JES-00536; No of Pages 10

J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX XXXX

Available online at www.sciencedirect.com

ScienceDirect
www.elsevier.com/locate/jes

Membrane fouling in ultrafiltration of natural water after

pretreatment to different extents

Lu Ao1,2 , Wenjun Liu1,, Lin Zhao1 , Xiaomao Wang1

1. School of Environment, Tsinghua University, Beijing 100084, China. Email: zyzzalou@163.com
2. Department of National Defence Architecture Planning & Environmental Engineering, Logistic Engineering University, Chongqing 401311, China

AR TIC LE I N FO ABS TR ACT

Article history: The combined fouling during ultrafiltration (UF) of surface water pretreated to different
Received 17 July 2015 extents was investigated to disclose the roles of polysaccharides, proteins, and inorganic
Revised 28 September 2015 particles in UF membrane fouling. Both reversible and irreversible fouling decreased with
Accepted 29 September 2015 enhanced pretreatment (biologically active carbon (BAC) treatment and sand filtration). The
Available online xxxx sand filter effluent fouled the membrane very slowly. The UF membrane removed turbidity
to less than 0.1 nephelometric turbidity unit (NTU), reduced polysaccharides by 25.4%
Keywords: 29.9%, but rejected few proteins. Both polysaccharides and inorganic particles were
Ultrafiltration detected on the fouled membranes, but inorganic particles could be effectively removed
Membrane fouling by backwashing. The increase of turbidity in the sand filter effluent to 3.05 NTU did not
Biopolymer significantly increase the fouling rate, but an increase in the turbidity in the BAC effluent to
Polysaccharide 6.11 NTU increased the fouling rate by more than 100%. The results demonstrated that the
Pretreatment polysaccharide, not the protein, constituents of biopolymers were responsible for
membrane fouling. Membrane fouling was closely associated with a small fraction of
polysaccharides in the feed water. Inorganic particles exacerbated membrane fouling only
when the concentration of foulinginducing polysaccharides in the feed water was
relatively high. The combined fouling was largely reversible, and polysaccharides were
the predominant substances responsible for irreversible fouling.
2015 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences.
Published by Elsevier B.V.

Introduction failure. An understanding of membrane fouling is therefore

Ultrafiltration (UF) has been increasingly applied for drinking
water production due to its consistently high performance in
terms of particle and microorganism removal (Boudaud et al.,
2012; Hagen, 1998; Xiao et al., 2012). Optimally, UF can be
operated as a stand-alone process with no requirement
for pretreatment operations. However, one major challenge
facing UF is membrane fouling, which impairs its cost efficien-
cy, in that membrane fouling can decrease filtration flux and
increase operational cost and, most severely, result in system
crucial for better design and operation of UF processes.
Membrane fouling is caused by the accumulation of any
substance (i.e., a foulant) on the membrane surface or in the
membrane material; a foulant restricts water flowing across
the membrane. The foulants in natural waters include
inorganic particles, natural organic matter (NOM), microor-
ganisms, mineral ions (e.g., calcium), and mixtures of these
substances due to mutual interactions (Buffle et al., 1998).
Nevertheless, NOM was demonstrated to be primarily respon-
sible for UF membrane fouling (Cho et al., 2000; Escobar et al.,

Corresponding author. E-mail: wjliu@tsinghua.edu.cn (Wenjun Liu).

