Journal of Membrane Science 415416 (2012) 250259

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

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

Novel polyethersulfone nanocomposite membrane prepared by PANI/Fe3O4

nanoparticles with enhanced performance for Cu(II) removal from water
Parisa Daraei a, Sayed Siavash Madaeni a,n, Negin Ghaemi b, Ehsan Salehi a,c, Mohammad Ali Khadivi a,
Rostam Moradian d, Bandar Astinchap d
a

Razi University, Membrane Research Center, Department of Chemical Engineering, Daneshgah Road, Tagh Bostan, Kermanshah 67149, Iran
Department of Chemical Engineering, Kermanshah University of Technology, Kermanshah, Iran
c
Faculty of Engineering, Department of Chemical Engineering, Arak University, Arak, Iran
d
Nano Science and Technology Research Center, Physics Department, Razi University, Kermanshah, Iran
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 2 March 2012
Received in revised form
26 April 2012
Accepted 1 May 2012
Available online 18 May 2012

A novel mixed matrix polymeric membrane was prepared from polyethersulfone (PES) and
self-produced polyaniline/iron(II, III) oxide (PANI/Fe3O4) nanoparticles by phase inversion method.
The core-shell structured PANI/Fe3O4 nanoparticles were veried and characterized using X-ray
diffraction (XRD), transmission electron microscopy (TEM) and fourier transform infrared spectroscopy
(FTIR). Three different amounts of nanoparticles were introduced into the casting solutions to obtain
the optimum value. According to the performance test, the membrane with 0.1 wt% nanoparticles
indicated the highest Cu(II) ion removal but the lowest pure water ux. This is caused by nanoparticles
located in the supercial pores of the membrane during preparation i.e., surface pore blockage.
Morphological analysis including eld emission scanning electron microscopy (FESEM) and atomic
force microscopy (AFM) as well as membrane performance tests revealed that adsorption is the most
possible separation mechanism by the membranes. For better investigation of the adsorption
mechanism, several isotherm models such as Langmuir, Freundlich and Redlich-Peterson were tested.
Based on the isothermal results, the Redlich-Peterson model offered superior tness indicating
relatively complex adsorption mechanism. The reusability of the nanocomposite membrane was
conrmed for several sequential adsorption-desorption processes using EDTA as regenerator.
& 2012 Elsevier B.V. All rights reserved.

Keywords:
PES
Nanocomposite membrane
PANI/Fe3O4 nanoparticles
Cu(II) removal
Adsorption isotherms

1. Introduction
Copper ion, as a toxic contaminant of potable water resources
at unauthorized dosages (more than 2 mg/l), should be eliminated
because of its dangerous risks for human being such as headache,
depression and learning problems [15]. Heavy metals including
copper are also toxic for plants and can affect the root growth [4].
There are several processes such as precipitation, coagulation,
adsorption, ion exchange, electro-dialysis, electro-coagulation,
and membrane separation processes for removing metal ions
from efuents [3,68]. Membrane processes offer several advantages compared with other separation methods such as high
removal efciency, low energy consumption, high ow rate, small
footprint and ease of scale up [9,10].
Elimination of copper ion by amphoteric polybenzimidazole
nanoltration hollow ber membrane has been reported [6]. The
removal of copper from efuents by Nanomax50 nanoltration

Corresponding author. Tel.: 98 831 4274530; fax: 98 831 4274542.

E-mail address: smadaeni@yahoo.com (S.S. Madaeni).

0376-7388/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.memsci.2012.05.007

membrane resulted around 35%Cu(II) rejection at 4 bars [5]. Many

studies describe preparation of adsorptive membranes by blending the polymers such as cellulose acetate, acrylonitrile butadiene
styrene and polyvinyl alcohol with an adsorptive polymer like
chitosan to enhance membrane performance for heavy metal
elimination from water [1114].
Adsorptive removal of copper ions is carried out not only by
membrane separation processes but also with various granola
adsorbents like silica gel and chitosan beads [1517]. Guolin et al.
used cross-linked magnetic chitosan beads prepared with a
magnetic uid containing iron salts [15]. Another study has
reported the application of chitosan coated polyvinyl chloride
beads to eliminate copper and nickel ions from water [16].
Utilizing the nanoparticles in membrane preparation is applied
to improve process effectiveness. For example, metal oxide nanoparticles are widely used as an additive for optimizing ceramic
membrane performance [18]. Hosseini et al. showed that addition
of proper amount of magnetic iron- nickel oxide particles can
improve performance of polyvinyl chloride based heterogeneous
ion exchange membranes [19]. Inuence of iron and aluminum
oxide coating layer on ceramic membranes performance for the

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

elimination of natural organic matter has been examined [20].

