10
0 (b)

-1 -1
10 10

Viable E. coli CFU, C/Co

Viable E. coli CFU, C/Co

-2 -2
10 10

0 g TiO2
-3 -3
10 0.05 g TiO2 10
0.1 g TiO2
-4 0.2 g TiO2 -4
10 10
0.4 g TiO2
0.6 g TiO2
-5 -5
10 0.8 g TiO2 10
Theoretical curve
-6 -6
10 Bacterial detection limit 10 Bacterial detection limit
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Irradiation time (min) Irradiation time (min)

10
0
(c) 10
0 (d)

-1 -1
10 10
Viable E. coli CFU, C/Co

Viable E. coli CFU, C/Co

-2 -2
10 10

-3 -3
10 10
0 g TiO2
-4
0.05 g TiO2 -4
10 10
0.1 g TiO2
0.2 g TiO2
-5 -5
10 0.3 g TiO2 10
Theoretical curve
-6 -6
10 Bacterial detection limit
10 Bacterial detection limit
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Irradiation time (min) Irradiation time (min)

Fig. 8. E. coli evolution (a) in tank and (b) in permeate stream for PSS-0.5 membrane and E. coli evolution (c) in tank and (d) in permeate stream for PSS-0.2 membrane in
SWTPE.

dependence of the reaction rate on the permeate flux for both 3.3. Microfiltration and inactivation of E. coli in batch recirculation
membranes, whereas Fig. 6b shows the expected increase in the operation
transmembrane pressure. The results show a significant effect in
the photocatalytic activity in agreement with the reports of other Fig. 7 show the results of experiments of E. coli removal by
authors [10,33], but interestingly, a maximum value of the reaction microfiltration using the PSS-0.5 membrane (without TiO2) in com-
rate is achieved for a relatively low value of permeate flux whereas parison with reference experiments without membrane, both in
a dramatic decrease is observed for higher values. Considering that deionized water (DW) and a simulated wastewater treatment
the system is operating under batch recirculation regime, differ- plant effluent (SWTPE). Without the membrane placed in the reac-
ences in the residence time per pass can be discarded as a suitable tor, the bacteria concentration remains constant with the irradia-
explanation of this behavior. In the lower range of fluxes, an tion time using both DW and SWTPE water, suggesting that the
increase in the permeate flow rate leads to a higher reaction rate mechano-osmotic stress induced by the pump recirculation and
probably due to the improvement in the mass transfer. However, the photolytic effect of UVA irradiation are negligible on the bacte-
after a certain value of the permeate flux, the increase in the trans- ria viability. On the contrary, when PSS membrane is placed on the
membrane pressure induce the compression of the TiO2 layer, system, a decrease in the concentration of viable bacteria in the
reducing the porosity and the available surface area and thus the tank of one order of magnitude in SWTPE and five orders in DW
photocatalytic activity. This effect is more pronounced in the after 60 min of irradiation is observed. In the case of DW, the evo-
PSS-0.2 membrane due to its lower porosity, showing the PSS-0.5 lution of viable bacteria is close to the theoretical curve calculated
based membrane higher reaction rates in all the studied range of assuming a full rejection of the cells in the membrane (correspond-
fluxes. ing to the expression C/C0 = exp( Q
t/V) = exp( 0.275
t) as

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
8 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx

Pressure limit detection

(a) 100 (b)
6
90

5 80
Transmembrane pressure (bar)
70

E. coli rejection (%)

4
60

50
3
40

2 30

20
1
PSS-0.2 10 PSS-0.2
PSS-0.5 PSS-0.5
0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
-2 -2
TiO2 loading (g·m ) TiO2 loading (g·m )

Fig. 9. Dependence on the TiO2 loading of the TMP (a) and E. coli rejection (b) for experiments with an initial concentration of bacteria of 106 CFU
mL 1
in SWTPE.