http://dx.doi.org/10.1016/j.jes.2015.09.008
1001-0742/ 2015 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
2 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX XXX
2005; Shao et al., 2011; Wang and Wang, 2006; Yamamura et
al., 2007; Yuan and Zydney, 2000). In particular, the biopoly-
mers or colloidal fractions of NOM were more closely
associated with membrane fouling than other NOM fractions,
such as humic substances (Carroll et al., 2000; Katsoufidou et
al., 2008; Kennedy et al., 2008). Measurements using a liquid
chromatograph-organic carbon detector (LC-OCD) demon-
strated that biopolymers were easily retained by UF mem-
branes (Halle et al., 2009; Huang et al., 2011; Kennedy et al., 2005;
Subhi et al., 2012). The major components of biopolymers are
polysaccharides and proteins, but the polysaccharide compo-
nent tends to contribute more to membrane fouling. Analyses
using Fourier transform infrared (FT-IR) spectroscopy showed
that polysaccharides were abundant on the fouled membrane
surface (Cho et al., 1998; Kimura et al., 2004). Many polysaccha-
rides in water, sometimes aided by calcium ions, can form a
gel-like structure, which usually has very low filterability.
Calcium ions are usually abundant in most natural waters and
exceed the quantity needed for gel formation. Nevertheless,
membrane fouling by polysaccharides can also be aggravated in
the presence of proteins and humic substances (Gray et al.,
2011). Fluorescence excitationemission matrix (EEM) measure-
ments showed that the protein peaks had some correlation
with the membrane fouling rate (Kimura et al., 2014). Peldszus
et al. (2011) found that protein-like substances contributed to
both reversible and irreversible fouling.
Recently, much attention has been paid to the combined
fouling caused by NOM coexisting with inorganic particles.
The inorganic particles that are naturally present in or
intentionally added to water include silts, clays, and precip-
itated crystals, as well as iron and aluminum oxyhydroxides
(Yiantsios and Karabelas, 1998). Although inorganic particles
alone usually do not lead to severe membrane fouling (Tian et
al., 2013) and whatever fouling does occur is mostly reversible
and easily removed by physical cleaning (Li et al., 2010; Peiris
et al., 2010), previous studies showed that inorganic particles
could aggravate membrane fouling caused by NOM. For
example, membrane fouling by NOM model compounds
(such as bovine serum albumin, humic acid and sodium
alginate) increased in the presence of silica particles (Munla et
al., 2012). Tian et al. (2013) also found that the membrane
fouling rate was proportional to the concentration of silica
particles in two size ranges (0.510 m and 45 m) in the
range of 1050 mg/L. However, a few studies showed that the
coexistence of inorganic particles with NOM alleviated
membrane fouling. Notwithstanding the above efforts, little
is known about the combined fouling of membranes by NOM
and inorganic particles found in natural water.
Pretreatment of water prior to filtration, usually using
conventional drinking water treatment processes, has been
proposed as a means to control membrane fouling. Coagula-
tion was found to be effective in reducing membrane fouling
(Chen et al., 2007; Citulski et al., 2008; Song et al., 2008) in that
NOM could be substantially adsorbed to flocs of hydrolyzed
coagulants that are highly filterable. Coagulation in combina-
tion with sedimentation was more effective than coagulation
alone in reducing membrane fouling (Liang et al., 2008),
because the pretreatment could effectively remove both
NOM and inorganic particles from the water (Liu et al., 2011).
Sand filtration was added following coagulation to control
Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
membrane fouling, but the additional benefit was found to be
minimal (Xia et al., 2004). Ozonation reduced membrane fouling
because of its effectiveness in degrading biopolymers to small
molecules that would be less likely to foul the membrane
(Geismar et al., 2012; Guo et al., 2013, 2014). Biologically active
carbon (BAC) filtration was shown to effectively decrease
membrane fouling due to the retention of biopolymers in the
filter (Guo et al., 2013; Halle et al., 2009; Huang et al., 2011).
In the study reported here, we investigated the combined
fouling of UF membranes by NOM and inorganic particles found
in natural water. The effects of various pretreatment tech-
niques were examined, and the relative importance of polysac-
charides and proteins in UF membrane fouling was evaluated.
The water samples used as feed to the UF membrane were
collected from a full-scale drinking water treatment plant
(DWTP) at various stages in the treatment process. The full
treatment process consisted of coagulation, followed by sedi-
mentation, ozonation, up-flow BAC (UBAC) degradation, sand
filtration, and chlorination (disinfection). The UBAC functioned
more as a biological reactor than as a conventional down-flow
BAC filter (Han et al., 2013). It was expected that both
biopolymers and inorganic particles would be gradually re-
moved throughout the treatment process. Special attention was
paid to the differences in fouling potential of different polysac-
charides in the feed water. A previous study showed that the
polysaccharides could be roughly grouped into gelling and
non-gelling components (Wang and Waite, 2009).
1. Materials and methods
1.1. Feed water
The feed water for the UF membrane experiments was from
grab samples collected from three stages of treatment in the
DWTP: after sedimentation (denoted as FW-S hereafter);
after UBAC treatment (denoted as FW-B); and after sand
filtration (denoted as FW-F). The dissolved oxygen (DO),
conductivity, and pH of samples were measured on site.
Thereafter, all samples were transported to the laboratory
within 24 hr, where other water quality parameters including
dissolved organic carbon (DOC), spectral absorption (UV254),
polysaccharide concentration, and turbidity were measured
immediately upon arrival. The feed water samples were then
stored in a refrigerator at 4C and warmed to room temper-
ature prior to use in the ultrafiltration experiments.
1.2. Ultrafiltration procedure
A laboratory-scale membrane filtration apparatus was used for
all filtration experiments. The apparatus consisted of a
home-made tubular filtration cell with an effective volume of
500 mL, two peristaltic pumps (BP100-1L, Longer, Hebei, China)
to add feed water and provide permeate suction, a pressure
transducer (CX-203, Futexin, Beijing, China) to measure
trans-membrane pressure (P), and other accessories (Fig. 1). A
hollow fiber UF membrane module (Litree, Hainan, China) with
an effective filtration area of 0.04 m2 was installed in the tubular
filtration cell. The membrane material was polyvinyl chloride
and the nominal membrane pore size was 0.01 m.
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX XXXX
Pressure Peristaltic
Transducer P Pump
Filtration
Cell
UF Membrane
Module
Peristaltic
Pump
Fig. 1 The apparatus diagram used for ultrafiltration (UF)
tests.
All filtrations were conducted in a constant-flux, dead-end
mode. Before filtrations, the membranes were cleaned by
filtering DI water. The filtration flux was set at 30 L/m2/hr
(LMH). Each feed water sample was filtered for six consecutive
cycles. Each cycle consisted of a filtration period of 1 hr and a
backwash period of 2 min with DI water. The backwash
process was conducted out of the filtration cell, for which
the flux was set at 60 LMH, and the backwash waste was
discharged. During the filtration, the trans-membrane
pressure was monitored and automatically recorded (XSR,
Futexin, Beijing, China). TheP were used to determine the
development of membrane fouling according to the Darcy
equation (Eq. (1)):
P dimensional fluorescence excitation-emission matrix (3D-EEM)
Rt Rm R f 1
J spectra (F-7000, Hitachi, Japan). The excitation wavelength (Ex)
where, Rt (1/m) is the total filtration resistance (equal to the sum
of the intrinsic membrane resistance (Rm) and the fouling layer
resistance (Rf)), (Pasec) the water viscosity and J (m3/sec) the
filtration flux. The total resistance increased with time after
filtration started, when it was equal to the intrinsic membrane
resistance. Every backwash restored the membrane filterability to
some extent. The removed and remaining fouling after back-
washes were termed as reversible and irreversible fouling,
respectively. The corresponding filtration resistances were de-
noted as Rr and Rir, respectively, and in combination determined
the fouling layer resistance (Rf) (Eq. (2)):
R f Rr Rir 2
On conclusion of filtration after six cycles, the membrane was
chemically cleaned; first backwashing with a 200 mg/L sodium
hypochlorite solution for 30 min and then backwashing with a
2% citric acid solution for another 30 min, with flux set at 60 LMH