Application of iron oxide as ller in polyvinyl alcohol nanocomposite pervaporation membrane was developed for dehydration of
organic solvents [21]. Compared with other metal oxides, conspicuous impact of iron oxide nanoparticles on membranes performance for arsenic removal is reported [22,23]. The observed effect
is attributed to the great afnity of iron oxides toward heavy
metals [2224].
In addition to metal oxides, some polymeric materials such as
polyaniline can remove toxic metal ions from water [17,2527]
due to the existence of nitrogen atom with a lone electron pair as
a reactive adsorption site. It has been revealed that saw dust can
efciently adsorb cadmium ions when it is coated with polyaniline [25]. Belaib et al. modied silica gel and some natural solid
materials by coating with polyaniline. They obtained signicant
enhancement in copper loading capacity as well as adsorption
kinetics, using this adsorbent [17]. A polyaniline/inorganic cationexchanger nanocomposite has been fabricated to obtain a high
capacity ion-exchanger with increased ion exchange rate [26].
With regard to these studies, polyaniline can be applied as a
modier to achieve higher heavy metal uptake onto adsorptive
nanoparticles. The properties of PANI blended membranes and
PANI modied nanoparticles in nanocomposite membrane are
crucial. These modied membranes represent signicant properties. For example, a novel mixed matrix pervaporation membrane
was successfully prepared by polyvinylalcohol and polyaniline
treated TiO2 nanoparticles [28]. This modication diminished the
ux and improved the selectivity in separation of water-isopropanol mixture. Furthermore, polyaniline nanobers were widely
used as a pore former and additive in improving protein retention
and antifouling properties of ultraltration membrane [2932].
In this work, iron oxide/polyaniline nanoparticles (NPs) as a
core-shell structured adsorbent was prepared and utilized in PES
matrix to obtain a new nanocomposite membrane with enhanced
afnity for copper ions. The performance of the prepared membranes was tested for removing low concentrations of Cu(II) ions
from water. The mechanism of copper ion elimination by the
membranes was investigated by application of adsorption
isotherms. The reusability of the membranes was examined using
EDTA as eluting agent. FESEM, AFM, TEM, XRD and FTIR were
applied for characterization of the prepared membranes and
nanoparticles.

2. Experimental
2.1. Materials
The chemicals used in the current study are presented in
Table 1. All reagents were used without further purication except
for aniline which was double-distilled to obtain pure aniline
monomers.

2.2. Preparation and characterization of PANI/Fe3O4 nanoparticles

Briey, FeCl3  6H2O and FeCl2  4H2O were dissolved in deionized water with 3:1 M ratio and stirred. The dissolved oxygen in
the solution was removed by means of a vacuum pump. Then, the
ammonium hydroxide solution was added to the orange-colored
solution to adjust the pH value at 10 under vigorous stirring. The
solution immediately became black due to formation of Fe3O4
precipitates. After that, the product (Fe3O4 nanoparticles) was
thoroughly washed with deionized water and subsequently with
ethanol for several times and dried at room temperature. The
obtained product was used for coating with polyaniline.

251

Table 1
Used chemicals identication.
Chemical

Supplier

Purication

FeCl3  6H2O
FeCl2  4H2O
NH4OH
Aniline

Merck,
Merck,
Merck,
Merck,

HCl
(NH4)2S2O8
DMAc
PES, Ultrason E6020P,
MW 58000 g/mol
PVP, MW 25,000 g/mol
EDTA

Merck, Germany
Merck, Germany
Merck, Germany
BASF, Germany

Twice
distillation

Germany
Germany
Germany
Germany

Merck, Germany
BDH Chemicals Ltd.,
England
Merck, Germany

NaOH

Table 2
Prepared membranes composition (values in wt%).
Name

PES

PVP

DMAc

PANI/Fe3O4

Pristine PES
FA0.01
FA0.1
FA1

18
18
18
18

1
1
1
1

81.00
80.99
80.90
80.00

0.01
0.1
1

The PANI/Fe3O4 core-sell nanostructure was prepared by

in situ chemical oxidative polymerization of aniline in the
presence of Fe3O4 particles. In this method, certain amount of
Fe3O4 nanoparticles (mass ratio of Fe3O4 to aniline monomer was
1:2) were added to 50 ml of 0.1 M HCl solution containing 3 ml of
distilled aniline monomer and ultrasonicated for 15 min. The
mixture was mechanically stirred for efcient dispersion. The
reaction initiator, ammonium persulfate (APS), (6 g of APS in
50 ml of 0.1 M HCl solution) was slowly added to the suspension
to initiate the polymerization of aniline under constant stirring in
an ice bath. After 60 min, the polymerization was completed and
the suspension color turned to dark green. The product was
obtained by ltering the solution. Afterward, it was washed with
deionized water, and dried in a vacuum oven for 48 h at 40 1C.
X-ray diffraction spectra (XPert PRO MPD apparatus with Cu(a)
radiation) was used for investigation of iron oxide type and grain
size. The structure and particle size of produced PANI/Fe3O4 was
monitored using transmission electron microscopy (TEM), Philips
CH 200, LaB6Cathode 160 kv. Finally, FTIR (ABB Bomem,
MB104, Canada) spectroscopy was applied for verication of
PANI formation on the iron oxide nanoparticles.
2.3. Preparation of PANI/Fe3O4/PES mixed matrix membrane
The conventional phase inversion method was used for fabrication of PES nanocomposite membrane. Three mixed matrix and
one pristine PES membranes were prepared. The composition of
casting solutions is detailed in Table 2. Distinct amount of PES and
polyvinylpyrrolidone (PVP), as pore former was dissolved in
N,N-dimethylacetamide (DMAc) under stirring at 400 rpm for
24 h. In the case of nanoparticles addition, it was impossible to
use magnetic stirrer because of magnetic properties of PANI/
Fe3O4 particles. Hence, nanoparticles rstly dispersed in 1 to 2 g
of total required DMAc with the aid of ultrasonication for 30 min
and then added to PES solution in DMAc. This mixture was
located on a magnet until noticeable diffusion of NPs suspension
to PES solution was observed. Afterward, the mixture was ultrasonicated for 45 min. The obtained homogeneous solution was

