Journal of Industrial and Engineering Chemistry 34 (2016) 198–203

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry

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

Phosphorous adsorption on synthesized magnetite in wastewater

Jeongyun Choi a, Jinwook Chung a, Wonhee Lee b, Jong-Oh Kim b,*
a
R&D Center, Samsung Engineering Co. Ltd., 415-10 Woncheon-Dong, Youngtong-Gu, Suwon 443-823, Gyeonggi-Do, Republic of Korea
b
Department of Civil and Environmental Engineering, Hanyang University, 222 Wangsimni-Ro, Seongdong-Gu, Seoul 133-791, Republic of Korea

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

Article history: The phosphorous adsorption on a synthesized magnetic iron oxide (MIO) was conducted using a batch
Received 18 August 2015 type experiment. A dry-form MIO was used in adsorption experiment to obtain useful information for
Received in revised form 14 October 2015 practical application. The Elovich model reasonably simulated the two-step adsorption process in the
Accepted 12 November 2015
kinetic experiments. Freundlich isotherm model was better than the Langmuir model for describing the
Available online 23 November 2015
phosphate isotherm on MIO. The maximum adsorption capacity of MIO was 15.2 mg P/g MIO which was
achieved at the pH less than 5. The adsorbed phosphorous was effectively detached from MIO within
Keywords:
30 min using 20 wt% NaOH solution.
Adsorption
ß 2015 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights
Desorption
Equilibrium isotherm reserved.
Magnetic iron oxide
Phosphorus recovery

Introduction
Phosphorus is an essential nutrient for the growth of
organisms in ecosystems and is the most important resource in
many fields, such as the manufacture of fertilizers, detergents, and
polishers. However, the increased discharge of phosphate into the
water system contributes to eutrophication, leading to environ-
mental problems in reservoirs and coastal areas [1,2]. Various
techniques have been successfully applied for phosphate removal,
such as chemical precipitation, adsorption, and biological
processes. However, biological processes are unreliable because
the quality of water has a significant impact on phosphate
removal. Chemical precipitation often consumes a large amount
of coagulant but requires sludge treatment and material disposal
[3]. Adsorption is an alternative and promising phosphate
removal method, especially for wastewater with a low phosphate
concentration [4,5], due to the production of less sludge, easy
operation, and high efficiency [6]. Therefore, the removal of
phosphate using various adsorbents, such as aluminum oxides
[7,8], iron oxides [9], mesoporous silicates [10], red mud [11],
alum sludge [12], apatite [13], and clinoptilolite [14], has received
increased attention.
* Corresponding author. Tel.: +82 2 2220 0325; fax: +82 2 2220 1945.
E-mail address: jk120@hanyang.ac.kr (J.-O. Kim).
On the other hand, despite its extensive use in various fields,
phosphorus is a non-renewable and non-interchangeable finite
resource. Therefore, the recovery and removal of phosphate should
be considered. The material used for removing phosphorus must
satisfy the following preconditions as an adsorbent: (1) high
adsorption rate, (2) easy to recycle, and (3) high potential for
practical application [15,16]. Magnetic iron oxide (MIO) was
selected here as an effective alternative for the removal and
recovery of phosphorus in solution because solid MIO can easily be
separated from a solution using magnets.
Many researches have been carried out to adsorb phosphorous
in solution using magnetic iron oxide [17]. The magnetic particles
used in the previous researches were mostly nano-sized particles
and were kept and used as suspended solid in the adsorption
experiment. However, there was a little study using micro-size or
dry-form magnetic solid. The relatively bigger size and dry-form
magnetic solids is more appropriate to the practical application
than the nano-sized suspended form because of cheap
manufacturing price and easy handling.
In this study, we investigated the basic characteristic of dry-
based MIO (20–100 mm) on adsorbing phosphorous from solution.
Kinetic and isotherm experiments were conducted to evaluate the
reaction kinetic behavior and the adsorption capacity of MIO. In
addition, the effect of operational factors (pH and temperature)
was examined to obtain the optimum condition of phosphorous
removal by MIO. Finally, the desorption experiment was con-
ducted to verify the reusability of MIO in practical application.

http://dx.doi.org/10.1016/j.jiec.2015.11.008
1226-086X/ß 2015 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
 J. Choi et al. / Journal of Industrial and Engineering Chemistry 34 (2016) 198–203 199