in both cases. The residual filtration resistance was determined
after each chemical cleaning; these measurements indicated the
effectiveness of chemical cleaning in restoring the membrane
filterability.
Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
3
1.3. Membrane and fouling layer characterization
Samples of membranes in their virgin, fouled, backwashed,
and cleaned states were examined using a scanning electron
microscope (SEM) (Quanta 200, FEI, The Netherlands) after
appropriate pretreatment. The virgin membrane samples were
rinsed with ultrapure water and subsequently oven-dried at 35C
for 12 hr. The fouled, backwashed, and cleaned membrane
samples were first gently rinsed with ultrapure water to remove
any loosely attached substances and then oven-dried at 35C for
12 hr. All samples were coated with a uniform layer of platinum
in a sputter-coating chamber (EM ACE600, Leica, Germany) prior
to SEM observation. The chemical elements in the fouling layers
on some membrane samples were also analyzed using an
elemental analysis system (Genesis, EDAX, USA) integrated with
the SEM.
The virgin, fouled, backwashed, and cleaned membrane
samples were also analyzed for chemical composition using
an FT-IR spectrometer (Nicolet 6700, Thermo Fisher, USA)
equipped with a single reflection attenuated total reflectance
(ATR) accessory (ZnSe crystal) at a nominal incident angle of
45. The FT-IR spectra were collected as the average of 32
scans in the wavenumber range from 4000 to 650 cm1 at a
4 cm1 interval.
1.4. Analytical methods
Polysaccharide concentrations in water were determined using
the phenol-sulfuric acid method with glucose as the standard.
Significant differences in polysaccharide rejection for different
feed waters by UF were analyzed by SPSS (SPSS, USA) software.
The absorbance at 490 nm was measured, and the polysaccha-
ride concentration was expressed as glucose equivalent.
Protein concentrations in water were determined using three-
range was between 220 and 450 nm, and the emission wave-
length (Em) was between 240 and 600 nm. The sampling interval
was 5 nm, the response time was 0.01 sec and the scanning
speed was 2400 nm/min. The spectral data were analyzed using
Origin (OriginLab, USA) software. Spectral subtraction was
performed to remove blank spectra that were mainly caused by
Raman scattering (Xiao et al., 2012).
The DO concentration, conductivity, and pH of water samples
were determined using a portable meter with multiple probes
(HQ30D, HACH, USA). The DOC content was determined using a
total organic carbon analyzer (Aurora 1030 W TOC, OI, USA) after
sample filtration through a 0.45 m membrane. Ultraviolet
absorbance at 254 nm (UV254) was determined using an ultravi-
oletvisible spectrophotometer (TU1901, Puxi, Beijing, China).
Turbidity was determined using a turbidimeter (2100Q, HACH,
USA).
2. Results and discussion
2.1. Effect of pretreatment on membrane fouling
The feed water for ultrafiltration experiments was collected
from a full-scale DWTP. Because the samples were collected at
4 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX XXX
three stages in the treatment process, each sample type
represented natural water that had been pretreated (relative
to UF) to a different extent. The extent of pretreatment was
the cumulative effect of all unit treatment processes in the
DWTP prior to and including the sample collection point. The
water qualities of samples, including turbidity and DOC
content, are listed in Table 1.
As shown in Table 1, the turbidity, DOC, UV254, and
polysaccharides and proteins (indications of biopolymers) all
decreased as the extent of water treatment increased. Generally,
inorganic particles were more easily removed than organic
compounds, and biopolymers were removed more than other
NOM. It was believed that the organic substances were removed
primarily by physical interception in the sand filter and by
biological degradation and physical interception in the UBAC,
aided by ozonation.
Ultrafiltration of the three feed waters showed that FW-S
(i.e., water with the least pretreatment) fouled the membrane
the most, and FW-F (i.e., the most highly pretreated water)
fouled the membrane the least (Fig. 2). The UF membrane
module was periodically backwashed during each filtration.
The intrinsic membrane resistance was 3.0 3.3 1012 1/m.
After operation for six consecutive filtration cycles, the
irreversible fouling resistance (Rir) increased to 2.6 1012,
2.1 1012 and 0.5 1012 1/m when FW-S, FW-B and FW-F
were filtered, respectively. In the sixth filtration cycle, the
reversible fouling resistance (Rr) that developed was 2.2 1012,
1.9 1012 and 0.5 1012 1/m, respectively. The beneficial
effect of water pretreatment on membrane fouling control
has been extensively reported in the literature (Citulski et al.,
2008; Halle et al., 2009; Peldszus et al., 2011; Zhao et al., 2010).
However, a previous study (Xia et al., 2004) showed that
sand filtration had no additional effect on membrane fouling
control (although the relevant water qualities were not
provided in the study), which was in a sharp contrast to the
significant positive effect observed in this study (Fig. 2). Water
pretreatment reduced both irreversible and reversible fouling.
After operation for six filtration circles, irreversible fouling
became the dominant form of fouling for all three of the feed
waters.
The order of membrane fouling propensity was in accordance
with that of the feed water qualities in terms of turbidity,
DOC content, UV254, polysaccharide and protein concentrations
(Table 1), with higher foulant concentrations generally causing
Table 1 Sources and water qualities of the three types of feed water for ultrafiltration.
Parameters FW-S
(after sedimentation)
Turbidity (NTU) 4.25 0.87
DOC (mg/L) 2.75 0.02
UV254 (1/cm) 0.036 0.001
pH 8.34 0.18
Polysaccharide (mg/L) 1.07 0.07
Peak intensity of protein (AU) 613.5 25.4
UBAC = up-flow biologically active carbon.
Concentration of polysaccharides is expressed as glucose equivalent; protein content is expressed as the arbitrary unit (AU) found using
fluorescence spectroscopy.
Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
faster membrane fouling (Zhang et al., 2012). Both inorganic
particles and organic substances were potential foulants.
Chemical cleaning of fouled membranes at the end of six
filtration cycles was conducted to determine its effect on
irreversible fouling. It was found that cleaning for 30 min with
200 mg/L sodium hypochlorite solution could largely recover
the membrane filterability (Fig. 2), recovering 26.1%, 32.2%,
and 34.4% of fouling layer resistance caused by FW-S, FW-B,
and FW-F, respectively, after backwash at the end of the last
filtration cycle. The effectiveness of sodium hypochlorite
in removing the hydraulically irreversible fouling was also
reported in a number of previous studies (Zhang et al., 2012;
Kimura et al., 2004; Yamamura et al., 2007). Sodium hypo-
chlorite was believed to effectively remove the hydrophilic
organic matter (Strugholtz et al., 2005). These results indicate
that in the present study, membrane fouling was mainly caused
by organic matter. Citric acid at a concentration of 2%, which was
used mainly to remove inorganic scales (Munla et al., 2012), could
remove 8.51%, 7.36%, and 6.90% more of the fouling layer
resistance that remained after cleaning with sodium hypochlo-
rite, respectively. After cleaning by citric acid, the total fouling
resistance was removed by 80.1% to 90.8%, which represented
most of the total fouling resistances.
2.2. Foulant retention on the membrane
Although a higher foulant concentration in the feed water to a
filtration membrane usually corresponds to a higher fouling
rate, membrane fouling occurs due to both the deposition and
accumulation of foulants on the membrane. As such, the
difference in water qualities between the feed water to, and
permeate from, a membrane was used to illustrate fouling.
The retention of biopolymers by the membrane was first
considered, in that a number of previous studies showed that
biopolymers were important in membrane fouling (Subhi et
al., 2012; Halle et al., 2009; Huang et al., 2011; Kennedy et al.,
2005). In the present study a significantly large portion of
polysaccharides were retained on the UF membrane (Fig. 3). The
rejected polysaccharide content was 0.32 0.02, 0.26 0.03 and
0.17 0.02 mg/L when FW-S, FW-B and FW-F samples were
filtered, respectively. The rejected polysaccharides constituted
29.9% 0.34%, 29.3% 1.69% and 25.0% 1.60% of the polysac-
charides contained in the FW-S, FW-B and FW-F feed water,
respectively, and were lower than those reported previously
FW-B FW-F
(after UBAC) (after sand filtration)
2.87 0.42 0.29 0.02
2.36 0.01 2.08 0.01
0.022 0.001 0.020 0.001
8.15 0.01 8.05 0.01
0.87 0.05 0.67 0.03
416.3 17.8 366.9 11.3
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX XXXX 5

