Chemical Engineering Journal 155 (2009) 101107

Contents lists available at ScienceDirect

Chemical Engineering Journal
j our nal homepage: www. el sevi er . com/ l ocat e/ cej
Optimizing the coagulant dose to control membrane fouling in combined
coagulation/ultraltration systems for textile wastewater reclamation
Bae-Bok Lee, Kwang-Ho Choo

, Daeic Chang, Sang-June Choi

Department of Environmental Engineering, Kyungpook National University, 1370 Sankyeok-Dong, Buk-Gu, Daegu 702-701, Republic of Korea
a r t i c l e i n f o
Article history:
Received 12 March 2007
Received in revised form 26 June 2009
Accepted 1 July 2009
Keywords:
Coagulation
Ultraltration
Membrane fouling
Wastewater reclamation
Textile wastewater
a b s t r a c t
Optimal coagulation conditions need to be re-examined when coagulation is coupled to membrane l-
tration for wastewater treatment. This work focused on the optimization of coagulant dosing in order
to control membrane fouling in ultraltration (UF), following coagulation for the reclamation of textile
wastewater. The effects of pore size and coagulant types and dosages on ux decline were investigated
using a stirred-cell UF unit. The ux was greatly enhanced for the UF membrane when a coagulant was
added, whereas for the microltration (MF) membrane the ux decreased. This could be attributed to
changes in the size of coagulated particles and their interaction with membrane pores. At a low dosage
(e.g., 0.0371mMas Al), the polyaluminumchloride (PACl) coagulant was found to control the ux decline
most effectively for low ionic-strength wastewater. The optimal dose minimized the fouling and cake
layer resistances, although it was sharp and dependent on inuent composition. The cake layer protected
the membrane from fouling, but it provided additional resistance to permeation. Analyses of turbidity,
particle size, and membrane surface exhibited the characteristics of coagulated particles and their cake
structures that are closely associated with ux behavior.
2009 Elsevier B.V. All rights reserved.
1. Introduction
Due to the shortage of water and more stringent regulations for
wastewater treatment, water recycling is not optional, but manda-
tory in many countries including Korea. Of the various advanced
treatment systems that have demonstrated wastewater reclama-
tion, membrane separation is one of the most attractive systems
regarding the removal of particles and dissolved contaminants
in wastewater. In particular, loose membrane processes such as
microltration (MF) or ultraltration (UF) can replace conventional
clarication methods such as settling and sand ltration, since they
can be operated at a high ux and/or a low pressure [1,2]. When
a membrane is used alone, however, the ux normally decreases
rapidlyduetomembranefoulingcausedbytheaccumulationof col-
loids and organics at the membrane surface. Thus, there has been
a need for the pretreatment of feed water to control membrane
fouling [3,4].
Several methods prior to MF or UF have been tested with regard
to minimizing membrane fouling, and to enhance membrane
performance, they include coagulation, adsorption, and ozona-
tion [5,6]. When ozonation is used before membrane treatment,
it helped to reduce membrane fouling caused by organic mat-

Corresponding author. Tel.: +82 53 950 7585; fax: +82 53 950 6579.
E-mail address: [email protected] (K.-H. Choo).
ter, since ozone degraded organics that can foul the membranes
[79]. The residual ozone, however, can damage the polymeric
membranes used and thereby, should be destroyed otherwise
ozone-resistant ceramic membranes are employed. Powdered acti-
vated carbon (PAC) adsorption improved the permeate ux in the
hybrid UF or MF membrane processes regarding wastewater reuse
[1012]. PAC is effective in removing organic foulants present in
feedwater, but the spent carbons that should be disposed are
generated. Some studies regarding the combination of MF or UF
with coagulation have also shown that coagulation was effective
in reducing membrane fouling caused by ne particles, as well
as in the removal of particulate contaminants [13]. The pretreat-
ment of coagulation is thus preferred if the wastewater contains
a large amount of ne particles. It has been reported that the
type and dosage of coagulants had an effect on the removal of
particulate contaminants from wastewater [14,15]. Also, the char-
acteristics of coagulated particles, such as surface charges and
size, were affected by the doses of coagulants. This is because the
cake layer formed at the membrane surface had different perme-
ability [16]. In our previous study, the addition of coagulants to
textile wastewater reduced the fouling of UF membranes substan-
tially [14]. The interaction of coagulated particles with membranes,
however, has not been fully understood and the optimization of
coagulant doses has not been established yet in hybrid membrane
systems.