Materials and methods Desorption of phosphate

Preparation of MIO
The Fe3O4 nanoparticles were synthesized using the co-
precipitation method [17], mixing 2 L of 1 M Fe SO47H2O and 2 L
of 2 M FeCl36H2O. The solution was then purged with nitrogen
in order to prevent Fe2+ ion oxidation. After 30 min of purging,
1 L of 6 M NaOH was added drop-wise into the solutions noted
above, and purging was continued for an additional 10 min.
During the process, the solution was gradually changed to black
and was washed 15 times with deionized water until the pH
value of the solution became neutral. The resulting black
solution was dried at 70 8C for 24 h in an electric furnace
(SF-03, SciLab Co., Korea) and then cooled at room temperature.
The dried magnetic iron oxide was crushed using a mortar and
pestle.
Characterization of MIO
The crustal structure of MIO was characterized using X-ray
diffractometry (XRD, D8-Advance, Bruker-AXS Co., Germany)
after excitement with Cu Ka radiation at 45 kV and 100 mA. The
Fourier-transform infrared spectra (FTIR) were recorded on a
Nicolet-380 Fourier-transform infrared spectrometer using
the KBr method. The MIO particles were classified by size
using the sieving method. The dried MIO was transferred to the
nest of sieves arranged from top to bottom with decreasing
size in the following order: 100-, 75-, 53-, 32-, 20-mm, and pan.
The sieves were then shaken on a sieve shaker for approximately
three min. Each MIO particle fraction was weighed, and the
average of five samples was calculated for use in further
analysis.
A desorption experiment was carried out by 2 steps. In the first
step, MIO was reacted with phosphorous at pH 4 and room
temperature. The phosphorous concentration on MIO was
4.9 mg P/g MIO. In the second step, 1 g of MIO was added to a
50 mL conical tube containing 20 mL of NaOH solution (3, 5, 10,
and 20 wt%). The MIO and alkaline solution was mixed using a
multi-rotator at room temperature. Sampling was performed for 1,
5, 10, 15, 30, and 60 min.
Results and discussion
Physicochemical characteristics of MIO
The properties of the synthesized MIO were evaluated using
X-ray diffraction (XRD), scanning electron microscopy (SEM), and
particle size distribution. The XRD result in Fig. 1a showed that
the d-space values obtained for the synthesized magnetite were
2.53, 1.48, and 1.61, which were analogous to the reference
power diffraction file (JCPDS No. 86-1354). There was no
noticeable peak for any of the other ferric or ferrous oxide-
hydroxides except MIO. The d-space value of the MIO-adsorbing
phosphate was almost the same as that of the synthesized MIO,
which indicated that no structural disruption of MIO or

Adsorption experiment for phosphate using MIO

The phosphate solution was obtained by dissolving potassium

dihydrogen phosphate (KH2PO4) in deionized water. The adsorp-
tion process for the single-factor experiments was conducted as
follows: 1 g of MIO was added to a 50 mL conical tube containing
20 mL of phosphate solution (40 mg/L). The sealed tubes were then
placed in a rotator shaker (GTR-100, Green Tech Co., Korea) at room
temperature for 2 h. After the reaction reached equilibrium
(approximately 2 h), the supernatant was filtered through 0.45-
mm syringe filters. The phosphate concentration was analyzed
using a HACH DR2800 spectrophotometer (method 8048 with a
detection limit of 0.01 mg-P/L) with PhosVer 3 reagent powder
pillows.
To determine the maximum adsorption capacities of the MIO,
equilibrium batch tests were conducted using the aforemen-
tioned procedures with various initial concentrations (165, 325,
490, 650, 815, 980, 1140, 1305, 1,470, and 1510 mg-P/L) at an
initial pH fixed at 4.0. In addition, the adsorption kinetic
experiment was conducted as follows: 1 g MIO was added to a
50 mL phosphate solution with 50 mg-P/L at pH 4. At regular
time intervals (1, 5, 10, 15, 30, 60, and 120 min), liquid samples
were withdrawn, and the phosphate concentration was analyzed.
The additional experiments were performed at 10 and 40 8C in
order to examine the effect of temperature on phosphate removal
and at an initial phosphate concentration of 50 mg-P/L and initial
pH 4 with a MIO dose of 0.25–2.0 g in 30 mL of solution in order
to evaluate the effect of adsorbent dosage. To investigate the
effect of pH on adsorption, the change in initial pH was controlled
at 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0 using 1 M/L NaOH and HCl
solutions for 120 min at an initial phosphate concentration of
50 mg-P/L. Fig. 1. Magnetic iron oxide used in this study: (a) XRD image and (b) FTIR pattern.
200 J. Choi et al. / Journal of Industrial and Engineering Chemistry 34 (2016) 198–203

Table 1
Adsorption characteristics of MIO with different particle sizes.