9 FW-S, FW-B and FW-F by UF are shown in Table 2. There were no

FW-S FW-B FW-F significant differences between FW-S and FW-B, a remarkable
8
significant difference between FW-B and FW-F, and some
7 Backwished with DI water significant difference between FW-S and FW-F. However,
Rt 1012 (1/m)

all studies, including this one, indicated that only a fraction

6 of the polysaccharides in feed water are retained by the
membrane. In other words, not all polysaccharides in the feed
5 water cause membrane fouling. The organic matter was settled
by coagulation-sedimentation in the sedimentation tank, elim-
4
inated by the metabolism of microorganisms in the BAC (Kim et
3 al., 1997), and intercepted in sand filtration. Therefore, the
Cleaned with NaClO solution polysaccharides in FW-S, FW-B and FW-F that caused severe
2 membrane fouling were different. Nevertheless, our results
0 60 120 180 240 300 360
showed that as more polysaccharides were retained on the
Time (min)
membrane, the extent of membrane fouling was more severe,
although a quantitative relationship between the fouling rate
Fig. 2 Membrane fouling development (as measured by and the retained polysaccharide content was not established.
total resistance, Rt) during six cycles of ultrafiltration of FW-S The retention of proteins by the UF membrane was also
(water after sedimentation), FW-B (water after up-flow investigated. The protein contents in the feed water and
biologically active carbon treatment) and FW-F (water after permeate were calculated from their respective 3D EEM
sand filtration). spectra (Fig. 4). According to the literature (Chen et al., 2003), the
EEM spectra for protein can be divided into five regions
representing aromatic protein I (tyrosine); aromatic protein II;
(Halle et al., 2009; Huang et al., 2011); this discrepancy could be fulvic acid-like substances; substances similar to soluble micro-
due to differences in the feed water qualities in the various bial by-products; and humic acid-like organic substances, respec-
studies. The significant differences of polysaccharide rejection for tively. The excitation and emission wavelength ranges in the EEM
spectra for proteins are 200250 nm and 330380 nm, respective-
1.20 ly. The highest spectral peaks of the FW-S, FW-B and FW-F feed
waters were 615.3, 416.3 and 366.9 (Em: 340 nm, Ex: 235 nm),
Feed Permeate a
1.00 respectively. After ultrafiltration, the highest peaks were 635.0
(Em: 335 nm, Ex: 230 nm), 432.5 (Em: 340 nm, Ex: 235 nm), and
Polysaccharide (mg/L)

379.6 (Em: 335 nm, Ex: 235 nm) for the FW-S, FW-B and FW-F feed
0.80
waters, respectively. A comparison of the fluorescence intensities
(for proteins) in the feed water and permeate showed that there
0.60
was little difference between the two for any of the feed waters,
clearly indicating that very few proteins were retained by the UF
0.40
membrane. Thus, compared with polysaccharides, proteins
contribute much less to membrane fouling. Dissolved organic
0.20
carbon is a gross indicator of the NOM concentration in water.
Results showed that the DOC content removed by the UF
0.00
FW-S FW-B FW-F membrane ranged from 0.194 to 0.271 mg/L, corresponding to a
rejection percentage from 9.4% to 9.9%.
5.00 Feed Permeate In contrast to organic substances, inorganic particles
b (denoted by turbidity) in feed water were almost completely
retained by the UF membrane (Fig. 3). Regardless of the feed
4.00
water turbidity (which was 4.25, 2.87 and 0.29 NTU for the
Turbidity (mNTU)

FW-S, FW-B and FW-F feed waters, respectively), the perme-

3.00 ate turbidity was always less than 0.1 NTU. This high
retention was due to the small nominal pore size (0.01 m)
2.00 of the UF membrane used. Although inorganic particles alone
do not significantly contribute to the overall membrane
1.00 fouling (Jermann et al., 2008; Law et al., 2010; Munla et al.,
2012; Tian et al., 2013), a previous study (Nakatsuka et al.,
1996) showed that membrane fouling increased as feed water
0.00
FW-S FW-B FW-F turbidity increased, and that a linear relationship existed
between the flux decrease and the logarithm of turbidity. A
Fig. 3 Removal of polysaccharides (a) and turbidity (b) of higher fouling rate was also observed for the feed water of
each feed water treated by UF. FW-S (water after sedimen- higher turbidity in the present study. However, this observa-
tation), FW-B (water after up-flow biologically active carbon tion might be a coincidence because the polysaccharide
treatment), and FW-F (water after sand filtration). concentration was also higher when the turbidity was higher.

Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
6 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX XXX

Table 2 The significant differences of polysaccharide spectrum scans of the fouled membranes. The FT-IR spectrum
rejection of FW-S, FW-B and FW-F by UF. can be used to identify the characteristic functional groups of
The pairs of feed water p-Value Significant difference the various foulants on the membrane. Characteristic bands
of polysaccharides, which include those at 30003600, 2927,
FW-S and FW-B 0.546 No
1240 and 1090 cm1, could be found in the spectra of
FW-B and FW-F 0.001 Remarkable
FW-S and FW-F 0.023 Exist membranes fouled by FW-S, FW-B and FW-F (Fig. 5). Polysac-
charides contain a significant number of hydroxyl groups,
which would exhibit a broad, rounded absorption peak above
The relative importance of polysaccharides and inorganic 3000 cm-1 (Howe et al., 2002). The peak at 2925 cm 1 corre-
particles in membrane fouling will be described in Section 2.4. sponds to OH stretching and the peak at 1074 cm1 is CO
stretching of polysaccharides (Lee et al., 2006). However, the
2.3. Fouling layer characterization peak at around 1650 cm 1 corresponds to carboxylate, car-
bonyl, or amide I (Howe et al., 2002), which might indicate the
The predominance of polysaccharides as organic substances existence of proteins, too (Jarusutthirak and Amy, 2006).
in the fouling layer was further confirmed by ATR-FTIR However, as mentioned in Section 2.2, protein cannot be