0 0
10 10
(a) (b)
Viable E. coli CFU, C/Co
Viable E. coli CFU, C/Co

-1 -1
10 10

-2 -2
10 10

Filtration Filtration
Photocatalysis +Filtration Photocatalysis +Filtration
-3 -3
10 10
5 10 15 20 25 30 5 10 15 20 25 30
Irradiation time (min) Irradiation time (min)

Fig. 10. E .coli evolution in permeate stream for (a) PSS-0.2 membrane and (b) PSS-0.5 membrane in SWTPE.

derived from the mass balance of the recirculation reactor consid- stream a viability only slightly lower than that in the tank. This
ering a flow rate of Q = 1.1 L
min 1 and a total volume of V = 4 L). In indicates that the decrease in the concentration of viable bacteria
addition to this, the concentration of viable bacteria in the perme- observed in DW is probably not due to the rejection by the mem-
ate stream of the DW experiment is very low (C/C0 in the order of brane but to the lethality of the membrane crossing under
10 4 to 10 6) (Fig. 7b) suggesting that a full rejection of the mechano-osmotic stress.
microorganism by the membrane is actually taken place. However, Fig. 8 shows the evolution of bacterial concentrations in the
the analysis of the concentration of viable bacteria in the SWTPE tank and permeate stream in experiments using TiO2–PSS mem-
experiments leads to relatively surprising results, as the removal branes with increasing titania loading. All experiments were car-
of E. coli cells is much lower than the expected from the microfil- ried out in SWTPE to avoid the interference of mechano-osmotic
tration process. Moreover, significant concentrations of viable bac- inactivation. Bacterial rejection in the range of 90–99% is con-
teria were detected on the permeate stream, being these results firmed by the 1–2 orders of magnitude of viable bacteria reduction
verified repeatedly under experimental conditions in which a full in the permeate stream in comparison with the tank. Incorporation
rejection of TiO2 particles is achieved. Consequently, despite the of TiO2 clearly improves the level of rejection, showing lower con-
bigger size of the E. coli cells (2.7  0.76 lm in comparison with centrations of bacteria as the titania loading increases up to certain
the average 0.7 lm size of the TiO2 agglomerates) these experi- values were results seem to be no longer dependent on the TiO2
ments suggest that bacteria shrink under pressure crossing the incorporated to the PSS. This effect can be reasonably explain by
membrane relatively easily (bacterial presence has been even the reduction of the effective pore size of the membrane upon
detected after ultrafiltration membranes [34]). When SWTPE is incorporation of the TiO2 layer both for PSS-0.2 and PSS-0.5
used, the bacteria are almost unaffected, showing in the permeate supports, but also for the increase in lethality of the membrane

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx 9

50 (a) 50 (b)
Batch Batch
Continuous Continuous

40 40

ht
lig
VA
U
o
N

Viable bacteria %
30
Viable bacteria %
30

ht
lig
ht

VA
lig

U
VA

ht

o
20

U
20

lig

ht
o

lig
VA
N

VA
U
o
ht

U
N
lig
VA
U
10 10
ht
lig
VA

ht
lig
U

VA
U
0 0

Fig. 11. Comparison of the recovered viable microorganisms retained in the membranes after 30 min of experiments for (a) PSS-0.5 membrane and (b) PSS-0.2 membrane in
SWTPE.