Therefore, the aim of this study was to optimize the coagulant
dose in combined coagulation/UF processes for the reclamation
1385-8947/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cej.2009.07.014
102 B.-B. Lee et al. / Chemical Engineering Journal 155 (2009) 101107
Fig. 1. Flow chart of (a) an existing textile wastewater treatment facility and (b) a proposed membrane-based wastewater reclamation system.
of textile wastewater from a water jet loom. The ultimate goal is
to be able to substitute for the existing treatment facility that is
composed of a series of complicated processes (e.g., otation, sand
ltration, carbon ltration, and diatomaceous ltration) (Fig. 1).
The effects of coagulant types and membrane pore sizes on fouling
control and turbidity removal were also evaluated by monitor-
ing the change of the particle size distributions of wastewater
at different coagulant dosages. The optimal dosage of polyalu-
minum chloride (PACl) and the mechanism of fouling control were
explored.
2. Materials and methods
2.1. Wastewater and coagulants
Textilewastewater samples werecollectedfroma water jet loom
in a textile company in Gyeongsan, Korea and were shipped to our
laboratory and stored at 4

C before use. The wastewater samples

for each experimental run were then equilibrated at room temper-
ature for a couple of hours. The key composition of the wastewater
is as follows: pH, 7.27.4; turbidity, 20.131.0NTU; conductivity,
5462930S/cm; and total organic carbon (TOC), 56mg/L. The
wastewater that had a low ionic strength of 546S/cm was used
if otherwise mentioned. Two different coagulants, such as ferric
chloride (Samchun Chemical, Korea) and PACl (Komax Chemical,
Korea), were used.
2.2. Coagulation tests
Jar tests were carried out to evaluate the coagulation efciency
of PACl. 500mL of textile wastewater was lled in each of six 1-L
beakers, respectively. PACl was added to each beaker in the wide
dosing range of 0.0055mM and the solution pH was adjusted to
7.0 using Na
2
CO
3
. The coagulant dose was based on the coagulation
diagram reported in the literature [17]. The wastewater samples
wererst agitatedvigorouslyat 170rpmfor 1minandsubsequently
B.-B. Lee et al. / Chemical Engineering Journal 155 (2009) 101107 103
mixedat a reducedspeedof 60 and30rpmfor 15 and5min, respec-
tively. After coagulation, the samples were settled for 90min and
then an aliquot of 50mL (supernatant) was sampled from the
clean top zone of the beaker for the measurement of residual
turbidity.
2.3. Membranes and membrane operation
The mixed cellulose ester MF and regenerated cellulose UF
membranes (Millipore, USA) used in this study had pore sizes
of 0.45m, 0.1m, and 100,000Da. An Amicon stirred-cell UF
unit (Model 8200, Millipore, USA) was used for the MF and UF
of coagulated wastewater. A predetermined volume of coagulant
solutions was added into the stirred cell containing 180mL of
wastewater, the mix was agitated vigorously at 200rpm (Reynolds
number =2.0810
4
) for 5min (which corresponds to G and Gt
values of 7120s
1
and 2.1410
6
, respectively and meets the cri-
teria required for rapid mixing), and then, membrane experiments
began under 0.5bar while mixing the solution at 200rpm and
20

C. During MF or UF, the permeate masses were monitored

using an electronic balance in order to check membrane ux and
the data were recorded on a personal computer connected to the
balance.
