Journal of Water Process Engineering 19 (2017) 8195

Contents lists available at ScienceDirect

Journal of Water Process Engineering

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

Chemical cleaning of mullite ceramic microltration membranes which are MARK

fouled during oily wastewater treatment

Ehsan Garmsiri, Yaser Rasouli, Mohsen Abbasi , Amir Abbas Izadpanah
Department of Chemical Engineering, Faculty of Oil, Gas and Petrochemical Engineering, Persian Gulf University, P.O. Box 75169-13798, Bushehr, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: In this study, chemical cleaning of home-made mullite ceramic membranes which are fouled during oily was-
Chemical cleaning tewater treatment was investigated. The high performance and low cost mullite membranes were prepared by
Mullite ceramic membrane local low cost kaolin clay using extrusion method and the characteristic of the fabricated membranes was studied
Microltration by dierent methods such as SEM, XRD, mean pore size and porosity analysis. Four types of chemical cleaning
Fouling
agents were selected for chemical in place cleaning of the fouled mullite membrane: Acid (sulfuric acid (H2SO4)),
Oily wastewater
surfactant (sodium dodecyl sulfate (SDS)), chelating agent (ethylene diamine tetra acetic acid (EDTA)) and
alkaline (sodium hydroxide (NaOH)). The fouled membranes were cleaned with single, binary and ternary so-
lution of these chemical agents with the concentrations of 5 mM and 10 mM under the best operating conditions.
After rst cleaning step, membranes cleaned with vinegar and sodium bicarbonate solution as a novel chemical
cleaning agent and named as second cleaning step. Results showed that by using single component chemical
agent, EDTA and SDS with concentration of 5 and 10 mM were the best cleaning agents which have ux recovery
about 31.265% and 57.778% respectively after two steps of cleaning. Binary solution of SDS + EDTA with the
concentration of 5 mM was the best cleaning agent among binary and ternary cleaning solution agents which led
to 41.802% and 65.163% ux recovery in the rst and second cleaning steps of chemical clanging process
respectively.

1. Introduction polarization. Fouling, which is common to all types of membrane se-

Oil, grease and organic emulsions in water are one of the great
pollutants of environment in the world and have traditionally been
great interest for wastewater treatment research due to their high vo-
lume and inherent complexity. This environmental pollution is due to
the release of industrial oily wastewaters from sources such as re-
neries, petrochemical plants, transportation and metallurgy [1,2].
Discharging them without treatment can pollute underground water,
soil and environment. The permitted oil and grease limits for dischar-
ging oily wastewater to environment is 10 mg/L and 20 mg/L for sur-
face and coastal waters instantaneous [3]. Among all the ways for
treating oily wastewaters, membrane techniques can meet these stan-
dards. Among membrane processes, microltration membranes espe-
cially ceramic microltration membranes are widely used for oily
wastewater treatment in the world. Ceramic membranes have many
advantages over polymeric membranes including thermal and chemical
tolerance, resistance to abrasion, mechanical stability and higher per-
meation ux due to the higher porosity [4,5]. Most important issues in
ceramic microltration membranes are fouling and concentration
paration methods, arises from a combination of chemical and physical
interactions [6,7]. Fouling during oily wastewater treatment generally
consist precipitation of oil, grease, solid suspension, colloidal and in-
organic particles on the membrane surface and into its pores [8,9]. Oily
wastewater contains some potential membrane fouling categories, lu-
bricants, cutting liquids, heavy hydrocarbons (tars, crude oils, grease
and diesel oil), and light hydrocarbons (kerosene, jet fuel and gasoline),
microbial (bacteria, viruses, etc.), and inorganic (minerals) contents
[10,11]. Fouling in membranes lead to permeation ux decline, de-
crease in the wastewater treatment eciency, increase in energy and
water consumption, treatment time and operation cost, therefore it is
very important to nd eective and ecient methods in order to con-
trol and minimization of membrane fouling. These methods such as
physical cleaning, chemical cleaning and physical and chemical com-
bined cleaning [11]. In the chemical cleaning process, choosing the best
type of cleaning agent is critical. The optimal selection of the cleaning
agent depends mainly on membrane material and type of foulant. These
agents must be able to dissolve most of the deposited materials on the
surface and remove them from the surface but not damaging membrane

Corresponding author.
E-mail address: m.abbasi@pgu.ac.ir (M. Abbasi).

http://dx.doi.org/10.1016/j.jwpe.2017.07.012
Received 11 March 2017; Received in revised form 14 July 2017; Accepted 15 July 2017
2214-7144/ 2017 Elsevier Ltd. All rights reserved.
E. Garmsiri et al. Journal of Water Process Engineering 19 (2017) 8195

Nomenclature VM Volume of the membrane (m3)

rm Mean pore radius of membrane (m)
EDTA Ethylenediaminetetraacetic acid Porosity of the membrane
SDS Sodium dodecyl sulfate Water viscosity at the operating temperature (Pa s)
TMP Transmembrane pressure (bar) L Membrane thickness (m)
CFV Cross ow velocity (m s1) daverage Average droplet size of oil in the emulsion (m)
T Temperature (C) di The each oil droplet diameter size in the emulsion (m)
t Time (min) n Volume fraction
V Volume of permeate (L) ln The natural logarithm
A Eectiveness area of membrane (m2) J PF at any arbitrary time (L m2 h1)
PF Permeation ux (L m2 h1) J0 PF at initial time (L m2 h1)
FR Flux recovery Jss Steady state PF (L m2 h1)
PFwi Pure water permeation ux of virgin membrane (new Kgl Constant parameter for cake/gel layer formation model
membrane) (L m2 h1) (s m2)
PFww Pure water permeation ux after fouling (after oily was- Ki Constant parameter for intermediate pore blocking model
tewater treatment) (L m2 h1) (m1)
PFwc Pure water permeation ux after chemical cleaning Kc Constant parameter for complete pore blocking model
(L m2 h1) (s1)
W1 Mass of dry membrane (g) Ks Constant parameter for standard pore blocking model
W2 The soaked membranes mass (g) (s0.5 m0.5)
m Water density at the experiments temperature (Kg m3)