252

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

casted on glass plates using a home-made lm applicator with

150 mm thickness. Distilled water (DW) was used as non-solvent
for membrane preparation. The membranes were retained in DW
for 24 h and then kept between lter papers to dry for another
24 h. Membrane fabrication was carried out at room temperature.
2.4. Membrane characterization
2.4.1. Membrane performance
This part of experiments was implemented to investigate the
membrane potential for copper ion removal and adsorption
mechanism. First, the prepared membranes were tested in terms
of Cu(II) ion retention using 20 mg/l of aqueous Cu(NO3)2 solution
as feed. A dead-end cell was selected for the tests. Ion removal
was monitored by atomic absorption spectroscopy (ame atomic
absorption spectrophotometer, Shimadzu AA-680, Japan) of feed
and every 10 min permeated solution. Rejection (R) percent can
be calculated as:
R%

C f C p
 100
Cf

Cf and Cp are copper ion concentration (mg/l) in feed and

permeate, respectively. The efciency of nanocomposite membrane was also investigated by applying relatively low concentration feeds (5 and 10 mg/l). Heavy metals including Cu are toxic
even at low concentrations. Accordingly, concentrations lower
than 100 mg/l of copper ion were selected to shows the membrane capability for up-taking low amounts of Cu(II) ions. Each
experiment was repeated 3 times using the prepared membranes.
The next step of performance test was conducted for understanding the copper ion removal mechanism. Membranes with
the effective area of 47.85 cm2 and specied weight were placed
in the ltration cell. Feed solutions with various ion concentrations (10 to 90 mg/l) were fed into the cell from the lowest
concentration to the highest without washing between the steps.
The details of the applied procedure were presented in another
study from the current authors [33]. Briey, the ltration process
was continued for 150 min until permeate concentration and
water ux remained almost unchanged i.e., dynamic (macroscopic) equilibrium was obtained. At the end of ltration time,
permeate concentration was analyzed to obtain dynamic adsorption capacity of membranes for isothermal study using the
following equation:
Qe

C 1 C e V
Mm

2.4.3. Membrane porosity

In addition to examining the pure water ux, another test was
performed for determining the membrane water content. The
procedure was dipping the certain weight of dry membrane in
water for 24 h and then drying the surface of species by lter
paper and immediately weighing. After that, the membranes were
dried in an oven at 50 1C for 24 h and weighed again. The
difference between two weights (dry and wet) indicates the
water content (weight rise percent) of each membrane. To reduce
the errors, this test was repeated three times.
Water content %

W w W d
 100
Wd

where Ww and Wd are wet and dry membrane weights (g),

respectively.
Moreover, the obtained value can be used for determining
the average (bulk) porosity of the prepared membranes by the
following equation:
Porosity %

W w W d
 100
rf V m

2.4.2. Membrane reusability

The best performed membrane in copper removal was selected
for reusability test. 25 ml of 10 mM EDTA solution [14], with the
pH value adjusted at 10.5 was used as chelating agent. The
membrane was immersed in EDTA solution and stirred for 1 h.
The membrane was washed with plenty of DW and placed in
dead-end cell to repeat the Cu(II) removal test using 20 mg/l
Cu(NO3)2 solution as feed. This procedure was sequentially
performed for four times with duration of 120 min for each cycle
of test.

where rf and Vm are water density (g/cm3) and membrane pieces

volume (cm3), respectively.
2.4.4. Membrane morphology
Field emission scanning electron microscopy (FESEM, Hitachi,
S-4160, Japan) was used to analyze morphology of the membrane.
The surface roughness was characterized by atomic force microscopy (AFM, Nanosurfs Mobile S scanning probeoptical microscope, Switzerland). Analysis of AFM images was conducted by
Nanosurfs Mobile S software (version 1.8), and roughness parameters obtained from this software were compared for the
prepared membranes.

where Qe (mg/g) is the equilibrium adsorbed amount of copper

ions per membrane mass. C0 and Ce (mg/l) are initial and
equilibrium copper ion concentration in the solution, respectively. V is the collected permeate volume (l) and Mm is the
membrane dry mass (g).
pH of Cu(II) solutions was adjusted at 5 for preventing
precipitation of Cu(OH)2, which could occur at pHs greater than 6.
Pure water ux was examined for each membrane at 4.5 bars
trans-membrane (TM) pressure and calculated in kg/m2/h.

Fig. 1. XRD pattern of the pure magnetic nanoparticles (Fe3O4).

Fig. 2. TEM image of PANI/Fe3O4 core-shell structure.

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

253

Fig. 3. FTIR spectra of Fe3O4 and PANI/Fe3O4 nanoparticles.

FA0.01

FA0.1

FA1

Pristine PES

100
90
80
70
60
50
40
30
20
10
0

Water flux (kg/m2.h)

Cu (II) rejection (%)

Pristine PES

20

40

60
Time (min)

80

100

120

FA0.01

FA0.1

FA1

55
50
45
40
35
30
25
20
15
10
5
0
0

10

20

30

50
40
Time (min)

60

70

80

90

Fig. 4. Prepared membranes Cu(II) ion rejection at 4.5 bars of TM pressure.

