Effect of membrane wettability on membrane fouling and chemical

durability of SPG membranes used in a microbubble-aerated

biolm reactor
Chun Liu

, Tai-Min Xiao, Jing Zhang, Lei Zhang, Jing-Liang Yang, Ming Zhang
Pollution Prevention Biotechnology Laboratory of Hebei Province, School of Environmental Science and Engineering, Hebei University of Science and Technology,
Shijiazhuang 050018, China
a r t i c l e i n f o
Article history:
Received 26 September 2013
Received in revised form 8 February 2014
Accepted 2 March 2014
Available online 12 March 2014
Keywords:
Microbubble aeration
SPG membrane
Surface wettability
Membrane fouling
Chemical durability
a b s t r a c t
Shirasu porous glass (SPG) membranes have been applied for microbubble aeration in aerobic wastewater
treatment successfully. In the present study, both hydrophilic and hydrophobic SPG membranes were
used for microbubble aeration in a biolm reactor with online chemical cleaning, and their membrane
fouling and chemical durability were determined to be strongly membrane wettability dependent during
a long-term operation. The fouling layer formed on the surface of both membranes was attributed to
organic fouling mainly, and the hydrophobic membrane showed a stronger resistance to the organic
fouling. The severe chemical corrosion of the hydrophilic membrane due to exposure to the alkaline
solution used for chemical cleaning was observed, and as a result, its median pore diameter and porosity
increased signicantly. On the other hand, the pore structure of the hydrophobic membrane changed
slightly when exposed to the alkaline solution, suggesting its strong alkali-resistance due to the
non-wetting surface; but its surface hydrophobic groups could be oxidized, making membrane surface
more wettable. The hydrophobic membrane also showed more efcient performance in the respects of
oxygen transfer, contaminant removal and energy-saving. Therefore, the hydrophobic membrane seems
more appropriate to be applied for microbubble aeration in aerobic wastewater treatment.
2014 Elsevier B.V. All rights reserved.
1. Introduction
Microbubbles provide the enhanced gasliquid mass transfer
due to their useful characteristics, including large gasliquid inter-
facial area, long residence time in the liquid phase and fast
dissolution rate. In recent years, microbubble technologies have
drawn great attention due to their wide applications in many elds
of science and technology, including wastewater treatment [1].
Most reported applications of microbubble technologies for waste-
water treatment focus on physicalchemical treatment processes
at present and the enhancement of ozone and oxygen gasliquid
mass transfer in microbubble aeration has been conrmed [26].
The aerobic biological treatment process is commonly utilized
for wastewater treatment where the contaminant digestion de-
pends on the available dissolved oxygen (DO). The fast oxygen sup-
ply rate is required to promote the aerobic biochemical reaction
because of oxygen feed limitation. Therefore, a highly efcient oxy-
gen supplier and more useful aerobic wastewater treatment
system are expected. Some microbubble aerators have been
developed for aerobic wastewater treatment to increase oxygen
transfer rate [7]. However, almost all microbubble aerators cause
a damage to activated sludge, resulting in broken sludge ocs
and poor sludge settleability [7,8], so that it is difcult to apply
microbubble aeration in the activated sludge-based bioreactors.
On the other hand, microbubble aeration has been applied success-
fully in a biolm reactor using a Shirasu porous glass (SPG)
membrane system, where the oxygen utilization efciency was
achieved as high as close to 100%, much higher than that in
conventional aeration processes [9].
SPG membrane is a kind of porous glass membrane which is
prepared based on phase separation and subsequent acid leaching
[10]. SPG membranes nd many applications as a dispersion med-
ium in both membrane emulsication [1113] and gas dispersion
process [1418] for the formation of uniformly sized droplets,
and microbubbles. Briey, a dispersed phase is permeated into a
continuous phase through the uniform pores of a SPG membrane:
in the gasliquid dispersion system, the gas phase is forced
through the porous membrane into the continuous liquid phase
and uniform-sized microbubbles are formed from the inner surface
http://dx.doi.org/10.1016/j.seppur.2014.03.001
1383-5866/ 2014 Elsevier B.V. All rights reserved.