surface, thus maintaining membrane properties. The most chemical eciency and was applied for the consecutive chemical cleaning. The
cleaning agents are commercially available, they are often mixtures of recovery eciency of the cleaning in place after rst, second, third and
compounds, and many of them are recommended by membrane man- fourth cleanings was 96.8%, 95.8%, 98.3% and 99.9%, respectively.
ufacturers according to the type of foulant and membrane, although in Chen et al. [19] used various composition of NaOH, NaOCl and sodium
the most cases the actual composition is not clearly specied [12,13]. dodecyl benzene sulfonate (SDBS) for chemical cleaning of ceramic
Anyway, in general Table 1 shows the common cleaning agents which membranes which were fouled by designing wastewater. Results
are used to cleaning membranes. In the recent years, the researchers showed that solution of 1.0 wt% NaOH + 0.1 wt% SDBS was the best
investigated chemical cleaning of dierent membranes, while by chemical solution which leads to 77% ux recovery. Silalahi and
looking up at the available literature; they are not any investigation Leiknes [20] were used dierent commercial products that are biode-
about chemical cleaning of mullite ceramic membranes which are fo- gradable i.e. Ultrasil 115, Ultrasil 73, Surfactron CD 50 and Derquim+
uled by oil and organic pollutants. Ogunbiyi et al. [14] investigated for chemical cleaning of ceramic membranes with nominal pore size
chemical cleaning of tubular ceramic membranes with nominal pore 0.1,0.2 and 0.5 m which were fouled by oil and particulate matter.
size 0.5 m which are fouled by yeast suspension. Chemical cleaning Results showed that at high temperature, the combination of alkaline
consisted of a sequential application of 1% caustic solution through the (Derquim+, Ultrasil 115) and acid (Surfactron CD 50,Ultrasil 73) gave
rig followed by 2% hypochlorite solution and 2% nitric solution, all at a good cleaning eciency except for the 0.5 mm membrane pore size.
50 C. Results indicated that permeation ux values increased with an In this study, mullite ceramic membranes were fabricated by ex-
increase in system pressure and it reduced by enhancing of feed con- trusion method using local low cost kaolin clay. These membranes were
centration. pH eects were also considered and the permeation ux fouled by oil and organic matters during oily wastewater treatment.
values were higher at lower pH values. Avet et al. [15] investigated Chemical cleaning using EDTA, SDS, NaOH and acid sulfuric was per-
chemical and hydraulic cleaning of a tubular ceramic microltration formed and ux recovery using single, binary and ternary solution of
membrane was fouled with a whey protein concentrate suspension. these chemical agents at the optimum conditions was obtained.
They employed a 0.1 m tubular ceramic microltration membrane Therefore, vinegar and sodium bicarbonate were used as a novel
which was fouled by 3.5 wt% whey protein concentrate suspension.
Results showed that cross ow velocity had no signicant eect on ux
recovery but ux recovery was greatly aected by trans-membrane Table 1
Common chemical cleaning agents used for chemical cleaning of ceramic membranes
pressure (TMP). Yin et al. [16] investigated chemical cleaning of [11].
ceramic ultraltration membranes with average pore size of 0.05 m
were fouled by desulfurization wastewater. Results indicates the best Family Examples General functions
eective solution was adding 1% (w/w) NaOH solution mixed with
Acids Strong: HCl, pH regulation, dissolution of inorganic
0.5% (wt%) NaClO, and then cleaning for 120 min at 50 C. Eventually,
HNO3 precipitates, acidic hydrolysis of certain
the ux recovery ratio was always higher than 98%. Ahmad et al. [17] Weak: H3PO4, macromolecules
investigated the water ux recovery following the chemical cleaning of citric
the commercial cellulose acetate membrane with pore size 1.2 m Alkalis Strong: NaOH, pH regulation, alteration of surface charges,
KOH alkaline hydrolysis of proteins, catalysing
which was fouled by micro algal biomass with the dierent chemical
Weak: Na2CO3 saponication of fats.
cleaning agents. Results showed that alkaline cleaning agents more Oxidants NaClO, H2O2 Oxidation of organics; disinfection
eectively removed the foulant layer on the membrane surface than the Surfactants Anionic: SDS Dispersion/suspension of deposits
acidic cleaning agents. In addition, among the tested alkaline agents, Cationic: CTAB
0.75% NaOCl exhibited the best cleaning performance, obtaining ap- Nonionic:
Tween 20
proximately 98% ux recovery. Woo et al. [18] applied oxalic acid and
Chelants EDTA Complexion with metals, removal of mineral
sodium hypochlorite as chemical cleaning agents for cleaning of hollow deposits.
ber ultraltration membranes. The cleaning in series of oxalic acid- Enzyms Proteases, Catalysing lysis of specic substrates (e.g.,
sodium hypochloriteoxalic acid showed the optimal cleaning lipases proteins, lipids)

82
E. Garmsiri et al. Journal of Water Process Engineering 19 (2017) 8195

Table 2 Gachsaran oil-eld, Iran. Physco-chemical analysis of crude oil is shown

Chemical analysis of kaolin that used for fabrication of ceramic membranes. in Table 3. Also distillated water with conductivity less than 5 s cm1
was used.
Component Percent Phases Percent

SiO2 61.62 Kaolinite 64

TiO2 0.4
2.2. Mullite membrane fabrication
Al2O3 2425 Illite 2.4
Fe2O3 0.450.65 Before mixing, massive kaolin samples were triturated by employing
K2O 0.4 Quartz 27 a crushing machine to obtaining ne powders with 200 mesh sieve size.
Na2O 0.5
Low cost mullite ceramic membranes were fabricated by mixing
L.O.I 9.510 Feldespare 6.6
Total 100 100 6772% kaolin clay and 2833% distillate water for 2 h using high
power mechanical mixer. The production cost of this ceramic mem-
brane is less than 70 USD m2. This mixture was extruded by an elec-
cleaning agent for cleaning of ceramic membrane after rst cleaning trical extruder in order to form tubular mullite ceramic membranes.
and called second cleaning. Flux recovery of mullite membranes after Then tubular shaped membranes were dried at room temperature
rst and second cleaning was reported and discussed. within 24 h and calcination was performed by programmed tempera-
ture in an electrical furnace (1200C MINI LAB ELECTRIC, T-Long,
China)at 1100 C for 3 h. Mullite ceramic membranes were 20 cm in
2. Material and methods long, outer diameter was 14 mm and inner diameter was 10 mm and
porosity of them were 42.6%. Temperature gradient from room tem-
2.1. Materials perature to 550 C was 5 C/min. Then temperature was kept x at
550 C for an hour. From 550 C to 975 C, temperature was increased
All chemical materials such as EDTA (C10H16N2O8 > 99.5% with the rate of 5 C/min. Finally, after 1 h that temperature is xed at
purity), SDS (NaC12H25SO4 > 99.5% purity), Sodium hydroxide 975 C, increased by 3 C/min from 975 C to 1100 C. Variation of
(NaOH > 99.9% purity), Tritton X-100 ((C14H22O(C2H4O)n) > 99% temperature with time during calcination of membranes is shown in
purity) and sulfuric acid (H2SO4 > 99.5% purity) supplied from Merck Fig. 1. In order to avoid breaking sintered membranes due to sudden
Company, Germany. Low cost kaolin clay for preparation of mullite change in temperature, membranes were taken out after 8 h cooling in
membranes was supplied from Zenooz mine, Tabriz, Iran. Chemical the furnace. To achieve higher porosity by removal of free silica in
analysis of kaolin clay is listed in Table 2. For preparation of oil-in membrane structure, sintered membranes are washed with strong alkali
water emulsion or oily wastewater, crude oil was purchased from solution (20 wt% NaOH) in an oven with 80 C for 5 h. Finally, ceramic
membranes were washed with enough distillated water in order to re-
Table 3 move remain caustic soda solution from membrane pores. Schematic
Chemical analysis of crude oil that used for oily wastewater treatment.
representation of procedures for mullite membranes preparation is
Specication unit value Specication unit value depicted in Fig. 2.