Fig. 5. Pure water ux of prepared membranes.
3. Results and discussion

100
3.1. Nanoparticles identication
The X-ray diffraction pattern of the powder sample is shown in Fig. 1. The
Bragg reection peaks at 2ys equal to 30.24, 35.64, 43.38, 53.84, 57.52, 63.02 and
74.53 are attributed to diffraction from the (2 2 0), (3 1 1), (4 0 0), (4 2 2), (5 1 1),
(4 4 0) and (5 3 3) planes of cubic inverse spinel Fe3O4, respectively, which are in
good agreement with the previously reported data (Magnetite: 011111). No
additional peak for other possible phases of iron oxide was observed (pure phase).
The average grain size (D) was calculated using DebyeScherrers equation:

Kl
bcosy

where D is the grain size of the crystal (nm), K is the Scherrer constant (0.89),

l and b are the X-ray wavelength and the peak width of half-maximum, and y is
the Bragg diffraction angle of the XRD spectra. The calculation revealed the
average grain size of about 14 nm.
Fig. 2 shows the TEM microphotograph of PANI/Fe3O4 nanoparticles with a
core-shell structure, which proved that iron oxide nanoparticles was successfully
entrapped in the polyaniline shell. The resulted core-shell PANI/Fe3O4 nanoparticles
retained their spherical structure with the average diameter of the magnetic core
about 1228 nm. The thickness of the polyaniline shell was determined around
8 nm (obtained from several microphotographs).
With respect to FTIR spectra of F3O4 and PANI/Fe3O4 depicted in Fig. 3, the
peaks at 1149, 1280 (CN), 1307 (QBB, QBQ), 1494 (NBN) and 1589 (NQN) cm
verify formation of PANI on iron oxide nanoparticles.

Copper ion rejection (%)

95
90
85
80
75
70

5 mg/l

65

10 mg/l

60
55
50
0

20

40

60
Time (min)

80

100

120

Fig. 6. Rejection of copper ions versus time at low concentrations (5 and 10 mg/l)
of feed solution for FA0.1.

3.2. Water ux and Cu(II) rejection

The results of the membrane performance are shown in Figs. 4 and 5. The
membrane containing 0.1 wt% of PANI/Fe3O4 nanoparticles (FA0.1) revealed the
highest copper removal around eight times greater than that of the pristine PES

254

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

Fig. 7. AFM images of (a) pristine PES, (b) FA0.01, (c) FA0.1 and (d) FA1.

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

membrane (Fig. 4). Furthermore, the membrane approximately removed 75% and
80% of copper ions from the feed with low concentrations of Cu(II) (5 and 10 mg/l,
respectively), after 2 h from the process commencement (see Fig. 6).
The pristine PES Membrane offered inferior copper rejection as well as
superior water ux. The highest ion rejection and the lowest water ux were
presented by FA0.1.
The AFM technique was used for justication of observed results. Fig. 7
indicates the AFM images of prepared membranes. AFM images present bulges
(bright points) and valleys (dark points) by scanning a sharp tip on membrane
surface. The more ups and downs result in higher roughness of membrane surface.
Surface pore size can control the intensity of these ups and downs. Comparing the
roughness parameters obtained from AFM images (Table 3) conrmed that the
surface roughness of the membrane decreased by increasing of the amount of NPs
in membrane matrix for FA0.01 and FA0.1. Decreasing the membrane surface
roughness by addition of nanoparticles might be due to the dispersion of NPs in
the polymer matrix and reduction of the surface pore size [34,35]. It means that
FA0.1 has a smoother surface due to superior dispersion of NPs in the membrane
structure. In FA0.01, the roughness was slightly declined (Sa descended from 12.4
to 11.6 nm by adding 0.01 wt% of NPs) because of low quantity of NPs in
membrane matrix which can provide lower continuity in membrane surface
relative to FA0.1. However, the roughness was increased in FA1 that it may be a
sign for more accumulation and less dispersion of NPs. The higher content of NPs
can produce the hunks because of decreasing the distance between nanoparticles
in casting solution. Besides, the viscosity of casting solution is increased by rising
NPs amount [35]. Increased viscosity of casting solution can slow down the phase
inversion process [35] and let the NPs to gather near each other due to delayed
demixing.
The efcient dispersion of NPs provides more accessible active sites for copper
ions to be adsorbed. Hence, the existence of proper amount of PANI/Fe3O4 NPs
accompanied with good dispersion can lead to higher elimination of Cu(II). By the
results of Cu(II) rejection (Fig. 4), FA0.1 showed the highest rejection and FA1 with
highest content of NPs revealed lower copper removal than that of both FA0.1 and
FA0.01. It seems that FA0.1 with smoother surface (Sa equal to 5.9 nm) and low
agglomerated NPs provides higher surface area of active NPs for adsorbing Cu(II)
ions relative to FA1. In spite of lower water ux of FA0.1 (due to decreasing surface
pores size as mentioned before), suitable dispersion (less accumulation) of NPs as
well as superior accessible active sites are the main causes for the higher rejection
of this membrane. The rejection of FA0.01 is less than that of FA0.1 indicating
effect of NPs quantity in membrane performance. It can be concluded that
adsorption mechanism is the dominant phenomenon involved in the removal of
copper ions where the amount and dispersion of NPs play a key role in prepared
membrane performance. Low rejection of FA1 may be attributed to the critical
agglomeration of NPs in the pores of the membrane surface. NPs agglomeration,
which increases in FA1 because of increasing NPs amount (about 10 and 100 folds
compared with FA0.1 and FA0.01, respectively), causes more defects and heterogeneity in the membrane (Fig. 8) Increased agglomeration of NPs at 1 wt% can also
decrease the effectiveness of the additive in membrane ion removal performance
by decreasing the available surface area of NPs. The depicted accumulated NPs in
membrane pores (Fig. 8) were observed only for FA1 in the scanning area of
membrane cross section (see Fig. 9). Moreover, an inappropriate dispersion and
formation of NPs hunk in membrane bulk can be observed in cross section
microphotographs of FA1 clearer than other membranes (Fig. 10).
Fig. 5 clearly indicates that adding the iron oxide NPs to the membrane matrix
leads to reduction of water ux compared to pristine PES membrane. While, the
water swelling tests indicated that water content of membranes enhanced by
increasing the NPs content of the membranes (Table 3). This can be described by
the mutual effect of iron oxide NPs on membrane structure (porosity increment)
against facial pore blockage by these NPs and reduction of pore size as concluded
from AFM results. It should be noticed that the iron oxide NPs tend to be
accumulated in the membrane surface and its supercial pores with regard to
this fact that the membrane surface is the rst place contacting with water
(nonsolvent) during phase inversion process [35]. The migration of NPs to the top
of the membrane causes to more changes in membrane surface pores [35].
Decreasing the water ux in the case of FA0.1 is because of surface pore blockage
as a result of nanoparticles dispersion in membrane surface and reduction of