2.4. Determination of hydraulic ltration resistance during MF/UF
The addition of coagulants, with respect to different levels of
membrane ux decline, was assessed using the resistance-in-series
model [18]. The model is provided in the following equation:
J =
P
(R
m
+R
f
+R
c
)
(1)
where J is the ux for pure water and wastewater during UF
(L/m
2
h), P is the transmembrane pressure (bar), is the dynamic
viscosity of the permeate (Pa s), R
m
is the intrinsic membrane
resistance (m
1
), R
c
is the cake layer resistance caused by parti-
cle deposition (m
1
), and R
f
is the fouling layer resistance caused
by internal pore clogging (m
1
). The initial water ux of a virgin
membrane was used to calculate the R
m
value, whereas the nal
water ux of the membrane was used to calculate the R
f
value.
Before the measurement of nal water uxes, the membrane cell
was emptied, theusedmembranewas placedupsidedownandthen
backwashedat 1.0bar with180mLof deionizedwater toremovethe
cake layer. The physical membrane cleaning was completed with
further water ushing as follows. The membrane was returned to
its proper position and washed with deionized water for 30s while
being agitated at 400rpm. The R
c
value was calculated using the
ux with raw or coagulated wastewater obtained at the end of
membrane ltration. The extrinsic resistance of R
c
plus R
f
, which
is a function of time or permeation volume, was calculated using
the membrane ux with wastewater samples and the initial water
ux. The extrinsic resistance (R
c
+R
f
) can be calculated by subtract-
ing the intrinsic membrane resistance (R
m
) fromthe total ltration
resistance (R
m
+R
c
+R
f
).
2.5. Analytical methods
The turbidity level was measured using a Hach 2100N tur-
bidimeter, whereas the TOC concentration was determined using
an Ionics Sievers 800 TOC analyzer. Particle size distribution mea-
surements were taken on a Beckman Coulter LS 13 320 particle size
analyzer. Membrane surfaces were visualizedusing a ledemission
scanning electron microscope (S-430, Hitachi, Japan).
3. Results and discussion
3.1. The effects of membrane pore size on fouling with the
addition of coagulant
In order to choose a suitable membrane for textile wastewa-
ter treatment with coagulation, the effects of membrane pore size
on relative ux (RF) decline and extrinsic ltration resistance were
evaluated and are shown in Fig. 2. When no coagulant was added to
raw wastewater, a sharp ux decline occurred with the 0.1-m MF
and UF membranes, but a relatively small ux reduction took place
withthe 0.45-mmembrane (Fig. 2a). Accordingly, the extrinsic l-
tration resistance (R
c
+R
f
) increased during MF and UF runs. When
0.0371mM of PACl as Al was dosed to wastewater, the ux pro-
le changed signicantly and the added coagulant was not always
helpful in improving the ux (Fig. 2b). A signicant improvement
in ux was observed with the addition of PACl for the UF mem-
brane, whereas for the MF membranes, the ux showed a severe
decline. The ltration resistance proles varied correspondingly. In
particular, the 0.45-mmembrane showed a further ux decline of
more than 25%. This can be explained by the pore blocking theory
as follows [19]. The ne colloids, which can clog the UF membrane,
increase in size with coagulation, so they may not foul the mem-
brane any more. Some of the coagulated particles, however, are still
Fig. 2. The variation of relative ux and extrinsic ltration resistance for MF and
UF membranes having different pore sizes: (a) without and (b) with dosing PACl
(0.0371mM as Al). Concentration factor is dened as the ratio of retentate volume
to initial volume.