Corresponding author. Tel.: +86 311 81668436; fax: +86 311 81668428.
E-mail address: liuchun@hebust.edu.cn (C. Liu).
Separation and Purication Technology 127 (2014) 157164
Contents lists available at ScienceDirect
Separation and Purication Technology
j our nal homepage: www. el sevi er. com/ l ocat e/ seppur
of the membrane. In this dispersion process, microbubble forma-
tion mainly consists of the following three processes: (1) micro-
bubble formation across the pore; (2) microbubble growth; and
(3) microbubble detachment. The prominent advantage of this
technique is that the resultant bubble size and void fraction are
mainly determined by the membrane pore size and membrane
area, respectively. This indicates that bubble size and void fraction
can be optimized for a large-scale application. The SPG membrane
surface is inherently hydrophilic due to the presence of hydroxyl
groups such as silanol groups. On the other hand, the surface of
SPG membranes can be hydrophobized by chemical modication
with organosilane compounds. The microbubble generation and
the gasliquid mass transfer are proven to be inuenced by SPG
membrane wettability. The larger microbubbles with a broad
diameter distribution would be generated from a hydrophobic
SPG membrane, compared to a hydrophilic SPG membrane with
the same pore diameter [16]. In addition, the resistance to gas mass
transfer in a hydrophilic SPG membrane is much stronger than that
of a hydrophobic SPG membrane because of the presence of water
in the pores [19].
Membrane fouling is one of the main problems in many mem-
brane applications. For inorganic membranes used in wastewater
treatment, inorganic deposition, organics adsorption and biolm
growth are considered to be responsible for membrane fouling
[2023]. When SPG membranes were used for oil emulsication
or demulsication process, the membrane fouling was observed
as a consequence of the accumulation of oil drops on the mem-
brane surface and inside the pores [24,25]. SPG membrane fouling
caused by the protein adsorption was also reported [26]. When a
hydrophilic SPG membrane was used for microbubble aeration in
a biolm reactor, the external membrane fouling was observed,
which would deteriorate microbubble generation and oxygen mass
transfer [27]. The thermal treatment is recommended for off-line
membrane cleaning by the manufacturer, which was conrmed
to be effective to remove organic foulants [27]. Moreover, the peri-
odic chemical cleaning is needed to control the development of
SPG membrane fouling. Although chemical cleaning is considered
to be an integral part of a membrane process operation, it may
damage the membrane material due to limited membrane resis-
tance to chemical corrosion, resulting in reduced membrane per-
formance and lifetime [28,29]. The chemical resistance of SPG
membranes is still poor understood, especially when a chemical
membrane cleaning is employed during a long-term operation. It
is expected that membrane fouling and chemical durability of a
SPG membrane should be inuenced by its surface properties such
as wettability, but this is still unclear.
In this study membrane fouling of hydrophilic and hydrophobic
SPG membranes used for microbubble aeration in a biolm reactor
was observed and characterized after a long-term operation, to
investigate the effect of surface wettability on membrane fouling.
In addition, the sodium hypochlorite solution and the hydrochloric
acid solution were used for online chemical cleaning during the
long-term operation. Then surface properties and pore structure
of both SPG membranes employed were observed and character-
ized after a frequent exposure to these chemical solutions to
compare their chemical durability.
2. Materials and methods
2.1. SPG membranes
Two types of tubular SPG membranes obtained from SPG Tech-
nology Co., Ltd. (Miyazaki, Japan) were used in this study, including
a hydrophilic membrane (PJ-500-006N) with the pore size of
0.6 lm and the contact angle of 27, and a hydrophobic membrane
(PJ-500-006U) with pore size of 0.6 lm and the contact angle of
107. The SPG membrane is prepared by phase separation of the
mother glass in the Na
2
OCaOMgOAl
2
O
3
B
2
O
3
SiO
2
system,
which can separate into an acid-insoluble (silica-rich) phase and
an acid-soluble (borate-rich) phase. The SPG membrane contains
SiO
2
of about 70 wt%, Al
2
O
3
of 1015 wt%, and other components
[14]. The detailed preparation procedure of SPG membranes has
been described elsewhere [10]. The hydrophobic SPG membrane
is prepared by surface chemical modication with organosilane
compounds.
2.2. Experimental set-up
Fig. 1 depicts schematic of a biolm reactor with a SPG mem-
brane microbubble aeration system. The newhydrophilic or hydro-
phobic membrane was used to generate air microbubbles for
aeration. The gasliquid dispersion system consisted of air as the
dispersed gaseous phase and the mixed liquor in the reactor as
the continuous water phase. A continuous water phase was owed
inside the membrane at a velocity of 0.731.08 m/s, corresponding
to a Reynolds number range of 1.01 10
4
to 1.49 10
4
, while the
compressed air was introduced on the outside and forced through
the membrane pores to form microbubbles in the water phase. The