Sample MIO synthesized in this study Daejung Chemical Co.

Particle size (mm) 20–32 32–53 53–75 75–100 1.7

Initial concentration (mg P/L) 200 200 200 200 200
MIO addition (g MIO/L) 33.3 33.3 33.3 33.3 33.3
Adsorbed phosphate in MIO (mg P/g MIO) 5.52 5.35 5.32 5.32 2.18
Removal efficiency (%) 89.1 89.0 88.6 88.5 36.35

formation of precipitates containing phosphorous occurred equilibrium (mg/g), k1 is the rate constant for the pseudo first-
during the adsorption reaction. In addition, the XRD results order (min1), k2 is the rate constant for the pseudo-second-order
indicated that the synthesized MIO was not easily oxidized to (g/mg min), a is the initial phosphate adsorption rate for the
ferric oxide-hydroxide under oxygen exposure. Elovich model (mg/g min), and b is the desorption constant during
Fig. 1b shows the results of FI-IR analysis of MIO before and any one of the experiments using the Elovich model (g/mg).
after phosphorous adsorption. The major peaks appeared at 1,632, Non-linear regression was performed using the least squares
1,070, 619, and 469 cm1, which corresponded to the bending method for each of the kinetic models, and the results are
vibration of H2O molecules, the stretching vibration of Fe–OH
groups, and Fe–OH vibration, respectively. A peak at 1119 cm1
was observed in the sample of MIO with phosphorous, which did
not appear in the sample before the reaction. This result was
evidence that the phosphorous was chemically bound to the
surface of MIO.
The size of MIO was one of the most important factors that
affected the adsorption capacity. The effect of size was initially
evaluated because the size of MIO synthesized in this study varied
from nanometers to micrometers. Four types of samples were
prepared, as shown in Table 1, and then the batch test was
conducted. The adsorption capacity was observed to be 5.5–
5.3 mg P/g MIO, and no measurable difference was not noted
between the samples. Commercial samples from Daejung Chemi-
cal Co. in Korea were also tested in order for a comparison of the
adsorption capability of the synthesized MIO. The particle size of
the commercial magnetite was 1.7 mm and the capacity was
2.2 mg P/g MIO.

Adsorption kinetic analysis

The kinetic experiments for phosphorous adsorption on MIO

were conducted at various initial phosphorous concentrations
(165, 490, 815, and 1140 mg P/L). Fig. 2 shows the adsorbed
phosphorous concentration over the reaction period in the batch
experiments. Approximately 80% adsorption was achieved during
the first five min in all of the cases, and then the phosphorous was
slowly adsorbed onto the MIO during the last reaction period. This
two-step adsorption process has been observed in previous studies
on phosphorous adsorption on different adsorbents including
magnetic iron oxide, iron-copper binary oxides, iron-manganese
binary oxides, and ferric sludge [17–21].
The kinetic analysis was conducted in order to determine the
proper kinetic model to describe the experimental data and to
evaluate the effect of initial phosphate concentration on adsorp-
tion kinetics [17]. The adsorption kinetic models applied in this
study were pseudo first-order, pseudo second-order, and Elovich
model, which are expressed in Eqs. (1)–(3), respectively, as
follows:
 
qt ¼ qe 1ek1 t (1)