Table. 3
Roughness parameters, water content and porosity of prepared membranes.
Membrane

PES
FA0.01
FA0.1
FA1
a

Roughness parameters (nm)a

Sa

Sq

Sz

12.4
11.6
5.9
10.3

17.2
15.5
7.6
13.1

286.4
187.3
51.2
136.7

Water content
(wt%)

Porosity
(v/v%)

285 7 2
293 7 2
3077 3
328 7 3

62
68
71
77

Data are the average value of three measurements.

255

Fig. 8. FESEM image of cross section of FA1.

surface pore size (as concluded from AFM results) due to tighter structure of
surface of nanocomposite membrane containing metal oxide NPs [28]. This
directly affects the water permeation capacity of the membrane. However, this
effect is diminished in FA1 membrane because of decrease of top-layer thickness
and resulted in higher water ux. Fig. 9 shows top-layer thicknesses for prepared
membranes, which were obtained from several measurements for each image.
Thinner top-layer allows superior water permeation through the membrane.
Considering low copper rejection of FA1 and the FESEM images (Fig. 7) it can be
proved that decrease of top-layer thickness also affects the ion rejection.
On the other hand, the FESEM images of cross section of membranes show the
changes in membrane sub-layer porosity and pore shape (see Fig. 10). According
to Fig. 10, sub-layer macro-voids have grown by increasing NPs amount in
membrane matrix. It can be seen that the highest quantity of NPs causes the
creation of largest macro-voids in the FA1 membrane (see Fig. 10d). Therefore, the
membrane with higher content of NPs can absorb and save much water into the
macro-voids leading to superior water content of the nanocomposite membrane
(but less water passage due to surface pore blockage). The data given in Table 3
verify that membrane with higher NPs content offers greater void capacity and
swells to a higher degree. This is attributed to the presence of NPs and their effect
on phase inversion kinetics i.e., the rate of membrane precipitation during
replacement of water and DMAc. Growing the macro-voids in sub-layer by
addition of metal oxides nanoparticles to membrane matrix has been proved
elsewhere [34,36].
Considering higher copper removal efciency of FA0.1 as well as acceptable
pure water ux (around 25 kg/m2/h) compared with other reported water ux
values (22 and 10 kg/m2/h) [34,35] for PES based nanocomposite membranes, the
optimum amount of PANI/Fe3O4 NPs can be selected as 0.1 wt% in casting solution.
3.3. Reusability and desorption efciency
Fig. 11 conrms that FA0.1 membrane can be reused even after 4 cycles of
ltration process with only 3% decrease in the rejection value. EDTA is a very
efcient agent for regeneration of adsorptive membranes [14]. Therefore, this
chemical was applied for reusing PANI/Fe3O4/PES nanocomposite membrane.
According to the pKa values (pKa1 1.66, pKa2 2.67, pKa3 6.16, pKa4 10.26) at
pH 10.5, EDTA is transformed to a hexadentate ligand and strongly reacts with
Cu(II) ions to form a stable octahedral complex. Fig. 12 shows how deprotonated
EDTA surrounds a copper ion. The nal copper ion rejection of FA0.1 membrane
was 82%. It should be mentioned that the membrane water ux had no noticeable
change during four cycles of test and was approximately equal to what presented
in Fig. 5 for FA0.1. This conrms signicant reusability and durability of prepared
nanocomposite membrane for copper removal from water.

3.4. Adsorption isotherms

As previously discussed, adsorption is the most likely mechanism offered by
the membranes to capture copper ions due to electrostatic interaction between
membrane and Cu(II) cations.
To verify this hypothesis, three widely used equilibrium adsorption
isotherms including Langmuir, Freundlich and Redlich-Peterson were applied.
Nonlinear regression method was used for tting the isotherm equations to the
equilibrium data.
The Langmuir isotherm is a traditional model postulated upon monolayer
adsorption coverage on a homogeneous surface without considering any interaction

256

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

Fig. 9. Comparison of top-layer thickness between (a) pristine PES, (b) FA0.01, (c) FA0.1 and (b) FA1.

Fig. 10. FESEM images of cross section of (a) pristine PES, (b) FA0.01, (c) FA0.1 and (d) FA1.

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

86
85

85

1.8
1.6

84

1.4

83

83

82

82

Qe (mg/g)

Cu (II) rejection (%)

257

82

81

1.2
1
0.8
0.6

80
1

y vs. x
langmuir

0.4

Run number

0.2

Fig. 11. Reusability of FA0.1 for four sequential runs.

0
0

10

15
Ce (mg/l)

20

25

30

Fig. 13. Langmuir isotherm tted to equilibrium data.