FW-S FW-S +UF

450 700.0 450 700.0
650.0 650.0
600.0 600.0
400 550.0 400 550.0
500.0 500.0
450.0 450.0
Ex (nm)

Ex (nm)
350 400.0 350 400.0
350.0 350.0
300.0 300.0
300 250.0 300 250.0
200.0 200.0
150.0 150.0
250 100.0 250 100.0
50.00 50.00
0 0
250 300 350 400 450 500 550 600 250 300 350 400 450 500 550 600
Em (nm) Em (nm)
FW-B FW-B +UF
450 700.0 450 700.0
650.0 650.0
600.0 600.0
400 550.0 400 550.0
500.0 500.0
450.0 450.0
Ex (nm)

Ex (nm)

350 400.0 350 400.0

350.0 350.0
300.0 300.0
300 250.0 300 250.0
200.0 200.0
150.0 150.0
250 100.0 250 100.0
50.00 50.00
0 0
250 300 350 400 450 500 550 600 250 300 350 400 450 500 550 600
Em (nm) Em (nm)
FW-F FW-F +UF
450 700.0 450 700.0
650.0 650.0
600.0 600.0
400 550.0 400 550.0
500.0 500.0
450.0 450.0
400.0
Ex (nm)

350 400.0
Ex (nm)

350
350.0 350.0
300.0 300.0
300 250.0 300 250.0
200.0 200.0
150.0 150.0
250 100.0 250 100.0
50.00 50.00
0 0
250 300 350 400 450 500 550 600 250 300 350 400 450 500 550 600
Em (nm) Em (nm)

Fig. 4 Three-dimensional excitationemission matrix (3D EEM) spectra of FW-S (water after sedimentation), FW-B (water after
up-flow biologically active carbon treatment), and FW-F (water after sand filtration) before and after ultrafiltration (UF).

Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX XXXX
100
90
Transmittance (%)
a waters are shown in Fig. 6ac. Images and spectra of the
80
b
c
d Inorganic particles were found on membranes fouled by
70
e FW-S and FW-B waters, and the main chemical elements of
f
60
500 1000 1500 2000 2500 3000 3500
Wavenumber (cm-1)
Fig. 5 Fourier transform infrared spectra of membranes
under different conditions. (Line a) Virgin membrane; (line b)
membrane fouled by FW-S (water after sedimentation) and
backwashed with sodium hypochlorite solution; (line c)
membrane fouled by FW-S and backwashed with ultrapure
water; (line d) membrane fouled by FW-F (water after sand
filtration); (line e) membrane fouled by FW-B (water after
up-flow biologically active carbon treatment); (line f) mem-
brane fouled by FW-S.
substantially retained on the membrane used in the present
study; instead, its presence may be as part of the building
structures of peptidoglycans that exist in the cell walls of the
few diatoms or bacteria that were deposited on the membrane
(Jarusutthirak and Amy, 2006).
The intensity of the transmittance in the spectral band at
30003600 cm1 was weakest for the fouled membranes filtering
FW-S and strongest for those fouled by FW-F samples. This
spectral response was in accordance with the order of fouling
rates shown in Fig. 2 and the retained contents of polysaccharides
shown in Fig. 3. Furthermore, transmittance in the bands at
around 30003600, 2927, 1650, 1240, and 1090 cm1 was increased
by backwashing with ultrapure water compared to the transmit-
tance for fouled membranes, but not totally restored to the value
for virgin membranes. This response demonstrated that poly-
saccharides were at least partially responsible for the hydrauli-
cally reversible fouling. After cleaning with sodium hypochlorite
solution, these bands were further increased, which demonstrat-
ed that this solution was more effective in removing more
polysaccharides than water alone, and indicated that polysac-
charides were also responsible for hydraulically irreversible
fouling. Similar phenomena were observed for the membranes
fouled by FW-B and FW-F waters. These results indicated that
polysaccharides were the main organic foulant and contributed
to both hydraulically reversible and irreversible fouling. A few
previous studies (Gray et al., 2011; Katsoufidou et al., 2010; Peiris
et al., 2010) demonstrated that the fouling caused by polysaccha-
rides was partially reversible and irreversible, while some other
studies (Xiao et al., 2012; Yamamura et al., 2007) regarded
polysaccharides as the major source of irreversible fouling. It is
worth noting that in the present study a noticeable spectral
response appeared at around 1550 cm1 on the membranes
fouled by FW-S and FW-B waters (Fig. 5), but not on the
Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
7
membrane fouled by FW-F, nor on the virgin membrane or any
cleaned membranes.
The fouled, backwashed, and cleaned membranes were
also analyzed using SEM equipped with EDX. The images and
spectra of membranes fouled by FW-S, FW-B, and FW-F
fouled (by FW-S) membranes after being cleaned with
ultrapure water and sodium hypochlorite solution are shown
in Fig. 6d and e, respectively.
these particles were silicon, aluminum, magnesium, sodium,
and calcium (Fig. 6a and b). Silicon is the main component of
4000 clay, which is commonly found in surface water. Silicon could
also be sourced from the few diatoms that were retained on
the membrane. Aluminum was a component in the coagulant
added in the DWTP coagulation process. However, few inorganic
particles were found on the membrane fouled by FW-F (Fig. 6c)
because sand filtration removed most inorganic particles from
the water. Moreover, few inorganic particles were observed on
the membranes after they were cleaned with ultrapure water
(Fig. 6d), indicating that water backwashing was very effective in
removing inorganic particles from the membrane fouling layer.
No inorganic particles were found on the membranes cleaned
with sodium hypochlorite solution (Fig. 6e). It appears that, in
contrast to polysaccharides, inorganic particles in the fouling
layer only contributed to reversible fouling. Peiris et al. (2010) also
found that colloidal and particulate substances were the major
contributors to reversible fouling.
2.4. Effect of increasing turbidity
The foregoing results revealed that both polysaccharides and
inorganic particles were responsible for reversible fouling, while
irreversible fouling was primarily caused by polysaccharides.
Previous studies showed that NOM (e.g. biopolymers) and
inorganic particles could behave synergistically in causing
membrane fouling (Munla et al., 2012; Tian et al., 2013).
However, the present study showed that not all polysaccharides
in the feed water were rejected or retained by the membrane. A
previous study also found that polysaccharides could be
grouped into gelling (or fouling-inducing) and non-gelling
fractions (Wang and Waite, 2009). To further clarify the
respective roles of polysaccharides and inorganic particles in
membrane fouling, kaolinite particles were added to FW-F and
FW-B feed waters to increase their turbidity. After amendment
with kaolinite, FW-F had the lowest fouling propensity and
contained the least fouling-inducing polysaccharide content.
FW-B contained more polysaccharides that could cause mem-
brane fouling than FW-F. The filtration resistance development
after adding kaolinite is shown in Fig. 7.
When the turbidity (after kaolinite addition) in the FW-F
increased to 1.063.05 NTU from its original value of 0.29 NTU,
membrane fouling only slightly increased (Fig. 7). In the third
filtration cycle, the total fouling resistance increased only to
1.0 1012, 1.0 1012 and 1.1 1012 1/m for the feed water with
turbidity at 1.06, 2.09 and 3.05 NTU, respectively. These fouling
resistances were close to those measured for non-amended FW-F
water. In comparison, when the turbidity in FW-B increased from
its original value of 2.87 NTU to 4.076.11 NTU through kaolinite
8 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX XXX