crossing under higher TMP conditions. On the other hand, Choi bacteria in permeate in comparison with the pure microfiltration
et al. [22] showed that the photocatalytic activity of a nonporous system in the dark. This improvement is more considerable in
TiO2 film with structure was much lower than that of porous the case of the TiO2–PSS-0.5 membrane, although in both cases
TiO2 film structures, which they attributed to the difference in sur- the efficiency decreases with the operation time and the photocat-
face area. alytic improvement tends to be minimized. It has to be notice that
Fig. 9 shows how the incorporation of TiO2 to the membrane the system is not working in steady state as dead-end filtration
increase the rejection, calculated as the ratio between viable bacte- imply the less favorable configuration in which the formation of
ria in permeate and feed streams. Fig. 9 also shows the increase in a cake in the filtering support takes place. This fact is confirmed
the TMP derived of the TiO2 loading. This higher pressure is also by the unceasing increase of the TMP value along the operation
the driving force for pushing the bacterial cells through the porous time due to the accumulation of bacterial cells. However these
structure, although under more stressful and therefore lethal con- results confirm the synergy achieved in the coupled system, sug-
ditions. However, it must be noticed that the presence of the gesting the feasibility of a really continuous production of disin-
microorganism also affect the value of the TMP pressure in com- fected permeate upon optimization of a cross-flow regime
parison with the values registered in methanol oxidation experi- operation.
ments due to the formation of an additional resistance of the
bacterial cells biofouling layer deposited on the membrane surface.
In this case, the pore size of the PSS support has a remarkable effect 3.5. Evaluation of viable E. coli in membranes
showing the PSS-0.2 membrane much higher TMP and rejection
values in comparison with the PSS-0.5.The rejection of bacteria The viability of the E. coli cells retained in the membranes have
by the uncoated membranes was around 40% for the PSS-0.5 and been quantified after 30 min of batch and continuous operation to
more than 80% for the PSS-0.2. After a 0.05 g of TiO2 loading the investigate whether the photocatalytic reaction could be effective
rejection increases to 75% and almost 100% for the PSS-0.5 and in preventing long-term membrane fouling. Fig. 11 shows the per-
the PSS-0.2, respectively. For higher TiO2 loadings, the rejection centages of the bacteria that are recovered in a viable state from
seems to slightly decrease, keeping values in the order of 90– the membrane over the total number of bacteria removed from
100% for both membranes. Consequently, despite the apparent the water. Results confirm that the amount of bacteria really inac-
higher photocatalytic activity showed by TiO2–PSS-0.5 membranes tivated is significantly higher when the system is under illumina-
for methanol oxidation, bacterial removal seems to be essentially tion, confirming the photocatalytic disinfection effect in
governed by the microfiltration rejection, enhanced in the case of comparison with the exclusively physical removal of still viable
the TiO2–PSS-0.2. cells achieved by filtration. This effect is understandably more pro-
nounced in batch recirculation operation where the amount of
viable bacteria is limited to the initial loading of the tank, in com-
3.4. Microfiltration and inactivation of E. coli in continuous operation parison with the constant supply of high concentrated suspensions
of E. coli cells provided in the continuous mode.
Inactivation of E. coli under continuous regime was studied with Previous reports also found an antibiofouling effect with TiO2/
both PSS membranes loaded with 8.50 g
m 2 of TiO2 and a con- Al2O3 composite membranes [22]. TiO2 membranes irradiated by
stant concentration of 106 CFU
mL 1 in the feed stream. The exper- UV exhibited less flux decline and no significant fouling over time
iments have been carried out both in the dark and under in comparison to the control experiment. On the other hand it has
illumination to elucidate independently the role of microfiltration been reported that the antifouling effect can be explained by the
on the combined microfiltration/photocatalytic process. Results production of reactive species by the hybrid membrane that kill
showed in Fig. 10 confirm that the photocatalytic functionality of bacterial cells and prevent bacterial attachment to membrane sur-
the membrane reduces significantly the concentration of viable face, which reduce the membrane biofouling [15,16,35]. In any

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091
10 C. Adán et al. / Chemical Engineering Journal xxx (2016) xxx–xxx