1.6
1.4

Qe (mg/g)

1.2
1
0.8
0.6
Fig. 12. Scheme of octahedral complex.

0.4

Qe a 

bC e
1 bC e

0.2
0

where Qe is the equilibrium adsorption (mg/g), a is the adsorption capacity (mg/g),

Ce is the equilibrium concentration in aqueous phase (mg/l) and b is the afnity
parameter.
The Freundlich model is used to describe multilayer adsorption on a heterogeneous surface considering interactions among adsorbates. Despite Langmuir
model, this isotherm cannot estimate the adsorption capacity for an adsorbent.
The model equation is expressed as [38,39]:

10

15
Ce (mg/l)

20

25

1.6

where k and m are relative adsorption constant and adsorption intensity

parameter, respectively. Adsorption is likely linear for m 1, chemisorption on a
relatively homogeneous surface if mo 1 and physical sorption on a rather
heterogeneous sorbent for m 41.
Langmuir and Freundlich models hypotheses are merged to obtain threeparameter Redlich-Peterson (R-P) isotherm with the potential to describe more
complex adsorption behavior. This model can be applied for describing adsorption
equilibrium on either homogeneous or heterogeneous sorbent surfaces. Interaction among adsorbing species is also taken into account in this isotherm. Equation
of the R-P model is represented as [3340]:
a  Ce
Qe
1 b  Ce m

Fig. 14. Freundlich isotherm tted to equilibrium data.

Figs. 1315 indicate Langmuir, Freundlich and R-P isotherms tted to the
equilibrium adsorption data, respectively. Based on the tting results (Table 4),
Freundlich isotherm conformity with the equilibrium data was somewhat better
than that of the Langmuir indicating multilayer adsorption coverage on the
membrane surface. Besides, being m-value higher than unity indicates that
physical adsorption occurred on a relatively heterogeneous surface. It is obvious
that Redlich-Peterson offers superior accordance with the equilibrium data on the
basis of the coefcient of determinations (R2  0.99). This suggests relatively

1.4
1.2
Qe (mg/g)

Q e kUC e m

y vs. x
Freundlich

among neighboring adsorption sites as well as adsorbing species [3337].

The Langmuir formula is given by:

1
0.8
0.6
0.4

y vs. x
R-P

0.2
0
0

10

15
Ce (mg/l)

20

25

Fig. 15. Redlich-Peterson isotherm tted to equilibrium data.

258

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

Table 4
Isotherms tting parameters and goodness of the ts.
Isotherm model

Parameters
a 1.632

Langmuir

b 1.525
k 0.9233

Freundlich

m 5.403
a 1.08
b 9.691

R-P

m 0.8662

Goodness of t
SSE 0.101
R-square 0.9659
Adjusted R-square 0.9617
RMSE 0.1123
SSE 0.06167
R-square 0.9792
Adjusted R-square 0.9766
RMSE 0.0878
SSE 0.03112
R-square 0.9895
Adjusted R-square 0.9865
RMSE 0.06668

complex mechanism of adsorption as well as the presence of lateral interaction

between adsorbates caused by electrostatic forces and probably related to metal
oxide core. On the other hand, existence of nitrogen atoms with a lone-pair
electron (related to PANI shell) along with Fe3O4 can make the NPs more efcient
in copper adsorption. It has been previously proved that the adsorption mechanism of copper on amino-functionalized Fe3O4 may be spontaneous, endothermic,
rapid and chemically natured [41].

4. Conclusion
Novel nanocomposite polymeric membrane was prepared by
PES and PANI/Fe3O4 nanoparticles. Characterization of nanoparticles using XRD and TEM proved that the prepared core-sell
structured particles had mean size of 1228 nm for iron(II,III)
oxide core and thickness of about 8 nm for the polyaniline shell.
The FTIR result of prepared nanoparticles veried formation of
PANI on iron oxide NPs. Copper ion removal was tested and the
results revealed that proper amount of nanoparticles in casting
solution and signicant dispersion of NPs are responsible for
efcient Cu(II) elimination. The membrane with 0.1 wt%NPs
(FA0.1) showed the highest ion rejection (around 85%) for
20 mg/l Cu(NO3)2 aqueous solution. The prepared membrane
was also able to remove 75%of Cu(II) ions from almost low
concentration feed solution (5 mg/l). Showing the minimum
agglomeration and high NPs dispersion in FESEM and AFM
images, the FA0.1 membrane composition was selected as the
optimum composition for mixed matrix membrane. The results
demonstrated that the adsorption mechanism was dominating
ion rejection of the membranes. By the isotherm conrmation
results, the most probable adsorption isotherm was
Redlich-Peterson isotherm expressing relatively complex adsorption
mechanism. The regeneration results conrmed that the prepared
nanocomposite membrane can offer excellent reusability as well as
durability in ltration process.