air ux was controlled at a range of 1.15 10
3
m
3
/(m
2
s) to
1.51 10
3
m
3
/(m
2
s). The diameter range of most generated
microbubbles in clean water was determined as 2060 lm under
these conditions, based on microscope observation and measure-
ment [4]. The average microbubble diameters were determined
statistically as 31.1 lm for the hydrophilic membrane and
41.7 lm for the hydrophobic membrane, respectively.
The reactor was a transparent plexiglass tank with a diameter of
250 mm and a depth of 600 mm. The work volume of the reactor
was 15 L. The commercial polypropylene ball lled with porous
polyurethane cubes with the volume of 3.2 cm
3
and the pore size
of about 1.5 mm (Shengxing, China), were used as carriers in the
reactor to support biolm growth. The diameter of these balls
was 8.0 cm and the carrier lling was about 27% in the reactor.
2.3. Experimental procedure
The activated sludge from a municipal wastewater treatment
plant was used as a microbial biomass inoculum for seeding the
reactor at an approximate concentration of 0.8 g/L during the
start-up period to promote biolm formation on the carriers. A
usual air distributor was used at this stage to aerate the reactor
and a synthetic wastewater was fed into the reactor at an organic
loading rate of 0.41 kgCOD/(m
3
d). Glucose, starch, peptone, NH
4
Cl
and KH
2
PO
4
were the main constituents in the synthetic wastewa-
ter to provide carbon, nitrogen and phosphorus sources [30]. Syn-
thetic wastewater was prepared daily and kept in a refrigerator to
avoid degradation. The start-up of the reactor was considered to be
completed when the biolm biomass on the carriers reached up to
0.8 g/L. After that, microbubble aeration using the hydrophilic or
hydrophobic SPG membrane was applied for the following long-
term operation.
The synthetic wastewater mentioned above was treated in the
microbubble-aerated biolm reactor and the operating conditions
were identical for both types of the membranes used during the
long-term operation, as shown in Table 1. The synthetic wastewa-
ter had an average chemical oxygen demand (COD) of 385.7 mg/L.
Feed wastewater was pumped to the reactor continuously and the
efuent from the bottom of the reactor was controlled by a valve.
The reactor was operated at room temperature.
Two chemical solutions, including the sodium hypochlorite
solution (1000 mg/L, pH > 11) and the hydrochloric acid solution
(0.5 mol/L), and water, were successively circulated inside the
158 C. Liu et al. / Separation and Purication Technology 127 (2014) 157164
SPG membrane for the intermittent online chemical cleaning dur-
ing the long-term operation. The chemical cleaning conducted
every two days, with the procedure of 30 min for the sodium hypo-
chlorite solution, 30 min for the hydrochloric acid solution and
15 min for water, respectively.
The performance of the biolm reactor aerated by a hydrophilic
or hydrophobic membrane was monitored, including air perme-
ation resistance, contaminant removal and energy consumption.
The surface properties and the pore structure of the new and used
SPG membranes were characterized after the operation, to illus-
trate their membrane fouling and chemical durability.
2.4. Analytical methods
The surface and the pores of the new and the fouled SPG mem-
branes were observed by scanning electron microscopy (SEM)
(S-4800-I, Hitachi, Japan). The energy dispersive spectroscopy
(EDS) technique was used associated to the SEM analysis to pro-
vide elemental distribution analysis of the SPG membrane surface
with a depth of 12 lm [31,32] and 6 spots of each membrane
sample were analyzed. The pore size distribution and porosity of
the membranes were determined using the mercury penetration
method with a mercury porosimeter (Poresizer 9320, Micromeri-
tics, USA). The surface functional groups of the SPG membranes
were also characterized by the infrared (IR) spectrometry. IR
measurements were performed using an IR-Raman spectroscopy
(Nicolet6700, ThermoFisher, USA) in the wavenumber range of
4004000 cm
1
with a resolution of 1 cm
1
.
The DO concentration was determined four times per day with
an electrochemical membrane electrode (WTW cellOx 325, Ger-
many) and a digital DO meter, to calculate the average value. Both
outside air pressure and inside liquid pressure of SPG membrane
were measured four times per day using pressure gauges to deter-
mine the transmembrane pressure, and the air ow rate was deter-
mined by a gas owmeter. The inuent and efuent water testing
was conducted on a daily basis for the duration of the biolm reac-
tor operation. Total COD and ammonia were measured in accor-
dance with the standard method. A single-phase kilowatt hour
meter was tted to the electricity supply to measure the power
consumptions of the air compressor served as an air supplier.
3. Results and discussion
3.1. SEM observation of SPG membrane surface
The surface pores of the new membrane seem clear (Fig. 2C1).
Fig. 2A1 and B1 show the SEM images of both fouled membrane
surfaces after a long-term operation, where a uniformly compacted
fouling layer is covering the fouled hydrophilic membrane surface
and a loosely bound fouling layer is covering the fouled hydropho-