Fig. 6 Elemental analyses and scanning electron microscopy images of membranes in different conditions. (a) Fouled by FW-S
(water after sedimentation); (b) fouled by FW-B (water after up-flow biologically active carbon treatment); (c) fouled by FW-F
(water after sand filtration); (d) fouled by FW-S and backwashed with ultrapure water; and (e) fouled by FW-S and backwashed
with sodium hypochlorite solution.

addition, the membrane fouling rate increased substantially
(Fig. 8). In the third filtration cycle, the total fouling resistance
caused by kaolinite-amended FW-B water increased to 3.3 1012,
3.7 1012, and 3.9 1012 1/m for the feed water with turbidity at
4.07, 5.06, and 6.11 NTU, respectively. These fouling resistances
were 29.9% 55.7% greater than those measured for
non-amended FW-B water. However, this increased fouling
resistance was largely reversible. After operation for three
filtration cycles, the reversible fouling resistances were
1.3 1012, 1.6 1012, and 1.7 1012 1/m, which were 63.2%
112.3% greater than those measured for non-amended FW-B
water, further demonstrating the insignificant role of inorganic
particles in irreversible fouling.
Thus, when feed water to a membrane contains a
relatively high concentration of membrane fouling-inducing
5 4.07 NTU
polysaccharides (as did FW-B), inorganic particles can exac-
erbate reversible membrane fouling, probably through the
synergistic fouling effect observed in a number of previous
studies (Munla et al., 2012; Tian et al., 2013). In the present
study, the polysaccharides acted as the gluing substances
that bound the inorganic particles to form a poorly filterable
cake layer on the membrane. There was probably a critical
ratio of inorganic particle content to polysaccharide content,
above which there would have been either a negligible effect
on, or even a reduction of, the fouling rate; this speculation
needs future verification. The foregoing results also indicated
that even when the inorganic particles were bound with
polysaccharides, the particles could be easily removed from
the fouling layer by backwashing, leaving predominantly
9
2.87 NTU 5.06 NTU
8
6.11 NTU
7
6
Rt 1012 (1/m)

4
5
Rt 1012 (1/m)

3 4
3
2
2
1.06 NTU
1
1 2.09 NTU
3.05 NTU 0
0 30 60 90 120 150 180
0 Time (min)
0 30 60 90 120 150 180
Time (min)
Fig. 8 Membrane fouling development (as measured by
Fig. 7 Membrane fouling development (as measured by total resistance, Rt) during three cycles of ultrafiltration of
total resistance, Rt) during three cycles of ultrafiltration of FW-B (water after up-flow biologically active carbon
FW-F (water after sand filtration) at different levels of treatment) at different levels of artificially induced turbidity
artificially induced turbidity (as measured by NTU). (as measured by NTU).

Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX XXXX
polysaccharides as residuals on the membrane. In compari-
son, when feed water contained few fouling-inducing poly-
saccharides (as did FW-F), the addition of inorganic particles
had a negligible effect on membrane fouling due to the lack
of the gluing effect of polysaccharides. In other words, the
cake layer consisting of inorganic particles and only a few
polysaccharides was relatively filterable, and the fouling
caused by the cake was hydraulically reversible.
3. Conclusions
This study of UF on water collected from different treatment
stages in a full-scale DWTP permits several conclusions.
Biopolymers (polysaccharides and proteins) and inorganic
particles in water contribute to the combined fouling of UF
membranes used in drinking water treatment. The extent to
which these substances foul UF membranes is affected by the
extent of pre-treatment afforded to the membrane feed water;
thus, the point at which UF is used in the drinking water
treatment process is an important operational consideration.
Membrane feed water collected after sedimentation, after BAC
treatment, and after sand filtration fouls UF membranes at
different rates and in different ways. Relative to more highly
treated feed water, water that has been subjected only to
coagulation and sedimentation has the highest biopolymer
concentration and turbidity, and causes the greatest mem-
brane fouling in terms of either hydraulically reversible or
irreversible fouling. In contrast, highly treated UF feed water
(such as that subjected to a combination of coagulation,
sedimentation, ozonation, BAC treatment, and sand filtration)
has little fouling effect on membranes, and the effect is very
slow to develop. Polysaccharides, not proteins, are primarily
responsible for membrane fouling. Polysaccharides are mod-
erately (25.4%29.9% in this study) rejected by UF membranes,
but proteins are negligibly rejected; in addition, the foulant on
membranes contains much more polysaccharides than pro-
teins. UF membranes can effectively remove turbidity (to less
than 0.1 NTU in this study), and inorganic particles can be
observed on membranes that filter water pretreated by only
coagulation and sedimentation as well as by coagulation and
sedimentation combined with ozonation and BAC. Both
polysaccharides and inorganic particles are responsible for
reversible fouling, but irreversible fouling is primarily caused
by polysaccharides, which remain on the membrane after
backwashing. Backwashing membranes with water can effec-
tively remove inorganic particles from the fouling layer. A
synergistic effect of polysaccharides and inorganic particles
on reversible fouling seems to occur only when the concen-
tration of fouling-inducing polysaccharides in the feed
water is sufficiently high. When the concentration of poly-
saccharides is high, increased turbidity in feed water has a
significant effect on the membrane fouling rate; in this study,
turbidity in the range 4.076.11 NTU increased membrane
fouling by over 100% for some samples. In contrast, inorganic
particles contribute little to reversible fouling when the
polysaccharide content of feed water is low; in this study, an
artificially produced increase (to 3.05 NTU) in the turbidity of
highly pretreated membrane feed water did not significantly
increase the fouling rate.
Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
9
Acknowledgments
This work was supported by the China Major Science and
Technology Program for Water Pollution Control and Treatment
(No. 2012ZX07404-002).
REFERENCES
Boudaud, N., Machinal, C., David, F., Bourdonnec, A., Jossent, J.,
Bakanga, F., et al., 2012. Removal of MS2, Q beta and GA
bacteriophages during drinking water treatment at pilot scale.
Water Res. 46 (8), 26512664.
Buffle, J., Wilkinson, K.J., Stoll, S., Filella, M., Zhang, J.W., 1998. A
generalized description of aquatic colloidal interactions: the
three-colloidal component approach. Environ. Sci. Technol. 32
(19), 28872899.
Carroll, T., King, S., Gray, S.R., Bolto, B.A., Booker, N.A., 2000. The
fouling of microfiltration membranes by NOM after coagulation
treatment. Water Res. 34 (11), 28612868.
Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003. Fluorescence
excitation emission matrix regional integration to quantify
spectra for dissolved organic matter. Environ. Sci. Technol. 37 (24),
57015710.
Chen, Y., Dong, B.Z., Gao, N.Y., Fan, J.C., 2007. Effect of coagulation
pretreatment on fouling of an ultrafiltration membrane.
Desalination 204 (1-3), 181188.
Cho, J.W., Amy, G., Pellegrino, J., Yoon, Y.M., 1998. Characterization
of clean and natural organic matter (NOM) fouled NF and UF
membranes, and foulants characterization. Desalination 118
(1-3), 101108.
Cho, J., Amy, G., Pellegrino, J., 2000. Membrane filtration of natural
organic matter: factors and mechanisms affecting rejection
and flux decline with charged ultrafiltration (UF) membrane.
J. Membr. Sci. 164 (1-2), 89110.
Citulski, J., Farahbakhsh, K., Kent, F., Zhou, H.D., 2008. The impact
of in-line coagulant addition on fouling potential of secondary
effluent at a pilot-scale immersed ultrafiltration plant.
J. Membr. Sci. 325 (1), 311318.
Escobar, I.C., Hoek, E.M., Gabelich, C.J., DiGiano, F.A., Le Gouellec,
Y.A., Berube, P., et al., 2005. Committee report: recent advances
and research needs in membrane fouling. J. Am. Water Works
Assoc. 97 (8), 7989.
Geismar, N., Berube, P.R., Barbeau, B., 2012. Variability and limits
of the unified membrane fouling index: application to the
reduction of low-pressure membrane fouling by ozonation and
biofiltration. Desalin. Water Treat. 43 (1-3), 91101.
Gray, S.R., Dow, N., Orbell, J.D., Tran, T., Bolto, B.A., 2011. The
significance of interactions between organic compounds on
low pressure membrane fouling. Water Sci. Technol. 64 (3),
632639.
Guo, J.N., Wang, L.Y., Zhu, J., Zhang, J.G., Sheng, D.Y., Zhang,
X.H., 2013. Highly integrated hybrid process with ceramic
ultrafiltration-membrane for advanced treatment of
drinking water: a pilot study. J. Environ. Sci. Health A 48 (11),
14131419.
Guo, J., Hu, J., Tao, Y., Zhu, J., Zhang, X., 2014. Effect of ozone on
the performance of a hybrid ceramic membrane-biological
activated carbon process. J. Environ. Sci. (China) 26 (4), 783791.
Hagen, K., 1998. Removal of particles, bacteria and parasites with
ultrafiltration for drinking water treatment. Desalination 119
(1-3), 8591.
Halle, C., Huck, P.M., Peldszus, S., Haberkamp, J., Jekel, M., 2009.
Assessing the performance of biological filtration as pretreatment
to low pressure membranes for drinking water. Environ. Sci.
Technol. 43 (10), 38783884.
10 J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX XXX
Han, L.N., Liu, W.J., Chen, M., Zhang, M.L., Liu, S.M., Sun, R.L., et
al., 2013. Comparison of NOM removal and microbial