Sp Gr @15.6 C 0.8750 R.V.P psi 4.83

API Gravity 30.21 Asphaltene Content Wt% 3.60 2.3. Oily wastewater preparation
Sulfur Content Wt.% 1.62 Wax Wt.% 6.06
H2S Wt. ppm < 1.0 D.Me.Point of Wax C 56 Oily wastewater with 1000 mg L1 oil concentration was synthe-
Total Nitrogen Wt/% 0.21 C.C.R Wt.% 5.19
sized in the laboratory by mixing crude oil, distillate water and Tritton
B.S & W Vol.% 0.05 Ash Content Wt.% 0.025
Water Content Vol.% 0.05 Acidity mg KOH/g 0.05 X-100 (0.1 wt% crude oil) as surfactant agent. For this purpose, 10 gr
Salt Content PTB 8 Nickel ppm 29 crude oil, 1 gr Tritton X-100 were added to 1 L distillated water. Then
Kin Vis @10 C cST 28.27 Vanadium ppm 105 this mixture was mixed for 30 min at speed of 19000 rpm using la-
Kin Vis @20 C cST 16.82 Iron ppm 2.6
boratory homogenizer (Wise Mix Homogenizer HG 15). Finally, the
Kin Vis @40 C cST 8.41 Lead ppm <1
Pour Point C 18 Sodium ppm 10
mixture was diluted by adding 9 L distillated water in order to
achieving 1000 mg L1 oil concentration.

Fig. 1. Variation of temperature with time during calcination of

mullite membrane.

83
E. Garmsiri et al. Journal of Water Process Engineering 19 (2017) 8195

Table 4
Experimental steps for each test of chemical cleaning.

step CFV TMP T (C) Time

(m/s) (bar) (min)

First Cleaning step

Distillate water permeation ux from 1.5 1 35 5
fresh membrane (Pwi)
Oily wastewater treatment 1.5 3 35 90
Distillate water ushing 1.5 1 35 5
Distillate water permeation ux (Pww) 1.5 1 35 5
Chemical cleaning 1.25 1 45 20
Distillate water ushing 1.5 1 35 5
Distillate water permeation ux (Pwc) 1.5 1 35 5

Second cleaning step

Cleaning with vinegar & sodium 1.25 1 45 10
bicarbonate
Distillate water ushing 1.5 1 35 5
Distillate water ux 1.5 1 35 5

control the operating conditions such as temperature, cross ow velo-

Fig. 2. Schematic representation of procedures for mullite membrane fabrication.
2.4. Experimental setup & operation
Fig. 3 presents schematic diagram of experimental microltration
setup. In order to control temperature of uids in tanks, a heater and
cooling water was placed into each tank. Setup was designed in order to
city (CFV) and trans-membrane pressure (TMP). The TMP was mon-
itored and controlled by the aid of pressure gage. A rotameter was used
to control the CFV of the passing uid. A centrifugal pump (Pentax,
Italy) with 1 hp power prepared the driving force for microltration
process. The membrane module was fabricated by a Teon tape and
had two O-rings in order to avoid mixing the feed (oily wastewater) and
the permeate ows.
The eect of dierent operating parameters such as TMP
(0.54 bar), CFV (02 m s1) and temperature (1555 C), on the PF,
fouling resistance and total organic hydrocarbon rejection of the mullite
membranes during treatment of the synthetic wastewater were in-
vestigated in previous work [1]. Thus the best operating conditions
(TMP = 3 bar, CFV = 1.5 m s1 and T = 35 C) based on the results of
our previous research were used in the present study. The best oper-
ating conditions in the chemical cleaning step were selected based on
the Salahi and his co-workers [21]. They recommended TMP = 1 bar,
CFV = 1.25 m/s and T = 45 C as the best operating conditions. Each
test consists of 10 steps. Table 4 shows steps and operating conditions
for each step. Distillate water, oily wastewater and chemical agent so-
lution in each step is 10 L. Vinegar and sodium bicarbonate was used as

Fig. 3. Schematic diagram of experimental

setup for cleaning mullite membrane.

84
E. Garmsiri et al.
Fig. 5. Schematic representation of blocking mechanisms: a) complete pore blocking, b)
intermediate blocking, c) standard blocking, d) cake layer formation.
second cleaning agent. Schematic representation of three cleaning steps
is shown in Fig. 4.
Flux recovery and permeation ux during cleaning were calculated
after rst and second cleaning steps.
Permeation ux is given by Eq. (1) as below:
V observed in the microscope.
PF =
A t
Which V is volume of collected permeate (liter), A is membrane
area (m2) and t is time (min). Flux recovery is calculated by Eq. (2)
[22]:
PF PFww
FR (%) = wc
PFwi PFww
In this equation PFwiis refers to initial distillate water ux. In the
other word, virgin membrane ux. PFww is distillate water ux after
fouling by oily wastewater. PFwc is distillate water ux after cleaning
with chemical agent. It must be noted that, during all tests operating
conditions were constant and didnt change.
Oil rejection percent is another important parameter in the se-
paration studies which shows percent of oil that is rejected by the
membrane. Oil rejection is dierence between oil concentrations in
Journal of Water Process Engineering 19 (2017) 8195
Fig. 4. Schematic representation of three cleaning steps.
feed (Cf) and permeates ow (Cp) divided by oil concentrations in feed
(Cf) which is calculated as follows:
Cp
oilrejection = 1 100
Cf (3)
2.5. Calibration curve of the oily wastewater
The emulsion had the maximum UV absorption at 254 nm wave-
length. A Shimadzu UV spectrophotometer UV1800 was utilized to
measure the oil concentration in the permeate ow. According to the
UVvis absorption curve, maximum absorption peak at 254 nm was
obtained. At 254 nm, the related max did not change with oil con-
centration in the emulsion. The linear calibration curve with a coe-
cient of correlation (R2) of 0.9985 was obtained.
2.6. Characterization techniques
The surface and cross section morphologies of virgin and fouled
mullite membranes were studied by scanning electron microscopy
(SEM) using a TESCAN Vega 3 model which is manufactured in the
Czech Republic microscope with the accelerating voltage 20 kV. The
membranes were cut into small pieces and cleaned then mounted on a
stub of metal with adhesive, coated with 4060 nm of Gold and then
Powder X-ray diraction (XRD) was conducted using a PW1800 X-
(1)
ray diractometer (Philips, Netherlands) using Cu K radiation with
tension = 40 kV and current = 20 mA in order to phase identication
in the sintered mullite membranes. Massive membrane samples were
triturated by employing a crushing machine to obtaining ne powders
which were suitable for XRD analysis.
Porosity of membranes was calculated by the simple water mass
(2)
conservation in the membrane soaking in water bath. The soaked in-
itially dried membrane (mass of W1) will uptake the water mass of
(W2 W1) after 24 h soaking in the pure water at room temperature
(W2 is the soaked membranes mass). The volume of uptake water can
be calculated as the(W2 W1)/m. Then the membrane porosity was
calculated as the follows [23,24]:
W W2
porosity = 1 100
M VM (4)
85
E. Garmsiri et al.
W2, W1 and VM are dry and wet masses and volume of the membrane,
respectively, and m is the water density at the experiments tempera-
ture.
The membrane average pore radius (rm) was measured by utilizing
two methods. As First, Guerout-Elford-Ferry equation was studied. For
this purpose, pure water ux for each TMP (0.254 bar) was obtained
for mullite membranes. rm is calculated by plotting the TMP versus pure
water ux and using the slope of the linear obtained equation.
Fig. 6. SEM micrographs of ceramic membrane (a) virgin membrane and (b) surface of fouled membrane (c) fouling layer which is formed on the membrane surface (d) cross section of
fouled membrane (f) membrane surface after chemical cleaning.
86
Journal of Water Process Engineering 19 (2017) 8195
(2.9 1.75 ) 8l PF
rm =
TMP (5)
where is the membrane porosity (%), is the water viscosity
(8.9 104 Pa s) at the operating temperature, l is the membrane
thickness (m) and PF is the pure water ux under the specic applied
TMP [25,26]. Second, the membrane average pore diameter was mea-
sured using Image J software (Version 1.44) by SEM analysis [27,28].
The membrane average pore diameter from SEM analysis was
E. Garmsiri et al. Journal of Water Process Engineering 19 (2017) 8195