bic membrane surface. The microorganisms are also observed in
the fouling layer on the fouled hydrophilic membrane surface
(Fig. 2A2). These results suggest that the chemical cleaning em-
ployed is not enough to control the membrane fouling entirely in
such a long-term operation.
Fig. 1. Schematic diagram of experimental apparatus.
Table 1
Operating conditions.
Item Value
Operation days (d) 97
SPG membrane area (m
2
) 1.57 10
3
Hydraulic retention time (HRT) (h) 12
Average inuent COD concentration (mg/L) 385.7
Average inuent ammonia concentration (mg/L) 49.3
SPG membrane cleaning Online chemical cleaning
C. Liu et al. / Separation and Purication Technology 127 (2014) 157164 159
In addition, both fouled membranes were off-line cleaned
further with acid treatment (hydrochloric acid solution of
0.5 mol/L, for 4 h at room temperature) and thermal treatment
(550 C, 2 h) [27]. The fouling layer on the hydrophobic membrane
surface was removed effectively after cleaned with acid treatment
(Fig. 2B2), but little fouling layer was removed from the hydro-
philic membrane surface (Fig. 2A3). This suggests more irreversible
fouling for the hydrophilic membrane. When the thermal treat-
ment is simultaneously applied, the fouling layer is removed com-
pletely (Fig. 2A4 and B3). The membrane fouling seems relatively
slight on the hydrophobic membrane, considering the friable and
weakly adhesive fouling layer, which is probably ascribed to its
non-wetting surface and the consequently reduced adsorption of
organic foulants.
Moreover, SEM observation also shows the obvious surface pore
destruction of the used hydrophilic membrane after the fouling
layer is removed (Fig. 3A), compared to the new membrane
(Fig. 3C). The three-dimensional network of the membrane formed
by the cylindrical tortuous pores seems to be corroded and the
pores become larger. For the used hydrophobic membrane, the sur-
face pore destruction seems slighter (Fig. 3B) than the used hydro-
philic membrane. The online chemical cleaning is considered to be
the reason for the surface pore destruction of both membranes due
to exposure to the chemical solutions. These SEM observations
Fig. 2. SEM observation of surface microstructure of SPG membranes: A14, the hydrophilic membrane surface, including the fouled, the microorganisms in the fouling layer,
the cleaned with acid solution, and the cleaned with both acid solution and thermal treatment, respectively; B13, the hydrophobic membrane surface, including the fouled,
the cleaned with acid solution, and the cleaned with both acid solution and thermal treatment, respectively; C1, the new membrane surface.
160 C. Liu et al. / Separation and Purication Technology 127 (2014) 157164
suggest that the hydrophobic membrane could provide stronger
fouling resistance and better chemical durability than the hydro-
philic membrane.
3.2. Surface elemental distribution of SPG membrane
Fig. 4 shows the surface elemental distributions of the new and
the fouled SPG membranes. For both new membranes, oxygen and
silicon are the dominant elements on the surface, which is consis-
tent with the membrane composition [14]. Carbon is also relatively
predominant on the hydrophobic membrane surface due to the
surface hydrophobized modication with organosilane compounds
having the hydrocarbon and uorocarbon functional groups.
A signicant increase in carbon weight is presented on the fo-
uled surfaces of both hydrophilic (Fig. 4a) and hydrophobic
(Fig. 4b) membranes as a consequence of the fouling layer covered.
At the same time, the decreased contents of oxygen and metal ele-
ments were also observed on the surface of the fouled membranes.
After the fouling layer is removed, the surface elemental distribu-
tion of the cleaned membranes is similar with the new membrane.
These results suggest that not the carbonate precipitation but the
organic fouling contributes to the fouling layer mainly for both
membranes.
3.3. Infrared (IR) spectrum analysis
Fig. 5 shows a comparison of IR spectra between the new and
the fouled membranes. The strong absorption peaks appeared in
all spectra include the peak at 1055 cm
1
which is due to SiAO
stretching bond [33] and the peak at 470 cm
1
which is generally
correlated with OASiAO bending mode [34]. This is consistent
with the dominance of SiO
2
in SPG membranes.
Compared to the hydrophilic membrane, one signicant differ-
ence observed in the IR spectrum of the hydrophobic membrane is
the much stronger absorption at 2930 and 2850 cm
1
, which can
be assigned to the CAH asymmetric vibration and symmetric
vibration of the methylene groups, respectively, due to the surface
hydrophobized modication [16].
No new adsorption peak and no shift of existing peaks are ob-
served in IR spectrum of the fouled hydrophilic membrane, com-
pared to the new membrane. However, a new strong adsorption
peak at 1718 cm
1
appears in the spectrum of the fouled hydro-
phobic membrane, which is related to carbonyl (C@O) stretching
Fig. 3. SEM observation of surface pore destruction of the membranes: (A) used hydrophilic membrane, (B) used hydrophobic membrane and (C) new membrane.
0
10
20
30
40
50
60
C O Na Al Si P Cl K Ca Fe
w
t