104 B.-B. Lee et al. / Chemical Engineering Journal 155 (2009) 101107
Fig. 3. The variations of inuent wastewater and permeate turbidity with and with-
out coagulation during MF and UF.
within the range of particle sizes that can plug the pores of the
MF membrane. Fig. 3 supports the above scenario. It can be pri-
marily assumed that when no coagulant was added to wastewater,
the turbidity removed was attributable to the rejection of parti-
cles only by the membrane. It was revealed that some ne particles
were still being rejected with a membrane pore size of 0.1m or
smaller. The coagulation process helped achieve further turbid-
ity removals. Especially, the greater turbidity removal by the UF
membrane occurred with coagulation than that of the 0.1-m MF
membrane, although there was no difference in permeate turbidity
between them without coagulation. The results implied that some
particles that had been even smaller than UF pores grew bigger
and rejected. Thus, the 0.1-m MF membrane should have had a
possibility of pore plugging by such coagulated particles leading to
further ux reduction. As a result, of the membranes tested, the UF
membrane was found to be the most improved in terms of fouling
control, along with the addition of coagulants for the hybrid mem-
brane treatment of textile wastewater. Therefore, the UF membrane
was used for the remainder of this study.
3.2. The effects of coagulant types and dosages
To begin with, jar tests were conducted with a range of PACl
doses (0.009270.371mM as Al) to identify the optimum dose
for coagulation (Fig. 4). The turbidity removal efciency increased
with higher PACl doses and nearly leveled off at a PACl dosage of
Fig. 4. Turbidity removal efciency as a function of PACl dosage in coagulation tests.
Fig. 5. The effects of the addition of ferric chloride on permeate ux during UF of
textile wastewater.
approximately 0.1mM as Al. Based on these data, therefore, UF
tests were conducted with different metal coagulant doses, such
as <0.1, 0.10.2, and >0.2mM. Figs. 5 and 6 show the effects of
the types and doses of coagulants on the ux for the UF of tex-
tile wastewater. The coagulants used affected the behavior of ux
decline during UF treatment. Without any addition of coagulants,
the uxes declined sharply by more than 85%. The ux, however,
increased by approximately 20, 60 and 80% when 0.05, 0.179 and
0.5mM of ferric chloride as Fe were added, respectively, whereas
there was no improvement in the ux with a very small dose of
ferric chloride (0.0137mM). Trends in ux decline changed signif-
icantly with the addition of PACl, as shown in Fig. 6. Although a
substantial increase in ux was achieved with 0.0371mMof PACl as
Al, the ux dropped with a larger dose of PACl; the addition of 0.185
and 0.371mM of PACl as Al had a negative effect on ux enhance-
ment, possibly due to the deposition of precipitated coagulants. In
summary, PACl controlled membrane fouling more effectively at a
minimal dose to neutralize charge and/or form bridges between
particles. In contrast, a substantial dose of iron, which belongs to a
dose to generate a sweep oc, was needed with respect to fouling
control. The degree of fouling control, withthe dosing of coagulants
to wastewater, should be associated with the combined effects of
Fig. 6. The effects of the addition of PACl on membrane ux during UF of textile
wastewater.
B.-B. Lee et al. / Chemical Engineering Journal 155 (2009) 101107 105
Table 1
The effects of wastewater particles and coagulants on hydraulic ltration resistance.
Feed Coagulant Dose (mM as Fe or Al) Rm (10
10
m
1
) Rc (10
10
m
1
) R
f
(10
10
m
1
)
Raw wastewater 42.4 39.5 11.0
Raw wastewater PACl 0.0371 42.6 11.9 1.94
Raw wastewater PACl 0.371 43.5 33.0 9.15
Raw wastewater FeCl
3
0.0179 43.3 37.8 10.5
Raw wastewater FeCl
3
0.179 42.8 15.7 1.40
Deionized water PACl 0.0371 43.1 0.324 0.288
Deionized water PACl 0.371 43.3 1.87 7.27
Deionized water FeCl
3
0.0179 42.8 1.92 6.95
Deionized water FeCl
3
0.179 43.2 1.30 0.180
coagulated wastewater particles and precipitated coagulants. As a
result, the dosage of a coagulant should be optimized in order to
minimize membrane fouling, rather than having this achieved by
adding a sufcient amount of a certain coagulant.