k2 q2e t
qt ¼ (2)
1 þ k2 qe t

1 1
qt ¼ lnðabÞ þ lnt (3)
b b
where qt is the amount of phosphate adsorbed onto MIO at time t Fig. 2. Kinetic model analyses: (a) pseudo-first-order model, (b) pseudo-second-
order model, and (c) Elovich model (experimental condition: pH 4.0, initial P
(mg/g), qe is the adsorption capacity of the phosphate at
concentration 165 mg-P/L, MIO concentration 33.3 g/L).
 J. Choi et al. / Journal of Industrial and Engineering Chemistry 34 (2016) 198–203
Table 2
Kinetic model parameters obtained from model fitting to the experimental data.
Initial P conc. (mg P/L) Pseudo-first-order model Pseudo-second-order model
2 2
qt (mg/g) k1 (h 1
) r qt (mg/g)
165 3.17 1.52 0.97 3.27
490 9.52 1.21 0.96 9.93
815 12.9 1.15 0.97 13.5
1140 13.4 1.36 0.95 13.9
presented in Fig. 2 and Table 2. Fig. 2a and b clearly illustrate that
the pseudo-first-order and the pseudo-second-order kinetic
models were not appropriate for the description of phosphate
adsorption in the late period of the reaction. The simulated
phosphate concentration was smaller than the measured concen-
tration, especially in the late reaction period. However, the Elovich
model effectively described the experimental kinetic data, and the
calculated coefficient of determination for the Elovich model was
greater than 0.99 in all of the cases, indicating that the mechanism
of phosphate adsorption on magnetite might be chemisorption
(see Fig. 2c) [22].
It has been clearly observed in other studies that the pseudo-
second-order model was better than the pseudo-first-order model
for phosphate adsorption on iron-bearing minerals. In addition, the
pseudo-second-order model provided a better description than the
Elovich model when kinetic experiments were conducted at a low
level of phosphate adsorption (less than 10 mg of phosphate was
adsorbed per 1 g of adsorbent) [17,20]. However, other studies
have reported that the pseudo-second-order kinetic model was not
a proper model when the adsorption capacity of the phosphate at
equilibrium (qe) exceeded dozens of mg phosphorous per g
adsorbent [20,22].
The initial phosphate adsorption rate (a) obtained from non-
linear regressions using the Elovich model was non-linearly
increased with respect to initial phosphate concentration or
phosphorous to magnetite ratio. The increase in adsorption rate
became smaller with an increase in phosphate concentration. In
addition, the adsorption capacity of the pseudo-second-order
kinetic model at equilibrium showed an analogous pattern to the
initial phosphate adsorption rate of the Elovich model. This result
indicates that the proper ratio of phosphorous to magnetite is an
important factor in the performance of an effective phosphorous
removal process.
Adsorption isotherm of MIO
The batch experiment for the adsorption isotherm study was
conducted to describe the adsorption of phosphate on MIO under
conditions of pH 7 and 23 8C, and the results are presented in
Fig. 3. In addition, the isotherm models, including the Langmuir
and Freundlich models, were used to simulate the experimental
data of phosphorus adsorption. The isotherm models are presented
in Eqs. (4) and (5)
kL C e
Langmuir equation : qe ¼ qmax (4)
1 þ kL C e
1=n
Freundlich equation : qe ¼ K F Ce (5)
where qe is the amount of adsorbate per gram of adsorbent at
equilibrium (mg/g), qmax is the maximum adsorption capacity (mg/
g), KL is the Langmuir isotherm constant (L/mg), Ce is the adsorbate
concentration in solution at equilibrium, KF is the Freundlich
isotherm constant describing the adsorption density (mg/g), and n
is the adsorption intensity. The non-linear regression was
conducted using the least squares method in order to justify
201
Elovich model
k2 (g/mg min) r a (mg/g min) b (g/mg) r2
0.84 0.98 489.4 3.30 0.98
0.18 0.98 1153 1.08 0.99
0.13 0.99 1555 0.80 0.99
0.15 0.97 1730 0.77 0.99
which models were appropriate to describe the experimental data
and to calculate the isotherm constants of phosphorus adsorption
by MIO.
The results showed that the Freundlich isotherm model was
more appropriate for describing phosphorus adsorption on MIO
than was the Langmuir model. The determination coefficient (R2)
value of the Freundlich model was 0.98, which was comparatively
higher than that of the Langmuir isotherm model at 0.759. The
estimated maximum adsorption capacity of the MIO using the
Langmuir model (qmax) was 15.2 mg P/g MIO, and the adsorption