%
Element
New membrane
Fouled membrane
Cleaned membrane
A
0
10
20
30
40
50
60
C O Na Al Si P Cl K Ca Fe
w
t

%
Element
New membrane
Fouled membrane
Cleaned membrane
B
Fig. 4. Surface elemental distributions of the new, the fouled and the cleaned
membranes: (A) hydrophilic and (B) hydrophobic.
C. Liu et al. / Separation and Purication Technology 127 (2014) 157164 161
vibration. This new peak is considered to be due to the oxidation of
the methylene groups on the hydrophobic membrane surface by
the sodium hypochlorite solution used for chemical cleaning
[35]. Then the hydrophobic membrane surface would become
more polar and wettable.
3.4. Pore structure
The pore size distributions and the cumulative pore volumes of
the new and the used membranes are determined after the fouling
layer was removed from the fouled membrane by acid treatment
and thermal treatment [27], to illustrate the change in the pore
structure inside the used membrane, as shown in Fig. 6. The new
hydrophobic membrane has almost the same pore structure with
the new hydrophilic membrane, since the pore structure is not
inuenced by the surface hydrophobized modication [19]. The
volume-based median pore diameter and the cumulative pore vol-
ume of both new membranes are approximately 0.52 lm and
0.53 cm
3
/g, respectively.
The median pore diameter and the cumulative pore volume of
the used hydrophilic membrane are determined as 0.80 lm and
2.06 cm
3
/g, respectively, much higher than the new membrane.
In addition, the bulk density of the membrane decreases from
1.04 g/cm
3
of the new membrane to 0.40 g/cm
3
of the used mem-
brane, which means that the weight loss of the used membrane is
as high as 61.5%. These results indicate the severe chemical corro-
sion happened inside the hydrophilic membrane when exposed
frequently to the chemical solutions used for online membrane
cleaning.
The severe chemical corrosion of the used hydrophilic
membrane is predictable when exposed to the alkaline sodium
hypochlorite solution, because SPG membrane has poor alkali-
resistance as a kind of glass membrane. When the membrane is ex-
posed to the alkaline solution, the SiAO bonds of the glass network
in the membrane skeleton break due to the attack of OH-ions, and
thus the silica network is gradually destroyed. As a result, the
remaining glass components are released, resulting in total glass
dissolution and larger pore size. The chemical corrosion of SPG
membranes caused by the alkaline solution is a problem for its
application because the alkaline solution is commonly used for or-
ganic fouling control [36]. The development of alkali-resistant SPG
membrane is suggested to prevent this chemical corrosion [37].
On the other hand, the hydrophobic membrane shows a strong
resistance to the chemical corrosion. The pore size distribution and
the cumulative pore volume of the used hydrophobic membrane
are similar with the new membrane, and its median pore diameter
and cumulative pore volume increase only to 0.59 lm and
0.60 cm
3
/g, respectively. The surface hydrophobicity could prevent
the alkaline solution entering inside the membrane, and then the
chemical corrosion could be avoided effectively. Therefore,
hydrophobical modication of SPG membrane surface could pro-
vide a feasible solution to resist the chemical corrosion inside the
membrane.
It should be noted that the alkaline sodium hypochlorite solu-
tion could oxidize the surface hydrophobic groups and make the
membrane surface more wettable. After this, the chemical resis-
tance of the hydrophobic membrane would be reduced. This might
be the reason for the slight pore destruction of the used hydropho-
bic membrane.
3.5. Air permeation resistance
The trans-membrane pressure is determined as the difference
between air pressure and the mean pressure of the continuous
water phase. The air permeation resistances of the hydrophilic
and hydrophobic membranes are then calculated using Darcys
equation during the long-term operation [38], as shown in Fig. 7.
The air permeation resistance of the hydrophilic membrane is
much higher than that of the hydrophobic membrane at the initial
0
20
40
60
80
100
400 800 1200 1600 2000 2400 2800 3200 3600 4000
T
r
a
n
s
m
i
t
t
a
n
c
e