The hydraulic ltration resistances were calculated using Eq. (1)
and the values are given in Table 1. All R
m
values were of the same
level, conrming that the membranes tested for each run showed
almost thesamepermeability. For wastewater, R
c
was always higher
than R
f
. This was likely due to the formation of cake layers with
particles present in wastewater. A decrease in R
c
occurred with
the dosing of coagulants to wastewater (most substantially with
0.0371mM of PACl as Al). The decrease in R
f
was also signicant at
a lower dose of PACl and a higher dose of ferric chloride. Overall,
the two coagulants tested showed different effects on ux decline,
depending on their dosages. The larger PACl dosage aggravated the
ux decline thus leading to a further increase in R
c
. A small dose
of ferric salt, however, did not seem to be enough to coagulate
wastewater colloids. Therefore, a larger dose was needed.
To check for any inuence of the coagulants themselves, PACl
and ferric chloride were added to the deionized water and then,
the solution pH was adjusted to 7. For deionized water with a
larger PACl dose of 0.371mM as Al and a smaller ferric chloride
dose of 0.0179mM as Fe, the R
f
contribution was more signi-
cant for the total resistance than that of R
c
. This indicates that the
self-precipitated PACl or ferric hydroxide particles themselves con-
tribute tothe clogging of UF membrane pores. This supports the fact
that the fouling control withthe dosing of coagulants towastewater
is linked to the combined consequence of coagulated wastewa-
ter particles and precipitated coagulants. Thus, it was found that
a small dose of PACl (e.g., 0.0371mM as Al) was most effective in
controllingR
c
andR
f
, as well as minimizingchemical sludgeproduc-
tion, and further investigation regarding the optimal PACl dosage
was needed (which will be discussed in a later section).
3.3. The optimization of PACl doses in terms of fouling control
To determine the optimum dosage of PACl in terms of fouling
control, the PACl dosage varied in the range of 05.0mM as Al and
its effects onux decline were examined, along withthe analyses of
the turbidity levels of supernatant and UF permeate (Figs. 7 and 8).
The supernatant turbidity decreased substantially with higher PACl
dosages, while the UF permeate turbidity remained stable within
a range of 0.070.25NTU not being inuenced signicantly by
the coagulant dosage. This justies dosing a small quantity of the
coagulant where membrane fouling can be minimized as well. As
expected, there existed an optimal PACl dosage with respect to
fouling minimization, but depending on the characteristics of feed
wastewater. For wastewater with a lowionic strength (546S/cm),
the highest ux was obtained at approximately 0.0371mM of
PACl as Al, whereas for wastewater with a high ionic strength
(2390S/cm), theuxwas maximal at a PACl doseof approximately
0.5mM as Al. The optimal PACl dose was shifted to a higher level
with the increased ionic strength. This is possible because parti-
Fig. 7. The variations of the turbidity of supernatant and UF permeate during com-
binedcoagulation/UFtreatment of textilewastewater. Thesupernatant was obtained
after 90-min settling with the addition of PACl.
cles that exist in a high ionic-strength environment are more stable
andthereby their coagulationoccurs withlarger coagulant dosages.
A previous nding demonstrated a similar phenomenon that high
alkalinity increased the coagulant dosage for charge neutralization
of particles [20]. Since the optimal range was sharp and varied in
response to changes in inuent composition, it might be hard to
achieve the optimum in a practical application. Nevertheless, the
aboveresults haveimportant implications that anexcessiveamount
of coagulant should not be added with regard to fouling control as
Fig. 8. The effects of PACl dosage on ux during UF of textile wastewater with
different ionic strengths.
106 B.-B. Lee et al. / Chemical Engineering Journal 155 (2009) 101107
Fig. 9. The size distributions of particles in wastewater with and without the addi-
tion of PACl.
well as turbidity removal. Further studies on the sensitivity and
control of the optimal dose are needed.
It appeared that the size of coagulated particles kept increas-
ing with higher PACl dosages (Fig. 9). As discussed above, however,
some small particles that were not detectable with the particle size
analyses still existed in coagulated wastewater and thereby can
cause membrane fouling during UF (refer to Figs. 2 and 3). Thus,
the coagulation conditions in the combined coagulation/UF system
should not be determined solely based on particle size, as this has
been often used as an indictor in determining optimal coagulation
conditions.