case, no evaluation of the viability of the bacteria removed by the [8] S. Leong, A. Razmjou, K. Wang, K. Hapgood, X. Zhang, H. Wang, TiO2 based
photocatalytic membranes: a review, J. Membr. Sci. 472 (2014) 167–184.
membranes was reported, confirming the relevance of the data
[9] S. Mozia, K. Szymanski, B. Michalkiewicz, B. Tryba, M. Toyoda, A.W. Morawski,
provided in this study about the mechanism of water disinfection Effect of process parameters on fouling and stability of MF/UF TiO2 membranes
with PSS–TiO2 membranes. in a photocatalytic membrane reactor, Sep. Purif. Technol. 142 (2015) 137–
148.
[10] W.Y. Wang, A. Irawan, Y. Ku, Photocatalytic degradation of acid red 4 using a
4. Conclusions titanium dioxide membrane supported on a porous ceramic tube, Water Res.
42 (2008) 4725–4732.
[11] B. Guo, E.V. Pasco, I. Xagoraraki, V.V. Tarabara, Virus removal and inactivation
The efficiency hybrid photocatalytic/microfiltration TiO2–PSS in a hybrid microfiltration–UV process with a photocatalytic membrane, Sep.
membrane systems have been evaluated both for the oxidation Purif. Technol. 149 (2015) 245–254.
[12] H. Choi, H.-S. Kim, I.-T. Yeom, D.D. Dionysiou, Pilot plant study of an
of methanol and the inactivation of E. coli bacteria. Optimal TiO2 ultrafiltration membrane system for drinking water treatment operated in
loading value of 8.5 g
m 2 of TiO2 supported on the membrane the feed-and-bleed mode, Desalination 172 (2005) 281–291.
inner surface provides a maximum photocatalytic activity keeping [13] A. Drews, Membrane fouling in membrane bioreactors-characterisation,
contradictions, cause and cures, J. Membr. Sci. 363 (2010) 1–28.
the transmembrane pressures in the range of 1.5–2 bar. An optimal
[14] R. Molinari, M. Mungari, E. Drioli, A. Di Paola, V. Loddo, L. Palmisano, M.
value is also found for the permeate flux, with a maximum value of Schiavello, Study on a photocatalytic membrane reactor for water purification,
the reaction rate for a relatively low value of permeate flux, due to Catal. Today 55 (2000) 71–78.
the improvement in the mass transfer, whereas a dramatic [15] B. Del Curto, M.F. Brunella, C. Giordano, M.P. Pedeferri, V. Valtulina, L. Visai, A.
Cigada, Decreased bacterial adhesion to surface-treated titanium, Int. J. Artif.
decrease is observed for higher values due to the compression of Organs 28 (2005) 718–730.
the TiO2 layer induced by the increase in the transmembrane pres- [16] R. Van Grieken, J. Marugán, C. Sordo, C. Pablos, Comparison of the
sure. As a result the porosity and the surface area of the TiO2 layer photocatalytic disinfection of E. coli suspensions in slurry, wall and fixed-
bed reactors, Catal. Today 144 (2009) 48–54.
is reduced and thus the photocatalytic activity. [17] J.J. Porter, R.S. Porter, Filtration studies of selected anionic dyes using
E. coli inactivation experiments suggested that bacteria are able asymmetric titanium dioxide membranes on porous stainless-steel tubes, J.
to adapt its shape to cross the membrane under pressure, even Membr. Sci. 101 (1995) 67–81.
[18] J. Grzechulska-Damszel, S. Mozia, A.W. Morawski, Integration of
with nominal pore sizes smaller than the bacteria diameter. In pure photocatalysis with membrane processes for purification of water
water this process leads to a drastic decrease in viable bacteria due contaminated with organic dyes, Catal. Today 156 (2010) 295–300.
to the lethality derived of the mechano-osmotic stress, whereas in [19] K. Azrague, P. Aimar, F. Benoit-Marquie, M.T. Maurette, A new combination of
a membrane and a photocatalytic reactor for the depollution of turbid wáter,
synthetic wastewater treatment plant effluents the bacteria are Appl. Catal. B Environ. 72 (2007) 197–204.
almost unaffected. The incorporation of TiO2 to the membrane [20] X. Zheng, Q. Wanga, L. Chen, J. Wanga, R. Cheng, Photocatalytic membrane
increases the rejection of bacteria although also increases the reactor (PMR) for virus removal in water: performance and mechanisms,
Chem. Eng. J. 277 (2015) 124–129.
transmembrane pressure specially in PSS-0.2. The addition of
[21] A. Rahimpour, M. Jahanshahi, B. Rajaeian, M. Rahimnejad, TiO2 entrapped
8.5 g
m 2 of TiO2 to the PSS membranes have demonstrated to pro- nano-composite PVDF/SPES membranes: preparation, characterization, anti-
vide a good efficiency in the removal of bacteria for water disinfec- fouling and antibacterial properties, Desalination 278 (2011) 343–353.
tion applications with low values of the transmembrane pressure. [22] H. Choi, E. Stathatos, D.D. Dionysiou, Photocatalytic TiO2 films and membranes