(
%
)
Wavenumber (cm
-1
)
New Fouled
3440
2930
2850
1600
1400
1055
470
714
788
0
20
40
60
80
100
400 800 1200 1600 2000 2400 2800 3200 3600 4000
T
r
a
n
s
m
i
t
t
a
n
c
e

(
%
)
Wavenumber (cm
-1
)
New Fouled
3440
2930
2850 1600
1400
1055
470
714
788
1718
A
B
Fig. 5. Infrared spectra of the new and the fouled membranes: (A) hydrophilic and
(B) hydrophobic.
0
10
20
30
40
50
0.01 0.1 1 10 100
d
V
/
d
l
o
g
D
p
o
r
e

v
o
l
u
m
e

(
m
L
/
g
)
Pore size (m)
Used hydrophilic membrane
Used hydrophobic membrane
New membrane
A
0
0.5
1
1.5
2
2.5
0.01 0.1 1 10 100
C
u
m
u
l
a
t
i
v
e

p
o
r
e

v
o
l
u
m
e

(
m
L
/
g
)
Pore size (m)
Used hydrophilic membrane
Used hydrophobic membrane
New membrane
B
Fig. 6. Pore size distribution (A) and cumulative pore volume (B) of the new and the
used (fouled and cleaned) hydrophilic and hydrophobic SPG membranes.
162 C. Liu et al. / Separation and Purication Technology 127 (2014) 157164
period, due to its wetting surface and the water-lled pores. Note
that, the air permeation resistance of the hydrophilic membrane
decreases gradually during the subsequent operation although
the membrane fouling has been developing. The increased porosity
of the membrane caused by the chemical corrosion, from the vir-
ginal of 54.8% to the nal of 81.9%, is considered to be responsible
for the decreased air permeation resistance.
For the hydrophobic membrane, its large contact angle tends to
facilitate the spreading of the gaseous phase and the microbubble
formation [16], which is the reason for its low air permeation resis-
tance. However, the air permeation resistance of the hydrophobic
membrane increased signicantly, due to the increase in its surface
hydrophilicity probably after a frequent exposure to the sodium
hypochlorite solution.
3.6. Performance of the treatment systems
Table 2 shows performance comparison of the treatment sys-
tems using the hydrophilic or hydrophobic membrane under the
same operating conditions. The higher DO concentration and more
efcient contaminant removal are achieved when using the hydro-
phobic membrane, but not signicant. The better performance of
the hydrophobic membrane is ascribed to its lower resistance to
oxygen transfer probably, compared to the hydrophilic membrane
[19]. In addition, the 33.6% lower energy consumption of the air
supplier is required for hydrophobic membrane due to its lower
air permeation resistance.
3.7. Proposed membrane type
The results presented in this study reveal the great inuence of
surface wettability on membrane fouling and chemical resistance
of the SPG membrane used for microbubble aeration in aerobic
wastewater treatment. The development of the organic fouling
could be decelerated when the membrane surface is hydropho-
bically modied. The foulant layer could be removed from the
hydrophobic membrane more easily due to its weaker adhesion,
compared to the hydrophilic membrane.
The poor chemical durability of the hydrophilic membrane
against the alkaline solution is also conrmed in this study. There-
fore, the alkaline solution should be applied cautiously for the
hydrophilic membrane cleaning. The hydrophobic membrane
shows a strong resistance to the chemical corrosion due to its
non-wetting surface. However, the oxidative chemical solution
would damage the surface hydrophobicity of the hydrophobic
membrane, resulting in a reduced chemical resistance. These nd-
ings should be considered when the chemical solutions are applied
for SPG membrane cleaning.
The hydrophobic membrane seems more applicable for micro-
bubble aeration in aerobic wastewater treatment, due to its stron-
ger fouling resistance, better chemical durability and more
efcient operational performance in the respects of oxygen trans-
fer, contaminant removal and energy requirement.
4. Conclusions
The membrane fouling was compared between the hydrophilic
and hydrophobic SPG membranes used for microbubble aeration in
a biolm reactor, and their chemical durability was also investi-
gated when the chemical solutions were applied for online clean-
ing. The resistances of a SPG membrane to membrane fouling
and chemical corrosion were found to be strongly membrane wet-
tability dependent. The organic fouling was identied to be the
main reason for the surface fouling layer formation for both types
of membranes by SEM observation and surface elemental distribu-
tion analysis. The hydrophobic membrane showed a stronger resis-
tance to the organic fouling than the hydrophilic membrane. In
addition, the severe chemical corrosion of the hydrophilic mem-
brane was observed after a frequent exposure to the alkaline so-
dium hypochlorite solution used for chemical cleaning, and as a
result, its median pore diameter and porosity increased signi-
cantly. On the other hand, the hydrophobic membrane showed
strong resistance to chemical corrosion and its pore structure
changed slightly. However, the surface hydrophobic groups of the
hydrophobic membrane could be oxidized when exposed to the so-
dium hypochlorite solution, resulting in the reduced surface
hydrophobicity and the increased air permeation resistance. The
hydrophobic membrane also showed more efcient performance
in the respects of oxygen transfer, contaminant removal and en-
ergy-saving. Therefore, the hydrophobic membrane seems more
appropriate to be applied for microbubble aeration in aerobic
wastewater treatment, considering its stronger fouling resistance,
better chemical durability and more efcient operation
performance.
Acknowledgments
The authors acknowledge the nancial supports from National
Natural Science Foundation of China (51008111) and Hebei Science
and Technology Agency, China (11966726D).
References
[1] A. Agarwal, W.J. Ng, Y. Liu, Principle and applications of microbubble and
nanobubble technology for water treatment, Chemosphere 84 (2011) 1175
1180.
[2] M. Takahashi, K. Chiba, P. Li, Free-radical generation from collapsing
microbubbles in the absence of a dynamic stimulus, J. Phys. Chem. B 111
(2007) 13431347.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 10 20 30 40 50 60 70 80 90 100
A
i
r