3.4. An analysis of the cake layer formation during the UF of
coagulated wastewater
The UF membrane surface was visualized using a SEM in order
to examine the changes of cake layer structures with coagulation
(Fig. 10). The open pores of the virgin membrane were observed
with SEM (Fig. 10a). When 0.00927mM of PACl as Al was added, a
large portion of the membrane surface was covered with the cake
layer. The degree of membrane surface coverage became severer,
with higher PACl dosages, as shown in Fig. 10bd. In particular, a
thick cake layer that grew on top of the membrane surface was
observed with a large PACl dose of 0.371mMas Al (Fig. 10d). A por-
tion of the top layer of the cake even had shown open pores. This
was possible because the size of the coagulated particles became
bigger with the addition of more PACl (see Fig. 9), but it did not
mean that the cake layer resistance would decrease. As discussed
above, the cake layer resistance increased with larger PACl doses,
due to the formation of thick cake layers. Thus, it was thought
that the cake layer must have had dual functions: it served to pro-
tect the membrane from pore plugging (internal fouling), but it
Fig. 10. SEM pictures of virgin and used membrane surfaces during the UF of wastewater with different PACl dosages: (a) virgin membrane; (b) 0.00927mM as Al; (c)
0.0371mM as Al; and (d) 0.371mM as Al.
B.-B. Lee et al. / Chemical Engineering Journal 155 (2009) 101107 107
also offered additional resistance to permeation. Consequently, this
phenomenon should affect the optimization of coagulant doses in
the combined coagulation/UF system.
4. Conclusions
The additionof ferric andpolyaluminumcoagulants was investi-
gated withregard to controlling membrane fouling during the UF of
textile wastewater generated from the water jet loom process. The
effects of membrane pore sizes and coagulant types and dosages
on membrane permeability were evaluated using a stirred-cell
membrane ltration unit. The examination of hydraulic ltration
resistances, particle size, and cake layer structures was conducted
based on the resistance-in-series model, particle size analysis, and
SEM. The following conclusions can be drawn.
Thedosingof coagulants totextilewastewater exhibitedthebest
control of fouling for the UF membrane, with a MWCO of 10kDa,
among the MF and UF membranes tested. The types and dosages
of coagulants applied affected the ux decline drastically during
the UF treatment of textile wastewater. Flux decline was mitigated
whenthedosageof ferric chloridecoagulant increasedupto0.5mM
as Fe, whereas PACl exacerbated fouling at a larger dosage (e.g.,
>0.1mM as Al for wastewater with a low ionic strength). The opti-
mal PACl dosage needed to minimize the ux decline during UF was
sharp and dependent on the characteristics of inuent wastewater.
The reduction in UF permeability, with the larger PACl dose, was
mainly attributed to the increase of the cake layer resistance. The
growth of cake layers was observed with SEM pictures. The cake
layer, however, must have exhibited dual functions while acting
as a protective barrier for membrane fouling in that it provided
additional resistance to permeation.
Acknowledgements
The authors wish to thank K.C. Kang and Y.J. Park of the
Department of Environmental Engineering, Kyungpook National
University for their assistance in the experiments. Financial sup-
port for Dr. Choos overseas research by the KMS Co., Ltd is greatly
appreciated.
References
[1] S.H. Kim, S.Y. Moon, C.H. Yoon, S.K. Yim, J.W. Cho, Role of coagulation in mem-
brane ltration of wastewater for reuse, Desalination 173 (2005) 301307.
[2] C.C. Teodosiu, M.D. Kennedy, H.A. Van Straten, J.C. Schippers, Evaluation of
secondary renery treatment using ultraltration membranes, Water Res. 33
(1999) 21722180.
[3] S.K. Kang, K.H. Choo, Use of submerged microltration membranes for glass
industry wastewater reclamation: pilot-scale testing and membrane cleaning,
Desalination 189 (2006) 170180.