The three-point bend test with ASTM C1505-01 was selected to

measure membrane mechanical strength based on SANTAM, STM-150
universal testing device with 150 kN capacity, Total grips distance
690 mm, dimensions (Width Depth Height) 1070 700 2430
and 200 V 5060 Hz power [29]. SANTAM universal testing machine is
used for industrial quality control laboratories and research centers.
The machine is able to test the mechanical properties of a wide range of
the materials such as metals and ceramics. Larger membrane was cut
into three smaller pieces with 5 cm length, inner and outer diameter
4 mm, then mechanical strength is measured using testing machine and
mean value of them are reported.
It must be noted that, the home-made mullite membranes have a
good resistance to chemical corrosion and can be employed in pH range
of 114 based on their good chemical stability and these membranes
have a long life time.
The size of oil emulsion droplets in the feed was measured by a
dynamic light scattering analyzer device (Microtrac, MN402-NS-0000-
0000-000-4M, USA). The emulsion sample (1 mL) was analyzed im-
mediately after preparation. The device measures particles ranging in
the size from 0.3 to 10,000 nm.

3. Results and discussion

3.1. Modeling

Hermias models are useful tools for description of fouling me-

chanism in cross ow MF process. These models are the most useful and
applicable models for MF ux decline prediction. Hermia [30] was
developed four equations for dead-end ltration that describe fouling
mechanisms. These models were cake layer formation model, inter-
mediate pore blocking model, standard pore blocking model and
complete pore blocking model. Schematic of these fouling models are
shown in Fig. 5. After that, these models were modied for cross ow
MF and the main equation is given by Eq. (7) as follows [31,32]:
dj
= K (j jss ) J 2 m
dt (7)
where m = 2 for complete pore blocking model; m = 1.5 for standard
pore blocking model; m = 1 for incomplete pore blocking model (in-
termediate pore blocking model) and m = 0 for cake layer formation
model. K is the model constant, Jss and J is steady state PF and PF at any
arbitrary time respectively.

3.1.1. Cake layer formation model

In the cake layer formation model, oil droplets are larger than
membrane pore sizes and do not enter into pores of membrane. They
form gel layer on the membrane surface and causes ux decline. Cake
layer formation model is given by below equation [33]:

1 J J0 J 1 1
t = 2
ln Jss

k J J J J J J
0 (8)
gl ss 0 ss

where J0 is the initial PF. The parameter kgl represents a ratio between
the characteristics of the gel layer and those of the un-foulded mem-
brane [31].
Fig. 6. (continued)
3.1.2. Intermediate pore blocking model
calculated using the following equation: Intermediate pore blocking model occurs when the oil droplet sizes
is similar to membrane pore size. Thus oil droplets obstruct membrane
n 0.5
n d 2 pore entrance without block them completely [31]. The resulting
i i
equation is Eq. (9) as follows:
daverage = i = 1n
J0 Jss e ki Jss t
n i J=
i = 1 (6) Jss + J0 (e ki Jss t 1) (9)
In this equation n is the number of pore, d average is the membrane the parameter Ki represents as well the membrane surface blocked per
average pore diameter (m). unit of total volume permeated through the membrane and unit of

87
E. Garmsiri et al. Journal of Water Process Engineering 19 (2017) 8195

Fig. 7. XRD pattern of mullite ceramic membrane.

Table 5 membrane pore size. It is assumed that the membrane pores have
Porosity, mean pore diameter and mechanical strength of home-made mullite mem- constant length and size and decreasing in pore volume is equal to
branes.
decreasing in their cross section area. The back transport diusion of
porosity Mean pore diameter using Mean pore diameter Mechanical solute molecules from the membrane surface to the bulk feed solution
Guerout-Elford-Ferry using Image J strength does not occur [31]. These assumptions are lead to Eq. (10):
equation
J0
J=
42.6% 0.451 m 0.867 18.3 MPa (J0 + J00.5 Ks t )2 (10)

where Ks is the standard pore blocking model constant.

3.1.4. Complete pore blocking model

Complete pore blocking model occurs when the size of the oil dro-
plets are greater than the size of the membrane pores. Therefore, pore
blocking takes place over the membrane surface and not inside the
membrane pores [34,35]. This model is given by Eq. (11):

J = Jss + (J0 Jss ) ekc J0 t (11)

The parameter Kc is equal to Ki in intermediate pore blocking model.

To determine whether the experimental data agree with any of
considered model data, total errors between experimental and model
data for each model were calculated by Eq. (12) as follows:

Jmod Jexp
Jexp
E=
Fig. 8. Droplet size distribution of synthetic oily wastewater. n (12)

where parameter n is the number of experimental point, Jmod is model

initial membrane surface porosity [31]. Ki can be obtained by linear-
predicted PF and Jexp is experimental PF data.
ization of Eq. (9) and tting linear equation on the experimental data.
Standard deviation () over the whole ltration time was calculated
as follows [36]:
3.1.3. Standard pore blocking model (Jmod Jexp)2
In the standard pore blocking model, oil droplets enter in the =
n1 (13)
membrane pores and deposit on the pore walls thus pore volumes re-
duce. This model occurs when the molecule size is smaller than Coecient of correlation (R ) and for each models were calculated
2