density of the Freundlich model (KF) was 4.27 L/mg P.
The obtained maximum adsorption capacity in this study was
analogous to that of other studies using iron-bearing oxides or slag
for phosphate removal. It has been reported that the maximum
capacity of magnetite or magnetic iron oxide was in the range of
2.6 to 27.15 mg/g MIO [17,23]. Iron-bearing binary oxide has been
shown to have a phosphate capacity of 10–40 mg/g solid
[19,20,22].
Effects of temperature and pH on phosphate adsorption by MIO
Fig. 4 shows the effects of temperature on phosphate adsorption
onto MIO. The batch kinetic experiments were carried out at
temperatures ranging from 283 K to 323 K. The rapid adsorption of
phosphate onto MIO was observed within five min, and then the
amount of phosphate adsorbed slowly increased during the last
55 min. The simulations using the pseudo-second-order kinetics
model were carried out, and the results are presented in Fig. 5. The
adsorption capacity of phosphate onto MIO at equilibrium (qe)
increased from 5.20 to 6.15 mg P/g MIO with an increase in
temperature. The endothermic nature of the adsorption process
was comparable to that in other studies focusing on phosphate
adsorption on iron oxide. In addition, this result indicates that the
phosphate was not physically but chemically adsorbed onto the
MIO, because physisorption is generally known to be an
exothermic process [17,19,22,23].
Fig. 3. Equilibrium isotherm model analyses with the Freundlich and Langmuir
models (experimental condition: pH 4.0, initial P concentration 165–1630 mg-P/L,
MIO concentration 33.3 g/L).
202 J. Choi et al. / Journal of Industrial and Engineering Chemistry 34 (2016) 198–203
Fig. 4. Adsorption capacities of phosphate under various temperature conditions
(experimental condition: pH 4.0, initial P concentration 165 mg-P/L, MIO
concentration 33.3 g/L, temperature 283–313 K).
Solution pH is one of the important factors in phosphorus
adsorption by iron oxides. Many researchers have investigated the
effects of pH on adsorption capacity and have reported that the
greater adsorption of phosphorus onto iron oxides was achieved at
lower pH. In this study, the phosphorus adsorption by MIO was
examined at various pH conditions (pH 3.5–pH 9), and the results
are shown in Fig. 5. The amount of adsorbed phosphate at pH 3.5 to
4.9 was relatively constant, and then the phosphorus removal
efficiency rapidly decreased with an increase in pH. The slope of
the adsorbed phosphate at pH in the range from 4.9 to 8.2 was
observed to be 0.56 mg P/g MIO/pH. The pH value at the inflection
point was 4.9, as shown in Fig. 5, which was lower than the
measured zero point charge (ZPC) of MIO (pH 5.8) in this study and
was located between the H3PO4 dissociation constant of pKa1 and
pKa2. Therefore, it can be presumed that the change in phosphorous
adsorption capacity near pH 4.9 was not associated with the
speciation of phosphorous but instead was related to the surface
charge of MIO. At pH less than 4.9, the phosphorous in the species
of H2PO4 was spontaneously bound to the positive charge of MIO.
On the other hand, the amount of adsorbed phosphate decreased at
pH values higher than 5.8 because the surface charge of MIO was
changed from positive to negative with an increase in the pH. The
similar trend in phosphorous adsorption dependence on pH has
been reported by researchers in previous studies [17,19,20],
Fig. 5. Adsorption capacity and removal efficiency of phosphate under various pH
conditions (experimental condition: pH 3.5–8, initial P concentration 165 mg-P/L,
MIO concentration 33.3 g/L).
Fig. 6. Desorption efficiency of phosphate-adsorbed MIO depending on
concentration of NaOH (experimental condition: pH 12.7–13.1, initial P
concentration on MIO 4.9 mg P/g MIO, NaOH concentration 3–20 wt%).
although with different pH inflection points. The pH values that
showed a rapid decrease in phosphorous adsorption capacity of
maghemite, Fe-Cu binary oxide, and Fe–Mn binary oxide were 6, 5,
and 7, respectively, and this result might have been due to
differences in the ZPC of each of the minerals.
Desorption of phosphate-adsorbed MIO
The concentration of phosphate adsorbed onto the MIO used for
the desorption experiments was 4.9 mg PO43-P/g. A NaOH
solution was selected as the desorption chemical, because several
studies have demonstrated the effectiveness of NaOH on desorbing