List of symbols

b
y
l

rf
a
AFM
APS
B
b
C0
Ce
Cf

Peak width of half-maximum

Bragg angle
X-ray wavelength (0.15 nm)
Fluid density (g/cm3)
Adsorption capacity (mg/g)
Atomic force microscopy
Ammonium persulfate
Benzenoid ring
Afnity parameter
Initial ion concentration (mg/l)
Equilibrium ion concentration (mg/l)
Ion concentration in feed (mg/l)

Cp
D
DMAc
DW
EDTA
FA0.01
FA0.1
FA1
FESEM
FTIR
K
k
m
Mm
NPs
PANI
PES
PVP
Q
Qe
R
RMSE
R-P
Sa
Sq
SSE
Sz
TEM
TM
V
Vm
Wd
Ww
XRD

Ion concentration in permeate (mg/l)

Average size of the crystal (nm)
N,N-dimethylacetamide
Distilled water
Ethylenediaminetetraacetic acid
Mixed matrix membrane with 0.01 wt%nanoparticles
Mixed matrix membrane with 0. 1 wt%nanoparticles
Mixed matrix membrane with 1 wt%nanoparticles
Field emission scanning electron microscopy
Fourier transform infrared spectroscopy
Scherrer constant (0.89)
Relative adsorption constant
Adsorption intensity parameter
Membrane dry mass (g)
Nanoparticles
Polyaniline
Polyethersulfone
Polyvinylpyrrolidone
Quinonoid ring
Equilibrium adsorbed amount of ions per membrane
mass (mg/g)
Ion rejection
Root mean square error
Redlich-Peterson
Average roughness (nm)
Root mean square of the Z data (nm)
Sum of squares due to error
Mean difference between highest peaks and lowest
valleys (nm)
Transmission electron microscopy
Trans membrane
Collected permeate volume (l)
Membrane pieces volume (cm3)
Dry membrane weight (g)
Wet membrane weight (g)
X-ray diffraction

References
[1] P.C. Eck, L. Wilson, Copper Toxicity, Eck Institute of Applied Nutrition and
Bioenergetics Ltd., 1989.
[2] N. Ratkevicius, J.A. Coreea, A. Moenne, Copper accumulation, synthesis of
ascorbate and activation of ascorbate peroxidase in Enteromorpha compressa
(L.) Grev. (Chlorophyta) from heavy metal-enriched environments in northern Chile, Plant Cell Environ. 26 (2003) 15991608.
[3] S. Ona, E. Toorisaka, M. Hirata, T. Hano, Adsorption and toxicity of heavy
metals on activated sludge, Sci. Asia 36 (2010) 204209.
[4] S.M. Reichman, The Responses of Plants to Metal Toxicity: a review focusing
on Copper, Manganese and Zinc, Australian Minerals & Energy Environment
Foundation, Melbourne, 2002.
[5] T. Chaabane, S. Taha, M. Taleb Ahmed, R. Maachi, G. Dorange, Removal of
copper from industrial efuent using a spiral wound modulelm theory
and hydrodynamic approach, Desalination 200 (2006) 403405.
[6] J. Lv, K.Y. Wang, T. Chung, Investigation of amphoteric polybenzimidazole
(PBI) Nanoltration hollow ber membrane for both cation and anions
removal, J. Membr. Sci. 310 (2008) 557566.
[7] Z. Ren, W. Zhang, H. Meng, J. Liu, S. Wang, Extraction separation of Cu(II) and
Co(II) from sulfuric solutions by hollow ber renewal liquid membrane,
J. Membr. Sci. 365 (2010) 260268.
[8] Y. Ku, S. Chen, W. Wang, Effect of solution composition on the removal of
copper ions by nanoltration, Sep. Purif. Technol. 43 (2005) 135142.
[9] A. Ghaee, M. Shariaty-Niassar, J. Barzin, T. Matsuura, Effects of chitosan
membrane morphology on copper ion adsorption, Chem. Eng. J. 165 (2010)
4655.
[10] R.S. Vieira, E. Guibal, E.A. Silva, M.M. Beppu, Adsorption and desorption of
binary mixtures of copper and mercury ions on natural and crosslinked
chitosan membranes, Adsorption 13 (2007) 603611.
[11] Z. Cheng, X. Liu, M. Han, W. Ma, Adsorption kinetic character of copper ions
onto a modied chitosan transparent thin membrane from aqueous solution,
J. Hazard. Mater. 182 (2010) 408415.