[4] S. Lee, J.H. Kweon, Y.H. Choi, K.H. Ahn, Effects of occulent aggregates on micro-
ltration with coagulation pretreatment of high turbidity waters, Water Sci.
Technol. 53 (2006) 191197.
[5] D. Abdessemed, G. Nezzal, Tertiary treatment of a secondary efuent by the
coupling of coagulationadsorptionultraltration for reuse, Desalination 175
(2005) 135141.
[6] Y. Zhao, Y. Zhang, W. Zing, N. Xu, Treatment of titanium white waste
acid using ceramic microltration membrane, Chem. Eng. J. 111 (2005) 31
38.
[7] A. Bes-Pi, J.A. Iborra-Clar, M.I. Mendoza-Roca, M.I. Iborra-Clar, Alcaina-
Miranda, Nanoltration of biologically treated textile efuents using ozone as
a pre-treatment, Desalination 167 (2004) 387392.
[8] B.S. Karnik, S.H.R. Davies, K.C. Chen, D.R. Jaglowski, M.J. Baumann, S.J. Masten,
Effects of ozonation on the permeate ux of nanocrystalline ceramic mem-
branes, Water Res. 39 (2005) 728734.
[9] B.S. Oh, H.Y. Jang, T.M. Hwang, J.W. Kang, Role of ozone for reducing fouling due
to pharmaceuticals in MF (microltration) process, J. Membr. Sci. 289 (2007)
178186.
[10] W.S. Guo, S. Vigneswaran, H.H. Ngo, H. Chapman, Experimental investigation
of adsorptionocculationmicroltration hybrid systemin wastewater reuse,
J. Membr. Sci. 242 (2004) 2735.
[11] C.W. Lee, S.D. Bae, S.W. Han, L.S. Kang, Application of ultraltration hybrid
membrane processes for reuse of secondary efuent, Desalination 202 (2007)
239246.
[12] H.K. Oh, S. Takizawa, S. Ohgaki, H. Katayahma, K. Oguma, M.J. Yu, Removal of
organics and viruses using hybrid ceramic MF system without draining PAC,
Desalination 202 (2007) 191198.
[13] S.

Sostar-Turk, I. Petrini c, M. Simoni c, Laundry wastewater treatment using
coagulation and membrane ltration, Resour. Conserv. Recycl. 44 (2005)
185196.
[14] K.H. Choo, S.J. Choi, E.D. Hwang, Effect of coagulant types on textile wastewater
reclamationina combinedcoagulation/ultraltrationsystem, Desalination202
(2007) 262270.
[15] A.L. Ahmad, S. Sumathi, B.H. Hameed, Coagulation of residue oil and suspended
solidinpalmoil mill efuent bychitosan, alumandPAC, Chem. Eng. J. 118(2006)
99105.
[16] J.D. Lee, S.H. Lee, M.H. Jo, P.K. Park, C.H. Lee, J.W. Kwak, Effect of coagulationcon-
ditions on membrane ltration characteristics in coagulationmicroltration
for water treatment, Environ. Sci. Technol. 34 (2000)
37803788.
[17] J. DeWolfe, B. Dempsey, M. Taylor, J. Potter, Guidance Manual for Coagulant
Changeover, Awwa Research Foundation, 2003.
[18] S.K. Kang, K.H. Choo, Use of MF and UF membranes for reclamation of glass
industry wastewater containing colloidal clay and glass particles, J. Membr. Sci.
223 (2003) 89103.
[19] K.H. Choo, C.H. Lee, Effect of anaerobic digestion broth composition on mem-
brane permeability, Water Sci. Technol. 34 (1996) 173179.
[20] C. Ye, D. Wang, B. Shi, J. Yu, J. Qu, M. Edwards, H. Tang, Alkalinity effect of
coagulation with polyaluminum chlorides: role of electrostatic patch, Colloids
Surf. A 294 (2007) 163173.