88
E. Garmsiri et al.
Fig. 9. Flux recovery percent by using a) SDS with concentration of 5 mM & 10 mM b) EDTA with concentration of 5 mM & 10 mM c) Sulfuric acid with concentration of 5 mM d) NaOH
with concentration of 5 mM & 10 mM.
and compared to other models.
3.2. Characterization results
SEM micrograph of virgin and fouled membranes by oily waste-
water is shown in Fig. 6. As micrograph shows, after treatment of oily
wastewater by mullite ceramic membranes, fouling layer of oil is
formed on membranes surface and pores ll by oil droplets. This re-
versible fouling layer is cleaned by distillated water ushing while
pores of membrane are cleaned by using chemical cleaning agents. Ir-
reversible fouling cannot be removed either with water ushing or with
a chemical cleaning agent. SEM observation showed that surface mor-
phology of membrane is porous and rough and membrane surface did
not have any defects in their structure.
XRD analysis pattern is carried out for (i) raw kaolin clay as starting
material and (ii) sintered membrane which are shown in Fig. 7. For raw
kaolin clay, kaolinite and quartz is the major phases and illite and
feldespare are the minor phases. In sintering temperature of 1050 C,
kaolin under this temperature is transformed to mullite, quartz and
cristobalite phase. Quartz and cristobalite phases can be washed by
strong alkali solution for increasing porosity of membrane.
3(Al2O32SiO2) 3Al2O32SiO2 (mullite) + 4SiO2 (quartz, cristobalite)
(14)
As XRD pattern shows, in mullite membranes, major phases are
quartz and mullite and minor phase is cristobalite. Other properties of
home-made mullite membrane such as porosity, mean pore size and
mechanical strength are shown in Table 5. As this table shows, mean
pore size which is calculated using Image J software is 0.867 m that is
in better agreement with SEM analysis than Guerout-Elford-Ferry
equation.
As observed in Fig. 8, average droplet size of synthetic oily is
89
Journal of Water Process Engineering 19 (2017) 8195
112 nm and this is small enough to remain emulsion stable for 12 h for
tests.
3.3. Selection of best single component chemical solution agent
3.3.1. Sodium dodecyl sulfate (SDS)
As shown in Fig. 9(a), by using SDS with 5 mM concentration, ux
recovery in the rst and second cleaning steps were 20.528% and
25.872% respectively. Increasing SDS concentration up to 10 mM in the
solution was lead to enhancing the ux recovery up to 20.602% and
57.778% in the rst and second cleaning steps respectively. SDS
cleaning eect can be attributed to cleaning strength of emulsiers and
overcome to surface tension of water [37]. The presence of anionic
surfactant molecules on the membrane surface after the cleaning could
also have increased repulsion between the membrane surface and the
negatively charged organic foulants in the oily wastewater [38]. Sur-
factants such as SDS reduce surface tension of in-contact molecules and
increases hydrophilic strength of membrane. Surfactants have both
hydrophilic and hydrophobic groups. They diuse in water and ad-
sorbed at interface between foulants and water [11]. Hydrophilic
groups adsorb water and hydrophobic groups adsorb foulants. In the
other word, bridging between foulants and water cause dissolving
foulants into water that lead to membrane is to be cleaned [39].
Fig. 10(a) shows permeation ux of mullite membrane during rst
chemical cleaning by using SDS with concentration of 5 and 10 mM as
cleaning agents. As this gure presents, by enhancing concentration of
SDS from 5 mM to 10 mM in the cleaning solution, descending trend
was observed in the ux recovery. The reason for this subject is that
higher concentration of SDS can act as foulant materials because extra
and unreacted SDS with foulants, re-deposit on membrane surface and
lead to pore blocking. Therefore, permeation ux during rst cleaning
is decreased by increasing SDS concentration from 5 to 10 mM. In the
E. Garmsiri et al.
Fig. 10. Permeation ux variation during rst cleaning by using a) SDS with concentration of 5 mM & 10 mM b) EDTA with concentration of 5 mM & 10 mM c) sulfuric acid with
concentration of 5 mM d) NaOH with concentration of 5 mM & 10 mM.
Fig. 11. Flux recovery percent by using dierent chemical agent.
rst 5 min of cleaning process, permeation ux is enhanced with time
due to chemical cleaning of membrane, while in the nal times of
cleaning process; permeation ux is almost xed and steady.
3.3.2. Ethylene diamine tetra acetic acid (EDTA)
Mullite ceramic membranes also cleaned using EDTA solution with
10 mM and 5 mM concentration under optimum conditions. As shown
in Fig. 9(b) ux recovery in the rst and second cleaning step was
31.265% and 43.481% respectively. Increasing concentration of EDTA
up to 10 mM was resulted to reduction of ux recovery to 2.364% in
rst cleaning step and 21.738% in the second cleaning step.
Fig. 10(b) also conrms these results. By using chemical solution of
EDTA with concentration of 10 mM, permeation ux is lower than
90
Journal of Water Process Engineering 19 (2017) 8195
EDTA with concentration of 5 mM which causes decrease in the ux
recovery percent. EDTA has six positions to bond with other material in
its own structure. In the other word it is a six dentate ligand. Therefore,
we expect EDTA form complex with foulants and separate them from
membrane surface. Chelants or sequestrates such as EDTA form strong
complexes with multivalent metal ions such as calcium and dispersal
minerals in general [4042].
3.3.3. Sulfuric acid
Acids are mainly used to dissolve precipitates of inorganic salt sand
metal oxides or hydroxides [43]. Mineral acids such as sulfuric acid
(H2SO4) can perform eciently and eectively for membrane washing
in order to increasing the permeation ux. Acid can separate mineral
foulant such as calcium from membrane surface [44]. They are also
have extremely eective for removing of metal cations on the fouled
members surfaces. However, they have relatively high cost and may
also be a concern that some reaction products of acids such as HNO3
and H3PO4 are of limited solubility, which may cause re-deposition on
the membrane surface [45]. Fig. 9(c) shows ux recovery using acid
sulfuric as chemical cleaning agent. By applying 5 mM concentration of
acid sulfuric in the cleaning solution, ux recovery was 17.867% and
32.232% at the rst and second cleaning steps respectively. As shown in
Fig. 10, permeation ux during rst cleaning of mullite membrane
using sulfuric acid with 5 mM concentration is increased from
185 L m2 h1 at the initial time of cleaning to 232 L m2 h1at the
nal time of cleaning.
3.3.4. Sodium hydroxide (NaOH)
Alkalis, such as NaOH, KOH or their mixtures have ecient neu-
tralization eect of acidic organics; decreasing the number of bonds
between the foulant and the membrane surface and increasing mass
E. Garmsiri et al.
Fig. 12. Flux recovery percent by using a) SDS + EDTA with concentration of 5 mM b) SDS + NaOH with concentration of 5 mM c) EDTA + NaOH + SDS with concentration of
5 mM & 10 mM d) sulfuric acid + EDTA with concentration of 5 mM.
transfer of cleaning agents to the membrane surface; saponication of
fats and oils and hence solubilization of the cleaning products; disper-
sion/emulsication of colloidal material; and regulation of pH to the
eective working condition for other chemical cleaners [41,44,46,47].
A concern with these hydroxide solutions is their lack of a buering
capacity (or hydroxide reserve). Therefore, most cleaning must begin
with high pH levels of 1112 in order to suciently neutralize all the
acidic components and saponify all fats [42,44]. Flux recovery after
rst and second cleaning steps was obtained 13.328% and 22.574%
using sodium hydroxide solution with 5 mM concentration. When
concentration of Sodium hydroxide solution was increased from 5 to
10 mM, the ux recovery was enhanced from 10.433% in rst cleaning
step to 41.955% in second cleaning steps as illustrated in Fig. 9(d).
Fig. 10(d) shows permeation ux of mullite membrane during
chemical cleaning using NaOH with 5 and 10 mM concentration in the
cleaning solution. Permeation ux of mullite membrane during
cleaning with 10 m MNaOH is higher than NaOH with 5 mM con-
centration. By using NaOH with 5 and 10 mM concentration, permea-
tion ux was increased from 145 to 172 L m2 h1 and 121 to
132 L m2 h1 respectively.
For summary of this part, as shown in Fig. 11, EDTA with 5 mM and
SDS with 10 mM concentration were the best chemical agents that had
the most ux recovery in rst and second cleaning steps. In all ex-
perimental tests, ux recovery in the second cleaning step was higher
than rst cleaning. This observation conrms eectiveness of vinegar