P. Daraei et al. / Journal of Membrane Science 415416 (2012) 250259

[12] A.G. Boricha, Z.V.P. Murthy, Acrylonitrile butadiene styrene/chitosan blend

membranes: preparation, characterization and performance for the separation of heavy metals, J. Membr. Sci. 339 (2009) 239249.
[13] W. Han, C. Liu, R. Bai, A novel method to prepare high chitosan content blend
hollow ber membranes using a non-acidic dope solvent for highly enhanced
adsorptive performance, J. Membr. Sci. 302 (2007) 150159.
[14] C. Liu, R. Bai, Adsorptive removal of copper ions with highly porous chitosan/
cellulose acetate blend hollow ber membranes, J. Membr. Sci. 284 (2006)
313322.
[15] H. Guolin, Y. Chuo, Z. Kai, S. Jeffrey, Adsorptive removal of copper ions from
aqueous solution using cross-linked magnetic chitosan beads, Chin. J. Chem.
Eng. 17 (2009) 960966.
[16] S.R. Popuri, Y. Vijaya, V.M. Boddu, K. Abburi, Adsorptive removal of copper
and nickel ions from water using chitosan coated PVC beads, Bioresour.
Technol. 100 (2009) 194199.
[17] F. Belaib, A.H. Meniai, M. Bencheikh-Lehocine, A. Mansri, M. Morcellet,
M. Bacquet, B. Martel, A macroscopic study of the retention capacity of
copper by polyaniline coated onto silica gel and natural solid materials,
Desalination 166 (2004) 371377.
[18] L.Y. Ng, A.W. Mohammad, C.P. Leo, N. Hilal, Polymeric Membranes Incorporated with Metal/Metal Oxide Nanoparticles: a Comprehensive Review,
Desalination, In press, Corrected proof, Available on line 23 December 2010.
[19] S.M. Hosseini, S.S. Madaeni, A.R. Heidari, A. Amirimehr, Preparation and
characterization of ion-selective polyvinyl chloride based heterogeneous
cation exchange membrane modied by magnetic ironnickel oxide nanoparticles, Desalination 284 (2012) 191199.
[20] B.I. Harman, H. Koseoglu, N.O. Yigit, M. Beyhan, M. Kitis, The use of iron
oxide-coated ceramic membranes in removing natural organic matter and
phenol from waters, Desalination 261 (2010) 2733.
[21] M. Sairam, B.V.K. Naidu, S.K. Nataraj, B. Sreedhar, T.M. Aminabhavi, Poly(vinyl alcohol)-iron oxide nanocomposite membranes for pervaporation
dehydration of isopropanol, 1,4-dioxane and tetrahydrofuran, J. Membr. Sci.
283 (2006) 6573.
[22] P. Sabbatini, F. Yrazu, F. Rossi, G. Thern, A. Marajofsky, M.M. Fidalgo de
Cortalezzi, Fabrication and characterization of iron oxide ceramic membranes for arsenic removal, Water Res. 44 (2010) 57025712.
[23] H. Park, H. Choi, As(III) removal by hybrid reactive membrane process
combined with ozonation, Water Res. 45 (2011) 19331940.
[24] P. Sabbatini, F. Rossi, G. Thern, A. Marajofsky, Iron oxide adsorbers for arsenic
removal: a low cost treatment for rural areas and mobile applications,
Desalination 248 (2009) 184192.
[25] M.S. Mansour, M.E. Ossman, H.A. Farag, Removal of Cd(II) ion from waste
water by adsorption onto polyaniline coated on sawdust, Desalination 272
(2011) 301305.
[26] L. Ai, J. Jiang, Ultrasonic-assisted synthesis of polyaniline nanosticks, and
heavy metal uptake performance, Mater. Lett. 65 (2011) 12151217.
[27] A.A. Khan, L. Paquiza, Characterization and ion-exchange behavior of thermally stable nano-composite polyaniline zirconium titanium phosphate: its

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]
[39]

[40]

[41]

259

analytical application in separation of toxic metals, Desalination 265 (2011)

242254.
S.A. Patil, T.M. Aminabhavi, Novel dense poly(vinyl alcohol)TiO2 mixed
matrix membranes for pervaporation separation of waterisopropanol mixtures at 30 1C, J. Membr. Sci. 281 (2006) 95102.
Z. Fan, Z. Wang, N. Sun, J. Wang, S. Wang, Performance improvement of
polysulfone ultraltration membrane by blending with polyaniline nanobers, J. Membr. Sci. 320 (2008) 363371.
Z. Fan, Z. Wang, M. Duan, J. Wang, S. Wang, Preparation and characterization
of polyaniline/polysulfone nanocomposite ultraltration membrane,
J. Membr. Sci. 310 (2008) 402408.
S. Zhao, Z. Wang, J. Wang, S. Yang, S. Wang, PSf/PANI nanocomposite
membrane prepared by in situ blending of PSf and PANI/NMP, J. Membr.
Sci. 376 (2011) 8395.
S. Zhao, Z. Wang, X. Xin Wei, J. Tian, S. Wang, Yang, S. Wang, Comparison
study of the effect of PVP and PANI nanobers additives on membrane
formation mechanism, structure and performance, J. Membr. Sci. 385386
(2011) 110122.
S.S. Madaeni, E. Salehi, Adsorption of cations on nanoltration membrane:
separation mechanism, isotherm conrmation and thermodynamic analysis,
Chem. Eng. J. 150 (2009) 114121.
N. Ghaemi, S.S. Madaeni, A. Alizadeh, H. Rajabi, P. Daraei, Preparation,
characterization and performance of polyethersulfone/organically modied
montmorillonite nanocomposite membranes in removal of pesticides, J.
Membr. Sci. 382 (2011) 135147.
V. Vatanpour, S.S. Madaeni, R. Moradian, S. Zinadini, B. Astinchap, Fabrication
and characterization of novel antifouling nanoltration membrane prepared
from oxidized multiwalled carbon nanotube/polyethersulfone nanocomposite, J. Membr. Sci. 375 (2011) 284294.
V. Vatanpour,S.S. Madaeni, R. Moradian, S. Zinadini, B. Astinchap, Novel
antibifouling nanoltration polyethersulfone membrane fabricated from
embedding TiO2 coated multiwalled carbon nanotubes, Sep. Purif. Technol.,
In Press, Accepted Manuscript., Available online 16 February 2012.
A.J. Prosser, E.I. Franses, Adsorption and surface tension of ionic surfactants at
the air-water interface: review and evaluation of equilibrium models,
Colloids Surf., A 178 (2001) 140.
D.M. Ruthven, Principles of Adsorption and Adsorption Processes, Fifth edn.,
John Wiley & Sons, 1984.
F. Gimbert, N. Morin-Crini, F. Renault, Adsorption isotherm models for dye
removal by cationized starch-based material in a single component system:
error analysis, J. Hazard. Mater. 157 (2008) 3446.
E. Salehi, S.S. Madaeni, V. Vatanpour, Thermodynamic investigation and
mathematical modelling of ion-imprinted membrane adsorption, J. Membr.
Sci. 389 (2012) 334342.
H. Yong-Mei, C. Mana, H. Zhong-Bo, Effective removal of Cu(II) ions from
aqueous solution by amino-functionalized magnetic nanoparticles, J. Hazard.
Mater. 184 (2010) 392399.