phosphate from iron oxides [18,19,24,25]. Four concentrations of
NaOH solution (3, 5, 10, and 20 wt%) were applied, and the results
are shown in Fig. 6. When a sufficient amount of NaOH was applied,
the maximum desorption efficiency at the given conditions was
achieved within 30 min. It was also shown that the desorption
efficiency was dependent on the concentration of NaOH, increasing
with an increase in NaOH concentration. More than 99% of the
phosphate was desorbed form the magnetite at 20% NaOH, which
indicated that 6 g of NaOH was needed to desorb 100% of
phosphate (4.9 mg P) from 1 g of MIO. The amount of chemical
used to desorb phosphate from MIO in this study was comparable
to that of other studies in which 5 g of NaOH was added to 1 g of
Fe–Cu binary oxide [19].
Conclusions
The characterization of synthesized MIO for recovering
phosphate in various conditions was carried out in this study.
The XRD and FT-IR results showed that the adsorbed phosphate did
not disturb the structure of MIO and formed a phosphate-based
complex on the surface of MIO. In the adsorption kinetic test, the
pseudo-second-order and Elovich models reasonably simulated
the phosphate adsorption on MIO. The kinetic model provided a
better description of the experimental data when MIO had smaller
adsorption capacity. The Freundlich isotherm model was better
than the Langmuir model at describing the phosphate isotherm on
MIO. The maximum adsorption capacity of MIO obtained from this
study was analogous to that of other magnetic iron oxides or iron-
bearing minerals. The adsorption capacity increased with an
increase in the temperature, which indicates that phosphate
adsorption on MIO is endothermic. In addition, the adsorption
capacity was also dependent on pH. The amount of adsorbed
 J. Choi et al. / Journal of Industrial and Engineering Chemistry 34 (2016) 198–203
phosphate was relatively constant in the pH range of 3–5 and
linearly decreased with an increase in pH from 5 to 8. In conclusion,
the MIO synthesized in this study can be practically applied for
phosphate recovery based on the complete desorption results.
Acknowledgements
This study was supported by Korea Ministry of Environment
(MOE) as Advanced Technology program for Environmental
Industry and was partially supported by National Research
Foundation of Korea (NRF) grant funded by the Korean Govern-
ment (NRF-2013R1A2A1A09007252).
References
[1] S.E. Rosenquist, W.C. Hession, M.J. Eick, D.H. Vaughan, Ecol. Eng. 36 (2010) 664.
[2] S. Benyoucef, M. Amrani, Desalination 275 (2011) 231.
[3] S.H. Lee, K.H. Yeon, H. Park, S.H. Lee, Y.M. Park, M. Iwamoto, Korean J. Chem. Eng.
25 (2008) 1040.
[4] Y. Bashan, L.E. de-Bashan, Water Res. 38 (2004) 4222.
[5] J.H. Park, D.I. Jung, Desalination 269 (2011) 104.
[6] X. Cheng, X. Huang, X. Wang, B. Zhao, A. Chen, D. Sun, J. Hazard. Mater. 169 (2009)
958.
203
[7] H.P. Grigoropoulou, D.A. Georgantas, J. Colloid Interf. Sci. 315 (2007) 70.
[8] S. Tanada, M. Kabayama, N. Kawasaki, T. Sakiyama, T. Nakamura, M. Araki, T.
Tamura, J. Colloid Interf. Sci. 257 (2003) 135.
[9] R. Stanforth, H.S. Zhao, Environ. Sci. Technol. 35 (2001) 4753.
[10] E.W. Shin, J.S. Han, M. Jang, S.H. Min, J.K. Park, R.M. Rowell, Environ. Sci. Technol.
38 (2004) 912.
[11] E. López, B. Soto, M. Arias, A. Nunez, D. Rubinos, M.T. Barral, Water Res. 32 (1998)
1314.
[12] E. Galarneau, R. Gehr, Water Res. 31 (1997) 328.
[13] M.L. Kandah, Sep. Purif. Technol. 35 (2004) 61.
[14] F.D. Tillman, S.L. Bartelt-Hunt, V.A. Craver, J.A. Smith, G.R. Alther, Environ. Eng. Sci.
22 (2005) 400.
[15] K. Urano, H. Tachikawa, Ind. Eng. Chem. Res. 30 (1991) 1893.
[16] M.S. Onyango, D. Kuchar, M. Kubota, H. Matsuda, Ind. Eng. Chem. Res. 46 (2007)
894.
[17] S.Y. Yoon, C.G. Lee, J.A. Park, J.H. Kim, S.B. Kim, S.H. Lee, J.W. Choi, Chem. Eng. J. 236
(2014) 341.
[18] Y.J. Tu, C.F. You, Chem. Eng. J. 251 (2014) 285.
[19] G. Li, S. Gao, G. Zhang, X. Zhang, Chem. Eng. J. 235 (2014) 124.
[20] G. Zhang, H. Liu, R. Liu, J. Qu, J. Colloid Interf. Sci. 335 (2009) 168.
[21] X. Song, Y. Pan, Q. Wu, Z. Cheng, W. Ma, Desalination 280 (2011) 383.
[22] Y.J. Tu, C.F. You, Chem. Eng. J. 251 (2014) 285.
[23] Y.J. Tu, C.F. You, C.K. Chang, M.H. Chen, J. Taiwan Inst. Chem. E 46 (2015) 148.
[24] I. de Vicente, A. Marino-Martos, L. Cruz-Pizarro, J. de Vicente, J. Hazard. Mater. 181
(2010) 375.
[25] A. Zach-Maor, R. Semiat, H. Shemer, J. Colloid Interf. Sci. 363 (2011) 608.