and sodium bicarbonate as a novel cleaning agent for chemical cleaning
of ceramic membrane that fouled by oil particles.
91
Journal of Water Process Engineering 19 (2017) 8195
3.4. Selection of best binary and ternary chemical solution agents
3.4.1. Binary solution of SDS & EDTA
As shown in Fig. 12(a), ux recovery using binary solution of SDS
and EDTA with concentration of 5 mM was 41.802% and 65.163% in
the rst and second cleaning steps respectively. Since SDS has both
hydrophilic and hydrophobic groups, it is miscible in the both organic
and aqueous solution. Surfactants dissolve large molecules by forming
micelle around them and separate foulants from membrane surface.
They enhances surface wettability causing prevent re-deposition of
fouling materials on the membrane surface. Also, EDTA removes di-
valent cations from complex organic molecules and increase ux re-
covery eciency. Generally, SDS remove oil and grease and EDTA re-
move inorganic materials from membrane surface [8,43,48,49].
Fig. 13(a) shows permeation ux during cleaning of mullite ceramic
membrane using binary solution of SDS+EDTA with 5 mM con-
centration. As this gure presents, permeation ux was increased ra-
pidly in 12 min of cleaning time and then almost steady. Permeation
ux during rst cleaning of membrane using SDS + EDTA with con-
centration of 5 mM was increased from 205 L m2 h1 to
240 L m2 h1.
3.4.2. Binary solution of SDS & NaOH
Fig. 12(b) presents the ux recovery using binary solution of SDS
+ NaOH with the concentration of 5 mM. Flux recovery in the rst
cleaning step was obtained 22.84% and in the second cleaning step was
30.26%. This result may be attributed to the synergistic eect of NaOH
and SDS on the foulants which sodium hydroxide changes pH of
E. Garmsiri et al.
Fig. 13. Permeation ux variation during rst cleaning by using a) SDS + EDTA 5 mM b) SDS + NaOH 5 mM c) EDTA + SDS + NaOH 5 mM & 10 mM d) acid+EDTA with con-
centration of 5 mM.
solution and causes both membrane surface and foulant material
charges become negatively and lead to repulsion between foulants and
membrane surface [19]. By increasing cleaning time as shown in
Fig. 13(b), permeation ux of mullite membrane was enhanced from
131 L m2 h1 to 156 L m2 h1.
3.4.3. Ternary solution of SDS, NaOH and EDTA
By using ternary solution of SDS + EDTA +NaOH with the con-
centration of 5 mM as cleaning agent, ux recovery was 35.895% and
49.322% in the rst and second cleaning steps respectively. Enhancing
the concentration of cleaning agent up to 10 mM, was resulted to ux
recovery decrease in the rst and second cleaning steps to 14.318% and
19.019% respectively according to Fig. 12(c). Fig. 13(c) conrms these
observations about ux recovery which permeation ux of mullite
membrane by using ternary solution of these chemical agents with
concentration of 5 mM is higher than 10 mM concentration, therefore,
ux recovery decreases by using 10 mM solution. NaOH changes pH of
the solution and provides a better condition for the highest removal of
foulants using EDTA and SDS [22].
3.4.4. Binary solution of sulfuric acid & EDTA
As shown in Fig. 12(d),by using binary solution of sulfuric acid and
EDTA with concentration of 5 mM as chemical cleaning agent, negative
ux recovery in rst cleaning step was obtained. This reason for this
phenomenon is that by employing this chemical cleaning agent, formed
gel layer on fouled membrane surface becomes denser under acidic
condition and removing them from the membrane surface become
harder [8]. In the second cleaning step, ux recovery was obtained
1.291%. As shown in Fig. 13(d), unlike prior cleaning agents,
92
Journal of Water Process Engineering 19 (2017) 8195
permeation ux of mullite ceramic membranes was decreased by in-
creasing cleaning time. As mentioned before, denser gel layer was
caused to decreasing in the permeation ux of membranes.
3.4.5. Binary solution of EDTA & NaOH
Flux recovery was reported 6.106% and 46.035% in the rst and
second cleaning steps by using binary solution of EDTA+ NaOH with
5 mM concentration as chemical cleaning agent. By increasing con-
centration of this cleaning agent up to 10 mM, ux recovery was 5.73%
and 9.805% in the rst and second cleaning steps as shown in Fig. 14.
Fig. 15 shows permeation ux of mullite membranes during
cleaning with EDTA + NaOH 5 mM and 10 mM as cleaning agents.
Permeation ux of membranes was increased by increasing
cleaning time. By using EDTA + NaOH with concentration of 5 mM and
10 mM, permeation ux of mullite membranes was increased
from 247 L m2 h1 to 270 L m2 h1 and 195 L m2 h1 to
287 L m2 h1respectively.
Fig. 16 shows ux recovery percent by using binary and ternary
solutions of dierent chemical cleaning agents. Binary solution of SDS
+ EDTA with the concentration of 5 mM was the best chemical agent
for chemical cleaning of mullite membrane which is fouled by oil and
organic matters. Ternary solution of SDS + EDTA + NaOH by 5 mM
concentration also showed moderate ux recovery while binary solu-
tion of sulfuric acid + EDTA with 5 mM concentration showed very
weak ux recovery. Similar results also reported in the literature [22].
3.5. Permeation ux & oil rejection during ltration
Fig. 17 presents variation of permeation ux of mullite membrane
E. Garmsiri et al.
Fig. 14. Flux recovery percent by using EDTA + NaOH with concentration of
5 mM & 10 mM.
and oil rejection percent with time during oily wastewater ltration. As
the curves in this gure show, fouling of membrane surface and lling
of membrane pores are very rapid in the rst 15 min of ltration pro-
cess. After that, descending rate of permeation ux decreased because
of lling smaller pores which have the least eect on permeation ux.
Initial and the last permeation ux of mullite during ltration of oily
wastewater are 650 L m2 h1 and 432 L m2 h1 respectively. In
fact, after 90 min ltration, permeation ux decreased about 33%. Also,
due to the fouling of membrane, oil rejection percent increases slightly
from 98.5 to 99.6%. in fact, cake layer which is formed on the mem-
brane surface, act as an additional resistance hence oil droplets cannot
pass through the membrane pores and reject by the membrane conse-
quently oil rejection percent is enhanced.
3.6. Permeation ux modeling
Hermia's model for cross ow microltration was employed in order
to nd out the fouling mechanism of mullite membrane during oily
wastewater ltration. As Fig. 18 shows, cake layer formation model is
the best mechanism for description of membrane fouling with the least
error among other models. Average error and standard deviation be-
tween experimental data and cake formation layer model are 4.49%
and 4.81 (Table 6). Intermediate pore blocking model is also showed
acceptable agreement with experimental permeation ux data with the
average error and standard deviation of 11.55% and 12.56 respectively.
Cake layer which is formed on the membrane surface is a kind of re-
versible fouling and can be cleaned by mechanical cleaning such as
water ushing, back washing and other physical methods while pore
blocking fouling is mostly irreversible and can be cleaned partially by
using chemical cleaning agents under specic conditions.
Fig. 15. Permeation ux variation during rst cleaning by using EDTA + NaOH with
concentration of 5 & 10 mM.
93
Journal of Water Process Engineering 19 (2017) 8195
Fig. 16. Flux recovery percent by using dierent chemical agent.
Finally, for knowledge of statistical analysis of the experimental
data, it must be noted that at least, one test for fouling reduction by
each chemical cleaning agent was repeated. Based on carefully con-
trolling of experimental conditions in all testes, the repeated results
showed that average error for accuracy and repeatability of the results
for ux recovery of each chemical cleaning was lower than 4%. This
error is obviously due to a human error in preparation of home-made
ceramic membranes and during wastewater treatment and chemical
cleaning process.
3.7. Cleaning mechanism
As mentioned before, cleaning procedure is consist of three steps
include (i) forward water ushing (ii) cleaning with chemical agents
and (iii) cleaning by using vinegar and bicarbonate sodium. Forward
ushing consists in pumping permeate water at high cross ow velocity
through the feed side in order to remove foulants from the membrane
surface. Because of more rapid ow and the resulting turbulence, par-
ticles absorbed to the membrane are released and discharged. When the
cleaning agents come into contact with the fouled layer, some physical
transformations and chemical reactions take place. Physical
Fig. 17. Permeation ux and oil rejection percent of mullite membrane during ltration.
E. Garmsiri et al. Journal of Water Process Engineering 19 (2017) 8195