p
e
r
m
e
a
t
i
o
n

r
e
s
i
s
t
a
n
c
e

(
1
0
1
0
m
-
1
)
Running time (d)
Hydrophilic membrane
Hydrophobic membrane
Fig. 7. Air permeation resistance of hydrophilic and hydrophobic SPG membranes
during a long-term operation.
Table 2
Performance of the treatment system using the hydrophilic or hydrophobic membrane.
Item Hydrophilic
membrane
Hydrophobic
membrane
Average DO concentration (mg/L) 2.13 2.84
Average efuent COD concentration
(mg/L)
45.6 34.7
Average COD removal rate (%) 88.2 91.0
Average efuent ammonia
concentration (mg/L)
16.2 13.5
Average ammonia removal rate (%) 67.1 72.6
Average organic loading rate removed
(kg COD m
3
d
1
)
0.68 0.70
Average energy consumption of air
supplier (kW)
0.053 0.035
C. Liu et al. / Separation and Purication Technology 127 (2014) 157164 163
[3] P. Li, M. Takahashi, K. Chiba, Enhanced free-radical generation by shrinking
microbubbles using a copper catalyst, Chemosphere 77 (2009) 11571160.
[4] L.B. Chu, X.H. Xing, A.F. Yu, Y.N. Zhou, X.L. Sun, B. Jurcik, Enhanced ozonation of
simulated dyestuff wastewater by microbubbles, Chemosphere 68 (2007)
18541860.
[5] L.B. Chu, X.H. Xing, A.F. Yu, X.L. Sun, B. Jurcik, Enhanced treatment of practical
textile wastewater by microbubble ozonation, Process Safety Environ. Protect.
86 (2008) 389393.
[6] S. Liu, Q.H. Wang, H.Z. Ma, P.K. Huang, J. Li, T. Kikuchi, Effect of micro-bubbles
on coagulation otation process of dyeing wastewater, Sep. Purif. Technol. 71
(2010) 337346.
[7] K. Terasaka, A. Hirabayashi, T. Nishino, S. Fujioka, D. Kobayashi, Development
of microbubble aerator for waste water treatment using aerobic activated
sludge, Chem. Eng. Sci. 66 (2011) 31723179.
[8] C. Liu, H. Tanaka, J. Ma, L. Zhang, J. Zhang, X. Huang, Y. Matsuzawa, Effect of
microbubble and its generation process on mixed liquor properties of activated
sludge using Shirasu porous glass (SPG) membrane system, Water Res. 46
(2012) 60516058.
[9] C. Liu, H. Tanaka, J. Zhang, L. Zhang, J.L. Yang, X. Huang, N. Kubota, Successful
application of Shirasu porous glass (SPG) membrane system for microbubble
aeration in a biolm reactor treating synthetic wastewater, Sep. Purif. Technol.
103 (2013) 5359.
[10] M. Kukizaki, T. Nakashima, Acid leaching process in the preparation of porous
glass membranes from phase-separated glass in the Na
2
OCaOMgOAl
2
O
3