for the development of efficient wastewater treatment and reuse systems,
Therefore, a synergistic effect derived of the physical removal of Desalination 202 (2007) 199–206.
microorganisms coupled to its photocatalytic inactivation proved [23] H. Choi, K. Zhang, D.D. Dionysiou, D.B. Oerther, G.A. Sorial, Influence of cross-
the antibiofouling effects of the combination of both processes flow velocity on membrane performance during filtration of biological
suspension, J. Membr. Sci. 248 (2005) 189–199.
due to a significant inactivation of the retained bacteria. [24] N. Ma, X. Fan, X. Quan, Y. Zhang, Ag–TiO2/HAP/Al2O3 bioceramic composite
membrane: fabrication, characterization and bactericidal activity, J. Membr.
Sci. 336 (2009) 109–117.
Acknowledgements [25] R.L. Ames, E.A. Bluhm, J.D. Way, A.L. Bunge, K.D. Abney, S.B. Schreiber, Physical
characterization of 0.5 lm cut-off sintered stainless steel membranes, J.
The authors gratefully acknowledge the financial support of the Membr. Sci. 213 (2003) 13–23.
[26] Z. Li, N. Qiu, G. Yang, Effects of synthesis parameters on the microstructure and
Spanish Ministry of Economy and Competitiveness (MINECO) phase structure of porous 316L stainless steel supported TiO2 membranes, J.
through the project WATER4FOOD (CTQ2014-54563-C3-1-R) and Membr. Sci. 326 (2009) 533–538.
Comunidad de Madrid through the program REMTAVARES [27] T. Nash, The colorimetric estimation of formaldehyde by means of the
Hantzsch reaction, J. Biochem. 55 (1953) 416–421.
(S2013/MAE-2716). Cintia Casado also acknowledges MINECO for
[28] M. Kositzi, I. Poulios, S. Malato, J. Cáceres, A. Campos, Solar photocatalytic
the FPI grant (BES-2012-056661). treatment of synthetic municipal wastewater, Water Res. 38 (2004) 1147–
1154.
[29] J. Marugán, R. van Grieken, O.M. Alfano, A.E. Cassano, Optical and
References physicochemical properties of silica-supported TiO2 photocatalysts, AIChE J.
52 (2006) 2832–2843.
[1] D. Rana, T. Matsuura, Surface modifications for antifouling membranes, Chem. [30] C. Adán, J. Marugán, S. Obregón, G. Colón, Photocatalytic activity of bismuth
Rev. 110 (2010) 2448–2471. vanadates under UV-A and visible light irradiation: Inactivation of Escherichia
[2] S. Mozia, Photocatalytic membrane reactors (PMRs) in water and wastewater coli vs oxidation of methanol, Catal. Today 240 (2015) 93–99.
treatment. A review, Sep. Purif. Technol. 73 (2010) 71–91. [31] J. Marugán, D. Hufschmidt, M.J. López-Muñoz, V. Selzer, D. Bahnemann,
[3] D.A. Athanasiou, G.Em. Romanos, P. Falaras, Design and optimization of a Photonic efficiency for methanol photooxidation and hydroxyl radical
photocatalytic reactor for water purification combining optical fiber and generation on silica-supported TiO2 photocatalysts, Appl. Catal. B Environ.
membrane technologies, Chem. Eng. J. (2015), http://dx.doi.org/10.1016/j. 201–207 (2006).
cej.2015.11.080. [32] C. Pablos, J. Marugán, R. van Grieken, C. Adán, A. Riquelme, J. Palma,
[4] A.W. Zularisam, A.F. Ismail, R. Salim, Behaviours of natural organic matter in Correlation between photoelectrochemical behaviour and
membrane filtration for surface water treatment-a review, Desalination 194 photoelectrocatalytic activity and scaling-up of P25-TiO2 electrodes,
(2006) 211–231. Electrochim. Acta 130 (2014) 261–270.
[5] D.S. Sholl, J.K. Johnson, Making high-flux membranes with carbon nanotubes, [33] C. Peng, Z. Xianzhi, Z. Minjie, W. Longde, Photocatalysis-membrane separation
Science 312 (2006) 1003–1004. coupling reactor and its application, Chin. J. Catal. 27 (2006) 752–754.
[6] M.T.M. Pendergast, E.M.V. Hoek, A review of water treatment membrane [34] H. Basri, A.F. Ismail, M. Aziz, Microstructure and anti-adhesion properties of
nanotechnologies, Energy Environ. Sci. 4 (2011) 1946–1971. PES/TAP/Ag hybrid ultrafiltration membrane, Desalination 287 (2012) 71–77.
[7] R. Molinari, L. Palmisano, E. Drioli, M. Schiavello, Studies on various reactor [35] S.H. Kim, S.-Y. Kwak, B.-H. Sohn, T.H. Park, Design of TiO2 nanoparticle self-
configurations for coupling photocatalysis and membrane processes in water assembled aromatic polyamide thin-film-composite (TFC) membrane as an
purification, J. Membr. Sci. 206 (2002) 399–415. approach to solve biofouling problem, J. Membr. Sci. 211 (2003) 157–165.

Please cite this article in press as: C. Adán et al., Bacterial inactivation and degradation of organic molecules by titanium dioxide supported on porous stain-
less steel photocatalytic membranes, Chem. Eng. J. (2016), http://dx.doi.org/10.1016/j.cej.2016.04.091