chemical cleaning agent and were named as second cleaning step.

These main results obtained in this research:

1. Among the single component chemical cleaning agents, SDS with

concentration of 10 mM utilize 57.78% ux recovery was re-
commended as the best cleaning agents which are lead to highest
ux recovery in the chemical cleaning of mullite membrane which is
fouled by oil and organic matters.
2. Among binary and ternary cleaning agents, binary solution of SDS
+ EDTA with concentration of 5 mM was the best cleaning agent
with the ux recovery of 41.802% and 65.163% in the rst and
second cleaning steps respectively. Also, ternary solution of SDS
+ NaOH + EDTA with the concentration of 5 mM as cleaning agent
had the ne ux recovery about 50%. In addition, sulfuric acid was
the weakest chemical cleaning agent which was led to very low and
negative ux recovery.

References

[1] M. Abbasi, M. Mirfendereski, M. Nikbakht, M. Golshenas, T. Mohammadi,

Performance study of mullite and mullitealumina ceramic MF membranes for oily
Fig. 18. Permeation ux modeling of mullite membrane during oily wastewater ltration. wastewaters treatment, Desalination 259 (2010) 169178.
[2] K. Suresh, G. Pugazhenthi, R. Uppaluri, Fly ash based ceramic microltration
membranes for oil-water emulsion treatment: parametric optimization using re-
Table 6 sponse surface methodology, J. Water Process Eng. 13 (2016) 2743.
Average errors and standard deviations of permeation ux modeling of mullite mem- [3] J.M. Ne, Bioaccumulation in Marine Organisms: Eect of Contaminants from Oil
Well Produced Water, Elsevier, 2002.
brane.
[4] S.R.H. Abadi, M.R. Sebzari, M. Hemati, F. Rekabdar, T. Mohammadi, Ceramic
membrane performance in microltration of oily wastewater, Desalination 265
model Cake layer Intermediate Standard Complete
(2011) 222228.
formation pore blocking pore pore [5] R.V. Kumar, L. Goswami, K. Pakshirajan, G. Pugazhenthi, Dairy wastewater treat-
model blocking blocking ment using a novel low cost tubular ceramic membrane and membrane fouling
mechanism using pore blocking models, J. Water Process Eng. 13 (2016) 168175.
Average error 4.49 11.55 13.73 23.19 [6] G. Pearce, Introduction to membranes: fouling control, Filtr. Sep. 44 (2007) 3032.
(%E) [7] G. Blandin, H. Vervoort, P. Le-Clech, A.R.D. Verliefde, Fouling and cleaning of high
Standard 4.81 12.56 13.49 22.61 permeability forward osmosis membranes, J. Water Process Eng. 9 (2016) 161169.
deviation [8] A. Salahi, T. Mohammadi, M. Abbasi, F. Rekabdar, Chemical cleaning of utaltra-
() tion membrane after treatment of oily wastewater, Iran. J. Chem. Eng. 7 (2010).
[9] L. Susan, S. Ismail, B. Ooi, H. Mustapa, Surface morphology of pvdf membrane and
its fouling phenomenon by crude oil emulsion, J. Water Process Eng. 15 (2017)
5561.
transformations include mechanical stress, wetting and soaking of fo- [10] S. Madaeni, S. Samieirad, Chemical cleaning of reverse osmosis membrane fouled
uled layer. Chemical reactions involved are hydrolisis, peptization, by wastewater, Desalination 257 (2010) 8086.
solubilization, dispersion, chelation, sequestering and suspending of the [11] X. Shi, G. Tal, N.P. Hankins, V. Gitis, Fouling and cleaning of ultraltration mem-
branes: a review, J. Water Process Eng. 1 (2014) 121138.
fouled layer that lead to reducing fouling and increasing permeation [12] W.S. Ang, S. Lee, M. Elimelech, Chemical and physical aspects of cleaning of or-
ux [50]. Sodium bicarbonate is a pure, natural product that is also a ganic-fouled reverse osmosis membranes, J. Membr. Sci. 272 (2006) 198210.
food; it is non-toxic, unlike many other chemical cleaners. It acts as a [13] J. Saqib, I.H. Aljundi, Membrane fouling and modication using surface treatment
and layer-by-layer assembly of polyelectrolytes: state-of-the-art review, J. Water
cleaning agent because it is a mild alkali and can cause oil and grease to
Process Eng. 11 (2016) 6887.
dissolve easily in water for eective removal. When it is not fully dis- [14] O.O. Ogunbiyi, N.J. Miles, N. Hilal, The eects of performance and cleaning cycles
solved, sodium bicarbonate is mildly abrasive and can lift fouled layer of new tubular ceramic microltration membrane fouled with a model yeast sus-
pension, Desalination 220 (2008) 273289.
for easy removal as a gentle scouring powder. When vinegar is mixed
[15] P. Blanpain-Avet, J. Migdal, T. Bnzech, Chemical cleaning of a tubular ceramic
with bicarbonate of soda, they are reacting with each other because of microltration membrane fouled with a whey protein concentrate suspen-
an acid-base reaction. Bicarbonate of soda as bicarbonate (NaHCO3) sioncharacterization of hydraulic and chemical cleanliness, J. Membr. Sci. 337
and vinegar as acetic acid (HCH3COO) were used. One of the products (2009) 153174.
[16] N. Yin, Z. Zhong, W. Xing, Ceramic membrane fouling and cleaning in ultraltration
of Bicarbonate of soda and vinegar reaction is carbon dioxide. This of desulfurization wastewater, Desalination 319 (2013) 9298.
reaction is lead to detachment of fouling layer on membrane surface [17] A. Ahmad, N.M. Yasin, C. Derek, J. Lim, Chemical cleaning of a cross-ow micro-
and pores. Carbon dioxide can help the mechanical cleaning of mem- ltration membrane fouled by microalgal biomass, J. Taiwan Inst. Chem. Eng. 45
(2014) 233241.
brane in which this gas is forced from the permeate side through the [18] Y.C. Woo, J.J. Lee, L.D. Tijing, H.K. Shon, M. Yao, H.-S. Kim, Characteristics of
membrane surface and out through them. membrane fouling by consecutive chemical cleaning in pressurized ultraltration as
pre-treatment of seawater desalination, Desalination 369 (2015) 5161.
[19] P. Chen, Z. Zhong, F. Liu, W. Xing, Cleaning ceramic membranes used in treating
4. Conclusion desizing wastewater with a complex-surfactant SDBS-assisted method, Desalination
365 (2015) 2535.
[20] S.H. Silalahi, T. Leiknes, Cleaning strategies in ceramic microltration membranes
In this research, chemical cleaning of home-made mullite ceramic fouled by oil and particulate matter in produced water, Desalination 236 (2009)
membranes were prepared via extrusion method which were fouled 160169.
[21] A. Salahi, T. Mohammadi, M. Abbasi, F. Rekabdar, Chemical cleaning of ultra-
during oily wastewater treatment, was investigated by single, binary ltration membrane after treatment of oily wastewater, Iran. J. Chem. Eng. 7
and ternary solution of four types of chemical cleaning agents (Acid (2010).
(sulfuric acid (H2SO4)), surfactant (sodium dodecyl sulfate (SDS)), [22] M. Kazemimoghadam, T. Mohammadi, Chemical cleaning of ultraltration mem-
branes in the milk industry, Desalination 204 (2007) 213218.
chelating agent (ethylene diamine tetra acetic acid (EDTA)) and alkali
[23] O. Bakhtiari, M. Samei, H. Taghikarimi, T. Mohammadi, Preparation and char-
(sodium hydroxide (NaOH))) with concentration of 5 mM and 10 mM acterization of mullite tubular membranes, Desalin. Water Treat. 36 (2011)
under best operating conditions. After rst cleaning step, membranes 210218.
cleaned with vinegar and sodium bicarbonate solution as a novel [24] M. Arzani, H.R. Mahdavi, O. Bakhtiari, T. Mohammadi, Preparation of mullite