B
2
O
3
SiO
2
system, Membrane 29 (2004) 301308.
[11] M. Yasuno, M. Nakajima, S. Iwamoto, T. Maruyama, S. Sugiura, I. Kobayashi, A.
Shono, K. Satoh, Visualization and characterization of SPG membrane
emulsication, J. Membr. Sci. 210 (2002) 2937.
[12] G.T. Vladisavljevic , H. Schubert, Inuence of process parameters on droplet
size distribution in SPG membrane emulsication and stability of prepared
emulsion droplets, J. Membr. Sci. 225 (2003) 1523.
[13] M. Kukizaki, Shirasu porous glass (SPG) membrane emulsication in the
absence of shear ow at the membrane surface. Inuence of surfactant type
and concentration, viscosities of dispersed and continuous phases, and
transmembrane pressure, J. Membr. Sci. 327 (2009) 234243.
[14] M. Kukizaki, M. Goto, Size control of nanobubbles generated from Shirasu-
porous-glass (SPG) membranes, J. Membr. Sci. 281 (2006) 386396.
[15] M. Kukizaki, M. Goto, Spontaneous formation behavior of uniform-sized
microbubbles from Shirasu porous glass (SPG) membranes in the absence of
water-phase ow, Colloids Surf. A: Physicochem. Eng. Aspects 296 (2007)
174181.
[16] M. Kukizaki, T. Wada, Effect of the membrane wettability on the size and size
distribution of microbubbles formed from Shirasu-porous-glass (SPG)
membranes, Colloids Surf. A: Physicochem. Eng. Aspects 317 (2008)
146154.
[17] M. Kukizaki, Y. Baba, Effect of surfactant type on microbubble formation
behavior using Shirasu porous glass (SPG) membranes, Colloids Surf. A:
Physicochem. Eng. Aspects 326 (2008) 129137.
[18] M. Kukizaki, Microbubble formation using asymmetric Shirasu porous glass
(SPG) membranes and porous ceramic membranes a comparative study,
Colloids Surf. A: Physicochem. Eng. Aspects 340 (2009) 2032.
[19] M. Kukizaki, K. Fujimoto, S. Kai, K. Ohe, T. Oshima, Y. Baba, Ozone mass transfer
in an ozonewater contacting process with Shirasu porous glass (SPG)
membranes a comparative study of hydrophilic and hydrophobic
membranes, Sep. Purif. Technol. 72 (2010) 347356.
[20] N. Cicek, D. Dionysiou, M.T. Suidan, P. Ginestet, J.M. Audic, Performance
deterioration and structural changes of a ceramic membrane bioreactor due to
inorganic abrasion, J. Membr. Sci. 163 (1999) 1928.
[21] M.D. Caric , S.D. Milanovic , D.M. Krstic , M.N. Tekic , Fouling of inorganic
membranes by adsorption of whey proteins, J. Membr. Sci. 165 (2000) 8388.
[22] J. Bentama, K. Ouazzani, Z. Lakhliai, M. Ayadi, Inorganic membranes made of
sintered clay for the treatment of biologically modied water, Desalination
168 (2004) 295299.
[23] G.Y. Chai, A.R. Greenberg, W.B. Krantz, Ultrasound, gravimetric, and SEM
studies of inorganic fouling in spiral-wound membrane modules, Desalination
208 (2007) 277293.
[24] G.T. Vladisavljevic , M. Shimizu, T. Nakashima, Preparation of monodisperse
multiple emulsions at high production rates by multi-stage premix membrane
emulsication, J. Membr. Sci. 244 (2004) 97106.
[25] M. Kukizaki, M. Goto, Demulsication of water-in-oil emulsions by
permeation through Shirasu-porous-glass (SPG) membranes, J. Membr. Sci.
322 (2008) 196203..
[26] K. Nakamura, K. Matsumoto, Protein adsorption properties on a microltration
membrane: a comparison between static and dynamic adsorption methods, J.
Membr. Sci. 285 (2006) 126136.
[27] C. Liu, H. Tanaka, L. Zhang, J. Zhang, X. Huang, J. Ma, Y. Matsuzawa, Fouling and
structural changes of Shirasu porous glass (SPG) membrane used in aerobic
wastewater treatment process for microbubble aeration, J. Membr. Sci. 421
422 (2012) 225231.
[28] S.K.S. Al-Obeidani, H. Al-Hinai, M.F.A. Goosen, S. Sablani, Y. Taniguchi, H.
Okamura, Chemical cleaning of oil contaminated polyethylene hollow ber
microltration membranes, J. Membr. Sci. 307 (2008) 299308.
[29] P. Blanpain-Avet, J.F. Migdal, T. Bnzech, Chemical cleaning of a tubular
ceramic microltration membrane fouled with a whey protein concentrate
suspension characterization of hydraulic and chemical cleanliness, J. Membr.
Sci. 337 (2009) 153174.
[30] X. Huang, R. Liu, Y. Qian, Behavior of soluble microbial products in a
membrane bioreactor, Process Biochem. 36 (2000) 401406.
[31] M.A.S. Rodrigues, C. Korzenovski, E. Gondran, A.M. Bernardes, J.Z. Ferreira,
Evaluation of changes on ion-selective membranes in contact with zinc-
cyanide complexes, J. Membr. Sci. 279 (2006) 140147.
[32] M. Dogan, S. Kilicarslan, Effects of process parameters on the synthesis of
palladiummembranes, Nucl. Instrum. Meth. Phys. Res. B 266 (2008) 34583466.
[33] E.M.A. Khalil, F.H. ElBatal, Y.M. Hamdy, H.M. Zidan, M.S. Aziz, A.M. Abdelghany,
Infrared absorption spectra of transition metals-doped soda lime silica glasses,
Physica B 405 (2010) 12941300.
[34] Y. Zhang, M.Q. Wang, D. Wu, Study on the structure of SiO
2
B
2
O
3
xerogels with
Eu
3+
and sulfosalicylic acid as a probe, Mater. Lett. 35 (1998) 144150.
[35] M. Dobromir, G. Biliuta, D. Luca, M. Aori, V. Harabagiu, S. Coseri, XPS study of
the ion-exchange capacity of the native and surface oxidized viscose bers,
Colloids Surf. A: Physicochem. Eng. Aspects 381 (2011) 106110.
[36] Y.H. Mo, J.H. Chen, W.C. Xue, X. Huang, Chemical cleaning of nanoltration
membrane ltrating the efuent from a membrane bioreactor, Sep. Purif.
Technol. 75 (2010) 407414.
[37] M. Kukizaki, Large-scale production of alkali-resistant Shirasu porous glass
(SPG) membranes: inuence of ZrO
2
addition on crystallization and phase
separation in Na
2
OCaOAl
2
O
3
B
2
O
3
SiO
2
glasses; and alkali durability and
pore morphology of the membranes, J. Membr. Sci. 360 (2010) 426435.
[38] J.L. Wu, X. Huang, Effect of mixed liquor properties on fouling propensity in
membrane bioreactors, J. Membr. Sci. 342 (2009) 8896.
164 C. Liu et al. / Separation and Purication Technology 127 (2014) 157164