properties in up-flow/down-flow BAC filter. Water Res. 47 (14),
48614868.
Howe, K.J., Ishida, K.P., Clark, M.M., 2002. Use of ATR/FTIR
spectrometry to study fouling of microfiltration membranes by
natural waters. Desalination 147 (1-3SI), 251255.
Huang, G.C., Meng, F.G., Zheng, X., Wang, Y., Wang, Z.G., Liu, H.J.,
et al., 2011. Biodegradation behavior of natural organic matter
(NOM) in a biological aerated filter (BAF) as a pretreatment for
ultrafiltration (UF) of river water. Appl. Microbiol. Biotechnol.
90 (5), 17951803.
Jarusutthirak, C., Amy, G., 2006. Role of soluble microbial products
(SMP) in membrane fouling and flux decline. Environ. Sci.
Technol. 40 (3), 969974.
Jermann, D., Pronk, W., Boller, M., 2008. Mutual influences
between natural organic matter and inorganic particles and
their combined effect on ultrafiltration membrane fouling.
Environ. Sci. Technol. 42 (24), 91299136.
Katsoufidou, K., Yiantsios, S.G., Karabelas, A.J., 2008. An
experimental study of UF membrane fouling by humic acid
and sodium alginate solutions: the effect of backwashing on
flux recovery. Desalination 220 (1-3), 214227.
Katsoufidou, I.S., Sioutopoulos, D.C., Yiantsios, S.G., Karabelas,
A.J., 2010. UF membrane fouling by mixtures of humic acids
and sodium alginate Fouling mechanisms and reversibility.
Desalination 264 (3SI), 220227.
Kennedy, M.D., Chun, H.K., Yangali, V., Heijman, B., Schippers, J.C.,
2005. Natural organic matter (NOM) fouling of ultrafiltration
membranes: fractionation of NOM in surface water and
characterisation by LC-OCD. Desalination 178 (1-3), 7383.
Kennedy, M.D., Kamanyi, J., Heijman, B.G.J., Amy, G., 2008.
Colloidal organic matter fouling of UF membranes: role of NOM
composition & size. Desalination 220 (1-3), 200213.
Kim, W.H., Nishijima, W., Shoto, E., Okada, M., 1997. Pilot plant
study on ozonation and biological activated carbon process for
drinking water treatment. Water Sci. Technol. 35 (8), 2128.
Kimura, K., Hane, Y., Watanabe, Y., Amy, G., Ohkuma, N., 2004.
Irreversible membrane fouling during ultrafiltration of surface
water. Water Res. 38 (14-15), 34313441.
Kimura, K., Tanaka, K., Watanabe, Y., 2014. Microfiltration of
different surface waters with/without coagulation: clear
correlations between membrane fouling and hydrophilic
biopolymers. Water Res. 49, 434443.
Law, C., Li, X.Y., Li, Q.L., 2010. The combined colloid-organic
fouling on nanofiltration membrane for wastewater treatment
and reuse. Sep. Sci. Technol. 45 (7), 935940.
Lee, N., Amy, G., Crou, J., 2006. Low-pressure membrane (MF/UF)
fouling associated with allochthonous versus autochthonous
natural organic matter. Water Res. 40 (12), 23572368.
Li, Y.H., Zhang, W., Zhang, X.J., Chen, C., Wang, J., 2010.
Characterization of fouling in immersed polyvinylidene
fluoride hollow fibre membrane ultrafiltration by particles and
natural organic matter. Desalin. Water Treat. 18 (1-3), 309314.
Liang, H., Gong, W.J., Li, G.B., 2008. Performance evaluation of
water treatment ultrafiltration pilot plants treating algae-rich
reservoir water. Desalination 221 (1-3), 345350.
Liu, Z., Chu, H.Q., Dong, B.Z., Sun, F., Liu, H., 2011. Pretreatment to
enhance flux of ultrafiltration. Desalin. Water Treat. 30 (1-3),
8088.
Munla, L., Peldszus, S., Huck, P.M., 2012. Reversible and irreversible
fouling of ultrafiltration ceramic membranes by model solutions.
J. Am. Water Works Assoc. 104 (10), 4748.
Please cite this article as: Ao, L., et al., Membrane fouling in ultrafiltration of natural water after pretreatment to different extents,
J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2015.09.008
Nakatsuka, S., Nakate, I., Miyano, T., 1996. Drinking water
treatment by using ultrafiltration hollow fiber membranes.
Desalination 106 (1-3), 5561.
Peiris, R.H., Budman, H., Moresoli, C., Legge, R.L., 2010. Understanding
fouling behaviour of ultrafiltration membrane processes and
natural water using principal component analysis of fluorescence
excitation-emission matrices. J. Membr. Sci. 357 (1-2), 6272.
Peldszus, S., Halle, C., Peiris, R.H., Hamouda, M., Jin, X.H., Legge, R.L.,
et al., 2011. Reversible and irreversible low-pressure membrane
foulants in drinking water treatment: identification by principal
component analysis of fluorescence EEM and mitigation by
biofiltration pretreatment. Water Res. 45 (16), 51615170.
Shao, J.H., Hou, J.A., Song, H.C., 2011. Comparison of humic acid
rejection and flux decline during filtration with negatively
charged and uncharged ultrafiltration membranes. Water Res. 45
(2), 473482.
Song, K.G., Kim, Y., Ahn, K.H., 2008. Effect of coagulant addition on
membrane fouling and nutrient removal in a submerged
membrane bioreactor. Desalination 221 (1-3), 467474.
Strugholtz, S., Sundaramoorthy, K., Panglisch, S., Lerch, A.,
Brugger, A., Gimbel, 2005. Evaluation of the performance of
different chemicals for cleaning capillary membranes.
Desalination 179 (1-3), 191202.
Subhi, N., Leslie, G., Chen, V., Le-Clech, P., 2012. Organic fouling of
ultrafiltration membrane: detailed characterization by liquid
chromatography with organic carbon detector (LC-OCD). Sep.
Sci. Technol. 48 (2), 199207.
Tian, J.Y., Ernst, M., Cui, F.Y., Jekel, M., 2013. Effect of particle size
and concentration on the synergistic UF membrane fouling by
particles and NOM fractions. J. Membr. Sci. 446, 19.
Wang, X.M., Waite, T.D., 2009. Role of gelling soluble and colloidal
microbial products in membrane fouling. Environ. Sci.
Technol. 43 (24), 93419347.
Wang, J., Wang, X., 2006. Ultrafiltration with in-line coagulation
for the removal of natural humic acid and membrane fouling
mechanism. J. Environ. Sci. (China) 18 (5), 880884.
Xia, S.J., Nan, J., Liu, R.P., Li, G.B., 2004. Study of drinking water
treatment by ultrafiltration of surface water and its application
to China. Desalination 170 (1), 4147.
Xiao, P., Xiao, F., Wang, D.S., Qin, T., He, S.P., 2012. Investigation of
organic foulants behavior on hollow-fiber UF membranes in a
drinking water treatment plant. Sep. Purif. Technol. 95,
109117.
Yamamura, H., Kimura, K., Watanabe, Y., 2007. Mechanism
involved in the evolution of physically irreversible fouling in
microfiltration and ultrafiltration membranes used for
drinking water treatment. Environ. Sci. Technol. 41 (19),
67896794.
Yiantsios, S.G., Karabelas, A.J., 1998. The effect of colloid stability
on membrane fouling. Desalination 118 (1-3), 143152.
Yuan, W., Zydney, A.L., 2000. Humic acid fouling during
ultrafiltration. Environ. Sci. Technol. 34 (23), 50435050.
Zhang, Y., Tang, C., Li, G.B., 2012. The role of hydrodynamic
conditions and pH on algal-rich water fouling of ultrafiltration.
Water Res. 46 (15), 47834789.
Zhao, B.Q., Wang, D.S., Li, T., Chow, C., Huang, C., 2010. Influence
of floc structure on coagulation-microfiltration performance:
effect of Al speciation characteristics of PACls. Sep. Purif.
Technol. 72 (1), 2227.