94
E. Garmsiri et al.
ceramic microlter membranes using Response surface methodology based on
central composite design, Ceram. Int. 42 (2016) 81558164.
[25] E. Yuliwati, A.F. Ismail, T. Matsuura, M.A. Kassim, M.S. Abdullah, Characterization
of surface-modied porous PVDF hollow bers for renery wastewater treatment
using microscopic observation, Desalination 283 (2011) 206213.
[26] X. Zhang, W.-Z. Lang, H.-P. Xu, X. Yan, Y.-J. Guo, The eects of hydroxyapatite
nano whiskers and its synergism with polyvinylpyrrolidone on poly (vinylidene
uoride) hollow ber ultraltration membranes, RSC Adv. 5 (2015) 2153221543.
[27] K. Suresh, G. Pugazhenthi, Development of ceramic membranes from low-cost clays
for the separation of oilwater emulsion, Desalin. Water Treat. 57 (2016)
19271939.
[28] B. Nandi, R. Uppaluri, M. Purkait, Treatment of oily waste water using low-cost
ceramic membrane: ux decline mechanism and economic feasibility, Sep. Sci.
Technol. 44 (2009) 28402869.
[29] K. Li, Ceramic Membranes for Separation and Reaction, John Wiley & Sons, 2007.
[30] J. Hermia, Constant pressure blocking ltration law application to powder-law non-
Newtonian uid, Trans. Inst. Chem. Eng. 60 (1982) 183187.
[31] M.C.V. Vela, S.. Blanco, J.L. Garca, E.B. Rodrguez, Analysis of membrane pore
blocking models applied to the ultraltration of PEG, Sep. Purif. Technol. 62 (2008)
489498.
[32] R. Field, D. Wu, J. Howell, B. Gupta, Critical ux concept for microltration fouling,
J. Membr. Sci. 100 (1995) 259272.
[33] S. De Barros, C. Andrade, E. Mendes, L. Peres, Study of fouling mechanism in
pineapple juice clarication by ultraltration, J. Membr. Sci. 215 (2003) 213224.
[34] K.-J. Hwang, T.-T. Lin, Eect of morphology of polymeric membrane on the per-
formance of cross-ow microltration, J. Membr. Sci. 199 (2002) 4152.
[35] W. Bowen, J. Calvo, A. Hernandez, Steps of membrane blocking in ux decline
during protein microltration, J. Membr. Sci. 101 (1995) 153165.
[36] M. Abbasi, A. Taheri, Eect of coagulant agents on oily wastewater treatment
performance using mullite ceramic MF membranes: experimental and modeling
studies, Chin. J. Chem. Eng. 21 (2013) 12511259.
[37] S.S. Madaeni, T. Mohamamdi, M.K. Moghadam, Chemical cleaning of reverse
95
Journal of Water Process Engineering 19 (2017) 8195
osmosis membranes, Desalination 134 (2001) 7782.
[38] L. Masse, J. Puig-Bargus, M. Mondor, L. Deschnes, G. Talbot, Eciency of EDTA,
SDS, and NaOH solutions to clean RO membranes processing swine wastewater,
Sep. Sci. Technol. 50 (2015) 25092517.
[39] Z. Wang, J. Ma, C.Y. Tang, K. Kimura, Q. Wang, X. Han, Membrane cleaning in
membrane bioreactors: a review, J. Membr. Sci. 468 (2014) 276307.
[40] Q. Li, M. Elimelech, Organic fouling and chemical cleaning of nanoltration
membranes: measurements and mechanisms, Environ. Sci. Technol. 38 (2004)
46834693.
[41] N. Porcelli, S. Judd, Chemical cleaning of potable water membranes: a review, Sep.
Purif. Technol. 71 (2010) 137143.
[42] L.J. Zeman, A.L. Zydney, Microltration and Ultraltration: Principles and
Applications, M. Dekker, 1996.
[43] G. Trgrdh, Membrane cleaning, Desalination 71 (1989) 325335.
[44] N. D'souza, A. Mawson, Membrane cleaning in the dairy industry: a review, Crit.
Rev. Food Sci. Nutr. 45 (2005) 125134.
[45] A.W.W. Association, Microltration and Ultraltration Membranes for Drinking
Water: M53 vol. 3, American Water Works Association, 2005.
[46] A. Al-Amoudi, R.W. Lovitt, Fouling strategies and the cleaning system of NF
membranes and factors aecting cleaning eciency, J. Membr. Sci. 303 (2007)
428.
[47] H. Lee, G. Amy, J. Cho, Y. Yoon, S.-H. Moon, I.S. Kim, Cleaning strategies for ux
recovery of an ultraltration membrane fouled by natural organic matter, Water
Res. 35 (2001) 33013308.
[48] M. Cox, Surfactants, Detergents and Cleaners: A Handbook for Formulators, (1994)
p. 43.
[49] N.G. Marriott, R.B. Gravani, Principles of Food Sanitation, Springer
Science & Business Media, 2006.
[50] C. Regula, E. Carretier, Y. Wyart, G. Gsan-Guiziou, A. Vincent, D. Boudot, et al.,
Chemical cleaning/disinfection and ageing of organic UF membranes: a review,
Water Res. 56 (2014) 325365.